A EPA
United States
Environmental Protection
Agency
Development of Estimated Quantitation Levels
for the Third Six-Year Review of
National Primary Drinking Water Regulations
(Chemical Phase Rules)

-------
Office of Water (4607M)
EPA 810-R-16-002
October 2016
www.epa.gov/safewater

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Table of Contents
Executive Summary	ES-1
1	Introduction	1-1
1.1	Background	1-1
1.2	Estimated Quantitation Level Development	1-2
1.3	Contaminants	1-3
2	Data Sources	2-1
2.1	MRL Data	2-1
2.2	MDL Data	2-2
3	Threshold Development Method	3-1
4	Development of Individual EQLs	4-1
4.1	MCL Currently Limited by PQL	4-1
4.1.1	Benzo[a]pyrene	4-1
4.1.2	Chlordane	4-3
4.1.3	l,2-Dibromo-3-chloropropane (DBCP)	4-5
4.1.4	Di(2-ethylhexyl)phthalate (DEHP)	4-6
4.1.5	Ethylene Dibromide (EDB)	4-8
4.1.6	Heptachlor	4-9
4.1.7	Heptachlor Epoxide	4-11
4.1.8	Hexachlorobenzene	4-12
4.1.9	Pentachlorophenol	4-14
4.1.10	Polychlorinated Biphenyls (PCBs)	4-15
4.1.11	2,3,7,8-Tetrachlorodibenzo-p-Dioxin (Dioxin)	4-17
4.1.12	Thallium	4-18
4.1.13	Toxaphene	4-20
4.1.14	1,1,2-Trichloroethane	4-21
4.2	MCL Currently Limited by MCLG	4-23
4.2.1	Carbofuran	4-24
4.2.2	Cyanide	4-25
4.2.3	Endothall	4-27
4.2.4	Methoxychlor	4-28
4.2.5	Oxamyl	4-30
4.2.6	Styrene	4-31
5	Summary	5-1
5.1	MCL Currently Limited by PQL	5-1
5.2	MCL Greater than Possible Lower MCLG	5-2
6	References	6-1

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Table of Exhibits
Exhibit ES-1. EQL Threshold Results	2
Exhibit ES-2. Occurrence Threshold Results	3
Exhibit 1-1. Contaminants Where MCLs Limited by Analytical Feasibility	1-3
Exhibit 1-2. Contaminants Where MCLs are Greater than Possible Lower MCLGs	1-4
Exhibit 3-1. EQL Development Steps	3-2
Exhibit 4-1. Occurrence Analysis Threshold Selection Scenarios	4-1
Exhibit 4-2. Summary of MRL Data for Benzo[a]pyrene	4-2
Exhibit 4-3. MRL Distribution for Benzo[a]pyrene	4-3
Exhibit 4-4. Analytical Methods for Benzo[a]pyrene	4-3
Exhibit 4-5. Summary of MRL Data for Chlordane	4-4
Exhibit 4-6. MRL Distribution for Chlordane	4-4
Exhibit 4-7. Analytical Methods for Chlordane	4-5
Exhibit 4-8. Summary of MRL Data for DBCP	4-5
Exhibit 4-9. MRL Distribution for DBCP	4-6
Exhibit 4-10. Analytical Methods for DBCP	4-6
Exhibit 4-11. Summary of MRL Data for DEHP	4-7
Exhibit 4-12. MRL Distribution for DEHP	4-7
Exhibit 4-13. Analytical Methods for DEHP	4-8
Exhibit 4-14. Summary of MRL Data for EDB	4-8
Exhibit 4-15. MRL Distribution for EDB	4-9
Exhibit 4-16. Analytical Methods for EDB	4-9
Exhibit 4-17. Summary of MRL Data for Heptachlor	4-10
Exhibit 4-18. MRL Distribution for Heptachlor	4-10
Exhibit 4-19. Analytical Methods for Heptachlor	4-11
Exhibit 4-20. Summary of MRL Data for Heptachlor Epoxide	4-11
Exhibit 4-21. MRL Distribution for Heptachlor Epoxide	4-12
Exhibit 4-22. Analytical Methods for Heptachlor Epoxide	4-12
Exhibit 4-23. Summary of MRL Data for Hexachlorobenzene	4-13
Exhibit 4-24. MRL Distribution for Hexachlorobenzene	4-13
Exhibit 4-25. Analytical Methods for Hexachlorobenzene	4-14
Exhibit 4-26. Summary of MRL Data for Pentachlorophenol	4-14
Exhibit 4-27. MRL Distribution for Pentachlorophenol	4-15
Exhibit 4-28. Analytical Methods for Pentachlorophenol	4-15
Exhibit 4-29. Summary of MRL Data for PCBs	4-16
Exhibit 4-30. MRL Distribution for PCBs	4-16
Exhibit 4-31. Analytical Methods for PCBs	4-17
Exhibit 4-32. Summary of MRL Data for Dioxin	4-17
Exhibit 4-33. MRL Distribution for Dioxin	4-18
Exhibit 4-34. Analytical Methods for Dioxin	4-18
Exhibit 4-35. Summary of MRL Data for Thallium	4-19
Exhibit 4-36. MRL Distribution for Thallium	4-19
Exhibit 4-37. Analytical Methods for Thallium	4-20
Exhibit 4-38. Summary of MRL Data for Toxaphene	4-20

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-39. MRL Distribution for Toxaphene	4-21
Exhibit 4-40. Analytical Methods for Toxaphene	4-21
Exhibit 4-41. Summary of MRL Data for 1,1,2-Trichloroethane	4-22
Exhibit 4-42. MRL Distribution for 1,1,2-Trichloroethane	4-22
Exhibit 4-43. Analytical Methods for 1,1,2-Trichloroethane	4-23
Exhibit 4-44. Occurrence Analysis Threshold Selection Scenarios for Contaminants with New
Possible MCLGs	4-23
Exhibit 4-45. Summary of MRL Data for Carbofuran	4-24
Exhibit 4-46. MRL Distribution for Carbofuran	4-25
Exhibit 4-47. Analytical Methods for Carbofuran	4-25
Exhibit 4-48. Summary of MRL Data for Cyanide	4-26
Exhibit 4-49. MRL Distribution for Cyanide	4-26
Exhibit 4-50. Analytical Methods for Cyanide	4-27
Exhibit 4-51. Summary of MRL Data for Endothall	4-27
Exhibit 4-52. MRL Distribution for Endothall	4-28
Exhibit 4-53. Analytical Methods for Endothall	4-28
Exhibit 4-54. Summary of MRL Data for Methoxychlor	4-29
Exhibit 4-55. MRL Distribution for Methoxychlor	4-29
Exhibit 4-56. Analytical Methods for Methoxychlor	4-30
Exhibit 4-57. Summary of MRL Data for Oxamyl	4-30
Exhibit 4-58. MRL Distribution for Oxamyl	4-31
Exhibit 4-59. Analytical Methods for Oxamyl	4-31
Exhibit 4-60. Summary of MRL Data for Styrene	4-32
Exhibit 4-61. MRL Distribution for Styrene	4-32
Exhibit 4-62. Analytical Methods for Styrene	4-33
Exhibit 5-1. Threshold Information Summary: Potential Improvements in Analytical Feasibility
	5-1
Exhibit 5-2. EQL Threshold Results	5-2
Exhibit 5-3. Threshold Information Summary: Possible Lower MCLGs	5-3
Exhibit 5-4. Occurrence Threshold Results	5-4

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Abbreviations and Acronyms
2,4-D
2,4-dichlorophenoxyacetic acid
DBCP
l,2-dibromo-3-chloropropane
DEHP
di(2-ethylhexyl)phthalate
EDB
ethylene dibromide
EPA
U.S. Environmental Protection Agency
EQL
estimated quantitation level
ICR
Information Collection Request
MCL
maximum contaminant level
MCLG
maximum contaminant level goal
MDL
method detection limit
mg/L
milligrams per liter
Hg/L
micrograms per liter
MRL
minimum reporting level
NPDWR
National Primary Drinking Water Regulation
PCBs
polychlorinated biphenyls
PQL
practical quantitation level
PT
proficiency testing
RfD
reference dose
SDWA
Safe Drinking Water Act
SOC
synthetic organic compound
VOC
volatile organic compounds
iv

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Executive Summary
The U.S. Environmental Protection Agency (EPA) has completed its third Six-Year Review
(Six-Year Review 3) of national primary drinking water regulations (NPDWRs). The 1996 Safe
Drinking Water Act (SDWA) Amendments require the U.S. Environmental Protection Agency
(EPA or the Agency) to periodically review existing NPDWRs. Section 1412(b)(9) of SDWA
reads:
,..[t]he Administrator shall, not less often than every 6 years, review and revise, as
appropriate, each national primary drinking water regulation promulgated under this
subchapter. Any revision of a national primary drinking water regulation shall be
promulgated in accordance with this section, except that each revision shall maintain,
or provide for greater, protection of the health of persons.
The primary goal of the Six-Year Review process is to identify NPDWRs for possible regulatory
revision. Although the statute does not define when a revision is "appropriate," as a general
benchmark, EPA considered a possible revision to be "appropriate" if, at a minimum, it presents
a meaningful opportunity to:
•	improve the level of public health protection, and/or
•	achieve cost savings while maintaining or improving the level of public health protection.
For Six-Year Review 3, EPA obtained and evaluated new information that could affect a
NPDWR, including information on health effects (USEPA, 2016c), analytical feasibility
(USEPA, 2016b and 2009a), and occurrence (USEPA, 2016a). EPA identified new health effects
or analytical methods information that indicated it may be possible to revise NPDWRs for
several contaminants. Consequently, EPA conducted occurrence and exposure analyses at
threshold concentrations that are below current maximum contaminant levels (MCLs) to
determine if there is a meaningful opportunity to improve the level of public health protection by
reducing MCLs. This document describes the data and method EPA used to establish the
threshold values that it used for the occurrence analyses.
For some contaminants, new information on analytical feasibility could affect the NPDWR
because these are contaminants for which the MCL equals a practical quantitation limit (PQL).
EPA evaluated new information for performance testing data, method minimum detection limits
(MDL), and compliance data minimum reporting levels (MRL) to determine whether it could
develop an estimated quantitation level (EQL) threshold below the current PQL.EPA's method
for developing an EQL has essentially three steps - one for each of the three information
sources: PT data, MRL data, and MDL values. The first step is to review the conclusion of the
PT analysis. If the PT data indicate potential to revise the PQL, then the objective of the next
steps is to identify an EQL (or verify the use of a health-based threshold) for the occurrence
analysis. The second step is to determine whether the modal MRL is a feasible EQL and, if so,
the third step is to determine whether the MDL multiplier approach supports that EQL value. If
the modal MRL is not a feasible EQL, then EPA uses the MDL multiplier approach to establish
an EQL.
ES-1

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
If the PT data do not indicate potential to revise the PQL, then the objective of the next steps is
to determine whether the MRL and MDL data concur with this finding. When the MRL and
MDL data confirm the finding, there is no basis for an EQL that is less than the PQL. When
these data contradict the finding, however, EPA used these secondary data sources to derive an
EQL (or verify the use of a health-based threshold) for the occurrence analysis.
MCL Currently Limited by PQL
The summary in Exhibit ES-1 shows that these data sources did not support EQL development
for seven contaminants. EPA based EQLs on MDL data for five contaminants and MRL data for
one. The MDL data indicate the greatest potential to revise PQL values. EPA used the MDL data
to derive an EQL for the following contaminants: chlordane, heptachlor, heptachlor epoxide,
hexachlorobenzene, and toxaphene. EPA did not use MDL values to develop EQL values for
three contaminants despite there being an MDL lower than the PQL: benzo[a]pyrene, DBCP, and
pentachlorophenol. For benzo[a]pyrene, an EQL based on the MDL would be the same as the
PQL. For DBCP, an EQL based on MDL data was less than 70 percent of the MRL values in the
database. For pentachlorophenol, EPA did not develop an EQL because six of the seven MDL
values rounded to or exceeded the PQL.
Exhibit ES-1. EQL Threshold Results

