[Federal Register Volume 59, Number 63 (Friday, April 1, 1994)]
[Unknown Section]
[Page 0]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 94-7241]


[[Page Unknown]]

[Federal Register: April 1, 1994]


_______________________________________________________________________

Part II





Environmental Protection Agency





_______________________________________________________________________



40 CFR Parts 63 and 70




Hazardous Air Pollutants: Proposed Regulations Governing Constructed, 
Reconstructed or Modified Major Sources; Proposed Rule
ENVIRONMENTAL PROTECTION AGENCY

40 CFR Parts 63 and 70

[FRL-4849-5]
RIN 2060-AD06

 
Hazardous Air Pollutants: Proposed Regulations Governing 
Constructed, Reconstructed or Modified Major Sources

AGENCY: Environmental Protection Agency (EPA).

ACTION: Proposed rule.

-----------------------------------------------------------------------

SUMMARY: The proposed rule would implement the provisions in section 
112(g) of the Clean Air Act (Act). The section 112(g) requirements are 
new provisions of the 1990 amendments to the Act. Section 112(g) 
applies to the owner or operator of a constructed, reconstructed, or 
modified major source of hazardous air pollutants (HAP). After the 
effective date of a title V permit program in a State, all owners or 
operators of major sources that are constructed, reconstructed, or 
modified in that State would be required to install maximum achievable 
control technology (MACT). The proposed rule establishes requirements 
and procedures for the owners or operators to follow in order to comply 
with section 112(g). The proposed rule also contains guidance 
permitting authorities in implementing section 112(g). When no 
applicable Federal emission limitation has been promulgated, the Act 
requires the permitting authority (generally a State or local agency 
responsible for the program) to determine a MACT emission limitation on 
a case-by-case basis. Procedures are proposed for making these 
determinations. The proposed rule includes proposed de minimis emission 
rates for all of the listed HAP. These de minimis values are critical 
in defining the scope of the section 112(g) requirements. Pursuant to 
section 112(g), an owner or operator may provide emission offsets to 
avoid requirements for modifications. The proposed rule provides 
procedures for providing and reviewing offset demonstrations, including 
procedures for evaluating whether emission offsets are ``more 
hazardous'' than emission increases being offset. Finally, the proposal 
includes a number of clarifying amendments to previously proposed or 
promulgated regulations. These proposed amendments would clarify the 
relationship between those requirements and section 112(g) of the Act.

DATES: Comments. Comments must be received on or before June 30, 1994. 
The EPA does not intend to extend this date.
    Public Hearing. If anyone contacts the EPA requesting a public 
hearing by May 2, 1994, a public hearing may be held June 1, 1994 
beginning at 10 a.m.
    Request to Speak at Hearing. Persons wishing to present oral 
testimony must contact the EPA by May 2, 1994.

ADDRESSES: Comments. Comments should be submitted (in duplicate if 
possible) to: Air Docket (LE-131), Attention Docket Number A-91-64 (see 
Docket section below), room M1500, U.S. Environmental Protection 
Agency, 401 M Street, Southwest, Washington, DC 20460. The EPA requests 
that a separate copy also be sent to the contact person listed below. 
The docket is located at the above address in room M-1500, Waterside 
Mall (ground floor), and may be inspected from 8:30 a.m. to 12 p.m. and 
1 p.m. to 3 p.m., Monday through Friday. The proposed regulatory text 
and other materials related to this rule making are available for 
review in the docket. A reasonable fee may be charged for copying 
docket materials.
    Public Hearing. If anyone contacts the EPA requesting a public 
hearing, it will be held at the EPA's Office of Administration 
Auditorium, Research Triangle Park, North Carolina. Persons interested 
in attending the hearing or wishing to present oral testimony should 
notify Ms. Theresa Adkins, U.S. Environmental Protection Agency, 
Research Triangle Park, North Carolina 27711, telephone number (919) 
541-5502.
    Docket. Docket No. A-91-64, containing supporting information used 
in developing the proposed rule is available for public inspection and 
copying between 8:30 a.m. and 3:30 p.m., Monday through Friday, at the 
EPA's Air Docket, room M1500, U.S. Environmental Protection Agency, 401 
M Street, SW., Washington, DC 20460. A reasonable fee may be charged 
for copying.

