Tuesday
October 11, 1988
Part II
Environmental
Protection Agency
40 CFR Part 264
Statistical Methods for Evaluating
Ground-Water Monitoring From
Hazardous Waste Facilities; Final Rule
-------
39720 Federal Register / Vol. 53, No. 198 / Tuesday, October 11, 1988 / Rules and Regulations
ENVIRONMENTAL PROTECTION
AGENCY
40 CFR Part 264
[FRL-3356-2]
Statistical Methods for Evaluating
Ground-Water Monitoring Data from
Hazardous Waste Facilities
AGENCY: U.S. Environmental Protection
Agency. '
ACTION; Final rule.
SUMMARY: EPA regulations, promulgated
under the Resource Conservation and
Recovery Act (RCRA), require {{round-
water monitoring to detect
contamination of ground water at
permitted hazardous waste land
disposal facilities (40 CFR § 264.90 et
seq. Part 284 Subpart F). These
regulations specify that a statistical
method must be used to evaluate the
presence or increase of contamination.
Due to problems associated with the use
of Cocnran's Approximation to the*
Behrens-Fisher Student's West (CABF)
as such as statistical method, EPA
proposed amendments to the Part 264
Subpart F regulations on August 24,1987
(52 FR 31948). These amendments, which
EPA is today finalizing, specify five
different statistical methods that are
more appropriate to ground-water-
monitoring than the CABF method. The
amendments finalized today also outline
sampling procedures and performance
standards that are designed to help
minimize the event that a statistical
method will indicate contamination
when it is not present (Type I error), and
fail to detect contamination when it is
present (Type II error).
DATE: These final regulations become
effective April 11,1989, pursuant to
RCRA section 3010(b).
ADDRESSES: The official docket-for this-
rulemaking (Docket No. F-88-8GWF-
FFFFFJ is located in Room MLG100. U.S.
Environmental Protection Agency, 401M
Street SW« Washington, DC 20460, and
is available for viewing from 9:30 a.m. to
3:30 p.m., Monday through Friday,
excluding legal holidays. The public
must make an appointment to review
. docket materials. Call (202) 47S-9327 for
appointments. The public may copy a
maximum of 100 pages of docket
materials at no cost. Additional copies
costS.15/page.
FOR FURTHER INFORMATION CONTACT:
For general information contact: RCRA/
Superfund Hotline, Office of Solid
Waste (WH-S63C), U.S. Environmental'
Protection Agency, 401M Street. SW.,
Washington, DC 20460. telephone (800)
424-9346 or (202) 382-3000. For technical
information contact Jini Brown, (202)
382-4658.
SUPPLEMENTARY INFORMATION:
Preamble Outline
I. Authority
II. Background
A. Concerns About Existing Standards
B. Suggested Changes Published in NPRM
ID'. Public Comments on NPRM
A. Comments Solicited by EPA
B. Other issues
IV. Implementation
V. Miscellaneous
A. Deletion of Proposed 5 264.97{i)(3)
B. Demonstrations of Error Caused by Data
Variability
VL General Description of Statistical
Methods
A. Analysis-of-Variance
B. Tolerance Intervals
C. Prediction Intervals
D. Control Charts
VII. Glossary
VUI. Regulatory Arialysis
A. State Authority
B. Regulatory Impact Analysis
C. Regulatory Flexibility Act
DC. List of Subjects in 40 CFR Part 264
I. Authority
These regulations are issued under the
authority of sections 1006,2002(a). 3004,
and 3005 of the Solid Waste Disposal
Act (SWDA), as amended by the
Resource Conservation and Recovery
Act of 1976 (RCRA), as amended-(42
USC 6905, 6912(a), 6924', and 6925).
II. Background
Subtitle C of the Resource
Conservation and Recovery Act of 1976
(RCRA) creates a comprehensive
program for the safe management of
hazardous waste. Section 3004 of RCRA
requires owners and operators of
facilities that treat, store, or dispose of
hazardous waste to comply with
standards established by EPA that are
"necessary to protect human health and
the environment." Section 3005 provides
for implementation of these standards
under permits issued to owners and
operators Tjy EPA or authorized States.
Section 3005 also provides that owners
and operators of existing facilities that
apply for a permit and comply with
applicable notice requirements may
operate until a permit determination is
made. These facilities are commonly
known as "interim status" facilities.
Owners and operators of interim status
facilities also must comply with
standards set under section 3004.
EPA promulgated ground-water
monitoring and response standards for
certain land-based interim status
facilities in 1980 (45 FR 33232, May 19.
1980), codified in 40 CFR Part 265,
Subpart F, and permitted facilities in
1982 (47 FR 32274, July 26,1982), codified
in 40 CFR Part 264, Subpart F. These
standards establish programs for
protecting ground water from releases of
hazardous wastes from certain landfills,
surface impoundments, and land
treatment units, and, in the case of
permitting standards, to waste piles as
well. (See 40 CFR 264.90(a)(2) and
265.90(a)). Facility owners and operators
are required to sample ground water at
specified intervals and to use a
statistical procedure to determine
whether or not hazardous wastes or
constituents from these units are
contaminating ground water. As
explained in more detail below, the
Subpart F regulations regarding
statistical methods used in evaluating
ground-water monitoring data have
generated criticism. EPA is today
finalizing amendments to the Part 264
Subpart F regulations governing
statistical methods for RCRA permitted
facilities proposed August 24,1987 (52
FR 31948) to respond to these concerns.
Due to the fact that most interim status
land disposal facilities are expected to
receive RCRA permits by November
1988, EPA is not amending the Part 265
Subpart F regulations governing
statistical methods at interim status
facilities.
A. Concerns About Existing.Standards
The current Part 264 regulations
provide that the Cochran's
Approximation to the Behrens Fisher
Student's t-test (CABF) or an alternate
statistical procedure approved by EPA
be used to determine whether there is a
statistically significant exceedance of
background levels, or other allowable
levels, of specified chemical parameters
and hazardous waste constituents.
Although the existing 40 CFR Part 264
regulations have always provided
latitude for the use of an alternate
statistical procedure, concerns have
been raised that the CABF statistical
procedure in the current regulations may
not be appropriate to ground-water
monitoring. It has been pointed out that:
(1) The replicate sampling method
required under the current Part 264
Subpart F regulations is not appropriate
for the CABF procedure, (2) the CABF
procedure does not adequately consider
the number of comparisons that must be
made under these regulations, and (3)
the CABF does not control for seasonal
variation. Specifically, the concerns are
that the CABF procedure could result in
"false positives" (Type I error),
instances where contamination is
falsely indicated at the site. False
positives may require an owner or
operator unnecessarily to collect
additional ground-water samples, to
-------
Federal Register / Vol. 53. No. 196 / Tuesday. October 11. 1988 / Rules and Regulations 39721
>r characterize ground-water
"xajid to apply for a permit
on to begin more
"\e monitoring or corrective
OoNpit modification is then
. In addition, there
procedure may
CABFV
negatives\v
which are oix
statistical comj.
variable due to I
analytical,
. >e actual
*ed. The
B. Suggested Changes^
NPRM
As a result of these concern;
v
amending both the statistical mev 4
and the sampling procedures of the
regulations, by requiring (if necessary,
that owners or operators more \
accurately characterize the
hydrogeology and potential
contaminants at the facility, and by
including in the regulations performance
standards which all the statistical
methods and sampling procedures must
meet Statistical methods and sampling
procedures meeting these performance
. standards should have a low probability
of indicating contamination when it is
not present and of failing to detect
contamination that actually is present
III. Public Comments on the NPRM
A. Comments Solicited by EPA
In a notice of proposed rulemaking
(NPRM) issued on August 24,1987 (52
FR 31948), EPA solicited comments on
alternative statistical methods to the
CABF method as well as general
information that would help evaluate
approaches to determining if a facility is
contaminating the ground water.
1. Power of a Statistical Test
EPA first invited comments on the
issue of whether the power of a
statistical test should be specified
numerically. In the NPRM, however,
EPA stated that it was its view that a set
of specific numerical performance
standards that would achieve the proper
balance between false positives and
false negatives is not possible because it
would involve specifying every possible
minimum magnitude of difference for
each contaminant at each site. This
requires specifying concentration level
changes at each site to which the
statistical test must be sensitive. This is
not possible due to the current state of
knowledge about ground-water
contamination.