PQL
EQL

Contaminant
(nq/L)
(nq/L)
Basis
Benzo[a]pyrene
0.2
none
Data do not support EQL < PQL
Chlordane
2
1
Based on 10 x MDL
1,2-Dibromo-3-Chloropropane
0.2
none
Data do not support EQL < PQL
Di (2-ethylhexyl)phthalate
6
none
Data do not support EQL < PQL
Ethylene Dibromide
0.05
none
Data do not support EQL < PQL
Heptachlor
0.4
0.1
Based on 10 x MDL
Heptachlor Epoxide
0.2
0.04
Based on 10 x MDL
Hexachlorobenzene
1
0.1
Based on 10 x MDL
Pentachlorophenol
1
none
Data do not support EQL < PQL
Polychlorinated Biphenyls
0.5
none
Data do not support EQL < PQL
Dioxin
3.0 x 10-5
5.0 x 10-6
Based on MRL mode
Thallium
2
none
Data do not support EQL < PQL
Toxaphene
3
1
Based on 10 x MDL
1,1,2-T richloroethane
5
3
Based on MCLG (EQL < MCLG)
MCL Greater than Possible MCLG
For other contaminants, new health effects information indicates a possible lower maximum
contaminant level goal (MCLG), which is a non-regulatory, health protection goal. For these
contaminants, the MCL is currently equal to the MCLG. A lower MCLG is an opportunity to
lower the MCL. Therefore, EPA reviewed quantitation data to evaluate the feasibility of an MCL
as low as the potential MCLG.
ES-2

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit ES-2 provides a summary of the occurrence thresholds for this contaminant group.
EPA's analysis indicates that most of the thresholds can be set equal to corresponding possible
MCLG values, regardless of whether PQL values exceed possible MCLGs. In five cases,
alternative values must be used because analytical feasibility will most likely limit setting an
MCL equal to a possible MCLG.
For six contaminants - carbofuran, cyanide, endothall, methoxychlor, oxamyl, and styrene - the
PQL potentially limits setting an MCL equal to the possible MCLG. For carbofuran, cyanide,
and methoxychlor, the EQL was based on 10 x MDL and supported threshold values that were
less than the PQL. For endothall and oxamyl, although the PT data do not support a reduction of
the PQLs, the MRL and MDL data do support the use of the possible MCLG values as thresholds
for the occurrence analysis.
Finally, for styrene, the modal MRL meets the EQL criteria.
Exhibit ES-2. Occurrence Threshold Results


Occurrence


Possible
Threshold

Contaminant
MCLG (nq/L)
(nq/L)
Basis
Carbofuran
0.6
5
EQL based on 10 x MDL
Cyanide
4
50
EQL based on 10 x MDL
cis-1,2-Dichloroethylene
10
10
possible MCLG
Endothall
50
50
possible MCLG
Fluoride
900
900
Possible MCLG
Hexachlorocyclopentadiene
40
40
possible MCLG
Methoxychlor
0.1
1
EQL based on 10 x MDL
Oxamyl
10
10
possible MCLG
Selenium
40
40
possible MCLG
Styrene
0
0.5
EQL based on MRL mode
Toluene
600
600
possible MCLG
Xylene
1000
1000
possible MCLG
ES-3

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
1 Introduction
The U.S. Environmental Protection Agency (EPA or the Agency) has conducted its third Six-
Year Review ("Six-Year Review 3") of national primary drinking water regulations (NPDWRs).
The 1996 Safe Drinking Water Act (SDWA) Amendments require that the Agency periodically
review existing NPDWRs. Section 1412(b)(9) of SDWA reads:
,..[t]he Administrator shall, not less than every 6 years, review and revise, as
appropriate, each primary drinking water regulation promulgated under this title.
Any revision of a national primary drinking water regulation shall be promulgated
in accordance with this section, except that each revision shall maintain, or
provide for greater, protection of the health of persons.
The primary goal of the Six-Year Review process is to identify possible regulatory revisions.
Although the statute does not define when a revision is "appropriate," as a general benchmark,
EPA considered a possible revision to be "appropriate" if, at a minimum, it presents a
meaningful opportunity to:
•	improve the level of public health protection, and/or
•	achieve cost savings while maintaining or improving the level of public health protection.
For Six-Year Review 3, EPA implemented the protocol that it developed for the first Six-Year
Review (USEPA, 2003), as revised during the second Six-Year Review (USEPA, 2009c). EPA
obtained and evaluated new information on various factors that could indicate potential to revise
an NPDWR: health effects (USEPA, 2016c), analytical feasibility (USEPA, 2016b), and
occurrence (USEPA, 2016a). This document serves as a bridge between the findings of the
health effects and analytical feasibility studies, which identify opportunities for NPDWR
revisions, and the occurrence analysis, which identifies whether a revision is a meaningful
opportunity for health risk reduction.
1.1 Background
An NPDWR includes a maximum contaminant level (MCL), which is the regulatory limit for the
amount of a contaminant allowed in water distributed by public water systems. EPA establishes
MCLs after identifying a maximum contaminant level goal (MCLG). The MCLG is a
concentration at which no known or anticipated adverse human health effect occurs. For
carcinogens, the MCLG is often equal to zero because there is no known safe dosage. For other
contaminants, the MCLG is based on a reference dose (RfD) at which EPA does not expect
adverse health effects to occur.
After identifying the MCLG, EPA must set the MCL as close to the MCLG as feasible. For some
contaminants, it is not feasible to set the MCL equal to the MCLG because of limitations in
contaminant measurement capabilities at very low concentrations. EPA identifies a practical
quantitation limit (PQL) when it establishes an NPDWR, which is "the lowest achievable level of
analytical quantitation during routine laboratory operating conditions within specified limits of
precision and accuracy" (50 Federal Register 46902, November 13, 1985). Thus, a PQL reflects
1-1

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
both the physical limitation of approved analytical methods and the practical limitations of
variability in laboratory performance nationwide.
For a carcinogen, EPA often bases the MCL on the PQL because it is not possible to measure
concentrations all the way down to zero. Analytical feasibility can improve over time, however.
Consequently, the Six-Year Review process is an opportunity to evaluate whether new
information regarding quantitation shows that PQLs for carcinogens can be reduced, which
introduces the possibility of reducing the MCLs for carcinogens.
1.2 Estimated Quantitation Level Development
When analytical methods information indicates potential to revise an MCL, EPA estimates
occurrence to evaluate whether the revision could be a meaningful opportunity for health risk
reduction. The occurrence estimates provide information on the number of systems and people a
revision might affect. To derive these estimates, EPA identifies a threshold value below the
current MCL at which to estimate occurrence. The threshold represents an estimated quantitation
level (EQL).1 This report documents EPA's approach to identifying these thresholds.
EPA used these thresholds to estimate possible system and population impacts in the occurrence
and exposure analysis conducted for the third Six-Year Review (USEPA, 2016a). EPA compared
contaminant occurrence estimates for these thresholds (i.e., the number of systems with water
quality exceeding a threshold) with baseline occurrence estimates at current MCLs. The
difference between these two occurrence estimates indicates potential for health risk reduction of
an MCL revision. EPA based its determinations about whether a reduction in the MCL for a
contaminant would provide a meaningful opportunity to improve the level of public health
protection on these estimates.
Analyzing the feasibility of reducing a contaminant's current PQL was one of the review tasks of the
Six-Year Review 3. For the PQL assessment, EPA obtained and evaluated new information regarding
the potential to revise PQL values. The primary sources of information for the PQL assessment were
laboratory proficiency testing (PT) study results obtained during Six-Year Review 2 and Six-Year
Review 3. The PT studies involve the use of spiked samples to evaluate laboratory quantitation
capabilities. USEPA (2016b) describes the review method, PT data, and findings for the PQL
analysis. For Six-Year Review 3, EPA did not always have sufficient PT data below current PQLs to
actually recalculate any PQL or derive EQLs for the occurrence and exposure analysis. Instead, EPA
used the PT study passing rate results (i.e., the percent of laboratories passing a performance test for
a given study) at and below the current PQL and the result of a linear regression analysis to indicate
whether the PT data support a reduction in the PQL.
Because the PT results were either not available below the PQL or did not provide conclusive
indications regarding a potential to revise a PQL or how far below the PQL quantitation might be
feasible, EPA relied on two alternate approaches to estimate EQLs: an approach based on the
minimum reporting levels (MRLs) obtained as part of the Six-Year Review 3 Information Collection
1 Although the EQLs are estimates of quantitation capabilities below a PQL, they do not represent the Agency's
intent to promulgate new PQLs. Any revisions to regulatory monitoring requirements such as PQLs will be made as
part of future rule-making efforts.
1-2

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Request (ICR), and an approach based on method detection limits (MDL). While EPA prefers to use
laboratory performance data to calculate the PQL, the MRL and MDL information can be valuable to
indicate whether it is possible to quantitate at levels below the current PQL.
An MRL is the lowest level or contaminant concentration that a laboratory can reliably achieve
within specified limits of precision and accuracy under routine laboratory operating conditions using
a given method (USEPA, 2016a). The MRL values provide direct evidence from actual monitoring
results about whether quantitation below the PQL using current analytical methods is feasible. An
MDL is a measure of analytical method sensitivity (USEPA, 2016b). MDLs have been used in the
past to derive PQLs for regulated contaminants. In addition, EPA used MDLs to help identify
possible analytical feasibility levels for Six-Year Review 1 (USEPA, 2003b). Consequently, EPA
used the MDLs as a second input to the EQL development process. Both sources of data provide
additional information on the feasibility of revising PQLs. Therefore, the Agency also evaluated
whether MRL and MDL data confirmed or contradicted the conclusions of the PT data review. For
most contaminants, the MRL and MDL data supported EPA's conclusion based on PT data.
1.3 Contaminants
For most contaminants, EPA established an EQL, which is an estimate of the possible lower bound
for a PQL. The current PQL for a contaminant is based on historical analytical capabilities, generally
the quantitation capabilities at the time EPA promulgated the existing NPDWR for the contaminant.
When a contaminant has a PQL that is higher than its MCLG, the MCL cannot be lower than the
PQL. Thus, improvements in analytical feasibility indicate potential opportunity to lower the PQL for
some contaminants that have MCLs limited by PQLs, and, therefore, lower the MCL closer to
MCLG.
Exhibit 1-1 shows contaminants for which historical PQLs provided a lower bound on MCLs.
Most of the contaminants are carcinogens for which MCLGs are equal to zero. For two,
however, MCLGs are nonzero, but PQLs precluded setting MCLs as low as the MCLGs.
Findings on the PT data supporting PQL revision from the analytical feasibility studies (USEPA,
2016b) are also included in the able. EPA evaluated whether new information indicated possible
EQL values less than the PQLs shown in the table.
Exhibit 1-1. Contaminants Where MCLs Limited by Analytical Feasibility




Do PT Data


PQL
MCL
Support PQL
Contaminant
MCLG (|jg/L)
(Mg/L)
(M9/L)
Revision?
Benzo[a]pyrene
0
0.2
0.2
No
Chlordane
0
2
2
No
1,2-Dibromo-3-Chloropropane
0
0.2
0.2
No
Di (2-ethylhexyl)phthalate
0
6
6
No
Ethylene Dibromide
0
0.05
0.05
No
Heptachlor
0
0.4
0.4
No
Heptachlor Epoxide
0
0.2
0.2
No
Hexachlorobenzene
0
1
1
Yes
Pentachlorophenol
0
1
1
No
Polychlorinated Biphenyls
0
0.5
0.5
No
2,3,7,8-Tetrachlorodibenzo-p-Dioxin
0
3.0 x10-5
3.0x10-5
No
1-3