FOR FURTHER INFORMATION CONTACT: Mr. Timothy Smith (regulatory issues), 
telephone (919) 541-4718, Dr. Jane Caldwell-Kenkel (hazard ranking 
issues), telephone (919) 541-0328, or Ms. Lynn Hutchinson (MACT 
determination procedures), telephone (919) 541-5624, Emission Standards 
Division (MD-13), U.S. Environmental Protection Agency, Research 
Triangle Park, North Carolina 27711.

SUPPLEMENTARY INFORMATION: The information presented in this preamble 
is organized as follows:

I. Summary of Proposed Rule
A. Requirements for Constructed and Reconstructed Major Sources
B. Requirements for Modified Major Sources
C. Software Illustrating the Proposed Requirements
II. Background Discussion
A. Clean Air Act Amendments: Section 112
B. Clean Air Act Amendments: Provisions for Constructed, 
Reconstructed and Modified Major Sources of Hazardous Air Pollutants
C. Process To Develop the Proposed Rule
III. Summary and Rationale for Sec. 63.40 Through 63.47, and 
Sec. 63.49, of the Proposed Rule
A. Section 63.40--Applicability
B. Section 63.41--Definitions
C. Section 63.42--Requirements for Constructed and Reconstructed 
Major Sources
D. Section 63.43--Requirements for Modified Major Sources
E. Section 63.44--de minimis Levels
F. Section 63.45-- MACT determinations
G. Sections 63.46 and 63.47. Offset Demonstration
H. Section 63.49. Requirements for Emission Units Subject to a 
Subsequently Promulgated MACT Standard or MACT Requirement
IV. Proposed Approach for Demonstrating that Offsets are ``More 
Hazardous'': Summary and Rationale (Sec. 63.48)
A. Statutory Requirements for a ``More Hazardous'' Finding
B. Overview of the Alternatives Considered for a ``More Hazardous'' 
Finding
C. The Establishment of Relative Hazard Between Categories of 
Pollutants
D. The Determination of Relative Hazard Within Categories of 
Pollutants
E. The Determination of a ``More Hazardous'' Decrease in Emissions
F. Miscellaneous Hazard Ranking Issues
V. Discussion of the Relationship of the Proposed Requirements to 
Other Requirements of the Act
A. Relationship of section 112(g) Implementation to Title V Program
B. Relationship to section 112(l) Delegation Process
C. Section 112(i)(5) Early Reductions Program
D. Section 112(j) ``Hammer'' Provision
E. Subpart A ``General Provisions''
F. Section 112(g) Implementation During the Transition Period
VI. Administrative Requirements
A. Executive Order 12866
B. Regulatory Flexibility Act
C. Paperwork Reduction Act
VII. Suggest Format for Comments

    The proposed regulatory text is not included in the Federal 
Register notice, but is available in Docket No. A-91-64 or by request 
from the EPA contact persons designated earlier in this note. The 
proposed regulatory language is also available on the technology 
Transfer Network (TTN), of EPA's electronic bulletin boards. The TTN 
provides information and technology exchange in various areas of air 
pollution control. The service is free, except for the cost of a 
telephone call. Dial (919) 541-5742 for up to a 14,400 bps modem. If 
more information on the TTN is needed call the HELP line at (919) 541-
5384.
    The purpose of this document is to provide the public with an 
opportunity to comment on the proposed rule implementing the 
requirements of section 112(g) of the Act. This preamble is organized 
to serve readers needing: (1) an overview of the proposed requirements 
of the section 112(g) program, and (2) a detailed discussion of the 
alternatives considered in the developing the proposed requirements.
    Section I of the preamble provides an overview of the requirements 
of the regulations being proposed today.
    Section II provides background on section 112(g) in the context of 
the 1990 amendments to the Act.
    Section III provides a detailed discussion of the requirements of 
the proposed rule and the rationale for these requirements including 
other regulatory options that were considered.
    One of the most important and challenging provisions of section 
112(g) is the requirement that the EPA provide a ranking of HAP for 
purposes of offset demonstrations. Section IV of the preamble provides 
a detailed discussion of EPA's approach to this pollutant ranking. 
Section V of the preamble discusses the relationship between the 
requirements of the proposed rule and other important Act 
implementation activities. Section VI demonstrates that the proposed 
rulemaking is consistent with a number of Federal administrative 
requirements.
    This preamble makes use of the term ``State,'' usually meaning the 
State air pollution control agency which would be the permitting 
authority implementing title V or part 70 and the section 112(g) 
program. The reader should assume that use of ``State'' also applies, 
as defined in section 302(d) of the Act, to the District of Columbia 
and territories of the United States, and may also include reference to 
a local air pollution control agency. In some cases, the term 
``permitting authority'' is used and can refer to both State agencies 
and to local agencies (when the local agency directly makes the 
determinations or assists the State in making the determinations). The 
term ``permitting authority'' may also apply to the EPA, where the EPA 
is responsible for the program.

