&EPA

United States
Environmental Protection
Agency

Six-Year Review 4 Technical Support
Document for Microbial Contaminant

Regulations


-------
Office of Water (4607M)
EPA 815-R-24-022
July 2024

www, epa. gov/ safewater


-------
Disclaimer

This document is not a regulation. It is not legally enforceable and does not confer legal rights or
impose legal obligations on any party, including EPA, states or the regulated community.

While EPA has made every effort to ensure the accuracy of any references to statutory or
regulatory requirements, the obligations of the interested stakeholders are determined by statutes,
regulations or other legally binding requirements, not this document. In the event of a conflict
between the information in this document and any statute or regulation, this document would not
be controlling.

i


-------
Table of Contents

1	Introduction	1-1

2	EPA's Protocol for the Six-Year Review 4	2-1

3	History of Microbial Regulations	3-1

3.1	Filter Backwash Recycling Rule	3-1

3.2	Long-Term 2 Enhanced Surface Water Treatment Rule	3-2

3.3	Ground Water Rule	3-2

3.4	Revised Total Coliform Rule	3-3

3.5	Aircraft Drinking Water Rule	3-4

3.6	Summary of the Microbial Rules	3-5

4	Health Effects	4-1

4.1	Summary of Health Effects Review Outcome and Information Evaluated - Revised
Total Coliform Rule and Ground Water Rule	4-1

4.2	Long-Term 2 Enhanced Surface Water Treatment Rule	4-5

4.2.1 Summary of Health Effects Review Outcome and Information Evaluated - Long
Term 2 Enhanced Surface Water Treatment Rule	4-5

4.3	Aircraft Drinking Water Rule	4-7

4.3.1 Summary of Health Effects Review Outcome and Information Evaluated	4-7

5	Analytical Methods	5-1

5.1	Revised Total Coliform Rule Monitoring Requirements	5-1

5.2	Ground Water Rule Monitoring Requirements	5-1

5.3	Methods for Measuring Disinfectant Residuals in Ground Water	5-2

5.4	Long-term 2 Enhanced Treatment rule Analytical Methods Approved	5-2

5.5	Aircraft Drinking Water Rule	5-2

6	Occurrence and Exposure	6-1

6.1 Data Sources for Microbial Occurrence Analyses	6-4

6.1.1	Six-Year Review 4 Information Collection Request Data	6-4

6.1.1.1	Description of Data Collected Under Six-Year Review 4 Information
Collection Request	6-4

6.1.1.2	Limitations of the Six-Year Review 4 Data	6-7

6.1.1.3	SYR4 "Reduced" Total Coliform / E. coli Dataset to Support Analysis of
Ground Water Rule and Revised Total Coliform Rule	6-7

6.1.1.4	Additional Six-Year Review 4 Information Collection Request Data Records
to Support Review of the Long Term 2 Enhanced Surface Water Treatment Rule	6-11

6.1.2	Aircraft Drinking Water Rule Data	6-11

ii


-------
6.1.3	Safe Drinking Water Information System Data	6-13

6.1.4	Other Data Sources	6-13

6.1.4.1 Six-Year Review 3 "Reduced" Total Coliform / E. coli Dataset for Analysis of
Ground Water Rule	6-13

6.2	Analytical Results of Samples Taken from the Distribution System	6-15

6.2.1	Total Coliform / E. coli Occurrence in Context of Ground Water Rule and Revised
Total Coliform Rule	6-15

6.2.2	Analytical Results of Level 1 and Level 2 Assessments Under Revised Total
Coliform Rule	6-23

6.3	Microbial Contaminants in Raw Water under Long-Term 2 Enhanced Surface Water
Treatment Rule	6-25

6.3.1 Occurrence of Cryptosporidium in Source Water	6-25

6.4	Analyses Involving Undisinfected Ground Water Systems	6-26

6.4.1	Approaches to Identify Undisinfected Ground Water Systems	6-26

6.4.2	Modeling Total Coliform Positivity Rates	6-29

6.4.2.1	Statistical Techniques	6-29

6.4.2.2	Markov Chain Monte Carlo Modeling Results	6-30

6.4.3	Analytical Results of Triggered Source Water Monitoring under Ground Water
Rule 6-39

6.5	Analyses of Aircraft Drinking Water Rule	6-41

6.5.1 Occurrence of Total Coliforms and E. coli in Aircraft Systems	6-44

7 Treatment	7-1

7.1	Long Term 2 Enhanced Surface Water Treatment Rule	7-1

7.1.1	Description of Long Term 2 Enhanced Surface Water Treatment Rule
Requirements	7-1

7.1.2	Advances/Improvements/Innovations to Long Term 2 Enhanced Surface Water
Treatment Rule Microbial Toolbox Requirements	7-4

7.1.2.1	Ultraviolet	7-9

7.1.2.2	Other New Information not included in existing Long Term 2 Enhanced
Surface Water Treatment Rule Toolbox	7-12

7.1.2.3	Turbo Coagulation	7-13

7.1.2.4	Powdered activated carbon	7-13

7.2	Ground Water Rule	7-13

7.2.1	Sanitary Surveys	7-13

7.2.2	Treatment Technique Requirements under Ground Water Rule	7-14

7.2.3	Corrective Actions	7-14

iii


-------
7.2.4	Analytical Results Reflecting Ground Water Rule Impacts from Treatment
Techniques Requirements	7-14

7.2.4.1	Undisinfected Ground Water System Treatment	7-15

7.2.4.2	Ultraviolet Virus Inactivation of Adenovirus	7-15

7.2.5	Compliance Challenges for Small Systems	7-16

7.2.6	CT Tables for Log Inactivation of Viruses	7-16

7.3	Revised Total Coliform Rule	7-17

7.3.1	Description of Level 1 and Level 2 Assessments	7-18

7.3.2	Description of Corrective Actions	7-19

7.3.3	Advances/Improvements/Innovations to Revised Total Coliform Rule Treatment
Techniques	7-20

7.4	Aircraft Drinking Water Rule	7-21

7.4.1	Corrective Actions	7-21

7.4.2	New Information Available since Aircraft Drinking Water Rule Promulgation. 7-21

7.4.2.1	Potential for Bacterial Growth and Temperature of Water	7-22

7.4.2.2	Disinfection and Flushing	7-23

7.4.2.3	Restrict Public Access and Notification to Passengers and Crew	7-24

7.4.2.4	Repeat Sampling	7-25

7.4.2.5	Other Operation & Maintenance Topics	7-25

7.5	Filter Backwash Recycling Rule	7-25

8	References	8-1

9	List of Appendices	9-1

Appendix A. Additional Analyses on the Six-Year Review 4 Microbial Data	A-l

Appendix B. Additional Analyses on the Aircraft Drinking Water Rule Data	B-l

Appendix C. Revised Total Coliform Rule Level 1 and Level 2 Assessment Characterization and
Data Quality Considerations	C-l

Appendix D. Revised Total Coliform Rule Corrective Actions and Assessment of Data
Quality	D-l

iv


-------
List of Exhibits

Exhibit 2-1. Process for Identifying NPDWRs that are Candidates for Revision	2-3

Exhibit 3-1. Timeline for Selected Activities Associated with Microbial Regulations for Drinking
Water	3-1

Exhibit 3-2. NPDWRs for Microbial Rules	3-5

Exhibit 4-1. Outbreaks per year of waterborne disease caused by Cryptosporidium in drinking

water (1992-2021)	4-7

Exhibit 6-1. SYR4 Data Summary of the Count of Records Removed via the Quality Assurance
Measures Applied to Microbial Rule Contaminants	6-6

Exhibit 6-2. Summary of the Count of Records Removed via the Quality Assurance Measures for
Six-Year Review 4 Reduced Total Coliform / E. coli Dataset	6-8

Exhibit 6-3. Excerpt of Data from Six-Year Review 4 Reduced Total Coliform / E. coli

Dataset	6-10

Exhibit 6-4. Summary of the Count of Records Removed via the Quality Assurance Measures for
Six-Year Review 3 Reduced Total Coliform / E. coli Dataset	6-14

Exhibit 6-5. GWR and RTCR Implementation Milestones in the Context of the Six-Year Review
3 and Six-Year Review 4 Information Collection Request Timeframes	6-16

Exhibit 6-6. Yearly Trend of Percent Total Coliform Positive Results for Routine Sampling at

All Public Water Systems (GW and SW Systems)	6-16

Exhibit 6-7. Changes in Percent of GW Systems with Disinfection ("Common Systems" with

90% Completeness)	6-18

Exhibit 6-8. Changes of Percent RTTC+ Rates among All Ground Water Systems ("Common

Systems" with 90% Completeness)	6-18

Exhibit 6-9. Changes of Percent RTEC+ Rates among All Ground Water Systems ("Common

Systems" with 90% Completeness)	6-19

Exhibit 6-10. Changes of Percent RTTC+ Rates among Disinfecting and Undisinfected Systems
("Common Systems" with 90% Completeness)	6-19

Exhibit 6-11. Changes of Percent RTEC+ Rates among Disinfecting and Undisinfected Systems
("Common Systems" with 90% Completeness)	6-20

Exhibit 6-12. Summary of Changes of Percent RTTC+ Rates by System Categories (All Public
Water Systems, Disinfecting Ground Water systems, Undisinfected Ground Water
Systems)	6-21

Exhibit 6-13. Summary of Changes of Percent RTEC+ Rates by System Categories (All Public
Water Systems, Disinfecting Ground Water Systems, Undisinfected Ground Water
Systems)	6-21

Exhibit 6-14. Total Coliform Positive Rates for Undisinfected Ground Water Systems (2012-

2019)	6-22

Exhibit 6-15. E. coli Positive Rates for Undisinfected Ground Water Systems (2012-2019)... 6-23

v


-------
Exhibit 6-16. Summary of Level 1 and Level 2 Assessment Processes	6-24

Exhibit 6-17. Total Coliform and E. coli Positivity Rate for Months Before and After Level 1 or
2 Assessment (90% Completeness Applied1)	6-25

Exhibit 6-18. Bin Classification for Filtered Systems	6-26

Exhibit 6-19. Process to Identify Community Water Systems and Non-Transient Non-

Community Water Systems that are Undisinfected Ground Water Systems (Using
Year 2019 as Example)	6-28

Exhibit 6-20. Process to Identify Transient Non-Community Water Systems that are

Undisinfected Ground Water Systems (Using Year 2019 as Example)	6-29

Exhibit 6-21. Markov Chain Monte Carlo Samples Predicting Total Coliform Detection Rates in
Small Undisinfected Transient Non-Community Water Systems (2011 Data, Six-
Year Review 3 Definition of Disinfecting)	6-32

Exhibit 6-22. Detection Rate Distribution Functions for Small Public Water Systems (serving 25-
100) based on Markov Chain Monte Carlo Sample Mean Parameter Values (2011
Data, Six-Year Review 3 Definition of Disinfecting)	6-33

Exhibit 6-23. List of Data Sources and Disinfection Definition for Markov Chain Monte Carlo

Model Runs	6-34

Exhibit 6-24. Maximum Likelihood Estimates of National Total Coliform Occurrence using Six-
Year Review 3 and Six-Year Review 4 Definitions for Small Undisinfected
Transient Non Community Water Systems	6-36

Exhibit 6-25. Maximum Likelihood Estimates of National Total Coliform Occurrence using for
Small Undisinfected Non-Transient Non-Community Water Systems	6-37

Exhibit 6-26 Count of Systems by Size and Type with Total Coliform Positive Rates >5% (2011
Data; SYR4 Definition of Disinfecting)	6-38

Exhibit 6-27. Count of Systems by Size and Type with Total Coliform Positive Rates >5% (2019
Data; SYR4 Definition of Disinfecting)	6-38

Exhibit 6-28. Six-Year Review 4 Information Collection Request - Summary of E. coli Results
in Undisinfected Ground Water Systems Collected as Triggered Source Water
Samples (2012-2019)	6-40

Exhibit 6-29. Count of Total Coliform Samples and Aircraft Systems by Size; 2012-2019	6-41

Exhibit 6-30. Count of Total Coliform Samples and Systems by Size, Aircraft Manufacturer and
Model; 2012-2019	6-42

Exhibit 6-31. Count of Aircraft Total Coliform Samples by Sample Type; 2012-2019	6-43

Exhibit 6-32. Count of Aircraft Total Coliform and is. coli Samples, and Total Coliform and is.

coli Positives by Year; 2012-2019 	6-44

Exhibit 6-33. Aircraft Total Coliform and E. coli Sample Count and Positivity Rate, by Size;

2012-2019	6-45

Exhibit 6-34. Aircraft Total Coliform and E. coli Sample Count and Positivity Rate, by Location;

2012-2019	6-46

vi


-------
Exhibit 6-35. Aircraft Total Coliform and E. coli Sample Counts and Positivity Rate for Follow-
up Samples, by Air Carrier; 2012-2019 	6-46

Exhibit 7-1. Cryptosporidium Treatment Credits for all Toolbox Options under Long Term 2

Enhanced Surface Water Treatment Rule [40 CFR 141.715(b)]	7-3

Exhibit 7-2. Potentially Relevant New Studies since 2015 for Existing Long Term 2 Enhanced

Surface Water Treatment Rule Microbial Toolbox Options	7-5

Exhibit 7-3. Requirements for Ultraviolet Dose to Achieve Log Inactivation (millijoules per

centimeter squared (mJ/cm2)	7-10

Exhibit 7-4. Ultraviolet Sensitivity of Challenge Microorganisms - Reported Delivered

Ultraviolet Dose to Achieve Log Inactivation	7-10

Exhibit 7-5. Examples of Potentially Relevant Low Pressure Lamp Ultraviolet Doses for Log

Inactivation of Cryptosporidium and Male-Specific-2 Bacteriophage	7-11

Exhibit 7-6. Examples of Potentially Relevant Medium Pressure Lamp Ultraviolet Doses for Log
Inactivation of Male-Specific-2 Bacteriophage	7-11

Exhibit 7-7. Examples of Potentially Relevant Light Emitting Diode Ultraviolet Doses for Log

Inactivation of Male-Specific-2 Bacteriophage	7-12

Exhibit 7-8. Revised Total Coliform Rule Treatment Technique Triggers for Level 1 and Level
2 Assessments	7-18

Exhibit 7-9. Aircraft Drinking Water Rule - Number of Systems Performing Corrective Actions
by Year - 2012 to 2019	7-23

vii


-------
Acronyms

ACRP

Airport Cooperative Research Program

ADWR

Aircraft Drinking Water Rule

AGI

Acute gastrointestinal illness

AIDS

Acquired immunodeficiency syndrome

ANSI

American National Standards Institute

AOC

Administrative Orders on Consent

ARCS

Aircraft Reporting and Compliance System

ASDWA

Association of State Drinking Water Administrators

AWWA

American Water Works Association

BF

Bank Filtration

CCR

Consumer Confidence Report

CDC

Centers for Disease Control and Prevention

CFR

Code of Federal Regulations

cfu/mL

Colony forming unit per milliliter

CT

Concentration x Time

CWS

Community Water System

DBP

Disinfection by-product

DC

District of Columbia

DCTS

Data Collection and Tracking System

DNA

Deoxyribonucleic acid

DWSRF

Drinking Water State Revolving Fund

EPA

Environmental Protection Agency

ETV

Environmental Technology Verification

FAA

Federal Aviation Administration

FBRR

Filter Backwash Recycling Rule

FDA

Food and Drug Administration

FLA

Free-living amoebae

FR

Federal Register

GLUMRB

Great Lakes- Upper Mississippi River Board

viii


-------
GW

Ground Water

GWR

Ground Water Rule

GWUDI

Ground Water Under the Direct Influence of Surface Water

HPC

Heterotrophic plate count

HUS

Hemolytic uremic syndrome

IBS

Irritable bowel syndrome

ICR

Information Collection Request

IESWTR

Interim Enhanced Surface Water Treatment Rule

LED

Light-emitting diode

LF-EMF

Low-frequency electromagnetic field

LRV

Log reduction values

LT1

Long Term 1 Enhanced Surface Water Treatment Rule

LT2

Long Term 2 Enhanced Surface Water Treatment Rule

MCL

Maximum contaminant level

MCLG

Maximum contaminant level goal

MCMC

Markov Chain Monte Carlo

MD

Maryland

MDBP

Microbial and Disinfection Byproducts

MF

Microfiltration

mg/L

Milligrams per liter

MLE

Maximum Likelihood Estimate

MRDL

Maximum residual disinfectant levels

MRDLG

Maximum residual disinfectant level goal

MS2

Male-Specific-2 Bacteriophage

MT

Montana

MUG

4-methylumbelliferyl-B-D-glucuronide

NCWS

Non-community water system

NDWAC

National Drinking Water Advisory Council

NF

Nanofiltration

nm

Nanometer

NORS

National Outbreak Reporting System


-------
NPDWR

National Primary Drinking Water Regulation

NTNCWS

Non-transient non-community water system

NTU

Nephelometric Turbidity Unit

PAC

Powdered Activated Carbon

PBS

Phosphate buffered saline

PCR

Polymerase chain reaction

PN

Public notification

PWSID

Public Water System Identification Number

QA

Quality Assurance

QC

Quality Control

QMRA

Quantitative microbial risk assessment

qPCR

Quantitative polymerase chain reaction

RO

Reverse Osmosis

RPEC

Repeat E. coli Samples

RPTC

Repeat Total Coliform Samples

RED

Reduction Equivalent Dose

RTCR

Revised Total Coliform Rule

RTEC

Routine E. coli Samples

RTTC

Routine Total Coliform Samples

SCADA

Supervisory control and data acquisition

SDWA

State Drinking Water Act

SDWIS

Safe Drinking Water Information System

SDWIS/FED

Federal Safe Drinking Water Information System database

SOP

Standard operating procedure

SW

Surface Water

SWAT

Surface Water Analytical Tool

SWTR

Surface Water Treatment Rules

SYR

Six-Year Review

TCR

Total Coliform Rule

TG

Triggered

TN

Tennessee


-------
TNCWS	Transient Non-Community Water System

TOC	Total Organic Carbon

TT	Treatment technique

TX	Texas

UCFWR	Uncovered finished water reservoir

UDF	Unidirectional Flushing

UK	United Kingdom

US	United States

UV	Ultraviolet

UVA	Ultraviolet A

UVB	Ultraviolet B

UVC	Ultraviolet C

UVT	Ultraviolet Transmittance

WBDOSS	Waterborne Disease and Outbreak Surveillance System

WCP	Watershed Control Program

WSV	Water service vehicles

xi


-------
1 Introduction

The Safe Drinking Water Act (SDWA) requires the United States Environmental Protection
Agency (EPA) to review each National Primary Drinking Water Regulation (NPDWR) at least
once every six years and revise them, if appropriate. The purpose of the review, called the Six-
Year Review (SYR), is to evaluate current information for regulated contaminants to determine
if there is new information on health effects, treatment technologies, analytical methods,
occurrence and exposure, implementation and/or other factors that provides a health or technical
basis to support a regulatory revision that will improve or strengthen public health protection.
EPA completed and published the results of its first Six-Year Review ("Six-Year Review 1"), on
July 18, 2003 (USEPA, 2003) and the second Six-Year Review ("Six-Year Review 2"), on
March 29, 2010 (USEPA, 2010a), after developing a systematic approach, or protocol, for the
review of NPDWRs. During Six-Year Review 1, EPA identified the Total Coliform Rule (TCR)
as a candidate for revision. Four additional NPDWRs (acrylamide, epichlorohydrin,
tetrachloroethylene and trichloroethylene) were identified as candidates for revision during the
Six-Year Review 2.

EPA completed and published the results of its third Six-Year Review ("Six-Year Review 3"),
on January 11, 2017. Under the Six-Year Review 3, EPA concluded that eight NPDWRs are
candidates for regulatory revision. These eight NPDWRs are included in the Stage 1 and the
Stage 2 Disinfectants and Disinfection Byproducts Rules, the Surface Water Treatment Rule
(SWTR), the Interim Enhanced Surface Water Treatment Rule (IESWTR) and the Long Term 1
Enhanced Surface Water Treatment Rule (LT1). The eight candidates are Chlorite,
Cryptosporidium (under the SWTR, IESWTR and LT1), Haloacetic Acids, Heterotrophic
Bacteria, Giardia lamblia, Legionella, Total Trihalomethanes, and viruses (under the SWTR). As
of 2024, EPA is conducting analyses to further evaluate these eight NPDWRs for potential
regulatory revisions; therefore these eight NPDWRs are not subject for review under SYR4.

Under the fourth (and current) Six-Year Review ("Six-Year Review 4"), EPA reviewed the
regulated chemical, radiological and microbiological contaminants included in previous reviews.
However, this is the first time EPA has conducted a comprehensive Six-Year Review of the
following microbial contaminant regulations:

•	Revised Total Coliform Rule (RTCR)

•	Aircraft Drinking Water Rule (ADWR)

This document provides a summary of available information and data relevant to determining
which, if any, of the microbial contaminant regulations are candidates for revision under this Six-
Year Review. The information cutoff date for Six-Year Review 4 was December 2021. That is,
information published during or before December 2021 was considered as part of the Six-Year
Review 4. The Agency recognizes that scientists and other stakeholders are continuing to
investigate microbial contaminants and publish information subsequent to this cutoff date. While
not considered as part of the Six-Year Review 4, the Agency anticipates providing consideration
for that additional information in future six-year reviews.

1-1


-------
Chapter 2 of this document provides an overview of the protocol that EPA used in this review.
Chapter 3 provides an overview of the specific regulations addressed in this support document,
along with historical information about their development. Available information and data
relevant to the Six-Year Review 4 are provided in Chapter 4 (health effects), Chapter 5
(analytical methods), Chapter 6 (occurrence and exposure) and Chapter 7 (treatment).

1-2


-------
2 EPA's Protocol for the Six-Year Review 4

This chapter provides an overview of the process the Agency used to review the microbial
National Primary Drinking Water Regulations (NPDWRs) discussed in the Six-Year Review 4
(SYR4). The protocol document, "EPA Protocol for the Fourth Review of Existing National
Primary Drinking Water Regulations," contains a detailed description of the process the Agency
used to review all the NPDWRs (USEPA, 2024a). The foundation of this protocol was developed
for the Six-Year Review 1 based on the recommendations of the National Drinking Water
Advisory Council (NDWAC, 2000). The process undertaken for SYR4 was very similar to the
process implemented during the prior rounds of the Six-Year Review.

The review elements that the Environmental Protection Agency (EPA) considered for each
NPDWR include the following: initial review, health effects, analytical feasibility, occurrence
and exposure, treatment feasibility, and other regulatory revisions. Risk balancing is also a
review element considered in the Six-Year Review process, however, was not applicable to the
NPDWRs reviewed for SYR4. Further information about these review elements are described in
the protocol document (USEPA, 2024a).

Exhibit 2-1 presents an overview of the Six-Year Review protocol and major categories of
review outcomes. The protocol is broken down into a series of questions about whether there is
new information for a contaminant that suggests it is appropriate to revise one or more of the
NPDWRs. The two major outcomes of the detailed review are:

(1)	the NPDWR is not appropriate for revision and no action is necessary at this time, or

(2)	the NPDWR is a candidate for revision.

Individual regulatory provisions of NPDWRs that are evaluated as part of the Six-Year Review
are: maximum contaminant level goals (MCLGs), maximum contaminant levels (MCLs),
maximum residual disinfectant level goals (MRDLGs), maximum residual disinfectant levels
(MRDLs), treatment techniques (TTs), and other treatment technologies and regulatory
requirements (e.g., monitoring). The MCL provisions of the protocol are not applicable for
evaluation of the microbial contaminants regulations which establish TT requirements in lieu of
MCLs. Because mostly all the microbial regulations use TT in lieu of an MCL, the TT branch of
the protocol is a tailored review of the microbial regulations used to guide the review of the
SYR4 microbial regulations. The MRDLG and MRDL provisions are only applicable for
evaluation of the Disinfectants and Disinfection Byproducts Rules (D/DBP) rules as part of the
Six-Year Review.

The Initial Review branch or element of the protocol identifies NPDWRs with recent or ongoing
actions and excludes them from the review process to prevent duplicative Agency efforts
(USEPA, 2024a). The cutoff date for the NPDWRs reviewed under the Six-Year Review 4 was
December 2021. Based on the Initial Review for microbial regulations, for the first time EPA
included the Aircraft Drinking Water Rule (ADWR), which was promulgated in 2009, and the
Revised Total Coliform Rule (RTCR) (the revision of the 1989 Total Coliform Rule (TCR)),
which was promulgated in 2013. Since most of the 1989 TCR requirements were replaced by the

2-1


-------
2013 RTCR, the 1989 TCR was excluded from the Six-Year Review 4. In addition, the Filter
Backwash Recycling Rule (FBRR), the Long-Term 2 Enhanced Surface Water Treatment Rule
(LT2) and the Ground Water Rule (GWR) were included as in previous reviews.

During the previous round of Six-Year Review (the third Six-Year Review, or SYR3), EPA
determined that eight NPDWRs were candidates for regulatory revision. The eight NPDWRs
were included in the Stage 1 and the Stage 2 Disinfectants and Disinfection Byproducts Rules,
the Surface Water Treatment Rule, the Interim Enhanced Surface Water Treatment Rule, and the
Long Term 1 Enhanced Surface Water Treatment Rule. EPA has initiated the process to decide
whether a rulemaking to revise the regulations should be initiated. Since these actions, initiated
under SYR3, are still underway, these rules were not reviewed for SYR4.

2-2


-------
Exhibit 2-1. Process for Identifying NPDWRs that are Candidates for Revision

Methods Branch
Identifies potential to revise
practical quantitation level
based on new monitoring
data or previous data
analysis.

Occurrence Branch

Identifies whether MCL

revisions present a
meaningful opportunity to
reduce health risks or costs
based on occurrence
estimates.

Starting point for
all NPDWRs*

f	A

Initial Review Branch

Identifies NPDWRs
undergoing EPA action or
assessment that should be
exd uded from review and
those that should be
reviewed further.
>	/

Health Effects and
MCLG Branch

Identifies NPDWRs with
potential for MCLG revision
(MCL Branch 1) and those
without (MCL Branch 2).
V	/

Treatment Branch

Identifies potential treatment
technologies to meet
alternative MCLs.

f	~\

MCL Branch 1 or 2

Identifies NPDWRs with
potential for MCL revision
and those without. NPDWRs
with TT are directed to the
Treatment Technique
Branch.

Risk-Balancing Branch

For MDBP NPDWRs only,
identifies potential risk-
balancing considerations
among MDESPs (regulated
and unregulated).

/" \
Implementation Branch

Identifies potential impacts
on monitoring/reporting
requirements, requiring
further MCL antfor TT
revisions to be considered.

NPDWR

candidate for
revision

Treatment Technique
Branch

Identifies NPDWRs with
potential forTT revision and
those without.

At any point, an outcome of 'no action at this time' can be made
based on individual branch exclusion criteria, meaning the
NPDWR is not a candidate for revision for this review cycle.

Treatment Technique
Analysis Branch

Identifies whether TT
revisions present a
meaningful opportunity to
. reduce health risks or costs ,
based on treatment
technologies.

2-3


-------
3 History of Microbial Regulations

This chapter provides a brief history of the microbial contaminant regulations reviewed under
Six-Year Review 4 (SYR4). The microbial contaminants regulations covered in SYR4 include:
the Filter Backwash Recycling Rule (FBRR), the Long-Term 2 Enhanced Surface Water
Treatment Rule (LT2), the Revised Total Coliform Rule (RTCR), the Ground Water Rule
(GWR), and the Aircraft Drinking Water Rule (ADWR).

A timeline of selected events in the statutory and regulatory history and regulatory review
processes is shown in Exhibit 3-1.

Exhibit 3-1. Timeline for Selected Activities Associated with Microbial Regulations

for Drinking Water

1989

Final SWTR and TCR

SDWA
Amendments

M-DBP Advisory
Committee Established

2002

Final LT1ESWTR

1986

SDWA Amendments

Proposed recommended MCLs
for turbidity, Giardia lamblia,
and viruses

1994

Proposed IESWTR

\

Negotiating Committee
Established

Final IESWTR

2000

Proposed GWR

2003

/ Six-Year
Review 1

\

2001

Final FBRR

\ 2000

GWR Notice of
Data Availability

TCR Distribution System
Advisory Committee
Established

Proposed ADWR

Six-Year Review 3

Final
LT2ESWTR

\ 2006

Final
GWR

Proposed RTCR

1987

Proposed TCR and SWTR

IESWTR Notice of
Data Availability

Proposed
LT1ESWTR and
FBRR

Proposed
LT2ESWTR

2013

Final RTCR
\ 2010

Six-Year Review 2
2009

Final ADWR

2023

Six-Year Review 4

1996

Final Information Collection Rule

3.1 Filter Backwash Recycling Rule

The purpose of the Filter Backwash Recycling Rule (FBRR), promulgated June 8, 2001 (66 FR
31086), is to further protect public health by requiring public water systems (PWSs), where
needed, to institute changes to the return of recycle flows to a plant's treatment process that may
otherwise compromise microbial control. The rule addresses a statutory requirement of the 1996
SDWA amendments to promulgate a regulation that governs the recycling of filter backwash
water within the treatment process of PWSs. It applies to all surface water and ground water
under the direct influence of surface water (GWUDI) systems using direct or conventional
filtration.

3-1


-------
The FBRR requires that recycled filter backwash water, sludge thickener supernatant, and liquids
from dewatering processes be returned through the processes of a system's conventional or direct
filtration system or at an alternate location approved by the state. This requirement is codified in
40 CFR 141.76(i).

3.2	Long-Term 2 Enhanced Surface Water Treatment Rule

The Long-Term 2 Enhanced Surface Water Treatment Rule (LT2), promulgated on January 5,
2006 (71 FR 654, USEPA, 2006a), requires 2- to 3-log inactivation of Cryptosporidium in
unfiltered systems and additional treatment for Cryptosporidium in filtered systems based on the
results of source water monitoring. The rule includes a screening procedure to reduce monitoring
costs for small systems. The rule also requires covering of all uncovered finished water
reservoirs (UCFWR), unless systems treat reservoir effluent to provide at least 99.99 percent (4-
log) inactivation or removal of viruses, 99.9 percent (3-log) inactivation or removal of Giardia
lamblia and 99 percent (2-log) inactivation or removal of Cryptosporidium.

For the purposes of the LT2, filtered water systems are classified in one of four treatment
categories (bins) based on their monitoring results. The majority of systems are classified in the
lowest treatment bin, which carries no additional treatment requirements. Systems classified in
higher treatment bins must provide 90 to 99.7 percent (1.0- to 2.5-log) of additional treatment for
Cryptosporidium. Systems select from a wide range of treatment technologies, process
optimization techniques, and management techniques from what is known as the "microbial
toolbox" to meet their additional treatment requirements. Any system that fails to achieve
treatment credit in any month that is at least equal to the level of treatment required under LT2 is
considered in violation of the treatment technique (TT) requirement. All unfiltered water systems
must provide at least 99 or 99.9 percent (2- or 3-log) inactivation of Cryptosporidium, depending
on the results of their monitoring. All unfiltered systems must report to the state the arithmetic
mean of all Cryptosporidium reported from the completion of the initial and second round of
source water monitoring within six months. Additionally, if the monthly Cryptosporidium
sampling frequency varies, systems must first calculate a monthly average for each month of
monitoring. Then, systems must use these monthly average concentrations, rather than individual
sample concentrations from the initial and secondary reports, in the calculation of the mean
Cryptosporidium level. Lastly, unfiltered systems must supply a summary of the source water
monitoring data used for the calculation to the State. Unfiltered systems that fail to comply with
the aforementioned requirements are considered to be in violation of the TT requirement of LT2.

3.3	Ground Water Rule

EPA promulgated the Ground Water Rule (GWR) on November 8, 2006 (71 FR 65573, USEPA,
2006b) to provide for increased protection against microbial pathogens, specifically viral and
bacterial pathogens, in PWSs that use ground water sources. EPA was particularly concerned
about ground water systems that are susceptible to fecal contamination because these systems
may be at risk of supplying water that contains harmful microbial pathogens. Viral pathogens
found in ground water systems may include enteric viruses such as echovirus, coxsackieviruses,
hepatitis A and E, rotavirus, and noroviruses. Enteric bacterial pathogens may include E. coli
(most E. coli is harmless but a few strains are pathogenic, including E. coli 0157:H7),

Salmonella species, Shigella species and Vibrio cholerae.

3-2


-------
The GWR established a risk-targeted approach to identify ground water systems susceptible to
fecal contamination and requires action to correct significant deficiencies and source water fecal
contamination in ground water systems (USEPA, 2006b). This risk-targeting strategy includes
the following:

•	Regular ground water system sanitary surveys

•	A program for identifying higher risk systems through RTCR monitoring and state
determinations

•	Ground water source monitoring to detect fecal contamination at certain ground water
systems that do not provide 4-log treatment of viruses or as part of state assessment of a
ground water source

•	Measures to protect public health:

o TT requirements to address significant deficiencies and fecal contamination in ground
water and

o In systems providing treatment, compliance monitoring to ensure that 4-log treatment
of viruses is maintained

TT requirements consist of implementation of one or more of the following corrective action
options: correct all significant deficiencies; provide an alternate source of water; eliminate the
source of contamination; or provide treatment that reliably achieves at least 99.99 percent (4-log)
treatment of viruses (using inactivation, removal, or a state-approved combination of 4-log virus
inactivation and removal) for each ground water source (USEPA, 2006b). In addition, ground
water systems must inform their customers of any fecal indicator-positive ground water source
samples.

There are approximately 45,000 undisinfected (or "non-disinfecting") ground water systems in
the U.S., judging by SYR4 ICR data for systems with total coliform records. Most serve small
permanent populations or larger transient populations.

3.4 Revised Total Coliform Rule

EPA published the RTCR in the Federal Register (FR) on February 13, 2013 (78 FR 10269,
USEPA, 2013) and minor corrections on February 26, 2014 (79 FR 10665, USEPA, 2014a).

The RTCR upholds the purpose of the 1989 Total Coliform Rule (TCR) to protect public health
by ensuring the integrity of the drinking water distribution system and monitoring for the
presence of microbial contamination. The RTCR, which replaced the TCR, is the only current
microbial drinking water regulation that applies to all PWSs. EPA anticipated greater public
health protection under the RTCR, because it required PWSs that are vulnerable to microbial
contamination to identify and fix problems, and it established criteria necessary for PWSs to
qualify for and stay on reduced monitoring, thereby providing incentives for improved water
system operation.

3-3


-------
All PWSs, except aircraft water systems which are subject to ADWR, are required to collect total
coliform samples to comply with the RTCR. If a sample is total coliform positive, it must be
further analyzed for E. coli. If any total coliform positive sample is also E. coli positive, then the
E. coli positive sample result must be reported to the state by the end of the day that the PWS is
notified. Total coliform positive and E. coli positive samples initiate a find-and-fix approach to
prevent fecal contamination and other microbial pathogens from entering the distribution system.
Additionally, if any routine sample is total coliform positive, repeat samples are required. PWSs
on a quarterly or annual monitoring schedule must take a minimum of three additional routine
samples (known as additional routine monitoring) the month following a total coliform positive
routine or repeat sample. Another provision of the RTCR is that reduced monitoring may be
available for PWSs using only ground water and serving 1,000 or fewer persons that meet certain
additional PWS criteria.

Key provisions of the RTCR included:

•	Setting a Maximum Contaminant Level Goal (MCLG) and Maximum Containment Level
(MCL) for E. coli for protection against potential fecal contamination

•	Setting a total coliform TT requirement

•	Requirements for monitoring total coliforms and E. coli according to a sample siting plan
and schedule specific to the PWS

•	Provisions allowing PWSs to transition to the RTCR using their existing TCR monitoring
frequency, including PWSs on reduced monitoring

•	Requirements for seasonal systems (i.e., non-community water systems (NCWSs) not
operated on a year-round basis that start up and shut down at the beginning and end of
each operating season) to monitor and certify the completion of a state-approved start-up
procedures

•	Requirements for assessments and corrective action when monitoring results show that
PWSs may be vulnerable to contamination. Assessments can be triggered by total
coliform positive samples, E. coli MCL violations, and performance failures; the
assessments are graded ("Level 1" and "Level 2") depending on the severity or frequency
of the problem. Assessment results must be reported, and sanitary defects discovered
during an assessment must be corrected, within 30 days of the assessment being triggered

•	Public notification (PN) requirements for violations

•	Specific language for community water systems (CWSs) to include in their Consumer
Confidence Reports (CCRs) when they must conduct an assessment or if they incur an E.
coli MCL violation.

3.5 Aircraft Drinking Water Rule

Drinking water safety on aircraft is jointly regulated by EPA, the Food and Drug Administration
(FDA), and the Federal Aviation Administration (FAA). EPA's responsibility is to regulate
systems that supply water to airports and onboard aircraft. Aircraft PWSs are considered
transient non-community water systems (TNCWSs) and are subject to NPDWRs that apply to
TNCWSs. In 2004, EPA found all aircraft water systems to be out of compliance with NPDWRs.

3-4


-------
Subsequently EPA tested 327 aircraft and found that 15 percent of them were positive for total
coliforms. Since the existing NPDWRs were designed for traditional stationary PWSs and not for
mobile aircraft water systems that are operationally very different, EPA determined that an
aircraft-specific rule would provide a clearer and more implementable regulatory framework for
aircraft water systems. The final Aircraft Drinking Water Rule (ADWR) was promulgated on
October 19, 2009 (74 FR 53590, USEPA, 2009). The ADWR establishes barriers of protection
from disease-causing organisms targeted to the air carrier industry.

ADWR combines coliform sampling, best management practices, corrective action, PN, operator
training, and reporting and recordkeeping to improve public health protection. Air carriers are
required to develop a coliform sampling plan covering each aircraft they own or operate. The
frequency of coliform monitoring is tied to the frequency of disinfection and flushing. Two
coliform samples are required per monitoring period. One water sample to be tested for total
coliforms must be taken from a lavatory, and one sample from a galley. Any total coliform
positive sample must be further analyzed for the presence of E. coli. A positive finding of E. coli
triggers PN, corrective action, and flushing. Also, routine disinfection and flushing are required
at least once per year.

The ADWR applies only to aircraft with onboard water systems that provide water for human
consumption through pipes and regularly serve an average of at least twenty-five individuals
daily, at least 60 days out of the year, and that board finished water for human consumption.
Human consumption includes water for drinking, hand washing, food preparation, and oral
hygiene. Aircraft water systems include the water service panel, the filler neck of the aircraft
finished water storage tank, and all finished water storage tanks, piping, treatment equipment,
and plumbing fixtures within the aircraft that supply water to passengers or crew.

3.6 Summary of the Microbial Rules

Exhibit 3-2 provides a summary of the NPDWRs for the microbial rules. For each contaminant
or indicator, the table lists the MCLG, whether the NPDWR involves an MCL or TT, and the
rule(s) where it is referenced. The final column indicates whether the NPDWR is being reviewed
in SYR4.

Exhibit 3-2. NPDWRs for Microbial Rules

Microorganism/Indicator

MCLG

MCL or
TT

Rule(s)

Reviewed in
SYR4?

Giardia lamblia

Zero

TT

SWTR

No

Viruses

Zero

TT

SWTR, GWR

Yes

Legionella

Zero

TT

SWTR

No

Total coliforms

Zero

TT

RTCR, ADWR

Yes

E. coli

Zero

MCL

RTCR, ADWR

Yes

Cryptosporidium

Zero

TT

IESWTR, FBRR, LT1, LT2

Yes

Heterotrophic bacteria (by the HPC method)

N/A

TT

SWTR

No

3-5


-------
Microorganism/Indicator

MCLG

MCLor
TT

Rule(s)

Reviewed in
SYR4?

Turbidity

N/A

TT

SWTR, IESWTR, LT1

No

3-6


-------
4 Health Effects

This chapter summarizes the results of EPA's review of information related to human health
risks from drinking water exposure to the microbial contaminants reviewed in the Fourth Six-
Year Review. The review examined human health risks from microbial contaminants regulated
under the Revised Total Coliform Rule (RTCR), the Ground Water Rule (GWR), the Long Term
2 Enhanced Surface Water Treatment Rule (LT2), and the Aircraft Drinking Water Rule
(ADWR).

EPA performed a systematic review of literature that was published no later than December
2021, however, EPA did also include a few recently published findings pertinent to this chapter.
EPA evaluated whether any new (and older relevant) health effects information would suggest
that it is appropriate to revise the Maximum Contaminant Level Goal (MCLG), the Maximum
Containment Level (MCL), or the treatment technique (TT) associated with the microbial
contaminant regulation. An MCLG is a health goal set at a level at which no known or
anticipated adverse health effects occur, allowing an adequate margin of safety. An MCL is the
maximum level of a contaminant allowed in public drinking water systems. MCLs are set as
close to MCLGs as feasible using the best available treatment technology, and taking cost into
consideration. MCLs are enforceable standards. When there is no reliable analytical method that
is economically and technically feasible to measure a contaminant at concentrations to indicate
there is not a health concern, EPA sets a TT requirement instead of an MCL. The TT is an
enforceable procedure or level of technological performance that public water systems (PWSs)
must follow to ensure control of a contaminant. EPA's review of existing TT requirements for
SYR4 microbial contaminants is discussed in the occurrence and treatment chapters.

EPA's review of human health risks from exposure to microbial contaminants in drinking water
encompassed endemic disease and outbreaks. The scope of the review varied by Rule: the RTCR
review focused on general trends in outbreaks and endemic disease caused by fecal pathogens
(versus opportunistic pathogens), the GWR focused on viruses, the LT2 review focused on
Cryptosporidium, and the ADWR review focused on pathogens known or suspected to be of
concern in aircraft drinking water.

4.1 Summary of Health Effects Review Outcome and Information Evaluated - Revised
Total Coliform Rule and Ground Water Rule

Reduced rates of endemic disease, notably acute gastrointestinal illness (AGI), was an
anticipated benefit of RTCR and GWR. A number of papers were published in and around 2006
attempting to calculate rates of endemic AGI (and/or total AGI) attributable to drinking water
exposure in the United States. Using data from published randomized trials of drinking water
(surface water) interventions, Colford et al. (2006) estimated AGI attributable to public drinking
water systems in the United States to be in the range of 4.26 to 11.69 million cases annually.
Messner et al. (2006) used data from epidemiological studies to estimate the incidence of AGI
from drinking water in the U.S. at 0.06 cases per person per year, which translates to 16.4 million
AGI cases per year or 8.5 percent of annual AGI cases from all sources. Both Colford et al.
(2006) and Messner et al. (2006) studies used Canadian data to estimate AGI cases. Calderon

4-1


-------
and Craun (2006) reviewed available data from community intervention studies that could be
used to help develop a national estimate of endemic AGI incidence, but did not perform new
calculations. Reynolds et al. (2008) estimated that community ground water systems in the U.S.
are responsible for 10.7 million infections and 5.4 million illnesses annually, that non-
community ground water systems are responsible for 2.2 million infections and 1.1 million
illnesses annually, and that surface water systems are responsible for 26.0 million infections and
13.0 million illnesses annually, for a grand total of 19.5 million illnesses per year attributed to
drinking water. These illnesses include but are not limited to AGI. Colford et al. (2009)
addressed ground water and is discussed in the SYR3 technical support document. See the SYR3
technical support document for more discussion on Colford et al.'s (2009) study.

Applicable to RTCR, Collier et al. (2021) conducted a systematic study using structured expert
judgment to estimate the likely collective U.S. disease burden attributable to over a dozen
waterborne illnesses (vibriosis, campylobacteriosis, cryptosporidiosis, giardiasis, Legionnaires'
disease, otitis externa, pneumonia, septicemia, salmonellosis, and shigellosis, and norovirus)
from infectious pathogens.

Collier et al. (2021) estimated the total disease burden of waterborne illnesses domestically
acquired is approximately 7.15 million cases annually, and responsible for an estimated 118,000
hospitalizations and 6,630 deaths. In Collier et al.'s analysis, waterborne disease is understood to
include gastrointestinal, respiratory, and systemic disease attributable to both drinking-water and
non-drinking water exposure. Of the estimated 7.15 million infectious waterborne illnesses in
2014 in the United States, drinking water exposure caused 40 percent of hospitalizations and 50
percent of deaths. From further evaluation of this study's cases, Gerdes et al. (2023) determined
that 1.13 million (95% credible interval 255,000-3.54 million) of these illnesses were attributable
to drinking water. Among the 17 waterborne infectious diseases included in the Collier et al.
(2021) study, those caused by opportunistic pathogens (i.e., Legionairres' disease, non-
tuberculous mycobacterial infections, Pseudomonas pneumonia, Pseudomonas septicemia, and
the two-thirds of otitis externa attributed by the authors to Pseudomonas) account for a large
share of the diseases' public health burden, accounting for 34 percent of domestically acquired
waterborne cases, 80 percent of associated hospitalizations, and 95 percent of associated deaths.
When the calculations are limited to cases associated with drinking water exposure, per Gerdes
et al. (2023), the diseases caused by opportunistic pathogens account for only 10 percent of
cases, but fully 89 percent of associated hospitalizations and 98 percent of associated deaths. The
structured expert judgement approach used in these studies is employed when primary data are
not available, and therefore is subject to limitations, such as expert bias. Only those waterborne
infectious diseases for which data were available to quantify associated health outcomes were
included in the studies.

