TECHNICAL BACKGROUND DOCUMENT
TO SUPPORT RULEMAKING
PURSUANT TO CERCLA SECTION 103(f)(2)
A Report to
RELEASES CONTROL BRANCH
OFFICE OF RESEARCH AND DEVELOPMENT
And to
EMERGENCY RESPONSE DIVISION
OFFICE OF SOLID WASTE AND EMERGENCY RESPONSE
U.S. ENVIRONMENTAL PROTECTION AGENCY
Under
Contract No. 68-03-3182
Prepared by
ENVIRONMENTAL MONITORING & SERVICES, INCORPORATED
A WHOLLY OWNED SUBSIDIARY OF COMBUSTION ENGINEERING, INC.
NEWBURY PARK, CALIFORNIA 91320
June 1986
DRAFT
DO NOT CITE OR QUOTE
-------
54019861
TECHNICAL BACKGROUND DOCUMENT
TO SUPPORT RULEMAKING
PURSUANT TO CERCLA SECTION 103(f)(2)
A Report to
RELEASES CONTROL BRANCH
OFFICE OF RESEARCH AND DEVELOPMENT
And to
EMERGENCY RESPONSE DIVISION
OFFICE OF SOLID VASTE AND EMERGENCY RESPONSE
U.S. ENVIRONMENTAL PROTECTION AGENCY
Under
Contract No. 68-03-3182
Prepared by
ENVIRONMENTAL MONITORING & SERVICES, INCORPORATED
A WHOLLY OWNED SUBSIDIARY OF COMBUSTION ENGINEERING, INC.
NEWBURY PARK, CALIFORNIA 91320
June 1986
DRAFT
DO NOT CITE OR QUOTE
-------
DISCLAIMER
This document is a preliminary draft. It has not been formally released by
the U.S. Environmental Protection Agency and should not at this stage be
construed to represent Agency policy. It is being circulated for comments
on its technical merit and policy implications.
-------
FOREWORD
This report was submitted to the Releases Control Branch, formerly the Oil
and Hazardous Materials Spills Branch, Office of Research and Development,
Edison, New Jersey, and the Emergency Response Division of the Office of
Emergency and Remedial Response of the U.S. Environmental Protection Agency
(EPA) to provide the technical base for the proposed rule on notification of
continuous releases under CERCLA Section 103(f)(2). Comments on possible
approaches under consideration by the EPA had been solicited in the preamble
to a May 25, 1983 (NPRM) (48 FR 23559). This report was prepared by the
technical staff of Environmental Monitoring & Services, Incorporated (EMSI),
a wholly owned subsidiary of Combustion Engineering. Principal authors are
Mr. Franz J. Burmann, Dr. Milton Kirsch and Dr. Alan Messing. Much of the
material contained herein was presented in preliminary drafts by the staff
of ICF, Inc., EMSI's principal subcontractor.
The guidance of many members of the EPA staff was provided not only in a
continuous fashion, but also in special working group meetings and by
written and oral comments on the various drafts. The assistance of Mr. John
Riley, Dr K. Jack Kooyoomjian, Ms. Barbara Hostage and Ms. Nancy Parkinson
of the Emergency Response Division was especially helpful.
The EPA project officer for this project is Mr. Rich Field of the Release
Control Branch, Edison, New Jersey, who provided invaluable guidance during
the project. The support and helpful suggestions of Mr. James Janis,
Mr. Michael Goldman, Mr. James Bunting and other staff members of ICF,
Incorporated, subcontractor to Environmental Monitoring & Services,
Incorporated, are acknowledged.
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY ES-1
PURPOSE OF CONTINUOUS RELEASE REGULATION 1-1
OVERALL PURPOSE OF CERCLA SECTION 102 AND 103 1-1
NEED FOR CONTINUOUS RELEASE REGULATION 1-2
Need for Authority to Regulate Under CERCLA
Section 103(f)(2) 1-5
Need for Clarifying Statutory Language 1-5
Provision of Civil Penalties in the Superfund
Improvement Act 1-6
Usefulness of Continuous Release Data 1-7
EMISSIONS AND DISCHARGES FROM OPERATING FACILITIES 1-7
RATIONALE FOR NOTICE OF PROPOSED RULEMAKING 2-1
THE AFFECTED COMMUNITY 2-1
DEFINITION OF CONTINUOUS RELEASE 2-8
INTERPRETATION OF "STABLE IN QUANTITY AND RATE" 2-9
INTERPRETATION OF "FOR A PERIOD SUFFICIENT TO ESTABLISH
THE CONTINUITY, QUANTITY AND REGULARITY OF THE RELEASE" ...2-10
INTERPRETATION OF "STATISTICALLY SIGNIFICANT INCREASE" (SSI) 2-11
The Nonparametric Test 2-13
Other Tests Which Were Considered But Rejected 2-15
Control Chart Test 2-15
Dixon Criterion 2-17
Grubb' s Test 2-19
Student t Test 2-20
INITIAL, ANNUAL, AND SSI REPORTING 2-22
Annual Reporting of a Continuous Release 2-22
Notification of a Statistically Significant
Increase (SSI) 2-23
BASIS FOR ADDING TOGETHER CONCURRENT RELEASES 2-25
NEED TO MEET ALL FOREGOING REQUIREMENTS 2-25
DATA SOURCES EXAMINED TO ASSIST IN RULEMAKING 2-25
NPDES Permit Compliance Reports —2-26
Philadelphia Air Pollution Control Board Reports 2-26
Office of Air Quality Planning and Standards (OAQPS)
Data Review 2-27
Summary 2-27
REFERENCES , 3-1
APPENDIX A A-l
APPENDIX B B-l
APPENDIX C C-l
-------
LIST OF TABLES
TABLE 2-1 - RELEASES REPORTED TO PHILADELPHIA AIR CONTROL BOARD
THAT ARE POTENTIAL CANDIDATES FOR REDUCED REPORTING
UNDER CERCLA SECTION 103(F)(2) 2-4
TABLE C-l DIXON CRITERION REPORTING TRIGGERS C-8
TABLE C-2 GRUBB'S TEST REPORTING TRIGGERS C-9
TABLE C-3 STUDENT'S t TEST REPORTING TRIGGERS C-10
ii
-------
EXECUTIVE SUMMARY
Section 103(a) of the Comprehensive Environmental Response, Compensation and
Liability Act of 1980 (CERCLA or "Superfund") requires immediate
notification from any person in charge of a vessel or an offshore or an
onshore facility who releases an amount of a hazardous substance equal to or
greater than its reportable quantity (RQ). Under CERCLA Section 102(b), the
RQ of any hazardous substance defined in CERCLA Section 101(14) is 1 pound
unless a different RQ has been established pursuant to Section 311(b)(4) of
the Federal Water Pollution Control Act also known as the Clean Water Act
(CWA). These are the statutory RQs for the hazardous substances unless and
until the Environmental Protection Agency (EPA) promulgates regulations
adjusting them. CERCLA also permits the EPA to establish a single RQ for
each hazardous substance regardless of the environmental medium into which
the substance is released. Section 103(f)(2) of CERCLA provides a mechanism
for those who release CERCLA hazardous substances on a continuous basis to
reduce their reporting burden. The proposed regulation on the reduced
reporting requirements for continuous releases of hazardous substances
clarifies the undefined terms in CERCLA Section 103(f)(2) so that the
regulated community is fully apprised of its responsibility under the
statute. "Continuous" is defined to include "during operating hours" and
"during regularly occurring batch processes" as well as "without
interruption or abatement."
To qualify for the reduced notification provisions of CERCLA Section
103(f)(2), it must be demonstrated that the release is "continuous" and
"stable in quantity and rate." In addition, the release must either be; 1)
ESI
-------
from a facility for which notification has been given under Section 103(c);
or 2) one for which notification has been given under Section 103(a) or
103(b) for a "period sufficient to establish the continuity, quantity and
regularity of such release." Documents and other relevant information used
to support this qualification are to be kept on file at the
facility where they may be inspected by EPA as necessary.
The required annual notification shall be made in writing to the appropriate
EPA Regional Administrator. For each hazardous substance for which the
reduced reporting requirement is claimed, the annual report must include:
o the identity of the hazardous substance, the Chemical Abstracts
Service Registry Number, the media affected, the annual quantity
released in pounds;
o the number of times the amount released in a 24-hour period exceeded
the reportable quantity; and
o the size of both the mean release and of the largest release in
pounds.
Notification of a statistically significant increase, SSI, in the quantity
released shall be made to the National Response Center as set forth in 40
CFR Part 302.6. Determination that a release represents an SSI shall be
made using a simple Nonparametric test described herein, or a superior
alternative test provided by the releasor and approved by the EPA
Administrator. Multiple concurrent releases of the same hazardous substance
occurring at various locations within a single facility shall be added
together in deciding whether the releases constitute a continuous release or
a SSI.
ES2
-------
The proposed regulation is designed to provide the regulated community and
the Agency with the maximum flexibility and minimum reporting burden while
ensuring that sufficient information and timely notice is given to the
Federal Government so that a response may be taken, if necessary, to protect
the public health, welfare and the environment consistent with the CERCLA
statutory mandate, and the National Contingency Plan (NCP).
ESS
-------
SECTION 1
PURPOSE OF CONTINUOUS RELEASE REGULATION
OVERALL PURPOSE OF CERCLA SECTIONS 102 AND 103
The Comprehensive Environmental Response Compensation and Liability Act of
1980 (CERCLA or the Act) is an Act "to provide for liability, compensation,
cleanup, and emergency response for hazardous substances released into the
environment and the cleanup of inactive hazardous waste disposal sites."
Section 101(14) of the Act defines a set of "hazardous substances" by
reference to other environmental statutes, with authority also residing with
EPA to designate additional hazardous substances under CERCLA Section 102.
The CERCLA hazardous substance list currently contains 717 substances.
Sections 103(a) and (b) of CERCLA require that a person in charge of a
vessel or a facility from which a hazardous substance is released into the
environment in an amount equal to or greater than its reportable quantity
(RQ) must immediately notify the National Response Center (NRC). Under
CERCLA Section 102(b), the RQ of any hazardous substance defined in CERCLA
Section 101(14) is 1 pound, except for those hazardous substances for which
RQs have been established pursuant to Section 311(b)(4) of the Clean Water
Act (CWA). Section 102(a) authorizes EPA to adjust RQs for hazardous
substances if deemed appropriate.
The primary purpose of the CERCLA notification requirement is to ensure that
releasers notify the government so that the need for a federal response can
be evaluated and any necessary response can be undertaken in a timely
fashion to protect the public health or welfare or the environment. Details
of the notification requirements for release of hazardous substances
appeared in a notice of proposed rulemaking (NPRM) on May 25, 1983. (48 FR
23552) and in a final rule on! April 4, 1985 (50 FR 13456).