I. Summary of Proposed Rule

    The proposed rule would implement the requirements of section 
112(g) of the Act by adding new regulatory sections to 40 CFR part 63, 
subpart B. The new sections would appear as Secs. 63.40 through 63.49 
of subpart B. The requirements of section 112(g) are displayed in 
Figure 1. The program applies to major sources of hazardous air 
pollutants for which changes are proposed that would lead to increases 
in emissions. The program imposes control technology requirements on 
``constructed, reconstructed, or modified'' major sources of hazardous 
air pollutants.
BILLING CODE 6560-50-P

TP01AP94.000


BILLING CODE 6560-50-C

A. Requirements for Constructed and Reconstructed Major Sources

    Figure 2 displays the requirements of the proposed rule for 
constructed or reconstructed major sources.

BILLING CODE 6560-50-P

TP01AP94.001


BILLING CODE 6560-50-C
    The definition of ``major source'' can be found in a subpart A to 
40 CFR part 63. (This subpart is expected to be promulgated at roughly 
the same time as today's proposal implementing section 112(g). In this 
proposed rule, two alternative definitions of ``construct'' and 
``reconstruct'' are given. The EPA is taking public comment on the 
alternatives and intends to select one in the final rule. (Under both 
alternatives, construction of major-emitting equipment on a new site is 
considered ``construction;'' the alternatives differ in the treatment 
of new equipment adding ``major'' amounts of emissions to an existing 
site.) If equipment additions or overhauls meet the definition of 
``construct a major source'' or ``reconstruct a major source,'' then, 
pursuant to Sec. 63.42 of the proposed rule, the owner or operator must 
demonstrate that emissions will be controlled to a level consistent 
with the ``new source MACT'' definition in section 112(d)(3) of the 
Act. When no applicable MACT standard (i.e., promulgated under section 
112(d) of the Act) has been promulgated for the category, a case-by-
case determination must be made. The procedures for MACT determinations 
are set forth in Sec. 63.45 of the proposed rule and are further 
described in a draft guidelines document that is being released in 
tandem with the proposed rule. Guidelines for MACT Determinations under 
Section 112(g), EPA-450/3-92-007b). The EPA is also requesting comment 
on these guidelines, which can be obtained from the EPA library, 
telephone (919) 541-2777.

B. Requirements for Modified Major Sources

    The statutory requirements in section 112(g) of the Act for 
``modifications'' to a major source are more complex. Figure 3 displays 
the requirements in the proposed rule for modifications. If a plant 
meets the definition of ``major source'' in subpart A, then any 
``physical change or change in the method of operation'' increasing 
``actual emissions'' above a ``de minimis level'' at the plant is a 
``modification.'' This proposed rule contains definitions and 
procedures for addressing each of these terms. Section 63.44 provides a 
table of de minimis values for each of the HAP listed in section 112(b) 
of the Act.
BILLING CODE 6560-50-P