A consensus of the commenters
acknowledged this difficulty. Once
commenter offered an excellent
summary of the problems associated
with setting a numericaLperfonnance
standard: *The power of a statistical
test is not a value, but a function
involving sample sizes, sampling plans,
the statistical models on which the test
is based, the Type I error level, the
inherent variability and correlation
structure of the measurements, and the
amount of increase in the level of the
.constituent at which the power is
evaluated."
However difficult to quantify, the
Agency agrees with the consensus of the
"<">."n *$>** ^Njcal test can be unproved by a
&& ^U/" i?f methods, such as adequately
e hydrogeology and the
w
%;
i, and reducing
* ^ity by using proper
» f^ ^>>
<&fc
*$A
c numeric
of a
asstk -^
than e\ -3^'
standard^ -
statistical U
systems appKs
monitoring as rk
performance stanu
other components 01
2. Methods to Analyzed
Limit Data \
EPA also invited public co\
the methods available for analj/,
data where the background level\
constituent is either below the deteV
limit of the analytical method used or\
recorded as a trace level of the
constituent. This problem is often
encountered with (although not limited
to) synthetic organic compounds (e^.,
volatile and semivolatile organic
compounds). Many of these compounds
do not occur naturally hi ground water,
and therefore are not detected during
background sampling. This makes
comparing downgradient (compliance
well) concentrations with background
levels of these compounds especially
dificult
Several commenters requested EPA to
consider establishing national baseline
values for compounds that do not occur
naturally in ground water, and as a
result are frequently recorded as below
the limit of analytical detection in
background monitoring wells.
Specifically, the commenters suggested
that EPA conduct a round-robin study
involving several different certified
chemical laboratories to establish
national baseline values for these
' compounds.
The Appendix IX rule (52 FR 25942,
July 9,1987) listed practical
quantification limits (pql's) that were
established from 'Test Methods for
Evaluating Solid Waste" (SW-846). SW-
846 is the general RCRA analytical
methods manual, currently in its third
edition. The pql's listed were EPA's best
' estimate of the practical sensitivity of
the applicable method for RCRA ground-
water monitoring purposes. However,
some of the pql's may be unattainable
because they are based on general
estimates for the specific substance.
Furthermore, due to site-specific factors,
these limits may not be reached. For
these reasons the Agency feels that the
pql's listed in Appendix EX are not
appropriate for establishing a national
baseline value for each constituent for
determining whether a release to ground
water has occurred. Instead, the pql's
are viewed as target levels that
chemical laboratories should try to
achieve hi their analyses of ground
water. In the event that a laboratory
cannot achieve the suggested pql, the
owner or operator may submit a
justification stating the reasons why
these values cannot be achieved (e.g.,
specific instrument limitations). After
reviewing this justification, the Regional
Administrator may choose to establish'
facility-specific pql's based on the
technical limitations of the contracting
\boratory.
^us EPA is today clarifying
97(h) to allow owners or operators
"86 facility-specific pql's. These
be used with the statistical
^d in § 264.97(h) (e.g..
ANOVA), to comply with
,/on approval of the
..ninistratar. In addition,
V^ /o adding language to
§ 2*^ /$}($} to state that any pql
approved by the Regional Administrator
must be the lowest concentration level
that can be reliably achieved within
specified limits of precision and
accuracy during routine laboratory
operating conditions that are available
to the facility.
EPA believes it to be necessary that
the owner or operator utilize a statistical
method to account for data below the
detection limit of the analytical method
used. Although several commenters
submitted methods which could be
specified in the regulations, EPA
believes that no single method is
appropriate at all facilities. Accordingly,
EPA believes it is necessary to evaluate
the appropriateness of each method on a -
case-by-case basis. The fifth
performance standard of today's final
rule, found at § 264.97(i)(5), reflects this
belief by requiring that the statistical
-------
39722 Federal Register / Vol. 53, No. 196 / Tuesday. October 11, 1988 / Rules and Regulations
method chosen include procedures to
evaluate data that is below the limit of
analytical detection. Statistical methods
that are commonly applied use tests of
proportions, prediction and tolerance
intervals, and procedures which
characterize censored data
distributions. Examples of these
methods will be provided in a guidance
document which will be available later
this year.
3. Establishing Background
Concentrations With Downgradient
Wells
EPA also sought comments on the
utility of allowing the use of samples
from downgradient wells to establish
background concentrations at newer
units that have had no opportunity to
contaminate the ground water and that
are located in areas with little potential
to be influenced by external sources
unrelated to the unit. Four commenters
addressed this issue and supported the
Agency's proposal to use downgradient
wells to establish background
concentrations of constituents in
selected circumstances.
EPA realizes that this option is not a
new feature of the Part 254 Subpart F
regulations (see previous § 264.97{g)(3)).
EPA believes that discussion of this
option in this rulemaking is appropriate
because of the emphasis in today's
regulations on choosing statistical
methods and sampling procedures
appropriate for individual facilites.
Using downgradient wells to establish
background concentrations reduces
some of the components of spatial
variability for any statistical method
employed. In addition, unlike the CABF
method, the control chart statistical
method specified in today's
amendments can accommodate intra-
well comparisons. An intra-welH
comparison is a method that establishes
background concentrations from an
individual well, and compares future
monitoring data obtained from the well
to its own background concentration.
An mtra-well comparison method is
necessary if downgradient wells are to
be used to establish background
concentrations.
B. Other Issues ,
Other comments on the NPRM were
received from the public on a wide
variety of issues raised by the proposal,
some of which are discussed below.
These comments and the Agency's
formal responses are available through
the official docket for this rulemaking.
, 1. Guidance Document
Many commenters addressed the fact
that the Agency did not make available
its planned guidance document on
statistical methods during the comment
period on the NPRM. "the respondents
stated that more details on the proposed
statistical methods and procedures for
handling censored data, correlations,
and seasonal variation were needed
before complete comments could be
given. However, EPA requested
comments on the regulation and not on
specific protocols of the statistical
methods. Therefore, a guidance
document detailing the statistical
methods should not have been
necessary to review the NPRM.
However, a draft guidance document
addressing these concerns will be issued
after finalization of this rule.
2. Data Distribution Assumptions
Commenters also addressed the
assumptions made in the first
performance standard, or § 264.97(i)(l)
of the proposed rule, concerning the
distribution of data. As written, this
proposed performance standard
assumed that the data obtained through
ground-water monitoring are normally
distributed at all sites. Proposed
§ 264.97(i)(l) required that a goodness of
fit test be conducted to demonstrate that
the normal distribution assumption is
not appropriate to the data. Some
respondents suggested that owners and
operators be required to justify the
normal distribution of their data, just as
they are required to justify a different
data distribution (e.g., lognormal, non-
normal, etc.) under the proposed
performance standard. Others
commented that it would be more
reasonable to assume a lognormal
distribution. Still others suggested that
EPA should replace the assumption of
normality with a requirement that the
statistical method, including any
preliminary transformations, be
appropriate for the background data or
data expected on the basis literature.
EPA is retaining the assumption of
normality in the data distribution in
today's final rule (§ 264.97(i)(l)) because
many of the statistical procedures cited
in the regulation are robust for data that,
while not normally distributed, do not
significantly violate the normal
distribution assumption. Thus EPA
believes it is reasonable to assume
normality of data and to only require
demonstrations where the owner or
operator wishes to use a distribution-
free theory test. The statistical test will
be appropriate for most data under this
assumption and the owner or operator
will not in all cases be required to go
through the extra step of determining the
distribution of ground-water data. The
regulation's first performance standard
provides that the owner or operator may
use a distribution-free theory test or a
transformation, provided he or she
demonstrates that the data are
inappropriate for a normal theory test.
EPA requires this showing to prevent
increases in the Type II error rate, a
possible result of using distribution-free
theory tests or transformations in
inappropriate circumstances. When the
Type II error rate increases,
environmentally significant
contamination may go undetected. A
demonstration of a data distribution
may include both graphics and literature
as well as the conventionally used
statistical methods.
