United States
Environmental Protection

Agency

Technical Support Document for the Final Lead and Copper

Rule Improvements


-------
Office of Water (4607M)
EPA 815-R-24-028
October 2024


-------
Table of Contents

1	Introduction	1

2	Feasibility of the Final LCRI's Service Line Replacement Requirements	1

2.1	Per-Connection Rate Assessment of LSLR Programs	1

2.2	Projection of Systems Expected to Exceed Deferral Threshold Using Updated Needs Survey Data	6

2.3	Removing the Proposed Threshold for Maximum Number of Annual Replacements	11

2.4	Potential for Faster Replacement Rates in the Future	12

3	Feasibility of Service Line Inventory Requirement	13

3.1	States with Service Line Inventory and Replacement Requirements	13

3.2	Updated Inventory Information from the Needs Survey to Inform Inventory Development Feasibility	15

3.3	Systems with Completed LSLR Programs	17

3.4	Additional Opportunities for Inventory Development	18

3.5	Determination of the Inventory Validation Pool and Minimum Number of Validations Required	19

4	Supporting Information for the Lead Action Level Analysis	25

Appendices	28

References	32


-------
Table of Exhibits

Exhibit 2.1: System Characteristics and Replacement Rate Data from the Included LSLR Programs	2

Exhibit 2.2: Summary Statistics on the Annual Replacement Rates per 1,000 Service Connections Previously

Achieved by Water Systems, by Size Category	5

Exhibit 2.3: Low-bound and Best Estimates of Systems Likely to Exceed the Per-connection Replacement Threshold

in the Final LCRI Compared to the Proposed LCRI and the Alternate Per-household Scenario	8

Exhibit 3.1: Summary of Remaining Unknown Service Lines in Service Line Inventories Submitted to Illinois EPA

from 2017-2022	14

Exhibit 3.2: Summary of Remaining Percentage of Service Lines Classified as "Unknown" from the 2023 EPA

Needs Survey One-Time Update Among Systems Reporting Service Line Materials	16

Exhibit 3.3: Water Systems with Completed LSLR Programs	17

Exhibit 4.1: Percent of CWSs in Each Size Category Estimated to Have 90th Percentile Lead Levels Exceeding 0.015

mg/L, 0.010 mg/L, and 0.005 mg/L by System Size Under the Final LCRI	24

Exhibit 4.2: Percent of CWSs in Each Size Category Estimated to Have 90th Percentile Lead Levels Exceeding 0.015
mg/L, 0.010 mg/L, and 0.005 mg/L by LSL and CCT Status Under the Final LCRI	28

II


-------
1 Introduction

This technical support document provides detailed information on data sources, data collection,
and technical analyses considered was considered in the development of the final Lead and
Copper Rule Improvements (LCRI), specifically in the agency's analysis of the feasibility of lead
service line replacement requirements, service line inventory, and action level. Chapter 2 of this
document details the data and analysis considered in the feasibility of mandatory service line
replacement requirements in the final rule, including methods for data collection of available
lead service line replacement (LSLR) data and use of data from the One-Time Update to the
Drinking Water Infrastructure Needs Survey and Assessment (Needs Survey). Chapter 3 of this
document describes the data sources and analyses that were considered by the EPA in its analysis
of the feasibility of service line identification and inventory development, including data from
States that have already required the development and verification of service line inventories.
Chapter 4 describes the data and analysis considered by the EPA in its analysis of the feasibility
of lowering of the lead action level to 0.010 mg/L in the LCRI.

2 Feasibility of the Final LCRTs Service Line Replacement Requirements
2.1 Per-Connection Rate Assessment of LSLR Programs

The LCRI maintains a treatment technique approach to reduce lead in drinking water. In
establishing treatment technique requirements, the Safe Drinking Water Act (SDWA) authorizes
the EPA Administrator to identify treatment techniques that "prevent known or anticipated
adverse effects on the health of persons to the extent feasible" (SDWA 1412(b)(7)(A)). Thus,
service line replacement, part of the LCRI's treatment technique requirements, must prevent
adverse health effects to the extent feasible in accordance with the statute and legislative history
(see preamble section III.D).

To determine the mandatory service line replacement requirements that would protect the health
of persons to the extent feasible for the proposed LCRI, the EPA identified water systems with
service line replacement programs to replace both lead and galvanized requiring replacement
(GRR) service lines and obtained data on the replacement rates they achieved. From January
2023 to May 2023, the EPA compiled replacement rate data from official sources, including
State, local, and water system websites, peer-reviewed research articles, and conference
proceedings. Websites providing information on replacement programs, such as non-profit
websites (e.g., Environmental Defense Fund) were used to help identify which States and water
systems the EPA should evaluate further for inclusion in the analysis. The EPA also became
aware of replacement programs though queries of local media reporting. Each source describing
replacement data was manually reviewed for data quality by the EPA's staff. This search yielded
44 service line replacement programs with available replacement rate data that met the EPA's
exclusionary criteria. From this dataset of 44 replacement programs, the EPA excluded projected
replacement rates because they do not indicate rates that have been achieved by water systems,
which is stronger evidence of a replacement rate's feasibility. Additionally, the EPA excluded
service line replacement programs which replace service lines solely in coordination with
emergency repair or routine infrastructure because it is unlikely that these water systems are
replacing lead and galvanized requiring replacement service lines to the extent feasible. In the
proposed LCRI, the EPA sought comment requesting any additional service line replacement

1


-------
rate data achieved by systems in replacement programs meeting its criteria (i.e., excluding
programs that only replace service lines in coordination with main replacement or emergency
repair) (USEPA, 2023b). A Quality Assurance Protection Plan (QAPP) was developed for the
LCRI outlining how the Needs Survey dataset was processed, analyzed, and the quality
assurance processes conducted for this analysis. The QAPP additionally detailed data sources
and analytical techniques used for the service line inventory progress (Chapter 3) and the
underlying data and analysis for determination of the lead action level (Chapter 4).

For the final rule, the EPA updated the replacement rate dataset used in the proposed LCRI to
include data submitted to the EPA in the public comment period on the proposed LCRI. This
resulted in a total of 48 replacement programs with replacement rate data for the final rule's
feasibility analysis (Exhibit 2.1).

Exhibit 2.1: System Characteristics and Replacement Rate Data from the
Included LSLR Programs

City

State

Population

Served

Total
Number of
LSL/GRR

%of
Total

SLs

Duration of
Program

Avg.
Replaced
per year

Annual
Replacements
per 1,000 service
connections

Years To
Complete at
This Rate

Cleveland3

OH

1,308,955

185,409

41%

2023

4,000
(2.2%)

8.8

45.5

Denver

CO

1,287,000

74,000

22%

Jan 2020 to Jan
2024

5,581
(7.0%)

17

14.3

Fort Worth

TX

853,762

1,790

0.31%

2016 to 2021

233
(13%)

0.40

7.7

Louisville3

KY

764,769

790

0%

2023

272
(34%)

0.96

2.9

Cincinnati

OH

750,200

51,951

0.70%

2016-2030

1,200
(2.3%)

4.9

43.5

Detroit

MI

713,777

80,000

29%

2023

2,060
(2.6%)

7.5

38.5

Tucson

AZ

675,686

600

0.31%

2016 to 2018

47.3
(7.9%)

0.25

12.7

Washington

DC

632,323

28,000

20%

2019-2023

933
(3.3%)

6.7

30.3

Pittsburgh

PA

520,000

16,000

22%

2016 to present

1,446
(9.0%)

20

11.1

Central
Arkansas
Water

AR

330,667

175

0.10%

2016-2017(14
months)

115

(66%)

0.89

1.5

Saskatoon

Canada

313,000

4,582

6.10%

2017 to 2022

488
(11%)

6.5

9.1

Newark

N.T

294,274

23,189

64%

2019 to 2022

7,730
(33%)

212

3.0

Grand Rapids

MI

273,005

1,608

2.00%

2021 to 2022

304
(19%)

3.8

5.3

Spokane

WA

244,817

486

0.60%

2016 to 2018

162

(33%)

1.9

3.0

Trenton

N.T

217,000

20,000

32%

2017 to 2022

1,372
(8.0%)

22

12.5

2


-------
City

State

Population

Served

Total
Number of
LSL/GRR

%of
Total

SLs

Duration of
Program

Avg.
Replaced
per year

Annual
Replacements
per 1,000 service
connections

Years To
Complete at
This Rate

Aurora

IL

200,000

17,729

35.70%

2022

612
(3.5%)

12

28.6

Sioux Falls

SD

198,524

230

0.38%

2016-2017(32
months)

115

(50%)

1.9

2.0

York

PA

197,177

2,300

2.90%

2017-2021

380
(17%)

5.8

5.9

Kalamazoo

MI

192,992

10,000

24%

2015 to Present

505
(5.0%)

12

20.0

Lansing

MI

166,000

12,150

22%

2004 to 2016

1,013
(8.3%)

18

12.0

Lancaster3

PA

120000

N/A

N/A

2012-2023

23
(N/A)

0.51

N/A

Elgin

IL

113,911

13,500

40.70%

2017-2022

292
(2.2%)

8.8

45.5

Green Bay

WI

107,395

2,028

5%

.Tan 2016 to Sep
2020

357
(18%)

9.3

5.6

Quincy

MA

101,636

285

1.20%

April 2017-
September
2018

206
(72%)

8.7

1.4

Flint

MI

98,310

12,035

37%

2016 to 2022

1,946
(16%)

59

6.3

Newton

MA

89,103

433

1.70%

2017-2019

144

(33%)

5.8

3.0

Somerville

MA

81,045

449

3.60%

2021 -2022

86
(19%)

5.6

5.3

Evanston

IL

75,570

10,803

37.50%

2017-2022

184
(1.7%)

6.4

58.8

Framingham

MA

72,362

184

1.08%

2004-2016

1

(0.5%)

0.06

184.0

Madison

WI

71,160

8,000

9%

2000 to 2011

728
(9.1%)

8.4

11.0

St. Clair
Shores

MI

59,715

1227

4.80%

2020

100
(8.1%)

4.0

12.3

Revere

MA

59,075

350

2.90%

2019-2021

83
(24%)

7.0

4.2

Bozeman

MT

56,000

170

0.70%

2016-2019

35
(20%)

2.9

5.0

N/Ab

VT

N/A

N/A

N/A

2021-2023

167
(N/A)

N/A

N/A

Bloomfield

N.T

47,315

500

4.10%

2018-2021

130
(26%)

11

3.8

Battle Creek

MI

43,975

5,000

22.50%

2022

140
(2.8%)

6.3

35.7

Marlborough

MA

38,499

1,350

13%

May 2018-
Sept 2018

176
(13%)

17

7.7

Galesburg

IL

31,745

3,500

28%

2016 to present

530
(15%)

42

6.7

Village of
Montgomery

IL

28,956

106

1.20%

Fall 2019 to
Summer 2020

106
(100%)

12

1.0

3


-------
City

State

Population

Served

Total
Number of
LSL/GRR

%of
Total

SLs

Duration of
Program

Avg.
Replaced
per year

Annual
Replacements
per 1,000 service
connections

Years To
Complete at
This Rate

Norwood

MA

28,284

200

2.20%

2004-2008

40
(20%)

4.5

5.0

Sandusky

OH

25,793

N/A

N/A

2021

39
(N/A)

3.7

N/A

Winchester

MA

22,800

21

0.29%

Mar 2017 to
2019

7

(33%)

0.93

3.0

Birmingham

MI

20,472

730

8.60%

2020 to 2022

182

(25%)

21

4.0

Frankfort

IL

20,296

82

0.70%

2021-2022

41

(50%)

3.6

2.0

Menasha

WI

14,792

636

12%

2017 to 2023

106
(20%)

19

5.0

Stoughton

WI

13,078

700

14%

2021

700
(100%)

144

1.0

Mayville

WI

5,112

220

12%

2021

220
(100%)

116

1.0

Evart

MI

1,903

500

72.36%

2019

40

(8%)

58

12.5

a Replacement rate data identified in public comments to the proposed LCRI (USEPA, 2023b).

b System was reported anonymously in the Vermont Drinking Water and Groundwater Protection Division public comment on
the proposed LCRI; thus, any identifying information about the system, including the number of service connections, was not
available.

Water systems are sorted by population served. "N/A" is used for cells with no data available. "GRR" = galvanized requiring
replacement. "SL" = service line. See Appendix 1 for the data sources for each water system.

Of the 48 systems for whom the EPA identified replacement rate data, 46 systems serve
populations of more than 10,000 persons and 2 systems serve populations of 10,000 or fewer
persons ("small" systems, as referred to in SDWA section 1412(b) and as finalized as the small
water system threshold in the agency's Consumer Confidence Report regulation (63 FR 44524,
USEPA, 1998)).