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations




Do PT Data


PQL
MCL
Support PQL
Contaminant
MCLG (|jg/L)
(Mg/L)
(M9/L)
Revision?
Thallium
0.5
2
2
No
Toxaphene
0
3
3
No
1,1,2-T richloroethane
3
5
5
Yes
Source: USEPA, 2016b and 2009a.
For many other contaminants, EPA set the MCL equal to the MCLG. Because the MCLG is
based on health risk information, new information such as a new health risk study may indicate
that this value should be lower. Exhibit 1-2 shows contaminants for which new health effects
information since EPA promulgated the NPDWRs indicates possible MCLGs that are lower than
current MCLGs. For these contaminants, EPA determined whether the threshold for the
occurrence analysis could equal the possible MCLG and, if not, determined whether quantitation
information supported an EQL below the current MCLG.
Exhibit 1-2. Contaminants Where MCLs are Greater than Possible Lower MCLGs




Possible
Do PT Data

Current
PQL
MCL
MCLG
Support PQL
Contaminant
MCLG (ng/L)
(ng/L)
(ng/L)
(Mfl/L)
Revision?
Carbofuran
40
7
40
0.6
No
Cyanide
200
100
200
4
No
cis-1,2-Dichloroethylene
70
5
70
10
Yes
Endothall
100
90
100
50
No
Fluoride
4000
500
4000
900
No
Hexachlorocyclopentadiene
50
1
50
40
No
Methoxychlor
40
10
40
0.1
Yes
Oxamyl
200
20
200
10
No
Selenium
50
10
50
40
No
Styrene
100
5
100
0
Yes
Toluene
1,000
5
1,000
600
Yes
Xylene
10,000
5
10,000
1,000
No
Source: USEPA, 2016b and 2009a.
This report documents EPA's selection of thresholds for the occurrence analysis of these two
groups of drinking water contaminants and contains the following: descriptions of the available
data sources (Section 2); a description of the approaches EPA used to evaluate the data and select
occurrence thresholds (Section 3): detailed results by contaminant (Section 4); and a summary of
the thresholds selected for the occurrence analysis (Section 5).
1-4

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
2 Data Sources
An EQL is an estimate of a possible quantitation limit below a PQL. Therefore, EPA sought to
base EQL values on the same type of data that it used to derive PQLs. EPA developed PQLs
using two approaches (USEPA, 2009a). The first approach, which EPA prefers, requires
laboratory performance testing (PT) data. For a performance test, multiple laboratories quantitate
samples that a testing facility has spiked with a known contaminant concentration. The testing
facility reviews the results and determines how many laboratories estimate a value within an
accuracy range around the spiked value (e.g., plus or minus 20%). The percentage of laboratories
in the accuracy range is the passing rate (e.g., if 15 of 20 are in the range, the passing rate is
75%). A PQL based on PT data is the lowest value for which at least 75 percent of laboratories
tested can quantitate within prescribed accuracy limits.
When PT data were not available, EPA used a second approach to derive PQLs. This approach
utilizes minimum detection level (MDL) data for applicable analytical methods. For this
approach, EPA multiplies an MDL by a factor - usually 5 or 10 - to compute a PQL.
For Six-Year Review 3 and the second Six-Year Review, EPA obtained PT study results from
testing facilities (USEPA 2016b and 2009a). The value reported for each PT study is a passing
rate, which is the percent of laboratories that successfully quantitated samples spiked with a
particular concentration within prescribed accuracy limits. Although PT passing rates would
seem to be ideal data for developing EQL values, unfortunately the studies were rarely
conducted at spiked values that are less than the PQLs. Therefore, the PT data could only provide
a general indication of whether there is potential to derive an EQL below the PQL.
Because of insufficient PT data, EPA used minimum reporting levels (MRLs) from the Six-Year
Review 3 Information Collection Request (ICR) database along with the MDL approach to
derive EQLs. Section 2.1 describes the MRL data. Section 2.2 describes the source of MDLs.
2.1 MRL Data
The Six-Year Review 3 ICR database contains compliance monitoring data for 2006 through
2011. USEPA (2016a) provides a description of the data collection, data management, and
quality assurance methods the Agency used to establish a high quality, national contaminant
occurrence database consisting of data from 46 states plus Washington, D.C., American Samoa,
and many other primacy entities such as Tribes. This database contains several million drinking
water compliance monitoring samples.
This Six-Year Review 3 ICR database also contains a substantial number of MRL values. An
MRL is the lowest level or contaminant concentration that a laboratory can reliably achieve
within specified limits of precision and accuracy under routine laboratory operating conditions
using a given method (USEPA, 2016a). In other words, the MRL is the lowest contaminant
concentration that can be reliably quantified in the laboratory and reported to primacy agencies.
When compliance monitoring data are recorded, laboratories should report "
-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
performed a variety of data quality checks and data transformations on the MRL data in
consultation with state data management staff. USEPA (2016a) describes the data management
process, including measures taken to address data quality concerns that affect the occurrence and
exposure analysis.
The MRL values provide EPA with valuable insight into actual analytical capabilities across
laboratories and States. MRLs can vary across laboratories because of differences in the
analytical method used as well as differences in instrumentation, implementation, and reporting.
By examining the distribution of MRL values for a contaminant, EPA can identify whether
laboratory performance is relatively uniform (e.g., most MRLs are the same) or highly variable
(e.g., MRLs that vary by one or more orders of magnitude). In particular, the mode or most
frequently occurring value is a potential candidate for EQL when a substantial share of the MRL
values for a contaminant equal the modal MRL2.
2.2 MDL Data
The MDL multiplier approach for estimating an EQL applies a multiplier usually ranging from
five to ten to the MDL. An MDL is a measure of analytical method sensitivity (USEPA, 2016b),
defined in 40 CFR Part 136 Appendix B as "the minimum concentration of a substance that can
be reported with 99 percent confidence that the analyte concentration is greater than zero" for a
given method. Although EPA has used this method to establish PQLs in the past, EPA is not
using MDLs for this purpose during Six-Year Review 3. Instead, EPA is using the MDL
approach to help identify EQLs below current PQLs for occurrence and exposure analysis.
MDLs can vary by analytical method and contaminant. USEPA (2016b) and USEPA (2009a)
provide MDLs by contaminant and analytical method. The MDL values or ranges of values are
for the approved analytical methods developed by EPA for drinking water compliance
monitoring.
Summary data by contaminant and method in Section 4 of this document includes only upper
bound values for any MDL ranges reported in USEPA (2016b) or USEPA (2009a). EPA used
only upper bound values for a particular method and contaminant in an effort to derive an EQL
that would represent a level at which most laboratories should be able to quantitate; the lower
bound value could result in an EQL that is below the analytical capabilities of some laboratories.
The multiplier for MDLs is used to account for the variability and uncertainty that can occur at the
MDL. Historically, the MDL multiplier method was mostly used in the early years of rule
development for NPDWRs when insufficient PT data were available. Once sufficient data became
available, most of the PQLs that were developed using the MDL multiplier were validated using PT
data.
2 The modal MRL used in the EQL analysis is the mode across all reported MRL values for a contaminant in the
SYR3 ICR dataset. This mode may differ from the mode reported in The Analysis of Regulated Contaminant
Occurrence Data from Public Water Systems in Support of the Third Six-Year Review of National Primary Drinking
Water Regulations: Chemical Phase Rules and Radionuclides Rules (USEPA, 2016a), which reports the mode of the
state-level modes instead of the mode of all MRL value.
2-2

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
3 Threshold Development Method
This section provides an overview of the method EPA used to identify thresholds for the third
Six-Year Review occurrence analysis. For the contaminants shown in Exhibit 1-1 (current MCL
based on PQL), EPA evaluated available data to derive an EQL. For the contaminants shown in
Exhibit 1-2 (current MCL based on MCLG), EPA first determined whether the possible MCLG
(USEPA, 2015c) could be the threshold. When available information did not support quantitation
as low as the possible MCLG, EPA evaluated whether it could derive an EQL between the PQL
and possible MCLG.
As noted in Section 2, EPA used three sources of information to derive an EQL:
•	PT passing rates reported in the analytical methods analysis (USEPA 2016b and 2009a);
•	MRL values from the occurrence database; and
•	MDL values for EPA-developed analytical methods.
First, EPA evaluated whether the PT data indicated potential to revise the PQL. However, the PT
studies were rarely conducted at spiked concentrations lower than current PQLs and thus the data
are limited for identifying an EQL. Nevertheless, indications of potential to revise would add
credibility to EQLs based on the other two data sources. Therefore, EPA primarily considered
whether there were several studies for spiked values less than the PQL with passing rates greater
than 75%. This type of PT data would be clear indication of potential to reduce the PQL.
Second, EPA evaluated the MRL data using the analysis method developed for second Six-Year
Review (2009b). The Agency identified the mode and estimated the percentage of MRL values
less than or equal to the mode. When 80 percent or more of the MRL values were less than or
equal to the mode, it was a candidate EQL value as long as it was less than the corresponding
PQL.
If the modal MRL was not a feasible EQL candidate, then EPA reviewed the MDL data to
determine the feasibility of deriving an EQL by multiplying the MDL by a factor of 10 (or 5 for
EDB and dioxin based on the factor used for original PQL development). In some instances,
there were multiple MDL values. EPA based the EQL on the highest factor-adjusted MDL value
that was less than the PQL.
For the contaminants shown in Exhibit 1-1, if the available data did not support an EQL less than
the PQL, then EPA did not develop an EQL. For those shown in Exhibit 1-2, if the data
supported an EQL value that was less than the possible MCLG, then EPA noted this and used the
possible MCLG as the threshold for the occurrence analysis. Exhibit 3-1 provides a summary of
the EQL steps.
3-1

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 3-1. EQL Development Steps
PT data support
PQL reduction
Review PQL
Analysis
Findings
PT data do
not support
PQL reduction
Review MRL
data for
feasible EQL
Review MRL
data for
feasible EQL
dentified
EQL
EQL not
identified
Identified
EQL
EQL not
identified
Review MDL
data to
confirm EQL
Review MDL
data to
identify EQL
Review MDL
data to
confirm EQL
Review MDL
data to
identify EQL
Note: When the feasible EQL is less than a possible MCLG, then the occurrence threshold is the possible MCLG
3-2

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
4 Development of Individual EQLs
This section provides a discussion of the occurrence thresholds developed for the contaminants
addressed in this report. Where applicable, the discussion for each contaminant contains an
overview of the PQL review in USEPA (2016b and 2009a), followed by MRL summary data and
MDL values. There are two subsections - one for the contaminants shown in Exhibit 1-1 and one
for those shown in Exhibit 1-2.
4.1 MCL Currently Limited by PQL
Most of the contaminants for which the MCL equals the PQL are carcinogens for which MCLGs
are zero. Exhibit 4-1 illustrates the analysis objective for these contaminants - to identify an
EQL that is less than the current PQL to use as an occurrence threshold (case A). For two
contaminants, however, a PQL limits the MCL, which is greater than a nonzero MCLG. For
these contaminants, if data support an EQL that is less than the PQL, then the occurrence
threshold depends on whether the EQL is greater than the MCLG (case B) or is less than the
MCLG (case C).
Exhibit 4-1. Occurrence Analysis Threshold Selection Scenarios
A. EQL is the occurrence
analysis threshold when
MCLG = 0 and data support
EQL < PQL
B. EQL is the occurrence
analysis threshold if data
support
MCLG < EQL < PQL
MCLG

EQL
iz:
MCL=
PQL
Ojig/L
MCLG

EQL

MCL=




PQL

^z:
o (J-g/L
C. MCLG is the occurrence
analysis threshold when
data support
EQL < MCLG < PQL
Ojig/L
4.1.1 Benzo[a]pyrene
The MCL for benzo[a]pyrene equals the PQL of 0.2 (J,g/L. The MCLG is zero. Although a health
effects assessment is in progress, there is no new health effects information that suggests a
change in the MCLG. Consequently, the threshold for the occurrence analysis is based on
analytical feasibility.
There are no PT study results at spiked concentrations below the PQL and several passing rates
for the available PT studies at concentrations greater than the PQL are below 75 percent
EQL