3. Obligation of'Owner/Operator to
Propose Statistical Methods and
Sampling Procedures
Some commenters opposed a
provision in the proposed rule stating
that the Regional Administrator is
responsible for specifying the sampling
procedures and frequencies, and the
statistical methods that are required
under § 264.97 (General Ground-Water
Monitoring Requirements). The
commenters stated that the regulated
party, not the Regional Administrator,
should be responsible for designing and
proposing the statistical methods and
. sampling procedures. EPA agrees that it
would be more effective to ask the
owner or operator to undertake initial
design of methods and procedures.
Therefore, EPA has changed the
language of §§ 264.97 (g) and (h) to
require the owner or operator to propose
a respective sampling procedure and
statistical method which must then be
approved by the Regional
Administrator.
4. Data Variability and Sampling
Procedures
Commenters also addressed the need
for specific methods to handle
correlated data (see autocorrelation in
glossary) and the problems caused by
temporal and spatial variation. EPA
recognizes the possibility of the
correlation of errors, and temporal and
spatial variation affecting the data sets
and believes that certain provisions in
today's final rule enable owners and
operators to reduce these sources of
errors and control for data variability.
Choosing an appropriate sampling
interval that spans a sufficient amount
of time to allow one to obtain an
independent ground-water sample will
help reduce the effects of
autocorrelation. Under § 264.98(d) and
§ 264.99(f), owners and operators have
the latitude to choose such an interval,
provided that four samples .are taken
from each well at least semiannually.
-------
Federal Register / Vol. 53, No. 196 / Tuesday. October 11, 1988 / Rules and Regulations 39723
Also, sampling both background and
compliance wells at the same-point-in-
time should reduce temporal effects.
One-point-in-time comparison sampling
is also allowed under § 264.9.8(d) and
§ 264.99(f), which require that all wells,
background and compliance, be sampled
during the specified sampling interval.
The current regulations prevented
owners -and operators from performing
one-point-in-time comparisons by
requiring that background
concentrations be established prior to
the monitoring of compliance wells in
detection and compliance monitoring.
(See § 264.97(g).) To better characterize
spatial variability, an owner or operator
may wish to install and sample from
multiple background and compliance
wells. Additionally, if sufficient data are
available, statistical procedures such as
moving averages, in which a background
value is established by continually
updating the data, and trend analysis
may be used to reduce seasonal and
temporal effects.
5. Procedures at Interim Status Facilities
Some respondents requested that the
same regulatory changes should be
made in the Part 265 ground-water
monitoring regulations for interim status
facilities as were made for permitted
facilities. They added that all the
reasons for replacing the Student's t-test
at a permitted facility apply with equal
force at an interim status facility.
As discussed above this rule is
expected to be finalized by September
1988, and to become effective six
months after the date of promulgation,
or March 1989. By November 1988, the
majority of interim status land disposal
facilities are expected to be either
permitted-or closed. In the event .that a
significant number of facilities are still
operating after this date, EPA will
assess the need to amend 40 CFR 265 as
appropriate. The Agency recognizes that
some facilities may be subject to interim
status due to new listings of RCRA
hazardous wastes. EPA intends to move
expeditiously to permit these facilities
so that they may take advantage of
today's amendments to the statistical
procedures at permitted facilities under
Part 264.
6. Determining Background
Concentrations.
Determining the background
concentration of constituents was
another topic addressed by commenters.
These commenters argued that the
current regulation, which limits
background determinations to data
collected during a single year, is too
restrictive. Section 264.97(g]{l) of the
current regulations states that
background ground-water quality for a
monitoring parameter or constituent in
detection monitoring must be based on
data from quarterly sampling of
background (or, in certain
circumstances, compliance wells) for
one year. EPA agrees with this position.
As discussed above, EPA is therefore
requiring that monitoring under
§ 264.98{d) and § 264.99(f) be performed
at all wells, including background and
compliance wells. Thus the background
determination will not be limited to data
collected during a single year prior to
monitoring compliance wells as is
currently set forth in § 264.98(g)(2). This
will allow the mean concentration of a
constituent to be used in one-point-in-
time comparisons between background
and compliance wells, or to be used to
establish a "moving average" in the
background well data base for
comparison to the compliance well
values at a frequency required in the
facility permit.
EPA encourages owners and
operators to determine the
concentrations of a constituent in these
samples through the use of one-point-in-
time comparisons between background
and compliance wells. Some facility
owners or operators may want to use
the concentrations to establish a
"moving average" in the background
well data base for comparison to the
compliance well values at the frequency
required in the facility permit. While
using several background values to
establish a "moving average" is an
acceptable method of analysis, it
increases the number of degrees of
freedom, making this method more
sensitive to changes in constituent
concentrations. Further, this method
does not account for seasonal variation
as effectively as one-point-in-time
comparison procedures. Therefore, most
owners or operators should find one-
point-in-time comparisons to be a
preferred method of analysis. This
approach will help reduce the
components of seasonal variation by
providing for simultaneous comparisons
between background well and
compliance well monitoring data.
7. Sampling Required by Proposed
§264.98(g)(2)
Many commenters were opposed to
the provisions in proposed § 264.98(g)(2)
for detection monitoring which required
the owner or operator to, upon obtaining
statistically significant evidence of
contamination," sample the ground
water in all monitoring wells at the
waste management area of concern and
determine if there is a statistically
significant difference between the
compliance and background levels for
concentration of all constituents
identified in Appendix IX of Part 264."
The respondent's primary point of
concern was that this provision would
require extensive sampling and
statistical analysis to determine
background concentrations for all of the
Appendix DC compounds prior to
obtaining statistically significant
evidence of contamination at a facility.
Under the current regulation an owner
or operator is required only to determine
whether any Part 264 Appendix IX
constituent is present, and at what
concentration (§ 264.98{h)(2)). EPA has
reviewed this requirement and has
, found it to be one of technical oversight.
Therefore, acting in accordance with the
comments received on this matter, EPA
is replacing the proposed sections with
the previously existing language of
§ 264.98(h)(2); that is "immediately
sample the ground water in all
monitoring wells and determine whether
constituents identified in the list in
Appendix IX of Part 264 are present and,
if so, at what concentration."
8. Type I Experimentwise Error Rate
Many commenters addressed the
second performance standard finalized
in today's rule as § 264.97(i)(2). For
individual well comparisons in which a
compliance well is compared with
background, § 264.97(i](2) specified that
the Type I error level shall be no less
than 0.01 for each testing.period. In
other words, the probability of the test
resulting in a false positive is no less
than 1 in 100. EPA believes that this
significance level will sufficiently limit
the false positive rate and has retained
this provision of the second performance
standard in today's rule. Section
264.97(i)(2) also accounted for those
owners and operators of facilities that
have an extensive network of ground-
water monitoring wells who find it more
convenient to use a multiple well
comparisons procedure. Multiple
comparisons procedures control the
experimentwise error rate for
comparisons involving multiple
background and compliance wells.
Under today's final version of the
second performance standard, if this
method is used, the Type I experiment-
wise error rate for each constituent shall
be no less than 0.05 for each testing
period. Here, the probability of the test
resulting in a false positive is no less
than 5 in 100. Again, EPA is limiting the
Type I error rate for the purpose of
controlling the Type II error rate. In the
multiple well comparisons procedure, if
the overall test is shown to be
significant, then individual well
contrasts are performed to identify
-------
39724 Federal Register / Vol 53. No. 196 / Tuesday October 11. 1988 / Rules and Regulations
which differences are statistically
significant In conducting a multiple well
comparisons procedure, if the owner or
operator chooses to use a t-statiutic
rather than an F-statistic, the individual
well Type I error level of no less than
0.01 must be maintained. This provision
should be considered if a facility owner
or operator wishes to use a procedure
that distributes the risk of a false
positive evenly among all monitoring
wells and monitoring parameters at the
facility. This is reflected in the second
performance standard which requires
that if a multiple comparisons procedure
is used, the Type I error of no leas than
0.01 for individual well comparinons
must be maintained.
Several commenters expressed
concern that in prescribing a Type I
error rate of no less than 0.01 (0.05 for
multiple well comparisons) this second
performance standard would lend to
high false positive rates. Owners and
operators should note, however, changes
in the language of § 264.97{i)(2) of
today's final rule which specifies that
this Type I error level applies per single
testing period, not for the entire
operating life of the facility. Muttiple-
unit facility owners and operators may
generate a large number of comparisons
due to the large number of wells at their
facilities, and may potentially face a
Jorge number of false positives in their
'data evaluation. These owners and
operators are encouraged to implement
a unit-specific data evaluation approach
if they wish to keep the overall false
positive rate down to a lower level.