For the proposed LCRI, replacement rate data from 30 systems serving more than 50,000 persons
were used, and the 95th percentile rate (or 0.039 annual replacements per household served) was
determined to be the feasibility threshold on a per-household basis. In the final LCRI, the EPA
expanded the analysis to include 12 additional systems serving between 10,001 and 50,000
persons to increase the sample size and allow the EPA to better understand the feasibility of
service line replacement for a wider variety of large systems. Data for systems serving between
10,001 and 50,000 persons were originally documented in the supporting information for the
proposed Lead and Copper Rule Improvements (USEPA, 2023b). In addition, data for three
systems provided during the public comment period was added that had not previously been in
the proposed LCRI dataset. Data from Newark, NJ and one of the four systems obtained in the
LCRI public comments (due to being reported anonymously, see Exhibit 2.1) were excluded. In
total, the EPA used replacement rate data from 44 systems to inform the final rule feasibility
threshold expressed as a per-connection annual replacement rate. The 95th percentile value of this
dataset is 39 replacements per 1,000 service connections (Exhibit 2.2).

4


-------
Exhibit 2.2: Summary Statistics on the Annual Replacement Rates per 1,000
Service Connections Previously Achieved by Water Systems, by Size
Category



Median

75th Percentile

90th Percentile

95th Percentile

Maximum

System Size

Annual Replacements per 1,000 Service Connections

Serving more than
10,000 people
(n=44)

6.6

12

21

39

144

Serving 10,000 or
fewer people (n=2)

58

N/A

N/A

N/A

116

All Systems (n=46)

6.8

11

32

59

144

System Size

Annual Replacements per 1,000 Households Served

Serving more than
10,000 people
(n=44)

5.1

8.2

18

39

135

Serving 10,000 or
fewer people (n=2)

81

N/A

N/A

N/A

109

All Systems (n=46)

6.2

10

36

53

135

For the final LCRI, the EPA estimated the total time it would take each system identified for the
analysis in Exhibit 2.1 to complete their mandatory service line replacement program, assuming
systems consistently replace lead and GRR service lines at the average annual replacement rates
identified in Exhibit 2.1. until all their lead and GRR service lines are replaced. This was
calculated by dividing 100% by the average percent of service lines replaced per year based on
the documented period in Exhibit 1.1. For example, York, PA, replaced an average of 17% of
their lead and GRR service lines per year from 2017-2021, which equates to 100/17 or 5.9 years
to completion.

Among the 44 systems serving more than 10,000 people, 27 (61%) systems were, at the time of
this analysis, replacing service lines at a rate that would lead to the completion of their
replacement programs in 10 years or less.

Availability of LSLR Data in Small Systems

Although not included in the EPA's assessment of an annual per-service connection replacement
rate, the EPA did identify two small systems (serving <10,000 people) with data from official
data sources. The EPA does not expect this is due to a lack of small systems conducting LSLR,
but rather the EPA expects the prevalence of official data in the sources investigated is more
likely available for larger systems as opposed to smaller systems. This is likely due in part to the
2000 EPA Public Notification Rule requirement for all systems serving 100,000 or more persons
to post their annual water quality reports online, while there is no current requirement for
systems serving fewer than 100,000 persons (USEPA, 2000). The revised Consumer Confidence
Report requirements lowers this threshold to 50,000 persons (89 FR 45980). The EPA observed
in the data used to populate Exhibit 1 that systems serving approximately 30,000 or more people
were more likely to have websites that were updated frequently enough to post information on
LSLR rates, possibly reflecting the greater capacity of systems in this size range to maintain a

5


-------
website even when not specifically required to do so. For systems serving fewer than 30,000
people in this dataset however, the EPA observed that they were less likely to have updated their
own independent website to reflect LSLR rates. For example, the data for the only system
serving < 3,300 people was obtained from a storyboard prepared by a consulting firm on the
system's behalf, and official data from many other systems serving fewer than 30,000 people but
more than 10,000 people (Frankfort, Stoughton, Mayville, Winchester, Norwood, etc.) were all
obtained from reports prepared by their respective State agencies rather than by the systems
themselves.

Examining systems that have been awarded Bipartisan Infrastructure Law (BIL) funds for LSLR
provides evidence that additional small systems have service line replacement programs for
which information on replacement rates was not available from official sources used to identify
replacement rates. Of the 85 systems receiving BIL funds for LSLR, 20 systems (23.5%) served
fewer than 3,300 people and a further 23 (27.1%) served between 3,300 and 10,000 people.
Together, 50.6% of systems that have been awarded BIL funding serve 10,000 people or fewer
("BIL LSLR funded projects as of 9-25-23.xlsx").

Similarly, the EPA is aware of additional systems for which service line replacement programs
and activities are underway; however, official data sources on replacement rates are not
available. For example, replacement programs have been identified through news articles in
Massachusetts (City of Chicopee, n.d.; Rhodes, 2023), Ohio (Garner, 2022; Vasko, 2022),
Michigan (Schulwitz, 2022), and Wisconsin (Leischner, 2022) that mention upcoming lead
service line identification and replacement programs, but the accompanying water department
websites either do not exist or do not contain updated data on replacement rates.

Service Connections in Lieu of Households Served

For the final rule, the EPA evaluated replacement rate data on a per-service-connection basis,
rather than a per-household-served basis as in the proposed rule. The EPA received public
comment recommending the agency include service connections rather than households in this
evaluation to simplify and improve the implementability of the LCRI. In the EPA's Federal Safe
Drinking Water Information System (SDWIS) database, the term "Service Connections Count"
can be viewed to obtain the number of service connections for any water system nationwide. No
column exists for the number of households of each water system in SDWIS, leaving the
calculation of the household term up to each individual water system (USEPA, 2023b).

The EPA evaluated the effect that the change from using replacement rates based on households
served to service connections in the distribution system had on the results of the feasibility
analysis. Comparable summary statistics for per-household-served replacement rates show only
minor differences from per-service-connection rates in Exhibit 2.2, highlighting that calculating
the replacement rate threshold as per-service connection does not result in significant changes as
compared to a per household rate threshold. (Exhibit 2.2).

2.2 Projection of Systems Expected to Exceed Deferral Threshold Using Updated
Needs Survey Data

Projected Number of Eligible Systems for Deferred Deadlines

6


-------
For the final rule, the EPA estimated the number of systems eligible for a deferred deadline by
projecting the number of systems that are likely to exceed 39 annual replacements per 1,000
service connections. In its revised feasibility analysis, the EPA utilized updated data by
analyzing responses from the One-Time Update to the 7th Drinking Water Infrastructure Needs
Survey and Assessment (or Needs Survey), published in May 2024 (USEPA, 2024a). The Needs
Survey collected data to represent the DWSRF-eligible infrastructure projects that are necessary
in the 20-year period from January 2021 to December 2040. For the first time, the most recent
Needs Survey collected service line material information to include the cost to replace LSLs, in
accordance with the America's Water Infrastructure Act of 2018 (USEPA, 2023a). The service
line material questionnaire provided for water systems to list the number of service lines that
they have categorized as lead, galvanized lines downstream of lead pipe, lead connectors,
unknown pipes, or connectors, and service lines of other or unknown materials. Seventy-five
percent of water systems provided responses to the service line material questionnaire. The EPA
surveyed all large community water systems (CWSs) serving more than 100,000 people, a
random sample of CWSs serving between 3,301 and 100,000 people in each state, and national
random samples of CWS serving fewer than 3,300 people and non-public non- community water
systems. The EPA analyzed information from the One-Time Update on 3,588 water systems that
provided service line material information (USEPA, 2024a). The Updated Needs Survey dataset
includes responses from 1,786 out of approximately 4,500 community water systems (CWSs)
serving more than 10,000 people and 1,626 out of approximately 40,000 CWSs serving 10,000
or fewer people. Responses were also received from 126 non-community water systems
(including transient and non-transient systems); however, among these systems, no lead or GRR
service lines were reported. The response rate 2 (RR2) among CWSs according to AAPOR
standards (AAPOR, 2023) was 6.9%. The RR2 considering only systems serving greater than
10,000 people was 40%. This difference is due to the larger number of CWS serving 10,000
people or fewer. Of the 49,396 CWSs listed in the SDWIS database, 44,867 serve 10,000 people
or fewer (91%). Accordingly, the amount of the population served by participating CWSs is
much higher than the RR2 value. Specifically, among CWSs serving more than 10,000 people,
participating CWSs represent nearly 200 million people, approximately 74% of the total
population among these CWSs. Additionally, the AAPOR cooperation rate 2 (COOP2) was 97%,
highlighting the high degree of participation among systems which were contacted. For more
information on the Needs Survey process and methodology, please see USEPA, 2023a.

Furthermore, the EPA could not connect 50 water systems to the SDWIS/Fed data, possibly due
to changes in water system status or consolidation of systems. Therefore, responses from 3,412
CWS were used to determine their estimated number and proportion of lead and GRR service
lines, and their required annual replacement rate to meet a 10-year deadline.

Consistent with the approach in the proposed rule, the EPA accounted for the presence of
unknown service lines using two different methods (low-bound and best) to estimate the total
number of lead and GRR service lines in each system. For the low-bound estimate, the EPA
assumed that all reported unknown service lines are non-lead service lines when calculating each
system's annual replacement per service connection value. For the best estimate, the EPA
evaluated unknown lines using methodology consistent with the EPA's analysis of the Needs
Survey results for the Drinking Water State Revolving Fund (DWSRF) allocations (USEPA,
2023a). For this method, the EPA projected the composition of reported unknown lines by (1)

7


-------
applying the combined percentage of known lead and GRR service lines among all known
service lines for each State and (2) applying that rate to all unknown service lines within systems
in that State. This method allowed the EPA to account for unknown service lines when
estimating the number of systems eligible for a deferred deadline under the final rule.

The EPA estimated the total number of systems exceeding the per-connection rate threshold by
calculating the proportion of surveyed systems in each population size category estimated to
exceed the threshold. This allowed the EPA to extrapolate results to all remaining (i.e., non-
surveyed) systems in each size category. Exhibit 2.3 shows number and percent of all systems
estimated to exceed the eligibility threshold for a deferred deadline under the proposed and final
LCRI (Exhibit 2.3). The EPA calculated the low and best estimates for each scenario. The
scenarios include the proposed rule and the final rule. The EPA also included the number of
water systems that would have exceeded the eligibility threshold using the proposed per-
household rate with the updated Needs Survey data (alternate scenario). This analysis was
conducted to determine the difference in projected eligibility for deferred deadlines based upon
the choice of normalization variable (households vs service connections).

Exhibit 2.3: Low-bound and Best Estimates of Systems Likely to Exceed the
Per-connection Replacement Threshold in the Final LCRI Compared to the
Proposed LCRI and the Alternate Per-household Scenario

Threshold

Eligible

Systems
Serving
>10,000

Eligible

Systems
Serving <
10,000

Total Eligible

Systems
(All Sizes)

Percent of
Total Systems
Eligible
(All Sizes)

Final Rule:

101-116

(2.2-2.6%)

387-498
(0.9-1.1%)





39 replacements per
1,000 connections

488-614

1.0-1.2%

Alternate Scenario for









the Final Rule:
39 replacements per
1,000 households served

68-83
(1.5-1.8%)

415-609
(0.9-1.4%)

483-692

1.0-1.4%

Proposed Rule:
0.039 replacements per
household served3

146-437 A
(1.5-4.5%)

570-1,737 A
(1.4-4.3%)

716-2,174 A

1.4-4.4%

a Hie proposed rule used the results of the 7th Drinking Water Infrastructure Needs Survey and Assessment (Needs Survey),
while the final rule analysis used the results of the One-Time Update to the Needs Survey released in 2024.

The results in Exhibit 2.3 show that the 10-year replacement deadline for the final LCRI is
feasible for the vast majority of water systems nationwide. It also demonstrates that the EPA's
updated deferred deadline eligibility criteria based on the final rule's updated feasibility analysis
makes only a minor impact on the percentage of systems projected to exceed the deferred
deadline threshold as compared to the proposed rule.

The use of a per-service-connection basis in the final rule results in a slightly higher proportion
of systems serving more than 10,000 persons eligible for deferred deadlines as compared to the
alternate scenario per-household-served basis (Exhibit 2.3). These results demonstrate a very

8


-------
small impact on using a per-service-connection basis versus a per-household-served basis. The
EPA expects, however, that the use of a per-connection basis is more representative of the
feasibility of the required scale of service line replacement for large systems Using the average
size of a household is not an accurate predictor of the number of service lines in a system nor is it
a predictor of the number of bill paying households in a system. For example, for water systems
with many multifamily homes or complexes the average size of a household is not an accurate
predictor of the number of service lines in a system. Chicago, IL, has 484,979 reported service
connections. If the per-household-served basis was used, Chicago's population served of
2,746,388 would be divided by 2.53 persons per household (the average persons per household
from the 2020 Census Bureau used in the proposed rule), resulting in an estimated 1.1 million
households served. This is twice as many service connections that the city actually has and does
not accurately reflect the proportion of service lines that require replacement. As a result, when
the number of service connections is used to normalize Chicago's 430,067 projected service lines
requiring replacement, it results in a value of 89 annual replacements per 1,000 service
connections, as opposed to a value of 39 annual replacements per 1,000 households served.