MCLG

MCL=
PQL
4-1

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
(USEPA, 2009a). Because of the lack of data below the PQL and passing rate variability, EPA
determined that PT data do not support reduction of the PQL.
As shown in Exhibit 4-2, the modal MRL for benzo[a]pyrene is 0.02 (J,g/L. Summary data show
that 35.6 percent of the MRLs are equal to this value and 37 percent are equal to or less than it.
Exhibit 4-3 shows that there are multiple clusters of MRLs between the mode and the PQL of
0.2 (J,g/L. Unlike the PT data, the MRL. data indicate that there may be potential to lower the PQL
because over 99 percent of the MRL values are below the PQL. The percentage of the MRL
values that are less than or equal to the mode does not meet the 80 percent threshold, however.
Therefore, EPA did not base the EQL on the modal MRL. Consequently, EPA reviewed MDL
values to determine whether they support an EQL below the PQL.
Exhibit 4-2. Summary of MRL Data for Benzo[a]pyrene
MRL Value Category
Number of Records
Percentage of Records
All
60,569
100%
Less than mode
872
1.4%
Equal to mode (0.02 |jg/L)
21,563
35.6%
Greater than mode
38,134
63.0%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-2

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-3. MRL Distribution for Benzo[a]pyrene
100.0%
80.0%
0)
_D
5
Z 60.0%
¦£ 40.0%
y
o>

a
c\V

&
J* 0
MRL (jig/L)
O-

Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-4 shows EPA's approved methods for the detection of benzo[a]pyrene, and
corresponding MDLs. Multiplying the MDLs by 10 results in a possible EQL range from 0.16 to
2.3 (J,g/L. The lower bound of this range rounds to 0.2 (J,g/L, which is the PQL. Thus, the MDL
data do not support an EQL below the PQL.
Exhibit 4-4. Analytical Methods for Benzo[a]pyrene
Method
MDL (nq/L)
MDL x 10 (pxq/L)
525.2
0.23
2.3
550
0.029
0.29
550.1
0.016
0.16
Source: USEPA, 2009a (upper bound values when ranges are reported)
EPA concluded that although MRL values are generally below the PQL, the combination of PT
and MDL data do not support revision of the PQL for benzo[a]pyrene. Therefore, EPA did not
develop an EQL.
4.1.2 Chlordane
The MCL for chlordane equals the PQL of 2 (J,g/L. The MCLG is zero and there is no new health
effects information that suggests a change in the MCLG. Consequently, the threshold for the
occurrence analysis is based on analytical feasibility.
The PT data does not include studies with spiked concentrations less than the PQL. Passing rates
for the studies above the PQL are greater than 75 percent (USEPA, 2016b). Because there are no
4-3

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
studies at concentrations less than the PQL, EPA determined that PT data do not support
reduction of the PQL.
As shown in Exhibit 4-5, the modal MRL for chlordane is 0.2 (J,g/L. Almost 54 percent of the
MRL values are equal to or less than the modal value. The percentage of the MRL values that are
less than or equal to the mode does not meet the 80 percent threshold. Therefore, EPA did not
base the EQL on the modal MRL. Exhibit 4-6 shows that more than 99 percent of the MRL
values are less than the PQL of 2 (J,g/L. Consequently, EPA reviewed MDL values to determine
whether they support an EQL below the PQL.
Exhibit 4-5. Summary of MRL Data for Chlordane
MRL Value Category
Number of Records
Percentage of Records
All
59,923
100%
Less than mode
15,272
25.5%
Equal to mode (0.2 |jg/L)
16,932
28.3%
Greater than mode
27,719
46.3%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Exhibit 4-6. MRL Distribution for Chlordane
100.0%
80.0%
0>
_3
5
_i 60.0%
u
Q.
40.0%
20.0%
0.0%
4.1%
28.5%
16.:
2.2% 2.1%
23.5%
9.8% 9.8%
3.1%
i	1
0.1%
cf" <§>	ry^
£>¦	O	O*	O'	O-
oV <$>' <$>' \V'	^ o5*
MRL (jlg/L)
%
&
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-7 shows EPA's approved methods for the detection of chlordane and the MDLs.
Applying a multiplier of 10 would give a possible EQL range from 0.015 to 2.2 (J,g/L. One of
these values is greater than the PQL. EPA used the highest value below the PQL (1.4 (J,g/L) and
rounded to 1 [j,g/L to obtain an EQL. Almost 97 percent of the MRLs for chlordane in the Six-
Year Review 3 ICR database are less than or equal to 1 (J,g/L.
4-4

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-7. Analytical Methods for Chlordane
Method
MDL (nq/L)
MDLx10(nq/L)
505
0.14
1.4
508
0.0015
0.015
508.1
0.004
0.04
525.2
0.22
2.2
525.3
0.002
0.02
Source: USEPA, 2016b (upper bound values when ranges are reported)
4.1.3 1,2-Dibromo-3-chloropropane (DBCP)
The MCL for DBCP equals the PQL of 0.2 (J,g/L. The MCLG is zero and there is no new health
effects information that suggests a change in the MCLG. Consequently, the threshold for the
occurrence analysis is based on analytical feasibility.
The PT data show greater than 80 percent passing rates for all studies. There are, however, no
studies with spiked values below the PQL (USEPA, 2016b). Because there are no studies below
the PQL, EPA determined that PT data do not support reduction of the PQL.
As shown in Exhibit 4-8, the modal MRL for DBCP is 0.5 (J,g/L, which is greater than the PQL
of 0.2 (J,g/L. Therefore, EPA did not base the EQL on the modal MRL regardless of the large
proportion of MRL values below the mode. Exhibit 4-9 shows that almost 70 percent of the
MRL values are greater than the PQL. Consequently, EPA reviewed MDL values to determine
whether they support an EQL below the PQL.
Exhibit 4-8. Summary of MRL Data for DBCP
MRL Value Category
Number of Records
Percentage of Records
All
126,959
100%
Less than mode
49,261
38.8%
Equal to mode (0.5 |jg/L)
34,759
27.4%
Greater than mode
42,939
33.8%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-5

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-9. MRL Distribution for DBCP
100.0%
80.0%
5
60.0%
CC
s
*5
£ 40.0%
0)
u
i—
a>
20.0%
0.0%
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-10 shows EPA's approved methods for the detection of DBCP and the MDLs.
Applying a multiplier of 10 would give a possible EQL range from 0.09 to 2.6 (J,g/L. EPA
excluded the highest values, which exceed the PQL. The higher of the two remaining values
indicate a potential EQL of 0.1 (J,g/L.
Exhibit 4-10. Analytical Methods for DBCP
Method
MDL (nq/L)
MDL x 10 (pxq/L)
504.1
0.01
0.1
524.2
0.26
2.6
524.3
0.063
0.63
551.1
0.009
0.09
Source: USEPA, 2016b (upper bound values when ranges are reported)
Neither the MRL nor PT data support establishing an EQL value that is less than the PQL of 0.2
[j,g/L. Although the MDL data support an EQL of 0.1 (J,g/L, almost 70 percent of the MRL values
are greater than this value. Therefore, EPA did not develop an EQL.
4.1.4 Di(2-ethylhexyl)phthalate (DEHP)
The MCL for DEHP equals the PQL of 6 (J,g/L. The MCLG is zero. Although a health effects
assessment is in progress, there is no new health effects information that suggests a change in the
MCLG. Consequently, the threshold for the occurrence analysis is based on analytical feasibility.
Passing rates for several PT studies are below 75 percent, including two studies with spiked
concentrations below the PQL (USEPA, 2009a). Because of the low passing rates, EPA
determined that PT data do not support reduction of the PQL.
27.5%
27.8%
1 i% 2.0% 3.3% 4.5%
I i	1 I I
18.2%
12.2%
0.2% 2S% 0.7%
&
&	o
o
	>
N V ^
MRL (ug/L)
4-6

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
As shown in Exhibit 4-11 and, the modal MRL for DEHP is 0.6 (J,g/L. Summary data show that
31.8 percent of the MRLs are equal to this value, and 40.7 percent of the MRL values are equal
to or less than it. Exhibit 4-12 shows multiple clusters of MRLs between the mode and the PQL
of 6 (J,g/L. Unlike the PT data, the MRL data appear to indicate that there is potential to lower the
PQL because more than 99 percent of values are below the PQL. The percentage of the MRL
values that are less than or equal to the mode does not meet the 80 percent threshold. Therefore,
EPA did not base the EQL on the modal MRL. Consequently, EPA reviewed MDL values to
determine whether they support an EQL below the PQL.
Exhibit 4-11. Summary of MRL Data for DEHP
MRL Value Category
Number of Records
Percentage of Records
All
55,550
100.0%
Less than mode
4,942
8.9%
Equal to mode (0.6 |jg/L)
17,648
31.8%
Greater than mode
32,960
59.3%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Exhibit 4-12. MRL Distribution for DEHP
100.c
80.0%
>
_i 60.(
40.(
u
a.
20.C
0.0%

34.6%
1.1%
5.0%
I I
22.1% 22.£
9.7%
1.0% 2.4% 1.1% o.l% 0.2%
V> >
V
N*
^	^ 
-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-13. Analytical Methods for DEHP
Method
MDL (nq/L)
MDLx10(nq/L)
506
2.25
22.5
525.2
1.3
13
Source: USEPA, 2009a (upper bound values when ranges are reported)
EPA concluded that although MRL values are generally below the PQL, the combination of PT
and MDL data do not support revision of the PQL for DEHP. Therefore, EPA did not develop an
EQL.
4.1.5 Ethylene Dibromide (EDB)
The MCL for EDB equals the PQL of 0.05 (J,g/L. The MCLG is zero and there is no new health
effects information that suggests a change in the MCLG. Therefore, the threshold for an
occurrence analysis is based on analytical feasibility.
There are no PT study results with spiked concentrations below the PQL. The results for spiked
concentrations greater than the PQL are scattered throughout the range from 75 percent to 100
percent (USEPA, 2009a). Therefore, EPA determined that the PT data do not support PQL
reduction.
As shown in Exhibit 4-14, the modal MRL for EDB is 0.5 (J,g/L which is greater than the PQL of
0.05 (J,g/L. Therefore, EPA did not base the EQL on the modal MRL regardless of the large
proportion of MRL values below the mode. Exhibit 4-15 shows that about 56 percent of the
MRL values are greater than the PQL. Consequently, EPA reviewed MDL values to determine
whether they support an EQL below the PQL.
Exhibit 4-14. Summary of MRL Data for EDB
MRL Value Category
Number of Records
Percentage of Records
All
88,891
100%
Less than mode
55,401
62.3%
Equal to mode (0.5 |jg/L)
26,205
29.5%
Greater than mode
7,285
8.2%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-8

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-15. MRL Distribution for EDB
100.0%
« 80.0%
01
_3
5
_i 60.0%
te.
S
¦s
£ 40.0%
0)
u
i_
20.0%
0.0%
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-16 shows EPA's approved methods for the detection of EDB, and the MDLs.
Applying a multiplier of 5 would give a possible EQL range from 0.05 to 0.16 (J,g/L. This range
is equal to or greater than the PQL. Thus, the MDL data do not support an EQL below the PQL.
Exhibit 4-16. Analytical Methods for EDB
Method
MDL (nq/L)
MDL x 5 (nq/L)
504.1
0.01
0.05
551.1
0.032
0.16
Source: USEPA, 2009a (upper bound values when ranges are reported)
EPA concluded that all three information sources - PT, MRL, and MDL data - do not support a
reduction of the PQL for EDB. Therefore, EPA did not develop an EQL.
4.1.6 Heptachlor
The MCL for heptachlor equals the PQL of 0.4 (J,g/L. The MCLG is zero, and there is no new
health effects information that suggests a change in the MCLG. Consequently, the threshold for
the occurrence analysis is based on analytical feasibility.
There are only two PT studies with spiked values below the PQL, both of which have passing
rates greater than 75%. The PT data for spiked values greater than the PQL show passing rates
scattered throughout the range from 75 percent to 100 percent (USEPA, 2016b). Because there
are only a couple of studies below the PQL, EPA determined that the PT data do not support
PQL reduction.
34.0%
27.7%
9.6%
6.4%
I I
10.3%
3.8%
I I
8.1%
0.0% 0.0% 0.0% 0.0%

p.\ rfc	\ T. "b	Jfc
, > 0>	<,' V V V n
* o>' <$¦ °
MRL(Llg/L)
4-9

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
As shown in Exhibit 4-17, the modal MRL for heptachlor is 0.04 (J,g/L. Summary data show that
27.9 percent of the MRLs are equal to this value, and 43.4 percent of the MRL values are equal
to or less than it. The percentage of the MRL values that are less than or equal to the mode does
not meet the 80 percent threshold. Therefore, EPA did not base the EQL on the modal MRL.
Exhibit 4-18 shows that more than 99 percent of the MRL values are less than the PQL of 0.4
[j,g/L. Consequently, EPA reviewed MDL values to determine whether they support an EQL
below the PQL.
Exhibit 4-17. Summary of MRL Data for Heptachlor
MRL Value Category
Number of Records
Percentage of Records
All
63,810
100%
Less than mode
9,863
15.5%
Equal to mode (0.04 |jg/L)
17,794
27.9%
Greater than mode
36,153
56.7%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Exhibit 4-18. MRL Distribution for Heptachlor
100.0%
 80.0%
01
_3
5
—! 60.0%
40.0%
£ 20.0%
0.0%
30.0%
12.9%
0.4%
17.4%
21.2%
9.S
3.0% 4-4%
i	1 I I
0.2% 0.2% 0.4%
&>¦
& o