EPA realizes, however, that there still
may be situations where facilities will
generate large false positive rates,
especially those that monitor for a large
number of constituents over several
monitoring wells. Here, if the otvner or
operator suspects that a detection is a
false positive, he or she may wish to
make a demonstration under
§ 204.98(81(6) or S 254.99(5) of today's
final rule.
In these cases, a determination of
whether a leak has occurred may in
many cases be based on the Regional
Administrator's evaluation of the
hydrogeology, geochemistry, climatic
factors, and the relative magnitude of
the concentration of the constituents .
along with the results of the statistical
test In evaluating the relative
magnitude of the concentration of the
constituents, for example, if the
exceedanceis based on an observed
compliance well value that has the same
relative magnitude as the practical .
quantification limits (pql) or tho
background level, the exceedance is
more likely a fake positive and further
sampling and'testing may be
appropriate. If, however, the background
or an action level is exceeded by an
order of magnitude in any sample, then -
the exceedance may indicate a release
from the facility.
Many commenters stated that it was
hard to understand how to apply this
second performance standard
(especially-the Type I error level of 0.01
for individual well comparisons) to
control charts, tolerance intervals, and
prediction intervals. Several
commenters suggested mat, in setting a
Type I error level for control charts, EPA
should be consistent with, the research
projects that were conducted by the
Agency's laboratories. Specifically, the
commenters requested that EPA utilize a
combined Shewhart-CUSUM control
chart scheme to evaluate ground-water
monitoring data. °
EPA agrees that § 264.97(i)(2), or the
second performance standard, is not
directly applicable to control charts,
tolerance intervals and prediction
intervals. Accordingly, the Agency is
specifying hi § 264.97(i)(2) that this
performance standard does not apply to
these three statistical methods. EPA
would nevertheless like to retain these
statistical methods and has therefore
attempted to specify, in today's final
rule, when their use is appropriate as
well as applicable performance . .
standards.
Control charts have been employed by
industry for many quality control
applications. Because of their
widespread use, EPA is generally
allowing their use as a statistical
method for ground-water monitoring
under § 264.97(h)(4), so long as they
comply with the performance standard
specified in § 264,.97(i)(3). There are a
variety of control charts available for
applications to ground-water
monitoring. Each procedure has different
parameters that need to be specified
based on various features of the data
such as .the mean, variance, sample size,
decision interval value (h), reference
value (k), and control limits. EPA does
not believe it to be appropriate to
specify numerical values for these
parameters in a performance standard
in today's final rule, because they are-
dependent on site-specific factors such
as the constituents being monitored for
and the facility's hydrogeology..
Therefore, the Agency is requiring the -
owner or operator to propose values for
these parameters that are appropriate
for the type of control chart used. If the
Regional Administrator finds the type of
control chart and the associated
parameters to be appropriate for the
facility that proposed them and . ,
protective of human health and the
environment, then he or she will
approve and include them in the
facility's operating permit This is
reflected in the third performance
standard of today's final rule.
In evaluating the control chart, the
owner or operator should also consider
ther average run lengths, in and out of
control, before a decision regarding a
suspected release is made. Guidance
addressing control charts will be issued
after finalization of this rule.
Tolerance intervals and prediction
intervals have not been widely used by
the Agency to evaluate ground-water
monitoring data. However, the Agency
is aware of recent publications that
have employed these statistical methods
to evaluate ground-water monitoring
data, especially in evaluating certain
classes of chemical compounds (e.g.,
volatile organic compounds). Several
commenters suggested' that the Agency
incorporate this research into today's
final rule, noting that these procedures
may be the best way to evaluate data
that is below the limit of analytical
. detection.
While EPA does not believe it
appropriate to specify the confidence
levels for prediction and tolerance '
intervals (or in the case of tolerance-
intervals the percentage of the
. population that the interval must
contain) in today's final rule, the Agency
is nevertheless adding a performance
standard relating to the use of these
, procedures hi today's final rule. Because
the parameters of confidence levels and
population percentages may vary due to
site-specific factors, § 264.97(i)(4) states
that the facility owner or operator must
submit parameters that are protective of
human health and the environment to
the Regional Administrator for approval.
In evaluating these parameters, the
Regional Administrator may consider
the number of samples' in the
background data base and the range of
the concentration values for each
constituent of concern.
9. Time Intervals for Ground-Water
Sampling
There was some confusion expressed
hi the comments regarding the time
intervals within which ground-water
samples are to be collected (i.e.,
sampling procedures). EPA proposed in
§ 264.97(g) that a sequence of samples
be taken at either daily, weekly, or
monthly intervals. Providing the owner
or operator with a flexible sampling
schedule will allow him or her to choose
a sampling procedure that will reflect ,
site-specific concerns. The intent was to
seta sampling frequency that allows
-------
Federal Register / Vol. 53, No. 196 / Tuesday. October 11, 1988 / Rules and Regulations 39725
sufficient time to pass between sampling
events to assure, to the greatest extent
technically feasible, that an independent
sample is taken from each well. In this
final rule, the language of § 264.97(g) (1)-
(4) has been consolidated into one
provision, § 264.97(g), which specifes
that the owner or operator shall obtain a
sequence of at least four samples from
each well, based on an interval that is
determined after evaluating the aquifer's
effective porosity, hydraulic
conductivity and hydraulic gradient
(which govern rates of flow), and the
fate and transport characteristics of the
potential contaminants.
The minimum number of samples that
are to be collected each testing period is
four. This minimum number was
selected by the Agency to maintain
consistency with the prior requirements
that specified that the owner or operator
collect one sample from each well and
divide it into four replicate samples for
laboratory analysis. Therefore, requiring
the owner or operator to collect four
samples from each well for laboratory
analysis should not impose an increase
in the number of analyses. There may,
however, be an increase in field
' sampling efforts associated with this
sampling procedure. However, the
quality of the ground-water monitoring
data will be significantly improved.
In order.to maintain a complete
annual record of ground-water data, the
facility owner or operator may find it
desirable to obtain a sample each month
of the year. This will help identify
seasonal trends in the data and permit
evaluation of the effects of
autocorrelation and seasonal variation if
present in the samples.
Several commenters noted that the
number and kinds of samples collected
to establish background should be
appropriate to the form of statistical test
employed, following generally accepted
statistical principles. EPA agrees. Thus,
for example, the use of control charts
presumes a well defined background of
perhaps 16 to 30 samples. By contrast,
ANOVA alternatives might require only
4 to 6 samples. A performance standard
stating that the number and kinds of
samples collected to establish
background be appropriate for the form
of statistical test employed was
incorporated into § 264.97(g) of today's
final rule. In addition, a guidance '
document under development includes
scenarios for which each sampling
procedure would be most appropriate.
IV. Implementation
. In addition to changes made in this
final rule pursuant to public comments,
the Agency is also promulgating a series
of changes to clarify the implementation
of these regulations. The Agency
recognizes that some discussion of the
implementation of these changes may be
beneficial prior to the issuance of the
guidance document. However,
additional information concerning
implementation will be addressed in the
guidance document
Because today's amendments to the
statistical methods and sampling
procedures require that an owner or
operator institute methods that conform
to the unit's site-specific characteristics
and eliminate the CABF method as the
default method, compliance with today's
regulations requires detailed knowledge
of the site. Thus, an important
implementation issue concerns die
source of this site-specific information.
Such information should be available to
owners or operators at a sufficient level
of specificity to allow these regulations
to easily be implemented at all regulated
units subject to these regulations. For
new units, or units operating under
interim status, the gathering of the
applicable site-specific data is a
requirement of Part B of a RCRA permit
application under § 270.14(c)(2). Under
this provision, owners and operators o£
hazardous waste surface impoundments,
waste piles, land treatment units, and
landfills must identify the uppermost
aquifer and aquifers hydraulically
interconnected beneath the facility
property, including ground-water flow
direction and rate, and the basis for that
identification. Units currently operating
with a RCRA permit should also have
site-specific data, obtained either from
on-going ground-water monitoring or to
fulfill the § 270.14(c){2) requirement for
the Part B permit application.