1-1
-------
NEED FOR CONTINUOUS RELEASE REGULATION
Section 103(f)(2) of CERCLA modifies the general notification requirements
of Sections 103(a) and 103(b) for certain releases. Releases may be
reported less frequently than otherwise would be required if they are
"continuous," "stable in quantity and rate," and notification has been given
under Sections 103(a) and (b) "for a period sufficient to establish the
continuity, quantity, and regularity" of the release or under Section
103(c), which relates to notification of the existence of certain facilities
that are or have been used for storage, treatment, or disposal of hazardous
wastes. The alternative reduced reporting requirements imposed by Section
103(f)(2) are that notification of releases that are continuous and stable
in quantity and rate must be given "annually, or at such time as there is
any statistically significant increase" in the quantity of the hazardous
substance being released.
The main function of CERCLA's notification requirements is to alert
government officials to the existence of a situation that may require a
government response to protect the public health or welfare or the
environment. Because episodic releases are almost always "unanticipated,"
they must be reported as they occur. Continuous releases, on the other
hand, may be predictable and can either be literally continuous or
recurring. Congress recognized in Section 103(f)(2) that CERCLA's
notification objectives would be satisfied by less frequent reporting in the
case of continuous releases. Thus, instead of reporting every release as it
occurs, persons in charge are allowed to report certain continuous releases
less often under Section 103(f)(2).
1-2
-------
The purpose of this section of the act is to reduce unnecessary reports of
releases. This provision is designed to decrease the burden of reporting on
both the regulated community and on the Federal Government. The rationale
for this approach is that when a release occurs regularly and in relatively
stable amounts, the Agency does not need to be notified each time such a
release occurs to have the information necessary to decide whether a
response to the release is necessary. In addition, this provision allows
attention to be focused on those releases most likely to require a response
by the government. Section 103(f)(2) provides, however, that notification
must still be given (1) under Section 103(c) or "for a period sufficient to
establish the continuity, quantity, and regularity of such release" and (2)
"annually, or at such time as there is any statistically significant
increase in the quantity of any hazardous substance or constituent thereof
released, above that previously reported or occurring."
These reports will provide EPA with sufficient information to evaluate a
release to determine whether a response should be taken to protect public
health, welfare and the environment.
The annual report of continuous release will be evaluated by the Regional
Administrator of the EPA region in which the releasing facility or vessel is
located in order to determine whether or not a .response action is necessary.
If no response is found necessary at this point, the notification of each
statistically significant increase to the NRC will provide another
opportunity for evaluation of that release and the necessity for response.
1-3
-------
The terms in Section 103(f)(2), continuous; stable in quantity and rate; for
a period sufficient to establish the continuity, quantity and regularity of
the release; and statistically significant increase, were not defined in
CERCLA or in its legislative history. On May 25, 1983, EPA published a
Notice of Proposed Rulemaking (NPRM) in 48 FR 23552, to clarify procedures
for reporting releases and to adjust reportable quantities for 387 of the
then 698 CERCLA hazardous substances (see Appendix A). In the preamble to
that proposed rule, the Agency discussed the reduced reporting requirements
for continuous releases and set forth a number of ideas that were under
consideration in the development of EPA's position on such releases. The EPA
specifically requested comments on the most feasible approach for continuous
release notification, the information to be required, what the Agency should
consider as a statistically significant increase in the release, and any
other relevant issues. The Agency stated at that time it intended that such
information would enable it to develop a system which would impose a minimal
burden on both the regulated community and the government, while achieving
the underlying statutory objectives. The EPA received 45 letters with
comments on the reduced reporting requirement for continuous releases. (See
Appendix B for a summary of these comments and a list of commenters.) In the
preamble to a final rule adjusting RQs published on April 4, 1985 (50 FR
13456), the EPA noted that due to the complexity of the issues involved, the
Agency had decided to study the continuous release reduced reporting
requirement further. The Agency's definition of these terms associated with
continuous releases are defined in Section 2 of this document and in the
proposed rule.
1-4
-------
Need for Authority to Regulate Under CERCLA Section 103(f)(2)
For many of the provisions of CERCLA that mandated action by the President,
rulemaking authority was delegated through Executive Order 12316 either to
the Administrator of the EPA, the Attorney General or other officials of the
Executive Branch. Since Section 103(f)(2) was assumed originally to be
self-implementing, no rulemaking authority was delegated to EPA for this
section. Not only has the regulatd community expressed uncertainty in the
operation of the reporting requirements in Section 103(f)(2) in comments
submitted on the May 25, 1983 NPRM, but the Agency has received many letters
and telephone calls requesting clarification of the provision. Furthermore,
letters received by the Agency that assert to be annual notifications of
continuous releases significantly lack uniformity in substance, level of
detail, and form. The apparent confusion within the regulated community in
regard to this section of the statute shows the need for regulation and, in
turn, the need for rulemaking authority for the EPA to promulgate a
regulation pursuant to this subsection.
Need for Clarifying Statutory Language
Because so many terms are undefined, and because of the desirability of
exacting uniform notification requirements under Section 103(f)(2),
clarification of the statutory language is essential for the regulated
community to understand fully what and how continuous releases are to be
reported to the NRC. The primary purposes of this document and of the
proposed rule it supports, are:
1-5
-------
o to explain the meaning of the undefined terms in Section
103
-------
Usefulness of Continuous Release Data
The major reason for a continuous release regulation is to clarify how the
regulated community can comply with the provisions of CERCLA Section
103(f)(2) and thereby take advantage of the reduced reporting requirements
for releases of RQs of hazardous substances under the circumstances
described in that subsection of the statute. Peripherally, increased use of
the reduced reporting requirement under CERCLA Section 103(f)(2) should
supply, in the form of the required annual reports, data useful to air and
water permit programs in determining levels of environmental pollution, and
information for the Chemical Emergency Preparedness Program (CEPP).
The EPA developed the CEPP to address accidental releases of acutely toxic
chemicals as part of a comprehensive strategy to deal with the problem of
air toxics in the environment. A notice announcing the availability of the
CEPP Interim Guidance appeared at 50 FR 51451, December 17, 1985. This
document is consistent with the right-to-know provisions of the CERCLA
Reauthorization Bills and with the desirability of local contingency
planning to handle releases for air toxic substances.
EMISSIONS AND DISCHARGES FROM OPERATING FACILITIES
Continuous releases may occur in operating facilities as a result of
emissions from stationary and rotating seals. Pressure vessels, pipe
joints, and pump seals are a few sources of emissions that contribute to
continuous releases. Chimney stacks, scrubber vents, and baghouses may be
larger contributors. Discharges of liquid wastes to Publicly Owned
Treatment Works (POTWs) or waterways may contribute other continuous
releases.
1-7
-------
Continuous releases may be expected to increase with time as a facility ages
and fittings of all kinds loosen. Without increasing vigilance and
continued improvements in maintenance, the sources of continuous releases
will increase. Gradual creep in the quantity released continuously is part
of normal operation.
1-8
-------
SECTION 2
RATIONALE FOR NOTICE OF PROPOSED RULEMAKING
The large number and wide variety of comments received in response to the
May 25, 1983 NPRM suggested that clarification was essential for uniform
interpretation of CERCLA Section 103(f)(2). Several comments requested that
this subsection be implemented soon to avoid the ambiguity that was
perceived. Some commenters also attempted to associate the "federally
permitted release" exemption from notification defined in CERCLA Section
101(10) with the reduced notification requirement allowed by CERCLA Section
103(f)(2) for continuous releases. The EPA is hoping to make perfectly
clear that a federally permitted release as defined in CERCLA Section
101(10) is indeed exempt from the reporting requirements of CERCLA Section
103(a) and (b). The reduced reporting allowed by CERCLA Section 103(f)(2)
is applicable to releases that are not federally permitted as defined by
CERCLA Section 101(10).
THE AFFECTED COMMUNITY
Since the enactment of CERCLA in 1980, it has become apparent that industry
is somewhat confused concerning the types of release that qualify for the
reduced reporting requirement under CERCLA Section 103(f)(2). Thirteen
applications have been submitted to the Agency from the regulated community
since enactment of CERCLA requesting continuous release status and
clarification of the law concerning continuous releases. In addition, a
total of 45 public comment letters were received in the months of July
and August, 1983, in response to the continuous release discussion of the
May 25, 1983 NPRM on CERCLA notification requirements. This number of
comments represents only a tiny fraction of those potentially able to claim
reduced reporting status under CERCLA Section 103(f)(2).
-------
The modification to the general reporting requirements of CERCLA allowed by
Section 103(f)(2) includes releases that are "continuous," "stable in
quantity and rate," and for which notification has been given under Section
103(c) of CERCLA or under Sections 103(a) and (b) of CERCLA "for a period
sufficient to establish the continuity, quantity, and regularity" of the
releases. In addition, in order for the facility to be affected by the
provisions of Section 103(f)(2), the release:
o must contain a CERCLA hazardous substance;
o must not be a federally permitted release as defined
in Section 101(10) of CERCLA; and
o must equal or exceed the RQ for the hazardous substance
in a 24-hour period.
The Agency intends to publish a proposed rule which will address the
federally permitted release exemption.
The types of facilities which are included in this estimate and which will
be affected by this proposed regulation include chemical and petrochemical
manufacturers, metal smelters, dry cleaning establishments, users of CERCLA
hazardous substances, such as aerospace industries, ordinance manufacturers,
insesticide distributors, pharmaceutical manufacturers, degreasers, and
others.
2-2
-------
Continuous releases occur as a result of processing, storing and
transferring hazardous substances. Indeed many releases are a result of
fugitive emissions. An analysis of the Philadelphia Air Pollution Control
Board data is shown in Table 2-1, which illustrates the types of hazardous
substances and sources of their release. These data show that over a
24-hour period the RQs of many substances are exceeded, in some cases by a
great deal.
Industrial practices involving solvent degreasers are obvious examples of
continuous releases of hazardous substances. . In many industries,
Metal-fabricated articles must be washed or degreased before electroplating,
painting or other surface finishing. Most degreasing operations are carried
out in units, called degreasers, in which a chlorinated solvent, such as
trichloroethylene or perchloroethylene, either in the gaseous or liquid
state, is used to wash the parts free of grease and oil. Some measureable
amounts of solvent are released as a vapor or a liquid from even the
smallest degreaser, and the large number of these units in any single big
manufacturing organization makes the combined release substantial. In the
manufacture of rocket engines vapor degreasers are used. Parts to be plated
are suspended in conveyorized units. Solvent is vaporized and rises to fill
the vapor space below water cooled condensers. At the condenser there is a
definite vapor line. Condensed solvent runs through a collection trough and
into a storage tank for later reuse. Articles to be degreased are conveyed
into the vapor space. Vapors condense on the metal parts and hot condensate
rinses oil and grease into the receptacle. It has been reported that daily
emissions of solvents (trichloroethylene and/or perchloroethylene) vary from
a few pounds to as high as 1300 pounds. Solvent is lost from the degreaser
-------
TABLE 2-1
RELEASES REPORTED TO PHILADELPHIA AIR CONTROL BOARD THAT ARE POTENTIAL
CANDIDATES FOR REDUCED REPORTING UNDER CERCLA SECTION 103(f)(2)
Hazardous Substance
Company Source Reported RQ
Maximum Daily (Promulga-
Release ted or
Pounds Proposed
Pounds)
Benzene
Carbon tetrachloride
Carbon tetrachloride
Chloroform
Chloroform
Chromium & compounds
Ethylene oxide
Ethylene oxide
Lead
Lead
Lead
Methyl chloride
Methylene chloride
Methylene chloride
Nickel & compounds
Gulf Refining
Smithkline
Smithkline
Smithkline
Smithkline
Gulf Refining
Penna. Hospital
St. Agnes Med.