TP01AP94.002


BILLING CODE 6560-50-C
    Past Federal air quality regulations have excluded a number of 
activities from the definition of ``physical change or change in the 
method of operation.'' In paragraph 63.43(c) of the proposed rule, the 
EPA proposes to provide a similar list of exclusions for purposes of 
section 112(g) of the Act. There are a number of possible approaches to 
the determination of an ``actual emission increase.'' In paragraph 
64.33(d), the EPA proposes a calculation procedure for ``actual 
emissions.''
    The Act requires the owner or operator of a major source 
``modification'' to demonstrate that an ``existing source MACT'' level 
will be met. The language in section 112(g) is ambiguous regarding the 
extent of coverage at a plant site when a modification has occurred. 
Paragraph 63.43(b) in the proposed rule is intended to clarify the 
ambiguity by describing the equipment that would require MACT when a 
``modification'' occurs.
    Section 63.45 of the proposed rule outlines the principles and 
procedures for the ``existing source MACT'' demonstration. If no 
applicable standard has been promulgated by the Administrator, then a 
case-by-case determination of MACT must be made. More detail on the 
procedures described in Sec. 63.45 is given in a draft MACT Guidelines 
Document, Guidelines for MACT Determinations under section 112(g) (EPA-
450/3-92-007b). This document includes a process for demonstrating that 
the control technology recommended by the owner or operator is 
consistent with minimum requirements described in section 112(d) of the 
Act.
    One important provision of section 112(g) of the Act is that an 
owner or operator wishing to avoid the MACT demonstration requirement 
may provide emission ``offsets.'' The Act provides little specific 
guidance on these offsets, and therefore the proposed regulation must 
address a number of complex issues related to the offsets. In 
Secs. 63.46 and 63.47 of the proposed rule, the EPA provides two 
optional approaches for defining the types of emission decreases that 
would be credited as offsets. The owner or operator wishing to provide 
an offset demonstration could use either approach. The approach in 
Sec. 63.46 is the more complex approach and resembles the ``netting'' 
process used for the criteria pollutant ``prevention of significant 
deterioration'' program (40 CFR 52.21) and nonattainment new source 
review provisions (40 CFR 51.165 and 166). The approach in Sec. 63.47 
is a more simplified approach which imposes greater restrictions on the 
types of decreases that would be creditable.
    Section 112(g)(1) allows for offsetting between pollutants. This is 
a major departure from other ``offset'' or ``netting'' programs which 
allow for decreases to be credited only towards increases of the same 
pollutant or pollutants within broad classes. Section 112(g)(1)(B) 
requires that the EPA provide guidance for determining that offsetting 
decreases are ``more hazardous'' than the increase being offset. In 
Sec. 63.48 of the proposed rule, the EPA requests comment on a possible 
method for making a ``more hazardous'' demonstration.

C. Software Illustrating the Proposed Requirements

    The EPA recognizes the complexity of this proposed rule and the 
need for assistance in clarifying the provisions for potentially 
affected plant operators and permitting authorities. In order to help 
communicate the requirements, the EPA is developing software that 
should provide this assistance. This software is designed to provide 
users with the opportunity to explore for example cases how to 
determine whether ``construction,'' ``reconstruction,'' or 
``modifications'' requirements apply and whether an example pollutant 
is considered ``more hazardous'' than another. The EPA hopes that this 
software can facilitate an improved review of the proposed rule.

II. Background

A. Act Amendments. Section 112

    The Act amendments of 1990 [Public Law 101-549] contain major 
changes to section 112 of the Act pertaining to the control of HAP 
emissions. Section 112(b) includes a HAP list that is composed of 189 
chemicals, including 172 specific chemicals and 17 compound classes. 
Section 112(c) requires publication of a list of source categories of 
major sources emitting these HAP, and of area sources that warrant 
regulation. Section 112(d) requires promulgation of emission standards 
for each listed source category according to a schedule set forth in 
section 112(e).

B. Act Amendments. Provisions for Constructed, Reconstructed, and 
Modified Major Sources of HAP

    The amendments to section 112 include a new section 112(g). This 
section is entitled ``Modifications,'' but it contains control 
technology requirements for constructed and reconstructed major sources 
as well as major source modifications.
    1. Statutory Requirements for Constructed and Reconstructed Major 
Sources. Section 112(g)(2)(B) contains requirements for constructed and 
reconstructed major sources, as follows:

    After the effective date of a permit program under title V in 
any State, no person may construct or reconstruct any major source 
of hazardous air pollutants, unless the Administrator (or the State) 
determines that the maximum achievable control technology emission 
limitation under this section for new sources will be met. Such 
determination shall be made on a case-by-case basis where no 
applicable emission limitations have been established by the 
Administrator.

    This section mandates a more stringent minimum level of control for 
``constructed'' and ``reconstructed'' major sources than for 
``modified'' sources. In addition, this section mandates the setting of 
a case-by-case emission limitation based on a technology determination 
for major sources that are constructed or reconstructed after the 
effective date of a title V permit program, but before the 
establishment of Federal emission limitations.
    2. Statutory Requirements for Modifications. The requirements for 
major source modifications differ from those for constructed and 
reconstructed major sources. Section 112(g)(2)(A) states that:

    After the effective date of a permit program under title V in 
any State, no person may modify a major source of hazardous air 
pollutants in such State, unless the Administrator (or the State) 
determines that the maximum achievable control technology emission 
limitation under this section for existing sources will be met. Such 
determination shall be made on a case-by-case basis where no 
applicable emission limitations have been established by the 
Administrator. (emphasis added)