,<3°
oS>
A


A
S>

MRL (jlg/L)
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-19 shows EPA's approved methods for the detection of heptachlor, and the MDLs.
Applying a multiplier of 10 to the MDL values results in a possible EQL range from 0.015 to 3.4
[j,g/L. Three of these values are greater than the PQL. EPA used the highest value below the PQL
(0.05 (J,g/L) and rounded up to 0.1 [j,g/L to establish an EQL. Almost 92 percent of the MRLs in
the Six-Year Review 3 ICR database are less than or equal to this value.
4-10

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-19. Analytical Methods for Heptachlor
Method
MDL (nq/L)
MDLx10(nq/L)
505
0.003
0.03
508
0.0015
0.015
508.1
0.005
0.05
525.2
0.15
1.5
525.3
0.34
3.4
551.1
0.081
0.81
Source: USEPA, 2016b (upper bound values when ranges are reported)
4.1.7 Heptachlor Epoxide
The MCL for heptachlor epoxide equals the PQL of 0.2 (J,g/L. The MCLG is zero, and there is no
new health effects information that suggests a change in the MCLG. Consequently, the threshold
for the occurrence analysis is based on analytical feasibility.
There are no PT studies with spiked values below the PQL. The PT data above the PQL show
passing rates close to 100 percent for most of the studies although one study has a passing rate
less than 75 percent (USEPA, 2016b). Given the lack of data below the PQL, EPA determined
that the PT data do not support a reduction of the PQL.
As shown in Exhibit 4-20, the modal MRL for heptachlor epoxide is 0.02 (J,g/L. Summary data
show that 28.9 percent of the MRLs are equal to this value, and 40.2 percent of the MRL values
are equal to or less than it. The percentage of the MRL values that are less than or equal to the
mode does not meet the 80 percent threshold. Therefore, EPA did not base the EQL on the modal
MRL. Exhibit 4-21 shows that more than 99 percent of the MRL values are less than the PQL.
Consequently, EPA reviewed MDL values to determine whether they support an EQL below the
PQL.
Exhibit 4-20. Summary of MRL Data for Heptachlor Epoxide
MRL Value Category
Number of Records
Percentage of Records
All
63,667
100%
Less than mode
7,184
11.3%
Equal to mode (0.02 |jg/L)
18,370
28.9%
Greater than mode
38,113
59.9%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-11

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-21. MRL Distribution for Heptachlor Epoxide
100.0%
« 80.0%
-j 60.0%
Z 40.0%
£ 20.0%
0.0%
29.2%
26.6%
9.8%
0.1% 0.6% 0.4% 0.0%
17.3%
10.4%
4.9%
y y y y ^ ^ ^ *
Oy	Oy "
MRL (|.lg/L)
0.6%
0^ Cr'
o%' 0>'
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-22 shows EPA's approved methods for the detection of heptachlor epoxide, and the
MDLs. Applying a multiplier of 10 to the MDL values results in a possible EQL range from
0.001 to 2.02 (J,g/L. Two of these values are greater than the PQL and one is approximately the
same. EPA used the highest value below the PQL (0.04 (J,g/L) to establish an EQL.
Exhibit 4-22. Analytical Methods for Heptachlor Epoxide
Method
MDL (nq/L)
MDL x 10 (pxq/L)
505
0.004
0.04
508
0.015
0.15
508.1
0.0001
0.001
525.2
0.13
1.3
525.3
0.0026
0.026
551.1
0.202
2.02
Source: USEPA, 2016b (upper bound values when ranges are reported)
4.1.8 Hexachlorobenzene
The MCL for hexachlorobenzene equals the PQL of 1 (J,g/L. The MCLG is zero, and there is no
new health effects information that suggests a change in the MCLG. Consequently, the threshold
for the occurrence analysis is based on analytical feasibility.
There are several PT studies with a spiked value below the PQL and passing rates greater than
80%, although one study has a passing rate below 75%. Above the PQL, the PT data show
greater than 75 percent passing rates for most of the studies (USEPA, 2009a). EPA determined
that the PT data support reduction of the PQL.
4-12

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
As shown in Exhibit 4-23, the modal MRL for hexachlorobenzene is 0.1 (J,g/L. Approximately
71 percent of the MRL values are equal to or less than the modal value. The percentage of the
MRL values that are less than or equal to the mode does not meet the 80 percent threshold.
Therefore, EPA did not base the EQL on the modal MRL. Exhibit 4-24 shows that more than 99
percent of the MRL values are less than the PQL. Consequently, EPA reviewed MDL values to
determine whether they support an EQL below the PQL.
Exhibit 4-23. Summary of MRL Data for Hexachlorobenzene
MRL Value Category
Number of Records
Percentage of Records
All
62,752
100%
Less than mode
13,418
21.4%
Equal to mode (0.1 |jg/L)
31,338
49.9%
Greater than mode
17,996
28.7%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Exhibit 4-24. MRL Distribution for Hexachlorobenzene
100.0%
80.0%
S
Z 60.0%
tc
2
o
? 40.0%
v
a.
20.0%
0.0%
56.3%
10.0%
3.0%
1.0% 1.0%
0.1%
20.1%
8.3%
0.0% 0.2% 0.1%

O-



c$>
o
\
o
.0^ <9-
o- oy oy oy
A
o- o-
MRL ((ig/L)

-7^
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-25 shows EPA's approved methods for the detection of hexachlorobenzene, and the
MDLs. Applying a multiplier of 10 would give a possible EQL range from 0.01 to 1.3 (J,g/L. One
of these values (1.3 (J,g/L) is greater than the PQL. EPA used the highest value below the PQL
(0.077 (J,g/L) and rounded up to 0.1 [j,g/L to establish the EQL.
4-13

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-25. Analytical Methods for Hexachlorobenzene
Method
MDL (nq/L)
MDLx10(nq/L)
505
0.002
0.02
508
0.0077
0.077
508.1
0.001
0.01
525.2
0.13
1.3
551.1
0.003
0.03
Source: USEPA, 2009a (upper bound values when ranges are reported)
4.1.9 Pentachlorophenol
The MCL for pentachlorophenol equals the PQL of 1 (J,g/L. The MCLG is zero, and a recent
health effects assessment did not indicate a change in the MCLG. Consequently, the threshold
for the occurrence analysis is based on analytical feasibility.
There were no PT studies with spiked concentrations less than the PQL. Above the PQL, passing
rates ranged from 70 percent to 100 percent (USEPA, 2016b). Because of the lack of results
below the PQL, EPA determined that the PT data do not support reduction of the PQL.
As shown in Exhibit 4-26 the modal MRL for pentachlorophenol is 0.04 (J,g/L. Summary data
show that 33.1 percent of the MRLs are equal to this value, and 38.8 percent of the MRL values
are equal to or less than it. The percentage of the MRL values that are less than or equal to the
mode does not meet the 80 percent threshold. Therefore, EPA did not base the EQL on the modal
MRL. Exhibit 4-27 shows that 98 percent of the MRL values are less than the PQL.
Consequently, EPA reviewed MDL values to determine whether they support an EQL below the
PQL.
Exhibit 4-26. Summary of MRL Data for Pentachlorophenol
MRL Value Category
Number of Records
Percentage of Records
All MRL Values
63,532
100%
Value < Modal MRL
3,649
5.7%
Value = Modal MRL (0.04 ug/L)
21,012
33.1%
Value > Modal MRL
38,871
61.2%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-14

-------
\A

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
The only PT study with a spiked concentration below the PQL had a passing rate below 75%.
The passing rates at higher concentrations ranged from 80 percent to 100 percent (USEPA,
2009a). Because of the low passing rate below the PQL, EPA determined that the PT data do not
support reduction of the PQL.
As shown in Exhibit 4-29, the modal MRL for PCBs is 0.5 (J,g/L, which equals the PQL.
Summary data show that 32 percent of the MRLs are equal to this value, and 99.2 percent of the
MRL values are equal to or less than it. As shown in Exhibit 4-30, the MRL data appear to
indicate that there is potential to lower the PQL because most of the MRL values are below the
PQL. Consequently, EPA reviewed MDL values to determine whether they support an EQL
below the PQL.
Exhibit 4-29. Summary of MRL Data for PCBs
MRL Value Category
Number of Records
Percentage of Records
All
32,755
100%
Less than mode
21,999
67.2%
Equal to mode (0.5 |jg/L)
10,478
32.0%
Greater than mode
278
0.8%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
40.0%
« 30.0%
5
_i
C£
2 20.0%
a. 10.0%
0.0%
Exhibit 4-30. MRL Distribution for PCBs
30.5%
32.4%
9.8%
11.5%
9.3%
5.5%
°-7% 0.1% 0.0% 0.0% 0.1%
* 0" *¦>' of	%	"
MRL ((ig/L)
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL. Percentages shown here may not match summary data in the
prior table because of independent rounding.
4-16

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-31 shows EPA's approved method for the compliance monitoring of PCBs (as
decachlorobiphenyl), and the MDL. Applying a multiplier of 10 would give a possible EQL of
0.8 |ig/L, which is greater than the PQL. The MDL data do not support an EQL below the PQL.
Exhibit 4-31. Analytical Methods for PCBs
Method
MDL (nq/L)
MDLx10(nq/L)
508A
0.08
0.8
Source: USEPA, 2009a. This document also reports methods and MDLs for aroclors, but these screening methods
are not sufficient for compliance monitoring.
EPA concluded that although MRL values are generally below the PQL, the combination of PT
and MDL data do not support revision of the PQL for PCBs. Therefore, EPA did not develop an
EQL.
4.1.11 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (Dioxin)
The MCL for dioxin equals the PQL of 3 x 10"5 (J,g/L. The MCLG is zero and there is no new
health effects information that suggests a change in the MCLG. Consequently, the threshold for
the occurrence analysis is based on analytical feasibility.
There is only one PT study. It has a passing rate greater than 75 percent and the spiked
concentration is greater than the PQL (USEPA, 2016b). Given the lack of data, EPA determined
that the PT data do not support revision of the PQL.
As shown in Exhibit 4-32 the modal MRL for dioxin is 5x 10"6 (J,g/L. Summary data show that 52
percent of the MRLs are equal to this value, and 93.3 percent of the MRL values are equal to or
less than it. Because more than 80 percent of the MRL values are less than or equal to 5x 10"6
[j,g/L, EPA identified the mode as the EQL. In Exhibit 4-33, the MRL data indicate that there is
potential to lower the PQL because most of the MRL values are below the PQL. EPA also
reviewed MDL values to determine whether they support an EQL below the PQL.
Exhibit 4-32. Summary of MRL Data for Dioxin
MRL Value Category
Number of Records
Percentage of Records
All
2,620
100%
Less than mode
1,082
41.3%
Equal to mode (5x10-6 |jg/L)
1,362
52.0%
Greater than mode
176
6.7%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-17