The difference in projections for the 0.039 annual replacements per household served scenario
from the proposed rule to the alternative scenario of 39 annual replacements per 1,000
households served in the final rule is due to the use of the One-Time Update to the Needs Survey
results. Prior to the One-Time Update to the Needs Survey, several potential errors in the dataset
had been documented, such as no response from New York City (who had documented the
number of service lines of each type in their water system online) and the likely erroneous
reporting of 200,000 LSLs in Houston, TX (Alkafaji, 2023). These issues were corrected in the
One-Time Update to the Needs Survey data, in addition to other changes in responses from
various systems and the addition of responses from other systems that did not respond previously
(USEPA, 2024a). The updated data had only a minor effect on the low-bound estimate because
unknown lines were assumed to be non-lead for the low-bound estimate and only changes in
reported number of lead or GRR service lines caused changes in low estimates. The best estimate
had more changes with the One-Time Update to the Needs data. For example, the removal of the
200,000 erroneously reported LSLs in Houston impacted every Texas system because unknown
service lines were projected based on the percentage of reported lead and GRR service lines
statewide (USEPA, 2023a). Removing 200,000 LSLs from the statewide pool reduced the State
percentage of reported lines that were lead or GRR service lines, thus many unknown lines in
Texas systems that were projected as lead or GRR service lines in the proposed rule are no
longer projected as lead or GRR service lines in the final rule. Corrections and refinements such
as these resulted in significantly fewer systems projected to exceed the specified eligibility
thresholds in the final rule analysis.

Length of Deferred Deadlines

The EPA calculated the range of maximum length of the deferred deadlines for all systems with
available data from the Needs Survey, including the One-Time Update, that would have to
replace at least 39 service lines per 1,000 service connections per year, based upon the best-
available estimate of the number of lead and GRR service lines. The calculated value does not
take into account the potential for the State to set a faster rate where feasible upon regular check-
ins with the State, therefore the actual deferred deadlines may be shorter than the maximum
length. The maximum length in years of each system's deferred deadline was calculated by

9


-------
multiplying the number of service connections by 0.039 (equivalent to 39 replacements per 1,000
service connections). The total number of lead and GRR service lines was then divided by this
value, in years, to determine the length of the deferred deadline for each system. Thus, for a
given system, both the number of service connections and the number of lead and galvanized
requiring service lines in the system are required to be known. For this reason, the length of
deferred deadlines could only be calculated for systems included in the One-Time Update to the
Needs Survey, and the EPA could not determine a reliable way to calculate the distribution of
deferred deadline lengths among systems expected to be eligible. Since the Needs Survey aimed
to capture a representative sample of systems nationwide, a similar distribution of deferred
deadline lengths is expected for systems nationwide. Under the best estimate, a total of 64 out of
the 3,412 surveyed CWSs would be eligible for a deferred deadline, providing a maximum
deadline of 1 to 19 additional years to complete mandatory service line replacement. This range
of deferred deadlines corresponds to a total maximum timeline to complete service line
replacement ranging from 11 to 29 years. Most systems evaluated, however, that are eligible for
a deferred deadline would be required to complete mandatory service line replacement in 15 or
fewer years (59%). In the Needs Survey sample dataset, fewer systems serving 10,000 or fewer
people were found to be eligible for deferred deadlines (n = 3, median = 15.4 years) than systems
serving more than 10,000 people (n = 61, median =13.4 years), however the small systems that
were eligible had longer deadlines.

For any system which is aware of their number of service lines requiring replacement and service
connections, the length of the deferred deadline can be determined by dividing the percentage of
lines requiring replacement (number of lead and GRR service lines/total number of service
connections) by 0.039 replacements per service connection. This calculation applies to systems
of all sizes. Thus, for any system, the length of a potential deferred deadline available can be
visualized in Exhibit 2.4.

10


-------
Exhibit 2.4: Length of Deferred Deadline for a Generic System of Any Size
with a Given Proportion of Service Lines Requiring Replacement

30



















ZD

S

i§ 20
a


-------
To evaluate the potential impacts of the 10,000 replacements per year threshold, the projections
based upon the best estimates from the One-Time Update to the Needs Survey data were
analyzed. The number of reported lead and GRR service lines, the number of unknown service
lines, and the best estimate case (incorporating projected unknowns in the same manner as the
above analysis) were all examined.

All systems from the Needs Survey dataset were examined, and all 32 water systems with a
combined total of more than 100,000 lead, GRR, and unknown service lines are posted to the
docket ("List of Large Systems.xlsx"). In the Needs Survey, the best estimate includes three
water systems over 100,000 lead and GRR service lines. Two of these systems, Chicago, IL and
Cleveland, OH, will qualify for a deferred deadline via exceedance of the 39 replacements per
1,000 service connections criteria. The remaining system, New York, NY, has a required
replacements per year value less than 39 replacements per 1,000 service connections, and thus is
expected to be feasible based on the EPA's feasibility analysis.

While the best estimate projection of Needs Survey systems accounts for unknown service lines
based upon the percentage of service lines classified as lead or GRR at the State level, it is
possible some systems have a higher rate of lead and GRR service line occurrence than the
projected value. For this reason, the EPA examined the number of reported unknown service
lines for each water system and calculated what percentage of remaining unknown service lines
would have to be classified as a lead or GRR service line for each water system to exceed
100,000 lead and GRR service lines ("List of Large Systems.xlsx"). The percentage of unknowns
classified as a lead or GRR service line to exceed 100,000 total lines requiring replacement
ranged from 21% to 97% (average=61%).

2.4 Potential for Faster Replacement Rates in the Future

Several factors may facilitate accelerated service line replacement rates in the future that would
not have been available for past replacement programs evaluated in the feasibility analysis.
Therefore, it is possible that in the future systems may be able to replace service lines faster than
the final LCRI's feasibility threshold. States are required to set a faster replacement rate where
feasible for a system.

The EPA anticipates that as water systems and their contractors gain more experience conducting
service line replacement, they are likely to become more efficient in conducting replacements
and engaging with customers and therefore be able to complete them more quickly and using
fewer resources. For example, Denver Water has released annual reports documenting the
progress of their LSLR program since January 2020 and has documented key lessons learned to
streamline the LSLR process in each reporting (Denver Water, 2023b). Over this time, LSLR
conducted have increased from 5,287 in 2020 to 6,891 in 2023.

Advancements in service line replacement have been and are also expected to continue to be
documented in various sources. Resources on replacing service lines, such as lessons learned

12


-------
from systems with proactive replacement plans or experiences implementing a new technology,
are available from several sources to support systems to implement their replacement programs
(see Appendix 2 for examples). These resources can help water systems save time and resources
by providing information and best practices through the lessons learned from other systems.
Similarly, as systems develop their service line inventories following the 2021 LCRR
compliance date of October 16, 2024, and submission of initial inventories, availability on
service line inventory information is likely to increase (see Appendix 3 for examples).

3 Feasibility of Service Line Inventory Requirement

The EPA analyzed service line inventory data from five States in which submission of service
line inventories had been required, along with data from the Needs Survey, to evaluate progress
made in past efforts to identify unknown service lines. Updated service line inventory data from
Illinois and the data resulting from the One-Time Update of Needs Survey was available and
incorporated into the final rule's feasibility analysis.

3.1 States with Service Line Inventory and Replacement Requirements
In 2016, both California and Ohio passed legislation requiring systems to develop LSL
inventories, where Ohio required systems to submit a map identifying areas in the system that are
known or likely to contain LSLs by 2017, and California required systems to compile an
inventory of known lead user service lines1 and likely lead user service lines by 2018 (California
Legislature 2017; Voltzer, 2016). A 2019 survey conducted by 120Water received responses
from 2,811 out of 4,402 CWSs and NTNCWSs in California and documented that only 8.7
percent of the service lines inventoried were classified as unknown (860,962 out of 9,852,421
lines) after three years of LSL investigations (Walker and Jacquette-Morrison, 2019). The
120Water survey also recorded that, by early 2020 in Ohio, 1,919 CWSs and NTNCWSs had
submitted maps of areas known or likely to contain LSLs to the State even though the data itself
was not publicly accessible2 (Walker and Jacquette-Morrison, 2019).

In 2018, both Michigan and Wisconsin adopted rules requiring submission of an LSL inventory,
both requiring initial submissions by 2020. The Wisconsin Public Service Commission published
the results of the systems inventory data in 2020, in which only five percent of reported system-
side service lines and six percent of reported customer-side service lines were characterized as
unknown (Public Service Commission of Wisconsin, 2021). The Michigan Department of
Environment, Great Lakes, and Energy (EGLE) also published inventory data in 2020, showing

1	California refers to a service line as a user service line. According to the California Code of Regulations (CCR), a
user service line is the "pipe, tubing, and fittings connecting a water main to an individual water meter or service
connection" (22 CCR § 64551.60). The definition of a lead user service line includes lead goosenecks or pigtails
connected to the user service line on the water system side of the meter. This definition does not include the service
line from the meter to individual homes (the customer side), so water systems can classify what is considered a
partial LSL as a lead user service line.

2	The Ohio Environmental Protection Agency created a webpage that hosts all maps submitted by systems:
https://epa.ohio.gov/divisions-and-offices/drinking-and-ground-waters/reports-and-datMead-lines-mapping. This
webpage was last updated January 6, 2022, and includes 1,997 map entries.

13


-------
44 percent of all service lines are listed as unknown across the state (Michigan Department of
Environment, Great Lakes, and Energy, 2020). Of these 44 percent of unknown service lines,
however, 25 percent of unknowns were reported as "unknown-likely not lead" and seven percent
were reported as "unknown4ikely lead. This indicates that systems have an indication of where
LSLs and non4ead service lines may be located. The EPA notes all such lines would be
considered unknown under the proposed LCRI.

The EPA is also aware of systems not required by their State who have worked to develop
service line inventories, such as the following systems: the City of Grand Forks, ND; the City of
Lincoln, NE; the City of Somerville, MA; the City of Troy Department of Public Utilities, NY;
DC Water, DC; Marshfield Utilities, WI; Memphis Light, Gas, and Water, TN; Milwaukee
Water Works, WI; Pittsburgh Water and Sewer Authority, PA; and Saint Paul Regional Water
Services, MN (USEPA, 2023b; USEPA, 2024c).

Updated Inventory Information from Illinois Database

In 2017, Illinois passed a law that required CWSs to develop and submit a water distribution
system material inventory (or service line inventory), to the Illinois Environmental Protection
Agency (Illinois EPA) beginning in 2018 (Illinois Municipal Code 2017). In 2022, this law was
repealed by a statewide LSLR regulation, which requires systems to identify all unknown lines
by their applicable replacement deadline (between 15 and 50 years) (Illinois General Assembly

2021).	The Illinois EPA posts system inventory progress to their website each, starting with
2017, and provides a downloadable dataset that contains the inventory data of each system. As of
2024, the State had received inventories from a total of 1,769 unique systems (Illinois EPA,

2022).	Over five years of recorded data, the percentage of unknown service lines reported
statewide decreased from 41% in 2017 to 16% in 2022 (Exhibit 3.1).

Exhibit 3.1: Summary of Remaining Unknown Service Lines in Service Line
Inventories Submitted to Illinois EPA from 2017-2022







Percent

Percent







Percent

Unknowns of

Unknowns of

Percent



Number of Systems

Unknown-

Median

75th Percentile

Unknowns of 90th

Year

Reporting

Statewide1

System

System

Percentile System

2017

1,660

41% (41%)

5.8%

83%

100%

2018

1,750

28% (31%)

0%

61%

100%

2019

1,755

21% (29%)

0%

42%

97%

2020

1,757

21% (28%)

0%

39%

97%

2021

136

18% (20%)

0%

20%

73%

2022

1,430

16% (26%)

0%

14%

72%

:In 2018, systems were permitted to classify some unknowns as "Unknown-Not Lead." The statewide percent unknown is
provided: not including these lines as unknown and (including "Unknown-Not Lead" with other unknowns).