    The underlined phrases signal important differences between section 
112(g) and the way modifications have been treated previously under 
sections 111 and 112 of the Act. Previously, a ``modification'' has 
been treated as a change to an existing air pollution source which 
caused it to be subject to an emission standard or level of control 
that would be required for new equipment. Under section 112(g), 
``modified'' equipment need only meet an existing source level of 
control which was envisioned to be potentially less demanding. Congress 
was apparently concerned that treating modifications as ``new sources'' 
would be overly stringent. Senator Lautenberg summarized the 
Congressional discussion as follows:

    * * * one of the differences between the House and Senate air 
toxics provisions was their treatment of modifications to existing 
sources. The House bill included modified existing sources in its 
definition of new source, while the Senate bill had limited the new 
source definition to new and reconstructed sources.
    As the Author of the Senate provision, I was concerned that the 
House definition would have unduly hampered routine operations of 
many manufacturing facilities that may make frequent operational or 
physical changes which may result in increased and different mixes 
of air pollutants. For example, many pharmaceutical and electronic 
manufacturing facilities in my State make frequent changes in their 
operations which result in variation in their air emissions. Simply 
substituting one hazardous air pollutant for another more hazardous 
air pollutant in a process could have caused the source to be 
considered a new source. Or, simply initiating the manufacturing of 
new or different products causing certain alterations and increases 
in the emissions, could have triggered the new source definition.
    The implications of the new source definition in the House bill 
would have been substantial for an existing major source. This could 
have caused time-consuming delays as well as imposing the 
unreasonable burden of retrofitting the modification to bring it 
into compliance with new source MACT.\1\
---------------------------------------------------------------------------

    \1\It has been held that although ``such statements by 
individual legislators should not be given controlling effect, * * * 
when they are consistent with the statutory language and other 
legislative history, they provide evidence of Congress' intent.'' 
Brock v. Pierce County, 476 U.S. 253, 263 (1986). While Senator 
Lautenberg's statement alone would not be sufficient to impose 
requirements not grounded in the statute itself or limit the 
Agency's discretion, it is a helpful explanation of Congressional 
intent.

---------------------------------------------------------------------------
    136 Cong. Rec. S 17124-5 (October 26, 1990).

    A second important difference between section 112(g) and the pre-
amendment treatment of modifications is that case-by-case control 
technology judgments must be made where no applicable emission 
limitations have been established by the Administrator. Modifications 
that occur after the ``effective date'' of the operating permits 
program (see discussion below in section II.A of this preamble), but 
before the promulgation by the EPA of a section 112(d) standard for a 
given source category, must comply with emission limitations that 
reflect a case-by-case judgment on the part of the reviewing authority.
    The Act establishes a number of complex considerations in 
determining what constitutes a ``modification.'' The definition of a 
major source ``modification'' is given in section 112(a)(5) and 
additional considerations are given in section 112(g)(1)(A). Section 
112(a)(5) states that:

    The term `modification' means any physical change in, or change 
in the method of operation of, a major source which increases the 
actual emissions of any hazardous air pollutant emitted by such 
source by more than a de minimis amount or which results in the 
emissions of any hazardous air pollutant not previously emitted by 
more than a de minimis amount. (emphasis added)

    This definition of modification is very similar to the definition 
contained in section 111 of the Act. One important difference in this 
definition is that it allows for an exception for modifications that 
result in an increase which is less than a de minimis amount, while 
section 111 regulates ``any emission increase'' caused by the physical 
change or change in the method of operation.
    Section 112(g)(1)(A) provides for consideration of emission offsets 
in the definition of ``modification.'' Section 112(g)(1)(A) states 
that:

    A physical change in, or change in the method of operation of, a 
major source which results in a greater than de minimis increase in 
actual emissions of a hazardous air pollutant shall not be 
considered a modification, if such increase in the quantity of 
actual emissions of any hazardous air pollutant from such source 
will be offset by an equal or greater decrease in the quantity of 
emissions of another hazardous air pollutant (or pollutants) from 
such source which is deemed more hazardous.