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-33. MRL Distribution for Dioxin
100.C
80.0%
5
H 60.0%
£ 40.0%
01



4.4%
I I
^ & cj? <& r§P
o° cP
C> Cy Ov jy
c?v

jF	cF c!
ov ov ov
MRL (flg/L)
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-34 shows EPA's approved method for the detection of dioxin, and the minimum
detection level (MDL). Applying a multiplier of five would give a possible EQL of 2.2x 10"5
[j,g/L, which is less than the PQL, but not as low as the modal MRL. EPA instead used the modal
MRL to establish the EQL.
Exhibit 4-34. Analytical Methods for Dioxin
Method
MDL (nq/L)
MDL x 5 (nq/L)
1613
4.4x10-6
2.2x10-5
Source: USEPA, 2016b
4.1.12 Thallium
The MCL for thallium equals the PQL of 2 (J,g/L. The MCLG is 0.5 (J,g/L, and a recent health
effects assessment did not indicate any changes to the MCLG. Therefore, the threshold for an
occurrence analysis depends on analytical feasibility.
There are no studies with spiked concentrations less than the PQL. The passing rates for the PT
studies above the PQL generally range from 80 percent to 100 percent (USEPA, 2016b). Given
the lack of data below the PQL, EPA determined that the PT data do not support revision of the
PQL.
As shown in Exhibit 4-35, the modal MRL for thallium is 1 (J,g/L. Summary data show that 48.3
percent of the MRLs are equal to this value, and 74.5 percent of the MRL values are equal to or
4-18

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
less than it. The percentage of the MRL values that are less than or equal to the mode does not
meet the 80 percent threshold. Therefore, EPA did not base the EQL on the modal MRL. Exhibit
4-36 shows that more than 99 percent of the MRL values are less than or equal to the PQL.
Consequently, EPA reviewed MDL values to determine whether they support an EQL less than
the PQL.
Exhibit 4-35. Summary of MRL Data for Thallium
MRL Value Category
Number of Records
Percentage of Records
All
75,776
100%
Less than mode
19,855
26.2%
Equal to mode (1 |jg/L)
36,589
48.3%
Greater than mode
19,332
25.5%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Exhibit 4-36. MRL Distribution for Thallium
100.0%
80.0%
>
Z 60.0%
DC
2
*6
? 40.0%
v
a.
20.0%
0.0%
48.3%
9.7%
12.4%
2.7%
0.9%
0.4%
24.8%
0.1% 0.3% 0.0% 0.3%
£¦
> ~0-

\
0>'	o*>'

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-37. Analytical Methods for Thallium
Method
MDL (nq/L)
MDLx10(nq/L)
200.7
no MDL
no MDL
200.8
0.3
3
200.9
1.0
10
Source: USEPA, 2016b (upper bound values when ranges are reported)
4.1.13 Toxaphene
The MCL for toxaphene equals the PQL of 3 (J,g/L. The MCLG is zero, and there is no new
health effects information that suggests a change in the MCLG. Consequently, the threshold for
the occurrence analysis is based on analytical feasibility.
One PT study has a spiked value below the PQL and a passing rate just above 75%. The passing
rates for the PT studies generally exceed 75 percent although the rates are below this threshold
for several studies (USEPA, 2016b). Given the single data point below the PQL, EPA
determined that the PT data do not support reduction of the PQL.
As shown in Exhibit 4-38, the modal MRL is 1 (J,g/L. Approximately 66.5 percent of the MRL
values are equal to or less than the modal value. The percentage of the MRL values that are less
than or equal to the mode does not meet the 80 percent threshold. Therefore, EPA did not base
the EQL on the modal MRL. Exhibit 4-39 shows that more than 99 percent of the MRL values
are less than the PQL. Consequently, EPA reviewed MDL values to determine whether they
support an EQL below the PQL.
Exhibit 4-38. Summary of MRL Data for Toxaphene
MRL Value Category
Number of Records
Percentage of Records
All
57,208
100%
Less than mode
14,117
24.7%
Equal to mode (1 |jg/L)
23,918
41.8%
Greater than mode
19,173
33.5%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-20

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-39. MRL Distribution for Toxaphene
100.0%
80.0%
(A
OJ
>
z 60.0%
Q£
I
o
£ 40.0%
v
u
b.
a
Q.
20.0%
0.0%
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-40 shows EPA's approved methods for the detection of toxaphene, and the MDLs.
Applying a multiplier of 10 would give a possible EQL range from 1.3 to 17 (J,g/L. Three of the
values are greater than the PQL. EPA used the value below the PQL (1.3 (J,g/L) and rounded
down to 1 [j,g/L to establish an EQL.
Exhibit 4-40. Analytical Methods for Toxaphene
Method
MDL (nq/L)
MDL x 10 (pxq/L)
505
1.0
10
508
no MDL
no MDL
508.1
0.13
1.3
525.2
1.7
17
525.3
0.32
3.2
Source: USEPA, 2016b (upper bound values when ranges are reported)
4.1.14 1,1,2-Trichloroethane
The MCL for 1,1,2-trichloroethane equals the PQL of 5 (J,g/L. The MCLG is 3 (J,g/L, and there is
no new health effects information that suggests a change in the MCLG. Therefore, the threshold
for an occurrence analysis depends on analytical feasibility.
There are several studies with spiked concentrations less than the PQL that have passing rates
greater than 90%. The PT results above the PQL also have passing rates in the 90 to 100 percent
42.3%
10.3%
4.5% 3.3% 3.0% 	 3.1%
22.5%
10.9%
0.0% 0.0% 0.1%
«S>	^	0*	.> V
«>' »>' «*•'
¦•>' V

MRL (lig/L)
4-21

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
range (USEPA, 2009a). Given the high passing rates below the PQL, EPA determined that the
PT data support reduction of the PQL.
As shown in Exhibit 4-41, the modal MRL is 0.5 (J,g/L, which is less than the MCLG. More than
99 percent of MRL values are less than the mode. Exhibit 4-42 shows that more than 99.9
percent of MRL values are less than or equal to the MCLG. Although the MRL mode meets
criteria to be an EQL, the mode is less than the MCLG. Consequently, the MCLG is the
appropriate threshold for the occurrence analysis.
Exhibit 4-41. Summary of MRL Data for 1,1,2-Trichloroethane
MRL Value Category
Number of Records
Percentage of Records
All
137,544
100%
Less than mode
18,378
13.4%
Equal to mode (0.5 |jg/L)
117,947
85.8%
Greater than mode
1,219
0.9%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Exhibit 4-42. MRL Distribution for 1,1,2-Trichloroethane
100.0%
80.0%
Z 60.0%
tc
*6
t; 40.0%
01
Q.
20.0%
0.0%
85.9%
0.2%
2.0% SJ/° 3-5% 1.9%
,	i I I I	1 i	L
0.7% 0.1% 0.0% 0.1% 0.0%
&
T, O)	b, <-,	\ t,		^	?
' J?	^	°
MRL (ng/L)
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-43 shows EPA's approved methods for the detection of 1,1,2-trichloroethane, and the
MDLs. Applying a multiplier of 10 would give a possible EQL range from 0.17 to 1 (J,g/L. This
range is below the current MCLG, which further supports use of the MCLG as the threshold in
the occurrence analysis.
4-22

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-43. Analytical Methods for 1,1,2-Trichloroethane
Method
MDL (nq/L)
MDL x 10 (nq/L)
502.2
0.04
0.4
524.2
0.10
1
551.1
0.017
0.17
Source: USEPA, 2009a (upper bound values when ranges are reported)
4.2 MCL Currently Limited by MCLG
For each contaminant addressed in this section, new health effects information indicates potential
to lower the MCLG (USEPA, 2016c). Therefore, EPA's objective was to determine whether this
possible MCLG could be used as the threshold for the occurrence analysis. When it could not be
used, EPA identified an alternative threshold. Exhibit 4-44 illustrates four possible outcomes. In
each case, the blue boxes show that the current MCL equals the current MCLG and the current
PQL is a lower value. The green boxes show new information - the possible MCLG and an
EQL.
Exhibit 4-44. Occurrence Analysis Threshold Selection Scenarios for
A. Possible MCLG is the
occurrence analysis
threshold when
PQL < possible MCLG
B. Possible MCLG is the
occurrence analysis
threshold if data support
EQL< possible MCLG
C. EQL is the occurrence
analysis threshold when
possible MCLG < EQL < PQL
D. PQL is the occurrence
analysis threshold when
possible MCLG < PQL < EQL
Ong/L
The top case (A) shows that the PQL is less than the possible MCLG. In this case, current
analytical feasibility does not limit setting an MCL equal the possible MCLG. Therefore, the
possible MCLG can be the threshold for the occurrence analysis.
Contaminants with New Possible MCLGs
PQL
3Z1
Possible
MCLG
MCLG
MCL
0 M-g/L
EQL

Possible
MCLG
PQL
MCLG
MCL

0 M-g/L
Possible
MCLG
EQL
PQL
MCLG
MCL
v v
Ojig/L
Possible
MCLG
PQL
EQL or
no EQL
MCLG
MCL

4-23

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
The possible MCLG can still be the threshold for the occurrence analysis when it is less than the
PQL. This is possible if EPA can identify an EQL that is less than the possible MCLG (case B).
If, however, data analysis results in an EQL that is greater than possible MCLG, then EPA used
the EQL as the threshold for the occurrence analysis when it was less than the PQL (case C). If
available data did not support deriving an EQL less than the current PQL, then EPA used the
PQL as the occurrence threshold (case D).
As Exhibit 1-2 shows, case A (PQL < possible MCLG) applies to the following contaminants:
cis-l,2-dichloroethylene, fluoride, hexachlorocyclopentadiene, selenium, toluene, and xylene.
For these contaminants, EPA can use the possible MCLG values as occurrence thresholds
without analyzing PT, MRL, or MDL data.
The six remaining contaminants - carbofuran, cyanide, endothall, methoxychlor, oxamyl, and
styrene - require further analysis. To establish an occurrence threshold, EPA used the available
PT, MRL, and MDL data and an analysis method similar to the one in section 4.1.
4.2.1 Carbofuran
The MCL for carbofuran equals the MCLG of 40 (J,g/L. EPA based the promulgated MCLG on a
reference dose (RfD) of 0.005 mg/kg-day. New health effects information indicates a revised
RfD of 0.0003 mg/kg-day. The corresponding possible MCLG is 0.6 [j,g/L (2016c), which is less
than the PQL of 7 (J,g/L. Because the PQL would not allow setting the MCL equal to the possible
MCLG, EPA evaluated how low an occurrence threshold could be.
There are no PT results at spiked concentrations below the PQL. In fact, none of the spiked
concentrations are below 15 (J,g/L, which is two times the PQL. Most of the passing rates are
above 75 percent; only one is less than 75 percent (USEPA, 2016b). Because of a lack of PT data
below the PQL, EPA determined that the PT data do not support reduction of the PQL.
As shown in Exhibit 4-45, the modal MRL for carbofuran is 0.9 (J,g/L, which is less than the
PQL of 7 (J,g/L, but greater than the possible MCLG. Exhibit 4-46 shows that a majority of MRL
values exceed 0.6 (J,g/L, which means the possible MCLG cannot be used for the occurrence
analysis without substantial upward bias in the occurrence estimates. Summary data show that
28.4 percent of the MRLs are equal to the mode, and 56.9 percent of the MRL values are equal to
or less than it. Therefore, a threshold cannot be based on the mode. EPA reviewed MDL values
to determine whether they support a threshold between the possible MCLG and the PQL.
Exhibit 4-45. Summary of MRL Data for Carbofuran
MRL Value Category
Number of Records
Percentage of Records
All
50,018
100%
Less than mode
14,273
28.5%
Equal to mode (0.9 |jg/L)
14,219
28.4%
Greater than mode
21,526
43.0%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-24

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-46. MRL Distribution for Carbofuran
100.0%
in
01
3
IB
>
—J
cc
2
o