Center
Halpern Stein
Assoc. lead
Metal Bank
Rohm & Haas
Publicker
Smithkline
Gulf
floating roof
storage tanks
fan exhauster
from pellet pans
fan & plenum from
pelltet & copper
pans
fan exhauster
from pellet pans
fan & plenum
from pellet &
copper pans
cooling tower
gas sterilizers
central supply
baghouse
Barton operation
wire incineration
process vents
storage tank
vent
film coating
stacks
100
50
100
106
351
17
18.5
U.33
1
10.32
2.8
210
1,355
1,295
20
10
10
10
10
10
1
10
10
1
1
1
100
1000
1000
10
7-4
-------
Contd.
>•
Hazardous Substance Company Source Reported RQ
Maximum Daily (Promulga-
Release ted or
Pounds Proposed
Pounds)
Nickel & compounds General Electric metallizing 17.6 10
spray
Perchloroethylene Coyne dry cleaning 165 100
Perchloroethylene Devon evap. dry 860 100
cleaning
Trichloroethylene Botany dry cleaning 681 100
2-5
-------
tanks in essentially two ways: (1) vaporization (including diffusion) from
the unit, and carryout with degreased articles. Estimates of 0.05 pound of
solvent lost by vaporization per hour per square foot of open tank area have
been made (Ref. 2).
Although the releases discussed above exceed the RQs, the hazards posed by
the operation need not be assessed every day by an On-Scene Coordinator
(OSC). Indeed, the OSC needs to know the daily output only once a year to
determine if a government response is necessary. However, if that daily
release exceeds the anticipated value then the OSC must be apprised so that
he can determine appropriate response action, if any.
Because of the difficulty in obtaining the precise number of facilities
continuously releasing CERCLA hazardous substances into any of the
environmental media, it has been necessary to make estimates. The estimates
are based upon the number of facilities regulated under the Clean Water Act,
the Clean Air Act, and the Resource Conservation and Recovery Act, but do
not include non-regulatd facilities that continuously release CERCLA
hazardous substances for which the Agency does not have any data.
Regardless, the Agency has attempted to make reasonable estimates
approximating the number of facilities that may be affected by the proposed
continuous release reporting regulation. The companion Regulatory Impact
Analysis (RIA) to this report estimates that approximately 15,300 facilities
in the United States which discharge CERCLA hazardous substances on a
continuous basis to surface water in an amount equaling.or exceeding an RQ
may qualify for CERCLA Section 103(f)(2) reduced reporting. In addition,
the RIA estimates that 9,700 facilities releasing CERCLA hazardous
-------
substances to air and 1000 facilities releasing CERCLA hazardous substances
to groundwater may also qualify for the reduced reporting provisions of
CERCLA Section 103(f)(2).
The Federal government is also likely to benefit from the reduced reporting
requirements set forth in the continuous release regulation. In the absence
of such a regulation, continuous releasers must notify the National Response
Center each time a release of a hazardous substance equals or exceeds its RQ
until a demonstration has been made that the release is "continuous" and
"stable in quantity and rate." Once the above mentioned demonstration has
been made, telephone notification to the National Response Center will be
required when there is a statistically significant increase in the
continuous release. A written report containing summary information as
described below must be sent to the EPA Regional Administrator. The EPA
Regional Administrator will receive and process these annual reports. The
National Response Center will also receive and process each telephone
notification of a statistically significant increase in these releases.
Though this represents an increase in the work load for the EPA regional
office, this should be more than offset by the dramatic decrease in reports
of individual releases to the National Response Center, and lead to a net
cost-savings to Government.
1. Processing by the National Response Center includes the relay of
information on the statistically significant increase to the
appropriate EPA Regional Office or U.S. Coast Guard District
Office for response.
-------
DEFINITION OF CONTINUOUS RELEASE
Section 103(f)(2) of CERCLA provides a reduced reporting requirement for
releases which are "continuous" and "stable in quantity and rate." In order
to demonstrate that the release is continuous in nature and stable in
quantity and rate and therefore eligible for the reduced reporting
requirement, a releaser must have:
o notified the Administrator under Section 103(c) of CERCLA; or
o notified the National Response Center under Sections 103(a)
and (b) of CERCLA "for a period sufficient to establish the
continuity, quantity, and regularity of the release."
Releases that meet these criteria need only be reported annually, or when a
statistically significant increase in the quantity of release occurs.
Members of the regulated community have expressed concern in their comments
on the May 25, 1983 NPRM over the definition of "continuous." (See Appendix
B). Many commenters expressed the desire that the word "continuous" be
defined as broadly as possible and that it contain any anticipated routine
intermittent release that is necessarily associated with a manufacturing or
treatment process. In response to these concerns, the EPA has developed a
proposed regulation that would clarify the meaning and operation of the key
terms and phrases of the reduced reporting requirement for continuous
releases, but has chosen not to broaden the definition to include all
"anticipated intermittent" releases. The Agency feels that such a broad
definition may include releases that could involve a large quantity of
hazardous substance (e.g., popping off of a relief valve) that are much
larger than the quantity normally released. Indeed, since "accidents are
bound to happen," and are thus anticipated, such releases of an RQ or more
2-8
-------
of hazardous substance might not be reported to the NRC resulting in a
potentially significant threat to human health, welfare, or the environment.
For this reason, the definition of a continuous release is, therefore,
limited to a release that is:
o literally continuous, without interruption or abatement;
o continuous during operating hours; or
o continuous during regularly occurring batch processes.
Releases that are anticipated are not to be considered continuous for that
reason alone, but must conform to the above definition.
INTERPRETATION OF "STABLE IN QUANTITY AND RATE"
The EPA has decided that the phrase "stable in quantity and rate" refers to
a release that is predictable during normal operations and not the result of
malfunction or upset. This interpretation of the phrase in qualitative
terms is preferable to a quantitative definition because it allows the
flexibility demanded by the wide variety of facilities, plants and vessels
from which continuous releases can occur. Quantitative definitions based on
an absolute magnitude of variation (e.g., 5 pounds, 10 pounds or one RQ from
the daily average) or a relative magnitude of variation such as 5 or 10
percent from the daily average could be developed but would suffer from lack
of flexibility and would be much more costly and burdensome both to the
regulatd community and to the Agency with no apparent increased benefit in
protecting human health or the environment.
2-9
-------
The person-in-charge must report in writing at least annually to the EPA
Regional Administrator the magnitude of the continuous release that is
stable in quantity and rate. A decision can be made at this time if a
response to this quantity and rate is necessary. If no response is
necessary to this base rate of release, then a response decision can be made
at the time a statistically significant increase in the magnitude of the
release is reported to the National Response Center.
EPA has chosen a qualitative measure of the phrase "stable in quantity and
rate." This is defined as a release that is predictable during normal
operations and is not the result of malfunctions or upsets.
INTERPRETATION OF "FOR A PERIOD SUFFICIENT TO ESTABLISH THE CONTINUITY,
QUANTITY AND REGULARITY OF THE RELEASE"
In the May 25, 1983 NPRM, the Agency asked for comments on approaches to
defining the "period sufficient to establish the continuity, quantity and
regularity of the release." Most commenters on this qualifiation for
continuous release reporting were in favor of leaving the specific definition
to the releaser on a case-by-case basis. Other commenters supported the
establishment of a single time period during which all releases would need to
be reported, while others proposed a specified number of release reports to
determine whether the release was continuous and stable in quantity and rate.
Since there are many types of facilities, plants and vessels involved, each of
which differ in the kinds of operations which may produce releases, the EPA
agrees that the person-in-charge may decide how to determine the "period
2-10
-------
sufficient." The Agency also believes that this period need not be long since
the facility owner/operator is familiar with the operation and can draw on
many sources to determine the "continuity, quantity and regularity of the
release." It might be possible to use recent data calculated from mass
balance, release monitoring, National Response Center reports or other sources
for this determination. The Agency recognizes the need for flexibility
because it would be impossible to anticipate all the possible conditions under
which the various facilities operate that could legitimately claim the reduced
reporting provisions of the continuous release exemption regulation. Just as
the term "continuous" encompasses several types of operations and their
associated releases, the "period sufficient" to demonstrate the continuity,
stability and regularity must also be allowed to vary so as to be compatible
with the type of release.
In order to reduce the information collection burden on the regulated
community and the EPA, the Agency has also decided to allow the
owner/operator to record and file at the facility the data and other
information upon which the demonstration that the release is continuous and
stable in quantity and rate is based. The EPA reserves the right to inspect
the releaser's file when necessary or desirable.
INTERPRETATION OF "STATISTICALLY SIGNIFICANT INCREASE" (SSI)
In the preamble to the May 25, 1983 proposed rule (48FR 23559), EPA reported
that three alternatives were being considered for defining statistically
significant increase:
2-11
-------
o Requiring reporting whenever a release falls outside some expected
range based on statistical tests, such as the Student's t test;
o Requiring reporting whenever the amount released exceeds the amount
ordinarily released by some pre-established factor, such as 2, 5,
or 10 times the daily average; or
o Letting the releaser determine what is a statistically significant
increase.
Commenters were nearly equally divided in preferring each of these
alternatives. The EPA has decided, however, to propose a method that is
simple, understandable and easy to use in determing whether a SSI has been
reached. This Nonparametric test identifies the top 5 percent of releases,
and the Agency defines these to be statistically signifant increases. The
Agency investigated a number of statistical tests including the Control
r
Chart test, Dixon Criterion, Grubb's test and Student's t test in order to
determine their appropriateness for identifying a SSI. These tests are
described below.
The statistical tests described in this report do not represent an
exhaustive list of potential statistical tests which might be suitable for
identifying a SSI. As noted in the proposed continuous release reporting
regulation, applicants who wish to use an alternative method may do so if
they can demonstrate to the Agency's satisfaction that it is superior to the
Nonparametric test (see Section 302.8(d) of the proposed rule).
2-12
-------
The Nonparametrie Test
The Nonparametrie test is a simple method to determine whether an increase
in a release is significant. The Agency defines a SSI to be any release
that exceeds the largest of the 19 most recent releases. If fewer than 19
release events have been recorded, then the Agency defines a SSI to be the
largest release to date. Each statistically significant increase in
release, so defined, must then be reported immediately to the National
Response Center. After the initial 19 releases, the newest release is
compared to the immediately previous 19 releases in determining a
statistically significant increase.
For an example of the application of the Nonparametric test, assume that a
facility has a daily release of a CERCLA hazardous substance with an RQ of
10 pounds. The following pattern' of daily releases are recorded. Release
reports are made to the NRC as indicated.