Notably, in 2017, the year the service line inventory law was passed, the median percentage of
unknown service lines remaining was 5.8%, indicating that many systems had already
determined all or nearly all of the service line materials and completed or nearly completed their
service line inventory. Additionally, even systems who did have unknown service lines at the

14


-------
start of the program were generally able to identify materials relatively quickly. The 75th
percentile system, which had 83% of service lines characterized as unknown in 2017, was able to
reduce the percentage of unknowns remaining by 69%, down to only 14% remaining in 2022.
Thus, for most systems, more than half of remaining unknown service lines were able to be
identified within the first 4 years of service line identification efforts. Further, Illinois data as of
2022 does not represent final inventories (which are not due according to Illinois State law until
2024), and thus some systems had not yet begun identifying service lines, but unknown service
lines in these systems are still included in these statistics. Analysis of data at the individual
system level shows that many systems have already made substantial progress on their
inventories and achieved substantial rates of service line material identification. For example, the
median system in Illinois had zero percent of unknown lines reported in 2018, which is only one
year after the previous inventorying law went into effect. Thus, more than half of all systems
identified all their remaining unknown service lines within the first year. This shows that in
Illinois, one of the states with the highest estimated number of LSLs, the median system has
already determined the material of all their service lines. This suggests that water system records
for the many systems are available, and prior identification work may have taken place.
Additionally, analysis of the number of systems with unknown service lines remaining in 2022,
the deadline for the Illinois initial inventory, can provide the best available approximation of the
number of unknown service lines expected to remain nationally following the 2021 LCRR initial
inventory. At the Illinois initial inventory deadline in 2022, only 44 of the 1,430 (3.1%) systems
reporting service line inventories in 2022 had not yet begun identifying service lines and half of
reporting systems had zero unknown service lines remaining. Of these 44 systems, 40 were small
systems, however these systems represented only 3.6% of all small systems. For the LCRI, the
analysis of eligibility for deferred deadlines will not occur until the LCRI compliance date in
2027, three years after the due date for the LCRR initial inventory in 2024. Thus, these statistics
from Illinois, which are comparable to the LCRR initial inventory, do not account for the same
amount of time for unknown investigation that will occur prior to the LCRI compliance date and
LCRI baseline inventory due date used for deferred deadline eligibility. Therefore, the vast
majority of systems of all sizes had made progress identifying unknown service lines by the
initial inventory deadline in Illinois and would have several additional years for service line
identification prior to the point where eligibility for deferred deadlines would be evaluated.

3.2 Updated Inventory Information from the Needs Survey to Inform Inventory
Development Feasibility

The 7th Drinking Water Infrastructure Needs Survey and Assessment (or Needs Survey) was
administered by the EPA in 2021, followed by a one-time update to the data collected in 2023
(USEPA, 2024a). The Needs Survey provides data to evaluate service line materials nationwide.
The EPA utilized the more recent 2023 data to update the 2021 data and determine the progress
of service line inventories as of 2023.

For the purposes of this analysis, the EPA evaluated responses from 3,588 water systems,
representing 1,786 from systems serving 10,000 or more people, 1,626 systems serving fewer

15


-------
than 10,000 people, 126 non-community water systems, and 50 systems which were not able to
be mapped to SDWIS information. While the Needs Survey is nationally representative across
system sizes due to the random selection of water systems serving fewer than 100,000 people to
be surveyed, there is a higher degree of uncertainty in applying service line information to non-
surveyed systems serving fewer than 10,000 people. This is due to the relatively larger number
of systems serving 10,000 or fewer people (approximately 45,000 in SDWIS) compared to those
serving 10,000 or more people (approximately 4,500 in SDWIS).

Of the 3,412 community water systems with available data in the 2023 data, the median system
had 65% of service lines reported as a known material, leaving 35% of service lines not
classified. This 35% of lines without a known material type includes responses from 760 systems
(22%) that did not report the composition of any service lines in their survey response. Of the
2,652 systems who reported the status of at least one service line, the median system had only
6.5% of service lines categorized as unknown (Exhibit 2.2). This suggests that, for most systems
who have started their service line identification programs as of 2023, service line records are
available or other methods are being effectively used. It is important to note that the one-time
update was collected in 2023, one year before the 2021 LCRR initial inventory deadline, and
thus systems with many remaining unknowns may still begin identifying unknowns in their
service line inventory prior to the 2021 LCRR initial inventory deadline in 2024. Thus, this
analysis is likely an overestimate the number of unknown service lines which will remain after
all systems have completed their initial inventories.

Exhibit 3.2: Summary of Remaining Percentage of Service Lines Classified as
"Unknown" from the 2023 EPA Needs Survey One-Time Update Among Systems
Reporting Service Line Materials

System Size

Percent Unknown of
Median System

Percent Unknown of
75th percentile system

Number of Systems Not
Reporting Any Lines

Serving < 10,000 persons
(n= 1.626)

0%

93%

443A

Serving >10,000 persons
(n= 1,786)

13%

79%

317B

All systems (n=3,412)

6.5%

84%

760

AAn additional 258 systems reported entirely unknown service lines
BAn additional 234 systems reported entirely unknown service lines.

Data was received in December 2022 and one-time update was collected in fall 2023.

While the median system serving fewer than 10,000 people had identified all of their unknown
lines, there were also relatively more systems serving fewer than 10,000 people who reported
entirely unknown service lines and who did not report any lines. (Exhibit 3.2). This could be due
to the relatively fewer number of lines to identify in systems serving fewer than 10,000 people.
In other words, systems serving fewer than 10,000 people were more likely to have finished their
programs, however systems serving more than 10,000 people were more likely to have programs
that have been started and are currently in progress. Additionally, the percentage of remaining
unknowns in Exhibit 3.2 was determined by excluding systems which did not report any service
lines but including any systems which reported entirely unknown service lines. This was done to
reflect the difference between systems that didn't respond to the survey and those that did
respond to indicate they had not yet begun identifying unknowns. In either case however, the

16


-------
EPA expects that progress will be made prior to the 2021 LCRR initial inventory 2024 deadline
and further progress prior to the 2027 LCRI compliance date.

The data show that in all States in which LSL inventories were required (C A, OH, IL, MI, WI),
which represent the states with among the highest counts of LSLs in the nation (USEPA, 2024a),
no more than 50 percent of all lines in the state were classified as unknown within two years of
the requirement (See "Michigan LSLR Analysis.xlsx", "New Jersey LSLR Analysis.xlsx",
"Illinois Downloaded Inventory Analysis.xlsx" in EPA-HQ-OW-2022-0801). This suggests that
initial records of service line material were robust, and that any service line material
investigations that have taken place, have resulted in the majority of service line materials having
been identified.

3.3 Systems with Completed LSLR Programs

The EPA is aware of several water systems who have fully replaced LSLs from their distribution
systems. Replacing all LSLs requires an inventory where all unknowns have been identified;
therefore, analysis of completed LSLR programs can provide information about a feasible
timeline that systems were able to identify the material of all unknown service lines in their
distribution systems. Exhibit 3.3 includes data about system LSLR programs which have
completed their replacement programs and have thus identified all unknown lines.

Exhibit 3.3: Water Systems with Completed LSLR Programs

City

State

Population

Total
Number of
LSL/GRR

%of
Total
SLs

Duration of
LSLR
Program

Avg. # Replaced Per
Year (% of Total
LSL/GRR)

Tucson

AZ

675,686

600

0.31%

2016 to 2018

47 (7%)

Newark

NJ

294,274

23,189

64%

2019 to 2022

7,730 (33%)

Spokane

WA

244,817

486

0.60%

2016 to 2018

162 (33%)

Lansing

MI

166,000

12,150

22%

2004 to 2016

1,013 (8%)

Green Bay

WI

107,395

2,028

5%

Jan 2016 to
Sep 2020

357 (18%)

Newton

MA

89,103

433

1.7%

2017-2019

144 (33%)

Framingham

MA

72,362

184

1.08%

2004-2016

1 d%)

Madison

WI

71,160

8,000

9%

2000 to 2011

727 (9%)

Village of
Montgomery

IL

28,956

106

1.20%

Fall 2019 to
Summer 2020

106 (100%)

Norwood

MA

28,284

200

2.2%

2004-2008

40 (20%)

Winchester

MA

22,800

21

0.29%

Mar 2017 to
2019

7 (33%)

Frankfort

IL

20,296

82

0.70%

2021-2022

41 (50%)

Menasha

WI

14,792

636

12%

2017 to 2023

127 (20%)

Stoughton

WI

13,078

700

14%

2021

700 (100%)

Mayville

WI

5,112

220

12%

2021

220 (100%)

17


-------
The data shows that a range of systems have completed LSLR programs thus identified all
unknowns. The EPA notes that systems such as Tuscon, AZ, and Spokane, WA, despite having
relatively few LSLs at the start of their replacement program, are nevertheless strong indicators
of inventory feasibility because the total number of service lines in the distribution system is
more relevant to completing an inventory than the number of LSLs because the proposed LCRI
requires systems to identify the material of all service lines. While Madison, WI, and Lansing,
MI, took longer than the 10-year deadline to complete both their inventory and replacement
program (i.e., approximately 11 years for Madison and 12 years for Lansing), the EPA does not
interpret these or other examples of systems completing LSLR in over 10 years as evidence that
a 10-year deadline is necessarily infeasible. The agency is not aware that these systems were
necessarily replacing service lines in accordance with the SDWA requirement to "prevent known
or anticipated adverse health effects to the extent feasible" as is required of the LCRI, per
SDWA. Additionally, these systems initiated their LSLR and inventorying simultaneously, while
the LCRR initial inventory deadline in 2024 will give systems a head start prior to beginning the
LCRI replacement requirements. In addition, the EPA's inventory guidance (USEPA 2022), BIL
funding towards LSL identification and replacement, as well as new and emerging technologies
to identify service line materials may contribute to significant inventory development for many
systems.

3.4 Additional Opportunities for Inventory Development

In systems with remaining unknown lines to investigate, the EPA notes that routine water system
activities or emergency repairs provide opportunities to identify unknowns in the course of
normal operations. Water meters are generally replaced after 15 to 20 years (City of Pasadena,
n.d.; Stubbart, 2003); although, the EPA has found information suggesting their replacement may
be as frequent as eight years (Vollrath 2015). This information suggests that approximately five
percent (replacing all meters every 20 years = 100 percent/20 years) to 12.5 percent (100
percent/8 years) of a system's meters will be replaced annually, which presents a substantial
opportunity for service line materials to be encountered and inventoried, although materials are
not guaranteed to be encountered during each meter replacement. Replacement of water mains
has been documented between 0.8 percent and two percent per year, presenting additional
opportunities to identify the material of the connected service lines (Folkman, 2018). This rate
could also potentially to increase in coming years, as 82 percent of cast iron water mains are over
50 years old and seeing increasing rates of the need for emergency repairs (Folkman, 2018).
Overall, through routine infrastructure work and emergency-repair activities, water systems
could be expected to encounter up to six percent to eight percent of their service lines each year.
Thus, systems may have the potential to encounter between 60 percent and 80 percent of their
service lines within the 10-year replacement deadline, outside of any activities dedicated solely
to service line material identification.

18


-------
3.5 Determination of the Inventory Validation Pool and Minimum Number of
Validations Required

For the final LCRI, the EPA requires water systems to validate the methods used to categorize
service lines as "non-lead" in their inventories. Water systems must populate a validation pool of
non-lead service lines from their inventories, excluding non-lead service lines categorized
through (1) records showing the line was installed after June 19, 1986, the date their States or
local communities adopted standards or plumbing codes that meet the definition of lead free in
accordance with SDWA section 1417, (2) visual inspection of the pipe exterior at a minimum of
two points, or (3) previously replaced lead or GRR service lines are excluded from the validation
pool.

Water systems will then validate a subset of non-lead service lines from the validation pool
through visual inspection of at least two points of the pipe exterior to reach a 95 percent
confidence level, so the results of this inventory validation are representative of the entire
validation pool. The determination of the "subset" of service lines (or the sample size) for the
validation pool is determined using a similar methodology to the methodology developed by the
Michigan Department of Environment, Great Lakes, and Energy (Michigan EGLE, 2021).
Specifically, this method utilizes a test of proportions to determine the proportion of a population
(in this case, the validation pool size) that meets a given criteria (in this case, that the service line
is validated as non-lead) by sampling only some subset of this population. For example, if a
system visually inspected 100 non-lead lines, and 98 were confirmed to be non-lead, while the
other two lines were found to be lead or GRR service lines, the proportion of lines verified as
non-lead would be 98 out of 100. However, if the entire validation pool of that system consisted
of 10,000 non-lead lines, then only sampling 100 lines is not be enough to achieve a 95 percent
confidence level to ensure that the observed results are representative, and it would be possible
that the proportion of lines validated as non-lead could be different if the system instead
inspected a larger subset (i.e., 300, 1,000, 5,000) of their 10,000 lines. Therefore, a statistical test
is needed to determine the minimum number, out of the 10,000 lines, that the system would need
to validate to achieve a 95 percent confidence level, so the proportion of lines validated as non-
lead in their sample is representative of the entire validation pool.