    Congress recognized the difficulty in interpreting the phrase 
``which is deemed more hazardous.'' In section 112(g)(1)(B), the EPA is 
required to provide a hazard ranking of the chemicals, as follows:

    The Administrator shall, after notice and opportunity for 
comment and not later than 18 months after the date of enactment of 
the Clean Air Act Amendments of 1990, publish guidance with respect 
to implementation of this subsection * * * [i.e, section 112(g)] * * 
* Such guidance shall include an identification, to the extent 
practicable, of the relative hazard to human health resulting from 
emissions to the ambient air of each of the pollutants listed under 
* * * [section 112(b)] * * * sufficient to facilitate the offset 
showing authorized by * * * [section 112(g)(1)(A)] * * * Such 
guidance shall not authorize offsets between pollutants where the 
increased pollutant (or more than one pollutant in a stream of 
pollutants) causes adverse effects to human health for which no 
safety threshold for exposure can be determined unless there are 
corresponding decreases in such types of pollutants.

    Finally, Congress directed the EPA to prevent unnecessary delays in 
the review of modifications, particularly where no case-by-case control 
decision is needed. Section 112(g)(3) says that:

    The Administrator (or the State) shall establish reasonable 
procedures for assuring that the requirements applying to 
modifications under this section are reflected in the permit.

    Again citing Senator Lautenberg,

    As long as the permit provides that the existing source MACT 
standard will be complied with in the event of a modification, the 
objectives of the modification provision in the law will have been 
satisfied. If there is no existing source MACT standard, then the 
Congress expects an expeditious determination of what emission 
limitations the modification must meet.

    3. Requirement for Guidance. Need for Rulemaking. Section 
112(g)(1)(B) of the Act directs the EPA to ``publish guidance with 
respect to the implementation of this subsection.'' The EPA believes 
that ``subsection'' refers to all of the requirements of ``subsection 
112(g)'' of the Act, and that guidance is required to provide 
consistency in implementing all of the section 112(g) requirements.
    The EPA requests comment on an alternative reading that would 
require guidance only for the ranking of pollutants. The EPA believes 
that although section 112(g)(1)(B) states that the pollutant ranking is 
to be included in the guidance, the EPA does not believe that this 
language means that the guidance should be limited to the pollutant 
ranking.
    In any case, there is no requirement in section 112(g) that the EPA 
publish a rule to implement the requirements. At a minimum, the EPA 
must issue ``guidance'' after ``notice and opportunity for comment.'' 
The EPA believes that there are sound policy reasons for promulgating a 
rule rather than issuing informal guidance. First, the requirement for 
``opportunity for comment'' on the guidance suggests that any guidance 
that is issued would require a review process similar to a rulemaking. 
Second, the EPA believes that a rulemaking would serve to provide a 
consistent basis for interpreting a number of ambiguous phrases in the 
statute. In the absence of such a rule, a consistent interpretation of 
the Federal requirements would not exist and the potential for 
litigation and delays could increase.

C. Process To Develop the Proposed Rule

    The EPA has undertaken a substantial effort to obtain feedback from 
interested parties in the development of the proposed rule.
    During July 1991, a 2-day meeting was held with an ad hoc group 
consisting of representatives of environmental organizations, 
industrial trade groups, and State and local air quality agencies. This 
meeting served to introduce the principal issues involved with section 
112(g) implementation. Written comments were received from a number of 
participants. These comments are included in the Docket for the 
proposed rule.
    During October 1991, the EPA consulted an independent panel of 
scientific experts for input into the hazard ranking process. This 
panel of the EPA's Science Advisory Board (SAB) was apprised of the 
EPA's theoretical outline for hazard ranking in a public meeting held 
on October 28 and 29, 1991. The consultation meeting provided members 
of the SAB an opportunity to provide verbal feedback on several 
approaches.
    On November 19, 1991, ideas for developing section 112(g) guidance 
were discussed at the meeting of the National Air Pollution Control 
Techniques Advisory Committee (NAPCTAC). At the NAPCTAC, staff of the 
EPA presented a number of preliminary positions on a number of section 
112(g) issues. A copy of the EPA presentation and a summary of the 
NAPCTAC meeting are also included in the Docket.
    Finally, the EPA has consulted with a subcommittee of the Clean Air 
Act Advisory Committee for input on approaches to implementing section 
112(g). Meetings were held on May 26 and June 29, 1992 to present a 
summary of section 112(g) issues and to provide the subcommittee with 
the EPA's staff thinking with respect to those issues. A third meeting 
was held on September 24, 1992 to obtain feedback on a draft of the 
proposed rule. A copy of the draft rule submitted to the subcommittee, 
and several comment letters on that draft, are included in the Docket. 
Additional meetings were held with the subcommittee on January 15, 
1993. In February 1993, a revised draft of the rule and a first draft 
of this preamble were circulated to subcommittee members. On March 19, 
a meeting was held to discuss procedures for case-by-case MACT 
determinations. A copy of the February draft of the rule and preamble, 
and associated feedback from subcommittee members, is included in the 
Docket.