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
Event
1 =
2 =
3 =
4 =
5 =
6 =
7 =
8 =
9 =
10 =
11 =
12 =
13 =
14 =
15 =
16 =
17 =
18 =
19 =
50
70
40
150
60
10
80
20
70
50
100
50
60
30
70
90
40
95
100
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
Report
Report
No
Report
Report
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Report
Report
Report
Report
Report
Report
Report
Report
Report
Report
Report
Report
Report
Report
Report
Event 20 = 50 pounds No Report
Event 21 = 30 pounds No Report
Event 22 = 40 pounds No Report
Event 23 = 90 pounds No Report
Event 24 = 60 pounds No Report
Event 25 = 110 pounds Report
— 1 7
-------
For release events 2 through 19, a release would be reported to the NRC if
it exceeds the largest previous release event. Thus, events 2 and 4 would
be reported to the NRC. Event 20 is then compared to releases in events 1
through 19 and is found not to be statistically significant because it is
less than 150 pounds. Release event 20 is then added to the release
history, and event 1 is deleted. Event 21 would be compared to releases in
events 2 through 20. Because 30 pounds released in event 21 is less than 150
pounds in event 4 no report is required. Event 21 is next added to the
release history and event 2 is deleted. For event 25, the relevant release
history includes events 6 through 24. The 110 pound release in event 25
must be reported to the NRC because it exceeds the largest of the most
recent 19 release events (i.e., 100 pounds).
The principal advantage of the Nonparametric test is that it is insensitive
to and indepenent of the underlying distribution of releases. Because the
Agency has defined the largest five percent of all continuous releases as
statistically significant increases in a release and therefore reportable to
the National Response Center, it was necessary to select a test that was
insensitive to the underlying distribution. The insensitivity of the
Nonparametrict test to the underlying distribution of releases helps
minimize underreporting and overreporting. The largest 5 percent of all
releases from all facilities with continuous and stable releases of CERCLA
hazardous substances will be reported to the National Response Center.
-------
A second advantage of the Non parametric test is its simplicity. The test
does not require any sophisticated calculations but is performed by
recording releases and comparing the newest release to the largest of the
preceding 19 releases. This analysis by inspection will not be burdensome to
the regulated community. It is understood that no additional monitoring of
emissions are required. Section 103(a) and (b) require notification of a
release of a hazardous substance when the operator has knowledge that a
release above athe RQ has occurred. Furthermore, if it was reasonable for
the owner/operator to have knowledge of the release, then failure to report
a release at or above the RQ may be a criminal offense. It follows,
therefore, that industry practices involving handling, manufacturing, using
and storing hazardous substances dictate knowledge of releases. Since the
reporting requirements of Section 103 were self-implementary upon CERCLA
enactment, then owners and operators of facilities involved with hazardous
substances must have the capabilities of determining their daily releases
in order to be in compliance with the law.
Because the Nonparametric test is simple, non burdensome and implementable
the Agency has selected this test as a means of defining statistically
significant increase in quantity and rate.
Other Tests Which Were Considered But Rejected
Control Chart Test
The Control Chart test requires a minimum history of 30 releases in order to
be used effectively. A report is made when an observation exceeds 1.645
-------
Standard Deviations above the mean. If the new recorded release is less than
1.645 standard deviations above the mean (1.645*5), then the release is not
considered a statistically significant increase and does not need to be
reported to the National Response Center. The release is then added to the
release history and a new mean and standard deviation are calculated.
n - 1
n
Where n is the number of releases; Y is the mean (average); and 2
1=1
represents the summation of the term in parentheses, which ranges from 1 to
The Control Chart test will dependably identify the largest 5 percent of all
releases only when the underlying distribution is normal. If the underlying
distribution is lognormal, for example, the control chart test will trigger
a" report for fewer than 5 percent of the releases. The number of reportable
releases is very sensitive to the underlying distribution of releases.
Because of this sensitivity, it is highly likely that this statistical test
would result in significant under or overreporting of releases to the NRC,
depending on the shape of each facility's release distribution.
-------
Relative to the Nonparametrict test, the Control Chart test is more time
consuming and difficult to perform. The test involves recalculation of the
mean and standard deviation of the release data with each nonstatistically
significant release. This recalculation (for some facilities, recalculation
would be expected to occur quite often), could be time consuming and
confusing, with the test yielding no apparent advantage over the
Nonparametric test. Further, an alternative test to the Control Chart test
would be required for facilities with fewer than 30 data points, resulting
in a more complicated regualtory scheme. An example of the application of
the Control Chart test can be found in Appendix C.
Dixon Criterion
The Dixon Criterion is a statistical test that can be used to test the
largest value in a release distribution. The Dixon Criterion compares the
newest release to the largest release in the data set and divides that
difference by the difference between the newest release and the smallest
release in the data set. That is:
2-17
-------
Where X is the most recent relese, X, _i\in the data set, X.. is the
f
smallest release in the data set, and C is the calculated value.
The Dixon Criterion thus derived is then compared to a table relating
reporting triggers to the number of releases in the data set. The Dixon
Criterion table is reconstructed in Table C-l. If the calculated value is
larger than the tabulated value, the release is considered statistically
significant and must be reported to the National Response Center. An
example of the use of the Dixon Criterion can be found in Appendix C.
The Dixon Criterion has several disadvantages. First, it assumes normality
of observations such that if the underlying distribution of releases is not
normal, the test will not be sufficiently sensitive in identifying outliers
in the distribution. Underreporting to the National Response Center of
statistically significant increases in release will result. Second, if the
largest release (X, _..) in the data set is actualy an extreme value, the
test will mask the identity of other extreme values. Extreme values are
more difficult to identify with smaller data sets and misclassification
could .occur relatively often with use of the Dixon Criterion.
The Dixon Criterion was not considered by the Agency for use with large data
sets. It was only considered by the Agency for situations when fewer than
30 observations were available. After 30 data points were accumulated, the
Aency would have required a differenct statistical test. Because such an
approach would have employed a relatively complicated regulatory scheme, the
use of the Dixon Criterion was rejected.
_T 0
-------
Grubb's Test
The Grubb's test is a statistical test used to identify outliers or
anomalies. The Grubb's test involves the calculation of a test value and a
comparison of this test value with the "Grubb's Statistic". The test value
is calculated by subtracting the mean (average) of all past releases (except
those determined to be statistically significant increases), from the newest
release, and dividing this difference by the sample standard deviation of
those same releases. That is,
X - X
Test Value • n
S
where X is the most recent release, X is the mean or average of all past
statistically insignificant releases, and S is the standard deviation for
all statistically insignificant releases. X and S are calculated using the
formulas:
+ (X.-X)2 + ... + (X -X)2
X - -i £ i 11 and S - ' * 2 n
n - 1 . h - 1
2-19
-------
where n represents the number of statistically insignificant releases in the
data set. If the Test Value is less than the corresponding Grubb's
Statistic shown in Table C-2, then the release is not considered a
statistically significant increase and need not be reported to the National
Response Center. If the Test Value is equal to or greater than the Grubb's
Statistic in Table C-2, then the release must be reported to the National
Response Center. After a report to the National Response Center, the value
reported should not be included in future Grubb's test calculations. An
example of the Grubb's test can be found in Appendix C.
The Grubb's test assumes the underlying distribution is normal. If this is
not the case, underreporting or overreporting to the National Response
Center of statistically significant releases could result.
As with the Dixon Criterion, the Grubb's test was not considered for use
with large data sets. The Agency believed there were better tests available
when a release history of 30 data points or more was available. The Grubb's
test is sensitive to the underlying distribution of releases and does not
perform so well as the Nonparametric test. A second statistical test would
have been employed when sufficient data were accumulated, resulting in
complicated the reporting requirements.
Student t Test
The Student t test is simlar to the Grubb's test. It involves the
calculation of a test vlaue and a comparison of this test value with the t
statistic. THe test value is calculated by subtracting the mean of all
2-20
-------
releases (excluding the newest release) from the newest release, and
dividing this difference by the standard deviation.
Student t Value
where X is the most recent release, X is the mean or average of all
releases excluding the newest release, and S is the sample standard
deviation. X and S are calculated using the formulas described for the
Grubb's test. The Student t test is identical to the Grubb's test except
the mean and standard deviation do not include the newest release. The
Student t value is then compared to the reporting triggers found in Table
C-3. If the Student t value equals or exceeds the reporting trigger, the
release is statistically significant and needs to be reported to the
National Response Center. An example of the use of the Student t test can
be found in Appendix C.
The burden of continually recalculating the mean and standard deviation to
formulate the t statistic for comparison to the tabulated value appears to
be much greater than the burden associated with the Nonparametric test, it
is a relatively well known statistical test, however, and facilities would
be generally more familiar with it than with either the Dixon Criterion or
the Grubb's test.
2-21
-------
INITIAL, ANNUAL, AND SSI REPORTING
For those facilities which qualify for reduced reporting the Agency proposes
that the annual report required pursuant to Section 103(f)(2)(B), be sent in
writing to the Regional Administrator of the EPA for the Region in which the
facility or vessel is located instead of to the NRC. The initial report is
due within 90 days of the effective date of the rule. Details of the
contents of these annual and initial reports are described below. Notice of
an SSI is to be made to the NRC pursuant to Section 103(a) and (b) as soon
as the person-in-charge has knowledge that such a release has occurred as
set forth in 40CFR Section 302.6.
Annual Reporting of a Continuous Release
The annual notification is to include for each hazardous substance for which
the CERCLA Section 103(f)(2) reduced reporting allowance is claimed—
o the identify of the hazardous substance, the Chemical Abstracts
Service Registry Number (CASRN) for the hazardous substance, the
media affected, and the annual quantity released in pounds,
o the number of times the amount of the release during any 24 hour
period exceeded the reportable quantity,
o the amount in pounds of the single largest daily release, and
o the amount in pounds of the mean daily release.
The Chemical Abstracts Service Registry Number (CASRN) is a unique
identification number that is universally recognized in the chemical
community. The number provides a useful code and insures accurate
identification of the hazardous substance being released.
-------
The Agency is requesting information on the media affected by the release.
The medium or media into which the release occurs may be significant when
evaluating whether the Federal government should respond to the release to
protect human health, welfare or the environment. For example, a release of
phosgene into air may be more hazardous than a release into water where
phosgene is rapidly hydrolyzed into less toxic products. If possible, the
releaser should not only identify the media to which the release is being
made, but also identify the distribution of the hazardous substance into all
media: air, surface water, groundwater and soil.
The total quantity of the hazardous substance released annually is necessary
to the Agency in assessing the potential long-term health and environmental
impact of the release. The Agency will use this data to determine whether
receiving waters are being overloaded or whether the air has become
unhealthful requiring a response by the Federal government.
Furthermore, the number of times the amount of the release during any 24
hour period exceeded the RQ will provide an additional trigger for
investigating whether a response action should be taken by the Fedral
government to protect human health, welfare and the environment. For
example, if a facility is continuously releasing a hazardous substance that
has an RQ of 1 pound (based on its potential carcinogenicity or chronic
toxicity) at an amount that is 100 times this RQ on a daily basis, the
Agency may determine that there is a need to respond to such a release.
Notification of a Statistically Significant Increase (SSI)
Notification of a SSI in the continuous release of a hazardous substance is
to be made to the National Response Center as described in 40 CFR 302.6.
The determination that a particular continuous release represents a SSI is
to be based on:
2-23
-------
o a Nonparametric test which identifies the top 5 percent of releases
as statistically significant, or
o any alternative methodology specifically approved by EPA that can
be demonstrated to the Agency's satisfaction to be superior to the
Nonparametrict test.