Ashbolt (2015) reviewed trends for viral, bacterial, protozoan, and fungal threats in drinking
water, including the emergence of opportunistic pathogens, and suggested that emerging
Polymerase chain reaction (PCR) and genome-sequencing techniques will over time enhance our
ability to detect and quantify pathogens that are currently not susceptible (at all, or in the life
stage present in water) to monitoring using culture-based techniques.

4-2


-------
In some studies, waterborne pathogens such as adenovirus, enteroviruses, hepatitis A, norovirus,
rotavirus, Salmonella, Giardia, Cryptosporidium, and Shigella have been found in untreated
ground water samples (Borchardt et al. 2012; Wallender et al. 2014; Stokdyk et al. 2020).
Infections from these pathogens can cause mild to severe illnesses. Illnesses may include AGI
with diarrhea, abdominal discomfort, nausea, vomiting, conjunctivitis, aseptic meningitis, and
hand-foot-and-mouth disease. Other more severe illnesses include hemolytic uremic syndrome
(HUS) (kidney failure), hepatitis, and bloody diarrhea (WHO, 2004). Infections from some
waterborne pathogens (e.g., Campylobacter) may cause sequelae with long-term implications,
such as reactive arthritis, Guillain-Barre syndrome, and irritable bowel syndrome (IBS) (Keithlin
et al. 2014).

Borchardt et al. (2023) reports the results of a community intervention human health study
(randomized controlled trial) to measure the proportion of AGI caused by undisinfected ground
water provided by 14 community PWS systems. Ultraviolet (UV) disinfection was installed on
supply wells of intervention communities, and in control communities, residents continued to
drink undisinfected ground water. Intervention and control communities switched treatments by
moving UV disinfection units at midpoint of the study (crossover design). Study participants
completed health diaries weekly during four 12-week periods and water supply wells were
analyzed monthly for pathogenic enteric viruses. The researchers compared AGI incidence
between intervention and control communities within the same period of time. They observed
that with norovirus contaminated wells, AGI attributable risk from well water was 19% (95%
confidence interval of-4% to 36%) for children less than 5 years old and 15% (95% confidence
interval of -9% to 33%) for adults. They also observed when echovirus 11 contaminated wells
that UV disinfection slightly reduced AGI in adults. Researchers found highly variable estimates
of AGI attributable risks from drinking undisinfected ground water due to exposure of various
types and quantity of viruses in supply wells changing through the study. However, AGI
attributable risks appeared greatest during times when supply wells were contaminated with
specific AGI etiologic viruses.

In Wallender et al.'s (2014) analysis of the reported waterborne outbreaks from the Centers of
Disease Control's (CDC) waterborne disease outbreak surveillance system, the researchers found
that among the 172 outbreaks associated with untreated ground water sources where contributing
factor data were available, the leading contamination sources associated with outbreaks with
ground water sources included human sewage (n = 57, 33,1%), animal contamination (n = 16,
9.3%>), and contamination entering via the distribution system (n = 12, 7.0%). Improper design,
maintenance, or location of the water system or a nearby septic tank was a contributing factor in
many cases (n = 116, 67.4%). Other contributing factors included rapid pathogen transport
through hydrogeologic formations (e.g., karst limestone; n = 45, 26.2%) and preceding heavy
rainfall or flooding (n = 36, 20.9%). Similarly, the Mattioli et al. (2021) study of a waterborne
norovirus outbreak at an inadequately disinfected campground identified high discharge septic
pollution, high yield well water demand, and unfavorable hydrogeology as factors. The
Wallender et al. (2014) and Mattoli et al. (2021) findings stress the importance of identifying
vulnerabilities of undisinfected and/or inadequately disinfected PWSs through frequent
inspection and routine maintenance, as recommended by protective regulations such as GWR,
and the need for consideration of the local hydrogeology. GWR allows, but does not require,
states to perform hydrogeological sensitivity assessments for ground water systems to identify
those most susceptible to contamination.

4-3


-------
Campylobacter is an example of a fecal bacterial pathogen that can cause outbreaks in
undisinfected ground water systems under conditions favorable to growth. A survey conducted
by Pitkanen (2013) found that 28 waterborne Campylobacter outbreaks were reported in 11
countries between 1978 and 2010. Most occurred in small ground water systems without
adequate disinfection. The probable causes of the outbreaks were cross contamination and breaks
in water treatment due to heavy rainfall or contamination by sewage. Two recent case studies are
described by Gilpin et al. (2020) and Pedati et al. (2019). Gilpin et al. (2020) reported on a 2016
outbreak of campylobacteriosis caused after heavy rain contaminated an untreated drinking water
supply in New Zealand. This was the largest outbreak of Campylobacter ever reported with
approximately 7,570 cases of diarrheal illness and four deaths. According to Gilpin et (2020), the
probable cause was sheep feces that contaminated a stream adjacent to the drinking water source.
This outbreak resulted in a recommendation that in New Zealand, all drinking water supplies,
including ground water, should be treated and that residual disinfection should be required.

Pedati et al. (2019) reported on an investigation of 39 cases of campylobacteriosis in Nebraska in
2017. Analysis showed a significant association between illness and consumption of untreated
municipal tap water from wells (odds ratio = 7.84, 95% confidence interval = 1.69-36.36).
According to Pedati et al. (2019), Campylobacter jejuni was determined to be the cause of the
illness, with six confirmed cases (either a stool culture or PCR-positive result for
Campylobacter) and 33 probable cases (a laboratory-confirmed probable illness in a nonresident
who worked, dined, or shopped in the city). The source was determined to be wastewater runoff
from an adjacent animal feeding operation after an irrigation system malfunctioned. The
wastewater runoff collected in a road ditch adjacent to two wells that supplied tap water to the
city. After the wells were permanently removed from service, no additional cases of illness were
reported.

The Economic Analysis for the GWR relied on a static factor or multiplier to account for
secondary transmission (person-to-person transmittal of illness initially caused by contaminated
drinking water). Soller (2009) demonstrated the application of a dynamic model for secondary
transmission. The author reports that "depending on the assumptions employed, the predicted
number of additional illnesses due to secondary transmission could be greater than that predicted
by the GWR base analysis by approximately an order of magnitude or could be as low as
effectively zero."

Although fecal indicator bacteria are useful for detecting fecal contamination, indicator bacteria
do not necessarily correlate with the presence of human pathogens (NRC, 2004). Studies by Fout
et al. (2017) and Stokdyk et al. (2020) found that total coliforms (and other indicators like E.
coli, somatic phage, HF183, and KacteroidalesAike HumM2) tend to have high specificity,
meaning that absence of the indicator provides relatively strong assurance that ground water is
free of viral and other pathogens, but low sensitivity, meaning that presence of the indicator does
not necessarily predict presence of pathogens.

Fout et al. (2017) found that hydrogeology affects the strength of the association between the
presence of indicators and the presence of pathogenic viruses. Specifically, sensitivities and
positive predictive values were higher for indicators measured in wells in hydrogeologically
susceptible areas (karst, fractured bedrock, and gravel/cobble settings) than in other wells. As a
methodological point, the authors also note that it can be particularly challenging to assign wells
to the right category of hydrogeological susceptibility.

4-4


-------
Noncommunity public water systems, which are systems that provide water to locations outside
consumers' residences (e.g., schools, restaurants), are subject to less stringent monitoring
requirements than community systems. Burch et al. (2022) found that noncommunity wells had
higher infection risk than community wells. Burch et al. (2022) along with Mattioli et al. (2021)
study show that PWSs served by ground water remain susceptible to contamination by viruses
and other pathogens.

The studies described above indicate that there could be opportunity to further protect public
health from microbial contamination of untreated ground water. However, these studies are
limited in number and the prevalence of endemic disease from microbial contamination of
untreated ground water is not well characterized. More studies are needed to further understand
the magnitude of the issue.

4.2 Long-Term 2 Enhanced Surface Water Treatment Rule

The purpose of LT2 (71 FR 654, USEPA, 2006a) is to reduce illness linked to Cryptosporidium
and other pathogenic microorganisms in drinking water. Under the Interim Enhanced Surface
Water Treatment Rule (IESWTR) (63 FR 69477, USEPA, 1998) and LT1 (67 FR 1812, USEPA,
2002), EPA established an NPDWR for Cryptosporidium and set an MCLG of zero. These
regulations established TT requirements for Cryptosporidium removal/inactivation rather than an
MCL. LT2 supplements these regulations by establishing additional Cryptosporidium treatment
requirements at higher risk systems. LT2 also contains provisions to reduce risks from uncovered
finished water reservoirs (UCFWR) and provisions to ensure that systems maintain microbial
protection when they take steps to decrease the formation of disinfection byproducts that result
from chemical water treatment.

4.2.1 Summary of Health Effects Review Outcome and Information Evaluated - Long
Term 2 Enhanced Surface Water Treatment Rule

Cryptosporidium is a protozoan parasite that can be found in surface waters used as drinking
water sources by PWSs. Cryptosporidiosis, the illness caused by the ingestion of infectious
Cryptosporidium oocysts, is excreted in the feces of infected humans or animals, can cause seven
to 14 days of diarrhea, and possibly be accompanied by low-grade fever, nausea, and abdominal
cramps in individuals with healthy immune systems (CDC, 2017; Juranek, 1995). Though the
illness is typically self-limiting, there have been cases where it recurs after initial clearance. In
the 1993 cryptosporidiosis outbreak in Milwaukee, Wisconsin, 39 percent of those with
laboratory-confirmed cryptosporidiosis experienced recurrences, which lasted for an average of
two days (Mac Kenzie et al., 1994). Longer-term effects have also been recently documented.
After a foodborne outbreak of cryptosporidiosis in the United Kingdom, those with confirmed
cases reported symptoms such as fatigue, diarrhea, IBS, joint pain, and eye pain at six months to
12 months following illness (Stiff et al., 2017).

People with acquired immunodeficiency syndrome (AIDS) are more likely to experience atypical
presentations of cryptosporidiosis, such as infections of the bile duct (cholecystitis), stomach
(enteritis), pancreas, or respiratory system (Hunter and Nichols, 2002). Severe cases of
cryptosporidiosis have also been reported in cancer patients undergoing chemotherapy and in a
few cases in patients being immunosuppressed following organ transplantation (Hunter and

4-5


-------
Nichols, 2002). The elderly are also at increased risk (Naumova et al., 2003), particularly those
older than 85 years of age (Mor et al., 2009). Children may also be at increased risk. In the
developing world, cryptosporidiosis was found to be twice as likely to cause death in toddlers
one to two years old compared to other pathogenic diarrheal diseases (Kotloff et al., 2013).

Several species of Cryptosporidium can infect and cause illness in humans, including C. parvum,
C. hominis, C. canis, C.felis, and C. meleagridis and C. cuniculus (Chalmers et al., 2011;
Koehler et al., 2014). C. parvum and C. hominis are thought to be responsible for most cases of
infection in humans (Chappell et al., 2011). Audebert et al. (2020) researched virulence
differences in C. parvum isolates. Comparing phenotypic differences between four different C.
parvum isolates (IOWA, DID, TUMI, and CHR), Audebert et al. found that C. parvum DID,
TUMI, and CHR isolates had higher virulence than the IOWA isolates. The researchers
inoculated mice with the four different isolates obtained from fecal samples of naturally infected
animals or humans and found that mice inoculated with the three more virulent isolates exhibited
a higher mortality rate, more severe clinical manifestations, and earlier onset of neoplastic
lesions, and only these mice showed extra gastrointestinal lesions.

A study by Messner and Berger (2016) investigated six different Cryptosporidium infectivity
dose-response models using human challenge data from earlier studies. The authors found that
three models that allowed for variability in human susceptibility (fractional Poisson, exponential
with immunity, and beta Poisson) fit the data better than earlier models that explicitly accounted
for virulence differences among the Cryptosporidium isolates rather than human susceptibility
differences. The authors reported that the three human-focused models predicted significantly
higher risk of infection from low-dose exposures than earlier models had predicted: for example,
a 72 percent probability of infection from a single ingested oocyst as opposed to a previous
estimate of 4 percent probability of infection.

However, Schmidt and Chappell (2016) disagreed with Messner and Berger's assumption that all
Cryptosporidium isolates share a single dose-response relationship and are equally infectious.
Their criticism was that Messner and Berger chose to analyze only the studies where subjects
had no serological evidence of prior infection and therefore did not include data that could have
provided a more representative depiction of the general population. Ethical considerations
prevent the gathering of experimental data on infective doses in immunocompromised persons.
An experiment comparing healthy mice and immunosuppressed mice, however, found that
Cryptosporidium infectivity rates were no higher in immunosuppressed mice than in healthy
mice, though illness was more severe and sometimes fatal in the immunosuppressed group
(Miller et al., 2007).

Exhibit 4-1 shows annual waterborne cryptosporidiosis outbreaks recorded in CDC's National
Outbreak Reporting System (NORS) as attributable to drinking water exposure (1992-2021).
Some of these outbreaks have been associated with private water wells, and others have occurred
at public water systems. Since 2012, there have been four reported outbreaks of cryptosporidiosis
from public water systems. One of the four outbreaks involved both Cryptosporidium and shiga-
toxin-producing E. coli. The four outbreaks together resulted in a total of 201 recorded illnesses,
two hospitalizations, and no deaths. Note that additional outbreaks may have gone unreported by
NORS or may have been recorded as of uncertain etiology. In addition, since NORS is
specifically focused on outbreaks it does not capture rates of endemic disease.

4-6


-------
Exhibit 4-1. Outbreaks per year of waterborne disease caused by
Cryptosporidium in drinking water (1992-2021)

Starting with CDC data on foodborne illnesses and factoring in estimated rates of underreporting
and underdiagnosis, Collier et al. (2021) calculated that in 2014 there were approximately

823.000	cases of cryptosporidiosis in the U.S., of which 322,000, were domestically acquired
waterborne cases. Of the approximately 322,000 cases, the authors estimate that 1,120 resulted in
hospitalizations and 24 in death.

4.3 Aircraft Drinking Water Rule

EPA promulgated the ADWR in 2009 (74 FR 53590; USEPA, 2009) to ensure that drinking
water provided to aircraft passengers and crew is protected from microbial contaminants. ADWR
requires aircraft carriers to develop operation and maintenance plans (O&M plans) and coliform
sampling plans and to conduct routine disinfection and flushing. Positive total coliform samples
are further analyzed for E. coli. If positive E. coli samples result, public notification and
corrective measures are required.

4.3.1	Summary of Health Effects Review Outcome and Information Evaluated

There is limited new literature available on the presence of microbial pathogens in aircraft
drinking water. Handschuh et al. (2015) found that long-haul flights were significantly poorer in
terms of microbial water quality than short haul flights, and that water service vehicles were a
significant source of increased microbial load in aircraft.

A follow-up study by Handschuh et al. (2017) demonstrated that there is a diversity of
microorganisms within the aircraft drinking water supply chain. The researchers sourced water
samples from long-haul and short-haul aircraft, the aircraft water source, and a water service

4-7


-------
vehicle. In total, 308 isolates were characterized and their identity was determined. Many of the
bacteria found were identified as known or potential human pathogens. Burkholderia
pseudomallei, for example, which was isolated from a water service vehicle, is highly pathogenic
and can be transmitted by skin contact with contaminated water. Opportunistic pathogens
capable of causing infections in vulnerable individuals were also found, such as thq Burkholderia
cepacian complex, Pseudomonas aeruginosa, Stenotrophomonas maltophilia, Acinetobacter
baumannii, Ralstonia pickettii, Pseudomonas fluorescens, and Sphingomonas paucimobilis.

Other studies have also found microbial contaminants present in aircraft drinking water,
including Pseudomonas aeruginosa, enterococci, Clostridia, and Salmonella (WHO, 2009;
Schaeffer et al. 2012).

Tracking an illness back to contaminated water served on an aircraft presents a technical
challenge. Most disease incubation periods are longer than the duration of a flight, and even if it
is possible to determine that a disease was incurred in air travel, it may be difficult to determine
if the route of transmission was from beverages, food, or close proximity of people, and to
determine whether transmission occurred on board the aircraft or at an air terminal. Possibly for
these reasons, there was no new health effects information identified that indicated increased
health risk compared to when ADWR was developed.

4-8


-------
5 Analytical Methods

This chapter summarizes the analytical methods approved for contaminant monitoring or
treatment technique (TT) requirements included under the Revised Total Coliform Rule (RTCR),
Ground Water Rule (GWR), Long-Term 2 Enhanced Surface Water Treatment Rule (LT2)
Round 2 and Aircraft Drinking Water Rule (ADWR). The Six-Year Review 4 (SYR4) also
includes the Filter Backwash Recycling Rule (FBRR), but because there are no monitoring
requirements under this rule, there are no analytical methods described for this rule in this
section.

5.1	Revised Total Coliform Rule Monitoring Requirements

Under the RTCR, samples are routinely collected by systems at sites which are representative of
water quality throughout the distribution system. These samples, also called routine samples, are
analyzed for the presence of total coliform bacteria. The number and frequency of samples
required by RTCR depend on the size of the system. If a sample tests positive for total coliform
bacteria, the sample is further analyzed for the presence of E. coli. The rule requires that the
presence of total coliform bacteria in any routine samples require the water system to collect
additional samples (called "repeat" samples) confirming evidence of fecal contamination.
Samples positive for total coliforms and E. coli will trigger the need for the system to take
corrective actions referred to as Level 1 and Level 2 assessments to identify sanitary defects that
could provide pathways for entry of microbial contamination into the distribution system.

For the RTCR, there have been new methods approved as well as revisions to existing methods
since SYR3. The Modified Colitag and Tecta methods were revised with one new method
approved. Other revised methods of RTCR can be identified by referencing updated versions
listed in the 23rd and 24th editions of "Standard Methods for the Examination of Water and
Wastewater."

A new method, Membrane Filtration procedure using REC2, was approved in 2021 and is a
newly approved method for RTCR. This method is similar to the other methods in terms of
providing microbial density values but this method uses a different media.

For all revised and new methods for RTCR, there are no new technologies or significant changes
to the detection methodology. Approved methods are specified in 40 CFR 141.852(a)(5).
Additional methods are listed in Appendix A to Subpart C of Part 141.

5.2	Ground Water Rule Monitoring Requirements

Under the GWR, ground water systems that do not provide 4-log treatment of viruses must
monitor their source water for a fecal indicator if there is a positive total coliform sample in the
distribution system. This positive total coliform sample triggers the source water monitoring
requirement of the GWR. The system must monitor their source water for either E. coli,
enterococci, or coliphage. Approved methods are specified in 40 CFR 141.402(i).

5-1


-------
Since the SYR3, there have been revisions to approved methods, with the Modified Colitag and
Tecta methods having been revised along with those from "Standard Methods for the
Examination of Water and Wastewater." In addition, there was one new method approved,

Rapid' E. coli 2 (REC2). The REC2 method is similar to other methods in terms of the technique
used and was shown to be equally effective in the recovery of total coliform bacteria and E. coli.

There have been no new technologies or significant changes to the detection methodologies used
to detect total coliforms oris, coli, which are detected under the RTCR. The Membrane Filtration
procedure using REC2 method, as with RTCR, is a newly approved method for the GWR.

5.3	Methods for Measuring Disinfectant Residuals in Ground Water

Ground water systems that provide 4-log inactivation, removal, or a state-approved combination
of 4-log virus inactivation and removal, must continue to conduct compliance monitoring to
show that they are providing 4-log treatment. The GWR requires that a system using a chemical
disinfectant to achieve the 4-log inactivation of viruses must use the analytical methods under the
SWTRin 40 CFR 141.74(a)(2).

5.4	Long-term 2 Enhanced Treatment rule Analytical Methods Approved

The purpose of LT2 is to reduce illness linked to Cryptosporidium and other pathogenic
microorganisms in drinking water. Under the Interim Enhanced Surface Water Treatment Rule
(IESWTR) (63 FR 69477, USEPA, 1998) and LT1 (67 FR 1812, USEPA, 2002), EPA
established an NPDWR for Cryptosporidium and set an MCLG of zero. The LT2 supplements
these existing regulations through additional treatment requirements in systems at higher risk for
Cryptosporidium. The LT2 also contains provisions to reduce risks from uncovered finished
water reservoirs and provisions to ensure that systems maintain microbial protection when they
take steps to decrease the formation of disinfection byproducts that result from water treatment.

The analytical methods for Cryptosporidium, E. coli, and turbidity have not changed, nor have
any new methods been approved for these analytes since LT2 was promulgated. The LT2
requires systems and/or laboratories to use either "Method 1622: Cryptosporidium in Water by
Filtration/IMS/FA" (USEPA, 2005a) or "Method 1623: Cryptosporidium and Giardia in Water
by Filtration/IMS/FA" (USEPA, 2005b). EPA Methods 1622, 1623, or 1623.1 can be used to
characterize Cryptosporidium levels in the source water of PWSs for the purposes of risk-
targeted treatment requirements under the LT2. Approved methods are specified in 40 CFR
141.704-707.

5.5	Aircraft Drinking Water Rule

ADWR was developed to protect against disease-causing microbiological contaminants through
the required development and implementation of aircraft water system operations and
maintenance plans. This includes routine disinfection and flushing of the water system, and
periodic sampling of the onboard drinking water. All aircraft water systems collect samples for
analysis of total coliform bacteria according to the frequency and procedures described in the
coliform sampling plan. Each routine, repeat, or follow-up sample that is positive for total
coliforms is tested for the presence of E. coli. If any sample is positive for E. coli, public

5-2


-------
notification and corrective disinfection and flushing are triggered. Approved methods are
specified in 40 CFR 141.852.

5-3


-------
6 Occurrence and Exposure

This chapter summarizes the results of EPA's occurrence analyses of regulated microbial
pathogens and indicators. The objectives of the occurrence analyses are to characterize national
occurrence baselines of the relevant microbial contaminants and related indicators and changes
to these baselines under the microbial rules covered during the SYR4. First, the chapter presents
occurrence analyses for total coliforms and E. coli relative to the Ground Water Rule (GWR) and
the Revised Total Coliform Rule (RTCR), using compliance monitoring data from the Fourth
Six-Year Review Information Collection Request (ICR) database (referred to as the "SYR4 ICR
microbial dataset" in this document, see USEPA, 2019a), and other sources, including the SYR3
ICR microbial dataset.

The SYR4 also includes the Filter Backwash Recycling Rule (FBRR), but because there are no
monitoring requirements under this rule, there is no occurrence analysis presented for this rule in
this chapter.

Next, the chapter presents and discusses the analytical results of the source water monitoring
data, related to the Long Term 2 Enhanced Surface Water Treatment Rule (LT2), primarily for
Cryptosporidium, that are contained in the SYR4 ICR microbial dataset. Third, EPA applies
statistical modeling to quantify uncertainty in total coliform occurrence at undisinfected ground
water systems (which mostly serve small populations), and also conducts a brief analysis of
triggered source water monitoring for E. coli at undisinfected ground water systems. Both of
these analyses depend on decision trees used to identify undisinfected systems. EPA presents the
systematic approach used to identify disinfection status in SYR4, and explains how this
definition differs from the definition used in SYR3.

Finally, the chapter presents the results of occurrence analysis for the Aircraft Drinking Water
Rule (ADWR) using the Aircraft Reporting and Compliance System (ARCS) database. Overall,
the analytical results presented and discussed in this chapter are intended to be helpful for
addressing one of the questions prescribed in the EPA Protocol for the Fourth Review of Existing
National Primary Drinking Water Regulations (USEPA, 2024a): Is there a significant increase in
health risk estimated from exposure to the contaminant?

EPA notes that (1) as presented below, the majority of the sampling records in both the SYR4
and SYR3 ICR microbial datasets are related to total coliforms and E. coli in distribution systems
in the context of the GWR and RTCR; (2) data on Cryptosporidium in source water (from round
2 monitoring under LT2) are also included in the SYR4 ICR dataset but are not present in the
SYR3 ICR dataset; (3) disinfectant residuals in distribution systems, which enable the evaluation
of paired records of total coliforms / E. coli and residuals, are being analyzed under a separate
effort to support potential revision of Microbial and Disinfection Byproducts (MDBP) rules, so
they are not discussed in this document; and (4) the compliance monitoring data presented for
ADWR show results for total coliform / E. coli sampling. Thus, the analytical results presented
and discussed in this chapter include total coliform / E. coli occurrence in distribution systems
and aircraft systems, E. coli in source ground water, and Cryptosporidium in source surface
water.

6-1


-------
Some of the goals of SYR4 are: to evaluate the possible differences in total coliform occurrence
between disinfected and undisinfected ground water systems, to suggest potential impacts of the
combined GWR and RTCR on the occurrence of total coliform and E. coli, and to identify the
sizes and types of undisinfected systems with the highest potential for public health
improvements, informed by robust statistical methods that identify total coliform detection rates
with low uncertainty. Here are some considerations that go into the analyses:

Identifying Disinfecting vs. Undisinfecting Systems

In SYR3, EPA identified and grouped disinfecting ground water systems based, in part, on the
level of disinfection for the purpose of assessing the impact of different levels of disinfectant
residuals on microbial contamination as indicated by total coliforms and E. coli. For SYR4, the
definition of disinfecting ground water systems (for community and non-transient non-
community water systems) was revised so that it was not based on disinfection residual levels
but focused instead on the inclusion of DBP data as an indicator of disinfection. A more detailed
justification for the change in definition is provided in section 6.4.1. For most analyses
performed in this document, systems from the SYR3 and SYR4 ICR were categorized as either
"disinfecting" or "undisinfecting" based on the SYR4 definition. There are a few instances where
SYR3 and SYR4 data are categorized using the SYR3 definition to enable comparison, to see
what difference the selected definition makes.

Evaluating Changes Attributable to the GWR and RTCR

To evaluate impacts of the GWR and RTCR on the occurrence of microbial indicators, it was
necessary to identify a subset of systems with data that span the timeline from before GWR, after
GWR, and after RTCR. To achieve this, a subset of systems that reported their routine total
coliform / E. coli monitoring samples for all years from 2007 to 2019 was identified. All data
were evaluated using the SYR4 definition of "undisinfected" (see above). Trends over time are
calculated using simple summary statistics of the total coliform / E. coli sampling results. EPA
did not attempt to investigate any combined GWR/RTCR attributable changes over time using
the statistical models applied to undisinfected systems. However, the years 2011 and 2019 are
compared using both definitions for undisinfected systems in Exhibit 6-24 and Exhibit 6-25.

Eliminating Outliers in Positivity Rate

To keep the size of the working data set manageable, all total coliform / E. coli occurrence data
(number of samples taken and the number of positive results) were aggregated by system, month,
and year. With this "reduced dataset," the percentage of positive samples for each system in any
given year could be calculated. The results were summarized using simple statistics. Note that
the number of systems in the data set varied year to year.

An important detail of the SYR3 and SYR4 ICR is that systems may not have collected or
reported all their required routine monitoring samples. This can lead to a misrepresentation of the
positivity rate (e.g., the 100% positivity rate of a system on quarterly monitoring that only
reported one sample for the year instead of four, and had a single positive result, might not be
representative of actual conditions at the system). To avoid including false representations of the
total coliform / E. coli positivity rate without overly excluding valuable data, a threshold of 90%

6-2


-------
completeness was created for certain analyses. In these cases, only data from the subset of
systems that reported a minimum of 90% of their routine monitoring samples were included. In
contrast to this 90% completion criterion, the statistical models used to assess the undisinfected
systems (Exhibit 6-24 and Exhibit 6-25) did not exclude data and presented the results with a
suitable display that captured the uncertainty associated with using all data.

Characterizing Systems that Have Highest Percentage of Total Coliform Positive Samples

Under SYR3, routine total coliform and E. coli sample data from undisinfected small ground
water systems were analyzed (Messner et al., 2017). For the year 2011, Messner et al. divided
the undisinfected systems into three types: (community, non-community transient, and non-
community non-transient. Each type was then further sub-divided into three population-served
size bins. When focusing the analysis on specific groups of systems (i.e., based on system size or
system type), the sample sizes become smaller, and this makes it challenging to generate reliable
estimates. For SYR3, a Bayesian Markov Chain Monte-Carlo (MCMC) model was implemented
to quantify the distribution of positive total coliform / E. coli samples and derive a mean
positivity rate with quantified uncertainty. This analysis was repeated using the SYR4 2019 data
and the SY4 definition of undisinfected systems for 2011 and 2019 (Exhibit 6-24 and Exhibit
6-25). EPA chose not to evaluate every SYR4 year (2012-2018) using the Bayesian MCMC
model. EPA reasoned that the common systems with 90% completeness summary statistics were
adequate to evaluate putative GWR/RTCR effects.

This chapter is organized as follows:

•	Section 6.1 describes the data sources used in the microbial occurrence analysis;

•	Section 6.2 presents a summary of the analysis related to the microbial contaminants in
distribution systems relative to the GWR and RTCR, focusing on total coliforms and E.
coli;

•	Section 6.3 presents a summary of the analysis to support the evaluation of LT2,,
focusing on Cryptosporidium in source water;

•	Section 6.4 presents EPA's methodology for identifying undisinfected ground water
systems, along with two sets of analyses involving these undisinfected ground water
systems: modeling of total coliform detection rates and characterization of triggered E.
coli source water monitoring results; and

•	Section 6.5 presents a summary of the analysis related to the ADWR.

6-3


-------
6.1 Data Sources for Microbial Occurrence Analyses

Data sources used in EPA's analyses of national-level microbial occurrence include national
datasets compiled by EPA such as the third and fourth Six-Year Review Information Collection
Request datasets (SYR3 ICR and SYR4 ICR, respectively), the Aircraft Reporting and
Compliance System (ARCS), and Safe Drinking Water Information System (SDWIS). Each of
those sources are described below along with how they were used in the microbial occurrence
analyses. See USEPA (2024b) for more information on the data files that were used for the
analyses summarized in this chapter.

6.1.1 Six-Year Review 4 Information Collection Request Data

This section provides a description of the SYR4 ICR database, which is the primary source of
data used in the SYR4 microbial analysis, and describes subsets of the database that were used
for the various analyses in this chapter. It is important to note that analyses described in this
report were conducted to inform the Six-Year Review and were not meant to assess compliance
with regulatory standards.

The 1996 SDWA Amendments require EPA to review each National Primary Drinking Water
Regulation (NPDWR) at least once every six years and revise it, if appropriate. As part of the
Six-Year Review, EPA evaluates any newly available data, information, and technologies to
determine if any regulatory revisions are needed. There is no national database that receives and
stores all relevant data on the occurrence of regulated contaminants in public drinking water
systems. To help support each Six-Year Review of NPDWRs, EPA conducts a voluntary data
call-in from the states and primacy entities (territories and tribes) to obtain compliance
monitoring data. EPA works with the states and other primacy agencies to receive their complete
records of compliance monitoring data (i.e., public drinking water system regulated contaminant
occurrence data). The compliance monitoring data are obtained through an ICR process. Under
the SYR4 ICR (EPA ICR No. 2574.01, USEPA, 2018), EPA requested compliance monitoring
data for the time period from 2012 through 2019 for the following microbial contaminants and
indicators: total coliforms, E. coli, fecal coliforms, enterococci, coliphage, Cryptosporidium, and
Giardia lamblia. In addition, the SYR4 ICR included treatment and disinfectant residual
information, sample-specific and system-specific information such as system type and source
water type, and corrective action information. In all, 46 states and thirteen other primacy
agencies provided compliance monitoring data records (USEPA, 2024b).

6.1.1.1 Description of Data Collected Under Six-Year Review 4 Information Collection
Request

EPA conducted data management and quality assurance (QA) evaluations on the data received
for contaminants evaluated for the SYR4 to establish a high quality, national compliance
monitoring and treatment technique (TT) dataset consisting of data from 59 states / primacy
entities (46 states plus territories, Washington, D.C., and tribes). The compliance monitoring data
for these 59 states / primacy agencies comprise more than 71 million analytical records from
approximately 140,000 public water systems (PWSs), which collectively serve more than 301

6-4


-------
million people nationally.1 This dataset is the largest and most comprehensive compliance
monitoring and TT dataset ever compiled and analyzed by EPA's Drinking Water Program. The
final SYR4 ICR dataset includes more than 25 million analytical records for microbial
contaminants and indicators (total coliforms, E. coh, fecal coliforms, heterotrophic bacteria (as
measured by heterotrophic plate count [HPC]), Giardia lamblia, Cryptosporidium, coliphage,
and enterococci). By comparison, the final SYR3 ICR dataset included almost 12 million
analytical records for microbial contaminants and indicators. For more details on the SYR4 ICR
Dataset, including descriptions of reviews for completeness, representativeness, and data
management and quality assurance / quality control (QA / QC), refer to USEPA (2024b).

Quality Assurance Activities

After the individual state datasets received under the SYR4 ICR were converted into a consistent
format, a significant effort was undertaken to ensure the quality of the data submitted. An
important objective regarding the data to be used for the SYR4 contaminant occurrence analyses
is development of a consistent and repeatable data management approach. Consistent data
editing and QA/QC assessments allow the individual state datasets to be aggregated and jointly
evaluated, to provide an overview of national occurrence patterns for individual contaminants.

Uniform, detailed QA/QC assessments conducted on the state compliance monitoring datasets
included comparisons of the number of systems with compliance monitoring data in each state
against total system inventory numbers from the Federal Safe Drinking Water Information
System database (SDWIS/Fed), and examination of the number of analytical records per system
(or per contaminant) to evaluate the completeness of the submitted analytical records. These
comparisons helped to understand the representativeness of the data provided by each state.
Identified errors or questionable results that did not have straight-forward explanations were
addressed through consultations with state data management staff to ensure consistent and
appropriate interpretations.

As described in the QA/QC document (USEPA, 2024b), the following QC measures were
applied to the Microbial Rule contaminants, including total coliforms, fecal coliforms, E. coli,
Cryptosporidium, Giardia lamblia, enterococci, and coliphage:

•	Removal of records from non-public water systems

•	Removal of records from systems with missing source water type or population served
data

•	Removal of records from outside the SYR4 ICR date range

•	Removal of records marked as being "not for compliance"

•	Removal of records marked with a sample type code other than routine, repeat, or
triggered

1 These statistics reflect the portion of the dataset representing compliance monitoring samples collected for
requested regulated contaminants. The full dataset (including data not specifically requested by EPA but submitted
voluntarily by some states) was comprised of over 83 million records from approximately 142,000 water systems.

6-5


-------
•	Removal of records with no data/results

•	Removal of records with irregular system type codes

The number of microbial records excluded per each of the steps described above is summarized
in Exhibit 6-1. Overall, 99 percent of microbial records are retained after these QA steps. Note
that very limited data were submitted for enterococci (8 records) and coliphage (3 records). For
the purpose of SYR4 occurrence analysis, EPA focused on total coliforms, E. coli (in distribution
systems and source water, examined separately), and Cryptosporidium in source water. Although
EPA requested data under SYR4 for Giardia lamblia, enterococci, and coliphage, those
contaminants were not included as part of the analysis due to low number of records received for
enterococci and coliphage and the fact that Giardia lamblia is part of an on-going potential
revision action. See Exhibit A-l in Appendix A for a state-level breakdown of the number of
records included in the analysis for total coliforms, E. coli, fecal coliforms, and
Cryptosporidium.

Exhibit 6-1. SYR4 Data Summary of the Count of Records Removed via the
Quality Assurance Measures Applied to Microbial Rule Contaminants

QA Step

Count of Records1

Included

Excluded

Original number of records

28,329,039

Stepl: Removal of data from non-public water systems.

28,315,533

13,506

Step 2: Removal of data from systems with missing source water type and/or
population served information.

28,236,298

79,235

Step 3: Removal of data with a sample collection date outside of the Six Year 4
ICR date range of 2012 - 2019.

28,114,841

121,457

Step 4: Removal of data marked as being "not for compliance."

27,985,027

129,814

Step 5: Removal of microbial data with sample type code other than "RT" (routine),
"RP" (repeat), or'TG" (triggered).

27,981,035

3,992

Step 6: Removal of records with no data/results

27,964,042

16,993

Step 7: Removal of records with irregular system type codes (specific to State of
PA where unknown system type codes were included)

27,962,474

1,568

Final number of records

27,962,474

Percent Included

99%

1 The following analytes are included in the counts: total coliforms, fecal coliforms, E. coli, Cryptosporidium, Giardia lamblia,
enterococci, and coliphage.

6-6


-------
6.1.1.2	Limitations of the Six-Year Review 4 Data

The SYR4 ICR microbial dataset consists of data from 57 primacy agencies (including 44 states
and 13 tribes and other entities). The SYR4 ICR does not include any data from four states
(Georgia, Michigan, Mississippi, and New Mexico) and three other entities (Guam, Puerto Rico,
and U.S. Virgin Islands). The SYR4 ICR microbial dataset includes data from approximately
127,000 PWSs, representing over 85 percent of the public water systems in the U.S. and 67
percent of the population served by public water systems in the U.S. In addition to the four states
that did not provide any SYR4 ICR data, two states did not provide microbial data that could be
utilized in the analysis. EPA recognizes a large degree of variability in the number of records
provided by water systems from state to state. Furthermore, the dataset is limited to the period of
2012 to 2019 and the number of samples and number of systems included in the SYR4 ICR
dataset from each state varies from year to year. There may also be varying levels of
completeness of data records for some compliance monitoring periods at some systems.

6.1.1.3	SYR4 "Reduced" Total Coliform / E. coli Dataset to Support Analysis of Ground
Water Rule and Revised Total Coliform Rule

As described above, the total coliform / E. coli / fecal coliform data received as part of the SYR4
ICR included records for individual samples. After EPA performed the QA/QC steps described
in Exhibit 6-1, the resulting dataset contained nearly 28 million individual monitoring records.
Such a large number of records poses great challenges for data management and analysis. To
better facilitate the analysis of how total coliform and fecal indicator (either E. coli or fecal
coliform) positivity rates varied by system size, system type, source water type, and disinfection
status over time, EPA created a "reduced total coliform / E. coli dataset." The "reduced dataset"
was formed by data aggregation. In this case, the individual sampling records were reduced to
summary counts for each month, year, and water system for the following measures: (a) the total
number of routine (RT) samples provided for total coliforms, (b) the number and percent of RT
samples testing positive for total coliforms, (c) the total number of RT samples provided for E.
coli, and (d) the number and percent of RT samples testing positive for E. coli / fecal coliforms.
Similar counts were generated for repeat (RP) samples of total coliforms and E. coli / fecal
coliforms as well. In other words, the reduced dataset includes, for each water system and month,
counts of the number of routine and repeat samples assayed and the number found to be positive
for total coliforms and for is. coli / fecal coliforms.

Note that because of the small count of fecal coliform monitoring results (as indicated in Exhibit
6-2 below), findings for fecal coliforms are not presented and discussed in this chapter. Also,
some additional QA steps beyond what are described in Section 6.1.1.1 were applied to the
individual sample results prior to the creation of the reduced dataset. These steps involved
considering the monitoring and reporting requirements for total coliforms / E. coli prescribed
under the RTCR. Exhibit 6-2 summarizes the number of records removed via each QA step
before generating the SYR4 reduced total coliform / E. coli dataset. EPA notes, in contrast to the
SYR3 QA steps, the SYR4 QA steps do not include this step. E. coli records that did not have a
corresponding total coliform positive record were not excluded from the reduced dataset.

6-7


-------
Exhibit 6-2. Summary of the Count of Records Removed via the Quality Assurance Measures for Six-Year

Review 4 Reduced Total Coliform / E. coli Dataset

OA Steps Applied

Total Coliforms

E. coli

Fecal Coliforms

Included

Excluded

Included

Excluded

Included

Excluded

Starting number of records

21,010,733

7,277,177

16,920

Number of records after QA measures applied

20,746,119

264,614

7,175,363

101,814

16,818

102

Removal of records with sample type code other than "RT" or "RP"

17,539,775

3,206,344

6,527,234

648,129

16,416

402

Removal of records with presence indicator code other than "A" or "P"

17,533,540

6,235

6,525,739

1,495

16,416

0

Removal of non-distribution system samples. (I.e., include only the records with
sample point type of "DS", "FC", "FN", "LD", "MD", or "MR" or records with water
facility type of "CC", "DS", "TP", or "TM" and sample point type of "WS" or null.)

16,538,009

995,531

5,852,233

673,506

11,193

5,223

Removal of records with a greater free chlorine concentration than total chlorine
concentration (Note: Kept records with <0.1 mg/L of absolute difference or
<30% of relative difference between free and total chlorine concentrations.)

16,462,870

75,139

5,846,138

6,095

11,192

1

Removal of records from states who confirmed that their microbial data are not
recorded as individual samples

16,454,914

7,956

5,839,855

6,283

11,190

2

Final Number of Records

16,454,914



5,839,855



11,190



6-8


-------
In the final "reduced" dataset, there are data for a total of 109,155 water systems located in 48
states/entities. Exhibit 6-3 provides an excerpt of the information included in the final "reduced"
dataset. Note that not all included systems have results for all 12 months of each year and not all
included system months have complete monitoring records relative to the regulatory monitoring
requirements of RTCR. Furthermore, there were some repeat samples that occurred in a different
month than their corresponding routine sample; thus, some system/month/year combinations
have repeat samples but no routine samples. Overall, the data "reduction" process reduced the
number of records from 22,305,959 individual sampling results to 5,327,984 records at the
system monthly level. This enabled EPA to more easily manage the data file and more
effectively conduct analyses on personal computers. See Appendix A for a table that includes the
field names and definitions for the reduced dataset.

6-9


-------
Exhibit 6-3. Excerpt of Data from Six-Year Review 4 Reduced Total Coliform / E. coli Dataset

PWSID

State

Population
Served

Pop
Cat

System
Type

Source
Water
Type

Year

Month

Disinfecting
Status

Number
of RTTC
Samples
Required

Percent
Completeness
of RTTC
Records

Number
of
RTTC

Number
of RTTC
Positive

Percent
RTTC
Positive

Number
of
RTEC

Number
of RTEC
Positive

Percent
RTEC
Positive

Number
of
RPTC

Number
of RPTC
Positive

Number

of
RPEC

Number

of
RPEC
Positive

OR4100601

OR

172

2

C

GW

2018

12

Disinfecting

1

100%

1

100%

1

1

0

0%

3

0

0

0

OR4100601

OR

172

2

C

GW

2018

6

Disinfecting

1

100%

1

100%

1

1

1

100%

3

3

3

3

OR4100601

OR

172

2

c

GW

2018

2

Disinfecting

1

100%

1

100%

1

1

0

0%

3

0

0

0

OR4100601

OR

172

2

c

GW

2018

11

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

10

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

9

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

8

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

7

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

5

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

4

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

3

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

OR4100601

OR

172

2

c

GW

2018

1

Disinfecting

1

100%

1

0%

0

0

0

0%

0

0

0

0

Notes: PopCat = system's population size category; RTTC = number of routine total coliform samples; RTEC = number of routine E. coli samples; RPTC = number of repeat total coliform samples; RPEC =
number of repeat E, coli samples; Percent Completeness = percent of routine total coliform samples taken by the system as compared to the total number of required routine total coliform samples based on
system size.

6-10


-------
6.1.1.4 Additional Six-Year Review 4 Information Collection Request Data Records to
Support Review of the Long Term 2 Enhanced Surface Water Treatment Rule

The LT2 provides for source water monitoring for Cryptosporidium and associated water quality
parameters. Under this provision, monitoring data from the Round 1 monitoring period (2006 to
2012) were analyzed by EPA and the results were presented in the Six-Year Review 3 Technical
Support Document for the Long-Term 2 Enhanced Surface Water Treatment Rule (USEPA,
2016b). Primacy agencies provided more than 19,000 monitoring records (sample analytical
results) for Cryptosporidium over the 8-year period of the SYR4 ICR (2012 - 2019). More than
99 percent of the Cryptosporidium records were from CWSs, with about 95 percent of those
records from surface water sources. Fewer than 0.3 percent of the 19,000 Cryptosporidium
monitoring records provided concentration levels.

In addition to these occurrence data, primacy agencies provided binning information for PWSs
based on monitoring conducted in response to LT2. Filtered systems serving at least 10,000
people were classified into a "bin" based on the results of their initial source water monitoring.
(See Exhibit 6-18 for more details on systems' bin classifications.) Bin classification determines
whether further treatment for Cryptosporidium is required. A second round of source water
monitoring was required six years after the initial bin classification per Round 1 monitoring
results and bin classifications were revised for some systems based on Round 2 monitoring
results.