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
As shown in Exhibit 4-48, the modal MRL for cyanide is 10 (J,g/L, which is greater than the
potential MCLG of 4 (J,g/L, but less than the PQL of 100 (J,g/L. Exhibit 4-49 shows that
approximately 14 percent of the MRL values are less than 4 (J,g/L, which means the possible
MCLG cannot be used for the occurrence analysis. Summary data show that 42.5 percent of the
MRLs are equal to this value, and 73.1 percent of the MRL values are equal to or less than it.
The percentage of the MRL values that are less than or equal to the mode does not meet the 80
percent threshold. Therefore, EPA did not base the EQL on the modal MRL. Exhibit 4-49 shows
that more than 99 percent of MRL values are less than the PQL. Therefore, EPA reviewed MDL
values to determine whether they indicate an EQL value that is less than the PQL.
Exhibit 4-48. Summary of MRL Data for Cyanide
MRL Value Category
Number of Records
Percentage of Records
All
56,219
100%
Less than mode
17,213
30.6%
Equal to mode (10 |jg/L)
23,865
42.5%
Greater than mode
15,141
26.9%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Exhibit 4-49. MRL Distribution for Cyanide
100.0%
80.0%
V)
_D
> 60.0%
_i
zc
2
o
§ 40.0%
u
I-
0)
Q.
20.0%
0.0%
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-50 shows EPA's method for the detection of cyanide and the corresponding MDL.
USEPA (2016b) identifies additional methods including several newer, proprietary methods that
have lower MDL values. Applying a multiplier of 10 gives a possible EQL of 50 (J,g/L, which is
greater than the potential MCLG, but less than the PQL.
42.5%
16.1%
1.2%
5.9% 6.9%
I	1 I I
0.4%
11.4
7.1%
4.1%	3.5%
I I
0.9%
i>	>	>
\ \
%'
MRL (lig/L)
4-26

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-50. Analytical Methods for Cyanide
Method
MDL (nq/L)
MDLx10(nq/L)
335.4
5.0
50
Source: USEPA, 2016b and NEMI, 2015.
The distribution in Exhibit 4-49 shows that more than 95 percent of the MRL values are less than
or equal to 50 |ig/L, Thus, an occurrence analysis at an EQL of 50 [j,g/L will have a relatively
small degree of bias introduced by the MRL values that are greater than the EQL.
4.2.3 Endothall
The MCL for endothall equals the MCLG of 100 (J,g/L. EPA promulgated the MCLG based on
an RfD of 0.02 mg/kg-day. New health effects information indicates a revised RfD of 0.007
mg/kg-day. The corresponding possible MCLG is 50 [j,g/L (USEPA, 2016c), which is less than
the PQL of 90 (J,g/L. Because the PQL would limit setting the MCL equal to the possible MCLG,
EPA evaluated whether the EQL can be as low as 50 (J,g/L.
There are no PT study results with spiked values below the PQL. Furthermore, some passing
rates for PT studies at spiked concentrations greater than the PQL are below 75 percent (USEPA,
2009a). Because of the lack of data below the PQL, EPA determined that the available PT data
do not support PQL reduction.
As shown in Exhibit 4-51, the modal MRL for endothall is 10 (J,g/L, which is less than the PQL.
Summary data show that 34.3 percent of the MRLs are equal to this value, and 79.6 percent of
the MRL values are equal to or less than it. The mode is also less than the possible MCLG of 50
[j,g/L. Exhibit 4-52 shows that more than 98 percent of the MRL values are less than or equal to
50 (J,g/L. Thus, the MRL data support use of the possible MCLG for the occurrence analysis.
Exhibit 4-51. Summary of MRL Data for Endothall
MRL Value Category
Number of Records
Percentage of Records
All
19,895
100%
Less than mode
9,004
45.3%
Equal to mode (10 |jg/L)
6,833
34.3%
Greater than mode
4,058
20.4%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-27

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
100.0%
80.0%
>
_1 60.0%
ec
*5
¦£ 40.0%
(V
Cl
20.0%
0.0%
Exhibit 4-52. MRL Distribution for Endothall
61.0%
15.5%
1.1% 0.5% 0.9%
0.6%
9.2%
1.2% 1.8%
7.1%
1.1%

\


Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-53 shows EPA's approved method for the detection of endothall, and the MDL.
Applying a multiplier of 10 gives a possible EQL 17.9 (J,g/L, which is less than 50 (J,g/L. Thus,
the MDL data support the use of the possible MCLG as a threshold in the occurrence analysis.
Exhibit 4-53. Analytical Methods for Endothall
Method
MDL (nq/L)
MDL x 10 (pxq/L)
548.1
1.79
17.9
Source: USEPA, 2016b (upper bound value when a range is reported)
Although the PT data do not support a reduction of the PQL, the MRL and MDL data do support
the use of the possible MCLG value of 50 [j,g/L as a threshold for the occurrence analysis.
4.2.4 Methoxychlor
The MCL for methoxychlor equals the MCLG of 40 (J,g/L. The promulgated MCLG was based
on an RfD of 0.005 mg/kg-day. New health effects information indicates a revised RfD of
0.00002 mg/kg-day. The corresponding possible MCLG is 0.1 [j,g/L (USEPA, 2016c), which is
less than the PQL of 10 (J,g/L. Because the PQL would limit setting the MCL equal to the
possible MCLG, EPA evaluated whether the EQL can be as low as 0.1 (J,g/L.
Four PT studies with spiked concentrations less than the PQL had passing rates above 75
percent. There are, however, studies with values greater than the PQL with passing rates at or
below 75 percent (USEPA, 2009a). Nevertheless, because of high passing rates for
4-28

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
concentrations less than the PQL, EPA concluded that the available PT data may support PQL
revision.
As shown in Exhibit 4-54, the modal MRL for methoxychlor is 0.1 (J,g/L, which equals the
possible MCLG. Summary data show that 44.3 percent of the MRLs are equal to this value, and
59.7 percent of the MRL values are equal to or less than it. The percentage of MRL values less
than or equal to the mode does not meet the 80 percent threshold. Therefore, the MRL data do
not support the use of the possible MCLG for the occurrence analysis. Exhibit 4-55 shows that
less than 1 percent of the MRL values are greater than the PQL of 10 (J,g/L. Therefore, EPA
evaluated MDL data.
Exhibit 4-54. Summary of MRL Data for Methoxychlor
MRL Value Category
Number of Records
Percentage of Records
All
70,142
100%
Less than mode
10,788
15.4%
Equal to mode (0.1 |jg/L)
31,060
44.3%
Greater than mode
28,294
40.3%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
Ol
a.
20.0%
0.0%
Exhibit 4-55. MRL Distribution for Methoxychlor
100.0%
80.0%
>
_i 60.0%
cc
E
¦s
t; 40.0%
45.3%
10.6%
1.2% 1.3% 1.1%
0.1%
26.3%
9.3%
2.4% 2.3%
0.1%
£y O- ,0 O
& ^
O"	V
\
o>
\	>	<7
\ 'V v
o-
MRL (j.ig/L)
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-56 shows EPA's approved methods for the detection of methoxychlor, and the MDLs.
Applying a multiplier of 10 would give a possible EQL range from 0.03 to 9.6 (J,g/L. This range
4-29

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
is below the PQL. The highest value, 9.6 (J,g/L, rounds to the PQL. The next highest value rounds
to 1.0 (J,g/L, which is less than the current PQL. Although this value is greater than the possible
MCLG, EPA established an EQL of 1.0 [j,g/L as the threshold for the occurrence analysis.
Exhibit 4-56. Analytical Methods for Methoxychlor
Method
MDL (nq/L)
MDLx10(nq/L)
505
0.96
9.6
508
0.022
0.22
508.1
0.003
0.03
525.2
0.13
1.3
551.1
0.026
0.26
Source: USEPA, 2009a (upper bound values when ranges are reported)
4.2.5 Oxamyl
The MCL for oxamyl equals the MCLG of 200 (J,g/L. The promulgated MCLG was based on an
RfD of 0.025mg/kg-day. New health effects information indicates a revised RfD of 0.0069
mg/kg-day. The corresponding possible MCLG is 10 [j,g/L (USEPA, 2016c), which is less than
the PQL of 20 (J,g/L. Because the PQL would limit setting the MCL equal to the possible MCLG,
EPA evaluated whether the EQL can be as low as 10 (J,g/L.
Two PT studies with spiked concentrations less than the PQL had passing rates at 75 percent.
There are also studies with values greater than the PQL with passing rates at or below 75 percent
(USEPA, 2016b). Because of limited number of studies below the PQL, EPA concluded that the
available PT data do not support PQL reduction.
As shown in Exhibit 4-57, the modal MRU for oxamyl is 2 (J,g/L, which is less than the possible
MCLG. Summary data show that 36 percent of the MRUs are equal to this value, and 85.4
percent of the MRU values are equal to or less than it. The fraction of MRU values less than or
equal to the mode meets the 80 percent threshold. Therefore, the MRU data also support the use
of the possible MCLG for the occurrence analysis. Exhibit 4-58 shows that less than 5 percent of
the MRL values exceed 10 (J,g/L.
Exhibit 4-57. Summary of MRL Data for Oxamyl
MRL Value Category
Number of Records
Percentage of Records
All
49,438
100%
Less than mode
24,422
49.4%
Equal to mode (2 |jg/L)
17,818
36.0%
Greater than mode
7,198
14.6%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
4-30

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-58. MRL Distribution for Oxamyl
100.0%
80.0%
0)
> 60.0%
cc
I
o
c 40.0%
v
Q-
20.0%
0.0%
36.0%
15.3% 14.7%
10.9%
4.1% 2.9% 3.5%
& ^	v" <.-fe
'	<*' O*' N>' ^
fe' vo' 7
MRL ((Ig/L)
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-59 shows EPA's approved methods for the detection of oxamyl, and the MDLs.
Applying a multiplier of 10 would give a possible EQL range from 0.65 to 8.6 (J,g/L. This range
contains the modal MRL and is less than the possible MCLG of 10 (J,g/L. Therefore, EPA
estimated an EQL of 10 |ig/L as a health-based threshold for the occurrence analysis.
Exhibit 4-59. Analytical Methods for Oxamyl
Method
MDL (nq/L)
MDL x 10 (^q/L)
531.1
0.86
8.6
531.2
0.065
0.65
Source: USEPA, 2016b (upper bound values when ranges are reported)
4.2.6 Styrene
The MCL for styrene equals the MCLG of 100 (J,g/L. The promulgated MCLG was based on an
RfD of 0.2 mg/kg-day. New health effects information indicates potential to revise the cancer
classification, resulting in a possible MCLG of zero (2016c). Because the PQL of 5 [j,g/L limits
setting the MCL equal to the possible MCLG, EPA evaluated how low an EQL can be.
There are several PT studies with spiked concentrations below the PQL and passing rates greater
than 90%. PT studies with spiked concentrations greater than the PQL consistently have passing
rates above 75 percent (USEPA, 2009a). Because of high passing rates for concentrations less
than the PQL, EPA concluded that the available PT data support PQL revision.
4-31

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
As shown in Exhibit 4-60, the modal MRL for styrene is 0.5 (J,g/L. Summary data show that 89.5
percent of the MRLs are equal to this value, and 99.5 percent of the MRL values are equal to or
less than it. The fraction of MRL values less than or equal to the mode meets the 80 percent
threshold. Therefore, the MRL data support the use of the modal MRL for the occurrence
analysis. Exhibit 4-61 shows that less than 1 percent of the MRL values exceed 0.5 (J,g/L.
Exhibit 4-60. Summary of MRL Data for Styrene
MRL Value Category
Number of Records
Percentage of Records
All
145,902
100%
Less than mode
14,589
10.00%
Equal to mode (0.5 |jg/L)
130,578
89.50%
Greater than mode
735
0.50%
Note: Percentages may not sum to 100 percent because of independent rounding. Aggregate percentages in the table may
differ from detail in the accompanying chart because of independent rounding.
Source: Six-Year Review 3 ICR database
100.0%
80.0%
I, 60.0%
tc
¦s
r 40.0%
01
a.
20.0%
0.0%
Exhibit 4-61. MRL Distribution for Styrene
89.5%
0.6% 1-896 ^6% |5-°°/° Q,9%
0.3% 0.1% 0.0% 0.0% 0.1%
cv>	 > > J* >
£ v v v n	\ v
'	o1*
o--
MRL (fig/L)
Note: The horizontal axis shows the percent of MRL values in each of 11 discrete ranges. The range with the modal
MRL as an upper bound includes MRL values throughout the range and, therefore, has a greater percentage than the
one reported in the preceding table for the modal MRL.
Exhibit 4-62 shows EPA's approved methods for the detection of styrene, and the MDLs.
Applying a multiplier of 10 would give a possible EQL range from 0.6 to 1.0 (J,g/L. This range
exceeds the modal MRL. Therefore, EPA established an EQL of 0.5 |ig/L based on the modal
MRL as a threshold for the occurrence analysis.
4-32