The Nonparametric test and the Agency's reasons for choosing it are
described above.
All SSIs in the amount of a continuous release that is stable in quantity
and rate are to be reported to the NRC. Because SSIs represent the release
of larger quantities of the hazardous substance resulting from an unusual
occurence at the facility (e.g., a malfunction or upset condition), it is
more likely that the Federal government will need to respond to the release
to protect human health, welfare and the environment. Notifying the NRC
initiates an assessment by the government of the circumstances surrounding
the release and leads to the determination of the need for a Federal
response.
If the person in charge prefers an alternative test to the Nonparametric
test to detect a SSI, the Agency may accept that alternative under the
conditions below. EPA will require applicants who choose to use an
alternatiave test to present two sets of background information to the EPA
Administrator:
o first, the applicant must describe any assumptions that its test
would make regarding the distribution of releases that would be
expected. EPA will consider the extent to which expectations
regarding the release distribution are releastic and whether the
alternative test depends on the accuracy of these expectations,
o second, the applicant must present evidence that the alternative
test is superior to the Nonparametric test described above for the
applicants facility. The alternative test's effects on the rates
of false positives and false negatives must be presented by the
applicant.
No alternative test is acceptable unless and until the EPA gives written
approval.
2-24
-------
BASIS FOR ADDING TOGETHER CONCURRENT RELEASES
Multiple concurrent releases of the same substance occurring at various
locations within plants or installations upon contiguous grounds that are
under common ownership or leasehold should be added together, both in
determining whether the facility qualifies as a continuous releaser and in
determining whether a SSI has occurred. Such additions are expected
to smooth out variations that occur at single emission or discharge points
such as valves, flanges, and unions. Realistic averaging of the releases to
the environment takes into account the possibly large changes that may
occur for a short period of time at a temporary leak in just one fitting
or other unit of the entire plant. Such averaging also eliminates the
otherwise burdensome need to keep detailed individual records for separate
portions of a large facility.
NEED TO MEET ALL FOREGOING REQUIREMENTS
The reduced frequency of reporting is available only when all the conditions
described in the above paragraphs of Section 2 are fully complied with.
Failure to do so subjects the releaser to all the notification requirements
for episodic releases of an RQ or more of a hazardous substance.
DATA SOURCES EXAMINED TO ASSIST IN RULEMAKING
In the development of proposed continuous release regulations, available
data sources were examined in a search for relevant information on
continuous releases. The information sought included ways of estimating the
magnitude of the variation in the quantity of particular hazardous
substances released to any medium, the time frame over which the release
remained essentially constant, the frequency with which major excursions
occurred in the amount released, and the size of those excursions. It was
expected that the data could be used to determine the stability of the rate
of release of some hazardous substances and possibly also the magnitude of
2-25
-------
significant increases in the quantity released. Although no precisely
applicable data sources were available, a number of sources were
investigated to determine what useful information could be extracted. They
are described below.
NPDES Permit Compliance Reports
In the expectation that information on the constancy of discharge of some
hazardous substance might be inferred from NPDES compliance reports, a
printout of such reports covering a period of several months was obtained
from EPA Region II. The reports were found to contain data that was very
difficult to relate to the magnitude of discharge of individual hazardous
substances or to the variation in the rates of discharge. Most of the data
were expressed in terms of conventional pollutants, e.g., BOD, total
suspended solids, rather than in terms of concentration or quantity of
particular hazardous substances.
Flow rate data were not generally relatable to concentration. Frequency of
monitoring or measurement of any particular parameter was usually unstated
and the measurements were apparently made at irregular intervals. Although
the Agency hoped to utilize the data to define "stable in quantity and rate"
and "statistically significant increase" it was not possible to determine
which numerical values corresponded to variations within permitted limits
that might be interpreted as continuous releases within normal operating
conditions and which values represented out-of-limit upsets.
Philadelphia Air Pollution Control Board Reports
The possible applicability of data collected over a two-year period by the
Philadelphia Air Pollution Control Board on 99 "toxic air contaminants" (84
of which were CERCLA hazardous substances) to continuous release regulation
was examined in order to arrive at an estimate of continuous releasers in
Philadelphia that could be extrapolated to the remainder of the United
States. The maximum daily emission rate reported was compared with the
proposed RQ for the CERCLA hazardous substances. It was assumed that
releases which were equal to or larger than the proposed RQs might qualify
2-26
-------
for the Section 103(f)(2) reduced reporting requirement. No information
existed on whether the reports represented releases which were "stable in
quantity and rate" or whether notification had been given for a "period
sufficient to establish the continuity, quantity and regularity." The data
was very useful in assisting the Agency to estimate the number of facilities
likely to qualify for reduced reporting.
Office of Air Quality Planning and Standards (OAQPS) Data Review
The OAQPS data provide information on air pollutants of concern to the
Agency and are presented on a pollutant by pollutant basis. For each
pollutant, facilities are identified which have emissions either from their
stacks or from other source. The emissions data are presented as an annual
total, with no information on the continuity (if any) of the process
involved. Thus, only a single data point is available for each emission
source, leaving no opportunity to determine the variability of the emission
rate over time which would have been useful to the Agency in defining what
is meant by the term "stable in quantity and rate." The data has proven
useful, however, in identifying the universe of facilities that will benefit
from a reduced reporting requirement for continuous releases.
Summary
As can be seen, the available data sources proved useful in identifying the
need for a regulation leading to reduced reporting. The Agency could not
use the data in deriving a quantitative definition of "stable in quantity
and rate," and could not apply the data in deriving a percentage from the
daily average for purposes of defining a statistically significant increase
in the amount released.
2-27
-------
SECTION 3
REFERENCES
ICF, Inc., Analysis of Continuous Release Notification Strategies for
the Superfund Program, February 1, 1982.
Air Pollution Engineering Manual, U.S. Department of Health, Education
and Welfare, 1967.
3-1
-------
APPENDIX A
TEXT OF MAY 25, 1983 NPRM
PREAMBLE ON CONTINUOUS RELEASES
50 FEDERAL REGISTER 23559-60
A-l
-------
D. Continuous Releases
Section 103(f)(2) of CERCLA exempts certain releases from the general
notification requirements of Section 103(a) and 103(b). Releases may be
exempted if they are "continuous," "stable in quantity and rate," and
notification has been given under Section 103(a) and (b) "for a period
sufficient to establish the continuity, quantity, and regularity" of the
release or under Section 103(c) (which relates to notification of the
existence of certain facilities that are or have been used for storage,
treatment, or disposal of hazardous wastes). Notification of continuous
releases must be given "annually, or at such a time as there is any
statistically significant increase" in the quantity of the hazardous
substance being released.
The main function of CERCLA's notification requirements is to alert
government officials to the existence of a situation that may require a
government response to protect the public health or welfare or the
environment. Since episodic releases are by definition "unanticipated,"
they must be reported as they occur. Continuous releases, on the other
hand, may be predictable and can either be literally continuous or
recurring. Congress recognized in Section 103(f)(2) that CERCLA's
objectives would be satisfied—in the case of continuous releases—by less
frequent notification.
Thus, instead of reporting every release as it occurs, persons in charge are
allowed to report certain continuous releases less often under Section
103(f)(2). The purpose of this section is to reduce unnecessary reports of
releases. The rationale for this approach is that when a release is regular
and of stable quantity, the Agency does not need to be notified each time
such a release occurs in order to have the information necessary to decide
whether a response to the release is necessary. Section 103(f)(2) provides,
however, that notification must still be given (1) under Section 103(c) or
"for a period sufficient to establish the continuity, quantity, and
regularity of such release" and (2) "annually, or at such time as there is
any statistically significant increase in the quantity of any hazardous
substance or constituent thereof released, above that previously reported or
occurring."
The Agency is considering development of a policy to identify:
• The types of releases that qualify for the limited exemption
under Section 103(f)(2), and
• The Section 103(f)(2) notification scheme.
Pending implementation of Section 103(f), EPA will focus its enforcement
efforts on episodic releases (e.g., accidental, one-time, non-routine
releases) exceeding the RQs proposed today which may present a serious risk
of harm to human health or welfare or the environment, rather than on
continuous releases (e.g., routine, continuous, or anticipated intermittent
releases which are incidental to normal manufacturing or treatment processes
or operations of facilities or vessels).
A-2
-------
1. Types of Releases. The Agency is considering defining releases that are
"continuous" and "stable in quantity and rate" to include:
• Literally continuous releases that enter the environment 24 hours a
day, 365 days a year. Examples include hazardous substances leaking
from pipes or lagoons into surface water, leaching into the soil or
groundwater, or evaporating.
• Release continuous during operating hours, such as releases from
some continuous indiustrial processes;
• Releases from batch operations; and
• Routine, anticipated, intermittent releases of hazardous substances
that are incidental to the normal manufacturing or treatment
processes or operations of facilities or vessels. Examples include
releases from relief valves, the maintenance of pollution control
equipment, charging of coke oven batteries, and tank cleaning
operations.
In order for a continuous release which is stable in quantity and rate to be
subject to Section 103(f)(2), notification must be given pursuant to Section
103(c), or under Section 103(a) and (b) for a "period sufficient to
establish the continuity, quantity, and regularity" of the release.
Reporting under Section 103(c) is addressed in 46 FR 22144 (April 15, 1981)
and will not be discussed further here. The Agency is considering several
approaches for defining the "period sufficient," including:
• Establishing a single period (such as a week, a month, six months,
or a year) of reporting that will be required for all releases; or
• Specifying the number of reports, instead of a specific time period,
to establish the "period sufficient."
EPA solicits comments on these or any other approaches for implementing the
Section 103(f)(2) notification provisions.
2. Section 103(f) Notification Scheme. The Agency intends to develop a
notification system for continuous releases. An approach under
consideration involves an annual written notification identifying the
hazardous substance being released, the quantity and rate of such release,
and the environmental media affected. The Agency does not anticipate that
this written notification would be necessarily costly or elaborate; it does
not anticipate that sophisticated analysis to determine the constituents of
a release or that special monitoring to determine the quantity or rate of a
release will be needde. A new notification (either written or by
telephone) would be required if there is a statistically significant
increase in the amount, or statistically significant change in the type of
hazardous substance being released. EPA is also considering whether
applications for federal permits containing information on the nature of
these releases might suffice to show that a release qualifies for the
continuous release exemption.
A-3
-------
Three alternatives for defining "statistically significant increase" are
under consideration.
• Requiring reporting whenever a release falls outside some expected
range based on statistical tests, such as the "Student's t" test;
• Requiring reporting whenever the amount released exceeds the amount
ordinarily released by some pre-established factor, such as 2, 5, or
10 times the daily averge; or
• Letting the releaser determine what is a statistically significant
increase.
EPA requests comments specifically on the most feasible approach for
continuous release notification, the information to be required, what the
Agency should consider a statistically significant increase in the release,
and any other relevant issues. EPA hopes that such information will enable
it to develop a system which imposes a minimal burden on both the regulated
community and the government, while achieving the underlying statutory
objectives.