6.1.2 Aircraft Drinking Water Rule Data

The Aircraft Reporting and Compliance System (ARCS) is a centralized web-based data
collection and management system that is used to facilitate the reporting of aircraft water system
data under the ADWR, for accountability and regulatory oversight. Air carriers subject to the
ADWR must report to EPA the following information in ARCS, unless an alternative reporting
method has been approved (see https://www.epa.gov/dwreginfo/aircraft-drinking-water-rule):

•	A complete inventory of aircraft water system fleet,

•	The date the Operations and Maintenance plan was developed,

•	The date the Coliform Sampling plan was developed,

•	The date the aircraft water system sampling plan(s) was/were incorporated into the
aircraft water system Operations and Maintenance plan,

•	The date the Operations and Maintenance plan(s) was/were incorporated into FAA-
accepted air carrier Operation and Maintenance program,

•	The frequency of routine disinfection and flushing, the corresponding routine total
coliform sampling frequency, and

•	The dates of routine disinfection and flushing, routine coliform sampling dates and
results, and corrective actions (when applicable).

For SYR4, EPA downloaded and reviewed compliance monitoring data available in ARCS as of
May 2021. Approximately 140,000 records of aircraft water system compliance monitoring data
for total coliform and E. coli samples were available in ARCS from February 2011 through May

6-11


-------
2021, including results reported for more than 70 different makes/models of aircraft. These
results were used to characterize the positivity rates of total coliforms and E. coli in aircraft
water system on an annual basis, as well as for all the years that data were available (2011-2021)
and for the subset of years 2012 through 2019. The evaluation of data for years 2012 through
2019 was performed to allow for a comparison with similar data for stationary PWSs as
described in Section 6.2. In addition, this approach removes potentially confounding
considerations associated with evaluating data for calendar year 2020, when a large number of
aircraft water systems were inactive due to COVID-19, as well as years 2011 and 2021, for
which the ARCS data evaluated at this time only represent partial years.

Aircraft inventory data, including manufacturer, model, and disinfection and flushing frequency,
were linked to the monitoring results by public water system identification number (PWSID).
Aircraft PWSs were categorized as small, medium, or large based on the seat capacity (small =
130 or fewer seats; medium = 131 - 250 seats; large = over 250 seats). Note that these categories
were developed specifically for this analysis, based on the dataset, and do not represent
regulatory categories. ADWR does not categorize aircraft water systems based on size. In
addition, the first three digits of the model number were used to summarize the make/model of
each aircraft.

A number of QA steps were applied to the ADWR dataset to identify the total coliform and E.
coli records suitable for analysis. Data were excluded via the following QA steps:

•	Records where [Location] was were excluded.

•	Records where [Total Coliform] was or "from" were excluded.

•	Records where the [Sample Taken On] date was incorrectly entered were excluded.

These incorrectly entered dates were as follows: 12/08/0014 00:00", "09/26/0201 03:52",
"09/13/0019 03:59", "09/09/0201 03:35", "07/22/0204 05:17", "07/16/0018 01:35",
"06/21/0018 01:40", and "02/02/0017 16:10."

•	Records where a [Total Coliform] result was entered as "absent" but [E. coli] was
positive.

The ADWR analyses were stratified in a variety of ways to summarize results, including the
number of total coliform samples and public water systems, by aircraft size, manufacturer,
model, air carrier, sample type, and more. It is important to note that all E. coli positivity rates
were calculated twice, under two different sets of assumptions:

1.	An E. coli sample was included in the analysis only if the E. coli result was listed as
"Present" or "Absent."

2.	An E. coli sample was included if the E. coli result was listed as "Present" or "Absent"
(i.e., the same as the first set of assumptions), but with an added assumption that an E.
coli sample could be considered "Absent" and included in the analysis as such if the
associated total coliform result was reported as "Absent" and there was no E. coli result
provided. These results are labeled in the file as "E. coli (Alternative Approach) "

6-12


-------
After the QA steps were applied, there were 140,502 total coliform results used in this
evaluation, provided by 8,093 PWSs and covering the full range of years for which ARCS data
were collected (i.e., February 2011 - May 2021). Under the first set of assumptions listed above,
there were 92,994 E. coli results provided by 7,091 PWSs (i.e., 66 percent of the number of total
coliform results and 88 percent of the number of aircraft water systems providing total coliform
results), with a total of 241 results (0.26 percent) positive for E. coli. Under the second set of
assumptions listed above for theE. coli analysis, there were 140,485 E. coli results provided by
8,093 aircraft water systems, with 241 results (0.17 percent) positive for E. coli.

6.1.3	Safe Drinking Water Information System Data

EPA used inventory information from SDWIS/Fed to identify the number of systems and the
population served by systems nationally, as well as the breakdown by source water type, system
type, and system size. SDWIS inventory tables were filtered to include all active water systems
in 2019 (USEPA, 2019b). In addition, SDWIS data were used in the process for identifying the
undisinfected ground water systems as discussed in Section 6.4.1.

6.1.4	Other Data Sources

6.1.4.1 Six-Year Review 3 "Reduced" Total Coliform / E. coli Dataset for Analysis of
Ground Water Rule

A SYR3 reduced total coliform / E. coli dataset was also generated using the same method and
the same QA steps described above for the SYR4 reduced dataset. As described above, the data
for each system and month / year were reduced to a small number of summary counts: (a) the
total number of routine (RT) samples provided for total coliforms, (b) the number and percent of
RT samples testing positive for total coliforms, (c) the total number of RT samples provided for
E. coli and (d) the number and percent of RT samples testing positive for E. coli. Similar counts
were generated for repeat (RP) samples of total coliforms and E. coli as well. In other words, the
reduced dataset includes, for each water system and month, counts of the routine and repeat
samples assayed for and found to be positive for total coliforms, E. coli, and fecal coliforms.
Note that at the time of the SYR3 analyses (USEPA, 2016a), a separate "reduced total coliform /
E. coif' dataset had been created using some different QA steps. A new SYR3 reduced total
coliform / E. coli dataset was created here for use in direct comparisons with the SYR4 reduced
total coliform / E. coli dataset. Exhibit 6-4 summarizes the number of records removed via each
QA steps in order to generate the SYR3 reduced dataset.

6-13


-------
Exhibit 6-4. Summary of the Count of Records Removed via the Quality Assurance Measures for Six-Year

Review 3 Reduced Total Coliform / E. coli Dataset

OA Steps Applied

Total Coliforms

E. coli

Fecal Coliforms

Included

Excluded

Included

Excluded

Included

Excluded

Starting number of records

9,953,551

1,833,281

281,642

Number of records after QA measures applied under the SYR3 effort (Note: Used
text files posted on SYR3 website.)

9,766,686

186,865

1,804,329

28,952

264,090

17,552

Removal of non-distribution system samples, (i.e., include only the records with
sample point type of "DS", "FC", "FN", "LD", "MD", or "MR" or records with water
facility type of "CC", "DS", "TP", or "TM" and sample point type of "WS" or null.)

9,018,655

748,031

1,545,146

259,183

111,059

153,031

Removal of records with a greater free chlorine concentration than total chlorine
concentration (Note: Kept records with <0.1 mg/L of absolute difference or <30%
of relative difference between free and total chlorine concentrations.)

8,869,163

149,492

1,544,418

728

111,034

25

Removal of records from one state (South Carolina) as the data are not
representative of the full state monitoring results.

8,864,250

4,913

1,543,025

1,393

110,067

967

Removal of records with sample type code other than "RT" or "RP"

8,850,363

13,887

1,543,025

0

110,067

0

Removal of records with presence indicator code other than "A" or "P"

8,526,333

324,030

1,543,025

0

110,067

0

Final Number of Records

8,526,333



1,543,025



110,067



6-14


-------
In the final "reduced" dataset, there are data for a total of 84,389 water systems located in 39
states/entities. Note that not all included systems have results for all 12 months of each year and
not all included system months have complete monitoring records relative to the regulatory
monitoring requirements of RTCR. Furthermore, there were some repeat samples that occurred
in a different month than their corresponding routine sample; thus, some system/month/year
combinations have repeat samples but no routine samples. Overall, the data "reduction" process
reduced the number of records from 10,179,425 individual sampling results to 3,024,834 records
at the system monthly level.

6.2 Analytical Results of Samples Taken from the Distribution System

This section presents and discusses results of EPA's analysis of total coliform and E. coli
contamination in source water and distribution systems based on samples measured in
distribution systems. In the smallest systems with little or no distribution system, distribution
system samples can be used to infer source water contamination. It is important to note that a
total coliform oris, coli positive sample collected in the distribution system should not be viewed
solely as a distribution system contamination event because there is an unknown source water
contamination component. To evaluate the occurrence of total coliforms and E. coli in the
distribution system, EPA used a statistical summary approach.

6.2.1 Total Coliform / E. coli Occurrence in Context of Ground Water Rule and Revised
Total Coliform Rule

To evaluate the potential impact of the GWR and the RTCR on the occurrence of microbial
indicators throughout the United States and whether there is an opportunity to make further
improvements in public health, EPA conducted an analysis focused on the occurrence of total
coliforms and E. coli using compliance monitoring data from the SYR4 and SYR3 ICRs. The
SYR3 ICR dataset included total coliform and E. coli compliance monitoring from 2006 to 2011
and the SYR4 ICR dataset included total coliform and E. coli compliance monitoring data from
2012 to 2019. As indicated in Exhibit 6-5, three regulatory milestones fell within the SYR3 and
SYR4 periods of record:

•	Beginning of GWR implementation (2009)

•	First round of Sanitary Surveys completed under the GWR (2014)

•	Beginning of the implementation of the RTCR, including Level 1 and Level 2
assessments (2016)

E. coli source water monitoring under the GWR is triggered by RTCR monitoring positive total
coliform results for undisinfected ground water systems or ground water systems with less than 4
log inactivation/removal of viruses. EPA notes that in addition to E. coli results, the routine
distribution system monitoring from TCR/RTCR itself (for total coliforms) can also be used as a
potential indicator of risk, e.g. Messner et al, 2017. The results of triggered source water
monitoring for E. coli under GWR are discussed in section 6.4.2.3.

6-15


-------
Exhibit 6-5. GWR and RTCR Implementation Milestones in the Context of the Six-
Year Review 3 and Six-Year Review 4 Information Collection Request Timeframes

Began GWR Implementation

Began Implementation
1st Round of SS of RTCR (including
was completed Leve, 1/2 Assessment)

under GWR

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

SYR3 ICR

SYR4 ICR

The GWR and the RTCR apply in combination to protect public health. All systems are required
to take routine total coliform samples under the RTCR. If an undisinfected system's RTCR
sample is also positive for E. coli then the GWR and the RTCR require additional samples and
actions. Given the overlapping requirements, it is difficult to assess the protections of each
individual rule based on total coliform detections.

To start, EPA examined yearly total coliform detections (using the SYR4 disinfection definition-
-see Section 6.4.1, below) by including all monitoring records from all PWSs (including both
ground water and surface water systems) contained in the SYR4 ICR and SYR3 ICR databases
to ensure the maximum representation of all PWSs (all of which are subject to the RTCR
requirements) by the datasets. This analysis focused on results from eight separate years, two
from each of the four periods identified in Exhibit 6-5 (viz., 2007/2008, 2010/2011, 2014/2015,
and 2018/2019). EPA used two-year averages from each of the periods to reduce the effect of
background yearly variation. Exhibit 6-6 shows the result of this analysis. The number of
systems and the associated number of routine total coliform records varied substantially from
year to year but there was an overall apparent decreasing trend in total coliform occurrence from
2006 to 2019, and particularly after the implementation of RTCR (i.e., after 2016). Inspection of
the underlying data indicates that the apparent observed decreasing trend among all PWSs is
mainly driven by an apparent decreasing trend among ground water systems, which constitute a
majority of PWSs in the nation. This is explored further with a more focused analysis of
common systems that also has reduced bias, described below.

Exhibit 6-6. Yearly Trend of Percent Total Coliform Positive Results for Routine
Sampling at All Public Water Systems (GW and SW Systems)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems
(GW + SW)

70,685

70,278

74,953

74,691

86,237

85,415

95,516

95,099

#RTTC

1,226,098

1,327,476

1,589,336

1,647,311

1,861,738

1,883,681

2,231,731

2,304,040

#RTTC+

19,589

20,364

20,056

19,474

21,787

21,804

20,455

19,058

%RTTC+

1.60%

1.53%

1.26%

1.18%

1.17%

1.16%

0.92%

0.83%

6-16


-------
Year

2007 2008

2010 2011

2014 2015

2018 2019

Ave @ 2 years

1.57%

(Right Before GWR)

1.22%
(Right after GWR)

1.16%

(Right after completing
SS)

0.87%
(Few years after RTCR)

Relative
difference

-21.96% -4.76% -25.09%

Overall
Difference

-44.32%

As noted above and as indicated by Exhibit 6-6, the number of systems with total coliform
records in the SYR3 and SYR4 ICR datasets varied from year to year (e.g., the number of
systems in 2007 and 2019 was 70,685 and 95,099, respectively). In addition, the included
systems may have different degrees of completeness of total coliform records (relative to the
number of routine total coliform samples that need to be taken for the given population served
under TCR/RTCR) in different years. These differences can contribute to the background yearly
variability of the percent total coliform positive findings.

To reduce the background yearly variability due to these factors, EPA took some additional
steps. EPA analyzed the yearly trends using only common systems (i.e., systems with data for all
eight of the years between 2007 and 2019 included in the analysis) with at least 90 percent of
completeness of routine total coliform monitoring records for a given system month (i.e., those
system months with less than 90 percent of completeness were excluded for yearly trend
analysis). (See Appendix A for further explanation on the use of the 90 percent completeness
threshold, and a sensitivity analysis on the difference between requiring 100 percent and 90
percent completeness and on the decision to use common systems only.)

EPA notes that the smallest systems take either quarterly or monthly total coliform samples, so
they are more likely to be excluded from the common systems analysis.

After applying these additional screening criteria, EPA focused the analysis on ground water
systems. Because the small systems are usually undisinfected ground water systems, this focus
may preferentially exclude these small ground water systems. EPA determined which systems
are disinfecting using the systematic method described in Section 6.4.1. Results are shown in
Exhibit 6-7. Overall, an apparent increase in the percentage of disinfecting systems was observed
overtime (i.e., steadily increasing from 45.3 percent in 2007/2008 to 53.1 percent in 2018/2019).
Although it is difficult to associate this apparent trend with regulatory actions, it appears to
confirm improved public health protections over time, at least for the larger undisinfected
systems.

See Appendix A for similar tables that present results separately for the following size
categories: systems serving fewer than 1,000 people; systems serving 1,000 people or more but
less than 10,000 people; and systems serving 10,000 people or more (Exhibit A-3, Exhibit A-4,
and Exhibit A-5, respectively). Note that in Section 6.4, EPA describes the use of a Bayesian
MCMC statistical approach for analysis of undisinfected ground water systems.

6-17


-------
Exhibit 6-7. Changes in Percent of GW Systems with Disinfection ("Common

Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Common Systems

42.822

42.822

42.822

42.822

42.822

42.822

42.822

42.822

#Disinfecting
Systems

20,067

18,740

21,494

20,091

21,384

20,789

22,729

22,769

#Non-disinfecting
Systems

22,755

24,082

21,328

22,731

21,438

22,033

20,093

20,053

%disinfecting
Systems

46.9%

43.8%

50.2%

46.9%

49.9%

48.6%

53.1%

53.2%

%disinfecting

Systems
(Ave @ 2 years)

45.3%
(Before GWR)

48.6%
(Right after GWR)

49.2%
(Right after SS)

53.1%

(After few years of RTCR)

Relative Change

7.2%

(Right after GWR)

1.4%
(Right after SS)

7.9%

(Ater few years of RTCR)

Overall Change

17.24%

EPA then evaluated the total coliform positive rate and the E. call positive rate (as a percentage
of total coliform samples) for the common ground water systems. As shown in Exhibit 6-8 and
Exhibit 6-9, the percent of total coliform positives and E. coli positives consistently decreased
over the years.

Exhibit 6-8. Changes of Percent RTTC+ Rates among All Ground Water Systems
("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

42,822

42,822

42,822

42,822

42,822

42,822

42,822

42,822

#RTTC

524,701

568,901

565,792

573,023

568,412

572,193

566,225

566,448

#RTTC+

10,915

11,212

10,751

10,414

9,243

9,660

7,868

7,691

%RTTC+

2.1%

2.0%

1.9%

1.8%

1.6%

1.7%

1.4%

1.4%

Periods

Period 1:
Right before GWR

Period 2:
Right after GWR

Period 3:

Right after completion of
Sanitary Survey under GWR

Period 4:
Few years after RTCR

Ave @ 2 years

2.0%

1.9%

1.7%

1.4%

Relative difference



-8.2%
(Right afterGWR)

-10.9%

(Right after Completion
of Sanitary Survey)

-17.1%
(After RTCR)



Overall Difference

-32.2%

6-18


-------
Exhibit 6-9. Changes of Percent RTEC+ Rates among All Ground Water Systems
("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

42,822

42,822

42,822

42,822

42,822

42,822

42,822

42,822

#RTTC

524,701

568,901

565,792

573,023

568,412

572,193

566,225

566,448

#RTEC+

492

485

441

409

360

316

341

294

%RTEC+

0.09%

0.09%

0.08%

0.07%

0.06%

0.06%

0.06%

0.05%

Ave @ 2 years

0.09%

0.07%

0.06%

0.06%

Relative difference



-16.6%

-20.6%

-5.4%



Overall Difference

-37.4%

EPA also broke out the assessed total coliform and E. coli detection rates for disinfecting and
undisinfected ground water systems. As indicated in Exhibit 6-10 and Exhibit 6-11, both total
coliform detection rates and E. coli detection rates are apparently 2-3 times higher among
undisinfected systems than among disinfecting systems. Messner et al, 2017 reported that, for
small undisinfected ground water systems, five percent of total coliform detections were E. coli
positive in all system sizes and types.

At disinfecting ground water systems, total coliform detection rates decreased after the GWR
rule implementation and in the periods that followed. A small apparent increase (5 percent) in
total coliform detection rates was observed in undisinfected ground water systems after the initial
implementation of the GWR (which to some extent could be attributable to background yearly
variation) and then there was an apparent decreasing trend after the Sanitary Survey completion
and the RTCR implementation. Overall, Exhibit 6-10 shows an apparent decrease of total
coliform detection rates.

Exhibit 6-10. Changes of Percent RTTC+ Rates among Disinfecting and
Undisinfected Systems ("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Disinfecting Systems

















#RTTC

355,296

374,135

411,575

399,646

417,688

415,336

421,189

421,961

#RTTC+

4,656

4,591

4,817

4,167

3,999

4,322

3,971

3,885

%RTTC+

1.31%

1.23%

1.17%

1.04%

0.96%

1.04%

0.94%

0.92%

Ave @ 2 years

1.27%

1.11%

1.00%

0.93%

Relative difference



-12.79%

-9.72%

-6.73%



Overall Difference

-26.56%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Undisinfected Systems

22,755

24,082

21,328

22,731

21,438

22,033

20,093

20,053

#RTTC

169,405

194,766

154,217

173,377

150,724

156,857

145,036

144,487

#RTTC+

6,259

6,621

5,934

6,247

5,244

5,338

3,897

3,806

%RTTC+

3.69%

3.40%

3.85%

3.60%

3.48%

3.40%

2.69%

2.63%

6-19


-------
Year

2007

2008

2010

2011

2014

2015

2018

2019

Ave @ 2 years

3.55%
(Right before
GWR)

3.73%
(Right after GWR)

3.44%
(Right after
completing SS)

2.66%
(Few years after
RTCR)

Relative difference



5.03%

-7.63%

-22.68%



Overall Difference

-24.99%

Exhibit 6-11 presents changes to E. coli positive rates among disinfecting and undisinfected
systems. At disinfecting systems, E. coli detection rates apparently decreased after GWR
implementation and in the period that followed, and then apparently increased a few years after
RTCR implementation. An apparent decrease of E. coli detection rate was observed in
undisinfected systems across the four periods. As noted earlier, E. coli records that did not have a
corresponding total coliform positive record were not excluded from the reduced dataset. The
total number of E. coli samples at common systems across the 8 years where there was no
corresponding total coliform positive record was approximately 0.20 percent of the total. Thus,
not excluding these E. coli records will probably not have a significant effect on the observed
trends.

Exhibit 6-11. Changes of Percent RTEC+ Rates among Disinfecting and
Undisinfected Systems ("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Disinfecting Systems

















#RTTC

355,296

374,135

411,575

399,646

417,688

415,336

421,189

421,961

#RTEC+

217

217

221

186

158

148

181

172

%RTEC+

0.06%

0.06%

0.05%

0.05%

0.04%

0.04%

0.04%

0.04%

Ave @ 2 years

0.06%

0.05%

0.04%

0.04%

Relative difference



-15.82%

-26.71%

13.99%



Overall Difference

-29.68%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Non Disinfecting Systems

22,755

24,082

21,328

22,731

21,438

22,033

20,093

20,053

#RTTC

169,405

194,766

154,217

173,377

150,724

156,857

145,036

144,487

#RTEC+

275

268

220

223

202

168

160

122

%RTEC+

0.16%

0.14%

0.14%

0.13%

0.13%

0.11%

0.11%

0.08%

Ave @ 2 years

0.15%

0.14%

0.12%

0.10%

Relative difference



-9.55%

-11.12%

-19.23%



Overall Difference

-35.07%

EPA assessed the annual average percent of total coliform positives and E. coli positives for
three groups of systems: all PWSs, all disinfecting ground water systems, and all undisinfected
ground water systems. Exhibit 6-12 presents a summary of the results for total coliforms. Similar
to the findings in Exhibit 6-8 and Exhibit 6-10 above, there is generally an apparent declining
trend after each regulatory milestone, with the exception of the undisinfected ground water
systems from 2007-2010. Overall, there was an apparent reduction of the percent total coliform
positive rate after the collective implementation of the GWR and RTCR. For E. coli positives, as

6-20


-------
shown in Exhibit 6-13 (in part summarizing results presented in Exhibit 6-11), EPA observed an
apparent declining trend after each of regulatory milestones except the last. Overall, there was
more than an apparent 35 percent reduction in E. coli positives in undisinfected ground water
systems after implementation of GWR and RTCR. Because the total number of E. coli detections
is small, there is high uncertainty associated with apparent changes over time. These apparent
changes will be examined further, using 95 percent confidence intervals, in the subsection that
follows, titled "Assessing the Changes in Total Coliform and E. coli Occurrence with All
Available Records."

Exhibit 6-12. Summary of Changes of Percent RTTC+ Rates by System Categories
(All Public Water Systems, Disinfecting Ground Water systems, Undisinfected

Ground Water Systems)

System Types

Year

2007

2008

2010 2011

2014

2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right after SS under
GWR

After few years of
RTCR

All PWSs

Relative Change



-8.8%

-8.5%

-16.3%



Overall Change

-30.2%

Disinfecting
GW systems

Relative Change



-12.8%

-9.7%

-6.7%



Overall Change

-26.6%

Undisinfected
GW Systems

Relative Change



5.0%

-7.6%

-22.7%



Overall Change

-25.0%

Exhibit 6-13. Summary of Changes of Percent RTEC+ Rates by System Categories
(All Public Water Systems, Disinfecting Ground Water Systems, Undisinfected

Ground Water Systems)

System Types

Year

2007

2008

2010 2011

2014

2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right after SS under
GWR

After few years of
RTCR

All PWSs

Relative Change



-19.2%

-10.1%

2.8%



Overall Change

-25.3%

Disinfecting

Relative Change



-15.8%

-26.7%

14.0%



Overall Change

-29.7%

Undisinfected
GW Systems

Relative Change



-9.6%

-11.1%

-19.2%



Overall Change

-35.1%

Assessing the Changes in Total Coliform and E. coli Occurrence with All Available Records

To evaluate the degree to which the occurrence of total coliform positives and E. coli positives
may have changed from year to year using undisinfected ground water systems, EPA utilized 95
percent confidence intervals2. If the confidence interval for one year and the confidence interval
for another year do not overlap, one can conclude with 95 percent certainty that there was an
actual decline (or increase) between the two years (i.e., the difference is "statistically

2 Upon considering various uncertainties in the underlying data such as weather, state policies, changes made by
systems, incomplete census of systems, etc. EPA utilized a 95 percent confidence interval analysis.

6-21


-------
significant"). EPA notes the following caveats about this analysis: Unlike analyses presented
earlier in the document that involved a "90% completeness" filter, all data from undisinfected
ground water systems were included in this analysis without any exclusions. The number of
systems with total coliform / E. coli records and, thus, the number of routine total coliform and
E. coli records, varied from year to year. The number of states included in each year also varied.
Furthermore, systems may have different degrees of completeness with respect to rule
requirements from one month to the next.

Exhibit 6-14 presents the annual total coliform positive rates for undisinfected ground water
systems from 2012 through 2019. There is a confidence interval overlap between each pair of
consecutive years, so the analysis does not show a statistically significant change from year to
year. But over the entire period (comparing 2012 figures to 2019 figures), there is a statistically
significant decline in total coliform positive results. The period of the statistically significant
decline in total coliform positives encompasses the period of GWR implementation (2012-2016)
and RTCR implementation (2014-2019).

Exhibit 6-24 and Exhibit 6-25 compare 2011 and 2019 total coliform detections in undisinfected
systems using Bayesian MCMC statistical models. In contrast to the statistical test described
above, the statistical model results (using the maximum likelihood estimate) do not support the
apparent conclusions from the summary statistics. Rather, the model results show that in general,
depending on system size and type, the smallest systems (large numbers of systems with
maximum likelihood estimates that have low uncertainty) have small increases or decreases
between the two years analyzed. Observing results using the two differing definitions for
undisinfected systems, it appears that the definition of an undisinfected system has a greater
effect on the maximum likelihood estimate than the comparison between years.

Exhibit 6-15 presents the annual E. coli positive rates for undisinfected ground water systems
from 2012 through 2019. As with the total coliform results, there is a confidence interval overlap
between each pair of consecutive years, so the analysis does not show a statistically significant
change from year to year. Furthermore, even the first and last year of the E. coli positive rates
data have overlapping confidence intervals, so no conclusions can be drawn with statistical
significance about trends in E. coli positive rates over the span of years from 2012 to 2019.

Exhibit 6-14. Total Coliform Positive Rates for Undisinfected Ground Water

Systems (2012-2019)

Year

RTTC+

RTTC

%RTTC+

STD

CI Lower

CI Upper

2012

2,568

73,859

3.48%

0.000674

3.345%

3.609%

2013

2,619

74,610

3.51%

0.000674

3.378%

3.642%

2014

2,672

74,527

3.59%

0.000681

3.452%

3.719%

2015

2,688

75,389

3.57%

0.000675

3.433%

3.698%

2016

2,347

75,386

3.11%

0.000633

2.989%

3.237%

2017

2,298

76,560

3.00%

0.000617

2.881%

3.122%

6-22


-------
Year

RTTC+

RTTC

%RTTC+

STD

CI Lower

CI Upper

2018

2,321

78,860

2.94%

0.000602

2.825%

3.061%

2019

2,285

78,970

2.89%

0.000596

2.777%

3.010%

Notes:

RTTC+ = Number of systems reporting positive total coliform
RTTC = Number of systems reporting routine total coliform samples
%RTTC+ = Percent of systems reporting positive total coliform
STD = Standard deviation

CI Lower Bound = Confidence interval lower bound
CI Upper Bound = Confidence interval upper bound

Exhibit 6-15. E. coli Positive Rates for Undisinfected Ground Water Systems

(2012-2019)

Year

RTEC+

RTTC

%RTEC+

STD

CI Lower

CI Upper

2012

75

73,859

0.10%

0.000117

0.079%

0.125%

2013

75

74,610

0.10%

0.000116

0.078%

0.123%

2014

112

74,527

0.15%

0.000142

0.122%

0.178%

2015

99

75,389

0.13%

0.000132

0.105%

0.157%

2016

81

75,386

0.11%

0.000119

0.084%

0.131%

2017

75

76,560

0.10%

0.000113

0.076%

0.120%

2018

108

78,860

0.14%

0.000132

0.111%

0.163%

2019

83

78,970

0.11%

0.000115

0.083%

0.128%

Notes:

RTTC+ = Number of systems reporting positive total coliform
RTTC = Number of systems reporting routine total coliform samples
%RTTC+ = Percent of systems reporting positive total coliform
STD = Standard deviation

CI Lower Bound = Confidence interval lower bound
CI Upper Bound = Confidence interval upper bound

6.2.2 Analytical Results of Level 1 and Level 2 Assessments Under Revised Total
Coliform Rule

Under the RTCR, a PWS whose sampling results show that it is vulnerable to microbial
contamination is required to conduct an assessment and take corrective action. They must
identify and correct any sanitary defects in the distribution system or treatment processes. Under
the RTCR, EPA uses total coliform occurrence as an indicator of the microbial integrity of the
distribution system, and E. coli as an indicator of the presence of fecal contamination. Exhibit
6-16 presents a summary of the assessment process if a Level 1 or Level 2 Assessment is
triggered.

6-23


-------
Exhibit 6-16. Summary of Level 1 and Level 2 Assessment Processes

What to do if you triggered an assessment?

Within 30 days of learning that your PWS triggereo an assessment, a completed state assessment form must be
submitted to your state. The process for completing and submitting the required form depends on the type of
assessment. In both cases, your state will review the completed assessment form to determine if the likely cause of
the trigger has been identified and to ensure the problem is corrected.

Level 2 Assessment

You have to do a level 2 Assessment if you have either:

1. £. coli MCL violation:

Level 1 Assessment

You have to do a Level 1 Assessment if you:

1.	Fail to collect and analyze at least 3 repeat
samples for each routine TC+; or

2.	Have two or more TC+ samples (use routine and
repeat results in your calculation) in one month.

Routine

Repeat

TC+ & EC-

£. co/i-positive (EC+)

TC+ & EC-

TC+ but not analyzed for EC

TC+& EC+

TC+

TO & EC+

One or more samples is missing

Your system conducts the assessment.

ir

2. Two Level 1 triggers in a rolling 12-month period or for
systems on annual monitoring, a Level 1 trigger in two
consecutive years.

Your state approves the party that will conduct the
assessment.

Further details about Level 1 and Level 2 assessments are found in Section 7.3.1 and Appendix
C.

To evaluate the impact of Level 1 and Level 2 assessments on total coliform positive / E. coli
positive rates, EPA compared total coliform positive rates before the completion of Level 1 and
Level 2 assessments and after the completion of corrective actions. EPA utilized RTCR Level 1
and Level 2 Assessment "event milestone" information available from SDWIS for the years
2016 through 2019 (i.e., after RTCR became effective) for systems with available total coliform
and E. coli data from the SYR4 ICR dataset. EPA recognized that systems' RTCR compliance
monitoring schedules may be monthly, quarterly, biannual, or yearly, and that most of the
systems with the relevant data records in SYR4 ICR are systems with monthly compliance
monitoring schedules. Thus, the analysis described here was focused on these systems. Exhibit
6-17 presents counts of routine total coliform and E. coli samples from systems with Level 1 or
Level 2 Assessments, counts of total coliform positive and E. coli positive routine samples, and
the total coliform and E. coli positivity rates, for the months before and after a Level 1 or 2
assessment. Exhibit 6-17 also presents the percent reduction in the total coliform positive rate
from the two months before the Level 1 or Level 2 Assessment to the two months after the Level
1 or Level 2 Assessment. The analysis makes use of 2-month averages, to reduce the effect of
month-to-month background variability.

The analysis shows a remarkable drop in total coliform and E. coli levels after corrective actions
were implemented. Overall, there was more than an 80 percent decrease in both total coliform
positives and E. coli positives after completion of RTCR assessments at systems having a
monthly monitoring schedule. It may be noted that some systems in some states could take
longer than two months to complete the corrective actions, and in such situations the positivity

6-24


-------
rates two months after the assessment would represent the period before corrective actions were
fully completed. Thus, the findings could be different from state to state.

Exhibit 6-17. Total Coliform and E. coli Positivity Rate for Months Before and After
Level 1 or 2 Assessment (90% Completeness Applied1)

Two months before L1 or L2
(e.g., April & May for the Assessment
month of June)

Two months after L1 or L2
(e.g., August & Sept.)

%Reduction in
%total coliform
positives (based on
2-month averages)

#RTTC

#RTTC+

%RTTC+

#RTTC

#RTTC+

%RTTC+

69,333

15,023

21.67%

63,064

2,175

3.45%

84.08%



#RTTC

#RTEC+

%RTEC+

#RTTC

#RTEC+

%RTEC+

% Reduction in %£.

coli positives
(based on 2-month
averages)

69,333

656

0.95%

63,064

77

0.12%

87.10%

1 For this analysis, EPA only included systems with at least 90% completeness in their total coliform sampling results for the "months
before" and the "months after," meaning included systems must have collected at least 90% of their required monthly total coliform
samples.

Notes:

L1 = Level 1 Assessment
L2 = Level 2 Assessment

#RTTC = number of systems reporting routine total coliform samples

#RTTC+ / #RTEC+ = number of systems reporting positive total coliform / E. coli samples

%RTTC+ / %RTEC+ = percent of systems reporting positive total coliform / E. coli samples

6.3 Microbial Contaminants in Raw Water under Long-Term 2 Enhanced Surface Water
Treatment Rule

Under the LT2, systems monitored their water sources to determine whether additional treatment
is needed for further removal of Cryptosporidium. This monitoring included an initial two years
of monthly sampling for Cryptosporidium in source water. To reduce monitoring costs, small,
filtered water systems (serving fewer than 10,000 people) first monitored for coli—a
bacterium that is less expensive to analyze for than Cryptosporidium—and then monitored for
Cryptosporidium if their coli results exceeded specified concentration levels. For more
information on source water monitoring provision, refer to US EPA (2006c).

6.3.1 Occurrence of Cryptosporidium in Source Water

EPA conducted an analysis of Cryptosporidium occurrence using compliance monitoring data
from SYR4 that reflect Round 2 monitoring results under LT2. Among records of raw water
samples (routine only) from surface water CWSs, about 9 percent reported that Cryptosporidium
was present. However, none of those 9 percent of records that reported the presence of
Cryptosporidium provided a concentration value. In the SYR4 dataset more broadly, fewer than
0.3 percent of the 19,000 Cryptosporidium monitoring records provided concentration levels
with units of oocysts/L.

6-25


-------
EPA examined available data to assess the accuracy of predictions for binning changes for a
group of systems from LT2 Round 1 to Round 2 monitoring. Under SYR3, Round 1 monitoring
data were available from the Data Collection and Tracking System (DCTS); these data are
described in the Six-Year Review 3 Technical Support Document for Long-Term 2 Enhanced
Surface Water Treatment Rule (USEPA, 2016b). EPA conducted a comparison of data from
systems with SYR4 Cryptosporidium binning information to data from systems with SYR3
Cryptosporidium binning information from the DCTS. Note that the DCTS data were limited to
systems identified as being in Bin 2. The bin classifications are described in Exhibit 6-18.

Exhibit 6-18. Bin Classification for Filtered Systems

Mean
Cryptosporidium
Concentration
(oocysts/L)

Bin

Classification

Additional Cryptosporidium Treatment Required

Alternative
Filtration

Conventional
Filtration

Direct
Filtration

Slow Sand or

Diatomaceous

Earth

< 0.075 oocysts/L

Bin 1

No additional
treatment

No additional
treatment

No additional
treatment

No additional
treatment

>0.075 to < 1.0
oocysts/L

Bin 2

1 -log treatment

1.5-log
treatment

1 -log treatment

(1)

>1.0 to < 3.0
oocysts/L

Bin 3

2-log treatment

2.5-log
treatment

2-log treatment

(2)

>3.0 oocysts/L

Bin 4

2.5-log
treatment

3-log treatment

2.5-log
treatment

(3)

1	As determined by the state (or other primacy agency) such that the removal/inactivation > 4.0-log.

2	As determined by the state (or other primacy agency) such that the removal/inactivation > 5.0-log.

3	As determined by the state (or other primacy agency) such that the removal/inactivation > 5.5-log.

Data provided in response to the SYR4 ICR showed that there were 309 systems serving
>10,000 people that provided binning results based on Cryptosporidium monitoring (presumed
Round 2 monitoring). Thirty-two (32) of those systems had at least some records in Bin 2. Only
one of those 32 systems appeared in the SYR3 "DCTS Binning Report" file. EPA determined
that approximately 10 percent of the systems serving >10,000 people (i.e., 31 out of 309
systems) would potentially move to an action bin (i.e., Bin 2, Bin 3, or Bin 4) based on Round 2
data. Under this analysis, EPA determined that the percentage of PWSs moving to an action bin
based on Cryptosporidium monitoring in Round 2 was not significantly higher than the 8 percent
of systems predicted to have moved to an action bin during the previous LT2 review.

Note that EPA did not examine E. coli co-occurrence with Cryptosporidium because of low
positive rates.

6.4 Analyses Involving Undisinfected Ground Water Systems

Similar to SYR3, EPA conducted an analysis to evaluate the possible differences in coliform
occurrence between disinfecting and undisinfected ground water systems as part of SYR4.

6.4.1 Approaches to Identify Undisinfected Ground Water Systems

For the SYR3 data analysis, disinfecting systems were defined in the following manner: Of
systems with SYR3 total coliform data that were evaluated, any of those systems with an

6-26


-------
indication of disinfecting per the SYR3 treatment data were considered to be disinfecting.
Remaining systems with free or total chlorine residual concentrations greater than 0.1 mg/L were
also considered to be disinfecting. Remaining systems were considered to be undisinfected. This
classification approach to differentiate systems can be called the "SYR3 definition of
disinfecting." See Appendix D of USEPA (2016a) for more details.

For the SYR4 ICR data analysis, a revised classification method was used to identify disinfecting
and undisinfected systems, as described in this section. This classification approach to
differentiate systems can be called the "SYR4 definition of disinfecting."

In SYR3, the focus was on distinguishing different levels of disinfection rather than specifically
identifying undisinfected systems. The purpose of the SYR3 approach was to understand if there
was an opportunity to balance disinfection byproduct risk with pathogen control with respect to
disinfectant residual level changes.

During the SYR4 process, EPA attempted to maximize the use of the available SYR4 ICR data
records for determining the disinfection status of systems included in the SYR4 ICR datasets by
making logical inferences from information pertaining to regulatory requirements. Exhibit 6-19
and Exhibit 6-20 illustrate the steps taken to achieve this purpose, using as an example the final
year (2019) of data records. The considerations involved are explained in the paragraphs that
follow.

Since the routine disinfectant by-product monitoring provisions under the Stage 1 and Stage 2
DBPRs apply to all of CWSs and NTNCWSs (including both GW and SW systems) that use
chemical disinfectants, two major steps were taken (See Exhibit 6-19): First, EPA considered all
of those GW CWSs and NTNCWSs reported any monitoring records of DBPs (i.e.,
TTHM/HAA5) in 2019 as disinfecting GW systems (assuming that the sole use of non-chemical
disinfectants such UV is rare). Second, among systems without any DBP records, EPA
considered any with disinfectant residual records to be disinfecting GW systems as well. With
this systematic approach, EPA was able to quantify the number of disinfecting versus
undisinfected GW systems in the SYR3 and SYR4 ICR datasets for the individual years3 shown
in Exhibit 6-10 and evaluate their yearly trends. (Note: Systems with no SYR3 or SYR4 data
could not be used in any of the analyses to evaluate trends in total coliform / E. coli occurrence;
however, systems could be classified as disinfecting or undisinfected based on SDWIS treatment
information (or lack thereof) for overall national counts of disinfecting systems, as shown in the
lower branch of the flowchart.)

Exhibit 6-20 shows how the process was adapted for transient non-community water systems
(TNCWSs). Since the routine disinfectant by-product monitoring provisions of the Stage 1 and
Stage 2 DBPRs do not apply to all TNCWSs (including both GW and SW systems) that use
chemical disinfectants (only TNCWSs using chlorine dioxide are required to monitor DBPs), the
step 1 described above was skipped.

3 The disinfection status of systems was identified separately for each year of data to enable year-to-year trends.
Exhibit 6-19 and Exhibit 6-20 show the process using the calendar year 2019 as an example.

6-27


-------
There were differences in the two definitions of undisinfected ground water systems considered
for use by EPA at different times. The SYR3 approach was included to enable a comparison with
the analytical results presented on the Messner et al. (2017) paper. The SYR4 approach focuses
on the inclusion of DBP data as a primary indicator of disinfection, incorporates disinfectant
residual records, and treatment information from SDWIS. Additionally, the data used for SYR4
ICR data analysis contained non-SDWIS state data while the SYR3 assessment in Messner et al.
(2017) did not. These differences may account for why the SYR4 approach identifies more
undisinfected ground water systems overall than are identified when using the SYR3 approach.

Exhibit 6-19. Process to Identify Community Water Systems and Non-Transient
Non-Community Water Systems that are Undisinfected Ground Water Systems

(Using Year 2019 as Example)

6-28


-------
Exhibit 6-20. Process to Identify Transient Non-Community Water Systems that
are Undisinfected Ground Water Systems (Using Year 2019 as Example)

AIIGWTNCWSs in
2019 in Nation

l/

All GWTNCWSs
Included in
2019SYR4ICR

All GWTNG
Not Include*
2019SYR411

National count of
Undisinfected systems
and pop served for GW
TNCWSs

6.4.2 Modeling Total Coliform Positivity Rates
6.4.2.1 Statistical Techniques

The analyses of total coliform / E. coli data presented above involved the use of summary
statistics to compute positivity rates. This approach (which is generally referred to as the
"frequentist" or "classical" statistical approach) has the advantage of being relatively
straightforward computationally. That is, the positivity rate is simply a proportion calculated as
the number of positives samples observed divided by the total number of samples taken. While
some characterization of uncertainty around that proportion can be obtained by computing
confidence intervals, the correct interpretation of those confidence intervals can be confusing in
that they do not conform to the common sense interpretation of confidence intervals, namely that
the "true" value of the proportion has a specified likelihood (e.g., 95%) of falling within the
confidence interval.4

An alternative statistical approach for obtaining estimates of positivity rates from data such as
the SYR4 data set is a Bayesian analysis using Markov Chain Monte Carlo (MCMC) simulation
methods. While computationally more involved than the summary statistics approach, the
Bayesian MCMC approach has the advantage of providing multiple estimates of the proportion
of positive samples, from which one can compute an overall mean estimate and also "credible
intervals" around that mean estimate to characterize uncertainty. The Bayesian credible intervals
also have an advantage over the frequentist confidence intervals in that they can be interpreted to
infer that the "true" value does have a specified likelihood (e.g., 95%) of falling within the range
of the interval.

4 The correct interpretation of a frequentist confidence interval is that if one were to perform multiple sampling of
the same number of samples from that population, then 95% of those confidence intervals would include the true
value.

6-29


-------
A Bayesian analysis was conducted by Messner et al. (2017) using data from SYR3; this work
serves as a model of how such an analysis could be performed to better inform the goals of the
SYR4 microbial data analysis (i.e., to evaluate the possible differences in total coliform
occurrence between disinfecting and undisinfected ground water systems, to assess the potential
impact of the GWR and RTCR on the occurrence of microbial indicators, and to characterize the
systems with the highest potential for public health improvements).

Messner et al. (2017) used the Bayesian MCMC method to estimate the positivity rate of total
coliform detections in routine sampling at small (serving <4,100 people) undisinfected ground
water systems for the year 2011, using data from SYR3. That Bayesian MCMC analysis
involved the estimation of parameters for beta distributions from which the mean positivity rates
could be derived. Uncertainty around those mean estimates was also obtained and displayed
graphically as scatter plots around those mean estimates (see Exhibits F.2 - F.5 in Appendix F
from USEPA (2016a)). Similarly, an expanded Bayesian MCMC analysis could be conducted
using SYR4 data to more specifically address the three goals of SYR4 microbial data analysis
stated in the preceding paragraph.

6.4.2.2 Markov Chain Monte Carlo Modeling Results

Systems that use undisinfected ground water may benefit the most from public health
improvements due to the lack of any disinfection barrier. As part of the data analysis conducted
under SYR3, routine total coliform / E. coli sample data from undisinfected small ground water
systems were analyzed to characterize the PWSs with high total coliform detection rates
(Messner et al., 2017). For SYR4, this analysis was repeated using data from the SYR4 ICR. The
goal was to investigate patterns and differences in the total coliform / E. coli positivity rates of
undisinfected small ground water systems in the SYR3 and SYR4 ICR datasets. The results
section below first presents the modeling output from the SYR3 analysis as an introduction to the
format of the output and then describes the results gathered from the analysis performed under
the SYR4 effort as well as a comparison of results using differing definitions (flow charts) for
identifying undisinfected systems in 2011 and 2019 data

Data Analysis

Limited total coliform data for the large number of small PWSs taking monthly or quarterly total
coliform samples are available from the SYR4 ICR. This makes it challenging to precisely
estimate the detection rate at a given system. Under SYR3, EPA used a Bayesian statistical
analysis approach to estimate mean total coliform detection rates. The probability distribution
was estimated using a likelihood function of the available data. One advantage of using the
Bayesian approach to determine the mean total coliform detection rate with Markov Chain
Monte Carlo (MCMC) simulations based on the available data is that the mean total coliform
positive detection rates of the different PWS groups being modeled can be compared without the
necessity of having identical sample sizes or common systems when comparing between
different years of the same system population served sizes and types. A more detailed
explanation of the MCMC modeling approach is outlined in Appendix F of USEPA (2016a).