The second major implementation
issue concerns when and how the
sampling frequency and statistical
method will be specified in the facility's
RCRA permit. Under § 270.14(c)(7)(vi),
owners and operators must submit a
description of their proposed sampling,
analysis and statistical comparison
procedures to be used in evaluating
ground-water monitoring data as a
requirement of their Part B permit
application. While most new units or
units operating under interim status
should have the data necessary to
propose a sampling frequency, they may
not have sufficient data to propose a
statistical comparison procedure. The
Agency does not believe this will pose
an implementation problem, however.
Where this is the case, the owner or
operator shall propose a'contingency
plan under § 270.14(c)(7)(vi) in which
several statistical methods and the
conditions under which the method
would be appropriate-at the site is
specified. The Agency notes that the .
ANOVA statistical method specified in
§ 264.97(h)(l) can be performed with six
months of ground-water monitoring
data, and thus owners and operators
with this amount of data would not need
to propose a contingency plan under
§ 270.14(c}(7)(vi), but could propose use
of the ANOVA statistical method.
Should an owner or operator who
incorporates a contingency plan into his
or her permit wish to use a statistical
method not specified in the contingency
plan at a later date, he or she may
propose a permit modification to
incorporate this method in their RCRA
permit under § 270.1(a)(3). Owners and
operators currently operating under a
RCRA permit and employing the CABF
method or another statistical method or
sampling procedure not appropriate at
their unit may of course also apply for a
permit modification under § 270.41(a}(3)
to institute an appropriate sampling
procedure and statistical method.
V. Miscellaneous
A. Deletion of Proposed § 264.97(i)(3)
The third performance standard that
appeared in § 264.97(i)(3) of the
proposed rule required that the
monitoring well system be in
accordance with the natural features of
the site. Although this requirement is a
very'important component .of a ground-
water monitoring system, it was out of
place as a performance standard in that
it does not describe requirements that
are directly related to the statistical
methods or sampling procedures.
Further, it is redundant with § 264.97(a)
of the regulations. For these reasons it
does not appear as a performance
standard in today's final rule.
B. Demonstrations of Error Caused by
Data Variability
Section 264.97(k) of the proposed
regulations included a provision
allowing the Regional Administrator to
specify statistical tests of trend,
seasonal variation and autocorrelation
should the owner or operator suspect
that the contamination detected by any
of the statistical tests was caused by
some feature of the data other than
contamination. The Agency is retaining
the substance of this provision in the
final rule. However, because § 264.98(1}
and § 264.99(j) °f the regulation
currently provide for demonstrations of
error by owners and operators pursuant
to a detection of contamination
suspected to be caused by some other
feature, this final rule amends these
sections to incorporate the substance of
the proposed § 264.97(k). Thus, under
today's final rule, as part of a
-------
39726 Federal Register / Vol. 53, No. 196 / Tuesday. October 11, 1988 / Rules and Regulations
- .,_____^.^^^^^^.^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^MMBiM^^^MM^^^^HB^^MBB^^^^^^*^^^^^^^^^^^^*^^^
demonstration that the detection of
contamination at a unit during detection
monitoring (§ 264.98{g)(6)) or during
compliance monitoring (§ 264.99{i)) was
an error or caused by another source,
the owner or operator may perform
statistical tests to evaluate trends,
seasonal variation, or autocorrelation.
VI. General Description of Statistical
Methods
A. Analysis-of-Variance
The analysis-of-variance (ANOVA) is
a statistical method for analyzing; data
from ground-water monitoring wells. It
is a special case of a more, general
procedure referred to as a general linear
model (GLM) and as such is a veiry
flexible analysis system.
Analysis-of-variance is a method for
partitioning the total variation in a set of
data into the different sources of
variation that are present. It results in a
summary table that provides a
convenient form for summarizing and
presenting information contained! in a
set of data. Analysis-of-variance models
are used tb analyze the effects of the
independent variable or variables under
study on the dependent variable. In the
context of ground-water monitoring,
wells or groups of wells represent the
independent variables. The
concentration of hazardous constituents
is the dependent variable. The analysis-
of-variance would determine whether
different wells (or groups of wells) had
significantly different concentrations of
the hazardous constituents.
Contrasts are used to investigate
where any differences occur. In this case
the contrasts of interest are the pairwise
contrasts between the background Wells
and the compliance wells. In a
parametric analysis-of-variance, the
contrast* of interest is the comparison
between the mean concentration of the
background wells and the mean
concentration of each compliance well.
In ground-water monitoring, the
analysis of variance is generally -
appropriate in situations where a
background concentration for a
particular constituent can be
established. If there are data from
several wells for one or more time
periods for a water quality parameter
that are not normally distributed, and
not transformable to normality, then an
analysis-of-variance based upon ranks
(nonparametric ANOVA) may be
appropriate.
B, Tolerance Intervals
Tolerance intervals define, with a
specified probability, a range of values
that contain a discrete percentage of the
population. Tolerance intervals are
simple to construct, requiring a
calculator and a table of tolerance
factors. Because of their simple '
construction, tolerance intervals are
easy to understand and apply to a
ground-water monitoring scenario.
Tolerance intervals can be used in a
detection monitoring program when
individual compliance wells are
compared to a group of background
wells in order to detect ground-water
contamination. Tolerance intervals can
be constructed from the background
well concentrations and expressed as an
interval centered at the mean
background well concentration.
Compliance well hazardous constituent
concentrations found to fall outside of
the tolerance interval limits signal
possible ground-water contamination.
Tolerance intervals may also be
applied to a hazardous waste site in a
compliance mentoring program.
Tolerance intervals can be constructed
from the compliance well hazardous
constituent concentrations, starting
when the facility entered the compliance
monitoring program. The objective of
this procedure is to construct a^tolerance
interval based on the background well
constituent concentrations, testing each
compliance well concentration to
determine if it lies within the tolerance
interval. If the present concentration of '
a compliance well hazardous
constituent is greater than the historical
tolerance interval limits, it indicates that
the ground-water quality has
deteriorated to such a point that further
action may be warranted.
C. Prediction Intervals
A prediction interval is an interval in
which one-is confident at a specified
percentage that the next observation
will lie within the interval. Lake
tolerance intervals, prediction intervals
are simple to construct, requiring only a
calculator and a table of prediction
factors.
Parametric prediction intervals can be
constructed for constituents that follow
a normal distribution. In some cases.
prediction intervals can be constructed
for constituents that have non-normal
distributions (e.g., Poisson or binomial
distributions). It should be noted.
however, that most other distribution-
free prediction intervals cannot be
constructed with a specified probability,
and therefore their use is not
recommended.
Prediction intervals are used in a
detection monitoring program when
individual compliance well
concentrations are compared to one or
more background wells. The mean
concentration and standard deviation
are estimated from the background well
sample, and prediction intervals are
constructed on the basis of the number
of previous observations, the number of
new measurements, and the levels of
confidence that one wishes to obtain.
Future compliance well hazardous
constituent -concentrations found to fall
outside of the prediction limit(s) signal
possible ground-water contamination.
In a compliance monitoring program,
prediction intervals are constructed
from compliance well concentrations,
starting when the facility entered the
compliance monitoring program. Each
future compliance well observation is
tested to determine if it lies within the
prediction interval If the present
concentration of a compliance well
hazardous constituent is greater than
the historical prediction limits, it
.indicates that the ground-water quality
'has deteriorated to such a point that
further action may be warranted.
D. Control Charts
Control charts are widely used as a
statistical tool in industry as well as
reasearch and development
laboratories. From the population
distribution of a given variable, such as
concentrations of a given constituent,
repeated random samples are taken at
intervals over time. Statistics, for
example the mean of replicate values at
a point in time, are computed and
plotted together with upper and/or
lower predetermined limits on a chart,
where the X-axis represents time. If a
result falls outside these boundaries,
then the process is declared to be "out
of control"; otherwise, the process is
declared to be "in control." The
widespread use of control charts is due
to their ease of construction and the fact
that they can provide a quick visual
evaluation of a situation.
In the context of ground-water
monitoring, control charts can be used
to monitor the inherent statistical
variation of the data collected and to
flag anomalous results. Further
investigation of data points lying outside
the established boundaries will be
necessary before any direct action is
taken.