A-4
-------
APPENDIX B
SUMMARIES OF PUBLIC COMMENTS
ON THE MAY 25, 1983 NPRM'S CONTINUOUS RELEASE PROVISIONS
INTRODUCTION
A total of 45 public comment letters were received by EPA during the months
of July and August, 1983 in response to the continuous release provisions of
the May 25, 1983 NPRM on Superfund notification requirements The majority of
commenters were members of the regulated community, with four government
agencies and two environmental groups comprising the balance.
Most comments centered on the definition or implementation of such terms as
"statistically significant," "period sufficient," "continuous," and "stable
in quantity and rate." Most industry commenters urged EPA to adopt a broad
interpretation of these terms to lessen the frequency of required reporting.
Many industry commenters also recommended that responsibility for defining
these conditions rest with the releaser. Comments by government agencies,
environmental organizations, and one member of industry, however, indicated
that this approach was unacceptable because of the inconsistencies and
possible evasions that would result.
Summaries of all of the comments are arranged in this Appendix according to
the order in which the comments were received.
KENDALL COMPANY
Kendall suggested that "continuous releases covered by either state or
federal permits should be left unchanged. There is no need for additional
reporting." This comment seemed to confuse the reduced reporting
requirement for continuous releases with the complete reporting exemption
for federally permitted releases because any release covered by a federal
permit is by definition a federally permitted release. In such instances,
no reporting would be required unless the release exceeded the permitted
level by an RQ or more. If such exceedences of permit conditions occurred
repeatedly and the releaser could show continuity and regularity, the
releases might be considered to be continuous releases.
AIR PRODUCTS
Air Products recommended that if a facility is covered by existing
monitoring rules, it should not be covered by Section 103(f) of CERCLA,
particularly in the case of continuous releases of volatile organic
compounds (VOCs) from facilities governed by the SIP process or NSPS
guidelines. Air Products requested that EPA make adjustments to 103(f)
reporting requirements to reflect the wide variety of reporting rules
already in place or in the process of rulemaking to avoid duplication of
existing programs.
B-l
-------
VULCAN CHEMICALS
Vulcan maintained that facilities which emit non-photoreactive VOCs, if not
entirely exempt from the notification requirements of CERCLA under the
federally permitted release provision, should at least fall within Section
103(f). To qualify for the continuous release exemption, such facilities
should be required to file only one report stating that they emit these
solvents.
ADOLPH COORS
Adolph Coors generally endorsed EPA's approach to the continuous release
issue. Coors made the following recommendations:
o EPA should give the releaser primary responsibility to determine
the period of time sufficient to establish continuity. Coors
suggested that such a practice will not lead to abuse because the
Agency has adequate authority to request additional data, if
required.
o Whether there has been a "statistically significant increase"
should be determined by the releaser.
THE LIGHT COMPANY
The Light Company supported EPA's proposal to emphasize reporting
requirements for episodic releases exceeding RQs rather than continuous,
routine, or anticipated intermittent releases incidental to normal
operations. Such an approach would, in its opinion, reduce redundant
reporting and promote efficiency in both the Agency and the regulated
community.
RSR CORPORATION
RSR Corporation suggested that to avoid duplicative reporting requirements,
continuous releases of substances subject to other legislation (e.g., CAA,
CWA) should be governed by amendments to the appropriate legislation rather
than by a new reporting system under CERCLA. Information contained in
permits covering such releases should be sufficient to satisfy the
notification requirements of Section 103(f)(2) of CERCLA.
PENNZOIL
Pennzoil approved of EPA's plan to broaden the definition of continuous
release to include not only literally continuous releases but also those
that are continuous during operating hours, those from batch operations, and
routine, anticipated, or intermittent releases that are incidental to normal
manufacturing or treatment processes or operations.
B-2
-------
NATIONAL AGRICULTURAL CHEMICALS ASSOCIATION (NACA) —
THE FERTILIZER INSTITUTE (TFI)
With the exception of the last paragraph of each comment, the comments on
continuous releases by NACA and TFI were identical. The identical portions
of their comments included the following recommendations:
o EPA's continuous release policy is heading in the right direction
because it will provide the Agency with relatively few but
comprehensive reports that analyze rates of release and project
the likelihood of future releases.
o EPA should define "routine, anticipated intermittent releases"
broadly to cover all routine discharges "that can be reasonably
estimated for the future based on experience." (The comment,
however, did not specify whose experience.)
o The Agency should take an ad hoc approach toward determining the
"continuity and quantity" of releases.
o A rule requiring a certain number of reports or period of time
for reporting is inappropriate because some facilities already
have substantial data from past operations enabling them to
accurately anticipate future releases.
o The determination of what constitutes a "statistically
significant increase" should be left to the releaser.
o EPA can correct any abuse of this discretion since the annual
reports, which must be prepared in any event, would reveal the
actual release history for the year.
In addition, NACA urged EPA to design its final regulations to allow
facilities to show that they are entitled to annual reporting on the basis
of existing as well as new data.
The Fertilizer Institute made a separate request that when continuous
releases can be estimated reliably on the basis of material balances or
other procedures, that continuous monitoring not be required.
CONOCO
Conoco.expressed the following views on the continuous release issue:
o EPA has correctly construed congressional intent in its examples
of releases considered to be "continuous" and "stable in quantity
and rate" for purposes of Section 103(f)(2).
o Under 103(f), the Agency should create a one-time notification
system, because continuous releases by their nature are expected
to be similar in composition, quantity, and rate from year to
year.
B-3
-------
OLIN
Olin seemed to suggest that, in effect, almost any continuous release could
be considered a federally permitted release. Olin's reasoning was as
follows:
o If a continuous release was not already federally permitted, the
data necessary to establish a continuous release exemption could
be used instead by the "permitting authority . . . to decide
whether or not the release warrants permit limits, monitoring
only, or no action.
o "Unless the decision is 'no action,' regulation by a permit would
qualify the release to be treated as a federally permitted
release."
o "If 'no action' is appropriate, this should satisfy the Agency
that no further notification is needed."
Coupled with this approach was a suggestion by Olin that a statistically
significant increase be defined as "an increase great enough to change the
release from 'no action' to 'federally permitted' status." Although Olin
did not state the following explicitly, the net effect of its proposals
would seem to be the elimination of all continuous release reporting
requirements.
AMERICAN TEXTILES MANUFACTURERS INSTITUTE (ATMI)
ATMI made the following suggestions:
o The "period sufficient" should consist of a single report where
releases are "continuous" and "stable in quantity and rate."
o As to "statistically significant releases," EPA should use a
factor of 2, 5, 10, or 100 times the daily average, depending on
the potential impact on the environment of the release.
o The National Response Center (NRC) should receive and distribute
to the appropriate state and federal agencies all hazardous
substance reports.
AMERICAN PAPER INSTITUTE AND NATIONAL FOREST PRODUCTS ASSOCIATION
(API/NFPA)
API/NFPA made the following comments:
o EPA has properly decided to focus enforcement efforts on episodic
releases exceeding proposed RQs.
B-4
-------
o The Agency should propose implementing regulations for Section
103(f)(2) continuous releases as soon as possible.
GENERAL BATTERY
General Battery stated that EPA's proposed application of reporting
requirements to routine operations (e.g., releases from relief valves,
maintenance of pollution control equipment, charging of coke oven batteries,
tank cleaning operations) will be "extremely difficult to compute in of
reportabterms le quantities and will be extremely difficult to enforce."
DOW CHEMICAL
Dow made the following recommendations:
o Continuous releases should be established through notification or
from information already in EPA's possession;
o EPA should not require additional and duplicative reporting;
o EPA should focus on sudden and accidental releases, not routine
releases; and
o EPA should ensure that routine releases "contemplated as the
normal occurrence of manufacturing operations" are exempt from
reporting.
STANDARD OIL COMPANY (INDIANA)
Standard Oil suggested that EPA promulgate simple and flexible continuous
release reporting requirements, and that for relatively stable or
predictable releases, a one-time notification should suffice.
AMERICAN PETROLEUM INSTITUTE (API)
API made the following recommendations:
o EPA should rely on the discretion of the releaser to determine
when a "statistically significant release" has occurred. API
maintained that this approach is supported by the discretion
given the releaser under Sections 103(a) and (b) to determine
whether a given release exceeds an RQ.
o Any anticipated release should be eligible to qualify for a
continuous release exemption.
o The releaser should be permitted to establish the continuity,
quantity, and regularity of the release through the use of
available monitoring data or estimates based on best engineering
-------
judgment. API stated that because an error in initial quantifica-
tion will lead to a "change" in the release and since this
"change" must be reported to the Agency, the releaser will have
sufficient incentive to quantify accurately the release.
o EPA should not set an arbitrary required number of notifications
to establish continuity and regularity because in most cases the
releaser will be able to identify the nature and quantity of the
release in a single notification.
o The Agency should not require releasers to file an annual report.
Because the primary purpose of notification to the NRC is to
facilitate a decision regarding emergency response, an annual
report would be of little value.
o The concept of "statistically significant increase" should be
applied through a trigger mechanism; specifically, when an
"upset" condition occurs (operational parameters exceed a
predetermined range), and in the releaser's discretion a
significant increase in emissions results, the NRC must be
notified.
o Whether a significant increase has occurred should be a
case-by-case determination, dependent upon the substance
released, media of release, normal amount of release, and
increment of increase.
o In determining an operational definition for "statistically
significant," API recommended that formal statistical
measurements and arbitrary multiples not be used because they do
not accurately reflect the environmental significance of an
increase in a release.
ELI LILLY
Eli Lilly made the following recommendations:
o EPA should promulgate regulations for continuous releases as soon
as possible.
o The Agency's proposed definition of "continuous" and "stable in
quantity and rate" is satisfactory.
o The "period sufficient" should be one year, and the reporting
frequency"should be annual.
o In their reports to the NRC, releasers should not be required to
include any quantities permitted by federal, state, or local
authorities.
B-6
-------
CHEMICAL MANUFACTURERS ASSOCIATION (CMA)
CMA made the following comments:
o EPA's proposed definition of releases that qualify as continuous
(e.g., literally continuous releases, releases continuous during
operating hours, releases from batch operations, and routine,
anticipated, intermittent releases) is fair and comprehensive.
o The "period sufficient" requirement should be satisfied by an
initial report containing a description of the types of releases
and an estimation of the levels of those releases for the year.
o If a releaser underestimates significantly any release, he should
be required to report such a release under the "statistically
significant increase" provision of Section 103(f).
o The terms "statistically significant increase" and "period
sufficient" should be determined on a site-specific and release-
specific basis using "best engineering judgment."
SOUTHERN CALIFORNIA EDISON (SCE)
SCE made the following suggestions concerning air emissions and the reduced
reporting requirement for continuous releases:
o Those substances (most stack gases) for which no de minimis
policy has been established under CAA regulation but which are
emitted under federal permits continuously and are stable in
quantity and rate, should qualify for the 103(f)(2) reduced
reporting requirement for continuous releases.
o Only those CERCLA-designated hazardous substances released on an
accidental, one-time or non-routine basis should be subject to
103(f)(2) reporting requirements. These releases, however,
should not include releases incidental to normal operations.
o These suggested exceptions are consistent with congressional
intent to minimize reporting burdens on the regulated community
and focus resources only on the most serious releases.