Messner et al. (2017) analyzed 2011 data from approximately 38,000 small (serving fewer than
4,101 individuals) undisinfected public water systems. Their statistical modeling results showed

6-30


-------
that the smallest undisinfected systems have significantly higher total coliform detection rates,
with low uncertainty in the estimate. Exhibit 6-22 shows, from Messner et al. 2017, the total
coliform detection rates for the three types of undisinfected systems. For the system type with the
greatest numbers of undisinfected systems (transient, non-community), the graph is annotated
with values to illustrate the total coliform detection magnitude. This annotated curve reports 25%
of undisinfected systems had total coliform detection rates of at least 5% (8% of transient
systems had total coliform detection rates greater than 15%) (Exhibit 6-22). In this document,
EPA calculates that these percentages translate to about 7000 systems having detection rates of
5% or more (Exhibit 6-26 and Exhibit 6-27).

Total coliform positivity rates modeled in Messner et al. (2017) for TNCWSs serving
populations of 25-100, 101-1,000, and 1,001-4,100 are shown in Exhibit 6-21. Similar modeling
was performed for CWSs and NTNCWSs. Exhibit 6-21 shows results of 10,000 simulations,
each providing an estimate of the total coliform detection rate for each of the three population
served groups. The maximum likelihood estimate for the total coliform detection rate is shown as
a small circle in the center of each of three "clouds" of "star" points. The spread in the "cloud"
displays the uncertainty in the maximum likelihood estimate of the total coliform detection rate.
Tightly grouped clouds have low uncertainty. The total coliform detection rate fields (e.g. 2%,
3%>, 4%) is shown by the vertical dashed lines. Note that the tight clouds almost completely
overlap so that the two population size clouds have very similar, low uncertainty, total coliform
detection rates.

6-31


-------
Exhibit 6-21. Markov Chain Monte Carlo Samples Predicting Total Coliform
Detection Rates in Small Undisinfected Transient Non-Community Water Systems
(2011 Data, Six-Year Review 3 Definition of Disinfecting)

t

1%	2%	3%	4%

-5.0	4.5	4.0	-3.5	-3.0

u = ln(a!pha / beta)

Source: USEPA (2016a)

Using the resulting scatter plot output, EPA determined the mean total coliform positivity rate
and the parameters (u and v) that describe the beta distribution associated with the mean. With
the mean and beta distribution parameters, EPA derived the probability distribution curve for the
detection rates within the beta distribution and thus counts of systems with specific total coliform
positive sample rates in routine sampling, as shown for systems in the smallest size category (25-
100) in Exhibit 6-22.

6-32


-------
Exhibit 6-22. Detection Rate Distribution Functions for Small Public Water
Systems (serving 25-100) based on Markov Chain Monte Carlo Sample Mean
Parameter Values (2011 Data, Six-Year Review 3 Definition of Disinfecting)

Routine TC Detection Rate

Source: USEPA (2016a)

Definitions of "Disinfecting" Systems

As discussed at the beginning of Chapter 6 and in Section 6.4.1, two definitions of disinfection
status have been developed. The process used for determining disinfection status in SYR4
(shown in the flowcharts in Exhibit 6-19 and Exhibit 6-20) is not the same as the process used in
SYR3 (USEPA, 2016a).

An example of the difference made by the choice of definition is displayed in Exhibit 6-25,
which offers a comparison of 2011 and 2019 total coliform maximum likelihood estimate (MLE)
detection rates using the two processes for identifying undisinfected NTNCWS systems. Using
2019 data (i.e., the graphs on the right-hand side of the exhibit), the total coliform MLE for the
smallest systems (serving 25-100 people), at the center of the green "cloud," was about 3.2%
using the SYR3 definition (the upper graph) and about 2.7% using the SYR4 definition (the
lower graph). The differences produced by choice of definition is less marked for other
undisinfected PWS categories. The horizontally tight statistical model results in the cloud around
the MLE show that there was low uncertainty associated with determining these specific MLE
values. Uncertainty is low because the available data included results from a very large number
of systems.

To confirm that the SYR4 definition of disinfecting applied to the SYR3 data yields the same
results and maintains the conclusions drawn in the Messner et al. (2017) paper, the SYR4 and

6-33


-------
SYR3 datasets were classified using both the SYR3 and SYR4 definitions of disinfecting to
create four different datasets: SYR3 data using the SYR3 definition of disinfecting, SYR3 data
using the SYR4 definition of disinfecting, SYR4 ICR data using the SYR3 definition of
disinfecting, and SYR4 ICR data using the SYR4 definition of disinfecting. MCMC modeling
was performed on all four variations of the data.

Average Total Coliform Detection Rate Comparison for Different System Sizes

The MCMC modeling was conducted by using as inputs the count of routine total coliform
samples and the count of these that were positive. These sample counts were generated
separately for several groupings, based on system population size category (systems serving 25-
100, 101-1,000, and 1,001-4,100 people), system type (CWSs, TNCWSs, and NTNCWSs), and
year (2011-2019). The output of the MCMC simulation modeling is a scatter plot of the
simulated mean total coliform positivity rates of the PWS subsets, with the sample mean total
coliform positivity rate in the center of the cluster.

Two different versions of the routine total coliform sampling data from 2011 and 2019 were
created: one using data from undisinfected systems that conform to the SYR3 definition of
disinfecting, and the other using data from undisinfected systems that conform to the SYR4
definition of disinfecting. These results were then compared to confirm that the conclusions
drawn in SYR3 do not need to be changed as a result of the additional analyses performed in
SYR4.

Accounting for the two different definitions of disinfecting systems and the data available for
each year, the MCMC total coliform results shown in Exhibit 6-23 are available.

Exhibit 6-23. List of Data Sources and Disinfection Definition for Markov Chain

Monte Carlo Model Runs

MCMC TC Results for Non-Disinfecting PWSs

Year(s)

Definition

PWSs

Data Source

2005

RTCR

60,000

EPA RTCR EA, 2012

2011

SYR3

38,000

Messneret al, 2017

2019

SYR3

28,000

Six Year4, 2023 based on Messner et al, 2017

2011-2019

SYR4

45,000

Six Year 4, 2023

In the scatter plots that follow (Exhibit 6-24 for small undisinfected TNCWs and Exhibit 6-25
for small undisinfected NTNCWSs), the size categories are color-coded: "small" systems
(serving 1,001-4,100) are represented by black points, "very small" systems (serving 101-1,000)
are represented by pink points, and "very, very small" systems (serving 25-100) are represented
by green points. These system sizes are also abbreviated as "S," "VS1," and "VS2," respectively.

Each point shown in the scatter plots is the results from one iteration of the model. The clouds of
points show the distribution of the MCMC modeled samples; the larger the spread of the points,
the greater the uncertainty. The black clouds are more spread out than the others because there

6-34


-------
aren't many systems that fall into this category, so the uncertainty is relatively high. The "S"
(black) plots represent a few hundred systems while the "VS1" (pink) and "VS2" (green) plots
represent approximately 30,000 systems. The Maximum Likelihood Estimate (MLE) of all
simulations for a system group is represented by a yellow dot in the center of the respective
cloud of points. The value of the MLE (interpreted as a projected total coliform detection rate) is
indicated by the vertical lines in the plot area.

For undisinfected transient-non-community systems (Exhibit 6-24), the modeling projects that
"VS2" and "VS1" systems had more than two times higher MLE total coliform positivity rates
than the "S" systems in 2011, regardless of definition of "undisinfected" used. Using 2019 data,
the MLEs for "S" and for "VS2" and "VS1" are closer together, using both definitions. Under
both definitions, the MLE for the "S" systems increased from 2011 to 2019 and the MLEs for the
"VS2" and "VS1" systems decreased.

6-35


-------
Exhibit 6-24. Maximum Likelihood Estimates of National Total Coliform
Occurrence using Six-Year Review 3 and Six-Year Review 4 Definitions for Small
Undisinfected Transient Non Community Water Systems

C

0

01

O

m
—

(z

£
X

J)

c
o

-}) = 0.03
in{Pr{+|) = 0.032

-5.0	4.5	4.0

u = ln(alpha / beta)

The plots in the top row represent data that were analyzed using the SYR3 definition of "disinfecting" and the plots
in the bottom row represent data that were modeled using the SYR4 definition of "disinfecting." The plots on the
left use data from 2011 and the plots on the right use data from 2019.

For undisinfected non-transient-non-community systems (Exhibit 6-25), the modeling projects
that in 2011 the MLE total coliform positivity rates ranged from slightly under 2% (for "S"
systems) to slightly over 3% (for "VS2" systems), with "VS1" systems in between, regardless of
the definition of "undisinfected" used. Using 2019 data, "VS2" and "VS1" systems have more
than two times higher MLE projected total coliform positivity rates than the "S" systems under
the SYR4 definition of disinfecting. Under the SYR3 definition of disinfecting, the 2019 MLEs
are closer together, without much change from 2011.

6-36


-------
Exhibit 6-25. Maximum Likelihood Estimates of National Total Coliform
Occurrence using for Small Undisinfected Non-Transient Non-Community Water

Systems

C

O

01

O

m
—

(z

£
X

J)

2011

Parameters for 2011 NTNC

H VS2, u =[-3.46, v = 1.72,[mean{Pr{+j) = 0.03

-5.5	-5.0	4.5	4.0	-3.5	-3.0

u = ln(alpha / beta)

2019

u = ln(alpha / beta)

c
o

0)

o

2

X

'J5

2011

Parameters for 2011 NTNC

S, u = -4.b2, v = 1.38, mtan(Pr{+}) =| 0.018 j
VS1, u =1-3.71, v = 1.93,imean{Pr{+J) = 0.024
B VS2, u =[-3.37, v = 1.71 ,[mean{Pr{+j) = 0.033

-5.5	-5.0	4.5	4.0	-3.5	-3.0

u = ln(alpha / beta)

2019

Parameters for 2019 NTNC

¦ S, u=-4.fe9, v = 3.01, mtan{Pr{+})=|0.01 j
i	¦ VS1, u =i-3.79, v = 2.53,intean{Pr{+}) = 0.022

] + 4. +O VS2, u = [-3.64, v = 1.91, [mean{Pr{+j) = 0.026

-5.5	-5.0	4.5	4.0	-3.5	-3.0

u = ln(alpha / beta)

The plots in the top row represent data that were analyzed using the SYR3 definition of "disinfecting" and the plots in the bottom
row represent data that were modeled using the SYR4 definition of "disinfecting." The plots on the left use data from 2011 and
the plots on the right use data from 2019.

Comparing the MCMC cloud plot outputs using the SYR3 definition of disinfecting (the top row
in Exhibit 6-24 and Exhibit 6-25) to the MCMC cloud plot outputs using the SYR4 definition of
disinfecting (the bottom row in Exhibit 6-24 and Exhibit 6-25) shows that MLE projected total
coliform positivity rates are, on the whole, very similar. EPA finds that the conclusions drawn in
the SYR3 analysis (Messner et al., 2017) do not need to be changed as a result of the adoption of
a new process for identifying undisinfecting systems developed in SYR4.

Average total coliform positive rates, as well as high total coliform positive rates in
approximately 7,000 vsl and vs2 undisinfected systems show an imbalance between small and
vsl/vs2 systems (Exhibit 6-24 and Exhibit 6-25). In one study, low average total coliform
positive rates (e.g., 2.5 percent) in small CWSs are thought to result in a 22 percent acute
gastrointestinal illness (AGI) attributable risk to drinking water from norovirus and enterovirus
(Borchardt et al., 2012). QMRA based on pathogens in undisinfected Minnesota drinking waters
show populations with high infection rates (Stokdyk et al., 2019; Stokdyk et al., 2020; Burch et
al., 2022). Note that these Wisconsin and Minnesota papers are representative of national disease

Parameters for 2019 NTNC

- ¦ S, u = [-3.97, v =
CI VS1, u = -3.59, v =
~ VS2, J = -3.44, v =

6-37


-------
burden in so far as undisinfected ground water systems in these two states constitute about 23
percent of undisinfected ground water systems nationally.

High Detection Rate Systems

From the MCMC modeling, the number of systems with total coliform positive sample rates
above a given percentage can be estimated. EPA developed counts of systems with total coliform
positive rates greater than 5 percent using the beta distributions for each of nine groups, based on
system size ("S," "VS1," and "VS2") and system type (abbreviated below as "C" for CWS,
"NC" for TNCWS, and "NTNC" for NTNCWS). Comparing results using the 2011 data (Exhibit
6-26) to results using the 2019 data (Exhibit 6-27), EPA finds that there has not been a
significant change in the count of systems with total coliform positive rates greater than 5
percent; the projected number of the TNCWSs and NTNCWSs serving less than 1001 persons
with total coliform positive rates greater than 5 percent is over 7000 in 2019, only slightly
higher than in 2011. The MCMC modeling projects no detection rates higher than 30 percent.
These findings were generated using the SYR4 definition of undisinfected systems.

Using the SYR4 ICR dataset, applying the SYR3 definition of undisinfected systems, EPA finds
that the counts of systems with total coliform positive rates greater than 5 percent do not change
significantly, though increase marginally, between 2011 and 2019.

Exhibit 6-26 Count of Systems by Size and Type with Total Coliform Positive
Rates >5% (2011 Data; SYR4 Definition of Disinfecting)

System Size	System Type

1 ] c

NC I

NTNC

|S (1001-4100)

43.82



13.98

Ivsi (101-1000) 1

356,27



235.74

[vS2 (25-100) [

429.22

4gog 61j

327.21

Exhibit 6-27. Count of Systems by Size and Type with Total Coliform Positive
Rates >5% (2019 Data; SYR4 Definition of Disinfecting)

System Size	System Type



C

NC

NTPJt I

S (1001-4100)

31.22

25.50

4.08

VS1 (101-1000)

295.91

2079.62



VS2 (25-100)

302.89

5067.64

329.62

Based on the findings from Messner et al. (2017) and the more recent findings presented in this
section, it appears that smaller PWSs (those serving less than 1,000 people) have significantly
higher average total coliform detection rates than the larger systems (serving 1,001-4,100).
While up to about 7,000 small (VS1 and VS2) undisinfected PWSs (non-transient and transient

6-38


-------
systems serving less than 1,001 people) have detection rates above 5 percent per year (5 percent
total coliform positive rates in a month triggers a Level 1 assessment per the RTCR), the total
coliform positive rates of these undisinfected PWSs do not exceed 30 percent per year. Average
total coliform positive rates in undisinfected PWSs have remained static or changed slightly
(increased or decreased) since 2011.

Caveats

The MCMC modeling was only conducted to evaluate the occurrence of total coliforms. There
were insufficient detection data for E. coli to support MCMC modeling of that contaminant. In
addition, EPA used the SYR4 definition of undisinfected ground water systems rather than the
SYR3 definition of undisinfected ground water systems for this recent modeling effort.

6.4.3 Analytical Results of Triggered Source Water Monitoring under Ground Water
Rule

Under the Ground Water Rule, if ground water systems that do not provide at least 4-log
treatment of viruses are notified of a routine total coliform positive sample collected in
compliance with the Revised Total Coliform Rule (RTCR) they must collect at least one source
water sample for E. coli from each ground water source (well) within 24 hours. This source
(well) sample is referred to as a "triggered source water sample." Results in the SYR4 ICR
database with a sample type code of "TG" (triggered) were evaluated for their /•]. coli positive
rate. Results are shown in Exhibit 6-28 for undisinfected ground water systems. Note that the
method used to identify undisinfected ground water systems is described in Section 6.4.1.

Overall, 270 triggered source water samples (1.42 percent) collected from undisinfected ground
water systems between 2012 and 2019 were E. coli positive. When evaluated by system type, the
E. coli positive rate was highest in TNCWSs. The rates of E. coli positives in CWSs,

NTNCWSs, and TNCWSs were 0.57 percent, 0.76 percent, and 1.82 percent, respectively.

EPA broke results down by system size, as shown in Exhibit 6-28. The highest E. coli positive
rate, and also the highest absolute count of E. coli positive samples, were found in the smallest
size category (i.e., systems serving <100 people). In this size category, 183 E. coli positive
samples were reported (representing two thirds of E. coli positive samples in water source), or
1.57 percent of samples from all systems in that size category.

When the results are broken down by system size and type, they show that in all three system
types (CWS, NTNCWS, TNCWS), the highest rate of E. coli positives was found in one of the
three smallest size categories (<100, 101-500, 501-1,000). Among all system categories, the
highest rate of E. coli positives (3.33 percent) was found in TNCWSs serving between 501 and
1000 people.

6-39


-------
Exhibit 6-28. Six-Year Review 4 Information Collection Request - Summary of E.
coli Results in Undisinfected Ground Water Systems Collected as Triggered

Source Water Samples (2012-2019)



Population
Served Size
Category

Undisinfected GW Systems

System Type

Number
of

Systems

Number
of E. coli
Samples

Number E.
coli
positive
Samples

Percent E.
coli
positive
Samples



<100

614

1,724

13

0.75%



101-500

458

1,723

14

0.81%



501-1,000

83

396

0

0.00%

Community

1,001-4,100

95

693

0

0.00%

Water Systems

4,101-33,000

14

163

0

0.00%



33,001-100,000

0

0

0

0.00%



>100,000

0

0

0

0.00%



Total GW

1,264

4,699

27

0.57%



<100

355

812

7

0.86%



101-500

283

657

4

0.61%



501-1,000

58

167

2

1.20%

Non-Transient
Non-Community
Water Systems

1,001-4,100

21

68

0

0.00%

4,101-33,000

0

0

0

0.00%

33,001-100,000

0

0

0

0.00%



>100,000

0

0

0

0.00%



Total GW

717

1,704

13

0.76%



<100

4,348

9,111

163

1.79%



101-500

1,437

3,092

54

1.75%



501-1,000

126

300

10

3.33%

Transient Non-

Community
Water Systems

1,001-4,100

48

151

3

1.99%

4,101-33,000

2

5

0

0.00%

33,001-100,000

0

0

0

0.00%



>100,000

0

0

0

0.00%



Total GW

5,961

12,659

230

1.82%



<100

5,317

11,647

183

1.57%



101-500

2,178

5,472

72

1.32%



501-1,000

267

863

12

1.39%

All system types

1,001-4,100

164

912

3

0.33%

4,101-33,000

16

168

0

0.00%



33,001-100,000

0

0

0

0.00%



>100,000

0

0

0

0.00%



Total GW

7,942

19,062

270

1.42%

6-40


-------
Regarding source water E. coli detections, there are three different ways to interpret the data: (1)
raw E. coli detections can represent E. coli sampling at seasonal systems, (2) E. coli sampling
following a total coliform positive from routine sampling under RTCR at systems that do not
have distribution systems, or (3) E. coli detections that may be follow up samples under RTCR
intended to be distribution system samples but are effectively source water samples because the
system does not have a distribution system (as is the case at some small systems). Therefore, it is
difficult to draw clear conclusions about the outcome of triggered source water monitoring other
than to say that even if all the raw E. coli detections were a result of the GWR Triggered Source
Water requirement, there is a small percent of positive detections and it appears that the vast
majority of E. coli detections were associated with undisinfected systems serving less than 1,000
people.

6.5 Analyses of Aircraft Drinking Water Rule

The ADWR dataset contains samples from a variety of system sizes, types, and aircraft
manufacturer/models. Exhibit 6-29 through Exhibit 6-31 provide a summary of the types of
information contained in the ADWR data. (Note that here and throughout this document
"system" is used in place of "aircraft public water system").

A count of total coliform samples broken down by system size for the years 2012-2019 is
presented in Exhibit 6-29. Approximately 6 percent of the samples were collected from large
systems, 53 percent from medium systems, and 41 percent from small systems. Large systems
represent 5 percent of the total systems with data; medium and small systems represent 55
percent and 40 percent of the total systems, respectively.

Exhibit 6-29. Count of Total Coliform Samples and Aircraft Systems by Size; 2012-

2019

Size

Seat Cap

Count of
Samples

Percent
of Total
Samples

Count of
Systems

Percent
of Total
Systems

Large

>250

7,003

5.93%

415

5.31%

Medium

>130 -
250

62,454

52.90%

4,323

55.31%

Small

<=130

48,613

41.17%

3,078

39.38%

Total

118,070

100.00%

7,816

100.00%

A breakdown of total coliform samples by size and aircraft manufacturer/model for the years
2012-2019 is presented in Exhibit 6-30. The average number of samples per size and
manufacturer/model category was 1,663, with a median of 94 and a maximum of 28,282. The
average number of systems per size and manufacturer/model category was 110, with a median of
13 and a maximum of 1,954.

6-41


-------
Exhibit 6-30. Count of Total Coliform Samples and Systems by Size, Aircraft

Manufacturer and Model; 2012-2019

Size

Manufacturer, Model

Count of Total

Coliform

Samples

Count of
Systems

Large

AIRBUS, 330

10

1

Large

AIRBUS, A330

1,724

82

Large

AIRBUS, A350

58

13

Large

BOEING, 747

606

42

Large

BOEING, 767

1,271

58

Large

BOEING, 777

3,072

180

Large

BOEING, 787

222

34

Large

BOEING, B747

14

1

Large

DOUG, DC1030

0

0

Large

DOUG, MD11

14

3

Large

EMB, ERJ170

12

1

Med

um

AIRBUS, 320

2,028

122

Med

um

AIRBUS, 321

20

2

Med

um

AIRBUS, A220

12

6

Med

um

AIRBUS, A231

10

1

Med

um

AIRBUS, A319

1,254

109

Med

um

AIRBUS, A320

7,502

521

Med

um

AIRBUS, A321

3,448

344

Med

um

AIRBUS, A330

306

12

Med

um

BOEING, 737

28,282

1,954

Med

um

BOEING, 747

30

2

Med

um

BOEING, 757

7,288

463

Med

um

BOEING, 767

3,825

215

Med

um

BOEING, 777

34

1

Med

um

BOEING, 787

522

54

Med

um

BOEING, B737

468

63

Med

um

BOEING, B757

8

1

Med

um

BOEING, DC982

4

1

Med

um

BOEING, DC983

2

1

Med

um

DOUG, DC950

6

1

Med

um

DOUG, DC982

946

118

Med

um

DOUG, DC983

1,515

131

Med

um

DOUG, MD83

280

18

Med

um

DOUG, MD88

3,066

118

Med

um

DOUG, MD90

482

23

Med

um

DOUG, MD9030

1,116

42

Small

ACE, 123

2

1

Small

ACE, 737

4

1

Small

ADAMS, 7897

16

1

Small

AIRBUS, 319

886

55

Small

AIRBUS, A220

48

22

Small

AIRBUS, A318

14

4

Small

AIRBUS, A319

3,934

212

Small

AIRBUS, A321

1,094

116

Small

BOEING, 707

4

1

6-42


-------
Size

Manufacturer, Model

Count of Total

Coliform

Samples

Count of
Systems

Small

BOEING, 717

2,601

111

Small

BOEING, 737

1,705

132

Small

BOEING, 757

94

12

Small

BOEING, 767

34

4

Small

BOEING, 777

16

1

Small

BOEING, B737

24

3

Small

BOMBDR, BD100

1,683

71

Small

BOMBDR, CL6002

17,070

933

Small

BOMBDR, CRJ900

0

0

Small

BOMBDR, DHC8402

356

33

Small

BOMBDR, Q400

2

1

Small

CNDAIR, CL6002

1,836

101

Small

DOUG, DC915

4

2

Small

DOUG, DC931

2

1

Small

DOUG, DC932

2

1

Small

DOUG, DC934

2

1

Small

DOUG, DC950

150

20

Small

DOUG, DC983

2

1

Small

DOUG, DC987

8

2

Small

DOUG, MD88

80

3

Small

EMB, 140

20

3

Small

EMB, EMB135

1,039

102

Small

EMB, EMB145

6,182

447

Small

EMB, EMB175

12

3

Small

EMB, ERJ170

8,203

584

Small

EMB, ERJ190

1,484

93

Total

118,070

7,816

A breakdown of total coliform samples by sample type (e.g., "Routine," "Repeat") for the years
2012-2019 is presented in Exhibit 6-31. Approximately 84 percent of the total coliform samples
were identified as routine.

Exhibit 6-31. Count of Aircraft Total Coliform Samples by Sample Type; 2012-2019

Sample Type

Number of

Total
Coliform
Samples

Percent of All
Total
Coliform
Samples

Routine

99,677

84.42%

Repeat

94

0.08%

Follow-up

11,001

9.32%

Special

7,298

6.18%

Total

118,070

100.00%

6-43


-------
6.5.1 Occurrence of Total Coliforms and E. coli in Aircraft Systems

Exhibit 6-32 through Exhibit 6-35 present summaries of total coliform and E. coli occurrence
from the ADWR data. Additional tables and information are included in Appendix B.

A breakdown of total coliform and E. coli samples and positivity rates by year for the years
2012-2019 is presented in Exhibit 6-32. The overall total coliform positivity rate was 5.46
percent, with a median annual rate of 5.6 percent, a minimum annual rate of 3.8 percent, and a
maximum annual rate of 7.0 percent. The annual total coliform positivity rate tended to decrease
over the 2012-2019 period.

For E. coli samples, two approaches were used to evaluate the positivity rate. One approach used
the E. coli samples present in the data set, while under the "alternative approach," additional
inferred data were added to the analysis: when there was a total coliform "Absent" and no
corresponding E. coli record, EPA assumed that the missing E. coli result was also "Absent."
Under the first approach, the average E. coli positivity rate was 0.26 percent and the median
annual E. coli positivity rate was also 0.26 percent, with a minimum of 0.17 percent and a
maximum of 0.33 percent. Regardless of approach, theE. coli positivity rate generally decreased
over the years, but this trend was less pronounced and less consistent than the decreasing trend
observed for total coliform positivity rate.

Exhibit 6-32. Count of Aircraft Total Coliform and E. coli Samples, and Total
Coliform and E. coli Positives by Year; 2012-2019

Year

Total Coliforms

E. coli

E. coli (Alternative Approach)1

Total
Samples

#

Positive

% Positive

Total
Samples

#

Positive

% Positive

Total
Samples

#

Positive

% Positive

2012

14,707

1,034

7.03%

10,283

33

0.32%

14,702

33

0.22%

2013

14,996

892

5.95%

9,493

31

0.33%

14,996

31

0.21 %

2014

15,658

890

5.68%

9,845

17

0.17%

15,656

17

0.11%

2015

15,436

861

5.58%

9,186

24

0.26%

15,433

24

0.16%

2016

15,823

933

5.90%

9,512

30

0.32%

15,823

30

0.19%

2017

13,648

651

4.77%

9,672

21

0.22%

13,647

21

0.15%

2018

13,903

665

4.78%

10,058

26

0.26%

13,900

26

0.19%

2019

13,899

522

3.76%

10,065

19

0.19%

13,899

19

0.14%

Total

118,070

6,448

5.46%

78,114

201

0.26%

118,056

201

0.17%

1 Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

A breakdown of total coliform and E. coli sample counts and positivity rates by system size is
presented in Exhibit 6-33. For total coliforms, small systems had a positivity rate nearly three
times higher than for medium systems and more than four times higher than large systems. A
comparison of total coliform positive rates for small aircraft with similar information for small
stationary PWSs (transient non-community) shows that both aircraft and stationery PWSs had
total coliform positive rates for small systems that were more than two times higher than for

6-44


-------
larger systems and that total coliform positive rates for both aircraft and stationary PWSs
generally declined over the 8-year period (2012-2019) of SYR4 ICR data.

For E. coh, using the standard approach, small systems had a positivity rate more than two times
higher than that of medium systems and more than four times higher than that of large systems.
Using the alternative approach, small systems had a positivity rate more than three times higher
than that of medium systems and more than nine times higher than that of large systems.

Exhibit 6-33. Aircraft Total Coliform and E. coli Sample Count and Positivity Rate,

by Size; 2012-2019





Total Coliforms

E. coli

E. coli (Alternative Approach)1

Size

Seat Cap

Total

#

%

Total

#

%

Total

#

%





Samples

Positive

Positive

Samples

Positive

Positive

Samples

Positive

Positive

L

>250

7,003

138

1.97%

2,569

2

0.08%

7,002

2

0.03%

M

>130-250

62,454

1,897

3.04%

37,600

59

0.16%

62,450

59

0.09%

S

<=130

48,613

4,413

9.08%

37,945

140

0.37%

48,604

140

0.29%

Total

118.070

6.448

5.46%

78.114

201

0.26%

118.056

201

0.17%

1 Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

A breakdown of total coliform and E. coli sample counts and positivity rates by sample location
(galley vs. lavatory) is presented in Exhibit 6-34. Similar information is presented in Appendix
B, Exhibit B-4, broken down by year. Note that Exhibit B-4 also includes data for the years
2011, 2020, and 2021. The total coliform positivity rate for lavatory samples was approximately
eight times higher than for galley samples. Exhibit B-4 shows that annual total coliform
positivity rates for lavatory samples tended to decrease during the period of interest. That trend
was not as apparent for total coliform positivity rates in galley samples.

The E. coli positivity rate for lavatory samples was five times higher than for galley samples
under the standard approach, and more than five times higher under the alternative approach.
There was little to no observable decreasing trend in E. coli positivity rate in galley or lavatory
samples for either set of E. coli assumptions during the period of interest.

The total coliform positivity rates for galley samples from small aircraft water systems (1-2
percent, varying by year) were slightly lower than the total coliform positivity rates for small
stationary PWSs (2-3 percent in 2019, the most recent year of data on stationary small TNCWSs
serving <1,000 people).

6-45


-------
Exhibit 6-34. Aircraft Total Coliform and E. coli Sample Count and Positivity Rate,

by Location; 2012-2019

Location

Total Coliforms

E. coli

E. coli (Alternative Approach)1

# Samples

# Total
Coliform
Positive

% Total
Coliform
Positive

# Samples

# E. coli
positive

% E. coli
positive

# Samples

# E. coli
positive

% E. coli
positive

Galley

54,277

635

1.17%

34,512

27

0.08%

54,275

27

0.05%

Lavatory

63,793

5,813

9.11%

43,602

174

0.40%

63,781

174

0.27%

Total

118,070

6,448

5.46%

78,114

201

0.26%

118,056

201

0.17%

1 Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

Exhibit 6-35 presents total coliform and E. coli sample counts and positivity rates for follow-up
samples, broken down by air carrier. Note that only data for air carriers with at least one follow-
up sample are presented. For total coliforms, the average carrier-specific positivity rate for
follow-up samples was 9.2 percent. The median was 7.8 percent, with a minimum of 0 percent
and a maximum of 34.6 percent. The 90th percentile rate was approximately 21 percent.

For E. coli, using the standard approach, the average carrier-specific positivity rate for follow-up
samples was 0.28 percent. The median and minimum were both 0 percent, and the maximum was
7.9 percent. The 90th percentile rate was approximately 0.36 percent. Using the alternative
approach, the average E. coli positivity rate for follow-up samples was 0.07 percent and the 90th
percentile rate was approximately 0.34 percent.

Exhibit 6-35. Aircraft Total Coliform and E. coli Sample Counts and Positivity Rate
for Follow-up Samples, by Air Carrier; 2012-2019

Carrier1

Total Coliforms

E. coli

E. coli (Alternative Approach)2

# Follow-

Up
Samples

#

Positive
Follow-

Up
Samples

%

Positive

#

Follow-

Up
Samples

#

Positive
Follow-

Up
Samples

%

Positive

#

Follow-

Up
Samples

#

Positive
Follow-

Up
Samples

%

Positive

AIR WISCONSIN

AIRLINES

CORPORATION

394

47

11.93%

393

1

0.25%

394

1

0.25%

AIRTRAN AIRWAYS
INC

16

0

0.00%

16

0

0.00%

16

0

0.00%

ALASKA AIRLINES INC

142

11

7.75%

141

0

0.00%

142

0

0.00%

ALLEGIANT AIR LLC

76

5

6.58%

76

0

0.00%

76

0

0.00%

AMERICAN AIRLINES
INC

1,412

102

7.22%

1,142

5

0.44%

1,412

5

0.35%

AMERISTAR AIR
CARGO INC

8

2

25.00%

8

0

0.00%

8

0

0.00%

ATLAS AIR INC

16

1

6.25%

14

0

0.00%

16

0

0.00%

CHAUTAUQUA
AIRLINES INC

158

40

25.32%

134

0

0.00%

158

0

0.00%

COLGAN AIR INC

2

0

0.00%

2

0

0.00%

2

0

0.00%

6-46


-------
Carrier1

Total Coliforms

E. coli

E. coli (Alternative Approach)2

# Follow-

Up
Samples

#

Positive
Follow-

Up
Samples

%

Positive

#

Follow-

Up
Samples

#

Positive
Follow-

Up
Samples

%

Positive

#

Follow-

Up
Samples

#

Positive
Follow-

Up
Samples

%

Positive

COMAIR INC

26

9

34.62%

26

0

0.00%

26

0

0.00%

COMPASS AIRLINES
LLC

142

34

23.94%

142

0

0.00%

142

0

0.00%

DELTA AIR LINES INC

730

37

5.07%

38

3

7.89%

730

3

0.41 %

ENDEAVOR AIR INC

888

113

12.73%

888

3

0.34%

888

3

0.34%

ENVOY AIR INC

424

73

17.22%

366

0

0.00%

424

0

0.00%

EXPRESSJET
AIRLINES INC

1,510

287

19.01%

1,510

2

0.13%

1,510

2

0.13%

FALCON AIR EXPRESS
INC

2

0

0.00%

2

0

0.00%

2

0

0.00%

FRONTIER AIRLINES
INC

30

0

0.00%

30

0

0.00%

30

0

0.00%

GOJET AIRLINES LLC

150

0

0.00%

150

0

0.00%

150

0

0.00%

HAWAIIAN AIRLINES
INC

20

0

0.00%

14

0

0.00%

20

0

0.00%

JETBLUE AIRWAYS
CORPORATION

290

21

7.24%

290

0

0.00%

290

0

0.00%

MESA AIRLINES INC

176

7

3.98%

167

1

0.60%

176

1

0.57%

MIAMI AIR

INTERNATIONAL INC

18

2

11.11%

18

0

0.00%

18

0

0.00%

OMNI AIR

INTERNATIONAL INC

8

1

12.50%

8

0

0.00%

8

0

0.00%

PIEDMONT AIRLINES
INC

18

0

0.00%

18

0

0.00%

18

0

0.00%

PSA AIRLINES INC

222

26

11.71 %

157

0

0.00%

222

0

0.00%

REPUBLIC AIRWAYS
INC

294

31

10.54%

150

0

0.00%

294

0

0.00%

SHUTTLE AMERICA
CORPORATION

102

14

13.73%

39

0

0.00%

102

0

0.00%

SKYWEST AIRLINES
INC

2,407

366

15.21%

2,228

7

0.31 %

2,406

7

0.29%

SOUTHWEST
AIRLINES CO

234

6

2.56%

234

0

0.00%

234

0

0.00%

SPIRIT AIRLINES INC

96

7

7.29%

24

0

0.00%

96

0

0.00%

SUN COUNTRY
AIRLINES

54

0

0.00%

54

0

0.00%

54

0

0.00%

SWIFT AIR LLC

4

0

0.00%

4

0

0.00%

4

0

0.00%

TEM ENTERPRISES
INC

6

1

16.67%

6

0

0.00%

6

0

0.00%

TRANS STATES
AIRLINES LLC

184

37

20.11%

184

0

0.00%

184

0

0.00%

UNITED AIRLINES, INC

412

34

8.25%

267

0

0.00%

412

0

0.00%

US AIRWAYS INC

290

25

8.62%

96

1

1.04%

290

1

0.34%

VIRGIN AMERICA INC

36

3

8.33%

36

0

0.00%

36

0

0.00%

VISION AIRLINES INC

2

0

0.00%

2

0

0.00%

2

0

0.00%

WORLD AIRWAYS INC

2

0

0.00%

2

0

0.00%

2

0

0.00%

1	Only the counts for carriers with at least one follow-up sample are presented in this table

2	Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

6-47


-------
7 Treatment

This chapter summarizes the results from Environmental Protection Agency's (EPA's) Six-Year
Review 4 (SYR4) of new information related to the treatment of microbial contaminants in
drinking water. For this SYR4, EPA conducted a scientific review of available information,
published in or before December 2021, to determine if new information has the potential to
present a meaningful opportunity to revise treatment technique (TT) requirements.

This chapter provides a brief overview of major TT requirements in the microbial contaminant
regulations, provides a description of recent regulatory implementation impacts for rules covered
by the scope of SYR4, and highlights new technical information that has become available since
SYR3. Additional background about the technical basis of the Ground Water Rule (GWR) is
provided in Chapters 3 and 7 of the Six-Year Review 3 Technical Support Document for
Microbial Contaminant Regulations (USEPA, 2016a), while the basis of Long Term 2 Enhanced
Surface Water Treatment Rule (LT2) is covered in Chapters 3 and 7 of the Six-Year Review 3
Technical Support Document for Long-Term 2 Enhanced Surface Water Treatment Rule
(USEPA, 2016b).

Information in this chapter is organized as follows:

•	Section 7.1 presents treatment information on the LT2

•	Section 7.2 presents treatment information on the GWR

•	Section 7.3 presents treatment information on the Revised Total Coliform Rule (RTCR)

•	Section 7.4 presents treatment information on the Aircraft Drinking Water Rule (ADWR)

Overall, the treatment information presented and discussed in this chapter are intended to be
helpful for addressing one of the questions prescribed in the EPA Protocol for the Fourth Review
of Existing National Primary Drinking Water Regulations (USEPA, 2024a, see Section 2 for
more detail as well): Is there a significant improvement in analytical or treatment feasibility?

7.1 Long Term 2 Enhanced Surface Water Treatment Rule

EPA promulgated the LT2 on January 5, 2006 to increase protection against microbial
pathogens, specifically Cryptosporidium, in public water supplies that use surface water sources.
This section presents a summary of literature that has become available since 2015 and key new
information related to each LT2 microbial toolbox option's effectiveness and implementation.

7.1.1 Description of Long Term 2 Enhanced Surface Water Treatment Rule
Requirements

The purpose of the LT2 is to improve public health protection by reducing illness linked to
Cryptosporidium and other microbial contaminants in drinking water and focusing on systems
with elevated Cryptosporidium risk. The LT2 defined a range of additional treatment

7-1


-------
requirements for inactivation of Cryptosporidium and built on pre-existing filtration
requirements for Subpart H water systems that practice conventional or direct filtration:

•	The Surface Water Treatment Rule (SWTR) requires 99.99 percent (4-log) removal for
viruses and 99.9 percent (3-log) removal and/or inactivation for Giardia lamblia

•	The Interim Enhanced Surface Water Treatment Rule (IESWTR) and Long Term 1
Enhanced Surface Water Treatment Rule (LT1) require 99 percent (2-log) removal of
Cryptosporidium

The LT2 required surface water and ground water under the direct influence of surface water
(GWUDI) systems to perform two different rounds of source water monitoring for a period of
one to two years for Cryptosporidium and/or /•]. coli. Large water systems were required to
monitor their source water for Cryptosporidium. Smaller systems serving fewer than 10,000
people could monitor for E. coli unless the E. coli levels exceeded a trigger level, at which point
they would then be required to conduct Cryptosporidium monitoring. Source water monitoring
was not required for filtered systems that provided or intended to install 5.5-log of treatment for
Cryptosporidium and unfiltered systems that provided or intended to install at least 3-log
treatment for Cryptosporidium.

Unfiltered systems were required to also monitor source water for Cryptosporidium
concentration and calculate a mean concentration to determine the appropriate treatment
requirements: 2- or 3-log inactivation of Cryptosporidium.

Filtered systems were classified into one of four "bins" that defined additional Cryptosporidium
treatment requirements based on the system's source water monitoring results. The lowest
treatment bin, Bin 1, has no additional treatment requirements. The bins specify the additional or
total required Cryptosporidium log treatment based upon the type of filtration treatment:
conventional treatment; direct filtration; slow sand or diatomaceous earth filtration; or alternative
filtration technologies. For conventional filtration treatment, the range was no additional
treatment to 2.5 log treatment, while the range for direct filtration was no additional treatment to
3-log additional treatment. In general, treatment requirements for Bin 2 through Bin 4 varied
from 1-log additional treatment to 5.5-log total Cryptosporidium treatment based upon source
water concentration and filtration type
[40 CFR 141.711(a)],

To meet the Cryptosporidium treatment requirements for each bin classification, water systems
were required to select from a "toolbox" of treatment or management options that prescribe the
amount of log treatment credit applicable to each tool, as listed in Exhibit 7-1. The Long Term 2
Enhanced Surface Water Treatment Rule Toolbox Guidance Manual describes the treatment and
management strategies that are necessary for implementing each toolbox option (USEPA,
2010b). Utilities are provided flexibility to perform a site-specific demonstration if an additional
credit is sought. EPA provided the Membrane Filtration Guidance Manual for clarification of
design and implementation of use of membrane filtration as an LT2 tool (USEPA, 2005c).

Systems must prove that they are meeting operational or performance criteria to receive toolbox
option credit. Systems that already use ozone, chlorine dioxide, UV light or membranes in
addition to conventional treatment prior to the promulgation of the LT2, can receive LT2 toolbox

7-2


-------
credit if they meet the performance criteria for the chosen technology. Systems currently using
chlorine or chloramine do not receive Cryptosporidium inactivation credits for these
disinfectants under the LT2.

Exhibit 7-1. Cryptosporidium Treatment Credits for all Toolbox Options under
Long Term 2 Enhanced Surface Water Treatment Rule [40 CFR 141.715(b)]

Toolbox Option

Cryptosporidium treatment credit with design and implementation criteria

Source Protection and Management Toolbox Options

Watershed Control
Program

0.5-log credit. Unfiltered systems are not eligible.

Alternative

Source/Intake

Management

No prescribed credit. System-specific case approval by primacy agency

Pre-Filtration Toolbox Options

Pre-sedimentation
basin with coagulation

0.5-log credit during any month that pre-sedimentation basins achieve a monthly
primacy agency-approved performance criteria.

Two-stage lime
softening

0.5-log credit for two stage softening where chemical addition and hardness
precipitation occur in both stages and all plant flow must pass thru both stages.

Bank filtration

0.5-log credit for 25-foot setback;

1.0-log credit for 50-foot setback;

Aquifer must be unconsolidated sand containing at least 10 percent fines; average
turbidity in wells must be less than 1 Nephelometric Turbidity Unit (NTU). Systems
using wells followed by filtration when conducting source water monitoring must sample
the well to determine bin classification and are not eligible for additional credit.

Treatment Performance Toolbox Options

Combined Filter
Performance

0.5-log credit for combined filter effluent turbidity < 0.15 NTU in at least 95% of
measurements each month.

Individual Filter
Performance

0.5-log credit (in addition to 0.5-log combined filter performance credit) if individual filter
effluent turbidity is < 0.15 NTU in at least 95% of samples each month in each filter and
is never greater than 0.3 NTU in two consecutive 15 minute measurements in any filter.

Demonstration of
Performance

Credit awarded to treatment process or treatment train based on a demonstration to
primacy agency

Additional Filtration Toolbox Options

Bag or cartridge filters
(individual filters)

Up to 2-log credit based on the removal efficiency demonstrated during challenge
testing with a 1.0-log factor of safety.

Bag or cartridge filters
(in series)

Up to 2.5-log credit based on the removal efficiency demonstrated during challenge
testing with a 0.5-log factor of safety.

Membrane filtration

Log credit equivalent to removal efficiency demonstrated in challenge test for device if
supported by direct integrity testing.

Second stage filtration

0.5-log credit for second separate granular media filtration stage if treatment train
includes coagulation prior to first filter.

Slow sand filters

2.5-log credit as a secondary filtration step;
3.0-log credit as a primary filtration process.
No prior chlorination allowed for either option.

Inactivation Toolbox Options

Chlorine dioxide

Log credit based on measured CT1 in relation to CT table

Ozone

Log credit based on measured CT in relation to CT table

UV

Log credit based on validated UV dose in relation to UV dose table; reactor validation
testing required to establish UV dose and associated operating conditions.

1 CT is defined as disinfectant residual concentration (C) multiplied by contact time (T). A CT value is a measure of disinfection
effectiveness for the time that microorganisms in the water are in contact with a disinfectant.