Control charts, when applied to the
properly adjusted and/or transformed
data, can be used to evaluate ground-
water monitoring data. A control chart
can be constructed for each constituent
in each well to monitor the
concentration of a constituent in a well
over time. A new sample for a given
well can be compared to the historical
data from that well, and conclusions can
be drawn on whether the well is in
control. This specific use of control
charts should be-encouraged regardless
-------
Federal Register / Vol. 53, No. 196 / Tuesday, October 11. 1988 / Rules and Regulations 39727
of the objectives of more refined data
analysis. It provides a quick and easy
means of checking the data for possible
outliners, quality control problems, or
data entry errors.
VH. Glossary
Autocorrelation
A measure of the relationship among
members of a series of observations
typically ordered in time or across
space.
F-Statistic
A statistic calculated on the basis of
the F-distribution. The F-statistic is used
in an analysis-of-variance to determine
if there is a relationship between factors
of interest The F-distribution is also
used to check the equality of variance
assumption in certain statistical tests.
Frequency Distribution
Used to described a set of
measurements and often expressed in a
graphical or tabular form. Several
arbitrary non-overlapping intervals are
established, the number of intervals is
based on the range and units of measure
of the data, and the number of
measurements falling within each
interval are recorded or plotted in
sequence. The resulting plot or table
describes the frequency distribution.
Lognormal Distribution
If the logarithms (to any base) of a set
of measurements are distributed
according to the normal distribution, the
original measurements, prior to the
logarithmic transformation, are said to
follow a lognormal distribution. A
lognormal frequency distribution
typically has a long narrow tail and is
often used to describe sets of
environmental data such as ground-
water concentration measurements.
Mean
The most common mean is the
arithmetic mean, which refers to the "
center or average of a set of
measurements. The arithmetic mean is
defined as equal to the sum of all the
observations divided by the number of
observations.
Non-Normal Distribution
A non-normal distribution refers to
any of the many distributions other than
the normal distribution. The lognormal
and exponential are examples of non-
normal distributions. Many parametric
statistical procedures require that the
data be selected from a population
following a normal distribution.
Nonparametric
Refers to statistical procedures which
do not necessitate the use of as many
assumptions,jfor example, that the data
be selected from a specific distribution,
as an equivalent parametric statistical
procedure. Nonparametric tests are
often called distribution-free tests.
Normal Distribution
A widely used, continuous frequency
distribution that approximates a
symmetrical bell-shaped curve hi
appearance. Parametric statistical
procedures often require that data
approximate a normal distribution.
Parametric
The mean and variance of an normal
distribution are examples of parameters.
Parametric statistical procedures rely on
estimates of the mean and variance and
often assume that the data were
selected from a population which
follows, a normal distribution.
Power
The power of the statistical
procedures used in detection and
compliance monitoring is .the probability
that contamination will be detected
(rejection of the null hypothesis of no
contamination) by the statistical
procedure when contamination is really
present For a given sampling protocol,
the power is greatest when the
downgradient concentrations are much
larger than background and the power is
least when dpwngradient concentrations
are only slightly larger than background.
The concept of power does not apply
when downgradient concentrations are
less than or equal to background
concentrations.
Practical Quantification Limits (pql's)
The lowest concentration level that
can be reliably achieved within
specified limits of precision and
accuracy during routine laboratory
operating conditions.
Robust
A testing procedure is robust in the
sense that small or moderate departures
from the assumptions required for a test,
such as normality or constant variance,
do not markedly affect its performance.
Seasonal Variation
A series of ground-water
measurements collected over time
exhibit seasonal variation when .the
measurements vary across sampling
events in a periodic or cyclical fashion
that can be explained by seasonal
effects such as the annual cycle-of
ground-water recharge.
Spatial Variation
The variation among a group of
measurements from samples obtained at
the same time from different horizontal
or vertical locations.
Standard Deviation
A measure of the dispersion, spread
or deviation of a set of observations
around the mean. It is the positive
square root to the variance and is
expressed in the same units of measure
as the original observations.
t-Statistic
A statistic calculated on the basis of
the t-distribution. The shape of the curve
for a ^distribution changes with the
number of observations in the sample
that are used to estimate the sample
populations. As the number of
observations in the sample approach
infinity, the t-distribution becomes
identical to the normal distribution.
Temporal Variation
The variation among a series of
measurements from samples obtained at
the same location but over time.
Variance
A measure of the dispersion, spread,.
deviation or variability of a set of
observations around the mean. The sum
of the squared deviations of the
observations from the arithmetic mean
divided by one less than the total
number of observations.
VIII. Regulatory Analysis
A. State Authority
Under section 3006 of RCRA, EPA
may authorize qualified States to
administer and enforce their State
hazardous waste management programs
in lieu of EPA operating the Federal
program in those States. Authorization,
either interim or final, may be granted to
State programs that regulate the
identification, generation;
transportation, or operation of facilities
that treat, store, or dispose of hazardous
waste. Upon authorization of the State
program, EPA suspends operation
within the States of those parts of the
ground-water monitoring requirements
for land-based hazardous waste.
management facilities applying for and
operating under permits. Since the
ground-water monitoring requirements
are not imposed under any of the
amendments made by the Hazardous
and Solid Waste Amendments of 1984,
final rules modifying the statistical
procedures would not take effect
directly in all States under section-
3006(g). States that have been granted
-------
39728 Federal Register / Vol. 53. No. 196 / Tuesday. October 11, 1988 / Rules and Regulations
final authorization will have to revise
their programs to cover the additional
requirements in today's announcement.
Generally, these authorized State
programs must be revised within one
year of the date of promulgation of such
standards, or within two years if the
State must amend or enact a statute in
order to make the required revision (see
40 CFR 271.21). However, States may
always impose requirements wlu'ch are
more stringent or have greater coverage
than EPA's programs.
Regulations which are broader in
scope, however, may not be enforced as
part of the federally-authorized RCRA
program.
B. Regulatory Impact Analysis
Executive Order 12291 (46 FR13191,
February 9,1981) requires that a
regulatory agency determine whether a
new regulation will be "major" and, if
so, that a Regulatory Impact Analysis be
conducted. A major rule is defined as a ,
regulation that ia likely to result in:
1. An annual effect on the economy of
$100 million or more;
2. A major increase in costs or prices
for consumers, individual industries,
Federal, State, or local government
agencies or geographic regions; or
3, Significant adverse effects on
competition, employment, investment,
productivity, innovation, or the ability of
United States-based enterprises to
compete with foreign-based enterprises
in domestic or export markets.
The Agency has determined that
today's regulation is not a major rule
because it does not meet the above
criteria. Today's action should produce
a net decrease in the cost of ground-
water monitoring at each facility. This
final rule has been submitted to the
Office of Management and Budget
(OMB) for review in accordance with
Executive Order 12291. OMB has
concurred with this final rule.
C. Regulatory Flexibility Act
Pursuant to the Regulatory Flexibility
Act, 5 U.S.C. 601 et seq., whenever an
agency is required to publish a general
notice of rulemaking for any proposed or
final rule, it must prepare and make
available for public comment a
regulatory flexibility analysis which
describes the impact of the rule on small
entities (e.g., small businesses, small
organizations, and small governmental
jurisdictions). The Administrator may
certify, however, that Jhe rule will not
have, a significant economic impact on a
substantial number of small entities. As
stated above, this final rule will have no
adverse impacts on businesses of any
aizo. Accordingly, I hereby certify that
this regulation will not have a
significant economic impact on a
substantial number of small entities.
This final rule, therefore, does not
require a regulatory flexibility analysis.
List of Subjects in 40 CFR Part 264
Hazardous material, Reporting and
recordkeeping requirements, Waste
treatment and disposal, Ground water,
Environmental monitoring.
Date: September 28,1988.
Lee M. Thomas,
Administrator. .
Therefore, 40 CFR Chapter I is -
amended as follows: . .
PART 264STANDARDS FOR
OWNERS AND OPERATORS OF
HAZARDOUS WASTE TREATMENT,
STORAGE, AND DISPOSAL
FACILITIES
1. The authority citation for Part 264
continues to read as follows:
Authority: Sees. 1006, 2002(a), 3004, and
3005 of the Solid Waste Disposal Act, as
amended by the Resource Conservation and
Recovery Act, as amended (42 U.S.C. 6905,
6912(a), 6924, and 6925).'
2. In § 264.91 by revising paragraphs
(a)(l) and (a)(2) to read as follows:
§ 264.91 Required programs.