MISSISSIPPI POWER AND LIGHT (MP&L)
MP&L suggested that EPA give authority to the releaser to determine what is
a statistically significant increase, because the "releaser has more
intimate knowledge of his facility and thus could better decide what is
significant."
B-7
-------
DETROIT EDISON
Detroit Edison made the following recommendations:
o EPA should interpret broadly Section 103(f)(2) for stack
emissions to exclude monitoring and reporting of day-to-day
emissions resulting from normal plant operations.
o In determining "period sufficient," only a very short time period
should be necessary for fossil-fuel fired power plants.
o To ease the reporting burden on industry while still gathering
adequate information, a few typical plants could be selected to
establish "continuity, quantity, and regularity." Then EPA would
only have to require reporting on an annual basis or when
emissions change significantly from the rest of the industry.
KIMBERLY-CLARK
Kimberly-Clark stated that:
o It supports EPA's intention to require only annual reporting of
continuous releases, unless there is a statistically significant
increase in the release.
o EPA needs to define more clearly the meaning of "continuous or
routine intermittent discharges, "taking into account both point
and nonpoint sources.
o The Agency should clarify the meaning of the phrase
"statistically significant increase" by using a pre-established,
scientifically based factor to trigger a reporting requirement.
DIAMOND SHAMROCK
Diamond Shamrock made the following suggestions:
o EPA should regulate continuous releases under existing permit
programs; if substantial continuous releases do not fit within
these programs, the programs should be modified.
o The concept of "statistically significant increase" should not be
based on the "Student t" test.
GENERAL MOTORS CORPORATION (GM)
GM believed that industrial discharges to POTWs are typically continuous
releases, stable in quantity and rate, and should be exempted from CERCLA
Section 103(a) and (b) notification requirements. Further, GM stated that
the less burdensome requirements of Section 103(f)(2) may be unnecessary to
B-8
-------
the extent that they duplicate reports required under local sewer ordinances
as well as state and municipal permit programs.
A yearly notification of continuous releases should be sufficient, coupled
with the provision for reporting "statistically significant increases."
However, of the three possible alternatives for defining "statistically
significant increase" discussed in the preamble to the NPRM (page 23559),
the third alternative, which allows the releaser to determine what is a
statistically significant increase, is objectionable because "unscrupulous
operators could be expected to take advantage of such a provision."
ATLANTIC RICHFIELD COMPANY (ARCO)
ARCO agreed that:
o Annual reporting should be sufficient, provided that the release
is "stable in quantity and rate," and that notification is not
provided through some other mechanism; and
o Releases which are "intermittent and anticipated" should be
considered "continuous" because, in most cases, they are an
unavoidable part of many industrial processes.
ARCO suggested that:
o The "period sufficient" to determine the continuity, quantity,
and regularity of a release should be determined by the person
responsible for reporting the release. ARCO suggested that EPA
set an upper time limit of one year for establishing the
qualifications of a release; and
o The person responsible for reporting a release should specify or
propose the method(s) by which a "statistically significant"
increase will be determined, and EPA should include the
statistical methods which would be acceptable in its final rule.
PENNSYLVANIA POWER.& LIGHT COMPANY (PP&L)
PP&L supported the Agency's continuous release policy but also suggested:
o EPA should explicitly consider recognizing distribution capacitor
failures (a "recurring" release) as a continuous release because
most incidents result in little or no environmental impact and
mechanisms exist to ensure that incidents needing EPA response
are reported; and
o The period for establishing that a release is continuous should
be short and should be based on the number of reports, not some
arbitrary time period. After a given number of occurrences
(regardless of time period), a determination can be made
concerning whether a release is continuous.
B-9
-------
SOHIO
The continuous release notification policy should be flexible and broadly
construed. SOHIO suggested that:
o Any anticipated release should be a candidate for the reduced
reporting requirement for continuous releases; this includes any
releases incidental to normal manufacturing or operation of a
facility;
o The releaser should be allowed to present the particulars of the
release in its initial notification;
o A twenty-four hour period for purposes of determining the amount
of the release should be employed;
o The releaser should not be required to aggregate releases across
all facilities within a plant; however, such an aggregation
should be allowed as long as there is an adequate description of
the sources which are included in the filing;
o The continuity, quantity, and regularity of the release should be
established by the releaser through the use of actual monitoring
data or estimates based on best engineering judgment. There will
be substantial incentive on the part of the releaser to quantify
accurately the release, since an error will necessitate reports
to notify the Agency of "changes" in the release;
o EPA should not specify an arbitrary number of notifications,
because in most instances the releaser will be able to pinpoint
the nature and quantity of the release in the initial
notification based on historic information and normal operating
practice. On the other hand, the Agency should not foreclose
multiple notifications in appropriate circumstances; and
o EPA should require the reporting of all upset conditions which
will cause a change in emissions. The term "upset," has a clear
meaning to those involved with industrial processes, whereas the
term "statistically significant," can mean different things to
different industries. This notification scheme, undertaken by
the releaser, identifies the types of unanticipated, episodic
changes which could prompt a government response.
POTOMAC ELECTRIC POWER COMPANY (PEPCO)
PEPCO supported the Agency's flexible definition of the kinds of releases
that are considered continuous, but urged EPA to provide guidance concerning
what period is sufficient to establish the continuity, quantity and
regularity of a release.
B-10
-------
o PEPCO suggested that a source's "potential to emit" a hazardous
air pollutant be considered when determining if a release
qualifies as "continuous."
o A "statistically significant" increase should be defined as any
emission of a hazardous pollutant from a source above its defined
"potential to emit." (Examples include emissions increases
resulting from changes in fuel characteristics or removal of
government restrictions on a source's hours of operation.)
ROOKS, PITTS, FULLAGER AND POUST on behalf of
ILLINOIS MANUFACTURERS' ASSOCIATION
o A release of a hazardous substance which is not a criteria
pollutant or is not otherwise regulated should be exempt from
notification requirements either under the federally permitted
release exemption or under the reduced reporting requirement for
continuous releases.
o The continuous release definition should include: literally
continuous releases that enter the environment 24 hours a day,
365 days a year; releases continuous during operating hours; and
routine, anticipated, intermittent releases incidental to normal
manufacturing or treatment processes or operations of facilities
or vessels.
o Facilities should have to notify EPA only once concerning their
releases, with additional notifications to be made if the release
substantially exceeds the continuous release rate.
o The releaser should be allowed to determine what is a
"statistically significant" increase.
MIDDLE SOUTH SERVICES, INC. (MSS)
MSS urged EPA to use other submissions besides NRC reports to determine the
eligibility of a release as continuous. MSS was concerned, however, that the
use of other submissions "may be somewhat constrained by the statute's
requirement that reporting be under specific CERCLA sections." In addition,
MSS suggested that:
o EPA not establish a specific time period in which to demonstrate
that a release is continuous, because some releases may be
reported more frequently than necessary, and some releases, due
to seasonal variations, may not be reported at all. Instead, the
releaser should be allowed to.provide the Agency with the
information it believes shows that a release qualifies for the
reduced reporting requirement for continuous releases; and
o The releaser should be allowed to determine what constitutes a
statistically significant increase based on "his knowledge of his
operation and what can be expected from it."
B-ll
-------
NATIONAL FOOD PROCESSORS ASSOCIATION (NFPA)
NFPA offered the following comments regarding the continuous release issue:
o NFPA concurred with EPA that episodic releases exceeding RQs
present a more serious risk of harm to the public health or
welfare or the environment than continuous releases.
o The Agency should simplify reporting requirements for continuous
releases by identifying specific releases which do not pose
significant risks to human health and the environment and
therefore do not need to be reported (e.g., releases of ammonia
to the air during routine maintenance of refrigeration units).
o The releaser should determine what is a statistically significant
increase that triggers the reporting requirement.
o Annual reports are unnecessary and burdensome.
PHILLIPS PETROLEUM COMPANY
Phillips reserved comment on continuous releases until specific regulations
are proposed by EPA.
COUNTY SANITATION DISTRICTS OF LOS ANGELES COUNTY (Sanitation Districts)
The Sanitation Districts recommended that notification of continuous
releases be in a written form because formal documentation of a problem is
more important than immediate telephone notification.
NATURAL RESOURCES DEFENSE COUNCIL, INC. (NRDC)
NRDC suggested that EPA:
o Provide specific guidelines for releasers on sampling plans to
establish stability and continuity of releases, and to provide
estimates of parameters (such as average flow rate and variation
in discharge) useful in determining if a statistically
significant increase has occurred; and
o Clearly define what a statistically significant increase is by
providing: "alternatives for sampling schemes for release
monitoring; procedures for checking distributional assumptions
about release data (such as the use of normal probability paper);
and suggestions for appropriate statistical tests with desired
level of significance." NRDC maintained that allowing releasers
to define their own monitoring and reporting standards would
almost certainly result in conflicting and inaccurate continuous
release reports.
B-12
-------
TEXAS DEPARTMENT OF WATER RESOURCES (Texas DWR)
The Texas DWR suggested that the continuous relese notification system be
structured as follows:
o Continuous releases should be documented through a registration
system; upon proper registration, all notification requirements
would be satisfied.
o The registration should document the location, source, impact,
and character of the discharge.
o If at any time after original notification a "significant change"
in discharge occurs, updates should be filed with the appropriate
regulatory agencies.
o Once registered, continuous releases should be individually
reviewed to determine whether periodic reporting or other
response actions are required.
EVANS, KITCHELL AND JENCKES, P.O. on behalf of PHELPS DODGE CORPORATION
Phelps Dodge provided the following recommendations:.
o EPA should give a broad interpretation to the term "continuous
release" so that notifications can be reduced or eliminated in
cases where a government response is unlikely.
o Because continuous releases are predictable and can be monitored,
a less frequent notification scheme is appropriate.
o Depending on the definition of "stable in quantity and rate,"
continuous air releases from stationary sources may not qualify
for the 103(f)(2) exemption.
o The unique circumstances surrounding any continuous release argue
in favor of determining "statistically significant" on a
case-by-case basis.
B-13
-------
HUNTON AND WILLIAMS on behalf of
UTILITY WATER ACT GROUP (UWAG)
UTILITY AIR REGULATORY GROUP (UARG)
UTILITY SOLID WASTE ACTIVITIES GROUP (USWAG)
EDISON ELECTRIC INSTITUTE (EEI)
NATIONAL RURAL ELECTRIC COOPERATIVE ASSOCIATION (NRECA)
AMERICAN PUBLIC POWER ASSOCIATION (APPA)
The UWAG, UARG, USWAG, EEI, NRECA, and APPA made the following comments
concerning continuous releases:
o EPA correctly rejected an extensive monitoring system to document
fluctuations in the nature or rate of releases.
o For purposes of Section 103(f)(2), accidental releases cannot
always be considered episodic.
o "Period sufficient" should be defined as a certain time period
(no longer than a month) or a certain number of reports,
determined on a case-by-case basis to maintain flexibility.
o Because Section 103(f)(2) calls for notification annually or when
there is a statistically significant increase, a releaser that
reports statistically significant increases should not be
required to report annually.
o As long as a source operates normally under the range of
conditions existing when it established the continuity of a
release, EPA should not require it to monitor for statistically
significant increases. Imposing an affirmative monitoring
requirement would defeat the purposes of Section 103(f).
o Once a release is established as continuous, further reporting
should be required only if existing monitoring reveals a
statistically significant increase or if a material change in
operations occurs.
o The Agency should use its rulemaking authority to implement
Section 103(f) for entire categories of sources based on the
experience of representative sources within the category. The
electric utility industry, for example, could provide representa-
tive data documenting the quantities of various hazardous
substances emitted under the full range of operating procedures
and techniques applicable throughout the industry. An individual
source could then establish the quantity of its continuous
B-14
-------
releases by stating its relevant operating procedures and
techniques and referring to the source-wide data.