7-3


-------
Under the LT2, public water systems with uncovered finished water reservoirs (UCFWR) must
either cover the storage facility or treat the water leaving the storage facility to achieve
inactivation and/or removal of 4-log virus, 3-log Giardia lamblia and 2-log Cryptosporidium
using a protocol approved by the state [40 CFR 141.714] (USEPA, 2006a).

Water systems were required to take measures to cover these reservoirs, treat the water leaving
the reservoirs, replace them with other storage facilities (e.g., ground level storage) or take them
out of service. All PWSs with UCFWRs in the United States are under administrative orders or
compliance agreements to cover or treat their UCFWR.

7.1.2 Advances/Improvements/Innovations to Long Term 2 Enhanced Surface Water
Treatment Rule Microbial Toolbox Requirements

This section presents a summary of literature that has become available since SYR3 and key new
information related to each microbial toolbox option's effectiveness and implementation. The
degree to which implementation issues have been identified varies by the toolbox option.

Exhibit 7-2 provides a summary of relevant new information for the LT2 toolbox options, with
the exception of UV which is discussed in more detail following Exhibit 7-2. For each toolbox
option, the table indicates whether the literature reviewed provided new information on
Cryptosporidium risk reduction or whether there was relevant new design and implementation
information. Overall, EPA has found that there are not any meaningful opportunities to revise the
treatment criteria prescribed in the Long Term 2 Enhanced Surface Water Treatment Rule
Toolbox Guidance Manual (USEPA, 2010b).

Several risk mitigation tools in the LT2 microbial toolbox have become better understood and
implementation improvements have been described in recent literature. Examples of modified
LT2 toolbox technologies include use of bauxite in slow sand filters; ceramic membrane filters
and UV treatment technologies such as light-emitting diode (LED) lamps.

Although some studies showed lower doses of UV treatment required for log inactivation of the
challenge organism male-specific-2 bacteriophage (MS2), others showed higher doses required
to achieve the same log inactivation. Since UV dose outcomes described in this section were
inconsistent in comparison to doses reported in previous EPA guidance, the new information
about MS2 inactivation by UV is considered to support the existing LT2 microbial toolbox
credits and basis of the original rule.

EPA provided new guidance:

•	Drinking Water Instrumentation Data Integrity Checklists (USEPA, 2022a)

•	Guidance Manual for compliance with the Surface Water Treatment Rules: Turbidity
Provisions (USEPA, 2020a)

•	Generating High-Quality Turbidity Data in Drinking Water Treatment Plants to Support
System Optimization and Monitoring (USEPA, 2019c)

7-4


-------
Exhibit 7-2. Potentially Relevant New Studies since 2015 for Existing Long Term 2 Enhanced Surface Water

Treatment Rule Microbial Toolbox Options

Microbial Toolbox Option

New Information for LT2 Microbial Toolbox Options regarding Risk Reduction, Design or Implementation since SYR3

Watershed Control
Program (WCP)

Schijven et al. (2015) provided a computational model to simulate microbial water quality of Cryptosporidium based upon fecal
deposits from wildlife, birds and humans in the floodplain and ground water infiltration.

Moltz et al. (2018) provided a watershed comparison of forest protection and forest buffers on increased drinking water treatment
costs due to changes in microbial water quality.

Ahmed et al. (2019) provided a review of microbial contaminants in stormwater runoff and a summary of log removals achieved
with stormwater design mitigation strategies.

Alternative Source / Intake
Management

EPA found no new information in the literature on this particular tool.

Pre-sedimentation Basin
with Coagulation

EPA found no new information in the literature on this particular tool.

Two-stage Lime Softening

EPA found no new information in the literature on this particular tool.

Bank Filtration (BF)

Berger et al. (2018) found six aerobic spore samples paired as surface water/ground water are sufficient to meet uncertainty
constraints in alluvial aquifers with large volume surface water induced recharge, for use as a surrogate to demonstrate
performance of Cryptosporidium oocyst log reduction. There is no EPA-approved standard method for total aerobic spore assay
in drinking water samples in alluvial aquifers.

Mustafa et al. (2021) found that the correlation of higher well contaminant concentrations due to larger relative stream width can
be neglected when the distance from the pumping well to the nearest river edge is more than twice the stream width.

Low-frequency electromagnetic field (LF-EMF) treatment in a lab setting intended in conjunction with riverbank filtration for the
removal of E. coii indicated removal rates correlated positively with increased strength: 100% E. coii removal at 6, 8, and 10
milliteslas (Selamat et al., 2019)

Oudega et al. (2022) reported the results of a study that involved the attenuation of Bacillus subtilis spores, as a surrogate for
Cryptosporidium and Campylobacter, in a sandy gravel aquifer. The purpose of the study was to estimate required setback
distances for drinking water wells from potential sources of contamination, such as a river. Hydraulic gradients were controlled by
varying the pumping rates in the subsurface at 1 L/s, 5L/s and 10 L/s. Observed removal rates were 0.2 - 0.3 log/m, with higher
removals observed at lower pumping rates. A setback distance of approximately 700 m at the highest pumping rate was
estimated.

Combined Filter
Performance

Schmidt et al. (2020) cautions against misinterpretation and misuse of averaged log-reduction values since these values
characteristically overstate performance that it represents, and recommended use of effective log reduction which averages
reduction and then expresses this as log reduction.

Ramsay et al. (2021) quantified grain displacement during filter backwash and found that grain movement during backwash is
highly inhomogenous in three dimensions and the elapsed time of backwash. Significant displacement of tracer grains in all types

7-5


-------
Microbial Toolbox Option

New Information for LT2 Microbial Toolbox Options regarding Risk Reduction, Design or Implementation since SYR3



of backwash tests, including backwash concluding with subfluidization wash, may affect hydraulic and biological filter function.
Ramsay et al. (2021) also concluded that extended air scours are without significant cleaning value.

The 10 States Standards revised its policy statement for optimization of rapid rate filtration at surface water treatment plants.
(GLUMRB, 2018).

Nix and Taylor (2018) summarized new procedures for granular media filters for addressing filter-clogging algae and for
suspected air binding.

Individual Filter
Performance

Pang et al. (2022) assessed the efficiencies of three different filter media for removal of C. parvum surrogate (using glycoprotein
coated 4.5 |jm polystyrene microspheres) and found that despite the ceramic sand filter achieving log removal values greater
than three consistently, the peak turbidity levels exceeded 0.30 NTU in 17% of the trials, which highlights the need to introduce
supplementary tools alongside turbidity to monitor filter performance more sensitively.

The sensitivity of biologically active filter performance associated with backwash was studied and it was found that the turbidity
spike during ripening of 0.35 NTU significantly improved with the addition of single and double stage extended terminal
subfluidization wash, with maximum turbidity values of 0.14 and 0.09 NTU, respectively (Piche, 2019). Piche also measured the
particle size distribution passing thru biologically active filters and conventional filters, as measured before backwash and after
backwash.

Monis et al. (2017) evaluated Cryptosporidium surrogates for conventional coagulation and dual media filtration and found that all
of the surrogates tested [modified microspheres, spores, high red fluorescent particles (algae), low red fluorescent particles
(bacteria and other material), total particle counts using on-line particle meter, and on-line particle meter count of particles in the
3-6 |jm range] were conservative indicators of oocyst removal with modified microspheres most closely matched oocysts in terms
of removal behavior.

Demonstration of
Performance (DOP)

EPA found no new information in the literature on demonstration of performance.

Bag or Cartridge Filters

Harmsco's HC/90-LT2 cartridge filter was found to reduce Cryptosporidium (using 2 |jm spheres as surrogate) by 3.53 log and
3.72 log in two challenge tests (Harmsco, 2014).

USEPA (2012) provided its generic verification protocol for product-specific challenge testing of full-scale bag and cartridge filters
for Cryptosporidium removal credits, The protocol was developed by the previous EPA's Environmental Technology Verification
(ETV) Drinking Water Systems Center, which is no longer certifying drinking water treatment effectiveness since certifications are
now provided by third-party certification programs.

Membrane Filtration

Chen et al. (2021) conducted a review of 1,060 research papers from the Web of Science database and found that membrane
filtration achieves a broad range of virus removal efficiency from 0.5-7 log removal values.

Ceramic membranes are a type of artificial membrane made from inorganic materials such as alumina, titania, zirconia oxides or
some glassy materials. Pore size can vary but is typically 0.1 |jm. The first water treatment plant using ceramic membranes was
placed into service in 2015 (Kinser, 2021). Since then, the use of ceramic membranes has expanded. Jaferey and Galjaard
(2020) describe benefits of ceramic membranes as not having fiber integrity issues (fiber breakage), easier cleaning and
disinfection, and having higher permeability which equates to lower energy consumption.

7-6


-------
Microbial Toolbox Option

New Information for LT2 Microbial Toolbox Options regarding Risk Reduction, Design or Implementation since SYR3

Although silver nanoparticles can be added to ceramic filters to improve disinfection performance, Abebe et al. (2015) found that
there was no statistical difference between ceramic filtration Cryptosporidium parvum removal efficiency and log mean reduction
due to in vivo silver deactivation. Similarly, Venis and Basu (2021), found that that ceramic water filters with silver only perform
significantly better if there is storage time after filtration in the presence of the silver ceramic filter and they caution of unknown
potential metallic influence on the biofilm layer internal to the filter over the filter's lifespan.

Sharma et al. (2020) summarized advances in nanocellulose filtration technologies comprised of self-standing membranes, thin
film nanofibrous composite membranes and nanocomposite barrier layers on differing scaffold.

Barbhuiya et al. (2021) evaluated the electrochemical antimicrobial and antifouling surface effects of direct current applied to
sulfur-doped laser induced graphene (LIG) filters and found that viral destruction of Vaccinia lister virus at 4-log removal, requires
higher electrical potential due to smaller viral size than bacterial pathogens they previously studied at 6-log removal for
Pseudomonas aeruginosa and mixed bacterial culture (Singh et al., 2018).

Malkoske et al. (2020) reviewed optimal coagulation / flocculation prior to low pressure membrane filtration by comparing
processes with (study Type 3) and without (study Type 2) settling prior to membrane filtration and found accumulated foulants
with settling (study Type 3) may include lower concentrations of hydrolytic coagulants which could result in greater irreversible
fouling.

Patterson et al. (2021) also reviewed membrane manufacturer pilot-scale data to determine allowable flux values for different
influent water quality conditions.

Jacangelo et al. (2019) compared fluorescent dyes for integrity monitoring of reverse osmosis (RO) membranes that could be
employed at full scale to establish and monitor for virus log removal values >3, which could also be extended to protozoa or
bacteria assuming size exclusion as mechanism of removal. Marker-based direct integrity tests are increasingly being approved
by state regulators for nanofiltration (NF) and RO processes (Alspach, 2019).

Vickers (2018) introduced a proposed methodology for establishing pathogen removal credit for RO membranes that are primarily
used for desalination or other applications. The proposed direct integrity testing methodology uses conductivity data to determine
if the RO unit integrity is operating within established limits.

Second Stage Filtration

EPA found no new information in the literature on this particular tool.

Slow Sand Filters

New sand filters by using a water extract of Moringa oleifera (MO) seeds, termed functionalized sand (f-sand) filters can achieve
~7 log MS2 bacteriophage removal (Samineni et al., 2019) and achieve > 8 log removal of E. coii (Xiong et al., 2018).

Slow sand filters with a 30 cm layer of bauxite performed with ~1 year of continuous filtration prior to E. coii breakthrough
represents a significant improvement of the performance of slow sand filters (Urfer, 2017).

Silica columns receiving water dosed with 10 mg/L chitosan coagulant achieved 4.75 log and 4.43 log reductions for E. coii and
MS2, respectively (Holmes, 2019).

7-7


-------
Microbial Toolbox Option

New Information for LT2 Microbial Toolbox Options regarding Risk Reduction, Design or Implementation since SYR3

Chlorine Dioxide

Gallandat et al. (2019) studied the maintenance of disinfectant residual of chlorine dioxide over 24 hours and found that chlorine
dioxide decayed more rapidly in the distribution system across all of the tested conditions and required a dose of 4 mg/L to
maintain a minimum of > 0.2 mg/L except at zero turbidity.

Ozone

Carvajal et al. (2017) showed that change in total fluorescence, a surrogate for dissolved ozone, achieved better fit (at 1 LRV) of
coliforms, C. perfringens spores and somatic coliphages, than the other surrogate measures studied: change in UV254
absorbance; and ozone to Total Organic Carbon (TOC) ratio (03:T0C). This study also cautioned that site-specific analysis
would be more accurate to measure system performance because microbial reductions based upon seeded microorganisms
could lead to overestimation of log credits due to less reduction in autochthonous microorganisms than for seeded
microorganisms.

Wolf et al. (2019) studied proxies to measure virus inactivation by ozone treatment and found that both carbamazepine and UV254
could be used to sufficiently track virus inactivation.

Silva and Sabogal-Paz (2020) studied bench-scale ozone treatments of filter backwash water and found that regardless of
condition no Cryptosporidium oocysts were found in the disinfected samples.

7-8


-------
7.1.2.1 Ultraviolet

New information shows that there have been improvements in UV technology for low- and
medium-pressure lamps and light-emitting diodes. Further, EPA has provided guidance about
validation approaches of UV for drinking water systems and a protocol for state review of UV
disinfection treatment plans.

Since SYR3, EPA has provided the following new guidance documents providing clarification
for ultraviolet treatment technology:

•	The Innovative Approaches for Validation of Ultraviolet Disinfection Reactors for
Drinking Water Systems document described how UV dose monitoring algorithms that
use the combined variable can be used to provide direct predictions of pathogen
inactivation, thereby eliminating the need to apply the reduction equivalent dose (RED)
bias factor, considerably simplifying the validation of UV disinfection. The document
also described a calculated dose approach that does not require an on-line UVT monitor
but calculates log inactivation and RED by the reactor using UV sensor readings, flow
through the reactor, and UV sensitivity of the microbe whose log inactivation and RED is
predicted (USEPA, 2020b).

•	The UV Treatment Toolkit provides a protocol for state review of UV disinfection
treatment plans and templates that address UV design, validation, operations, sensor
calibration, and factors for awarding disinfection credit (USEPA, 2022b). In addition, it
includes a recent update of alternative challenge microorganisms for demonstrating virus
inactivation.

This SYR4 process reviewed new UV literature that has become available since SYR3.
Previously published EPA guidance is shown in Exhibit 7-3 and Exhibit 7-4. Exhibit 7-5 through
Exhibit 7-7 provide a summary of new findings of UV technology results for Cryptosporidium
and the challenge organism type MS2 used as a surrogate with low pressure lamps, medium
pressure lamps and light-emitting diodes, respectively. This new literature regarding UV studies
challenged with Cryptosporidium and MS2 shows inconsistent log inactivation performance
results when compared to doses reported in previous EPA guidance. With some studies
achieving the same log inactivation at doses lower than those reported in previous EPA guidance
and other studies achieving log inactivation at doses higher than those contained in guidance, the
new information is not showing a consensus indication that research is achieving log inactivation
at levels significantly lower than EPA prior published guidance. Additional information
describing the dose outcome of the new articles is included in the discussion of the separate UV
technologies below.

Exhibit 7-3 shows the UV doses required to achieve log inactivation (40 CFR 141.720(d)(1))
while Exhibit 7-4 shows the UV dose sensitivity for challenge microorganisms as reported in the
Ultraviolet Disinfection Guidance Manual for the Final Long Term 2 Enhanced Surface Water
Treatment Rule (USEPA, 2006d).

7-9


-------
Exhibit 7-3. Requirements for Ultraviolet Dose to Achieve Log Inactivation
(millijoules per centimeter squared (mJ/cm2)

T arget
Pathogens

UV Dose
(mJ/cm2)
for

0.5 Log

UV Dose
(mJ/cm2)
for

1.0 Log

UV Dose
(mJ/cm2)
for

1.5 Log

UV Dose
(mJ/cm2)
for

2.0 Log

UV Dose
(mJ/cm2)
for

2.5 Log

UV Dose
(mJ/cm2)
for

3.0 Log

UV Dose
(mJ/cm2)
for

3.5 Log

UV Dose
(mJ/cm2)
for

4.0 Log

Cryptosporidium

1.6

2.5

3.9

5.8

8.5

12

15

22

Giardia

1.5

2.1

3.0

5.2

7.7

11

15

22

Virus

39

58

79

100

121

143

163

186

Source: 40 CFR 141.720(d)(1)

Exhibit 7-4. Ultraviolet Sensitivity of Challenge Microorganisms - Reported
Delivered Ultraviolet Dose to Achieve Log Inactivation

Microorganism

UV Dose
(mJ/cm2) to
Achieve 1-log
Inactivation

UV Dose
(mJ/cm2) to
Achieve 2-

log
Inactivation

UV Dose
(mJ/cm2) to
Achieve 3-log
Inactivation

UV Dose
(mJ/cm2) to
Achieve 4-log
Inactivation

Reference

Bacillus subtilis

28

39

50

62

Sommer et al., 1998

MS2 phage

16

34

52

71

Wilson et al., 1992

QI3 phage

10.9

22.5

34.6

47.6

Mackey et al., 2006

PRD-1 phage

9.9

17

24

30

Meng and Gerba, 1996

B40-8 phage

12

18

23

28

Sommer et al., 1998

4>x174 phage

2.2

5.3

7.3

11

Sommer et al., 1998

E. coli

3.0

4.8

6.7

8.4

Chang et al., 1985

T7

3.6

7.5

11.8

16.6

Mackey et al., 2006

T1

~5

-10

-15

-20

Wright, 2006

Source: Ultraviolet Disinfection Guidance Manual for the Final Long Term 2 Enhanced Surface Water Treatment Rule
(USEPA, 2006d)

This SYR4 review did not find any peer-reviewed articles correlating Cryptosporidium log
inactivation for ultraviolet treatment of drinking water to a new challenge microorganism other
than those types of challenge microorganisms reported in Exhibit 7-4. Some articles presented
ancillary information regarding UV challenge microorganisms as indicators of protozoa log
inactivation. These challenge microorganisms are briefly listed here: Clostridiumperfringens, a
spore forming bacteria used as challenge microorganism for solar ultraviolet treatment
(Gutierrez-Alfaro et al., 2015); and PR772 bacteriophage, a surrogate of adenovirus inactivation
was reported for sequential ultraviolet-chlorine disinfection of wastewater (Gao et al., 2023).
Any use of a new UV challenge microorganism would require reactor-specific validation testing
to establish the UV dose required, since UV inactivation credits for LT2 are not defined by the
regulation based specifically on the challenge microorganism.

Exhibit 7-5 provides a summary of the findings of UV technology results for Cryptosporidium
and MS2 surrogate for low pressure lamps. Studies achieving the same log inactivation at doses
lower than those reported in previous EPA guidance using low pressure lamps included Busse
(2019) and Hull and Linden (2018), while studies of low pressure lamps achieving log

7-10


-------
inactivation at doses higher than those contained in guidance were Baldasso (2021), Younis
(2019), Zyara (2016), and Blatchley (2019).

Exhibit 7-5. Examples of Potentially Relevant Low Pressure Lamp Ultraviolet
Doses for Log Inactivation of Cryptosporidium and Male-Specific-2 Bacteriophage

Wavelength
(nm)

Microorganism/
Surrogate

Medium

UV
dose
(mJ/
cm-')

Log

inactivation

Inactivation
Rate

Constant
(cm'/mJ)

Reference

Study Type

253.7

Cryptosporidium
(S. chilensis)

Phosphate
buffered
saline
(PBS)

7

2.2



Blyth et al.,
2021

Bench scale

254

Cryptosporidium

Water

2

1.1



Busse et al.,
2019

Bench scale

285

MS2

Water

60

3

(interpolated
Table S1 B)



Hull et al.,
2019

Bench scale /
Demo

254

MS2

Turbid
Water

40

(calc)

2

0.05

Baldasso et
al., 2021

Bench scale

254

MS2

Turbid
Water

153
(calc)

4

0.026

Baldasso et
al., 2021

Bench scale

253.7

MS2

Particle
free CaCI2
solution
(2mM)

40

2.8



Feng et al.,
2016

Bench scale

253.7

MS2

Particle
free CaCI2
solution
(200 mM)

40

1.2



Feng et al.,
2016

Bench scale

253.7

MS2

Water

35.7

2

0.0561

Mbonimpa et
al., 2018

Bench scale

254

MS2

Irrigation
Water

82

4



Younis et al.,
2019

Bench scale

253.7

MS2

Dechlorin.
Tap Water

117

3.35



Zyara et al.,
2016

Bench scale

253.7

MS2

PBS

40

2.49 (calc)

0.062

Sholtes and
Linden, 2019

Bench scale

254

MS2

Peptone
buffered
saline

40

2

(interpolated)



Blatchley et
al., 2019

Full scale

255 or 265

MS2

PBS

40

3

(interpolated)



Hull and
Linden, 2018

Bench scale

Exhibit 7-6 provides a summary of the findings of UV technology results for MS2 surrogate for
medium pressure lamps. A study using medium pressure lamps achieving the same log
inactivation at doses lower than those reported in previous EPA guidance included Wang (2019).

Exhibit 7-6. Examples of Potentially Relevant Medium Pressure Lamp
Ultraviolet Doses for Log Inactivation of Male-Specific-2 Bacteriophage

Wavelength
(nm)

Microorganism/
Surrogate

Medium

UV
dose
(mJ/
cm')

Log

inactivation

Inactivation
Rate

Constant
(cnv'/mJ)

Reference

Study
Type

Notes

254

MS2

PBS

40

3



Wang et
al., 2019

Bench
scale

Collimated
Beam

220

MS2

PBS

25

3.1-5.2



Wang et
al., 2019

Bench
scale

Collimated
Beam

7-11


-------
Exhibit 7-7 provides a summary of the findings of UV technology results for MS2 used as a
surrogate for light-emitting diode lamps. Studies using LED lamps achieving log inactivation at
UV doses lower than those reported in previous EPA guidance included: Beck et al. (other than
280 nanometer (nm) only), 2017; Sholtes et al., 2016; Jarvis et al., 2019; and Hull and Linden,
2018, while studies of LED lamps achieving log inactivation at doses higher than those
contained in guidance were Kesharvarzfathy et al., 2020; Oguma et al., 2019, Beck et al. (LED
280 nm only), 2017; Oguma et al., 2016; Hull et al., 2019.

Exhibit 7-7. Examples of Potentially Relevant Light Emitting Diode
Ultraviolet Doses for Log Inactivation of Male-Specific-2 Bacteriophage







UV



Inactivation







Wavelength
(nm)

Microorganism/
Surrogate

Medium

dose
(mJ/
cm')

Log

inactivation

Rate

Constant
(cnvVmJ)

Reference

Study
Type

Notes

254

MS2

PBS

42.3

1.94



Keshavarzfathy
et al., 2020

Bench
scale

Collimated
Beam

255

MS2

PBS

40

3.16 (calc)

0.079

Sholtes and
Linden, 2019

Bench
scale



260

MS2

Water

30.3

2



Beck et al.,
2017

Bench
scale

Collimated
Beam

260

MS2

Buffered
Water

58

4



Sholtes et al.,
2016

Bench
scale

Collimated
Beam

265

MS2

Buffered
Water

40

2.5



Song et al.,
2018

Bench
scale

Pulsed

265

MS2

PBS

105.2

4



Oguma et al.,
2019

Bench
scale



266

MS2

Water

9

7



Kim et al.,
2017

Bench
scale

Collimated
Beam

260 & 280

MS2

Water

32.8

2



Beck et al.,
2017

Bench
scale

Collimated
Beam &
Combined

265

MS2

PBS

20

1.6



Song et al.,
2019

Bench
scale

Combined

275

MS2

PBS

9.2

2



Jarvis et al.,
2019

Demo

Collimated
Beam

280

MS2

Water

38.5

2



Beck et al.,
2017

Bench
scale

Collimated
Beam

280

MS2

PBS

122.1

4



Oguma et al.,
2019

Bench
scale



285

MS2

Water

103.4

3



Oguma et al.,
2016

Bench
scale



285

MS2

PBS

60

2

(interpolated
Table S1B)



Hull et al.,
2019

Bench
scale/
Demo

Collimated
Beam

255 or 265

MS2

PBS

40

3

(interpolated)



Hull and
Linden, 2018

Bench
scale

Combined

255 or 265

MS2

PBS

40

4

(interpolated)



Hull and
Linden, 2018

Bench
scale

Combined

7.1.2.2 Other New Information not included in existing Long Term 2 Enhanced Surface
Water Treatment Rule Toolbox

Through the SYR4 process, EPA reviewed whether other new information pertaining to
emerging technologies, which have not been included in the existing LT2 toolbox guidance
manual, may be helpful for removal or inactivation of protozoa including Cryptosporidium. EPA

7-12


-------
summarizes that information here and finds that this information appears insufficient to develop
quantification criteria for inactivation and removal credit for Cryptosporidium.

7.1.2.3	Turbo Coagulation

Turbo coagulation was described in the Surface Water Analytical Tool (SWAT) model guidance
(USEPA, 2000) as removal of TOC ranging from 30-84 percent, which is generally superior to
the values established by regulatory 3x3 matrix TOC removals as 15-50 percent. No new journal
articles were found during the SYR4 period of review that referenced the term turbo coagulation
or correlated Cryptosporidium removal at turbo coagulation performance levels.

7.1.2.4	Powdered activated carbon

Campinas et al. (2021a) studied the application of powdered activated carbon (PAC) in
combination with other treatment methods such as membrane filtration by conducting pilot-scale
research on the pressurized PAC/coagulation/ceramic microfiltration (MF) hybrid system
(PAC/Alum/MF) with low turbidity and low natural organic matter surface water spiked with
organic microcontaminants. Results indicated that PAC/Alum/MF was a full barrier against
aerobic endospores as an indicator of protozoan/ Cryptosporidium (oo)cysts.

Also, there has been increased emphasis on optimizing the PAC treatment process through
factors such as PAC dosage. Campinas et al. (2021b) conducted pilot trials of
PAC/coagulation/sedimentation with low-turbidity surface waters, and four sets of operating
conditions were considered to test different PAC types, doses, and contact times. The result
indicated that the PAC dosage above 10 mg/L hampered the clarification of the studied waters
with aerobic endospores used as indicators of protozoan oocysts.

7.2 Ground Water Rule

EPA promulgated the GWR on November 8, 2006, to increase protection against microbial
pathogens, specifically viral and bacterial pathogens, in public water supplies that use ground
water sources (USEPA, 2006b). The GWR established a risk-targeted approach to identify
ground water systems susceptible to fecal contamination and requires action to correct significant
deficiencies and fecal contamination identified by triggered source water monitoring, assessment
source water monitoring or additional source water monitoring. (USEPA, 2006b). This approach
involves a multifaceted strategy including sanitary surveys, source monitoring, high risk system
identification, and appropriate treatment and compliance monitoring. Following a brief recap of
the TT requirements under the GWR, EPA presents and discusses the analytical results with the
SYR3 and SYR4 ICR data for assessing the national impacts collectively from those
requirements in this section. This section also discusses new information relevant to the
treatment provisions of the GWR.

7.2.1 Sanitary Surveys

As a condition of primacy delegation by EPA, primacy agencies must conduct on-site sanitary
surveys of each ground water system by reviewing the adequacy of water source, facilities,
equipment, operation, and maintenance of a PWS.

7-13


-------
A sanitary survey is defined by the NPDWR as an onsite review of the water source, facilities,
equipment, operation and maintenance of a public water system for the purposes of evaluating
the adequacy of such source, facilities, equipment, operation and maintenance for producing and
distributing safe drinking water (40 CFR 141.2(d)). The sanitary survey is intended to identify
significant deficiencies (USEPA, 2019d) including deficiencies which may make a system
susceptible to microbial contamination. Primacy agencies must conduct sanitary surveys every
three years for most CWSs and every five years for NCWSs and CWSs that meet certain
performance criteria. The systems need to take corrective actions to fix the identified
deficiencies, as described in section 7.2.3.

7.2.2	Treatment Technique Requirements under Ground Water Rule

The GWR TT requirements require ground water systems to implement corrective action if a
sanitary survey significant deficiency is identified or if the initial source sample or one of the
five additional ground water source samples tests positive for fecal contamination.

7.2.3	Corrective Actions

A ground water system must take corrective action within 120 days, or within the approved
corrective action plan schedule, upon receiving notification of a significant deficiency from the
primacy agency or written notice from a laboratory that a ground water source sample collected
was fecal indicator-positive (USEPA, 2006b).

Ground water systems must implement at least one of the following corrective actions:
correction of significant deficiencies; providing an alternate source of water; eliminating the
source of contamination; or providing treatment that reliably achieves at least 99.99 percent (4-
log) treatment of viruses for each contaminated ground water source (USEPA, 2006b).

7.2.4	Analytical Results Reflecting Ground Water Rule Impacts from Treatment
Techniques Requirements

As described and discussed in the Occurrence and Exposure chapter (Chapter 6), EPA analyzed
the national compliance monitoring data records in SYR3 and SYR4 ICR collectively to assess
the changes that occurred after the implementation of GWR and/or RTCR. Considering the
related regulatory timelines indicated in Exhibit 6-5 in the Occurrence and Exposure chapter, the
changes observed among ground water systems from years 2007 and 2008 (right before GWR
became effective) to years 2014 and 2015 (after the Sanitary Survey was completed during the
first round and right before RTCR became effective) may indicate changes driven by
implementation of GWR especially among undisinfected ground water systems. After 2016, the
changes may have been collectively driven by the GWR and RTCR among ground water
systems, including undisinfected ground water systems. Although the SYR3/SYR4 ICR datasets
do not allow EPA to evaluate the extent to which the individual corrective actions taken could
lead to the overall changes, the datasets have enabled EPA to do the following:

1. Systematically identify disinfecting versus undisinfected ground water systems for each
of the individual years in the datasets (see Section 6.4 in the Occurrence and Exposure
chapter). Note that this approach does not allow determination of whether disinfecting
systems are achieving 4-1 og inactivation/removal.

7-14


-------
2.	Evaluate changes in the number and percentages of disinfecting (vs. undisinfected)
systems over time (see Section 6.4 of the Occurrence and Exposure chapter). It is difficult
for EPA to determine the universe of ground water systems that must maintain 4-log
treatment as treatment information is not always current in SDWIS/Fed.

3.	Assess changes in total coliform positive / E. coli positive rates from pre-GWR to post-
GWR (See Section 6.2.1 of the Occurrence and Exposure chapter.

As discussed in Chapter 6, since the GWR became effective, there is an increasing trend of
number or percentages of ground water systems that are disinfecting. That may be somewhat
attributable to the regulatory element that adding treatment of 4-log virus inactivation and
removal is one of the corrective actions under the GWR. The decreasing trends of both total
coliform positive and E. coli positive rates indicate that the GWR appeared to reduce microbial
occurrence in the distribution system. Such a decreasing trend could be extended to the period
after the RTCR became effective in 2016. However, the modeling results (see Section 6.4.2)
suggested that some very small undisinfected ground water systems might continue having high
total coliform positive / E. coli positive rates. Potential compliance challenges among small
ground water systems are discussed in section 7.2.6 below.

7.2.4.1	Undisinfected Ground Water System Treatment

Undisinfected ground water systems that choose to install treatment to correct significant
deficiencies or fecal contamination must provide at least 99.99 percent (4-log) inactivation
(disinfection) or removal (filtration) of viruses. Treatment technologies or combination of
technologies that have demonstrated the ability to achieve 4-log inactivation or removal of
viruses are chlorine (chlorine gas or sodium hypochlorite), chlorine dioxide, ozone, ultraviolet
(UV) radiation, anodic oxidation, reverse osmosis (RO), and nanofiltration (NF).

Since disinfection technology regulatory improvements are being considered the MDBP revision
effort, this section does not include discussion of new treatment information that has come
available other than those included in the LT2 section of this document (specifically, in section
7.1.2): membrane filtration, bag or cartridge filtration options, and disinfection options including
chlorine, chlorine dioxide, ozone, and UV.

Information about disinfection practices for ground water systems is available online:
https://www.epa.gOv/dwreginfo/drinking-water-distribution-svstem-tools-and-resources#dbps. A
limited discussion of the relationship between UV dose and virus inactivation of adenovirus
follows.

7.2.4.2	Ultraviolet Virus Inactivation of Adenovirus

UV disinfection is less effective at inactivating some viruses, particularly adenovirus, due in part
that UV-induced deoxyribonucleic acid (DNA) damage may be repaired during cell culture
assays (Eisheid et al., 2009).

To demonstrate 3- or 4-log inactivation for viruses, validation testing would need to demonstrate
greater than 6-log inactivation of MS2 phage. Such a demonstration requires an extremely high
concentration in the reactor influent to allow for enumeration of the organisms in the effluent

7-15


-------
samples. Because of the need for serial dilutions, these high concentrations are difficult to
measure and can introduce error into the experiment (USEPA, 2022b).

Research to find alternative challenge microorganisms for demonstrating virus inactivation is
ongoing (USEPA, 2022b). Some recent validations have included B. pumilus spores, A.
brasiliensis, and Adenovirus type-2 as high resistant test microorganisms for virus UV
inactivation applications, but there are differing practices for applying these test microbes in
validation testing and differing acceptance criteria are often encountered in validation reports
(USEPA, 2022b).

The UV Treatment toolkit (USEPA, 2022b) clarifies that the Innovative Approaches for
Validation of Ultraviolet Disinfection Reactors for Drinking Water Systems (USEPA, 2020b)
document describes four alternative calculated dose procedures that do not require the use of B.
pumilus spores, A. brasiliensis, and Adenovirus type-2, which often have higher observed dose
response variability and no established QA/QC dose-response-bounds criteria.

EPA stated at the time of the publication of the final GWR (USEPA, 2006b) that it believed that
UV technology could be used in a series configuration or in combination with other inactivation
or removal technologies to provide a total 4-log treatment of viruses to meet the GWR's
requirements. EPA also stated its belief that a UV reactor dose verification procedure for 4-log
inactivation of a range of viruses may be developed in the future. With future development of
UV validation procedures since the rule publication, it is now feasible for systems to demonstrate
that they can achieve 4-log inactivation of viruses with UV treatment alone.

Linden et al. (2015) described that adenovirus is very resistant to UV light requiring a relatively
high UV dose of 186 mJ/cm2 based on LP UV light at 254 nm for 4-log inactivation credit; and
that many of the test microbes used for UV reactor testing (e.g., MS2 phage, B. subtilis spores,
and T1 phage) are too sensitive to UV light to demonstrate such high UV dose values.

7.2.5	Compliance Challenges for Small Systems

EPA has found in Chapter 6 that small undisinfected ground water systems are more likely to
have higher total coliform rates than larger undisinfected ground water systems, and that
thousands of them have total coliform rates above 5 percent. These findings coincide with an
understanding of technical, managerial and financial limits for small system compliance and
suggest a need to improve those compliance factors.

7.2.6	CT Tables for Log Inactivation of Viruses

As discussed earlier, one of the corrective actions appropriate for response to source water fecal
contamination is providing 4-log viral inactivation/removal during treatment. Since the SYR3
process, EPA has recognized that some literature indicated the CT5 tables for viruses existing in
the Surface Water Treatment Rule (SWTR) Guidance Manual might not be sufficiently

5 CT is defined as disinfectant residual concentration (C) multiplied by contact time (T). A CT value is a measure of
disinfection effectiveness for the time that microorganisms in the water are in contact with a disinfectant.

7-16


-------
protective. Since the SWTRs are currently under consideration for a revision, the CT tables for
log inactivation of viruses are not discussed in detail further here.

Malayeri et al. (2016) provides a summary of recent research regarding fluence UV dose
required to achieve incremental log inactivation of bacteria, protozoa, viruses, and algae.

7.3 Revised Total Coliform Rule

EPA promulgated the RTCR on February 13, 2013 (and minor corrections on February 26, 2014)
and the rule became effective on April 1, 2016. The RTCR is the revision to the 1989 TCR and
protects public health through the reduction of potential pathways of entry for fecal
contamination into distribution systems. The RTCR ensures the integrity of all PWSs6 by
requiring PWSs that are vulnerable to microbial contamination to find and fix operations and
maintenance problems. Key RTCR provisions include: setting a MCLG and MCL for E. coli]
setting a total coliform TT requirement; requiring PWSs to monitor for total coliforms and E.
coli according to a specific sampling schedule and site plan; requiring assessments and corrective
actions to identify and eliminate sanitary defects when monitoring results indicate that a system
may be vulnerable to contamination due to the total coliform and E. coli positives; and requiring
seasonal systems (i.e., NCWSs not operated on a year-round basis that start up and shut down at
the beginning and end of each operating season) to monitor and certify the completion of state-
approved start-up procedures. The RTCR also requires PN for violations and requires CWSs to
include specific language in their CCRs when they must conduct an assessment or if they incur
an E. coli MCL violation.

Systems that exceed a specified frequency of total coliform positive sample occurrences or incur
an E. coli MCL violation trigger the rule's TT requirements and must 1) conduct a Level 1 or
Level 2 assessment to determine if any sanitary defects exist; and 2) correct any sanitary defects,
which are defects that provide a pathway of entry for microbial contamination into the
distribution system or failure of a barrier already in place. This approach is referred to as the
"find and fix approach." The RTCR specifies two levels of TT triggers and corresponding levels
of assessment (Level 1 and Level 2) in response to those triggers. The degree and depth to which
a PWS must examine its system depend on the TT trigger's potential impact on public health.

The "find and fix" regulatory framework is supported by the Revised Total Coliform Rule
Assessments and Corrective Actions Guidance Manual: Interim Final (USEPA, 2014b). As
discussed in section 6.2.1 of the Occurrence and Exposure chapter, EPA analyzed the SYR4 ICR
data and observed that both total coliform positive rates and routine E. coli positive rates have
decreased after the implementation of RTCR between 2015 and 2018. Note that RTCR became
effective on April 1, 2016 and the records were included from 2015 as a pre-RTCR
implementation baseline. As shown in Exhibit 6-12, this decreasing trend is applicable to
systems in different categories, i.e., source water types.

6 The RTCR applies to all PWSs except aircraft water systems, which are subject to the Aircraft Drinking Water
Rule (40 CFR 141 Subpart X).

7-17


-------
7.3.1 Description of Level 1 and Level 2 Assessments

Level 1 and 2 assessments are conducted in order to identify the possible presence of sanitary
defects and defects in distribution system monitoring practices, including those defects that may
have caused total coliform positive samples and triggered the assessment (40 CFR 141.2). Both
assessments must include a review and identification of atypical events that could affect
distributed water quality or indicate that distributed water quality was impaired; changes in
distribution system maintenance and operation that could affect distributed water quality
(including water storage); source and treatment considerations that impact distributed water
quality, where appropriate (e.g., whether a ground water system is disinfecting); existing water
quality monitoring data; and inadequacies in sample sites, sampling protocol, and sample
processing.

Exhibit 7-8 lists the RTCR TT triggers for Level 1 and Level 2 assessments.

Exhibit 78-9. Revised Total Coliform Rule Treatment Technique Triggers for Level

1 and Level 2 Assessments

RTCR Assessment
Level

Triggers

Level 1 Assessment
Triggers

For systems taking 40 or more samples (including routine and repeat samples) per
month, the PWS exceeds 5.0 percent total coliform positive samples for the month.

For systems taking fewer than 40 samples (including routine and repeat samples) per
month, the PWS has two or more total coliform positive samples in the same month.

The PWS fails to take every required repeat sample after any single total coliform
positive sample.

Level 2 Assessment
Triggers

The PWS has an E. coli MCL violation.

The PWS has a second Level 1 TT trigger within a rolling 12-month period unless the
state has determined that the PWS found the sanitary defect that likely caused the
first Level 1 TT trigger, and the PWS corrected or fixed the sanitary defect before the
second Level 1 TT trigger occurred. With the state's approval, the system would not
trigger a Level 2 assessment but would need to conduct a second Level 1
assessment.

For PWSs with approved reduced annual monitoring, the system has a Level 1 TT
trigger in two consecutive years.

Source: 40 CFR 141.859 (a)

A system must conduct a Level 1 or Level 2 assessment consistent with State requirements if the
system exceeds one of the TT triggers discussed in Exhibit 7-8. The system must comply with
any expedited actions or additional actions required by the State in the case of an E. coli MCL
violation. A Level 2 assessment involves a more comprehensive investigation and review of
available information, additional and internal and external resources, and other relevant practices
(USEPA, 2020c).

When a PWS has a second Level 1 TT trigger within a rolling 12-month period, this triggers a
Level 2 TT unless the state has "reset" the PWS by determining that the PWS found the sanitary
defect that likely caused the first Level 1 TT trigger, and corrected or fixed the sanitary defect
before a second Level 1 TT trigger occurred. When a PWS is "reset" with the state's approval,
the system would not trigger a Level 2 assessment due to a second Level 1 TT trigger within a

7-18


-------
rolling 12-month period, but would instead need to conduct a second Level 1 assessment (40
CFR 141.859(a)(2)(ii).

Primacy agencies provide PWSs specific directives and forms for conducting and documenting
assessments. A Level 1 assessment must be conducted by the PWS, unless the state specifies
otherwise. Level 2 assessments must be conducted by the state or parties approved by the state.
Some examples of such approved parties could include state personnel; an operator certified by
the state to operate a system of similar size, type and complexity; technical assistance provider
such as a circuit rider; a supervisor or manager from the water system, supported by other
experts or employees of the system; and/or a consultant/consulting engineer.

All sanitary defects identified during a Level 1 or Level 2 assessment must be corrected within
30 days of the date that the PWS learns it has exceeded a TT-trigger or within a schedule
approved by the primacy agency.

A discussion of data quality considerations regarding the reporting of Level 1 and Level 2
assessments is included in Appendix C. Also included in Appendix C is a characterization of the
recurrence of multiple RTCR assessments at individual PWS; frequency that RTCR assessments
were unable to identify a sanitary defect or corrective action; and types of sanitary defects
identified during RTCR assessments.

7.3.2 Description of Corrective Actions

Corrective actions are measures taken to address or fix any sanitary defect(s). The type of
corrective action that a system performs will depend on the type of sanitary defect identified. The
system must complete corrective actions within 30 days of triggering the assessment or on a
timetable approved by the State. The objective of RTCR Level 1 and Level 2 assessments is to
identify pathways of microbial drinking water contamination and to correct sanitary defects
identified. Sanitary defects can be resolved using distribution system corrective actions specified
by the RTCR corrective action guidance, tools of the distribution system toolbox or as approved
by the primacy agency.

It should also be noted that in addition to corrective actions taken within distribution systems,
treatment enhancements in the treatment train for improving the quality of water entering
distribution systems can also be helpful to reduce total coliform / E. coli detections in
distribution systems and can be part of considerations of corrective actions. For instance, the
enhanced TOC removal or operation of filters in a biological mode will help improve the
biological stability of the treated water and thus maintain residual levels throughout distribution
systems, resulting in lower total coliform/ E. coli detection rates. However, these types of
treatment enhancements are currently part of ongoing considerations for potential revisions of
microbial and disinfection by-product rules and will not be covered in this document any further.

Since the RTCR, EPA has compiled information about tools that have been used to help the
public water systems to manage and improve the water quality throughout the distribution
system (see https://www.epa.gov/dwreginfo/drinking-water-distribution-system-tools-and-
resources), which will be described herein as the distribution system toolbox. A discussion of the

7-19


-------
characterization of RTCR assessments and occurrence of sanitary defects in DS and corrective
actions is highlighted below in Section 7.3.3.

7.3.3 Advances/Improvements/Innovations to Revised Total Coliform Rule Treatment
Techniques

USEPA has developed many resources and tools related to distribution system water quality
which may be helpful to PWSs and others seeking to address total coliform positives and E. coli
positives as related to RTCR sanitary defects. These resources, which are summarized on
USEPA's distribution system website, are referred to here as the distribution system toolbox
(USEPA, 2023). The distribution system toolbox includes techniques such as disinfection,
flushing, repair and replacement of distribution system components, pressure management, water
age, storage facility maintenance, cross-connection control and backflow prevention, sampling,
monitoring, operations plan, and corrosion control. Many of these distribution system
improvements also are being considered under the MDBP rule revisions effort currently under
consideration by the National Drinking Water Advisory Council, and they are not described
further in this SYR4 support document.

ASDWA conducted a survey about distribution system practices that was completed by drinking
water representatives from 41 states and territories (2020). The survey covered many of the same
techniques as described in the toolbox. ASDWA found that at least 12 states (30 percent of
respondents) have flushing requirements written in their state legislation to better ensure a safe
and reliable distribution system (ASDWA, 2020). Of the 70 percent of state respondents that do
not require a flushing program, many strongly recommend it. Some of the methods to encourage
a flushing program include either requiring a flushing plan to be eligible for Drinking Water
State Revolving Fund (DWSRF) funding or encouraging it to be a part of the water system's
operations and maintenance plan during sanitary surveys (ASDWA, 2020). ASDWA also found
that most states specify a requirement of 20 psi as a minimum pressure limit and that some states
also have a maximum pressure limit; typically between 60 and 150 psi. They found that 53
percent of responding states require a cross-connection survey and half of those require the
survey to include all water use equipment (e.g., cooling towers, spray misters, spas, and pools)
(ASDWA, 2020).