The following information is included in this technical support document to provide clarification
on the inventory validation calculations, specifically what data was used to determine the
expected sample proportion, what the relevant comparison is between the number of validations
required and the validation pool (and what is significance of that comparison), and where the
agency derived its formulas for determining the appropriate sample size (as well as to provide
citations).

To determine the minimum number of validations required for a certain validation pool size, the
EPA derived formulas from the methodology used to determine the sample size for estimating
means from chapters 5.3, 6.7, and 6.8 in the textbook, "Biostatistics: A Foundation for Analysis
in the Health Sciences, tenth edition" (Daniel and Cross, 2013). Online sample size calculators
(Select Statistical Services, n.d.) and other EPA materials (USEPA, 1997) have used similar
methodologies (Cochran, 1977). The following information walks through the determination of
the sample size for estimating means and proportions from the textbook and how the EPA
simplified and applied those formulas to the inventory validation calculations.

19


-------
Determining the sample size for estimating means (chapter 6.7, Daniel and Cross, 2013)
The EPA utilizes a method for determining the sample size, n, required for estimating a
population mean, which is applied to the case of sample size determination when the parameter
to be estimated is a population proportion.

The objectives in interval estimation are to obtain narrow intervals with high reliability. First, the
width of the interval is determined by looking at the components of a confidence interval:

since the total width of the interval is twice this amount. This quantity is referred to as the
precision of the estimate or the margin of error. For a given standard error, when reliability is
increases, the reliability coefficient becomes larger. However, a larger reliability coefficient for a
fixed standard error makes for a wider interval. On the other hand, by fixing the reliability
coefficient, the only way to reduce the width of the interval is to reduce the standard error. The
standard error is equal to 
-------
Nz2o2
U d2(N — 1) + z2a2

The formulas for sample size require knowledge of o2 (or the population variance). However, the
population variance is, as a rule, unknown. Therefore, o2 must be estimated.

Finite population correction (chapter 5.3, Daniel and Cross, 2013)

As seen above, for sampling without replacement from a finite population, the finite population

(TV-7l)

correction is required. The finite population correction factor is — and can be ignored when
the sample size is small in comparison with the population size. When the population is much
larger than the sample, the difference between a2/n and fr)K] will be negligible. The finite

population correction is not used unless the sample is more than 5% of the size of the population.
That is, the finite population correction is typically ignored when n/N < 0.05.

Determining sample size for estimating proportions (chapter 6.8, Daniel and Cross, 2013)
The method of sample size determination when a population proportion is to be estimated
utilizes the same method that described for estimating a population mean. One-half of the desired
interval, may be set equal to the product of the reliability coefficient and the standard error.

It is assumed that random sampling and conditions permitting approximate normality of the
distribution of p leads to the following formula for n when sampling is with replacement from an
infinite population, or when the sampled population is large enough that the finite population
correction can be ignored:

z2pq

where q = 1 -p, and p is the sample proportion. If the finite population correction cannot be
disregarded, the proper formula for n is as follows:

Nz2pq
U d2(N — 1) + z2pq

Both formulas require knowledge of p. Since this is the parameter that the EPA is trying to
estimate, it will be unknown. One solution is to take a pilot sample and compute an estimate to
be used in place of p in the formula for n. Sometimes, an investigator will have some notion of
an upper bound for p that can be used in the formula. If it is impossible to come up with a better
estimate, setp equal to 0.5, and solve for n. Sincep = 0.5 in the formula yields the maximum
value of //, this procedure will give a large enough sample for the desired reliability and interval
width. However, it may be larger than needed and utilize more resources during sampling than if
a better estimate of p had been available. This procedure is used only if a better estimate of p is
unavailable.

21


-------
Simplifying the determination of the sample size for a finite population when the population
proportion is estimated

Start with the formula for sample size n for sampling without replacement for a finite population
as derived above.

Nz2pq
71 d2(N — 1) + z2pq

Multiply the numerator and denominator by d2.

Nz2pq	d2

71 d2(N — 1) + z2pq X d2

Substituted (denoted as X instead of n to reduce confusion) for a sample size for sampling with
replacement from an infinite population.

NXd2
U d2(N — 1) + z2pq

Rearrange the formula.

d2(N — 1) + z2pq NX
d2	n

Simply the left side of the formula.

d2(JV -
d2

(JV

1) z2pq NX
+ d2 n

NX

1)+X =	

n

Solve for n.

NX

n =	

X + N-l

Where, as mentioned above, Xis the formula for a sample size for sampling with replacement
from an infinite population:

„ _z2pq
d2

22


-------
Connecting the sample size for a finite population when the population proportion is estimated to
the minimum number of validations required

To calculate the minimum number of samples required to achieve a 95 percent confidence level,
the simplified formula for calculating the sample size n for a finite population when the
population proportion is estimated is used:

„	.	NxX

Equation 1.	n = 	

M	X+N-l

Where N is the size of the validation pool, and X is the sample size for an infinite population. X\s
further calculated through Equation 2:

Equation 2.	X = 2^^2^ = 384.16

The EPA assumes sampling follows a normal distribution and specifies the confidence level of
95 percent, which means the value from the standard normal distribution at this confidence (z) is
a constant value of 1.96. The precision of the estimate or margin of error, J, (denoted as "MOE")
is also already specified at 0.05 (5%), and the validation pool size (TV) is calculated based on
summing the total number of non-lead service lines eligible for validation. Therefore, once the
sample pool is established, the only remaining unspecified variable is the estimate of the sample
proportion (p).

In the Michigan EGLE method, the sample proportion is assumed to be 50%, meaning that 50%
of the non-lead service lines investigated are expected to be confirmed as non-lead. In practice,
this proportion will most likely be much higher since most non-lead lines will end up being
confirmed as non-lead. However, as Daniel and Cross (2013) noted, an assumption of 0.50
provides the most conservative (or largest) sample size, meaning that any system validating this
number of lines will be guaranteed to have met or exceeded the minimum number for statistical
significance regardless of the true sample proportion. The EPA assumes a sample proportion of
0.50 because the agency does not have sufficient data to estimate a sample proportion specific to
discovering a non-lead service line as a lead or GRR service line and, therefore, used a
conservative sample proportion to ensure the minimum number of validations required is
statistically significant in all systems nationwide regardless of the possibility for a more precise
sample proportion at an individual system's level. The procedure of using a sample proportion of
0.5 is used when one cannot arrive at a better estimate (Daniel and Cross, 2013). It is also
important to note that, although a sample proportion of 0.5 may overestimate the sample size,
since these equations are based on interval estimation and do not take into account power, some
of the equations used may also underestimate the necessary sample size (Kupper and Hafner,
1989).

Exhibit 3.4: Sample size required for a system with varying validation pool sizes to achieve 95%
confidence in the validation with varying assumptions of the sample proportion.

23


-------
450
400
1 350
I 300
1 250
£ 200

(LI

"5. 150

I 100

50
0

Assumed Sample Proportion

Flexibility for small systems with the minimum number of validations required
Similar to the Michigan EGLE method, both the proposed and final rules provide a flexibility for
systems with fewer than 1,500 non-lead lines in their validation pool to only be required to
validate 20 percent of their eligible non-lead lines rather than the number required for statistical
significance. This provision is designed primarily for small systems for whom a 95 percent
confidence level may not be technically possible given possible staffing limitations and burden
associated with simultaneously complying with other parts of the LCRI as well as the increased
proportion of validations smaller systems would be required to complete compared to larger
systems. For example, at a 95 percent confidence level, a system with a validation pool of 100
lines would be required to validate 80 lines (80 percent of validation pool) to achieve statistical
significance, whereas a system with a validation pool of 10,000 lines would be required to
validate only 370 of them (3.7 percent) (Exhibit 3.4).

Display of the minimum number of validations required in tabular form
To decrease rule complexity and ease compliance tracking, in the final LCRI, the EPA presents
the validation requirements in tabular form rather than as an equation (Exhibit 3.5). In this way,
each individual system will not calculate their specific number of validations; they only need to
establish the size of their validation pool and conduct the corresponding number of validations
specified in the table. The size of increments in the table were chosen to ensure that the number
of validations required would be no greater than 10 away from the number required to achieve a
95 percent confidence level. Thus, as the validation pool increases in size, the additional number
of validations required increases, and larger differences between rows are employed (i.e., the
first row only captures sizes from 1,500 to 2,000, whereas the last rows capture all systems from
10,001 to 50,000 or greater than 50,000).

Exhibit 3.5: Number of validations required based on the validation pool size

Size of Validation Pool

Number of Validations Required

Fewer than 1,500

20 percent of validation pool

N=100
N=1,000
N=10,000
N=100,000

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

24


-------
Size of Validation Pool

Number of Validations Required

1,500-2,000

322

2,001-3,000

341

3,001-4,000

351

4,001-6,000

361

6,001-10,000

371

10,001-50,000

381

More than 50,000

384

4 Supporting Information for the Lead Action Level Analysis

In the final LCRI, the EPA used data from the 6,551 community water systems of all sizes with
known corrosion control treatment (CCT) and lead service line (LSL) status and reported 90th
lead percentile values in EPA's Safe Drinking Water Information System (SDWIS) from 2012-
2020. The EPA used this data to evaluate a range of action level alternatives for the final LCRI
that are generally representative of OCCT (see section IV.F.4 of the final LCRI preamble for
details). The EPA updated the number of systems from 6,529 to 6,551 from the proposed to the
final LCRI. Under the proposed LCRI, 6,529 systems with known CCT and LSL status and
reported 90th percentile values were evaluated but the agency was able to include an additional
22 systems with known LSL status based on additional data through the one-time update to the
7th Drinking Water Infrastructure Needs Survey and Assessment from May 2024.

The LCRI includes new tap sampling requirements that are likely to result in higher 90th
percentile lead levels compared to the LCR. To account for differences in the tap sampling
requirements under the LCR and the LCRI, the lead 90th percentile data was adjusted using a
multiplier approach. Specifically, the EPA adjusted for the requirement for systems with lead
service lines to collect all samples at lead service line sites, and to collect and use the higher lead
concentration of the first- and fifth-liter samples at each lead service line site in the 90th
percentile calculation. The EPA found that these 6,551 systems are representative of water
systems nationally. See sections 3.3.3-3.3.5 of the final LCRI Economic Analysis (USEPA,
2024b) for additional information about how the EPA identified these systems and their CCT and
LSL status, details about the multiplier approach, representativeness of the data nationally, and
the associated uncertainties.

To further inform whether the level of 0.010 mg/L supports the action level's purpose of
addressing the technical feasibility for the CCT treatment technique, in the proposed LCRI the
EPA estimated what percentage of CWSs are likely to exceed various potential action levels
nationally. The EPA conducted this analysis to illustrate the estimated percentages of systems
that would be required to make a detailed OCCT demonstrations, which could require more
effort and resources from systems, and therefore, States (see Exhibit 3, 88 FR 84941, USEPA,
2023b). For the final LCRI, the EPA is updating the analysis presented at proposal and including

25


-------
an additional analysis to clarify how these values were calculated and correspond to the analysis
presented in the final LCRI preamble (see Exhibit 2 in the final LCRI preamble). The EPA used
the 6,551 systems grouped by system size and CCT and LSL status characteristics for these
analyses. The additional analysis and updated data presented here do not support an alternate
conclusion from the LCRI proposal.

To further inform the selection of an action level that is representative of OCCT, the EPA
evaluated the 6,551 systems grouped by CCT and LSL status and estimated the percentage of
systems that would meet each level (see Exhibit 2, section IV.F.4 of the final LCRI preamble).
For the final LCRI, the EPA evaluated the 6,551 systems grouped by LSL and CCT status to
clarify how the estimates of percentages of systems that would be required to conduct CCT
demonstrations correspond to Exhibit 2 in the final LCRI preamble (Exhibit 4.1).

