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           DO NOT CITE  OR QUOTE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           DO NOT CITE  OR QUOTE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

-------