The American Water Works Association (AWW A) also has developed guidance that may be
helpful with addressing RTCR sanitary defects. For example, the AWWA M68 manual suggests
that, to remove sediment and biofilm that may harbor nitrifying bacteria, storage facilities should
be inspected and cleaned at least every five years (AWWA, 2017). AWWA conducted its fifth
disinfection survey in 2017 which collected information from water systems on their common
treatment practices (AWWA, 2018; AWWA, 2021). Survey responses were summarized for a
total of 375 water systems, distributed across 44 states and one United States territory, and
represented 0.7 percent of the approximately 52,700 community water systems in the United
States. Systems noted that meeting minimum chlorine target levels is more challenging in the
distribution system than it is for meeting the targets for primary disinfection. The report showed
12 percent of systems reported frequent difficulties in meeting their chlorine residual targets in
the distribution system while the majority of the respondents reported having difficulty on
occasion (AWWA, 2018). Gibson and Bartrand (2021) evaluated publicly available data for

7-20


-------
secondary disinfection practices of 3,823 CWSs and found that 831 systems do not provide
residual (secondary) disinfection.

A more detailed explanation of new information regarding the RTCR corrective actions is
included in Appendix D.

7.4 Aircraft Drinking Water Rule

The ADWR became effective in October 2011 and established barriers of protection from
disease-causing organisms in the air carrier industry and to ensure that safe and reliable drinking
water is provided to aircraft passengers and crew. The ADWR applies only to aircraft with
onboard water systems that provide water for human consumption through pipes and regularly
serve an average of at least 25 individuals daily, at least 60 days out of the year, and that board
only finished water for human consumption. The ADWR assumes that no additional treatment is
necessary for aircraft water systems if finished water is boarded. Air carriers develop and
implement an operations and maintenance plan for each aircraft water system in active service
that identifies the frequency of routine disinfection and flushing for aircraft water systems. In
addition, air carriers develop a coliform sampling plan covering each aircraft owned or operated
by the carrier and routinely disinfect and flush aircraft water systems at the frequency
recommended by the water system manufacturer. The minimum frequency of sampling can
range from 1 to 12 times per year and is based upon the chosen minimum frequency of routine
disinfection and flushing of the aircraft [40 CFR 141.803(b)(3)],

7.4.1	Corrective Actions

The Interim Final Guidance Manual for the Aircraft Drinking Water Rule (ADWR) clarifies the
appropriate corrective actions for the occurrence of total coliform positives and E. coli positives
[40 CFR 141.803(c)(2&3)] (USEPA, 2010c). These corrective actions can include disinfection
and flushing, follow-up sampling, restricting public access, and repeat sampling depending upon
the water quality or operation and maintenance trigger. When restricting public access,
notification is to be provided within 24 hours to the crew and posted in the aircraft to notify
passengers of the restricted access.

Corrective actions required for E. coli positive include restriction of public access; conducting
corrective action with disinfection and flushing; and follow-up sampling. Corrective disinfection
and flushing are triggered by failure to collect appropriate coliform samples and the following
positive sample situations: any repeat total coliform positive oris, coli positive sample; routine
or follow-up E. coli positive sample; and routine total coliform positive sample if corrective
disinfection and flushing option is selected. Corrective action disinfection and flushing is also
required for boarded water that does not comply with FDA regulations, meet TNCWS standards
or is otherwise determined to be unsafe.

7.4.2	New Information Available since Aircraft Drinking Water Rule Promulgation

Several studies have been identified since ADWR promulgation that address bacterial growth,
disinfection and flushing, restricting public access, and other operations and maintenance topics.
These studies underscored the benefits of corrective actions under ADWR and identified aspects
that could improve communication and water transport.

7-21


-------
7.4.2.1 Potential for Bacterial Growth and Temperature of Water

Handschuh et al. (2015) conducted a study that analyzed water quality from short-haul and long-
haul aircraft and water service vehicles on a weekly basis. The study found a significant
difference in that the long-haul aircraft showed a degraded microbial quality in comparison to
short-haul aircraft, with the long-haul aircraft having a mean viable microbial count of 12,000
colony forming unit per milliliter (cfu/mL) as compared to short-haul count of 4800 cfu/mL
during the September through December 2010 study period. The study suggested that long-haul
aircraft may require more operations and maintenance activities to maintain water quality.

Handshuh et al. (2015) also found that for both long- and short-haul aircraft, there were higher
levels of chlorine (free and total) discovered in the water service vehicle and water source than
on the aircraft itself. The study also reported that the temperature of both the aircraft water and
water service vehicle, and viable bacterial count of the water service vehicle, had a significant
impact on on-board aircraft bacterial levels. The water service vehicle was identified as a
significant source of increased microbial load within aircraft water tanks having a maximum
viable microbial count of 140,000 cfu/mL during the Sept. - Dec. 2010 study period (Handshuh
et al., 2015).

A study by Platkin (2019a), found that the number of ADWR violations by all aircraft in 2018
was significantly less than the number in 2012, the first year after the ADWR was implemented.
For major aircraft, violations have decreased 69 percent (262 to 81) while violations among
regional aircraft have decreased 71 percent (351 to 103).

EPA found that disinfection/flushing and follow-up sampling performed as a corrective action
under ADWR (i.e., in response to total coliform positives / E. coli positives) occurred 39.5
percent less frequently in 2019 than in 2012 (See Exhibit 7-9) (USEPA, 2022c). Caution should
be used when interpreting these data since a reduction in the occurrence of disinfection/flushing
and follow-up sampling corrective actions does not necessarily imply there is a reduction in fecal
risk since disinfection and flushing and follow-up sampling could be performed in situations of
total coliform positive and E. coli negative sample results.

7-22


-------
Exhibit 710-11. Aircraft Drinking Water Rule - Number of Systems Performing
Corrective Actions by Year - 2012 to 2019

900
800
700
600
500
400
300
200
100
0

2012 2013 2014 2015 2016 2017 2018 2019

Disinfection &
Flushing and Follow-
up Sampling

Restrict Public Access

Repeat Sampling

Source: ADWR Compliance Public Reports - Aircraft Drinking Water Rule System Operations (USEPA, 2022c)

7.4.2.2 Disinfection and Flushing

Szabo et al. (2019) studied a mock pilot-scale aircraft drinking water system to measure the
effectiveness of routine disinfection and flushing in preventing coliform persistence. The authors
found that coliform bacteria are not persistent on aircraft plumbing surfaces and following
routine flushing using disinfectants consisting of chlorine dioxide (i.e., Purogene®), ozone or
mixed oxidant. While coliform bacteria were not detected in the bulk mock aircraft plumbing
surfaces and distribution plumbing, one exception was that the aerator installed in the lavatory
faucet continued to be total coliform positive following the routine disinfection and flushing
procedure.

Szabo et. al. (2019) concluded that faucet aerators could be a source of coliform contamination
that may result in total coliform positive samples. The study proceeded to conduct experiments
with the total coliform positive aerators and found that 30 minutes of soaking disinfection
yielded coliform negative results (i.e., no detectable coliforms). The aerators soaked in
commercially available and commonly used cleaning agents during the study contained 100 mL
of either Glyco-San® or quaternary ammonium compounds (e.g., Lysol®).

EPA considers disinfection and flushing to be a more protective and pro-active public health
measure than monitoring. For the final ADWR rule, EPA re-aligned the disinfection, flushing
and monitoring frequencies to emphasize the importance of disinfection and flushing in
comparison to monitoring. As a result, those air carriers that conduct more frequent disinfection
and flushing do not have to monitor as frequently.

The ADWR requires that the air carrier conduct disinfection and flushing of the aircraft water
system in accordance with, or is consistent with, the water system manufacturer's
recommendations. The air carrier may conduct disinfection and flushing more frequently, but not
less frequently, than the manufacturer recommends [40 CFR 141.804(b)(2)(i)].

7-23


-------
Review of the ADWR PWS inventory as of June 2021 indicated that 97.6 percent of aircraft
were scheduled to conduct routine disinfection and flushing a minimum of four times per year.

While it may seem prudent to establish requirements for sampling, disinfection and flushing for
vehicles used to transport water, in fact, the sampling performed in the aircraft is intended to be
representative of the entire aircraft water system including water received from transport. One
aspect that adds complexity to the representative sampling of the aircraft is that the aircraft water
tank may not be completely drained between fillings and therefore water from more than one
source is usually commingled in the aircraft water system. Water tanks are generally only
completely emptied and refilled when the water system is serviced, when the water on board has
been completely consumed, and when the aircraft is not in operation during cold winter days to
avoid the system freezing (Handschuh et al., 2015).

7.4.2.3 Restrict Public Access and Notification to Passengers and Crew

Restricted public access is required by ADWR as a corrective action for E. coli positive samples
and is optional in total coliform positive / E. coli negative sample result situations (USEPA,
2010c). Restricted public access is defined as: physically disconnecting or shutting off the
aircraft water system or otherwise preventing the flow of water through the tap(s); providing
public notification via delivery methods such as broadcast via aircraft announcement,
prominently displayed in lavatories for passengers and prominent notice in the galley for crew,
or via hand delivery of notice; and providing alternatives to water from the aircraft water system,
such as bottled water, antiseptic hand gels or wipes and other feasible measures that reduce or
eliminate the need to use the aircraft water system (40 CFR 141.803(d)).

This SYR4 review identified ADWR restricted public access corrective actions by year for the
period of 2012 to 2019. As shown in Exhibit 7-9, corrective actions for restricted public access
remained relatively constant throughout the SYR4 review period.

Platkin (2019b) reported that representatives from the Air Line Pilots Association, International,
expressed written comment to EPA at the time of ADWR rule proposal that although public
notice of a coliform oris, coli positive result would be required, this public notice would not be
provided to persons who may have already been sickened before the discovery of the water
sample.

The ADWR does not require public notice to passengers of coliform or E. coli positive results
that are obtained from an aircraft they have flown on the date of or before the date of water
sample collection having positive results for coliform or E. coli. As mentioned previously with
regard to representative sampling, one aspect that adds to the complexity to retroactive passenger
notification of positive results is that the aircraft water tank may not be completely drained
between fillings and therefore water from more than one source is usually commingled in the
aircraft water system.

7-24


-------
7.4.2.4	Repeat Sampling

This SYR4 review identified ADWR repeat sampling corrective actions by year for the period of
2012 to 2019. As shown in Exhibit 7-9, corrective actions for repeat sampling remained
relatively constant throughout the SYR4 review period.

No additional new information was found regarding the ADWR corrective action of repeat
sampling.

7.4.2.5	Other Operation & Maintenance Topics

The Airport Cooperative Research Program (ACRP), sponsored by the Federal Aviation
Administration, provided a synthesis report of airport practice regarding airport community,
water quality events and the ADWR in 2018. This report summarized numerous observations
and conclusions regarding the implementation of ADWR including but not limited to the
following operation and maintenance challenges: emergency communications, situational
notification, and contractual and regulatory obligations including those to the U.S. FDA (ACRP,
2018).

A significant challenge identified by the ACRP is that airports and aircraft are not receiving
prompt and accurate notification of a community PWS water quality issue for the water serving
an airport. The airport contact stakeholders are currently often notified in the same manner and
time frame as the general public, which often means notice is received late or not at all. The
report suggested that appropriate airport personnel be added to the community PWS critical
customer contact list. Additional suggestions were provided for communication of stakeholders
within the airport affected entities (ACRP, 2018). Some synthesis participants suggested that
water quality events be addressed by the airport similar to communicable disease responses
(ACRP, 2018).

The report suggested additional research or collaboration regarding water cabinet quality and
maintenance: connective hose condition and exposure to sunlight; water stagnation; water
boarding procedures without proper flushing of equipment; options for airport water system
treatment; and international terminal water quality maintenance (ACRP, 2018).

7.5 Filter Backwash Recycling Rule

EPA promulgated the Filter Backwash Recycling Rule (FBRR) on June 8, 2001 (66 FR 31086).
The rule aimed to increase public health protection by addressing microbial contaminant risks
associated with filter backwash recycling practices. The rule required certain systems to return
recycled filter backwash water, sludge thickener supernatant, and liquids from dewatering
processes to a location in the system such that all filtration processes of a system are employed,
or at an alternate location if approved by the State. In addition, the rule required systems that
employ conventional filtration or direct filtration to notify States of their recycling practices by
June 8, 2004, and after then to keep and retain records on file about their recycle flows for
subsequent review and evaluation by the State. There are no ongoing monitoring requirements
associated with the FBBR.

7-25


-------
EPA reviewed available State data collected under the ICR; however, the EPA did not identify
any new and relevant information that would indicate that revisions to the NPDWR are
appropriate at this time.

7-26


-------
8 References

Abebe, L.S., Y-H Su, R.L. Guerrant, N.S. Swami, and J.A. Smith. 2015. Point-of-Use Removal
of Cryptosporidium Parvum from Water: Independent Effects of Disinfection by Silver
Nanoparticles and Silver Ions and by Physical Filtration in Ceramic Porous Media. Environ. Sci.
Technol. 49:21, 12958-12967.

Airport Cooperative Research Program (ACRP). 2018. ACRP Synthesis 88 - Airport
Community, Water Quality Events, and the Aircraft Drinking Water Rule, National Academy of
Sciences.

Ahmed, W., K. Hamilton, S. Toze, S. Cook, and D. Page. 2019. A Review on Microbial
Contaminants in Stormwater Runoff and Outfalls: Potential Health Risks and Mitigation
Strategies. Science of the Total Environment, 692, 2019, 1304-1321.

Alspach, B. 2019. Charting the Reverse Osmosis Renaissance. Journal AWWA, 111:8, 85-87.

Association of State Drinking Water Administrators (ASDWA). 2019. A Study Examining
Contamination of Total Coliform Samples, https://www.asdwa.org/2019/04/15/a-studv-

examining-contamination-of-total-coliform-samples/.

ASDWA. 2020. State Drinking Water System Survey, https://www.asdwa.org/wp-
content/uploads/2020/09/A.SDW A-Distribution-System-Survev-White-Paper-
2Q2Q.pdf#:~:text=Tn%2QMav%2Q2Q20%2C%20the%2QAssociation%2Qof%2QState%2QDrinkin
g.that%20could%20be%20used%20to%20inform%20future%20regulations.

Ashbolt, Nicholas J. 2015. Microbial contamination of drinking water and human health from
community water systems. Current Environmental Health Reports, 2(1): 95-106.

Audebert, C., F. Bonardi, S. Caboche, K. Guyot, H. Touzet, S. Merlin, N. Gantois, C. Creusy, D.
Meloni, A. Mouray, E. Viscogliosi, G. Certad, S. Benamrouz-Vanneste, and M. Chabe. 2020.
Genetic basis for virulence differences of various Cryptosporidium parvum carcinogenic isolates.
Scientific Reports. 10(1):7316.

American Water Works Association (AWW A). 2016. Water Treatment.

https://www.awwa.Org/portals/0/files/piiblications/dociiments/wsogradelexcerpt.pdf.

AWWA. 2017. M68. Water Quality in Distribution Systems. Denver, CO: AWWA.

AWWA. 2018. 2017 Water Utility Disinfection Survey Report. April 2018. Denver, Colo.:
AWWA.

AWWA. 2020. Disinfection of Water Storage Facilities. ANSI/AWWA Standard C652. Denver,
Colo.: AWWA.

AWWA. 2021. Emerging Trends in Disinfection: Lessons from AWWA's Disinfection Survey.
Journal AWWA, 113(1)20-28. https://doi.org/! 0.1002/awwa. 1648

8-1


-------
Baldasso, V., H. Lubarsky, N. Pichel, A. Turolla, M. Antonelli, M. Hincapie, L. Botero, F.
Reygadas, A. Galdos-Balzategui, J. A. Byrne, and P. Fernandez-lb anez. 2021. UVC inactivation
of MS2-phage in drinking water - Modelling and field testing. Water Research, 203: 117496.

Barbhuiya, N. H., S.P. Singh, A. Makovitzki, P. Narkhede, Z. Oren, Y. Adar, and C.J. Arnusch.
2021. Virus inactivation in water using laser-induced graphene filters. Materials, 14(12), 3179.

Beck, S.E., H. Ryu, L.A. Boczek, J.L. Cashdollar, K.M. Jeanis, J.S. Rosenblum, O.R. Lawal, and
K.G. Linden. 2017. Evaluating UV-C LED disinfection performance and investigating potential
dual-wavelength synergy. Water Research, 109: 207-216.

Berger, P., M.J. Messner, J. Crosby, D.V. Renwick, and A. Heinrich. 2018. On the use of total
aerobic spore bacteria to make treatment decisions due to Cryptosporidium risk at public water
system wells. International Journal of Hygiene and Environmental Health, 221, 704-711

Blatchley, E.R, C. Shen, and K. Scheible. 2019. Design and Validation Protocols for UV
Disinfection Systems. Water Research Foundation, Project No. ENER16C15/4791.

Blyth, J., M.R. Templeton, S-J Court, C. Luce, W. Cairns, and L. Hazell. 2021. Assessment of
indigenous surrogate microorganisms for UV disinfection dose verification. Water and
Environment Journal, 35(4): 1384-1392.

Borchardt, M.A., S.K. Spencer, B.A. Kieke, Jr., E. Lambertini, and F.J. Loge. 2012. Viruses in
Nondisinfected Drinking Water from Municipal Wells and Community Incidence of Acute
Gastrointestinal Illness. Environmental Health Perspectives, 120(9): 1272-1279.

Borchardt, M.A., B.A. Kieke, Jr., S. K. Spencer, E. Lambertini, T.R. Burch, and F.J. Loge. 2023,
Community intervention trial for estimating risk of acute gastrointestinal illness from
groundwater-supplied non-disinfected drinking water. J Water Health (2023) 21 (9): 1209-1227.

Burch, T.R., J.P. Stokdyk, N. Rice, A.C. Anderson, J.F. Walsh, S.K. Spencer, A.D. Firnstahl,
and M. A. Borchardt. 2022. Statewide Quantitative Microbial Risk Assessment for Waterborne
Viruses, Bacteria, and Protozoa in Public Water Supply Wells in Minnesota. Environmental
Science and Technology, 56(10): 6315-6324.

Busse, M., M. Becker, B.M. Applegate, J.W. Camp Jr., and E.R. Blatchley III. 2019. Responses
of Salmonella typhimurium LT2, Vibrio harveyi, and Cryptosporidium parvum to UVB and
UVA radiation. Chemical Engineering Journal, 371: 647-658.

Calderon, R.L., and G.F. Craun. 2006. Estimates of endemic waterborne risks from community-
intervention studies. Journal of Water and Health, 4(S2): 89-99.

Campinas, M., R.M.C. Viegas, R. Coelho, H. Lucas, and M.J. Rosa. 2021a. Adsorption/
Coagulation/ Ceramic Microfiltration for Treating Challenging Waters for Drinking Water
Protection. Membranes, 11(2): 91.

Campinas, M., C. Silva, R.M.C. Viegas, R. Coelho, H. Lucas, and M.J. Rosa. 2021b. To what
extent may pharmaceuticals and pesticides be removed by PAC conventional addition to low-

8-2


-------
turbidity surface waters and what are the potential bottlenecks? Journal of Water Process
Engineering, 40: 101833.

Carvajal, G., A. Branch, P. Michel, S.A. Sisson, D.J. Roser, J.E. Drewes, and S.J. Khan. 2017.
Robust evaluation of performance monitoring options for ozone disinfection in water recycling
using Bayesian analysis. Water Research, 124: 605-617.

Centers for Disease Control and Prevention (CDC). 2017. Parasites - Cryptosporidium (also
known as "Crypto"). Page last reviewed: August 16, 2017.
https://www.cdc.gov/parasites/crypto/illness.html

Chalmers, R.M., K. Elwin, S.J. Hadfield, and G. Robinson. 2011. Sporadic human
cryptosporidiosis caused by Cryptosporidium cuniculus, United Kingdom, 2007-2008. Emerg
InfectDis., 17(3): 536-8.

Chang, J.C.H., S.F. Osoff, D.C. Lobe, M.H. Dorfman, C.M. Dumais, R.G. Quails, and J.D.
Johnson. 1985. UV inactivation of pathogenic and indicator microorganisms. Applied and
Environmental Microbiology, 49: 1361-1365.

Chappell, C.L., P.C. Okhuysen, R.C. Langer-Curry, D.E. Akiyoshi, G. Widmer, and S. Tzipori.
2011. Cryptosporidium meleagridis: infectivity in healthy adult volunteers. American Journal of
Tropical Medicine and Hygiene, 85(2): 238.

Chen, C., L. Guo, Y. Yang, K. Oguma, and L.A. Hou. 2021. Comparative effectiveness of
membrane technologies and disinfection methods for virus elimination in water: A review.
Science of the Total Environment, 801: 149678.

Colford, Jr., J.M., S. Roy, M.J. Beach, A. Hightower, S.E. Shaw, and T.J. Wade. 2006. A review
of household drinking water intervention trials and an approach to the estimation of endemic
waterborne gastroenteritis in the United States. Journal of Water and Health, 4 (S2): 71-88.

Collier, S.A., L. Deng, E.A. Adam, K. M. Benedict, E.M. Beshearse, A.J. Blackstock, B.B.
Bruce, G. Derado, C. Edens, K.E. Fullerton, J.W. Gargano, A.L. Geissler, A.J. Hall, A.H.
Havelaar, V.R. Hill, R.M. Hoekstra, S.C. Reddy, E. Scallan, E.K. Stokes, J.S. Yoder, and M.J.
Beach. 2021. Estimate of Burden and Direct Healthcare Cost of Infectious Waterborne Disease
in the United States. Emerging Infectious Diseases. 27(1): 140-149.

Eisheid, A.C., J.N. Meyer, and K.G. Linden. 2009. UV Disinfection of Adenoviruses: Molecular
Indications of DNA Damage Efficiency. Applied and Environmental Microbiology, 75(1): 23-28.

Feng, Z., R. Lu, B. Yuan, Z. Zhou, Q. Wu, and T.H. Nguyen. 2016. Influence of solution
chemistry on the inactivation of particle-associated viruses by UV irradiation. Colloids and
Surfaces B Biointerfaces, 148: 622-628.

Fout, G.S., M.A. Borchardt, B.A. Kieke, and M.R. Karim. 2017. Human virus and microbial
indicator occurrence in public-supply groundwater systems: meta-analysis of 12 international
studies. Hydrogeology Journal, 25(4): 903-919.

8-3


-------
Gallandat, K., D. Stack, G. String, and D. Lantagne. 2019. Residual Maintenance Using Sodium
Hypochlorite, Sodium Dichloroisocyanurate, and Chlorine Dioxide in Laboratory Waters of
Varying Turbidity. Water, 11(6): 1309.

Gao, Y., Z. Sun, Y. Guo, Z. Qiang, and W. Ben. 2023. Virus inactivation by sequential
ultraviolet-chlorine disinfection: Synergistic effect and mechanism. Chemosphere, 314: 137632

Gerdes ME, Miko S, Kunz JM, Hannapel EJ, Hlavsa MC, Hughes MJ, Stuckey MJ, Francois
Watkins LK, Cope JR, Yoder JS, Hill VR, Collier SA. 2023. Estimating Waterborne Infectious
Disease Burden by Exposure Route, United States, 2014. EmergInfectDis. Jul;29(7): 1357-1366.

Gibson, M.C. and T.A. Bartrand. 2021. Understanding Community Water System Disinfection
Practices in the United States. Journal AWWA, 113(6):38-46.

Gilpin, B.J., T. Walker, S. Paine, J. Sherwood, G. Mackereth, T. Wood, T. Hambling, C.
Hewison, A. Brounts, M. Wilson, P. Scholes, B. Robson, S. Lin, A. Cornelius, L. Rivas, D.T.S.
Hayman, N.P. French, J. Zhang, D.A. Wilkinson, A.C. Midwinter, P.J. Biggs, A. Jagroop, R.
Eyre, M.G. Baker, and N. Jones. 2020. A large scale waterborne Campylobacteriosis outbreak,
Havel ock North, New Zealand. Journal of Infection. 81: 390-395.

Great Lakes Upper Mississippi River Board of State and Provincial Public Health and
Environmental Managers (GLUMRB). 2018. Recommended Standards For Water Works-
Policies for the Review and Approval of Plans and Specifications for Public Water Supplies.
Water Supply Committee, GLUMRB.

https://www.health.nY.eov/environmental/water/drinkine/reeulations/docs/2018 recommended
standards.pdf.

Gutierrez-Alfaro, S., A. Acevedo, J. Rodriguez, E.A. Carpio, andM.A. Manzano. 2015. Solar
photocatalytic water disinfection of Escherichia coli, Enterococcus spp. and Clostridium
Perfringens using different low-cost devices. J. Chem. Technol. Biotechnol., 91: 2026-2037.

Handschuh, H., J. O'Dwyer, and C.C. Adley. 2015. Bacteria that travel: the quality of aircraft
water. International Journal of Environmental Research and Public Health, 12(11): 13938-
13955.

Handschuh, H., M.P. Ryan, J. O'Dwyer, and C.C. Adley, 2017. Assessment of the bacterial
diversity of aircraft water: identification of the frequent fliers. PloS One, 12(1): e0170567.

Harmsco 2014. Cryptosporidium surrogate HC/90-LT2 Filter Performance Test: 2.0 micron latex
spheres, IBR Laboratories, May 2014.

Heinrich, A., D.V, Renwick, R. Weisman, A. Greene, S. Regli, K. Roland, and K. Rotert. 2022.
Using Sanitary Survey Findings to Identify Risk Management Challenges. JAWWA. 114:5: 34-
45.

Hill, C., D. Owen, S. Trussell, A. Evans, E. Houtz, S. Sriboonlue, S.Triolo, andK. Thompson.
2018. Alternative Water Source Requirements for Conventional Drinking Water Treatment.
Water Research Foundation. Project #4665.

8-4


-------
Holmes, E. 2019. Development and evaluation of chitosans as transformative coagulants-
flocculants to improve sand filter drinking water treatment. University of North Carolina at
Chapel Hill.

Hull, N.M. and K.G. Linden. 2018. Synergy of MS2 disinfection by sequential exposure to
tailored UV wavelengths. Water Research, 143: 292-300.

Hull, N., W.H. Herold, and K.G. Linden. 2019. UV LED water disinfection: Validation and
small system demonstration study. AWWA Water Science, July/August 2019, el 148.

Hunter P.R. and G. Nichols. 2002. Epidemiology and clinical features of Cryptosporidium
infection in immunocompromised patients. Clin Microbiol Rev. 15(1): 145-54.

Jacangelo, J.G., A. Subramani, S. Gray, and M. Cran. 2019. Methods for Integrity Testing and
Online Monitoring of NF and RO Membranes. Water Research Foundation. Project No. Reuse-
12-07/4757.

Jaferey, I. and G. Galjaard. 2020. Back to Ceramics: Centuries old material proves its mettle in
maximizing financial and operational performance. Water Environment & Technology, June
2020, 33-37.

Jarvis, P., O. Autin, E.H. Goslan, andF. Hassard. 2019. Application of Ultraviolet Light-
Emitting Diodes (IV-LED) to Full-Scale Drinking Water Disinfection. Water, 11(9): 1894.

Juranek, D.D. 1995. Cryptosporidiosis: sources of infection and guidelines for prevention. Clin
Infect Dis. 21 Suppl 1: S57-61.

Keithlin, J., J. Sargeant, M.K. Thomas, and A. Fazil. 2014. Systematic review and meta-analysis
of the proportion of Campylobacter cases that develop chronic sequelae. BMC Public Health, 14:
1203.

Keshavarzfathy, M., A.H. Malayeri, M. Mohseni, and F. Taghipour. 2020. UV-LED fluence
determination by numerical method for microbial inactivation studies. Journal of Photochemistry
& Photobiology A: Chemistry, 392: 112406.

Kim, D-K., S-J. Kim, and D-H. Kang. 2017. Inactivation modeling of human enteric virus
surrogates, MS2, QP, and OX174, in water using UVC-LEDs, a novel disinfecting system. Food
Research International, 91: 115-123.

Kinser, K. 2021. Membrane Filtration: Past, Present and Future. Journal AWWA. 113: 72-75.

Koehler, A.V., M.J. Whipp, S.R. Haydon, and R.B. Gasser RB. 2014. Cryptosporidium
cuniculus - new records in human and kangaroo in Australia. Parasit Vectors. 7: 492.

Kotloff, K.L., J.P. Nataro, W.C. Blackwelder, D. Nasrin, T.H. Farag, S. Panchalingam, Y. Wu,
S.O. Sow, D. Sur, R.F. Breiman, A.S. Faruque, A.K. Zaidi, D. Saha, P.L. Alonso, B. Tamboura,
D. Sanogo, U. Onwuchekwa, B. Manna, T. Ramamurthy, S. Kanungo, J.B. Ochieng, R. Omore,
J.O. Oundo, A. Hossain, S.K. Das, S. Ahmed, S. Qureshi, F. Quadri, R.A. Adegbola, M.

8-5


-------
Antonio, M.J. Hossain, A. Akinsola, I. Mandomando, T. Nhampossa, S. Acacio, K. Biswas, C.E.
O'Reilly, E.D. Mintz, L.Y. Berkeley, K. Muhsen, H. Sommerfelt, R.M. Robins-Browne, and
M.M. Levine. 2013. Burden and aetiology of diarrhoeal disease in infants and young children in
developing countries (the Global Enteric Multicenter Study, GEMS): a prospective, case-control
study. Lancet. 382(9888): 209-22.

Linden, K.G., H.B. Wright, and S. Beck. 2015. Guidance for Implementing Action Spectra
Correction with Medium Pressure UV Disinfection. Water Research Foundation Report #4376.

Mac Kenzie, W.R., N.J. Hoxie, M.E. Proctor, M.S. Gradus, K.A. Blair, D.E. Peterson, J.J.
Kazmierczak et al. 1994. A massive outbreak in Milwaukee of Cryptosporidium infection
transmitted through the public water supply. New England Journal of Medicine, 331(3): 161-
167.

Mackey, E.D., H.B. Wright, T. Hargy, J. Clancy, and A. Cabaj. 2006 (in press). Optimization of
UV Reactor Validation. American Water Works Association Research Foundation, Denver, CO.

Malayeri, A.H., M. Mohseni, B. Cairns, and J.B. Bolton. 2016. Fluence (UV Dose) Required to
Achieve Incremental Log Inactivation of Bacteria, Protozoa, Viruses and Algae, IUVA News,
2016.

Malkoske, T.A., P.R. Berube, and R.C. Andrews. 2020. Coagulation/flocculation prior to low
pressure membranes in drinking water treatment: a review. Environ. Sci.: Water Res. Technol., 6:
2993.

Mattioli, M.C., K.M. Benedict, J. Murphy, A. Kahler, K.E. Kline, A. Longenberger, P.K.
Mitchell, S. Watkins, P. Berger, O.C. Shanks, and C.E. Barrett. 2021. Identifying septic pollution
exposure routes during a waterborne norovirus outbreak-A new application for human-associated
microbial source tracking qPCR. Journal of Microbiological Methods, 180: 106091.

Mbonimpa, E.G., E.R. Blatchley III, B. Applegate, and W.F. Harper, Jr. 2018. Ultraviolet A and
B wavelength-dependent inactivation of viruses and bacteria in the water. Journal of Water and
Health, 16(5): 796-806.

Meng, Q.S., and C.P. Gerba. 1996. Comparative inactivation of enteric adenovirus, poliovirus
and coliphages by ultraviolet irradiation. Water Research, 30(11): 2665-2668.

Messner, M.J., and P. Berger. 2016. Cryptosporidium Infection Risk: Results of New Dose-
Response Modeling. Risk Anal. 36(10): 1969-1982.

Messner, M., S. Shaw, S. Regli, K. Rotert, V. Blank, and J. Soller. 2006. An approach for
developing a national estimate of waterborne disease due to drinking water and a national
estimate model application. Journal of Water and Health, 4(S2): 201-240.

Messner, M.J., P. Berger, and J. Javier. 2017. Total coliform and E. coli in public water systems
using undisinfected ground water in the United States. International Journal of Hygiene and
Environmental Health, 220(4): 736-743.

8-6


-------
Miller, T.A., M.W. Ware, L.J. Wymer, and F.W. Schaefer III. 2007. Chemically and genetically
immunocompromised mice are not more susceptible than immunocompetent mice to infection
with Cryptosporidium muris. Veterinary Parasitology, 143(2): 99-105.

Moltz, H.L.N., R. Mandel, K.R. Bencala, J.B. Palmer, A. Nagel, and S. Kaiser. 2018. Forest
Cover Impacts on Drinking Water Utility Treatment Costs in a Large Watershed. Water
Research Foundation, Project 4651, 2018.

Monis, P., M. Lau, M. Harris, D. Cook, and M. Drikas. 2017. Risk-based management of
drinking water safety in Australia: Implementation of health based targets to determine water
treatment requirements and identification of pathogen surrogates for validation of conventional
filtration. Food and Waterborne Parasitology, 2017(8-9): 64-74.

Mor, S.M., A. DeMaria, Jr., J.K. Griffiths, and E.N. Naumova. 2009. Cryptosporidiosis in the
elderly population of the United States. Clin Infect Dis. 48(6): 698-705.

Mustafa, S., A. Bahar, A.R.Z. Abidin, Z.A. Aziz, and M. Darwish. 2021. Three dimensional
model for solute transport induced by groundwater abstraction in river-aquifer systems.
Alexandria Engineering Journal. 60: 2573-2582.

Naumova, E. N., A.I. Egorov, R.D. Morris, and J.K. Griffiths. 2003. The Elderly and Waterborne
Cryptosporidium Infection: Gastroenteritis Hospitalizations before and during the 1993
Milwaukee Outbreak. Emerging Infectious Diseases, 9(4): 418-425.

National Drinking Water Advisory Council (NDWAC). 2000. Working Group Meeting on
Contaminant Candidate List Regulatory Determinations and the 6-Year Review of Existing
Regulations. Available online at: https://www.epa.eov/sites/prodiiction/files/2015-
I I >locuments/march \ . .^00 meeting on ccl am! year review.pdf

Nix, D.K., and J.S. Taylor. 2018. Filter Evaluation Procedures for Granular Media. Second
Edition. AWWA, Denver, CO.

National Research Council (NRC). 2004. Indicators for Waterborne Pathogens.

https://www.ncbi.nlm.nih.eov/books/ 1 4 _ s i 53/pdf/Booksh^lt _ x < Hv _ I -n 63.pdf.

Oudega, T.J., G. Linder, R. Sommer, A.H. Farnleitner, G. Kerber, J. Derx, M.E. Stevenson and
A.P. Blaschke. 2022. Transport and removal of spores of Bacillus subtilis in an alluvial gravel
aquifer at varying flow rates and implications for setback distances. Journal of Contaminant
Hydrology, 251: 104080.

Oguma, K., S. Rattanakul, and J. Bolton. 2016. Application of UV Light-Emitting Diodes to
Adenovirus in Water. Journal Environmental Engineering, 142(3): 04015082.

Oguma, K., S. Rattanakul, and M. Masaike. 2019. Inactivation of health-related microorganisms
in water using UV light-emitting diodes. Water Supply, 19(5): 1507-1514.

Pang, L., A. Tham, P. Nilprapa, A. Cocker, P. MacDonald, R. Adams, B. Robson, D. Wood, V.
Ward, and C. Nokes. 2022. Cryptosporidium surrogate removal in pilot-scale rapid sand filters

8-7


-------
comprising anthracite, pumice or engineered ceramic granular media, and its correlation with
turbidity. Journal of Water Process Engineering, 46: 102614.

Patterson, M.E., Z. Bohrerova, T.A. Wolfe, and L.K. Weavers. 2021. Developing Design Criteria
for Emerging Technologies in Ohio, Journal AWWA, 113, September 2021, 24-34.

Pedati, C., S. Koirala, T. Safranek, B.F. Buss, and A.V. Carlson. 2019. Campylobacteriosis
outbreak associated with contaminated water supply - Nebraska, 2017. Centers of Disease
Control and Prevention, Morbidity and Mortality Weekly Report, February 22, 2019.

Piche, A., A. Campbell, S. Cleary, I. Douglas, and O.D. Basu. 2019. Investigation of backwash
strategy on headloss development and particle release in drinking water biofiltration. Journal of
Water Process Engineering, 32: 100895.

Pitkanen, T. 2013. Review of Campylobacter spp. In drinking and environmental waters. Journal
of Microbiological Methods. 95: 39-47.

Platkin, C. 2019a. Aircraft Water Study. Diet Detective. New York City Food Policy Center,
September 13, 2019.

Platkin, C. 2019b. Background Aircraft Drinking Water Rule and Other Helpful Information,

Diet Detective, New York City Food Policy Center, August 29, 2019.

Puzon, G.J., H.C. Miller, N. Malinowski, T. Walsh, and M.J. Morgan. 2020. Naegleria fowleri in
drinking water distribution systems. Current Opinion in Environmental Science & Health, 16:
22-27.

Ramsay, L., F. Du, M. Lund, H. He, and D.A. S0borg. 2021. Grain Displacement during
Backwash of Drinking Water Filters. Water Supply, 21(1): 356-367

Reynolds, K.A., K.D. Mena, and C.P. Gerba. 2008. Risk of waterborne illness via drinking water
in the United States. Reviews of Environmental Contamination and Toxicology, 192: 117-158.

Samineni, L., B. Xiong, R. Chowdhury, A. Pei, L. Kuehster, H. Wang, R. Dickey, P.E. Soto, L.
Massenburg, T.H. Nguyen, C. Maranas, D. Velegol, M. Kumar, and S. Velegol. 2019. 7 Log
Virus Removal in a Simple Functionalized Sand Filter. Environmental Science & Technology,
53(21): 12706-12714.

Schaeffer, F., K. Tower, and A.S. Weiss'eld. 2012. What's Up with Aircraft Drinking
Water? Clinical Microbiology Newsletter, 34(2): 9-13.

Schijven, J., J. Derx, A.M.R. Husman, A.P. Blaschke, and A. Farnleitner. 2015. QMRAcatch:
Microbial Quality Simulation of Water Resources including Infection Risk Assessment. Journal
of Environmental Quality, 44: 1491-1502.

Schmidt, P.J. and C.L. Chappell. 2016. Comment on "Cryptosporidium Infection Risk: Results
of New Dose-Response Modeling" — Discussion of Underlying Assumptions and Their
Implications. Risk Analysis, 36(12): 2189-2192.

8-8


-------
Schmidt, P.J.; W.B. Anderson, and M.B. Emelko. 2020. Describing water treatment process
performance: Why average log reduction can be a misleading statistic. Water Research, 176:
115702.

Selamat, R., I. Abustan, M.R. Arshad, and N.H.M. Kamal. 2019. Removal of Escherichia Coli
Using Low-Frequency Electromagnetic Field in Riverbank Filtration, chapter in Water and
Wastewater Treatment. London, UK: IntechOpen.

Shaheen, M., C. Scott, and N.J. Ashbolt. 2019. Long-term persistence of infectious Legionella
with free-living amoebae in drinking water biofilms. International Journal of Hygiene and
Environmental Health, 222(4): 678-686.

Sharma, P.R., S.K. Sharma, T. Lindstrom, and B.S. Hsiao. 2020. Nanocellulose-Enabled
Membranes for Water Purification: Perspectives. Advanced Sustainable Systems, 4:5: 1900114.

Sholtes, K. and K.G. Linden. 2019. Pulsed and continuous light UV LED: microbial inactivation,
electrical and time efficiency. Water Research, 165: 114965.

Sholtes, K.A., K. Lowe, G.W. Walters, M.D. Sobsey, K.G. Linden,and L.M. Casanova. 2016.
Comparison of ultraviolet light-emitting diodes and low-pressure mercury-arc lamps for
disinfection of water. Environmental Technology, 37(17): 2183-2188.

Silva, H.G. and L.P. Sabogal-Paz. 2020. Filter backwash water and floated residue containing
pathogenic protozoa: detection method and treatment alternatives. Water, Air & Soil Pollution,
231: 129.

Singh, S.P., Y. Li, J. Zhang, J.M. Tour, and C.J. Arnusch. 2018. Sulfur-Doped Laser-Induced
Porous Graphene Derived from Polysulfone-Class Polymers and Membranes. ACSNano, 12(1);
289-297.

Soller, J. A. 2009. The potential implications of person-to-person transmission of viral infection
for US EPA's Groundwater Rule. J Water Health, 7(2): 208-223.

Sommer, R., T. Haider, A. Cabaj, W. Pribil, and M. Lhotsky. 1998. Time dose reciprocity in UV
disinfection of water. Water Science and Technology, 38: 145-150.

Song, K, F. Taghipour, and M. Mohseni. 2018. Microorganisms inactivation by continuous and
pulsed irradiation of ultraviolet light-emitting diodes (UV-LEDs). Chemical Engineering
Journal, 343: 362-370.

Song, K, F. Tahipour, and M. Mohseni. 2019. Microorganisms inactivation by wavelength
combinations of ultraviolet light-emitting diodes (UV LEDs). Science of the Total Environment,
665: 1103-1110.

Stiff, R.E., A.P. Davies, B.W. Mason, H.A. Hutchings, and R.M. Chalmers. 2017. Long-term
health effects after resolution of acute Cryptosporidium parvum infection: a 1-year follow-up of
outbreak-associated cases. Journal of Medical Microbiology, 66(11): 1607-1611.

8-9


-------
Stokdyk, J.P., S.K. Spencer, J.F. Walsh, J.R. de Lambert, A.D. Firnstahl, A.C. Anderson, L.-I.
Rezania, and M.A. Borchardt. 2019. Cryptosporidium Incidence and Surface Water Influence of
Groundwater Supplying Public Water Systems in Minnesota, USA. Environmental Science and
Technology. 53(7): 3391-3398.

Stokdyk, J.P., A.D. Firnstahl, J.F. Walsh, S.K. Spencer, J.R. de Lambert, A.C. Anderson, L.W.
Rezania, B.A. Kieke, Jr., and M.A. Borchardt. 2020. Viral, bacterial, and protozoan pathogens
and fecal markers in wells supplying groundwater to public water systems in Minnesota, USA.
Water Res., 178: 115814.

Szabo, J., M. Rodgers, J. Mistry, J. Steenbock, and J. Hall, 2019. The effectiveness of
disinfection and flushing procedures to prevent coliform persistence in aircraft water systems.
Water Supply, 19(5): 1339-1346.

Urfer, D. 2017. Use of bauxite for enhanced removal of bacteria in slow sand filters. Water
Science & Technology: Water Supply, 17(4): 1007-1015.

United States Environmental Protection Agency (USEPA). 1998. National Primary Drinking
Water Regulations: Interim Enhanced Surface Water Treatment; Final Rule. 63 FR 69477.
December 16, 1998.

USEPA. 2000. Surface Water Analytical Tool (SWAT): Program Description and Assumptions,
Version 1.1, Office of Ground Water and Drinking Water, Standards and Risk Reduction
Division, November 2000, prepared with Malcom Pirnie under Cadmus Group.

USEPA. 2002. National Primary Drinking Water Regulations: Long Term 1 Enhanced Surface
Water Treatment Rule. 67 FR 1812. January 14, 2002.

USEPA. 2003. National Primary Drinking Water Regulations; Announcement of Completion of
EPA's Review of Existing Drinking Water Standards. 68 FR 42907. July 18, 2003.

USEPA. 2005a. Method 1622: Cryptosporidium in Water by Filtration/IMS/FA. EPA 815-R-
05-001. December 2005.

USEPA. 2005b. Method 1623: Cryptosporidium and Giardia in Water by Filtration/IMS/FA.
EPA 815-R-05-002. December 2005.

USEPA. 2005c. Membrane Filtration Guidance Manual. EPA 815-R-06-009. November 2005.

USEPA. 2006a. National Primary Drinking Water Regulations: Long-Term 2 Enhanced Surface
Water Treatment Rule; Final Rule. 71 FR 654. January 5, 2006.

USEPA. 2006b. National Primary Drinking Water Regulations: Ground Water Rule; Final Rule.
71 FR 65574. November 8, 2006.

USEPA. 2006c. Source Water Monitoring Guidance for Public Water Systems. EPA 815-R-06-
005. February 2006.

8-10


-------
USEPA. 2006d. Ultraviolet Disinfection Guidance Manual for the Final Long Term 2 Enhanced
Surface Water Treatment Rule. EPA 815-R-06-007. November 2006.

USEPA. 2009. National Primary Drinking Water Regulations: Drinking Water Regulations for
Aircraft Public Water Systems. 74 FR 53590. October 19, 2009.

USEPA. 2010a. National Primary Drinking Water Regulations; Announcement of the Results of
EPA's Review of Existing Drinking Water Standards and Request for Public Comment and/or
Information on Related Issues. 75 FR 15499. March 29, 2010.