Exhibit 4,1: Percent of CWSs in Each Size Category Estimated to Have 90th
Percentile Lead Levels Exceeding 0 g/L, 0.010 mg/L, and 0.005 mg/L by
System Size Under the Final LCRI





LSL and CCT Status



P901

System Size

No LSL/

LSL /

No LSL/

LSL /

Total



CCT

CCT

No CCT

No CCT

(n=6,551)1





(n=2,062)1

(n=l,277)J

(n=2,73iy

(n=48iy





< 3,300 (40,113 systems)2

4.1%

4.5%

4.3%

8.9%

4.6%



3,301- 10,000 (5,026)2

0.7%

5.5%

0.2%

6.0%

1.8%

0.015
mg/L

10,001 -50,000 (3,374)2

0.0%

9.9%

0.1%

5.4%

2.4%

> 50,000 (1,016)2

0.1%

6.7%

0.0%

0.0%

1.4%



TOTAL (49,529)2

5.0%

26.7%

4.6%

20.4%

10.2%



<3,300

6.0%

5.4%

8.3%

14.8%

7.5%



3,301- 10,000

1.3%

8.5%

0.7%

10.4%

3.1%

0.010
mg/L

10,001-50,000

0.3%

16.4%

0.2%

10.4%

4.2%



> 50,000

0.2%

9.7%

0.0%

0.0%

2.0%



TOTAL

7.9%

40.1%

9.3%

35.6%

16.8%



<3,300

10.8%

7.1%

18.5%

23.3%

14.2%



3,301- 10,000

3.7%

13.2%

2.6%

18.9%

6.2%

0.005
mg/L

10,001-50,000

2.2%

25.1%

1.0%

21.0%

7.5%



> 50,000

1.5%

17.1%

0.0%3

0.0%3

3.8%



TOTAL

18.1%

62.5%

22.2%

63.2%

31.8%

Notes:

1 Data from 6,551 community water systems with known CCT and LSL status used in the analysis. See
"Analysis of reported 90th percentile values from 2012-2020 for final LCRI.xlsx" in EPA-HQ-OW-2022-
0801. Systems categorized based on their highest P90 value reported (SDWIS 2012-2020).

2Total number of CWSs in each size category nationally as reported to SDWIS in fourth quarter 2020. See
USEPA, 2024b, Chapter 3, Exhibit 3-2. The data used in the analysis was determined to represent systems
nationally (see USEPA, 2024b, Chapter 3, section 3.3.5.1.3). The total number of systems in each size category
is included to inform the impact to all systems in each size category.

3Systems serving > 50,000 people without CCT must have 90thpercentile levels below 0.005 mg/L in accordance
with § 141.82(b)(3) under the LCR.

26


-------
Exhibit 4.1 shows the percent of systems in each LSL and CCT status category estimated
to have 90th percentile values (P90) exceeding the evaluated levels of 0.015 mg/L, 0.010 mg/L,
and 0.005 mg/L as adjusted for the final LCRI tap sampling requirements. The LSL and CCT
status headings include the number of systems in each of the categories evaluated. The number
of systems in each size category nationally are also provided for reference to inform the potential
national impact of each of the action levels evaluated (see above discussion on the
representativeness of the data evaluated nationally). The bolded values in the rows labeled
"total" represent the percent of systems in each column that exceed the various P90 values. For
example, 26.7 percent of LSL and CCT systems exceed 0.015 mg/L. This is the inverse of the
percentage of CCT and LSL systems that meet 0.015 mg/L in Exhibit 2 in section IV.F.4 of the
LCRI preamble after rounding (73.3 percent). Exhibit 4.1 shows the percent of systems within a
specific LSL and CCT size category that exceed a P90 and are systems of a given size (e.g., X
percent of systems with CCT and LSLs exceed 0.005 mg/L and are systems serving < 3,300
persons). In other words, within each LSL and CCT status category, the percent of systems that
are expected to exceed the action levels evaluated and belong to a system size category are
shown. Therefore, the percentage of exceedances across system sizes for each P90 level
evaluated adds up to the total exceedance percentage for each CCT and LSL status category (i.e.,
the bolded values). These non-bolded percentages can be used to identify which system size
categories have the greater number of system exceedances in each LSL and CCT status group.
The EPA presented a similar analysis to Exhibit 4.1 in the LCRI proposal using the same data
but based on system size category instead of CCT and LSL status (Exhibit 3, 88 FR 84941,
USEPA, 2023b). The EPA included this analysis at proposal because the discussion of technical
factors related to CCT feasibility are explained based on system size rather than LSL and CCT
status. While the EPA is including Exhibit 4.1 in the final LCRI to clarify interpretation of the
data relative to Exhibit 2 in the final LCRI preamble, the EPA also updated the analysis
presented in the proposed LCRI using the 6,551 systems for completeness and transparency.
Exhibit 4.2 does not offer different conclusions to Exhibit 4.1, but the values differ because the
percentages are based on the number of systems in each system size category rather than LSL
and CCT status group. Exhibit 4.2 shows the percentage of each system size category estimated
to have P90 values exceeding the evaluated levels of 0.015 mg/L, 0.010 mg/L, and 0.005 mg/L
by LSL and CCT status as adjusted for the final LCRI tap sampling requirements. Similar to
Exhibit 4.1, the number of systems evaluated in each size category are included along with the
number of systems in each size category nationally for reference. Each bolded row shows the
percent of systems in each size category that exceed each P90. Within each size category, the
percent of systems that are estimated to exceed the action levels evaluated and have a specific
CCT and LSL status are shown. As in Exhibit 4.1, the EPA is clarifying that the non-bolded
values do not represent the percentage of the subset of systems exceeding each P90 value by
CCT and LSL status (e.g., X percent of the systems serving < 3,300 persons that exceed 0.005
mg/L are CCT and LSL systems, but rather X percent of systems serving <3,300 persons exceed
0.005 mg/L and are CCT and LSL systems). Therefore, the percentage of exceedances across

27


-------
each LSL and CCT status group for each P90 level evaluated adds up to the total exceedance
percentage for each system size category (i.e., the bolded values).

Exhibit 4.2: Percent of CWSs in Each Size Category Estimated to Have 90th
Percentile Lead Levels Exceeding 0 g/L, 0.010 mg/L, and 0.005 mg/L by
LSL and CCT Status Under the Final LCRI





System Size



P90

LSL and CCT
Status

< 3,300
11=3,798!
(40,113
systems)2

3,301-
10,000
0=1,121!

(5,026
systems)2

10,001-
50,000
n= 1,099!

(3,374
systems)2

> 50,000
11=533!
(1,016
systems)2

Total
n=6,551i
(49,529
systems)2



No LSL/No CCT

3.1%

0.5%

0.2%

0.0%

1.9%



No LSL/CCT

2.2%

1.3%

0.1%

0.6%

1.6%

0.015
mg/L

LSL/No CCT

1.1%

2.6%

2.4%

0.0%

1.5%

LSL/CCT

1.5%

6.2%

11.6%

16.1%

5.2%



TOTAL

8.0%

10.7%

14.2%

16.7%

10.2%



No LSL/No CCT

6.0%

1.7%

0.5%

0.0%

3.9%



No LSL/CCT

3.3%

2.4%

0.6%

0.9%

2.5%

0.010
mg/L

LSL/No CCT

1.9%

4.5%

4.5%

0.0%

2.6%

LSL/CCT

1.8%

9.7%

19.1%

23.3%

7.8%



TOTAL

13.0%

18.3%

24.8%

24.2%

16.8%



No LSL/No CCT

13.3%

6.4%

2.5%

0.0%3

9.2%



No LSL/CCT

5.9%

6.8%

4.1%

5.6%

5.7%

0.005
mg/L

LSL/No CCT

2.9%

8.1%

9.2%

0.0%3

4.6%

LSL/CCT

2.4%

15.1%

29.1%

40.9%

12.2%



TOTAL

24.5%

36.4%

44.9%

46.5%

31.8%

Notes:

1	Data from 6,551 community water systems with known CCT and LSL status used in the analysis. See
"Analysis of reported 90th percentile values from 2012-2020 for final LCRI.xlsx" in EPA-HQ-OW-2022-
0801. Systems categorized based on their highest P90 value reported (SDWIS 2012-2020).

2	Total number of CWSs in each size category nationally as reported to SDWIS in fourth quarter 2020. See
USEPA, 2024b, Chapter 3, Exhibit 3-2. The data used in the analysis was determined to represent systems
nationally (see USEPA, 2024b, Chapter 3, section 3.3.5.1.3). The total number of systems in each size
category is included to inform the impact to all systems in each size category.

3	Systems serving > 50,000 people without CCT must have 90thpercentile levels below 0.005 mg/L in
accordance with § 141.82(b)(3) under the LCR.

Appendices

Appendix 1.0: List of references for service line replacement data from Exhibit 2.1

28


-------
City

State

Citation in References

Cleveland

OH

City of Cleveland LCRI Public Comment

Denver

CO

Denver Water, 2023a

Fort Worth

TX

Fort Worth, n.d.

Louisville

KY

Louisville Water Company LCRI Public Comment

Cincinnati

OH

Greater Cincinnati Water Works, 2023

Detroit

Ml

City of Detroit, n.d.

Tucson

AZ

City of Tucson, 2019

Washington

DC

DC Water, 2023

Pittsburgh

PA

Pittsburgh Water and Sewer Authority, n.d.

Central Arkansas
Water

AR

Sweeney, 2020

Saskatoon

Canada

City of Saskatoon, n.d.

Newark

NJ

City of Newark, n.d.

Grand Rapids

Ml

City of Grand Rapids, 2022

Spokane

WA

Feist, 2018

Trenton

NJ

Trenton Water Works, n.d.

Aurora

IL

IEPA, 2022

Sioux Falls

SD

Kelley, 2017

York

PA

The York Water Company, 2023

Kalamazoo

Ml

City of Kalamazoo, 2023

Lansing

Ml

Lansing Board of Water and Light, n.d.

Lancaster

PA

City of Lancaster LCRI Public Comment

Elgin

IL

IEPA, 2022

Green Bay

Wl

Green Bay Water, n.d.

Quincy

MA

MWRA, 2023

Flint

Ml

City of Flint, n.d.

Newton

MA

MWRA, 2023

Somerville

MA

City of Somerville, n.d.

Evanston

IL

IEPA, 2022

Framingham

MA

MWRA, 2023

Madison

Wl

City of Madison, 2014

St. Clair Shores

Ml

City of St. Clair Shores, n.d.

Revere

MA

City of Revere, 2023

Bozeman

MT

City of Bozeman, 2020

NA

VT

Vermont DWGPD LCRI Public Comment

Bloomfield

NJ

Bloomfield Water Department, 2021

Battle Creek

Ml

Battle Creek, 2023

Marlborough

MA

MWRA, 2023

Galesburg

IL

IEPA, 2022

Village of
Montgomery

IL

Village of Montgomery, n.d.

Norwood

MA

MWRA, 2023

Sandusky

OH

City of Sandusky, 2021

29


-------
City

State

Citation in References

Winchester

MA

MWRA, 2023

Birmingham

Ml

City of Birmingham, n.d.

Frankfort

IL

IEPA, 2022

Menasha

Wl

Menasha Utilities, 2023

Stoughton

Wl

City of Stoughton Utilities Committee, 2022

Mayville

Wl

City of Stoughton Utilities Committee, 2022

Evart

Ml

City of Evart, 2020

Appendix 2.0: Examples of types of documents available to provide additional guidance to
water systems beginning or implementing their service line replacement programs

•	LSLR Collaborative-Established in 2017 following the Flint Water Crisis, the
Collaborative is a joint effort of 28 national public health, water utility, environmental,
labor, consumer, housing, and State and local government organizations to accelerate full
removal of lead pipes. The Collaborative has provided a roadmap for starting a service
line replacement program, an ongoing webinar series focused on service line
replacement, and links to additional related resources. The Collaborative also notes that
service line replacement practices "may evolve with new information" and that they will
continually update resources to reflect any changes in technology or processes over time.
(LSLR Collaborative, n.d.).

•	Principles for Lead Service Line Replacements-Released in 2022, NRDC and colleagues
proposed a set of basic principles for replacing LSLs based upon their experience with
efforts to replace pipes from Flint and Benton Harbor, MI, Newark, NJ, Pittsburgh, PA,
Washington, DC, and many other locations. The principles cover varying aspects of a
LSLR program, from community involvement to consent of property owners, to the
approach and methods used for replacement. (Alliance of Nurses for Healthy
Environments et. al., n.d.).

•	Denver Water Lead Reduction Program-Since the first year of Denver Water's
replacement program in 2021, Denver has released semi-annual reports detailing the
progress of the replacement program, inventory updates, lead sampling, and distribution
and sampling of water filters. These reports include a section "Learning by Doing" in
which the system reflects and shares lessons learned the previous year including
adjustments to their procedures or policies to increase efficiency and effectiveness of
their program, (e.g. Denver Water, 2023b).

•	Central Arkansas Water's Lead Service Line Replacement Program: Investigation,
Communication, Implementation-Published in April 2020, Central Arkansas Water
shares a service line replacement protocol developed in 2016 and details the success of
the protocol and challenges faced in their replacement program after conducting
approximately 100 replacements in October 2017. (Sweeney, 2020).

•	Halifax Water's Lead Service Line Replacement Program Gets the Lead Out-Published
March 2022, Halifax Water details their process for service line replacement, including a
research partnership with nearby Dalhousie University and insights resulting from this
partnership. This paper details customer outreach and funding strategies and the evolution
of the LSLR program in recent years. (Krkosek et. al., 2022).