AMERICAN MINING CONGRESS (AMC)
AMC generally endorsed EPA's approach to continuous releases. AMC made the
following recommendations:
o "Period sufficient" should not be measured by a specific time
period or number of reports. Where a releaser knows, based on
past experience, the identity, frequency, and rate of its
releases, a single notification should satisfy the requirements
of Section 103(f)(2).
o EPA should allow releasers to qualify for the reduced reporting
requirement for continuous releases based on information
contained in applications for federal, state, and regional
permits. Such an approach would reduce reporting burdens on both
the Agency and the regulated community.
KOPPERS COMPANY
Koppers expressed the following views on the continuous release issue:
o EPA should define "period sufficient" as an initial report
describing the types of releases the releaser anticipates in the
next 12 months; no single time period will work well for all
types of possible releases.
o The definitions of both "period sufficient" and "statistically
significant increase" should incorporate the concept of "best
engineering judgment."
o Sites of past industrial activity which have been reported under
Section 103(c) of CERCLA should be exempt from reporting
continuous releases.
AMAX ENVIRONMENTAL SERVICES
AMAX generally supported EPA's continuous release policy. AMAX made the
following suggestions:
o EPA should define "continuous release" by type of activity and
"significant increase" by significant change in the activity
because of the difficulty of quantifying the amounts released by
many activities.
o Until the Agency establishes reporting procedures, it should not
take enforcement action for failure to report continuous
releases.
B-15
-------
U.S. DEPARTMENT OF THE INTERIOR (DPI)
i'
DOI made the following suggestions:
o "Period sufficient" must be clearly defined as a specific number
of periodic reports rather than an arbitrary time period.
o "Statistically significant increase" should be defined by a
statistical test because such tests are specifically designed to
determine significant variances from an average.
ARMY CORPS OF ENGINEERS
o The definition of a "period sufficient" should encompass both a
specific time period and a specified number of reports.
o "Statistically significant increase" should be defined in terms
of a pre-determined range, not a pre-established factor, because
even a factor of only two times the "daily average of certain
hazardous substances" can be lethal for some aquatic biota.
ENVIRONMENTAL DEFENSE FUND (EOF)
EOF urged EPA to propose immediately a notification system for continuous
releases. EOF believed that data on hazardous substance releases of a
routine nature are extremely valuable in setting priorities for future
regulatory activity.
U.S. NUCLEAR REGULATORY COMMISSION
The Nuclear Regulatory Commission recommended that for purposes of defining
"statistically significant increase," EPA permit concentration levels to be
averaged over periods of one year.
-------
COMMENTERS ON CONTINUOUS RELEASES
Kendall Company
Air Products and Chemicals
Vulcan Chemicals
Adolph Coors
The Light Company
RSR Corporation
Pennzoil
National Agricultural Chemicals Association and
The Fertilizer Institute
Conoco
Olin
American Textile Manufacturers Institute
American Paper Institute and National Forest
Products Association
General Battery
Dow Chemical
Standard Oil Company (Indiana)
American Petroleum Institute
Eli Lilly
Chemical Manufacturers Association
Southern California Edison
Mississippi Power and Light
Detroit Edison
Kimberly-Clark
Diamond-Shamrock
General Motors
Atlantic Richfield
Pennsylvania Power and Light
SOHIO
Potomac Electric Power Company
Rooks, Pitts, Fullagar and Poust on behalf of
Illinois Manufacturers' Association
Middle South Services
National Food Processors Association
Phillips Petroleum
County Sanitation Districts of Los Angeles County
Natural Resources Defense Council
Texas Department of Water Resources
Evans, Kitchell and Jenckes, P.C. on behalf of
Phelps Dodge Corp.
Hunton and Williams on behalf of
Utility Water Act Group;
Utility Air Regulatory Group;
Utility Solid Waste Activities Group;
Edison Electric Institute;
B-17
-------
National Rural Electric Cooperative Association;
American Public Power Association
American Mining Congress
Koppers Company
AMAX Environmental Services
U.S. Department of the Interior
Army Corps of Engineers
Environmental Defense Fund
U.S. Nuclear Regulatory Commission
B-18
-------
APPENDIX C
EXAMPLES OF USING STATISTICAL TESTS CONSIDERED
Control Chart Test
The Control Chart test should be used when at least 30 releases have been
recorded. The Control Chart test is a dynamic test in that it uses prior
releases to form a data set in which the mean and standard deviation are
calculated. When a new release is recorded, its value is compared to 1.645
times the standard deviation. If the new recorded release is less than
(1.645)*(S), then the release is not considered statistically significant
and does not need to be reported to the NRC. The release is then added to
the release history and a new standard deviation is calculated. If the next
new release is greater than 1.645 times the newly calculated standard
deviation, then the release must be reported to the NRC and the release is
not included in the release history; there is no recalculation of the
standard deviation.
Example
Release
Release
Release
Release
Release
Release
Release
Release
Release
Release
1 =
2 =
3 =
4 =
5 =
6 =
7 =
8 =
9 =
10 =
10
15
70
30
60
50
40
80
10
50
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
Release
Release
Release
Release
Release
Release
Release
Release
Release
Release
11
12
13
14
15
16
17
18
19
20
= 30
= 40
= 5
= 90
= 47
= 12
= 25
= 33
= 75
= 10
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
Release
Release
Release
Release
Release
Release
Release
Release
Release
Release
21 =
22 =
23 =
24 =
25 =
26 =
27 =
28 =
29 =
30 =
110
50
105
80
40
20
60
10
90
40
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
pounds
C-l
-------
For the 30 releases shown above, the mean is 46.2 and the standard deviation
is 30.25 (the general method for calculating a mean and standard deviation
is shown in the section entitled "Girubb's Test"). Any release greater than
49.76 pounds (1.645 times 30.25) must be reported to the NRC.
If release 31 equals 40 pounds, it is less than the trigger value of 49.76
pounds and is not reported to the NRC. This release is added to the data
set and a new trigger value is calculated. The standard deviation now
equals 29.77 and 'the trigger value is equal to 48.96.
If release 32 is equal to 50 pounds, it must be reported to the NRC because.
it is greater than the trigger value of 48.96 pounds; The trigger value
remains the same for the next release because the 50 pound release is not
added to the release history and no new standard deviation is calculated.
Dixon Criterion
The Dixon Criterion compares the newest release to the largest release in
the data set and divides that difference by the difference between the
newest release and the smallest release in the data set, as shown in the
equation below:
C-2
-------
The number so derived is compared to Table C-l relating reporting triggers
to the number of releases in the data set. If the number is larger than the
appropriate number in the table, then a report must be made to the NRC.
Example
If the first three releases by a continuous releaser were 10, 15 and 20
pounds, in that order, then:
20 - 15
C - 0.5
20 - 10
The reporting trigger for the third release is not exceeded and the release
need not be reported to the NRC. A fourth release of 60 pounds would,
however, exceed the reporting trigger for a fourth release and must be
reported to the NRC.
60 - 20
C =—: = 0.8
60 - 10
C-3
-------
Grubb's Test
The Grubb's test involves the calculation of a test value and a comparison
of this test value with the "Grubb's Statistic." The test value is
calculated by subtracting the mean of all past releases (except those
determined to be statistically significant increases) from the newest
release, and dividing this difference by the standard deviation of all
releases. That is,
where X is the most recent release, X is the mean or average of all past
n
nonstatistically significant releases and S is the standard deviation for
all statistically significant releases. If the test value is less than the
corresponding Grubb's statistic shown in Table C-2, then the release is not
considered a statistically significant increase and need not be reported to
the NRC. If the test value is equal to or greater than the Grubb's
statistic in Table C-2, then the release must be reported to the NRC. A SSI
value should not be included in future Grubb's test calculations. The
general methods for calculation of X and S are shown below.
n - 1
C-4
-------
Example
If the first three releases by a continuous releaser were 10, 15 and 20
pounds, in that order, the standard deviation (S) is 5, the mean is 15 and
the Grubb's test value is calculated as follows:
A (10 - 15)2 + (15 - 15)2 + (20 - 15)2
!___________ L. 5
(3- 1)
20-15
Grubb's Test Value = — - 1
5
Since Table C-2 gives a Grubb's statistic of 1.15 for the third observation
and the Grubb's test value just calculated is 1, no report need be made to
the NRC. IF the fourth release is 60 pounds, the mean equals 26.25 and the
standard deviation is 22.87. The Grubb's test value is calculated below.
60 - 26.25
Grubb's Test Value = = 1.48
22.87
C-5
-------
The Grubb's test value is 1.48 while the Grubb's statistic is 1.46
indicating that this 60 pound release must be reported to the NRC.
Student t Test
The Student t test is identical to the Grubb's test except that the mean and
standard deviation do not include the newest release. All releases in the
release history are used to calculate the mean and standard deviation.
Student t Value -.
S
Example
If the first three releases by a continuous releaser were 10, 15 and 20
pounds, and it was desired to calculate the Student t value for the fourth
release of 25 pounds and the standard deviation is 5.
25 - 15
Student t Value = = 2
C-6
-------
The Student t value is less than the value in Table C-3 for the fourth
release, and this release need not be reported ''to the NRC. A fifth release
of 40 pounds yields a Student t value of 3.48, calculated in the same way,
and would require a report to the NRC.
40 - 17.50
Student t Value = = 3 48
6.45
C-7
-------
TABLE C-l
DIXON CRITERION REPORTING TRIGGERS
Trigger Value at
Number of Observations 95 Percent Confidence Level
3
4
5
6
7
8
9
10
12
14
16
18
20
25
30
0.941
0.765
0.642
0.560
0.507
0.468
0.437
0.412
0.376
0.349
0.329
0.313
0.300
0.277
0.260
C-8
-------
TABLE C-2
GRUBB'S TEST REPORTING TRIGGERS
Number of Observations Grubb's Statistic
3
4
5
6
7
8
9
10
12
14
15
16
18
20
30
1.15
1.46
1.67
1.82
1.94
2.03
2.11
2.18
2.29
2.37
2.41
2.44
2.50
2.56
2.74
C-9
-------
TABLE C-3
STUDENT'S t TEST REPORTING TRIGGERS
Number of Observations Student's t Statistic
3
4
5
6
7
8
9
10
12
14
15
16
18
20
30
2.353
2.132
2.015
1.943
1.895
1.860
1.833
1.812
1.782
1.761
1.753
1.746
1.734
1.725
1.697
C-10
------- |