USEPA. 2010b. Long Term 2 Enhanced Surface Water Treatment Rule Toolbox Guidance
Manual. EPA 815-R-09-016. April 2010.

USEPA. 2010c. Interim Final Guidance Manual for the Aircraft Drinking Water Rule. EPA 816-
R-10-020. October 2010.

USEPA. 2012. Environmental Technology Verification Protocol: Protocol for the Product
Specific Challenge Testing of Bag and Cartridge Filter Systems, Drinking Water Systems
Center, 12/01/EPADWCTR, prepared by NSF International, March 2012.

USEPA. 2013. National Primary Drinking Water Regulations: Revisions to the Total Coliform
Rule; Final Rule. 78 FR 10269. February 13, 2013.

USEPA. 2014a. National Primary Drinking Water Regulations: Minor Corrections to the
Revisions to the Total Coliform Rule. 79 FR 10665. February 26, 2014.

USEPA. 2014b. Revised Total Coliform Rule Assessments and Corrective Actions Guidance
Manual: Interim Final. EPA 815-R-14-006. September 2014.

USEPA. 2016a. Six-Year Review 3 Technical Support Document for Microbial Contaminant
Regulations. EPA-810-R16-010. December 2016.

USEPA. 2016b. Six-Year Review 3 Technical Support Document for Long-Term 2 Enhanced
Surface Water Treatment Rule. EPA-810-R-16-011. December 2016.

USEPA. 2016c. Revised Total Coliform Rule (RTCR) Data Entry Instructions with Examples.
EPA 816-B-16-005. December 2016.

USEPA. 2016d. Quick Guide to Drinking Water Sample Collection. September 2016.

USEPA. 2019a. Information Collection Request Submitted to OMB for Review and Approval;
Comment Request; Contaminant Occurrence Data in Support of the EPA's Fourth Six-Year
Review of National Primary Drinking Water Regulations. 84 FR 58381. October 31, 2019.

USEPA. 2019b. Safe Drinking Water Information System (SDWIS) Federal Reports Advanced
Search. Available on the Internet at: https://ordspub.epa.gov/ords/sfdw/f?p=108:l:::NO:l::.

8-11


-------
USEPA. 2019c. Generating High-Quality Turbidity Data in Drinking Water Treatment Plants to
Support System Optimization and Monitoring. EPA 815-B-19-010. June 2019.

USEPA. 2019d. How to Conduct a Sanitary Survey of Drinking Water Systems. EPA 816-R-17-
001. August 2019.

USEPA. 2020a. Guidance Manual for Compliance with the Surface Water Treatment Rules:
Turbidity Provisions. EPA 815-R-20-004. June 2020

USEPA. 2020b. Innovative Approaches for Validation of Ultraviolet Disinfection Reactors for
Drinking Water Systems. EPA 600-R-20-094. April 2020.

USEPA. 2020c. The Revised Total Coliform Rule (RTCR) State Implementation Guidance—
Final. EPA 816-R-20-003. June 2020.

USEPA. 2021. Understanding the significance and potential growth of pathogens in pipes water
systems, https://www.epa.eov/water-research/imderstandine-sienificance-and-potential-erowth-
pathogens-piped-water-svstems. Page last updated: August 19, 2021.

USEPA. 2022a. Drinking Water Instrumentation Data Integrity Checklists. EPA 815-B-22-002.
January 2022.

USEPA. 2022b. Ultraviolet (UV) Treatment Toolkit: Technical Resource for States using EPA 's
Ultraviolet Disinfection Guidance Manual to Evaluate UV Technology. EPA 815-B-21-007. May
2022.

USEPA. 2022c. ADWR Compliance Public Reports - Aircraft Drinking Water Rule System
Operations, https://sdwis.epa.eov/ords/arcs/f?p=l30:109. Accessed Online May 23, 2022.

USEPA. 2023. Drinking Water Distribution System Tools and Resources.

https://www.epa.eov/dwTeeinfo/drinkine-water-distribiition-SYStem-tools-and-resources.
Accessed Online April 18, 2022. Site last updated: April 12, 2023

U SEP A. 2024a. EPA Protocol for the Fourth Review of Existing National Primary Drinking
Water Regulations. EPA-815-R-24-018. February 2024.

USEPA. 2024b. The Data Management and Quality Assurance/Quality Control Process for the
Fourth Six-Year Review Information Collection Rule Dataset. EPA-815-R-24-017. February
2024.

Venis, R.A. and O.D. Basu. 2021. Mechanisms and efficacy of disinfection in ceramic water
filters: A critical review. Environmental Science and Technology, 51(24): 2934-2974.

Vickers, J.C. 2018. Proposed methodology for determining reverse osmosis unit integrity for
pathogen removal credit. Journal-American Water Works Association, 110(8), E27-E41.

8-12


-------
Wallender, E.K., E.C. Ailes, J.S. Yoder, V.A. Roberts, and J.M. Brunkard. 2014. Contributing
factors to disease outbreaks associated with untreated groundwater. Ground Water. 52(6): 886-
97.

Wang, Y., E. Araud, J.L. Shisler, T.H. Nguyen, and B. Yuan. 2019. Influence of algal organic
matter on MS2 bacteriophage inactivation by ultraviolet irradiation at 220 nm and 254 nm.
Chemosphere, 214: 195-202.

World Health Organization (WHO). 2004. Guidelines for Drinking-Water Quality, Third
Edition. Volume 1: Recommendations. https://www.who.int/publications/i/item/978(

WHO. 2009. Guide to Hygiene and Sanitation in Aviation, 3rd Edition.

https://www.who.int/piiblications/i/item/978i 7772.

Wilson, B.R., P.F. Roessler, E. Van Dellen, M. Abbaszadegan, and C.P. Gerba. 1992. Coliphage
MS-2 as a UV water disinfection efficacy test surrogate for bacterial and viral pathogens.
Proceedings of the Water Quality Technology Conference, Toronto, November 15-19, pgs.
219-235.

Wolf, C., A. Pavese, U. von Gunten, and T. Kohn. 2019. Proxies to monitor the inactivation of
viruses by ozone in surface water and wastewater effluent. Water Research, 166: 115088.

Wright, H.B., D. Gaithuma, T. Dzurny, C. Schulz, A. Cabaj, and A. Schmalweiser. 2006. Design
and performance guidelines for UV sensor systems. American Water Works Association
Research Foundation, Denver, CO (in preparation).

Xiong, B., B. Piechowicz, Z. Wang, R. Marinaro, E. Clement, T. Carlin, A1 Uliana, M. Kumar,
and S.B. Velegol. 2018. Moringa oleifera f-sand Filters for Sustainable Water Purification.
Environmental Science & Technology Letters, 5(1): 38-42.

Younis, B.A., L. Mahoney, W. Schweigkofler, and K. Suslow. 2019. Inactivation of plant
pathogens in irrigation water runoff using a novel UV disinfection system. European Journal of
Plant Pathology, 153(3): 907-914.

Zyara, A.M., E. Torvinen, A-M Veijalainen, and H. Heinonen-Tanski. 2016. The Effect of UV
and Combined Chlorine/UV Treatment on Coliphages in Drinking Water Disinfection. J Water
Health, 14 (4): 640-649.

8-13


-------

-------
9 List of Appendices

Appendix A. Additional Analyses on the Six-Year Review 4 Microbial Data

Appendix B. Additional Analyses on the Aircraft Drinking Water Rule Data

Appendix C. Revised Total Coliform Rule Level 1 and Level 2 Assessment Characterization
and Data Quality Considerations

Appendix D. Revised Total Coliform Rule Corrective Actions and Assessment of Data Quality

9-1


-------

-------
Appendix A. Additional Analyses on the Six-Year Review 4 Microbial Data

This appendix provides additional summaries of the SYR4 microbial data that were not
presented within the body of Chapter 6.

Exhibit A-l presents a state-level summary of the total coliform, E. coll, fecal coliform, and
Cryptosporidium data from the SYR4 ICR dataset that passed QA and were included as part of
the SYR4 analyses presented in various analyses of Section 6. For the SYR4 process, EPA
evaluated total coliforms in finished water, E. coli / fecal coliforms in finished water and raw
water, and Cryptosporidium in source (raw) water for binning purposes. The E. coli / fecal
coliform results are presented separately for raw and finished water since different data were
used to support different analyses (i.e., finished water data as part of RTCR and raw water data
as part of GWR). The other analytes are not presented separately for raw water and finished
water results.

Exhibit A-1. Summary of Six-Year Review 4 Microbial Data, By State12

State

Total
Coliforms

E. coli

Fecal Coliforms

Cryptosporidium

Finished

Raw

Unknown

Total

Finished

Raw

Unknown

Total

AK

103,898

10,395

1,353

53,666

65,414

175

1

2647

2,823

0

AL

284,580

27,620

61,560

1,470

90,650

5

1

0

6

2,310

AR

394,314

5,468

580

41

6,089

0

0

0

0

0

AS

13,186

0

0

13,184

13,184

0

0

0

0

0

AZ

219,468

37,728

5,074

60

42,862

24

2

0

26

728

CO

352,349

0

0

204,889

204,889

0

0

24

24

2,012

CT

382,725

110,026

32,726

77,102

219,854

2

2

10

14

1,053

DC

13,693

1,803

48

7,797

9,648

0

0

0

0

48

DE

70,366

12,524

516

2

13,042

2

1

0

3

0

FL

2,342,672

0

0

350

350

0

0

21

21

0

HI

16,035

13,332

219

42

13,593

13

0

0

13

0

IA

425,813

200,743

79

6,465

207,287

2

0

1

3

498

ID

193,935

333

761

13,357

14,451

1

0

2

3

159

IL

1,526,019

365,549

273,277

12,218

651,044

49

113

73

235

1,494

IN

398,481

12,059

1,643

0

13,702

0

0

0

0

0

KS

279,741

203,682

5,221

59

208,962

11

0

0

11

927

KY

427,911

1,742

64

143

1,949

0

0

0

0

0

LA

179,619

142,670

4,388

359

147,417

11

3

0

14

0

MD

60,832

32,435

1,388

258

34,081

1,091

1

0

1,092

0

ME

145,575

73,458

3,921

379

77,758

2

0

0

2

56

MN

225,927

0

0

15,141

15,141

0

0

12

12

416

MO

601,095

213,536

2,287

67,050

282,873

1

0

0

1

0

MP

13,364

0

0

12,020

12,020

0

0

0

0

0

MT

260,675

166,580

34,956

15,116

216,652

4,238

490

214

4,942

0

NC

926,048

625,697

2,385

268

628,350

4

0

0

4

0

ND

95,674

804

71

61

936

1

0

0

1

0

NE

218,891

153,709

12

187

153,908

0

0

0

0

0

NH

155,791

0

0

156,191

156,191

0

0

0

0

0

A-l


-------
State

Total
Coliforms

E. coli

Fecal Coliforms

Cryptosporidium

Finished

Raw

Unknown

Total

Finished

Raw

Unknown

Total

NJ

935,126

13,886

8,109

689

22,684

55

9

0

64

793

NN

7,447

0

0

6,789

6,789

0

0

0

0

0

NV

81,129

10,045

730

2,724

13,499

0

0

0

0

359

NY

541,960

68,925

11,433

7,874

88,232

104

319

15

438

209

OH

1,022,164

107,099

4,226

1,443

112,768

106

6

0

112

9

OK

398,661

228,441

6,913

1,432

236,786

0

0

0

0

1,235

OR

477,951

8,062

8,000

16

16,078

0

1

0

1

821

PA

854,438



11,108

235,709

246,817

0

6

724

730

0

Rl

61,041

5,831

5,062

33,985

44,878

216

117

1,459

1,792

166

SC

9,563

772

2,602

4,136

7,510

0

0

2

2

0

SD

117,852

0

0

66,507

66,507

0

0

0

0

0

TN

91,984

0

0

84

84

0

0

1,449

1,449

0

TX

2,637,545

1,265,898

77,248

15,976

1,359,122

1,064

62

6

1,132

693

UT

297,343

79,741

11,742

769

92,252

5

5

0

10

417

VA

703,226

307,532

35,071

754

343,357

131

19

0

150

2,467

VT

126,345

95,904

2,612

7,968

106,484

1

0

0

1

192

WA

949,429

224,814

8

0

224,822

191

0

0

191

0

Wl

693,211

0

0

545,150

545,150

0

0

0

0

420

WV

187,869

966

2,505

611

4,082

4

0

7

11

1,716

WY

108,011

76,445

1,199

10,042

87,686

482

0

927

1,409

213

01

2,722

0

0

2,708

2,708

0

0

0

0

0

02

912

0

0

84

84

0

0

0

0

0

04

3,591

24

33

0

57

0

3

0

3

71

05

19,648

98

6

41

145

1

0

0

1

0

06

21,655

9,490

164

486

10,140

43

0

4

47

11

07

2,468

2,029

3

205

2,237

0

0

0

0

0

08

21,291

10,754

811

2,175

13,740

22

0

2

24

9

09

21,764

0

0

17,844

17,844

0

0

0

0

40

10

21,089

46

189

289

524

0

1

0

1

0

Total

20,746,112

4,928,695

622,303

1,624,365

7,175,363

8057

1162

7599

16,818

19,542

1	Six-Year Review 4 microbial data that passed quality assurance on which the Section 6 analyses were
based.

2	Under SYR4, very limited data were submitted for coliphage (3 records) and enterococci (8 records).

As part of the GWR and RTCR analysis, EPA evaluated annual trends through consideration of
"common systems" across the years, as well as only those systems with a high level of
completeness. "Common systems" refers to systems with data in all years. The inclusion of only
these systems eliminates year-to-year variation in the number of systems in the analysis. To
further reduce the annual variability in the underlying data, EPA also focused many of the
analyses on only those system-months with at least 90 percent of completeness of routine total
coliform monitoring records (i.e., those system months where a system collected at least 90
percent of their required routine total coliform samples based on system size). As shown in
Exhibit A-2, the focus on only the "common systems with 90% completeness" excludes a
significant portion of systems from the analysis.

A-2


-------
Exhibit A-2 Comparison of Summary of the Maximum Percentage Variation in the
Number of Systems and the Number of Annual Routine Total Coliform Samples
for all Public Water Systems without 90% completeness and Common Systems

with 90% completeness



#System

s or
#RTTC

2007

2008

2010

2011

2014

2015

2018

2019

Max %
Variation

All PWSs

for
Individual
Years

#Sy stems

70,685

70,278

74,953

74,691

86,237

85,415

95,516

95,099

35.9%

#RTTC
(without

90%
Complete
ness)

1,226,098

1,327,476

1,589,336

1,647,311

1,861,738

1,883,681

2,231,731

2,304,040

87.9%

Common
PWSs for
All Years

#Sy stems

48,292

48,292

48,292

48,292

48,292

48,292

48,292

48,292

0%

#RTTC
(with 90%
Complete
ness)

897,301

964,469

962,858

972,586

960,727

968,790

963,481

961,707

8.4%

Exhibit A-3 through Exhibit A-5 present a summary of the changes in the percent of ground
water systems with disinfection for various system sizes with common systems across the years
with records with 90% completeness. These results presented separately for systems of differing
sizes correspond to the analysis presented in Exhibit 6-7 for all system sizes.

Exhibit A-3. Changes in Percent of GW Systems with Disinfection Serving Fewer
than 1,000 People (Common systems with 8 years of data and >= 90%

completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

39,070

39,070

39,070

39,070

39,070

39,070

39,070

39,070

#Disinfecting Systems

















#Non Disinfecting
Systems

22,178

23,332

20,918

22,217

21,100

21,704

19,825

19,790

%disinfecting Systems

43.24%

40.28%

46.46%

43.14%

45.99%

44.45%

49.26%

49.35%

%disinfecting Systems
(Ave @ 2 years)

41.76%

44.80%

45.22%

49.30%

Relative Change



7.28%

0.95%

9.02%



Overall Change

18.07%

A-3


-------
Exhibit A-4. Changes in Percent of GW Systems with Disinfection Serving 1,000
People or More but Less Than 10,000 People (Common systems with 8 years of

data and >= 90% completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

3,455

3,455

3,455

3,455

3,455

3,455

3,455

3,455

#Disinfecting Systems

















#Non Disinfecting
Systems

568

734

401

502

331

324

264

258

%disinfecting Systems

83.56%

78.76%

88.39%

85.47%

90.42%

90.62%

92.36%

92.53%

%disinfecting Systems
(Ave @ 2 years)

81.16%

86.93%

90.52%

92.45%

Relative Change



7.11%

4.13%

2.13%



Overall Change

13.91%

Exhibit A-5. Changes in Percent of GW Systems with Disinfection Serving 10,000
People or More (Common systems with 8 years of data and >= 90%

completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

297

297

297

297

297

297

297

297

#Disinfecting Systems

















#Non Disinfecting
Systems

9

16

9

12

7

5

4

5

%disinfecting Systems

96.97%

94.61%

96.97%

95.96%

97.64%

98.32%

98.65%

98.32%

%disinfecting Systems
(Ave @ 2 years)

95.79%

96.46%

97.98%

98.48%

Relative Change



0.70%

1.57%

0.52%



Overall Change

2.81%

Exhibit A-6 through Exhibit A-8 present a summary of the annual total coliform detection rates
for disinfecting and undisinfected ground water systems for various system sizes. These results
presented separately for systems of differing sizes correspond to the analysis presented in Exhibit
6-10 for all system sizes.

A-4


-------
Exhibit A-6. Changes of %RTTC+ Rates among Disinfecting and Undisinfected
Systems Serving Fewer than 1,000 People ("Common Systems" with 90%

Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019



















#RTTC

174,235

181,321

202,355

191,731

203,224

197,845

203,058

203,789

#RTTC+

3,742

3,744

4,018

3,437

3,176

3,456

3,121

3,099

%RTTC+

2.15%

2.06%

1.99%

1.79%

1.56%

1.75%

1.54%

1.52%

Ave @ 2 years

2.11%

1.89%

1.65%

1.53%

Relative difference



-10.31%

-12.40%

-7.61%



Overall Difference

-27.41%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Non Disinfecting Systems

22,178

23,332

20,918

22,217

21,100

21,704

19,825

19,790

#RTTC

151,547

168,483

141,403

156,526

140,720

147,426

137,882

136,766

#RTTC+

6,033

6,313

5,755

6,055

5,090

5,135

3,767

3,672

%RTTC+

3.98%

3.75%

4.07%

3.87%

3.62%

3.48%

2.73%

2.68%

Ave @ 2 years

3.86%

3.97%

3.55%

2.71%

Relative difference



2.72%

-10.56%

-23.71%



Overall Difference

-29.90%

Exhibit A-7. Changes of %RTTC+ Rates among Disinfecting and Undisinfected
Systems Serving 1,000 People or More but Less Than 10,000 People ("Common

Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019



















#RTTC

98,373

106,355

119,597

118,509

125,074

126,416

126,541

125,988

#RTTC+

692

671

673

594

643

691

662

587

%RTTC+

0.70%

0.63%

0.56%

0.50%

0.51%

0.55%

0.52%

0.47%

Ave @ 2 years

0.67%

0.53%

0.53%

0.49%

Relative difference



-20.26%

-0.31%

-6.75%



Overall Difference

-25.88%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Non Disinfecting Systems

568

734

401

502

331

324

264

258

#RTTC

16,164

23,743

10,798

14,747

8,228

8,165

6,274

6,471

#RTTC+

218

296

172

177

137

186

115

131

Ave @ 2 years

1.35%

1.25%

1.59%

1.20%

1.67%

2.28%

1.83%

2.02%

Ave @ 2 years

1.30%

1.40%

1.97%

1.93%

Relative difference



7.62%

41.17%

-2.17%



Overall Difference

48.63%

A-5


-------
Exhibit A-8. Changes of %RTTC+ Rates among Disinfecting and Undisinfected
Systems Serving 10,000 People or More ("Common Systems" with 90%

Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019



















#RTTC

82,688

86,459

89,623

89,406

89,390

91,075

91,590

92,184

#RTTC+

222

176

126

136

180

175

188

199

%RTTC+

0.27%

0.20%

0.14%

0.15%

0.20%

0.19%

0.21%

0.22%

Ave @ 2 years

0.24%

0.15%

0.20%

0.21%

Relative difference



-37.99%

34.44%

7.02%



Overall Difference

-10.78%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Non Disinfecting Systems

9

16

9

12

7

5

4

5

#RTTC

1,694

2,540

2,016

2,104

1,776

1,266

880

1,250

#RTTC+

8

12

7

15

17

17

15

3

%RTTC+

0.47%

0.47%

0.35%

0.71%

0.96%

1.34%

1.70%

0.24%

Ave @ 2 years

0.47%

0.53%

1.15%

0.97%

Relative difference



12.22%

116.95%

-15.46%



Overall Difference

105.84%

Exhibit A-9 through Exhibit A-l 1 present a summary of the annual E. coli detection rates for
disinfecting and undisinfected ground water systems for various system sizes. These results
presented separately for systems of differing sizes correspond to the analysis presented in Exhibit
6-11 for all system sizes.

A-6


-------
Exhibit A-9. Changes of %RTEC+ Rates among Disinfecting and Undisinfected
Systems Serving Fewer than 1,000 People ("Common Systems" with 90%

Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019



















#RTTC

174,235

181,321

202,355

191,731

203,224

197,845

203,058

203,789

#RTEC+

178

180

175

154

121

108

149

143

%RTEC+

0.10%

0.10%

0.09%

0.08%

0.06%

0.05%

0.07%

0.07%

Ave @ 2 years

0.10%

0.08%

0.06%

0.07%

Relative difference



-17.19%

-31.58%

25.78%



Overall Difference

-28.74%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Non Disinfecting Systems

22,178

23,332

20,918

22,217

21,100

21,704

19,825

19,790

#RTTC

151,547

168,483

141,403

156,526

140,720

147,426

137,882

136,766

#RTEC+

268

263

214

219

200

167

160

121

%RTEC+

0.18%

0.16%

0.15%

0.14%

0.14%

0.11%

0.12%

0.09%

Ave @ 2 years

0.17%

0.15%

0.13%

0.10%

Relative difference



-12.52%

-12.31%

-19.93%



Overall Difference

-38.57%

Exhibit A-10. Changes of %RTEC+ Rates among Disinfecting and Undisinfected
Systems Serving 1,000 People or More but Less Than 10,000 People ("Common

Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Disinfecting Systems

















#RTTC

98,373

106,355

119,597

118,509

125,074

126,416

126,541

125,988

#RTEC+

33

28

38

27

31

33

27

23

%RTEC+

0.03%

0.03%

0.03%

0.02%

0.02%

0.03%

0.02%

0.02%

Ave @ 2 years

0.03%

0.03%

0.03%

0.02%

Relative difference



-8.88%

-6.72%

-22.20%



Overall Difference

-33.87%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Non Disinfecting Systems

568

734

401

502

331

324

264

258

#RTTC

16,164

23,743

10,798

14,747

8,228

8,165

6,274

6,471

#RTEC+

6

5

6

4

1

1

0

1

%RTEC+

0.04%

0.02%

0.06%

0.03%

0.01%

0.01%

0.00%

0.02%

Ave @ 2 years

0.03%

0.04%

0.01%

0.01%

Relative difference



42.13%

-70.49%

-36.67%



Overall Difference

-73.44%

A-7


-------
Exhibit A-11. Changes of %RTEC+ Rates among Disinfecting and Undisinfected
Systems Serving 10,000 People or More ("Common Systems" with 90%

Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019



















#RTTC

82,688

86,459

89,623

89,406

89,390

91,075

91,590

92,184

#RTEC+

6

9

8

5

6

7

5

6

%RTEC+

0.01%

0.01%

0.01%

0.01%

0.01%

0.01%

0.01%

0.01%

Ave @ 2 years

0.01%

0.01%

0.01%

0.01%

Relative difference



-17.81%

-0.83%

-16.88%



Overall Difference

-32.25%

Year

2007

2008

2010

2011

2014

2015

2018

2019

#Non Disinfecting Systems

9

16

9

12

7

5

4

5

#RTTC

1,694

2,540

2,016

2,104

1,776

1,266

880

1,250

#RTEC+

1

0

0

0

1

0

0

0

%RTEC+

0.06%

0.00%

0.00%

0.00%

0.06%

0.00%

0.00%

0.00%

Ave @ 2 years

0.03%

0.00%

0.03%

0.00%

Relative difference



-100.00%

'



-100.00%



Overall Difference

-100.00%

Exhibit A-12 through Exhibit A-14 present a summary of the annual total coliform detection
rates for all ground water systems for various system sizes. These results presented separately for
systems of differing sizes correspond to the analysis presented in Exhibit 6-8 for all system sizes.

Exhibit A-12. Changes of %RTTC+ Rates among All Ground Water Systems
Serving Fewer than 1,000 People ("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

39,070

39,070

39,070

39,070

39,070

39,070

39,070

39,070

#RTTC

325,782

349,804

343,758

348,257

343,944

345,271

340,940

340,555

#RTTC+

9,775

10,057

9,773

9,492

8,266

8,591

6,888

6,771

%RTTC+

3.00%

2.88%

2.84%

2.73%

2.40%

2.49%

2.02%

1.99%

Ave @ 2 years

2.94%

2.78%

2.45%

2.00%

Relative difference



-5.22%

-12.16%

-18.05%



Overall Difference

-31.78%

A-8


-------
Exhibit A-13. Changes of %RTTC+ Rates among All Ground Water Systems
Serving 1,000 People or More but Less Than 10,000 People ("Common Systems"

with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

3,455

3,455

3,455

3,455

3,455

3,455

3,455

3,455

#RTTC

114,537

130,098

130,395

133,256

133,302

134,581

132,815

132,459

#RTTC+

910

967

845

771

780

877

777

718

%RTTC+

0.79%

0.74%

0.65%

0.58%

0.59%

0.65%

0.59%

0.54%

Ave @ 2 years

0.77%

0.61%

0.62%

0.56%

Relative difference



-20.24%

0.83%

-8.87%



Overall Difference

-26.71%

Exhibit A-14. Changes of %RTTC+ Rates among All Ground Water Systems
Serving 10,000 People or More ("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

297

297

297

297

297

297

297

297

#RTTC

84,382

88,999

91,639

91,510

91,166

92,341

92,470

93,434

#RTTC+

230

188

133

151

197

192

203

202

%RTTC+

0.27%

0.21%

0.15%

0.17%

0.22%

0.21%

0.22%

0.22%

Ave @ 2 years

0.24%

0.16%

0.21%

0.22%

Relative difference



-35.90%

36.72%

2.76%



Overall Difference

-9.94%

Exhibit A-15 through Exhibit A-17 present a summary of the annual E. coli detection rates for all
ground water systems for various system sizes. These results presented separately for systems of
differing sizes correspond to the analysis presented in Exhibit 6-9 for all system sizes.

Exhibit A-15. Changes of %RTEC+ Rates among All Ground Water Systems
Serving Fewer than 1,000 People ("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

39,070

39,070

39,070

39,070

39,070

39,070

39,070

39,070

#RTTC

325,782

349,804

343,758

348,257

343,944

345,271

340,940

340,555

#RTEC+

446

443

389

373

321

275

309

264

%RTEC+

0.14%

0.13%

0.11%

0.11%

0.09%

0.08%

0.09%

0.08%

Ave @ 2 years

0.13%

0.11%

0.09%

0.08%

Relative difference



-16.42%

-21.47%

-2.79%



Overall Difference

-36.20%

A-9


-------
Exhibit A-16. Changes of %RTEC+ Rates among All Ground Water Systems
Serving 1,000 People or More but Less Than 10,000 People ("Common Systems"

with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

3,455

3,455

3,455

3,455

3,455

3,455

3,455

3,455

#RTTC

114,537

130,098

130,395

133,256

133,302

134,581

132,815

132,459

#RTEC+

39

33

44

31

32

34

27

24

%RTEC+

0.03%

0.03%

0.03%

0.02%

0.02%

0.03%

0.02%

0.02%

Ave @ 2 years

0.03%

0.03%

0.02%

0.02%

Relative difference



-4.05%

-13.57%

-21.96%



Overall Difference

-35.29%

Exhibit A-17. Changes of %RTEC+ Rates among All Ground Water Systems
Serving 10,000 People or More ("Common Systems" with 90% Completeness)

Year

2007

2008

2010

2011

2014

2015

2018

2019

All Systems

297

297

297

297

297

297

297

297

#RTTC

84,382

88,999

91,639

91,510

91,166

92,341

92,470

93,434

#RTEC+

7

9

8

5

7

7

5

6

%RTEC+

0.01%

0.01%

0.01%

0.01%

0.01%

0.01%

0.01%

0.01%

Ave @ 2 years

0.01%

0.01%

0.01%

0.01%

Relative difference



-22.89%

7.50%

-22.48%



Overall Difference

-35.74%

Exhibit A-18 through Exhibit A-20 present a summary of the average annual total coliform
detection rates for three groups of systems: all PWSs, all disinfecting ground water systems, and
all undisinfected ground water systems. These results presented separately for systems of
differing sizes correspond to the analysis presented in Exhibit 6-12 for all system sizes.

Exhibit A-18. Summary of Changes of %RTTC+ Rates by System Categories
Serving Fewer than 1,000 People (All PWS, Disinfecting Ground Water systems,

Undisinfected Ground Water Systems)

System Types

Year

2007

2008

2010

2011

2014 2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right afterSS under
GWR

Afterfew years of
RTCR

All PWSs

Relative Change



-6.2%

-10.8%

-18.0%



Overall Change

-31.4%

Disinfecting ground
water systems

Relative Change



-10.3%

-12.4%

-7.6%



Overall Change

-27.4%

Undisinfected GW
Systems

Relative Change



2.7%

-10.6%

-23.7%



Overall Change

-29.9%

A-10


-------
Exhibit A-19. Summary of Changes of %RTTC+ Rates by System Categories
Serving 1,000 People or More but Less Than 10,000 People (All PWS, Disinfecting
Ground Water systems, Undisinfected Ground Water Systems)

System Types

Year

2007 | 2008

2010

2011

2014 2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right afterSS under
GWR

Afterfew years of
RTCR

All PWSs

Relative Change



-15.6%

-0.4%

-5.6%



Overall Change

-20.7%

Disinfecting ground
water systems

Relative Change



-20.3%

-0.3%

-6.8%



Overall Change

-25.9%

Undisinfected GW
Systems

Relative Change



7.6%

41.2%

-2.2%



Overall Change

48.6%

Exhibit A-20. Summary of Changes of %RTTC+ Rates by System Categories
Serving 10,000 People or More (All PWS, Disinfecting Ground Water systems,

Undisinfected Ground Water Systems)

System Types

Year

2007

2008

2010

2011

2014 2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right afterSS under
GWR

Afterfew years of
RTCR

All PWSs

Relative Change



-22.7%

16.7%

1.2%



Overall Change

-8.8%

Disinfecting ground
water systems

Relative Change



-38.0%

34.4%

7.0 %



Overall Change

-10.8%

Undisinfected GW
Systems

Relative Change



12.2%

117.0%

-15.5%



Overall Change

105.8%

Exhibit A-21 through Exhibit A-23 present a summary of the average annual E. coli detection
rates for three groups of systems: all PWSs, all disinfecting ground water systems, and all
undisinfected ground water systems. These results presented separately for systems of differing
sizes correspond to the analysis presented in Exhibit 6-13 for all system sizes.

Exhibit A-21. Summary of Changes of %RTEC+ Rates by System Categories
Serving Fewer than 1,000 People (All PWS, Disinfecting Ground Water systems,

Undisinfected Ground Water Systems)

System Types

Year

2007

2008

2010

2011

2014 2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right afterSS under
GWR

Afterfew years of
RTCR

All PWSs

Relative Change



-17.7%

-16.4%

-4.9%



Overall Change

-34.6%

Disinfecting ground
water systems

Relative Change



-17.2%

-31.6%

25.8%



Overall Change

-28.7%

Undisinfected GW
Systems

Relative Change



-12.5%

-12.3%

-19.9%



Overall Change

-38.6%

A-ll


-------
Exhibit A-22. Summary of Changes of %RTEC+ Rates by System Categories
Serving 1,000 People or More but Less Than 10,000 People (All PWS, Disinfecting
Ground Water systems, Undisinfected Ground Water Systems)

System Types

Year

2007 | 2008

2010

2011

2014 2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right afterSS under
GWR

Afterfew years of
RTCR

All PWSs

Relative Change



-9.9%

2.4%

56.3"-.



Overall Change

44.2%

Disinfecting ground
water systems

Relative Change



-8.9%

-6.7%

-22.2',u



Overall Change

-33.9%

Undisinfected GW
Systems

Relative Change

42.1%

-70.5%

-36.7%



Overall Change

-73.4%

Exhibit A-23. Summary of Changes of %RTEC+ Rates by System Categories
Serving 10,000 People or More (All Public Water Systems, Disinfecting Ground
Water systems, Undisinfected Ground Water Systems)

System Types

Year

2007

2008

2010

2011

2014 2015

2018

2019

2-Year Period

Before GWR

Right after GWR

Right afterSS under
GWR

Afterfew years of
RTCR

All PWSs

Relative Change



-48.2%

84.6%

-8.9%



Overall Change

-12.9%

Disinfecting ground
water systems

Relative Change



-17.8%

-0.8%

-16.9%



Overall Change

-32.3%

Undisinfected GW
Systems

Relative Change



-100.0%

#DI V/0!

-100.0%



Overall Change

-100.0%

A-12


-------
Appendix B. Additional Analyses on the Aircraft Drinking Water Rule

Data

This appendix provides the analytical results for the Aircraft Drinking Water Rule (ADWR) data
that were not presented within the body of Chapter 6. Exhibit B-l presents a count of the number
of samples and systems for each air carrier for the years 2012-2019. The average sample count
per air carrier was 2,186 with a median of 389 and a maximum of 22,492. The average system
count per air carrier was 163 with a median of 54 and a maximum of 1,284.

Exhibit B-1. Number of Samples and Aircraft Systems by Air Carrier; 2012-2019

Air Carrier

Count of
Samples

Count of
Systems

AIR WISCONSIN AIRCRAFTS CORPORATION

1,683

71

AIRBORNE ENERGY SOLUTIONS LIMITED

4

1

AIRTRAN AIRWAYS INC

630

134

ALASKA AIRCRAFTS INC

2,848

271

ALLEGIANT AIR LLC

1,383

155

AMERICAN AIRCRAFTS INC

17,098

1,284

AMERISTAR AIR CARGO INC

180

5

ATLAS AIR INC

152

13

CHAUTAUQUA AIRCRAFTS INC

609

74

COLGAN AIR INC

14

6

COMAIR INC

94

34

COMPASS AIRCRAFTS LLC

1,742

62

CONOCOPHILLIPS ALASKA

50

5

DELTA AIR LINES INC

22,492

1,130

ENDEAVOR AIR INC

3,508

272

ENVOY AIR INC

3,886

314

EXECUTIVE AIR CRAFT LTD

22

3

EXPRESSJET AIRCRAFTS INC

4,983

453

FALCON AIR EXPRESS INC

142

8

FRONTIER AIRCRAFTS INC

1,290

143

GOJET AIRCRAFTS LLC

960

59

HAWAII ISLAND AIR INC

4

2

HAWAIIAN AIRCRAFTS INC

1,781

77

HORIZON AIR INDUSTRIES INC

138

30

JETBLUE AIRWAYS CORPORATION

3,851

254

MESA AIRCRAFTS INC

1,988

147

MIAMI AIR INTERNATIONAL INC

130

12

MSG FLIGHT OPERATIONS LLC

26

1

NATIONAL AIR CARGO GROUP INC

0

0

NORTH AMERICAN AIRCRAFTS

20

5

OMNI AIR INTERNATIONAL INC

168

17

ORANGE AIR LLC

6

2

PIEDMONT AIRCRAFTS INC

308

60

PSA AIRCRAFTS INC

2,024

151

REPUBLIC AIRWAYS INC

3,464

268

RYAN INTERNATIONAL AIRCRAFTS INC

6

3

SHUTTLE AMERICA CORPORATION

1,277

112

B-l


-------
Air Carrier

Count of
Samples

Count of
Systems

SIERRA PACIFIC AIRCRAFTS INC

36

4

SKYWEST AIRCRAFTS INC

9,148

541

SONGBIRD AIRWAYS, INC

12

5

SOUTHWEST AIRCRAFTS CO

11,075

920

SPIRIT AIRCRAFTS INC

1,564

147

SUN COUNTRY AIRCRAFTS

469

48

SWIFT AIR LLC

240

33

TEM ENTERPRISES INC

96

15

THE DOW CHEMICAL COMPANY

20

2

TRANS STATES AIRCRAFTS LLC

718

81

UNITED AIRCRAFTS, INC

12,057

897

US AIRWAYS INC

2,716

396

USA JET AIRCRAFTS INC

12

6

VIRGIN AMERICA INC

812

67

VISION AIRCRAFTS INC

32

6

WORLD AIRWAYS INC

14

3

WORLD ATLANTIC AIRCRAFTS

88

9

Total

118,070

7,816

Exhibit B-2 presents a count of the number of systems with total coliform and E. coli data, and
the associated total coliform and E. coli positivity rate, broken down by year for the years 2011-
2021. Approximately 34.8 percent of the systems reported at least one total coliform sample with
a positive result during the period 2011-2021. When looking only at the years 2012-2019,
approximately 32.8 percent of the systems reported at least one total coliform sample with a
positive result.

Approximately 2.74 percent of the systems with E. coli data reported at least one E. coli sample
with a positive result during the period 2011-2021 (2.40 percent when using the second set of E.
coli assumptions). Approximately 2.45 percent of the systems with /•]. coli data reported at least
one E. coli sample with a positive result when considering only the years 2012-2019 (2.12
percent when using the second set of E. coli assumptions).

These results imply that the positive samples are spread among a wide variety of aircraft rather
than a small number of aircraft.

Exhibit B-2. Number of Systems with Total Coliform and E. coli Data, and Total
Coliform and E. coli Positivity Rate by Year; 2011-2021

Year

Total Coliforms

E. coli

E. coli (Alternative Approach)1

# Systems
with Data

#

Systems
with Total
Coliform
Positives

%

Systems
with
Total
Coliform
Positives

# Systems
with Data

#

Systems
with E.

coli
positives

%

Systems
with E.

coli
positives

# Systems
with Data

#

Systems
with E.

coli
positives

%

Systems
with E.

coli
positives

20112

1,992

253

12.70%

1,331

9

0.68%

1,992

9

0.45%

2012

5,429

707

13.02%

4,359

29

0.67%

5,429

29

0.53%

2013

5,508

640

11.62%

3,945

24

0.61 %

5,508

24

0.44%

B-2


-------
Year

Total Coliforms

E. coli

E. coli (Alternative Approach)1

# Systems
with Data

#

Systems
with Total
Coliform
Positives

%

Systems
with
Total
Coliform
Positives

# Systems
with Data

#

Systems
with E.

coli
positives

%

Systems
with E.

coli
positives

# Systems
with Data

#

Systems
with E.

coli
positives

%

Systems
with E.

coli
positives

2014

5,602

584

10.42%

3,924

15

0.38%

5,602

15

0.27%

2015

5,687

576

10.13%

3,871

22

0.57%

5,687

22

0.39%

2016

5,753

602

10.46%

3,865

25

0.65%

5,753

25

0.43%

2017

5,835

414

7.10%

4,120

17

0.41 %

5,835

17

0.29%

2018

6,017

433

7.20%

4,293

22

0.51 %

6,017

22

0.37%

2019

6,105

353

5.78%

4,386

17

0.39%

6,105

17

0.28%

20203

5,538

210

3.79%

3,930

18

0.46%

5,538

18

0.33%

20214

1,926

66

3.43%

1,395

3

0.22%

1,926

3

0.16%

Total
(2011 -
2021)

8,093

2,820

34.84%

7,091

194

2.74%

8,093

194

2.40%

Total
(2012-
2019)

7,816

2,567

32.84%

6,776

166

2.45%

7,816

166

2.12%

1 Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

2	The 2011 data does not represent an entire calendar year as it represents the period of February to December 2011.

3	Due to the COVID-19 pandemic there were a large number of inactive aircraft in comparison to the preceding and following years.

4	The 2021 data does not represent an entire calendar year as it represents the period of January to May 2021.

Exhibit B-3 and Exhibit B-4 presents a count of the number of total coliform and E. coli samples
and the associated total coliform and E. coli positivity rates broken down by size and
manufacturer/model for the years 2012-2019. For total coliforms, an average of 3.6 percent of
samples provided for a given size and manufacturer/model category were positive. The median
was 1.2 percent, with a minimum of 0 percent and a maximum of 25 percent. The 90th percentile
was approximately 11 percent.

For E. coli, an average of 0.2 percent of samples provided for a given size and
manufacturer/model category were positive. The median was 0 percent, the minimum was 0
percent and the maximum was 2.7 percent. The 90th percentile was approximately 0.5 percent.
Theis. coli alternative approach showed an average of 0.1 percent and a 90th percentile of
approximately 0.3 percent.

