UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WSG70
Date Signed: December 16, 1993
Date Revised: June 1, 1998
MEMORANDUM
SUBJECT: Guidance and Clarification on the Use of Detection Limits in Compliance
Monitoring
FROM: James R. Elder, Director
Office of Ground Water and Drinking Water
TO: Water Management Division Directors
Environmental Services Division Directors
Quality Assurance Officers
Regions 1-10
Several Regions and States have requested guidance and clarification on the use of
detection limits in monitoring of drinking water samples for herbicides, pesticides and other
synthetic organic chemicals (SOCs). The basic concern seems to be that some laboratories are
having difficulty in achieving the detection limits specified in the regulations for some SOCs on
a regular basis. As a result, contaminants may occur in public water systems at detectable levels
even though laboratories report "no detect." Though EPA laboratory certification procedures
specify performance criteria for SOCs, detection levels are not one of these criteria. Therefore,
this guidance suggests an appropriate standard for determining when a finding of "no detect"
should be considered "acceptable" for the purposes of the trigger for decreased monitoring.
In the attachment to this memorandum, EPA is suggesting detection limits that a
laboratory must achieve in order to report an acceptable finding of "non-detect." By listing these
concentrations, EPA affirms that results that may occur below these specified concentrations
should not trigger continuing quarterly monitoring. Though existing regulations do not so
require, results of "no detect" from laboratories that cannot achieve the upper confidence limit of
the detection limits should not relieve the public water systems from the requirement for
quarterly monitoring.
Section 141.24(h)(19) of the regulations provides that analysis for drinking water
contaminants shall only be conducted by laboratories that have received certification from EPA
or the State. At a minimum, certified laboratories are required to satisfy criteria, specifically
relating to precision and accuracy. Laboratory certification requirements do not directly specify
detection limits that laboratories must be able to achieve for herbicides, pesticides and other
SOCs.
Section 141.24 requires that analysis for drinking water contaminants be conducted using
the EPA methods or their equivalents. The approved EPA methods require the laboratories to
use specific quality control procedures. One of the quality control procedures is the initial
demonstration of laboratory capability which includes the determination of detection limit. The
detection limit for a given contaminant by a specific method is associated with an inherent
variability of measurement or a confidence interval. The method for determining is specified in
-------
WSG70
40 CAR Part 136 Appendix B. The Appendix includes an explicit procedure for calculating the
lower confidence limit (LCL) and upper confidence limit (UC) based on seven aliquots: The
LCL = 0.64 times the detection limit and the UC = 2.2 times the detection limit.
In §141.24(h)(18), the Agency specifies detection limits for certain synthetic organic
contaminants. If a public water system detects a contaminant, §141.24(h)(7) provides for
continued quarterly monitoring. Although §141.24(h)(18) specifies detection limits for the
purpose of "detection," EPA inadvertently failed to specify limits for "non-detection."
In addition, these detection limits did not have specific confidence intervals associated
with them, even though these intervals are an essential part of the Part 136 Appendix B
procedure. This memorandum explains that, although the published detection limits in
141.24(h)(18) are the applicable standards for "detection," the UC to the detection limit suggests
an appropriate standard for a finding of "no detect" for those contaminants. By suggesting this
standard, this memorandum does not alter pre-existing legal standards or obligations.
In the attachment, we have listed the detection limits from 141.24(h)(18) as well as the
UCS that are appropriate for those herbicides, pesticides and other SOC analysts. If a regional,
state, utility, or a private laboratory has demonstrated that their detection limit for a specific
analyze, calculated by the Appendix B procedure, falls at or below the UC in the attachment,
they should be considered to have achieved the detection limit for the purposes of reporting an
acceptable finding of "no detect." If the laboratory reports "no detect," then the P.S. would be
relieved from continuing quarterly monitoring pursuant to 141.24(h)(7). Of course, the States do
have the option of enforcing more stringent requirements and are not in any way required to
lessen their own requirements to meet the interpretations in this memorandum. This
memorandum does not affect applicable standards when a laboratory reports "detects" at or
above the detection limits in 141.24(h) (18).
This approach may still pose problems for some laboratories and we urge that you give
the States and utilities as much technical assistance as you can in attempting to achieve the
required detection limits or to use their compliance results as part of a waiver application.
I hope this provides clarification on the use of detection limits in compliance monitoring
of drinking water samples. If you have any questions, please call James M. Conlon, Director,
Drinking Water Standards Division on (202) 260-7575. You may also contact Balded L. Bathija,
Ph.D., Chief, Methods and Monitoring Section, on (202) 260-3040.
Attachment
cc: James M. Conlon, DASD
Robert J. Blanco, ENID
Ramona E. Thrived, GWPD
Alan A. Stevens, T.D.
Frederick F. Stiehl, OE
Susan G. Lepow, O.C.
-------
ATTACHMENT
WSG70
Analyze
Alachlor
Atrazine
Benzo(a)pyrene
Carbofuran
Chlordane
Dalapon
Dibromochloropropane
Di(2 -ethylhexl)adipate
Di(2 -ethylhexl)phthalate
Dinoseb
Diquat
2,4-D
Endothal
Endrin
Ethylene dibromide
Glyphosate
Heptachlor
Heptachlor epoxide
Hexachlorobenzene
Hexachlorocyclopentadiene
Lindane
Methoxychlor
Oxamyl (Vydate)
Pentachlorophenol
Picloram
PCBs (as decachlorobiphenyl)
MCLG
mg/L
0.002
0.003
zero
0.04
zero
0.2
zero
0.4
0.006
0.007
0.02
0.07
0.1
0.002
zero
0.7
zero
zero
zero
0.05
0.0002
0.04
0.2
zero
0.5
zero
MCL
mg/L
0.002
0.003
0.0002
0.04
0.002
0.2
0.0002
0.4
0.006
0.007
0.02
0.07
0.1
0.002
0.00005
0.7
0.0004
0.0002
0.001
0.05
0.0002
0.04
0.2
0.001
0.5
0.0005
FR Detection
Limits
mg/L
0.0002
0.0001
0.00002
0.0009
0.0002
0.001
0.00002
0.0006
0.0006
0.0002
0.0004
0.0001
0.009
0.00001
0.00001
0.006
0.00004
0.00002
0.0001
0.0001
0.00002
0.0001
0.002
0.00004
0.0001
0.0001
UCL
mg/L
0.00044
0.00022
0.000044
0.00198
0.00044
0.0022
0.000044
0.00132
0.00132
0.00044
0.00088
0.00022
0.0198
0.000022
0.000022
0.0132
0.000088
0.000044
0.00022
0.00022
0.000044
0.00022
0.0044
0.000088
0.00022
0.00022
-------
WSG70
Simazine 0.004
Toxaphene zero
2, 3, 7, 8-TCDD (Dioxin) zero
2, 4, 5-TP 0.05
Aldicarb*
Aldicarb sulfoxide*
Aldicarb sulfone*
(0.001)
(0.001)
(0.001)
0.004
0.003
0.00000003
0.05
(0.003)
(0.004)
(0.002)
0.00007
0.001
0.000000005
0.0002
0.0005
0.0008
0.0003
CFR
Numbers
0.000154
0.0022
0.000000011
0.00044
0.0011
0.00176
0.00066
New
Suggested
"Non-detect"
*MCLGs and MCLs for these contaminants have been stayed pending further rulemaking.
------- |