30


-------
•	Safety and Affordability are Vital for LSLR-Published in August 2020, an Ohio water
utility describes their process to make service line replacement affordable for its
customers and safe for field crews. (Niranjan and McCarthy, 2020).

•	Lead in Drinking Water: A Permanent Solution for New Jersey-Published in October
2019, Jersey Water Works convened a 30-member task force of representatives from
local, State, and Federal governments, water utilities, academia, environmental, smart
growth and community advocates, and public health organizations to determine practical,
cost-effective, equitable, and permanent solutions to virtually eliminate lead in drinking
water within 10 years. (Jersey Water Works, 2019).

•	Identifying Information Gaps to Help Communities Navigate Lead Service Line
Replacement-Published in June 2024, The Joyce Foundation and the Federal Reserve
Bank of Chicago jointly convened 25 regional and national leaders in LSLR and
published a summary of key information and research gaps, including LSLR cost drivers,
financing needs, and how to address common equity issues. (Anderson et. al., 2024).

Appendix 3.0: Examples of types of documents available to provide additional guidance to
water systems beginning or implementing their service line identification programs

•	Planning and Developing a Service Line Inventory-EPA has released comprehensive
inventory guidance along with a LSL identification webpage housing all EPA guidance,
case studies, templates, fact sheets, and webinars related to service line identification.
(USEPA, 2022).

•	Developing and Maintaining a Service Line Inventory: Small Entity Compliance Guide-
EPA released the Small Entity Compliance Guide in 2023 to support water systems,
particularly small water systems, in complying with the LCRR initial inventory
requirements. This guide explains the inventory-related actions small community and
non-transient non-community water systems are required to take under LCRR. (USEPA,
2023c).

•	Lead service line identification: A review of strategies and approaches-Published in June
2021, scientists at the EPA's ORD summarized the current industry LSL identification
methods, including records screening, basic visual examination of indoor plumbing,
water sampling, excavation, and predictive data analyses. This analysis was used to
develop a stepwise approach to identify unknown service lines using a combination of
different available methods. (Hensley et. al., 2021).

•	Service Line Material Identification: Experiences from North American Water Systems-
Published in January 2022, an AWWA subcommittee interviewed 10 water systems to
learn about their processes for LSL inventory creation, material identification, customer
communication, and other aspects of their experiences. In this way they have compiled
lessons learned from various utilities into one article to peruse to aid in preparing water
utilities with what to expect in their own identification efforts. (Ligget et. al., 2022)

•	Predictive Modeling in Service Line Inventory Development-Published in December
2023, this article documents States which permit the use of predictive modeling and
discusses in detail various uses and applications of predictive modeling to the
development of a service line inventory. (Deheer et. al., 2023).

•	Identifying Lead Service Lines with Field Tap Water Sampling-Published in 2021,
researchers at the University of Pittsburgh worked with the PWSA to analyze historical
tap sampling data by utilizing a random forest statistical model. This study gauged the

31


-------
accuracy of the random forest model and considered sampling strategies which could lead
to even more accurate models. (Blackhurst, M., 2021).

•	Developing and Verifying a Water Service Line Inventory-Published in April 2021,
Pittsburgh Sewer and Water Authority (PWSA) discusses how they learned from the
efforts of similar utilities to inform their development of a service line inventory, even
when data and records have not been consistently collected and maintained in the past.
PWSA documented their inventory process and which inventory method were used, and
overall lessons learned to assist future utilities. (Duffy and Pickering, 2021).

•	Planning for Service Line Material Identification and Lead Service Line Replacement
Costs-Published in 2023, the study team from CDM Smith sent a survey to more than
4,500 utilities to obtain information about what service line identification methods are
being used and various associated costs of them at utilities nationwide. The summary of
costs provides utilities with realistic estimates for what to expect in terms of the cost of
their service line identification and replacement programs to aid in long-term planning
and financial assessments. (Kutzing et. al., 2023).

o Full associated report from CDM Smith is also publicly available

•	Development and optimization of a systematic approach to identifying lead service lines:
One community's success-Published in 2023, Engineers working with Bennington, VT
worked with EPA scientists to document the accuracy of various service line
identification methods during their development of a service line inventory. Service line
identification strategies arising from this information were developed and mapped for
future water systems (Smart et. al., 2023).

References

Alkafaji, R. (2023, June 29). Lead pipe survey flaws reveal that states may not be getting their
fair share. Environmental Defense Fund. Retrieved July 31, 2023 from

https://bloes.edf.ore/deepdives/2023/06/29/lead-pipe-siirvey-flaws/

Alliance of Nurses for Healthy Environments, Beyond Plastics, Clean Water Action,
Earthjustice, Environmental Advocates NY, Environmental Policy Innovation Center,
Environmental Protection Network, Gulf Coast Center for Law & Policy, Healthy Babies Bright
Futures, League of Conservation Voters, Little Village Environmental Justice Organization, MI
Welfare Rights Organization, Natural Resources Defense Council, New Jersey Future, Newark
Education Voters Caucus, New York League of Conservation Voters, People's Water Board
Coalition, Waterway Advocates, WE ACT for Environmental Justice, (n.d.). Principles for Lead
Service Line Replacements. Retrieved July 17, 2024 from

https://www.nrdc.org/sites/default/files/principles-for-lead-service-line-replacements-
20220228.pdf

American Association for Public Opinion Research (AAPOR), 2023. Standard Definitions: Final
Dispositions of Case Codes and Outcome Rates for Surveys. Retrieved 9-9-2024 from
https://aapor.org/wp-content/uploads/2023/05/Standards-Definitions-10th-edition.pdf

32


-------
Anderson, N., Castro, M., Cisar, E., Cuenca, K., Saxena, S. (2024, June). Identifying Gaps to
Help Communities Navigate Lead Service Line Replacement. Federal Reserve Bank of Chicago-
Chicago Fed Insights. Retrieved July 17, 2024 from

https://www.chicagofed.org/publications/chicago-fed-insights/2024/lead-service-line-
repl acement-inform ati on-gap s

Battle Creek, Michigan (2023, Mar. 29). Works to replace existing lead service lines throughout
city resumes. Public Works/Road Work News Flash. Retrieved August 9, 2023, from

https ://www.battlecreekmi. gov/Ci vie Alerts, aspx? A.ID= 1698

Blackhurst, M. (2021, Aug. 4). Identifying Service Lines with Field Tap Water Sampling. ACS
ES&T Water. Vol. 1, Issue 8, pp 1983-1991. Retrieved July 17, 2024 from

http s: //pub s.acs. ore/doi/ ful	acsestwater. 1 c00227#

Bloomfield Water Department (2021). Important information about lead in your drinking water.
Retrieved August 9, 2023 from

http://bloomfieldtwpni.com/DociimentCenter/View/7734/Bloomfield-Lead-Update—February -
2021-PDF

California Legislature (2017). Health and Safety Code, Division 104-Environmental Health, Part
12-Drinking Water, Chapter 5-Water Equipment and Control. Retrieved July 31, 2023 from

https://leginfo.legislature.ca.gov/faces/codes displaySection.xhtml?lawCode=HSC§ionNum
'885.

City of Birmingham (n.d.). Turn on the Tap-Residential Water Testing. Water and Sewer
Department. Retrieved August 9, 2023, from

https://www.bhamgov.org/about_birmingham/city_departments/public_services/water	sewer/1

eadtesting.php

City of Bozeman (2020). 2019 Water Quality Report. Bozeman Water Treatment Plant.

Retrieved August 9, 2023, from

https://www.bozeman.net/home/showpiiblisheddocument/9965/637169227578900000

City of Chicopee (n.d.). Lead Service Line Inventory. City of Chicopee Water Department.
Retrieved August 19, 2023 from http://chicopeema.gov/1031/Lead-Service-Line-Inventory

City of Detroit (n.d.). Capital Improvement Program. Detroit Water and Sewerage Department.
Retrieved August 9, 2023, from https://detroitmi.gov/departments/water-and-sewerage-
department/dwsd-projects/capital-improvement-program

City of Evart (2020, Mar. 31). City of Evart Lead & Copper Program. Retrieved August 9, 2023,
from https://storymaps.arcgis.com/stories/2be961871eb444629a87e8f7d70b6d95 28

33


-------
City of Flint, (n.d.). Progress Report on Flint Water. Michigan. Retrieved July 25, 2023, from
https://www.cityofflint.com/progress-report-on-flint-water/

City of Grand Rapids (2022, Nov.). Private Lead Service Line Replacements. Water System.
Retrieved August 9, 2023, from

https://www.grandrapidsmi.gov/Government/Departments/Water-System/Lead-in-Drinking-
Water/Private-Lead-Service-Line-Replacements#section-4

City of Kalamazoo (2023). City of Kalamazoo Lead Service Line Replacement Project.

Retrieved August 9, 2023, from

https://cityofkalamazoo.maps.arcgis.com/apps/dashboards/765de573218b406d8a9062d67286e0e
3

City of Madison (2014, Nov. 7). EPA seeks details of Madison's Lead Service Line Replacement
Program. Madison Water Utility. Retrieved August 9, 2023, from

https://www.cityofmadison.com/water/insidemwu/epa-seeks-details-of-madisons-lead-service-
replacement-program

City of Newark, New Jersey (n.d.). Lead Service Line Replacement Program.
NewarkLeadServiceLine.com Retrieved July 25, 2023, from
https://www.newarkleadserviceline.com/replacement

City of Pasadena, Texas (n.d.). Frequently Asked Questions-Water Meters. Water Services
Department. Retrieved August 1, 2023 from

https://www.pasadenatx.eov/DociimentCenter/View/2093/FreqiientlY-Asked-Qiiestions—Water-
Meters-PDF #: ~: text=m eter s%20beim g%20r eplaced1

.The%20tvpical%20useful%201ife%20for%20a%20water%20meter%20is%z ore%20prone
%20to%20maintenance%20issues

City of Revere (2023). Water and Sewer Department. Retrieved August 9, 2023, from
https://www.revere.org/departments/water-and-sewer

City of Sandusky (2021). 2021 Annual Report. Retrieved August 9, 2023, from

https://cms6.revize.com/revize/sanduskyoh/docs/City%20of%20Sandusky%20Annual%20Repor

t%202021.pdf

City of Saskatoon (n.d.). Lead Water Pipes. Saskatoon Power, Water, & Sewer. Retrieved
August 9, 2023, from https://www.saskatoon.ca/services-residents/power-water-sewer/drinking-
water/lead-water-pipes

City of Somerville (n.d.). Lead Service Line Replacement Program. Engineering Department.
Retrieved August 9, 2023, from https://www.somervillema.gov/departments/programs/lead-
service-line-replacement-program

34


-------
City of St. Clair Shores (n.d.). SCS Lead Safe. Water Maintenance Division. Retrieved August 9,
2023, from https://www.scsmi.net/892/SCS-Lead-Safe#:~:text=about%20this%20issue%3F-
,The%20City%20of%20St.,and%20send%20to%20the%20DPW 29

City of Stoughton Utilities Committee (2022, Aug. 15). Official Notice and Agenda. Meeting
Agendas. Retrieved August 9, 2023, from

https://www.stoughtonutilities.com/documents/ucpackets/2022-08-15-uc-packet_r.pdf

City of Tucson (2019, Jan. 18). Lead and Copper. Water Quality Monitoring. Retrieved August
9, 2023, from https://www.tucsonaz.gov/Departments/Water/About-Your-Water-Quality/Water-
Quality-Monitoring/Lead-and-Copper

Cochran, W. G. (1977). Sampling techniques. John Wiley & Sons.

Daniel, W. W., and Cross, C. L. (2013). Biostatistics: a foundation for analysis in the health
sciences. Tenth edition. John Wiley & Sons.