..(a)***
(1) Whenever hazardous constituents
under § 264.93 from a regulated unit are
detected at a compliance point under
§ 264.95, the owner or operator must
institute a compliance monitoring
program under § 264.99. Detected is
defined as statistically significant
evidence of contamination as described
in § 264.98(f);
. (2) Whenever the ground-water
protection standard under § 264.92 is
exceeded, the owner or operator must
institute a corrective action program
under § 264.100. Exceeded is defined as
statistically significant evidence of
increased contamination as described in
§ 264.99(d); '
* ' * * * *
3. Section 264.92 is revised to read as
follows: ...
§ 264.92 Ground-water protection
standard.
The owner or operator must comply
with conditions specified in the facility
permit that are designed to ensure that
hazardous constituents under § 264.93
detected in the ground water from a
regulated unit do not exceed the
concentration limits under § 264.94 in
the uppermost aquifer underlying the
waste management area beyond the
point of compliance under § 264.95
during the compliance period under
§ 264.96. The Regional Administrator
will establish this ground-water
protection standard in the facility permit
when hazardous constituents have been
detected in the ground water.
4. In § 264.97 by removing the word
"and" from the end of (a)(l),
redesignating and revising (g)(3) as
(a)(l)(i)> adding (a)(3), revising
paragraphs (g) and (h), and adding (i)
and (j), to read as follows:
§ 264.97 General ground-water monitoring
requirements.
(a) * * * -
(1)* * *
(i) A determination of background
quality may include sampling of wells
that are not hydraulically upgradient of
the waste management area where:
(A) Hydrogeologic conditions do not
allow the owner or operator to
determine what wells are hydraulically
upgradient; and
(B) Sampling at other wells will
provide an indication of background
ground-water quality that is
representative or more representative
than that provided by the upgradient
wells; and
*****
(3) Allow for the detection*of
contamination when hazardous waste or
hazardous constituents have migrated
from the waste management area to the
uppermost aquifer.
*****
(g) In detection monitoring or where
appropriate in compliance monitoring,
data on each hazardous constituent
specified in the permit will be collected
from background wells and wells at the
compliance point(s). The number and
kinds of samples collected to establish
background shall be appropriate for the
form of statistical test employed,
following generally accepted statistical
principles. The sample size shall be as
large as necessary to ensure with
reasonable confidence that a
contaminant release to ground water
from a facility will be detected. The
owner or operator will determine an
appropriate sampling procedure and
interval for each hazardous constituent
listed in the facility permit which shall
be specified in the unit permit upon
approval by the Regional Administrator.
This sampling procedure shall be:
(1) A sequence of at least four
samples, taken at an interval that
assures, to the greatest extent
technically feasible, that an independent
sample is obtained, by reference to the
uppermost aquifer's effective porosity,
hydraulic conductivity, and hydraulic
gradient, and the fate and transport
-------
Federal Register / Vol. 53, No. 198 / Tuesday, October 11, 1988 / Rules and Regulations 39729
characteristics of the potential
contaminants, or
(2) an alternate sampling procedure
proposed by the owner or operator and
approved by the Regional
Administrator.
(h) The owner or operator will specify
one of the following statistical methods
to be used in evaluating ground-water
monitoring data for each hazardous
constituent which, upon approval by the
Regional Administrator, will be
specified in the unit permit. The
statistical test chosen shall be
conducted separately for each
hazardous constituent in each well.
Where practical quantification limits
(pql's) are used in any of the following
statistical procedures to comply with
§ 264.97(i)(5), the pql must be proposed
by the owner or operator and approved
by the Regional Administrator. Use of
any of the following statistical methods
must be protective of human health and
the environment and must comply with
the performance standards outlined in
paragraph (i) of this section.
(1) A parametric analysis of variance
(ANOVA) followed by multiple
comparisons procedures -to identify
statistically significant evidence of
contamination. The method must
include estimation and testing of the
contrasts between each compliance
well's mean and the background mean
levels for each constituent.
(2) An analysis of variance' (ANOVA)
based on ranks followed by multiple
comparisons procedures to identify
statistically significant evidence of
contamination. The method must
include estimation and testing of the
contrasts between each compliance
well's median and the background
median levels for each constituent.
(3) A tolerance or prediction interval
procedure in which an interval for each
constituent is established from the
distribution of the background data, and
the level of each constituent in each
compliance well is compared to the
upper tolerance or prediction limit.
(4) A 'control chart approach that gives
control limits for each constituent.
(5) Another statistical test method
submitted by the owner or operator and
approved by the Regional
Administrator.
(i) Any statistical method chosen
under § 264.97(h) for specification in the
unit permit shall comply with the
following performance standards, as
appropriate:
(1) The statistical method used to
evaluate ground-water monitoring data
shall be appropriate for the distribution
of chemical parameters or hazardous
constituents. If the distribution of the
chemical parameters or hazardous
constituents is shown by the owner or
operator to be inappropriate for a
normal theory test then the data should
be transformed or a distribution-free
theory test should be used; If the
distributions for the constituents differ,
more than one statistical method may be
needed.
(2) If an individual well comparison
procedure is used to compare an
individual compliance well constituent
concentration with background
constituent concentrations or a ground-
water protection standard, the test shall'
be done at a Type I error level no less
than 0.01 for each testing period. If a
multiple comparisons procedure is used,
the Type I experimentwise error rate for
each testing period shall be no less than
0.05; however, the Type I error of no less
than 0.01 for individual well
comparisons must be maintained. This
performance standard does not apply to
tolerance intervals, prediction intervals
or control charts.
(3) If a control chart approach is used
to evaluate ground-water monitoring
data, the specific type of control chart
and its associated parameter values
shall be proposed by the owner or
operator and approved by the Regional
Administrator if he or she finds it to be
protective of human health and the
environment
(4) If a tolerance interval or a
prediction interval is used to evaluate '
ground water monitoring data, the levels
of confidence and, for tolerance
intervals, the percentage of the
population that the interval must .
contain, shall be proposed by the owner
or operator and approved by the
Regional Administrator if he or she finds
these parameters to be protective of
human health and the environment.
These parameters will be determined
after considering the number of samples
in the background data base, the data
distribution, and the range of the
concentration values for each
constituent of concern.
(5) The statistical method shall
account for data below the limit of
detection with one or more statistical
procedures that are protective of human
health and the environment. Any
practical quantification limit (pql)
approved by the Regional Administrator
under § 264.97(h) that is used in the
statistical method shall be the lowest
concentration level tha can be reliably
achieved within specified limits of
precision and accuracy during routine
laboratory operating conditions that are
available to the facility.
(6) If necessary, the statistical method.
shall include procedures to control or
correct for seasonal and spatial
variability as well as temporal
correlation in the data.
(j) Ground-water monitoring data
collected in accordance with paragraph
(g) of this section including actual levels
of constituents must be maintained in
the facility operating record. The
Regional Administrator will specify in
the permit when the data must be
submitted for review.
5. In § 284.98 by removing paragraphs
(i), (j) and (k), and by revising
paragraphs (c), (d), (f), (g), and (h) to
read as follows:
I 264.98 Detection monitoring program.
***.,*.*
(c) The owner or operator must
conduct a ground-water monitoring
program for each chemical parameter
and hazardous constituent specified in
the permit pursuant to paragraph (a) of
this section in accordance with
§ 264.97(g). The owner or operator must
maintain a record of ground-water
analytical data as measured and in a
form necessary for the determination of
statistical significance under § 264.97(h).
(d) The Regional Administrator, will
specify the frequencies for collecting
samples and conducting statistical tests
to determine whether there is
statistically significant evidence of
contamination for any parameter or
hazardous constituent specified in the
permit under paragraph (a) of this
section in accordance with § 264.97(g). A
sequence of at least four samples from
each well (background and compliance
wells) must be collected at least semi-
annually during detection monitoring.
*****
(f) The owner or operator must
determine whether there is statistically
significant evidence of contamination
for any chemical parameter of
hazardous constituent specified in the
permit pursuant to paragraph (a) of this
section at a frequency specified under
paragraph (d) of this section.
(1) In determining whether
statistically, significant evidence of
contamination exists, the owner or
operator must use the method(s)
specified in the permit under § 264.97(h).
These method(s) must compare data
collected at the compliance point(s) to
the background ground-water quality
data.