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 4-62. Analytical Methods for Styrene
Method
MDL (nq/L)
MDLx10(nq/L)
502.2
0.1
1.0
524.2
0.06
0.6
Source: USEPA, 2009a (upper bound values when ranges are reported)
4-33

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
5 Summary
This section provides a summary of the thresholds that EPA derived for analysis of occurrence.
5.1 MCL Currently Limited by PQL
Exhibit 5-1 provides a summary of the information EPA used to develop EQL values in cases of
potential improvements in analytical feasibility. The information includes the PQL values, which
limit current MCL values. The next column indicates whether the PT data indicate potential to
reduce the PQL, i.e., whether there are high passing rates for studies with spiked values below
the PQL. Next is the modal MRL values and the percentage of MRL values that are less than or
equal to the mode. Finally, the table contains the range of EQLs based on the MDL multiplier
method (10 x MDL values; 5 x MDL for 2,3,7,8-TCDD). Bold font indicates information
supporting PQL reduction and EQL development.
Exhibit 5-1. Threshold Information Summary: Potential Improvements in
Analytical Feasibility
Contaminant
PQL
PT Data
Support PQL
Reduction
Modal MRL1
(^q/L)
Range of 10 x MDL
Values2
(^g/L)
Benzo[a]pyrene
0.2
no
0.02 (37%)
0.16 to 2.3
Chlordane
2
no
0.2 (54%)
0.04 to 2.2
DBCP
0.2
no
0.5 (66%)
0.09 to 2.6
DEHP
6
no
0.6 (41%)
13 to 22.5
EDB
0.05
no
0.5 (92%)
0.1 to 0.32
Heptachlor
0.4
no
0.04 (43%)
0.015 to 3.4
Heptachlor Epoxide
0.2
no
0.02 (40%)
0.001 to 2.02
Hexachlorobenzene
1
yes
0.1 (71%)
0.01 to 1.3
Pentachlorophenol
1
no
0.04 (39%)
0.32 to 16
PCBs
0.5
no
0.5 (99%)
0.8
Dioxin
3.0 x 10-5
no
5.0 x 106 (93%)
2.2 x 10-5
Thallium
2
no
1 (75%)
3 to 10
Toxaphene
3
no
1 (67%)
1.3 to 17
1,1,2-T richloroethane
5
yes
0.5 (99%)
0.17 to 1
1.	Based on Six Year 3 ICR dataset. MRL mode is the most frequently reported value. Value in parenthesis is the percent of
MRL values that are less than or equal to the mode.
2.	For each contaminant, the range shown is 10 times the range of MDL values for the EPA-developed analytical methods.
The exception is 2,3,7,8-TCDD, which reflects a multiplier of 5 instead of 10.
The PT data are not sufficient to support PQL reductions for most of the contaminants. This
generally occurs because of the lack of PT studies at spiked concentrations below PQL values.
The three contaminants for which PT data indicate potential to reduce the PQL are
hexachlorobenzene and 1,1,2-trichloroethane.
Generally, the modal MRL values are less than the PQL values, often differing by an order of
magnitude. The exceptions are MRL values that exceed the PQL values for DBCP and EDB, and
5-1

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
the MRL for PCBs, which equals the PQL. Nevertheless, most of these modal MRL values are
not EQL candidates because less than 80 percent of MRL values are less than or equal to them.
Thus, only the MRL modes for dioxin and 1,1,2-trichloroethane meet criteria for EQL
development. The mode for 1,1,2-trichlorethane of 0.5 [j,g/L is less than the MCLG, which is 3
[j,g/L. Therefore, the occurrence threshold for this contaminant is the current MCLG instead of an
EQL.
The MDL data indicate the greatest potential to revise PQL values. The ranges in bold font
include at least one MDL that is less than the PQL. EPA used the MDL data to derive an EQL
for the following contaminants: chlordane, heptachlor, heptachlor epoxide, hexachlorobenzene,
and toxaphene.
The EQL summary in Exhibit 5-2 shows that EPA did not use MDL values to develop EQL
values for three contaminants despite there being an MDL lower than the PQL: benzo[a]pyrene,
DBCP, and pentachlorophenol. For benzo[a]pyrene, an EQL based on the MDL would be the
same as the PQL. For DBCP, an EQL based on MDL data was less than 70 percent of the MRL
values in the database. For pentachlorophenol, EPA did not develop an EQL because six of the
seven MDL values rounded to or exceeded the PQL.
Exhibit 5-2. EQL Threshold Results
Contaminant
PQL
EQL
Basis
Benzo[a]pyrene
0.2
none
Data do not support EQL < PQL
Chlordane
2
1
Based on 10 x MDL
DBCP
0.2
none
Data do not support EQL < PQL
DEHP
6
none
Data do not support EQL < PQL
EDB
0.05
none
Data do not support EQL < PQL
Heptachlor
0.4
0.1
Based on 10 x MDL
Heptachlor Epoxide
0.2
0.04
Based on 10 x MDL
Hexachlorobenzene
1
0.1
Based on 10 x MDL
Pentachlorophenol
1
none
Data do not support EQL < PQL
PCBs
0.5
none
Data do not support EQL < PQL
Dioxin
3.0 x 10-5
5.0 x 10-6
Based on MRL mode
Thallium
2
none
Data do not support EQL < PQL
Toxaphene
3
1
Based on 10 x MDL
1,1,2-T richloroethane
5
3
Based on MCLG (EQL < MCLG)
5.2 MCL Greater than Possible Lower MCLG
Exhibit 5-3 contains summary data for the contaminants for which EPA identified a lower
possible MCLG. The first two data columns contain the possible MCLG and PQL values. Bold
font indicates that seven PQL values are greater than corresponding possible MCLG values.
For the other contaminants, the PQL is lower than the possible MCLG. The MRL information
for these contaminants indicates the percent of MRL values that are less than the possible MCLG
value (instead of an MRL mode). In all instances, almost all of the MRL values are less than the
5-2

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
possible MCLG. The 10 x MDL ranges are generally less than the possible MCLG. Thus, the
possible MCLGs can be used as occurrence thresholds.
Exhibit 5-3. Threshold Information Summary: Possible Lower MCLGs

Possible

PT Data
Six Year 3 MRL
Range of 10 x

MCLG
PQL
Support PQL
Data1
MDL Values2
Contaminant
(nq/L)
(nq/L)
Reduction
(nq/L)
(nq/L)
Carbofuran
0.6
7
no
mode: 0.9 (57%)
0.58 - 5.2
Cyanide
4
100
no
mode: 10 (73%)
50
cis-1,2-Dichloroethylene
10
5
yes
**
**
Endothall
50
90
no
mode: 10 (80%)
17.9
Fluoride
900
500
no
**
**
Hexachlorocyclopentadiene
40
1
no
**
**
Methoxychlor
0.1
10
yes
mode: 0.1 (60%)
0.03-9.6
Oxamyl
10
20
no
mode: 2 (85%)
0.65-8.6
Selenium
40
10
no
**
**
Styrene
0
5
yes
mode: 0.5 (99.5%)
0.6-1.0
Toluene
600
5
yes
**
**
Xylene
1000
5
no
**
**
1.	Based on Six Year 3 ICR dataset. MRL mode is the most frequently reported value. Value in parenthesis is the percent of
MRL values that are less than or equal to the mode.
2.	For each contaminant, the range shown is 10 times the range of MDL values for the EPA-developed analytical methods.
**. Analysis not required because the PQL is less than the possible MCLG.
For six contaminants - carbofuran, cyanide, endothall, methoxychlor, oxamyl, and styrene - the
PQL potentially limits setting an MCL equal to the possible MCLG. The MRL. and MDL
summary information shown in the table indicate whether an EQL could be as low as the
possible MCLG.
The modal MRL. values for two contaminants, endothall and oxamyl, are less than the possible
MCLG values and meet EQL criteria. The MDL values are also less than the possible MCLG.
Therefore, the MRL and MDL data support using the possible MCLG as an occurrence threshold
for these two contaminants.
For styrene, the modal MRL meets the EQL criteria. The modal MRL is greater than the possible
MCLG, however. Therefore, EPA used the EQL instead of the possible MCLG for the
occurrence analysis.
For carbofuran, cyanide, and methoxychlor, the modal MRLs do not meet EQL criteria.
Furthermore, the MDL values did not support use of the respective possible MCLGs as
occurrence thresholds. Nevertheless, EPA could use 10 x MDL values to develop EQLs that are
less than current PQLs. The EQL for carbofuran is 5 (J,g/L; more than 98 percent of the MRL
values are less than 5 (J,g/L. The EQL for cyanide is 50 (J,g/L; 94 percent of the MRL values are
less than this value. Similarly, the EQL for methoxychlor is 1 (J,g/L; 86 percent of the MRL
values less than 1 (J,g/L.
5-3

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
Exhibit 5-4 provides a summary of the occurrence thresholds for this contaminant group. EPA's
analysis indicates that most of the thresholds can be set equal to corresponding possible MCLG
values, regardless of whether PQL values exceed possible MCLGs. In five cases, alternative
values must be used because analytical feasibility will most likely limit setting an MCL equal to
a possible MCLG.
Exhibit 5-4. Occurrence Threshold Results

Possible MCLG
Occurrence

Contaminant
(nq/L)
Threshold
Basis
Carbofuran
0.6
5
EQL based on 10 x MDL
Cyanide
4
50
EQL based on 10 x MDL
cis-1,2-Dichloroethylene
10
10
possible MCLG
Endothall
50
50
possible MCLG
Fluoride
900
900
Possible MCLG
Hexachlorocyclopentadiene
40
40
possible MCLG
Methoxychlor
0.1
1
EQL based on 10 x MDL
Oxamyl
10
10
possible MCLG
Selenium
40
40
possible MCLG
Styrene
0
0.5
EQL based on modal MRL
Toluene
600
600
possible MCLG
Xylene
1000
1000
possible MCLG
5-4

-------
Development of Estimated Quantitation Levels for the
Third Six-Year Review of National Primary Drinking Water Regulations
6 References
U.S. Environmental Protection Agency (USEPA). 2003. EPA Protocol for Review of Existing
National Primary Drinking Water Regulations. EPA 815-R-03-002.
USEPA. 2009a. Analytical Feasibility Support Document for the Second Six-Year Review of
Existing National Primary Drinking Water Regulations. EPA 818-B-09-003.
USEPA. 2009b. Development of Estimated Quantitation Levels for the Second Six-Year Review
of National Primary Drinking Water Regulations. EPA 815-B-09-003.
USEPA. 2009c. EPA Protocol for the Second Review of Existing National Primary Drinking
Water Regulations (Updated). EPA 815-B-09-002.
USEPA. 2016a. The Analysis of Regulated Contaminant Occurrence Data from Public Water
Systems in Support of the Third Six-Year Review of National Primary Drinking Water
Regulations: Chemical Phase Rules and Radionuclides Rules. EPA 810-R-16-014.
USEPA. 2016b. Analytical Feasibility Support Document for the Third Six-Year Review of
National Primary Drinking Water Regulations: Chemical Phase Rules and Radionuclides Rules.
EPA 810-R-16-005.
USEPA. 2016c. Six-Year Review 3 - Health Effects Assessment for Existing Chemical and
Radionuclide National Primary Drinking Water Regulations - Summary Report. EPA 822-R-16-
008.
6-1

-------