DC Water (2023, June 30). DC Water Releases Lead Free DC Plan Update and New Lead
Inventory Map. News. Retrieved August 9, 2023, from https://www.dcwater.com/whats-going-
on/news/dc-water-releases-lead-free-dc-plan-update-and-new-lead-inventory-map

Deheer, K., Epstein, K., Cummings, J. (2023, December, 6). Predictive Modeling in Service Line
Inventory Development. Journal AWWA. Vol. 115, Issue 10, pp 8-18. Retrieved July 17, 2024
from

https://awwa.onlinelibrary.wiley.com/doi/fult/10.1002/awwa.2193

Denver Water (2023a). Annual Report-2022. Denver Water Lead Reduction Program. Version
1.0: February 10, 2023. Retrieved July 25, 2023, from:

https://www.denvenyater.ore/sites/defaiilt/files/lead-rediiction-proeram-anniial-report-2022.pdf

Denver Water (2023b). Denver Water Lead Reduction Program-Learning by Doing-2022.
Version 1.0: February 10, 2023. Retrieved July 19, 2024 from:

https://www.denverwater.org/sites/default/files/lead-reduction-program-learning-by-doing-
2022.pdf

Duffy, D., Pickering, W. (2021, April 1). Developing and Verifying a Water Service Line
Inventory. Journal AWWA. Vol. 113, Issue 3, pp 6-13. Retrieved July 17, 2024 from

https://awwa.onlinelibrary.wiley.com/doi/10.1002/awwa. 1686

Feist, M. (2018, July 27). City Eliminates Remaining Lead Service Lines. Spokane City News.
Retrieved August 9, 2023, from https://my.spokanecity.ore/news/releases/2018/07/27/city-

eliminates-remainine-lead-service-lines/

Folkman, S. (2018). Water Main Break Rates in the USA and Canada: A Comprehensive Study.
Utah State University-Buried Structures Laboratory. Retrieved July, 25, 2023, from:

35


-------
https://digitalcommons.iisii.edu/cgi/viewcontent.cgi7article fccontext=mae facpub&source
=post page

Fort Worth (n.d.). Fort Worth's action plan. Water Department. Retrieved August 9, 2023, from
https://www.fortworthtexas.gov/departments/water/lead/protection

Garner, D (2022, May 17). Leetonia to receive portion of state grants to update water lines.
Morning Journal. Retrieved August 19, 2024 from

https://www.morningjournalnews.com/news/local-news/2022/05/leetonia-to-receive-portion-of-
state-grants-to-update-lead-water-lines/

Greater Cincinnati Water Works (2023). See Progress in your neighborhood. Retrieved August
9, 2023, from https://la.mygcww.org/neighborhoods/

Green Bay Water (n.d.). Fresh drinking water from Lake Michigan. Retrieved August 9, 2023,
from https://www.gbwater.org/services#gsc.tab=0

Hensley, K., Bosscher, V., Triantafyllidou, S., Lytle, D. (2021, June 10). Lead service line
identification: A review of strategies and approaches. AWWA Water Science. Vol. 3, Issue 3, pp
el226. Retrieved July 17, 2024 from

https ://awwa. onlinelibrary .wiley. com/doi/full/10.1002/aws2.1226

Illinois Environmental Protection Agency (IEPA) (2022). Lead Service Line Information. Public
Water Users. Retrieved June 24, 2024 from https://epa.illinois.gov/topics/drinking-water/public-
water-users/lead-service-line-information.html

Illinois General Assembly (2021). HB3739 Lead Service Line Replacement. 102nd General
Assembly. Retrieved July 31, 2023, from

https://ilga.gov/legislation/BillStatus. asp?DocNum=3739&GAID=16&DocTypeID=HB&LegID
=132788&SessionID=110&SpecSess=&Session=&GA=102

Illinois Municipal Code (2017). Public Act 099-0922. Retrieved July 31, 2023 from
https://www.ilga.gov/legislation/publicacts/99/099-0922.htm

Jersey Water Works. (2019, October). Lead in Drinking Water: A Permanent Solution for New
Jersey. Report of the Jersey Water Works Lead in Drinking Water Task Force. Retrieved July 17,
2024 from

https://cms.iersevwaterworks.org/wp-content/uploads/2019/10/JWW-Lead-Report.pdf

Kelley, C. (2017, Aug. 2022). Sioux Falls Replaces Final Lead Water Pipe. Retrieved August 9,
2023, from https://www.keloland.com/news/sioux-falls-replaces-final-lead-water-pipe/

36


-------
Krkosek, W., Healey, M., Sampson, C., and McKnight, A. (2022). Halifax's Lead Service Line
Replacement Programs Gets the Lead Out. Journal AWWA. Vol 114, Issue 2, Pg 10-19.
Retrieved July 17, 2024 from https://awwa.onlinelibrary.wiley.com/doi/full/10.1002/awwa.1862
Kupper, L. L., and Hafner, K. B. (1989). How Appropriate Are Popular Sample Size Formulas?
The American Statistician, 43(2), 101-105. https://doi.org/10.2307/26

Kutzing, S., Murray, N., Sayre, A., Gawlik, E. (2023, July 5). Planning for Service Line Material
Identification and Lead Service Line Replacement Costs. Journal AWWA. Vol. 115, Issue 6, p
22-33. Retrieved July 17, 2024 from

https://awwa.onlinelibrarY.wileY.com/doi/fult/10.1002/awwa.2123

Lansing Board of Water and Light (n.d.). Lead Information. Water Resource Center. Retrieved
August 9, 2023, from https://www.lbwl.com/customers/water-resource-center/lead-information

Leischner, M. (2022, May 31). City of Mosinee Receives Funding for Lead Service Line
Replacement. WSAU Radio. Retrieved AuguM l JO J I horn https://wsau.eom/2022/05/31/city-
of-mosinee-receives-funding-for-lead-service-line-replacement/

Ligget, J., Baribeau, H., Deshommes, E., Lytle, D., Masters, S., Muylwyk, Q., Triantafyllidou, S.
(2022, Jan. 18). Service Line Material Identification: Experiences from North American Water
Systems. Journal AWWA. Vol. 114, Issue 1, pp 8-19. Retrieved July 17, 2024 from

https://awwa.onlinelibrary.wiley.com/doi/fuM/10.1002/awwa. 1841

LSLR Collaborative (n.d.). About the Lead Service Line Replacement Collaborative. Retrieved
July 17, 2024 from https://www.lslr-collaborative.org/about-us.html

Massachusetts Water Resources Authority (MWRA). (2023, June 16). Lead Service Line
Replacement Loan Program or Lead Loan Program (LLP), an Enhancement to the Local Water
System Assistance Program (LWSAP) for Member Communities. MWRA Online. Retrieved
July 17, 2023, from https://www.mwra.com/comsupport/llp/llpprogram.html

Menasha Utilities (2023, May 2). Menasha Completes Replacement of all Lead Water Laterals.
Retrieved August 9, 2023 from

https://www.menashaiitilities.com/sites/menashautilities.com/files/Menasha%20Lead%20Lateral
%20Replacement%20Complete.pdf

Michigan Department of Environment, Great Lakes, and Energy (2020). Michigan Service Line
Materials Estimates-Preliminary Distribution System Materials Inventories. Retrieved August 1,
2023 from https://www.michigan.gov/egle/-

/media/Proj ect/W eb sites/egle/Documents/Program s/DWEHD/Community-W ater-
Supply/Sampling/pdsmi-summary.pdf

Michigan Department of Environment, Great Lakes, and Energy (2021). Minimum Service
LineMaterial Verification Requirements. EGLE Drinking Water and Environmental Health
Division. Retrieved August 1, 2023 from https://www.michigan.gov/egle/-

37


-------
/media/Proiect/Websites/egle/Documents/Programs/DWEHD/Communitv-Water-Supply/Lead-

Copper/Minimum-Service-Line-Material-Verification-Requirements.pdf

Niranjan, D., McCarthy, C. (2020, August). Safety and Affordability Are Vital for Lead Service
Line Replacement. Opflow. Vol. 46, Issue 8, pp. 10-13. Retrieved July 17, 2024 from
https://ui.adsabs.harvard.edu/abs/2020Opflo..46h..l0D/abstract

Pittsburgh Water and Sewer Authority (n.d.). Community Lead Response. Retrieved August 9,
2023, from https://lead.peh2o.com/

Public Service Commission of Wisconsin (2021). 2020 Number of Service Lines by Material
Type for Wisconsin Utilities. Lead Service Line (LSL) Replacement Line. Retrieved August 1,
2023 from https://psc.wi.eov/Paees/ServiceTYpe/Water/LeadServiceLine.aspx

Rhodes, B. (2023, Jan. 25). Cheshire to have Special Town Meeting March 6. iBerkshires.com.
Retrieved August 19, 2024 from https://www.iberkshires.com/story/70702/Cheshire4o-have-
Spec ;n -Meetin e-March-6.html

Schulwitz, S. (2022, April 25). Nearly 300 residents to have water lines inspected for
contaminants. The Alpena News. Retrieved August 19, 2024 from

https://www.thealpenanews.com/news/local~news/2022/Q4/nearly~3QQ~residents-to~have~water~
lines~inspected~for~contaminants/

Select Statistical Services, (n.d.) Population Proportion - Sample Size. Accessed July 2, 2024,
from https://select~statistics.co.uk/calculators/sample~size~calculator~population~proport.ion/

Smart, P., McRae, L., Formal, C., Lytle, D. (2023, Nov. 1). Development and optimization of a
systematic approach to identifying lead service lines: One community's success. Water
Research. Vol. 246. Retrieved July 17, 2024 from

https://www.sciencedirect.com/science/article/pii/S004313542

Stubbart, J., (2003). When Should Meters Be Replaced? American Water Works Association-
Opflow Question of the Month. Retrieved August 1, 2023 from

https://awwa.onlinelibrary.wiley.com/doi/eP'\ l l * J00 '> ihO \ i \

Sweeney, S. (2020, April 6). Central Arkansas Water's Lead Service Line Replacement
Program: Investigation, Communication, Implementation. Journal AWWA. 112(4), 32-38. Doi:
https://doi.org/l 0.1002/awwa. 1479

Trenton Water Works (n.d.) TWW LSLR Program Progress Metrics. TWW Lead Service Line
Replacement Program. Retrieved August 9, 2023, from
https://www.twwleadprogram.com/progress-3

38


-------
USEPA. (1997). Techniques for tracking, evaluating, and reporting the implementation of
nonpoint source control measures: Forestry. Office of Water. EPA 841-B-97-009.

https://www.epa.eov/sYStem/files/dociiments/20	;king Full%20Doc.pdf

USEPA. (1998). National Primary Drinking Water Regulations: Consumer Confidence Reports;
Final rule. Federal Register	August 19, 1998.

USEPA. (2000). Federal Register. Vol. 65. No. 87. National Primary Drinking Water Regulations: Public
Notification Rule; Final Rule. (Thursday. May 4, 2000), 25982-26049 (65 FR 25982).

USEPA. (2022). Guidance for Developing and Maintaining a Service Line Inventory. EPA
Headquarters. Office of Water. Document No. EPA 816-B-22-001. Retrieved July 17, 2024,
from https://www.epa.gov/svstem/files/documents/2022-

08/Inventory%20Guidance A.ugust%202022 508%20compliant.pdf

USEPA. (2023a). Drinking Water Infrastructure Needs Survey and Assessment-7th Report to
Congress. Retrieved July 17, 2024, from
https://www.epa.gov/svstem/files/documents/2023-
09/Seventh%20DWINSA September2023 Final.pdf

USEPA. (2023b). National Primary Drinking Water Regulations for Lead and Copper:
Improvements (LCRI). Proposed Lead and Copper Rule Improvements. Federal Register. 88 FR
84878. December 6, 2023.

USEPA. (2023c). Developing and Maintaining a Service Line Inventory: Small Entity
Compliance Guide. Retrieved July 17, 2024, from

https://www.epa.gov/ground-water-and-drinking-water/planning-and-developing-service-line-
inventory

USEPA. (2023d). Economic Analysis for the Proposed Lead and Copper Rule Improvements.
EPA 815-R-23-005.

USEPA. (2024a). EPA's 7th Drinking Water Infrastructure Needs Survey and Assessment.

https://www.epa.gov/dwsrf/epas-7th-drinking-water-infrastructiire-needs-siirvey-and-assessment

USEPA. (2024b). Economic Analysis for the Final Lead and Copper Rule Improvements.
USEPA. (2024c). Final Lead and Copper Rule Improvements.

Vasko, M. (2022, March 28). A Resolution authorizing the submission and support of an
application to State of Ohio Environmental Protection Agency and declaring an emergency.
Council of the Village of Swanton. Retrieved August 19, 2024 from

https://villageofswantonohio.iis/wp-content/iiploads/2022/04/VILSWA.N-2022-09-03282022.pdf

39


-------
Village of Montgomery (n.d.). Lead-Free. Public Works Department. Retrieved August 9, 2023.
from https://www.monteomervil.ore/711 /Lead-Free

Vollrath, M., (2015). Water Utilities Change AMR/AMI Systems More Frequently than
Previously Thought. Badger Meter. Retrieved August 1, 2023 from

https://www.pnveb.com/releases/2015/05/pnyebl2748042.htm

Voltzer, H., Grim, J., George, A. (2016). Sub. H.B. 512. 131st General Assembly. Ohio
Legislative Service Commission. Retrieved July 31, 2023 from

https://www.leeislatiire.ohio.eov/download7key fcformat=pdf

Walker, E., Jacquette-Morrison, A. (2019). State-Driven Lead Service Line Inventories. 120
Water. Retrieved July 31, 2023 from https://!20water.com/wp-

content/uploads/2020/10/Whitepaper-State-Driven-Lead-Service-Line-Inventories-l-l.pdf

The York Water Company (2023). Lead Information. Retrieved August 9, 2023. from

https://www.Yorkwater.com/lead-information/

40


-------