(2) The owner or operator must
determine whether there is statistically
significant evidence of contamination at
each monitoring well as the compliance
point within a reasonable period of time
after completion of sampling. The
Regional Administrator will specify hi
the facility permit what period of time is
reasonable, after considering the-
-------
39730 Federal Register / Vol. 53, No. 196 / Tuesday, October 11, 1988 / Rules and Regulations
complexity of the statistical test and the
availability of laboratory facilities to
perform the analysis of ground-water
samples.
(g) If the owner or operator
determines pursuant to paragraph (f) of
this section that there is statistically
significant evidence of contamination
for chemical parameters or hazardous
constituents specified pursuant to
paragraph (a) of this section at any
monitoring well at the compliance point,
he or she must:
(1) Notify the Regional Administrator
of this finding in writing within seven
days. The notification must indicate
what chemical parameters or hazardous
constituents have shown'statistically
significant evidence of contamination;
(2) Immediately sample the ground
water in all monitoring wells and
determine whether constituents in the
list of Appendix IX of Part 264 are
present, and if so, in what
concentration.
(3) For any Appendix IX compounds
found in the analysis pursuant to
paragraph (g)(2) of this section, the
owner or operator may resample within
one month and repeat the analysis for
those compounds detected. If the results
of the second analysis confirm the initial
results, then these constituents \vill form
the basis for compliance monitoring. If
the owner or operator does not resample
for the compounds found pursuant to
paragraph (g)(2) of this section, the
hazardous constituents found during this
initial Appendix DC analysis will form
the basis for compliance monitoring.
(4) Within 90 days, submit to the
Regional Administrator an application
for a permit modification to establish a
compliance monitoring program meeting
the requirements of § 264.99. The
application must include the following
information:
(i) An identification of the
concentration or any Appendix IX
constituent detected in the ground water
at each monitoring well at the
compliance point;
(it) Any proposed changes to the
ground-water monitoring system at the
facility necessary to meet the
requirements of § 264.99;
(iiJ) Any proposed additions or
changes to the monitoring frequency,
sampling and analysis procedures or
methods, or statistical methods used at
the facility necessary to meet the
requirements of § 264.99;
(iv) For each hazardous constituent
detected at the compliance point, a
proposed concentration limit under
§ 204.94(a) (1) or (2), or a notice of intent
to seek an alternate concentration limit
under § 264.94(b): and
(5) Within 180 days, submit to the
Regional Administrator:
(i) All data necessary to justify an
alternate concentration limit sought
under § 264.94(b); and
(ii) An engineering feasibility plan for
a corrective action program necessary to
meet the requirement of § 264.100,
unless:
(A) All hazardous constituents
identified under paragraph (g)(2) of this
section are listed in Table 1 of § 264.94
and their concentrations do not exceed
the respective values given in thajt
Table: or
(B) The owner or operator has sought
an alternate concentration limit under
§ 264.94(b) for every hazardous
constituent identified under paragraph
(g)(2) of this section.
(6) If the owner or operator
determines, pursuant to paragraph (f) of
this section, that there is a statistically
significant difference for chemical
parameters or hazardous constituents
specified pursuant to paragraph (a) of
this section at any monitoring well at
the compliance point, he or she may
demonstrate that a source other than a
regulated unit caused the contamination
or that the detection is an artifact
caused by an error in sampling,
analysis, or statistical evaluation or
natural variation 'in the ground water.
The owner operator may make a
demonstration under this paragraph in
addition to, or in lieu of, submitting a
permit modification application under
paragraph (g)(4) of this section;
however, the owner or operator is not
relieved of the requirement to submit a
permit modification application within
the time specified in paragraph (g)(4) of
this section unless the demonstration
made under this paragraph successfully
shows that a source other than a
regulated unit caused the increase, or
that the increase resulted from error in
sampling, analysis, or evaluation. In
making a demonstration under this
paragraph, the owner or operator must:
- (i) Notify the Regional Administrator
in writing within seven days of
determining statistically significant
evidence of contamination at the
compliance point that he intends to
make a demonstration under this
paragraph;
(ii) Within 90 days, submit a report to
the Regional Administrator which
demonstrates that a source other than a
regulated unit caused the contamination
or that the contamination resulted from
error in sampling, analysis, or
evaluation;
(iii) Within 90 days, submit to the..
Regional Administrator an application
for a permit modification to make any
appropriate changes to the detection
monitoring program facility; and
(iv) Continue to monitor in accordance
with the detection monitoring program
established under this section.
(h) If the owner or operator
determines that the detection monitoring
program no longer satisfies the
requirements of this section, he or she
must, within 90 days, submit an
application for a permit modification to
make any appropriate changes to the
program.
6. In § 264.99 by revising paragraph
(c), revising paragraphs (d), (f), and (g),
removing paragraph (h), redesignating
paragraph (i) as (h), (j) as (i) and (k) as
(j), revising the redesignated paragraphs
(h) introductory text and (i) introductory
text, and removing paragraph (1) to read
as follows:
§ 264.99 Compliance monitoring program.
*****
(c) The Regional Administrator will
specify the sampling procedures and
statistical methods appropriate for the
constituents and the facility, consistent
with § 264.97 (g) and (h).
(1) The owner or operator must
conduct a sampling program for each
chemical parameter or hazardous
constituent in accordance with
§ 264.97(g).
(2) The owner or operator must record
ground-water analytical data as
measured and in form necessary for the
determination of statistical significance
under § 264.97(h) for the compliance
period of the facility.
(d) The owner or operator must
determine whether there is statistically
significant evidence of increased
contamination for any chemical
parameter or hazardous constituent
specified in the permit, pursuant to
paragraph (a) of this section, at a
frequency specified under paragraph (f)
under this section.
(1) In determining whether
statistically significant evidence of
increased contamination exists, the
owner or operator must use the
method(s) specified in the permit under
§ 264.97(h). The methods(s) must
compare data collected at the
compliance point(s) to a concentration
limit developed in accordance with
§ 264.94.
(2) The owner or operator must
determine whether there is statistically
significant evidence of increased
contamination at each monitoring well
at the compliance point within a
reasonable time period after completion
of sampling. The Regional Administrator
will specify that time period in the
facility permit, after considering the
-------
Federal Register / Vol. 53, No. 196 / Tuesday, October 11, 1988 / Rules and Regulations 39731
complexity of the statistical test and the
availability of laboratory facilities to
perform the analysis of ground-water
samples.
*****
(f) The Regional Administrator will
specify the frequencies for collecting
samples and conducting statistical tests
to determine statistically significant
evidence of increased contamination in
accordance with § 264.97(g). A sequence
of at least four samples from each well
[background and compliance wells)
must be collected at least semi-annually
during the compliance period of the
facility.
(g) The owner or operator must
analyze samples from all monitoring
wells at the compliance point for all
constituents contained in Appendix IX
of Part 264 at least annually to
determine whether additional hazardous
constituents are present in the
uppermost aquifer and, if so, at what
concentration, pursuant to procedures in
. § 264.98(f). If the owner or operator finds
Appendix IX constituents in the ground
water that are not already identified in
the permit as monitoring constituents,
the owner or operator may resample
within one month and repeat the
Appendix IX analysis. If the second
analysis confirms the presence of new
constituents, the owner or operator must
report the concentration of these
additional constituents to the Regional
Administrator within seven days after
the completion of the second analysis
and add them to the monitoring list. If
the owner or operator chooses not to
resample, then he or she must report the
concentrations of these additional
constituents to the Regional
Administrator within seven days after
completion of the intiial analysis and
add them to the monitoring list
(h) If the owner or operator
determines pursuant to paragraph (d) of
this section that any concentration
limits under § 264.94 are being exceeded
at any monitoring well at the point of
compliance he or she must:
*****
(i) If the owner or operator
determines, pursuant to paragraph (d) of
this section, that the ground-water
concentration limits under this section
are being exceeded at any monitoring
well at the point of compliance, he or
she may demonstrate that a source other
than a regulated unit caused the
contamination or that the detection is an
artifact caused by an error in sampling,
analysis, or statistical evaluation or
natural variation in the ground water. In
making a demonstration under this
paragraph, the owner or operator must:
*****
[FR Doc. 88-22913 Filed 10-7-88; 8:45 am]
BILLING CODE 656O-50-M
-------
------- |