B-3


-------
Exhibit B-3. Number of Total Coliform and E. coli Samples, and Total Coliform and
E. coli Positives, by Size and Manufacturer and Model; 2012-2019

Si
ze

Manufacturer,
Model

Total Coliforms

E. coli

E. coli (Alternative Approach)1

Total
Samples

#

Positiv
e

%

Positiv
e

Total
Samples

#

Positiv
e

%

Positiv
e

Total
Samples

#

Positiv
e

%

Positiv
e

L

AIRBUS, 330

10

0

0.00%

0

0

0.00%

10

0

0.00%

L

AIRBUS, A330

1,724

37

2.15%

73

2

2.74%

1,723

2

0.12%

L

AIRBUS, A350

58

0

0.00%

0

0

0.00%

58

0

0.00%

L

BOEING, 747

606

34

5.61 %

294

0

0.00%

606

0

0.00%

L

BOEING, 767

1,271

29

2.28%

159

0

0.00%

1,271

0

0.00%

L

BOEING, 777

3,072

35

1.14%

1,799

0

0.00%

3,072

0

0.00%

L

BOEING, 787

222

1

0.45%

204

0

0.00%

222

0

0.00%

L

BOEING,
B747

14

0

0.00%

14

0

0.00%

14

0

0.00%

L

DOUG,
DC1030

0

0

0.00%

0

0

0.00%

0

0

0.00%

L

DOUG, MD11

14

2

14.29%

14

0

0.00%

14

0

0.00%

L

EMB, ERJ170

12

0

0.00%

12

0

0.00%

12

0

0.00%

M

AIRBUS, 320

2,028

41

2.02%

1,368

2

0.15%

2,028

2

0.10%

M

AIRBUS, 321

20

1

5.00%

12

0

0.00%

20

0

0.00%

M

AIRBUS, A220

12

0

0.00%

0

0

0.00%

12

0

0.00%

M

AIRBUS, A231

10

2

20.00%

4

0

0.00%

10

0

0.00%

M

AIRBUS, A319

1,254

24

1.91 %

887

2

0.23%

1,253

2

0.16%

M

AIRBUS, A320

7,502

214

2.85%

4,483

6

0.13%

7,501

6

0.08%

M

AIRBUS, A321

3,448

112

3.25%

1,812

9

0.50%

3,448

9

0.26%

M

AIRBUS, A330

306

5

1.63%

5

0

0.00%

306

0

0.00%

M

BOEING, 737

28,282

1,023

3.62%

22,239

25

0.11%

28,280

25

0.09%

M

BOEING, 747

30

0

0.00%

30

0

0.00%

30

0

0.00%

M

BOEING, 757

7,288

186

2.55%

2,472

4

0.16%

7,288

4

0.05%

M

BOEING, 767

3,825

143

3.74%

1,519

8

0.53%

3,825

8

0.21 %

M

BOEING, 777

34

3

8.82%

4

0

0.00%

34

0

0.00%

M

BOEING, 787

522

6

1.15%

365

0

0.00%

522

0

0.00%

M

BOEING,
B737

468

34

7.26%

232

2

0.86%

468

2

0.43%

M

BOEING,
B757

8

0

0.00%

0

0

0.00%

8

0

0.00%

M

BOEING,
DC982

4

0

0.00%

4

0

0.00%

4

0

0.00%

M

BOEING,
DC983

2

0

0.00%

2

0

0.00%

2

0

0.00%

M

DOUG, DC950

6

0

0.00%

0

0

0.00%

6

0

0.00%

M

DOUG, DC982

946

11

1.16%

573

0

0.00%

946

0

0.00%

M

DOUG, DC983

1,515

16

1.06%

1,186

0

0.00%

1,515

0

0.00%

M

DOUG, MD83

280

3

1.07%

279

0

0.00%

280

0

0.00%

M

DOUG, MD88

3,066

54

1.76%

105

1

0.95%

3,066

1

0.03%

M

DOUG, MD90

482

2

0.41%

2

0

0.00%

482

0

0.00%

M

DOUG,
MD9030

1,116

17

1.52%

17

0

0.00%

1,116

0

0.00%

S

ACE, 123

2

0

0.00%

2

0

0.00%

2

0

0.00%

S

ACE, 737

4

1

25.00%

4

0

0.00%

4

0

0.00%

S

ADAMS, 7897

16

0

0.00%

16

0

0.00%

16

0

0.00%

S

AIRBUS, 319

886

2

0.23%

613

0

0.00%

886

0

0.00%

S

AIRBUS, A220

48

1

2.08%

1

0

0.00%

48

0

0.00%

B-4


-------
Si
ze

Manufacturer,
Model

Total Coliforms

E. coli

E. coli (Alternative Approach)1

Total
Samples

#

Positiv
e

%

Positiv
e

Total
Samples

#

Positiv
e

%

Positiv
e

Total
Samples

#

Positiv
e

%

Positiv
e

S

AIRBUS, A318

14

0

0.00%

14

0

0.00%

14

0

0.00%

S

AIRBUS, A319

3,934

66

1.68%

1,387

8

0.58%

3,933

8

0.20%

s

AIRBUS, A321

1,094

19

1.74%

899

0

0.00%

1,094

0

0.00%

s

BOEING, 707

4

0

0.00%

4

0

0.00%

4

0

0.00%

s

BOEING, 717

2,601

21

0.81 %

459

0

0.00%

2,601

0

0.00%

s

BOEING, 737

1,705

33

1.94%

1,074

2

0.19%

1,705

2

0.12%

s

BOEING, 757

94

1

1.06%

84

0

0.00%

94

0

0.00%

s

BOEING, 767

34

1

2.94%

34

0

0.00%

34

0

0.00%

s

BOEING, 777

16

0

0.00%

14

0

0.00%

16

0

0.00%

s

BOEING,
B737

24

0

0.00%

24

0

0.00%

24

0

0.00%

s

BOMBDR,
BD100

1,683

180

10.70%

1,680

5

0.30%

1,683

5

0.30%

s

BOMBDR,
CL6002

17,070

2,228

13.05%

15,635

63

0.40%

17,066

63

0.37%

s

BOMBDR,
CRJ900

0

0

0.00%

0

0

0.00%

0

0

0.00%

s

BOMBDR,
DHC8402

356

84

23.60%

216

3

1.39%

356

3

0.84%

s

BOMBDR,
Q400

2

0

0.00%

2

0

0.00%

2

0

0.00%

s

CNDAIR,
CL6002

1,836

126

6.86%

1,391

10

0.72%

1,832

10

0.55%

s

DOUG, DC915

4

0

0.00%

4

0

0.00%

4

0

0.00%

s

DOUG, DC931

2

0

0.00%

2

0

0.00%

2

0

0.00%

s

DOUG, DC932

2

0

0.00%

2

0

0.00%

2

0

0.00%

s

DOUG, DC934

2

0

0.00%

2

0

0.00%

2

0

0.00%

s

DOUG, DC950

150

3

2.00%

3

0

0.00%

150

0

0.00%

s

DOUG, DC983

2

0

0.00%

2

0

0.00%

2

0

0.00%

s

DOUG, DC987

8

2

25.00%

8

0

0.00%

8

0

0.00%

s

DOUG, MD88

80

2

2.50%

2

0

0.00%

80

0

0.00%

s

EMB, 140

20

0

0.00%

14

0

0.00%

20

0

0.00%

s

EMB, EMB135

1,039

113

10.88%

793

16

2.02%

1,039

16

1.54%

s

EMB, EMB145

6,182

1,182

19.12%

5,656

21

0.37%

6,182

21

0.34%

s

EMB, EMB175

12

0

0.00%

12

0

0.00%

12

0

0.00%

s

EMB, ERJ170

8,203

307

3.74%

6,784

10

0.15%

8,203

10

0.12%

s

EMB, ERJ190

1,484

41

2.76%

1,108

2

0.18%

1,484

2

0.13%

Total

118,070

6,448

5.46%

78,114

201

0.26%

118,056

201

0.17%

1 Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

B-5


-------
Exhibit B-4. Total Coliform and E. coli Positivity Rate, by Location and Year; 2011-

2021

Year

Location

Total Conforms

E. coli

E. coli (Alternative Approach)1

# Samples

# Total
Colifor

m
Positiv
es

% Total
Coliform
Positive

s

# Samples

# E.
coli
positiv
es

% E. coli
positive
s

# Samples

# E. coli
positive
s

% E.
coli
positi
ves

2011

Galley

2,615

33

1.26%

1,077

2

0.19%

2,615

2

0.08%

2012

Galley

6,632

69

1.04%

4,352

4

0.09%

6,630

4

0.06%

2013

Galley

6,801

71

1.04%

4,103

4

0.10%

6,801

4

0.06%

2014

Galley

7,122

84

1.18%

4,294

2

0.05%

7,122

2

0.03%

2015

Galley

7,083

88

1.24%

4,029

2

0.05%

7,083

2

0.03%

2016

Galley

7,384

97

1.31 %

4,285

6

0.14%

7,384

6

0.08%

2017

Galley

6,389

82

1.28%

4,416

2

0.05%

6,389

2

0.03%

2018

Galley

6,426

83

1.29%

4,508

6

0.13%

6,426

6

0.09%

2019

Galley

6,440

61

0.95%

4,525

1

0.02%

6,440

1

0.02%

2020

Galley

5,922

54

0.91 %

4,172

2

0.05%

5,922

2

0.03%

2021

Galley

2,028

17

0.84%

1,471

2

0.14%

2,028

2

0.10%

2011

Lavatory

3,118

345

11.06%

1,736

8

0.46%

3,116

8

0.26%

2012

Lavatory

8,075

965

11.95%

5,931

29

0.49%

8,072

29

0.36%

2013

Lavatory

8,195

821

10.02%

5,390

27

0.50%

8,195

27

0.33%

2014

Lavatory

8,536

806

9.44%

5,551

15

0.27%

8,534

15

0.18%

2015

Lavatory

8,353

773

9.25%

5,157

22

0.43%

8,350

22

0.26%

2016

Lavatory

8,439

836

9.91 %

5,227

24

0.46%

8,439

24

0.28%

2017

Lavatory

7,259

569

7.84%

5,256

19

0.36%

7,258

19

0.26%

2018

Lavatory

7,477

582

7.78%

5,550

20

0.36%

7,474

20

0.27%

2019

Lavatory

7,459

461

6.18%

5,540

18

0.32%

7,459

18

0.24%

2020

Lavatory

6,653

275

4.13%

4,882

23

0.47%

6,653

23

0.35%

2021

Lavatory

2,096

78

3.72%

1,542

3

0.19%

2,095

3

0.14%

Total (All Years)

140,502

7,250

5.16%

92,994

241

0.26%

140,485

241

0.17%

Total (2012-2019)

118,070

6,448

5.46%

78,114

201

0.26%

118,056

201

0.17%

1 Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

A count of the number of total coliform samples and associated total coliform positivity rates
broken down by year and disinfection/flushing frequency for the years 2012-2019 is presented in
Exhibit B-5. A count of the number of E. coli samples and associated E. coli positivity rates
broken down by year and disinfection/flushing frequency for the years 2012-2019 are presented
in Exhibit B-6 and Exhibit B-7, for the first and second set of E. coli assumptions, respectively.

B-6


-------
More than 99 percent of total coliform samples were from systems that performed disinfection
and flushing four times per year. For those systems, the average total coliform positivity rates for
a given year was greater than 5 percent. Similarly, more than 99 percent of E. coli samples were
from systems that performed disinfection and flushing four times per year under both sets of E.
coli assumptions. For those systems, the average E. coli positivity rate for a given year was
approximately 0.26 percent under the first set of E. coli assumptions, and 0.17 percent under the
second set of E. coli assumptions.

Exhibit B-5. Total Coliform Sample Count and Positivity Rate, by Disinfection and

Flushing Frequency, by Year; 2012-2019

Total Coliforms

Year

2x per year

3x per year

4x per year

Unknown

Total
Sample
s

# Total
Coliform
Positive
s

% Total
Coliform
Positive
s

Total
Sample
s

# Total
Coliform
Positive
s

% Total
Coliform
Positive
s

Total
Sample
s

# Total
Coliform
Positive
s

% Total
Coliform
Positive
s

Total
Sample
s

# Total
Coliform
Positive
s

% Total
Coliform
Positive
s

2012

68

2

2.94%

180

6

3.33%

14,293

1,010

7.07%

166

16

9.64%

2013

66

0

0.00%

158

9

5.70%

14,650

871

5.95%

122

12

9.84%

2014

94

1

1.06%

32

0

0.00%

15,460

885

5.72%

72

4

5.56%

2015

60

8

13.33%

12

0

0.00%

15,354

853

5.56%

10

0

0.00%

2016

16

0

0.00%

0

0

0.00%

15,795

931

5.89%

12

2

16.67%

2017

16

0

0.00%

4

0

0.00%

13,620

651

4.78%

8

0

0.00%

2018

16

0

0.00%

0

0

0.00%

13,877

664

4.78%

10

1

10.00%

2019

16

0

0.00%

0

0

0.00%

13,877

522

3.76%

6

0

0.00%

Tota
1

352

11

3.13%

386

15

3.89%

116.926

6.387

5.46%

406

35

8.62%

Exhibit B-6. E. coli Positivity Rate, by Disinfection and Flushing Frequency, by

Year; 2012-2019

E. coli

Year

2x per year

3x per year

4x per year

Unknown

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

2012

68

0

0.00%

27

0

0.00%

10,124

33

0.33%

64

0

0.00%

2013

65

0

0.00%

21

0

0.00%

9,351

31

0.33%

56

0

0.00%

2014

94

0

0.00%

0

0

0.00%

9,713

17

0.18%

38

0

0.00%

2015

60

0

0.00%

0

0

0.00%

9,118

24

0.26%

8

0

0.00%

2016

16

0

0.00%

0

0

0.00%

9,484

29

0.31 %

12

1

8.33%

2017

16

0

0.00%

4

0

0.00%

9,644

21

0.22%

8

0

0.00%

2018

16

0

0.00%

0

0

0.00%

10,032

26

0.26%

10

0

0.00%

B-7


-------
E. coli

Year

2x per year

3x per year

4x per year

Unknown

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

2019

16

0

0.00%

0

0

0.00%

10,043

19

0.19%

6

0

0.00%

Tota
1

351

0

0.00%

52

0

0.00%

77,509

200

0.26%

202

1

0.50%

Exhibit B-7. E. coli Positivity Rate using Alternative Approach, by Disinfection
and Flushing Frequency, by Year; 2012-2019

E. coli (Alternative Approach)1

Year

2x per year

3x per year

4x per year

Unknown

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E. coli
positive

Total
Samples

# E. coli
positive

% E.
coli
positive

2012

68

0

0.00%

180

0

0.00%

14,288

33

0.23%

166

0

0.00%

2013

66

0

0.00%

158

0

0.00%

14,650

31

0.21%

122

0

0.00%

2014

94

0

0.00%

32

0

0.00%

15,458

17

0.11%

72

0

0.00%

2015

60

0

0.00%

12

0

0.00%

15,351

24

0.16%

10

0

0.00%

2016

16

0

0.00%

0

0

0.00%

15,795

29

0.18%

12

1

8.33%

2017

16

0

0.00%

4

0

0.00%

13,619

21

0.15%

8

0

0.00%

2018

16

0

0.00%

0

0

0.00%

13,874

26

0.19%

10

0

0.00%

2019

16

0

0.00%

0

0

0.00%

13,877

19

0.14%

6

0

0.00%

Total

352

0

0.00%

386

0

0.00%

116,912

200

0.17%

406

1

0.25%

1 Under the E. coli "Alternative Approach," any E. coli sample paired with a total coliform "Absent" was included as an E. coli
"Absent" sample.

B-8


-------
Appendix C. Revised Total Coliform Rule Level 1 and Level 2 Assessment
Characterization and Data Quality Considerations

Data Quality Considerations for Level 1 and Level 2 Assessments and Sanitary Defect
Designation

As described in EPA's Data Entry Instructions for RTCR, inaccurate and incomplete data limits
EPA's and the public's understanding of the state of compliance with the Safe Drinking Water
Act, as well as limits the review of RTCR effectiveness during this SYR4 review process
(USEPA, 2016c).

This subsection clarifies the reporting methods for TT triggers, sanitary defect identification and
RTCR corrective actions database accounting methods since discrete data extraction limited to
sanitary defects only may not appropriately capture the reported occurrence of sanitary defects.
The following factors affect the accounting of sanitary defects: whether Level 1 and Level 2
assessment site visits were reported by primacy agencies in Agency-specific databases and
reported to SDWIS/Fed; whether Level 1 and Level 2 assessments were performed during the
same site visit as a sanitary survey; and whether the SDWIS/Operational Data System
(SDWIS/ODS) appropriate accounts for sanitary defects reported in SDWIS State version 3.33
as significant deficiencies.

Although primacy agencies are not required by RTCR reporting requirements of 40 CFR 142 to
submit data elements regarding completion of Level 1 assessments or Level 2 assessments or
scheduled corrective actions, the RTCR Data Entry Instructions (USEPA, 2016c) clarify that
primacy agencies should report the following RTCR data elements for accurate and complete
data acceptance into the EPA national database of record including: RTCR TT trigger incurred,
primacy agency minimum requirement to satisfy RTCR TT trigger; and actual site
visit/assessment conducted in response to primacy agency RTCR TT Trigger requirement; and
expedited/corrective actions for assessments (USEPA, 2016c).

A Level 1 assessment, Level 2 assessment and sanitary survey each consists of a minimum
evaluation of eight site visit data object category evaluations, also known as elements. When
there are multiple findings within a site visit category/element, the primacy agency is required to
report the finding having the highest severity for that category/element. The ranking of highest to
lowest severity of findings is as follows: sanitary defect (highest severity), significant deficiency,
minor deficiency, recommendations made, no deficiencies or recommendations, not evaluated, or
not applicable (USEPA, 2016c). RTCR provides primacy agencies the authority in 40 CFR
141.859(b)(4)(iii-iv) to require expedited and additional actions, such as minor deficiencies or
recommendations, to be completed even if no sanitary defects are identified when there is an E.
coli MCL violation (USEPA, 2016c).

A database data quality error is assigned to the primacy agency data due to failure to report all
eight required RTCR assessment site visit category/elements (USEPA, 2016c) when a site visit is
entered for a RTCR assessment. However, the SDWIS database is limited on its alert of
discrepancies/errors for lack of reporting of the occurrence of an RTCR TT trigger event; lack of

C-l


-------
reporting a site visit for an assessment and lack of designation of corrective action event
schedule timeline (USEPA, 2016c).

Primacy agencies were required to provide examples of significant deficiencies for each of the
eight elements of a sanitary survey under the GWR and IESWTR. The difference between
significant deficiencies and sanitary defects can vary based on how the primacy agency
identified significant deficiencies. Sanitary defects are defined by the RTCR to be deficiencies
that could provide a pathway of entry for microbial contamination into the distribution system or
are indicative of a failure or imminent failure in a barrier that is already in place. Some sanitary
defects could also be significant deficiencies.

RTCR guidance describes that a complete sanitary survey covering all applicable elements is
allowed to fulfill the requirements of a Level 1 or Level 2 TT trigger if allowed by the primacy
agency, and the sanitary survey (including sanitary survey report and corrective actions) is
completed within 30 days of when the RTCR TT trigger happens (USEPA, 2016c). When a
RTCR Level 1 or Level 2 assessment is combined with a sanitary survey it is described as an
Integrated Assessment (USEPA, 2020c) and is described with site visit codes: L1SS, L2SS
(USEPA, 2016c). If a Level 1 or Level 2 assessment is combined with a partial sanitary survey it
is reported with site visit codes LIPS, L2PS.

When a primacy agency allows sanitary surveys to meet the RTCR Level 1 and/or Level 2 TT
triggers, then there is a potential that some sanitary defects could also be significant deficiencies.
When this happens, the primacy agency should use the report the category/element finding as a
sanitary defect, which is the highest severity finding (USEPA, 2016c).

SDWIS State 3.33 has a data element limitation where it will not allow the primacy agency to
report any RTCR assessment outcome of a sanitary defect. As a workaround, the Revised Total
Coliform Rule (RTCR) Data Entry Instructions with Examples (USEPA, 2016c) specified that
the finding be designated as a significant deficiency when a sanitary defect was identified. The
instructions further clarified that when the database user entered a significant deficiency and also
enters site visit as any type of Level 1 or Level 2 assessment (LV1A, LV2A, L1SS, L2SS, LIPS,
L2PS) that SDWIS/ODS would convert the values reported as significant deficiencies to sanitary
defect in EPA's national database (USEPA, 2016c). Although this workaround could be an
effective option for tracking sanitary defects, unfortunately the SDWIS/Fed site visit extract used
for the data period of April 1, 2016 to December 31, 2019, showed no significant deficiencies
were converted to sanitary defects for any element for Level 1 and Level 2 assessments reported
as integrated assessments or partial assessments. This would seem to mean that the only sanitary
defects maintained in the SDWIS/Fed national database were reported directly via state database
migration of other than SDWIS State 3.33.

For the SDWIS/Fed site visit dataset used below, the values of sanitary defects were based on the
combination of reported significant deficiencies and sanitary defects performed during Level 1
assessments and Level 2 assessments having site visit objects: LV1A, LV2A, L1SS, L2SS,

LIPS, L2PS, because significant deficiencies were not appropriately reported as sanitary defects
in SDWIS/Fed from SDWIS State 3.33 as described above.

C-2


-------
Characterization of Level 1 and 2 Assessments

In order to characterize the occurrence of Level 1 and Level 2 assessments, this SYR4 review
considered two datasets. For the first dataset of this analysis an extract of the SDWIS/Fed "event
and milestone" data was reviewed for the period of April 1, 2016 to December 31, 2019 to
determine the frequency that RTCR assessments were triggered. This data was used to identify
whether Level 1 and Level 2 assessments were triggered and how often assessments were
triggered by an individual PWS. When a RTCR Level 1 or Level 2 assessment is triggered it is
reported using data elements, RTL1 and RTL2 respectively, in the event and milestone database.
Reporting of a triggered RTCR assessment as an "event and milestone" is different from
reporting of RTCR violations in the SDWIS/Fed database

The second data set, SDWIS/Fed site visit dataset for the period of April 1, 2016 to December
31, 2019 covered 40 states (not including Alabama, District of Columbia, Georgia, Kentucky,
Mississippi, Minnesota, North Carolina, Pennsylvania, South Carolina, Tennessee, Washington,
Guam, Puerto Rico, U.S. Virgin Islands) and including Navajo Nation, American Samoa, and
Northern Mariana Islands. The SDWIS/Fed site visit dataset was used to show the occurrence of
sanitary defects and significant deficiencies identified during Level 1 and Level 2 assessments. A
more detailed explanation of data quality considerations regarding the reporting of sanitary
defects for SDWIS/Fed site visit is included in Appendix C above.

Exhibit C-l and Exhibit C-2 show the recurrence of multiple RTCR Level 1 and Level 2
assessments, respectively, at individual water systems by year. Exhibit C-l shows that there was
a decreasing trend in the total number of water systems having two or more Level 1 assessments
as time progressed from the initial RTCR implementation, when comparing years having a full
calendar year of data, 2017 to 2019.

Exhibit C-1. Summary of Frequency of Recurring Revised Total Coliform Rule
Level 1 Assessments1 Triggered by an Individual Public Water System by Year

Year

Total PWS
Triggering
Level 1 per
year

Count of
PWS with
exactly 1
Level 1
Trigger
per year

Count of
PWS with 2
or more
Level 1
T riggers
per year

Percent of
PWS with 2
or more
Level 1
Triggers
per year

Count of
PWS with
exactly 2
Level 1
Triggers
per year

Count of
PWS with
exactly 3
Level 1
Triggers
per year

Count of
PWS with
>4 Level 1
Triggers
per year

20162

1,967

1,836

131

6.7%

123

7

1

2017

2,369

2,290

79

3.3%

76

2

1

2018

2,506

2,418

88

3.5%

80

7

1

2019

1,861

1,809

52

2.8%

47

5

0

1	There was a total of 9081 Level 1 assessments reported as triggered in this dataset from April 1, 2016 to December
31, 2019. Source: SDWIS/Fed "Event and Milestone" data

2	Represents partial year from April 1 to December 31, 2016 due to RTCR implementation start date of April 1, 2016.

Exhibit C-l shows that for the period of April 1, 2016 to December 31, 2019, that there were
approximately 3 percent of water systems that incurred a second Level 1 assessment (reported as
a Level 1 assessment) in the same calendar year. Recurrence of Level 1 TT trigger within 12
months of the preceding Level 1 TT trigger is defined as a Level 2 TT trigger. Therefore,

C-3


-------
accurate data reporting of multiple Level 1 assessments in the same year may only occur due to
the primacy agency designating the water system as "reset" due to identifying and resolving the
underlying finding of the initial Level 1 assessment at that same water system. Alternatively,
data reporting of two recurring instances of Level 1 assessment may indicate improper data entry
of the second Level 1 TT trigger that the state should have reported as Level 2 assessment if the
"reset" provision did not apply.

Exhibit C-2. Summary of Frequency of Recurring Revised Total Coliform Rule
Level 2 Assessments1 Triggered by an Individual Public Water System by Year

Year

Total PWS
Triggering
Level 2 per
year

Count of
PWS with
exactly 1
Level 2
Trigger
per year

Count of
PWS with 2
or more
Level 2
Triggers
per year

Percent of
PWS with 2
or more
Level 2
Triggers
per year

Count of
PWS with
exactly 2
Level 2
Triggers
per year

Count of
PWS with
exactly 3
Level 2
Triggers
per year

Count of
PWS with
>4 Level 2
Triggers
per year

20162

602

504

98

16%

84

13

1

2017

1,152

911

241

21%

188

38

15

2018

1,351

1,028

323

24%

246

51

26

2019

1,013

803

210

21%

168

24

18

1	There was a total of 5268 Level 2 assessments reported as triggered in this dataset from April 1, 2016 to December
31, 2019. Source: SDWIS/Fed "Event and Milestone" data

2	Represents partial year from April 1 to December 31, 2016 due to RTCR implementation start date of April 1, 2016.

For Level 2 assessments shown in Exhibit C-2, there were approximately 20 percent of water
systems that incurred a second Level 2 assessment in the same calendar year for the years having
a full calendar year of data, 2017-2019. This means more than 20 percent of water systems
triggering Level 2 assessments in each of the calendar years 2017, 2018, and 2019 had recurring
E. coli positivity; or multiple coliform positivity (e.g., two LI triggers occurred within 12
months); or a recurrence of failure to perform appropriate coliform monitoring within the same
calendar year. This would indicate sanitary defect conditions for these PWS, since defects had
not been effectively resolved in accordance with the Level 2 corrective action requirements or
timeline.

Exhibit C-3 shows the number of Level 1 and Level 2 assessments occurring between April 1,
2016 to December 31, 2019 that did not identify sanitary defects or other corrective actions.

Exhibit C-3. Summary of Data Provided by Federal Safe Drinking Water
Information System Database Site Visit Entries Showing Revised Total Coliform
Rule Level 1 and Level 2 Assessments during April 1, 2016 to December 31, 2019
that did not Identify a Sanitary Defect or Other Corrective Action

Type of Findings Reported

Count of Level 1

Percent of

Count of Level 2

Percent of

for Required RTCR

Assessments

Level 1

Assessment

Level 2

Elements

without Identified

Assessments

without Identified

Assessments



Finding

without

Finding

without





Identified



Identified





Finding



Finding

C-4


-------
No Sanitary defect1 Found for
All Required Elements

5395

65%

2556

45%

No Finding of Any Type2 for
All Required Elements

3952

47.6%

1804

31.8%

1	Sanitary defect for this purpose includes: significant deficiencies as described in the data quality considerations
discussion in Appendix C above.

2	For this purpose, no finding of any type means that no sanitary defect, significant deficiency, minor deficiency or
recommendation was reported.

Source: SDWIS/Fed site visits extract, 40 states plus Navajo Nation and territories

Also, the frequency of recurring Level 2 assessments shown in Exhibit C-2 did not show a
decreasing trend for recurring Level 2 assessments when comparing full calendar years of data
from 2017 to 2018. This may be because 45 percent of Level 2 assessments that occurred during
the period of April 1, 2016 to December 31, 2019 did not find a sanitary defect to fix, as noted in
Exhibit C-3. It is not implicit that the 45 percent of Level 2 assessments without sanitary defect
correlate directly to the 20 percent of systems that triggered more than one Level 2 assessment in
the same calendar year, although the lack of an identifiable sanitary defect complicates a water
system's ability to implement a corrective action to resolve coliform positive detect(s).

A complicating factor for RTCR assessments is that they are performed on a lagging basis and
can be completed over a 30-day timeframe following collection of the initial coliform sample.
This can make identification of a sanitary defect difficult due to the delayed nature of the
assessment. The lagging nature of this process is due to several factors such as, up to 30 hours
for sample hold time, microbial incubation time of method analysis, time interval prior to repeat
sampling (i.e., 24 to 72 hours) and time interval between notifying the PWS of the sample result
and performing the assessment.

Exhibit C-4 shows the types of sanitary defects that were identified during Level 1 and Level 2
assessments between April 1, 2016 to December 31, 2019. Likewise, Exhibit C-5 shows the
types of corrective actions other than sanitary defects that were identified during Level 1 and
Level 2 assessments that occurred between April 1, 2016 to December 31, 2019.

Exhibit C-4. Summary of Data Provided by Federal Safe Drinking Water
Information System Database Site Visit Entries1 showing the Types of Sanitary
Defect (and Significant Deficiencies) identified by Revised Total Coliform Rule
Level 1 and Level 2 Assessments during April 1, 2016 to December 31, 2019

Sanitary Defect2 Element
Reported for RTCR
Assessment

RTCR

Element

Required

Reporting

Status

(USEPA,

2016c)

Count of
Level 1

Assessments
that

Reported
Sanitary
Defect
Element

Percent of
Level 1

Assessments
with Sanitary
Defect
Element1

Count of Level 2
Assessments
that Reported
Sanitary Defect
Element

Percent of
Level 2
Assessments
with Sanitary
Defect
Element1

Source Water

Required

974

11.7%

1644

29.0%

Management Operation

Required

587

7.1%

480

8.5%

Treatment

Required

554

6.7%

545

9.6%

C-5


-------
Finished Water

Required

351

4.2%

563

9.9%

Distribution

Required

973

11.7%

1183

20.9%

Data Verification

Required

580

7.0%

236

4.2%

Compliance

Required

40

0.5%

132

2.3%

Pump

Required

87

1.0%

128

2.3%

Security

Optional

6

0.1%

10

0.2%

Other

Optional

803

9.7%

530

9.4%

Financial

Optional

5

0.1%

2

0.0%

1	There was a total of 8305 Level 1 RTCR Assessments and 5668 Level 2 Assessments having SDWIS/Fed Site Visit
entries reported during April 1, 2016 to December 31, 2019 in this dataset. Source: SDWIS/Fed Site Visits, 40 states
plus Navajo Nation and territories.

2	Sanitary Defect for this purpose includes: significant deficiencies and sanitary defects identified during RTCR
assessments as described in the data quality considerations discussion in Appendix C above.

Issues with source water and issues in the distribution system represent the largest two types of
sanitary defects for both Level 1 and Level 2 RTCR assessments. This is also the same with
regard to other corrective actions identified by Level 1 and Level 2 RTCR assessments. The third
most relevant sanitary defect applicable to other corrective actions was management of the water
system as shown in
Exhibit C-5.

Exhibit C-5. Summary of Data Provided by Federal Safe Drinking Water
Information System Database Site Visit Entriesl showing the Types of Other
Corrective Actions, Including Recommendations, and Minor Deficiencies
identified by Revised Total Coliform Rule Level 1 and Level 2 Assessments
during April 1, 2016 to December 31, 2019

Element of Corrective
Action (Minor Deficiency or
Other Recommendation)
Reported for RTCR
Assessments

Count of Level 1
Assessments that
Identified Element
of Other

Corrective Action

Percent of
Level 1

Assessments

with Element of

Other

Corrective

Action

Identified1

Count of Level 2
Assessments that
Identified Element
of Other

Corrective Action

Percent of
Level 2
Assessments
with Element of
Other
Corrective
Action
Identified1

Source Water

565

6.8%

488

8.6%

Management Operation

532

6.4%

465

8.2%

Treatment

271

3.3%

267

4.7%

Finished Water

205

2.5%

273

4.8%

Distribution

742

8.9%

494

8.7%

Data Verification

231

2.8%

214

3.8%

Compliance

74

0.9%

86

1.5%

Pump

80

1.0%

105

1.9%

Security

20

0.2%

70

1.2%

Other

54

0.7%

68

1.2%

C-6


-------
Element of Corrective
Action (Minor Deficiency or
Other Recommendation)
Reported for RTCR
Assessments

Count of Level 1
Assessments that
Identified Element
of Other

Corrective Action

Percent of
Level 1

Assessments

with Element of

Other

Corrective

Action

Identified1

Count of Level 2
Assessments that
Identified Element
of Other

Corrective Action

Percent of
Level 2
Assessments
with Element of
Other
Corrective
Action
Identified1

Financial

6

0.1%

20

0.4%

1 There was a total of 8305 Level 1 RTCR Assessments and 5668 Level 2 Assessments having SDWIS/Fed Site Visit
entries reported during April 1, 2016 to December 31, 2019 in this dataset. Source: SDWIS/Fed Site Visits, 40 states
plus Navajo Nation and territories.

This SYR4 review did not include a review of available information to determine whether
detailed information of the exact nature of sanitary defects was identified in SDWIS for RTCR
assessments.

C-7


-------

-------
Appendix D. Revised Total Coliform Rule Corrective Actions and

Assessment of Data Quality

This SYR4 review studied new information regarding distribution system corrective actions as
well as SDWIS data quality limitations regarding RTCR assessments.

New Information Pertaining to Sanitary Defects and Corrective Actions

When a PWS triggers a Level 1 or Level 2 assessment because a sanitary defect was identified
that could be the cause of total coliform positive or E. coli positive samples, a corrective action is
required (40 CFR 141.859(c)).

¦	Failure to Disinfect After Maintenance

Existing practices used for repair and replacement of water mains pose potential risk of microbial
contamination. Available guidelines and industry standards outline proper planning and standard
operating procedures (SOPs) to address risk of contamination degradation of water quality
associated with water main repair and replacement.

If a finished water storage facility is drained for maintenance or inspection, disinfection must
occur before being placed into service. ANSI/AWWA C652 (AWW A, 2020), provides the
guidance for proper disinfection.

The Ten States Standards (GLUMRB, 2018) stresses the importance of a sufficient number of
valves in the distribution system to minimize sanitary hazards during repairs. In commercial
districts valves should not be located at greater than 500-foot intervals, in other districts one
block or 800-foot intervals, and in areas that serve widely scattered customers (or where there is
no expected future development) valves should not exceed one mile.

¦	Lack of Flushing Programs

Flushing entails allowing water to discharge from the distribution system by opening a
connection. Flushing can have many benefits in a distribution system including reduction of
water age and addressing water quality complaints (USEPA, 2022a). Flushing is listed as an
acceptable corrective action to address sanitary defects in the Revised Total Coliform Rule
Assessments and Corrective Actions Guidance Manual: Interim Final (USEPA, 2014b). Detail
on proper flushing techniques or applications is not provided by the manual and flushing is not
clearly listed as an appropriate corrective action for particular coliform response situations (Hill
et al., 2018).

The State drinking Water Distribution Survey conducted by ASDWA (2020) was completed by
drinking water representatives from 41 states and territories. At least 12 states (30 percent of
respondents) have flushing requirements to better ensure a safe and reliable distribution system

D-l


-------
written in their state legislation (ASDWA, 2020). Of the 70 percent of state respondents that do
not require a flushing program, many strongly recommend it. Some of the methods to encourage
a flushing program include either requiring a flushing plan to be eligible for Drinking Water
State Revolving Fund (DWSRF) funding or encouraging it to be a part of the water system's
operations and maintenance plan during sanitary surveys (ASDWA, 2020).

Hill et al. (2018) points out a lack of regular flushing programs as a common cause of total
coliform and E. coli detections in the distribution system while noting that specific causes will
vary from system to system.

Spot flushing, a conventional flushing technique, can be performed on particular areas of the
distribution system to decrease water age in areas such as dead-end water mains by drawing in
fresh water. Spot flushing can be triggered by a water quality issue or customer complaints, or
can be scheduled regularly in known areas of issue (USEPA, 2022a; Hill et al., 2018).

Unidirectional Flushing (UDF) starts at the point where clean water enters the system and
systematically flushes through the system. This flushing typically involves a higher velocity due
to closing nearby isolation valves to direct flow. UDF can be used as a regular maintenance
practice and can achieve complete turnover in the defined UDF area if applied to all mains. A
flushing velocity and terminating criteria can be defined for the specific utility in the flushing
program (Hill et al., 2018).

Automatic flushing stations can also be used as an on-going water age management technique.
These can be programmed to flush specific portions of the system at specified velocities and
intervals. They can be semi-permanent or portable (Hill et al., 2018).

Flushing and secondary disinfection can work together in the distribution system to maintain
adequate disinfection residual and control microbial activity on the pipe walls. Hill et al. (2018)
established a guidance toolbox for use of flushing under the RTCR which has four steps
discussed in terms of being a corrective action. The steps begin with conducting a coliform
assessment to determine the likely cause. Next a decision matrix is consulted to assess if flushing
is appropriate, then the matrix is used to decide the proper flushing technique. Last other
potential actions can be identified using a summary table for various coliform occurrence
pathways (Hill et al., 2018).

Hill et al. (2018) suggests that flushing is an appropriate corrective action for coliform events
from pipe wall biofilm regrowth and release. However, does not correct factors of coliform
events that arise from treatment breakthrough, direct DS contamination, or source water
contamination and should be used a secondary corrective action for those instances (Hill et al.,
2018).

D-2


-------
¦	Main Breaks

PWSs should have a written SOP for proper main repair and disinfection practices that meet
AWWA standard C651. Maintenance staff and contractors should have access to the SOP and
any resources needed to comply with it.

¦	Pressure Loss in Distribution System

The five best available technologies, TTs, or other means of achieving compliance with the MCL
for E. coli are identified in the RTCR. Distribution system pressure management is one of these
identified options as follows:

"Proper maintenance of the distribution system including appropriate pipe replacement and
repair procedures, main flushing programs, proper operation and maintenance
of storage tanks and reservoirs, cross connection control, and continual maintenance ofpositive
water pressure in all parts of the distribution system... " (40 CFR 141.63 (e)(3)).

Thirty-eight states have a design or operational standard for minimum pressure (ASDWA,
2020).). To ensure fire fighters have sufficient pressure in the system and to avoid instances of
negative pressure, most states specify 20 psi. Some states also have a maximum pressure limit,
typically between 60 and 150 psi (USEPA, 2024b).

The potential for contaminants to enter the distribution system during pressure events (i.e., main
breaks and pressure surges) through physical gaps can be reduced by an effective pressure
management strategy. E. coli, total coliform bacteria, worms, hydro-seed, propylene glycol,
ethylene glycol, and irrigation water can enter a water distribution system via backflow
(AWWA, 2017).

¦	Breaches in Finished Water Storage Facilities

Finished water storage facility vent and overflow screens can have significant physical gaps that
can lead to contamination of storage facilities and, consequently, distribution systems. This
stored water is delivered directly to the customers and thus creates a public health risk. Physical
gaps in vents and overflows can come from damaged screens, no screens, or screens with
openings too large to prevent intrusion of insects and animals. The water in the facility can be an
attraction for small animals (e.g., birds, bats, rodents, snakes and insects) and they can
potentially enter through these physical gaps causing contaminants such as opportunistic
waterborne pathogens to enter the system (USEPA, 2019d).

EPA's sanitary survey guidance states that overflow pipes should not discharge directly to the
ground or to any storm or sewer line to prevent contamination (USEPA, 2019d). The guidance
also states that rooftop tank vents should have a corrosion resistant fine mesh screen to prevent
entrance of birds, insects, and small debris as a flapper valve alone could be prevented from
closing completely by debris, ice, or snow. To prevent the storage facility from imploding caused

D-3


-------
by a vacuum effect of a clogged screen, the screen must be designed to fail. The vent should face
downward or be covered to protect the storage facility contents from rain (USEPA, 2019d).

Regular tank cleaning and inspection can remove the accumulated sediment and help locate
breaches in a finished water storage facility. EPA's review of state regulations in 2017 found that
some states require or recommend periodic comprehensive inspections but other states have no
such requirements. EPA's review of sanitary survey reports in the SDWIS database (Heinrich et
al., 2022) also confirmed that some water systems have no tank inspection program.

AWWA's M68 manual suggests that, in order to remove sediment and biofilm that may harbor
nitrifying bacteria, storage facilities should be inspected and cleaned at least every five years
(AWWA, 2017).

¦ Cross-Connection Control and Backflow Prevention Program

The AWWA M68 manual (2017) notes that only a robust and active cross-connection control
program can ensure that a distribution system is truly not affected by outside conditions. Possible
indicators of cross-connection and backflow incidents in distribution systems can include
(AWWA, 2017):

•	Drops in operating pressure

•	Customer complaints

•	Water meter anomalies

•	Drops in disinfectant residual

•	Detections of total coliform and HPC bacteria changes

The M68 manual emphasizes using cross-connection control and backflow-prevention programs
to prevent, eliminate, and/or control cross-connections (AWWA, 2017).

Due to potential for hydrants to be a source of cross-connection, the GLUMRB 10 States
Standards (2018) state that hydrants and flushing lines must be equipped with backflow
prevention devices.

The SDWIS analysis of data between 2010 and 2017 performed by EPA found that unprotected
existing or potential cross-connections were the most prevalent deficiencies identified at 26.9
percent of all surveys (Heinrich et al., 2022).

EPA found, as of May 2020, that 49 of the 50 US states (with the exception of Delaware) have
developed and implemented cross-connection control programs. According to the ASDWA State
Drinking Water Distribution System Survey, 53 percent of responding states require a cross-
connection survey and half of those included all water use equipment (e.g., cooling towers, spray
misters, spas, and pools) (ASDWA, 2020).

D-4


-------
¦ Inadequate Disinfectant Residuals

AWWA conducted its fifth disinfection survey in 2017 which collected information from water
systems on their common treatment practices (AWWA, 2018). Survey responses were
summarized for a total of 375 water systems, distributed across 44 states and one United States
territory.

Systems noted that meeting minimum chlorine target levels is more challenging in the
distribution system than it is for meeting the targets for primary disinfection. The report showed
12 percent of systems reported frequent difficulties in meeting their chlorine residual targets in
the distribution system while the majority of the respondents reported having difficulty on
occasion.

Gibson and Bartrand (2021) evaluated disinfection practices of CWSs used publicly available
data. Of the total 3,823 systems in the statistically representative sample, 831 reported that they
do not provide residual (secondary) disinfection (Gibson and Bartrand, 2021).

Many factors influence the concentration of the disinfectant residual in the distribution system,
and its effectiveness in controlling microbial growth and biofilm formation. These factors
include the assimilable organic carbon (AOC) level, the type and concentration of disinfectant,
water temperature, pipe material, and system hydraulics. The number of variables associated
with biofilm control has led researchers to reach differing conclusions regarding the
effectiveness of secondary disinfectants at controlling biofilm growth.

The ability to control (but not eliminate) biofilms using secondary disinfection is impacted by the
disinfectant residual concentration used in the system. If the concentrations are too low, the
disinfectant residual becomes ineffective at controlling biofilm growth. Several studies have
shown that biofilm growth is reduced when sufficient disinfectant residuals are maintained in the
bulk water passing through pipes.

This advantage was maintained at a higher dose of chlorine in the presence of organic matter that
could react with chlorine. Puzon et al. (2020) further characterized N. fowleri presence in
biofilm, noting a diverse community composition can contribute to the organism's ability to
successfully colonize a biofilm and demonstrate considerable resistance to chlorine disinfection.

The presence of sessile Legionella, particularly within Free-living amoebae (FLA), may be due
to biofilm build-up and conditions that favor its continued presence in a biofilm such as
inconsistent temperature, slower water velocity, and disinfectant concentration. Shaheen et al.
(2019) suggest that diminished or absent disinfectant residual in conjunction with other
environmental conditions such as optimal temperature and nutrient inflow may encourage FLA
growth.

To date, a range of amoebae that may support Legionella and mycobacteria cell growth have
been identified in drinking water. It has been found that slight water temperature changes can
influence the growth potential of different pathogenic strains of Legionella and their supporting

D-5


-------
host amoebae. These data describe critical numbers of Legionella in water pipe biofilms (slimes),
shower head water, and bathroom aerosols that could be inhaled (USEPA, 2021).

¦	Contaminated Sampling Taps

The detection of coliform bacteria in a water sample by any of the four analytical techniques is a
warning of possible contamination. One positive test does not conclusively prove contamination
AWWA, 2016). Samples are often contaminated by improper sampling technique, improperly
sterilized bottles, and laboratory error (AWWA, 2016).

Ouro Koumai voiced concerns that coliform-positive samples were sometimes erroneously
attributed to poor sampling technique rather than a call to action to identify and address the
source of contamination (ASDWA, 2019).

Heinrich et al. (2022) conducted a study which intended to identify the most frequent
deficiencies found by reviewing sanitary survey information collected by primacy agencies. This
study found that that 22.9 percent of all systems surveyed had a deficiency of "failure to monitor
according to system's monitoring plan(s) or established procedures". That placed it as the second
most commonly identified deficiency (Heinrich et al., 2022). EPA provided a guide for drinking
water sample collection (USEPA, 2016d).

¦	Addition or Upgrade of On-Line Monitoring and Control

Controlling and monitoring disinfectant dosages and water quality parameters can also be
performed through the use of a supervisory control and data acquisition (SCADA) system at the
treatment facility. Disinfectant dosing equipment can be monitored and analyzers can be placed
in the treatment process to monitor water quality parameters. Monitoring water quality
parameters via SCADA in a distribution system is possible; however, it can be costly.
Determining the number and location of the analyzers is challenging and highly dependent upon
the system size. Typically, analyzer equipment will draw samples from an above grade pipe or a
sample tap to an analyzer that is placed in a building. Sample locations will require analyzer
equipment, a building, electric power and, in the case of some systems, integration to the PWS's
existing SCADA system. Method requirements for on-line amperometric chlorine monitors are
more time intensive and difficult than grab sampling (USEPA, 2014).

Installing online pressure monitoring and control will help minimize future incidents of pressure
loss that can allow entry of contaminants into the distribution system. It can also help a PWS
determine if there are any physical problems in the system, e.g., a crack in a pipe, a leaking
valve, etc., that cause changes to the water quality of the system.

On-line distribution system monitoring through the SCADA system can alert operators if there
are possible issues with the distribution system; however, monitoring the water quality or
pressure will not identify the source of the contamination nor will it necessarily identify the
location of the contamination (USEPA, 2014)

D-6


-------
¦	Addition of Security Measures

PWSs may need to install security measures in circumstances where the assessment or onsite
inspection reveals vandalism or security breaches that could lead to water contamination.
Measures that PWSs may take to correct security breaches include installing a fence or locking
buildings to restrict access to the system. Other possible security measures include employing a
full time, on-site security staff and using alarms and cameras to detect security breaches
(USEPA, 2014).

PWSs should prioritize their security measures and concentrate on the most vulnerable parts of
their system, such as unstaffed facilities (e.g., finished water storage tanks). An important
implementation issue is determining the extent to which the water system needs to be secured.
This would depend on how widely spread the system/facility is, the number and complexity of
the treatment trains, the extent of the watershed, the distance of the treatment plant from the
influent wells, accessibility of the distribution system, etc. (USEPA, 2014)

¦	Development and Implementation of an Operations Plan

PWSs may need to develop an operations plan or improve their existing one when the
assessment identifies gaps in the way the system is operated that could have led to or contributed
to the sanitary defect identified. For example, a broken valve might have been prevented if
routine inspections were part of the operations plan. An operations plan can integrate all
operations and maintenance functions to meet the goals of flow, pressure and water quality. The
AWWA G200-04 standard describes the critical requirements for the effective operation and
management of drinking water distribution systems. According to this standard, a water system
should develop SOPs, comprehensive monitoring plans, routine inspections and emergency
response plans (USEPA, 2014).

¦	Corrosion control for Microbial Control

The distribution system toolbox factsheet, Impact of Corrosion Control on Disinfectant Residual,
covers the association of low disinfectant residual and the corresponding potential for microbial
growth with corroding metals and associated corrosion products in finished water. The factsheet
covers potential strategies to address corrosion-related considerations and find potentially
corrosive microbial growth (USEPA, 2024b). This SYR4 review was not intended to identify
any new information pertaining to distribution system corrosion control.

D-7


-------