III. Summary and Rationale for Section 63.40 Through 63.47, and Section 
63.49, of the Proposed Rule

    This section of the preamble is a detailed discussion of the 
provisions of the proposed rule. This discussion outlines the rationale 
for the decisions that were made, and describes other options that were 
considered. The overall structure of the proposed requirements for 
constructed and reconstructed sources is displayed in Figure 2. The 
overall structure of the proposed requirements for modifications is 
displayed in Figure 3.

A. Section 63.40--Applicability

    Section 63.40 describes the timing of the requirements of the 
proposed rule and the situations it is generally intended to address.
    1. 63.40(a)--Subpart B applicability. Paragraph 63.40(a) of the 
proposed rule indicates that the intent of the rule is to implement 
section 112(g) of the Act.
    2. 63.40(b)--Overall requirements. Paragraph 63.40(b) of the 
proposed rule indicates the overall applicability of section 112(g) to 
the owner or operator of a major source of HAP who constructs, 
reconstructs or modifies a major source after the ``effective date of a 
title V program'' in each State.
    (a) Effective date. The meaning of ``effective date of a title V 
permit program'' is indicated in the final regulations for 
implementation of title V of the Act, which are contained in 40 CFR 
part 70, and which were published on July 21, 1992 (57 FR 32250). Under 
these regulations, States are required to submit a permit program for 
review by the EPA on or before November 1993. The EPA is required to 
approve or disapprove the permit program within 1 year after receiving 
the submittal. The EPA's program approval date is termed the 
``effective date.''
    Congressional intent for using this effective date as the trigger 
date for section 112(g) requirements is clear. According to Senator 
Lautenberg (Congressional record, S. 17125, October 26, 1990):


    Requirements for modifications do not apply to a source until 
there is an approved permit program in that State. This should 
ensure that there is a permit program in place that is designed to 
expeditiously deal with modifications.


    The effective date of title V permit programs is defined in section 
502(h) of the Act, which says ``The effective date of a permit program, 
or partial or interim program, approved under * * * [title V] * * * 
shall be the effective date of approval by the Administrator. The 
effective date of a permit program, promulgated by the Administrator 
shall be the date of promulgation.'' This definition is incorporated 
into the operating permit regulations as 40 CFR 70.4(g).
    This language refers to two types of title V programs: one type 
where the EPA ``approves'' the title V program under 40 CFR part 70 and 
another type where the EPA ``promulgates'' a program under 40 CFR part 
71. Programs ``approved'' by the EPA under Part 70 will be developed by 
the State or local area and submitted to the EPA for approval. The 
language in section 502(h) of the Act makes these programs immediately 
effective upon EPA approval. Programs ``promulgated'' by the EPA under 
part 71 are anticipated to be rare, and they occur only where a State 
failed to submit a program or submitted a program that EPA could not 
approve. The EPA is required by section 502(d)(3) of the Act to 
promulgate and administer a title V program if, by November 1995, the 
EPA has not approved the State program. The language in section 112(g), 
because it refers to the effective date of a title V program in any 
State (and not by any State), means that the program will apply to both 
the EPA ``approved'' and ``promulgated'' programs.
    The title V regulations provide for approval of ``interim'' and 
``partial'' programs in certain limited circumstances. The EPA believes 
that, because partial programs must ensure compliance with ``all 
requirements established under section 112 applicable to `major 
sources' and `new sources','' and interim programs must ``substantially 
meet the requirements of [title V],'' an interim or partial program 
would trigger the requirements of section 112(g).
    A significant issue in the beginning of a section 112(g) program is 
to define the activities that would ``grandfather'' a project that is 
already underway. As described in the proposed rule, ``construction, 
reconstruction, or modification'' are triggered from the ``onsite 
fabrication, erection, or installation'' of a project. If such 
activities occur after the effective date, then the proposed rule would 
be applicable. The EPA requests comment on other alternatives such as: 
(1) Grandfathering projects for which a complete application has been 
submitted to the permitting authority, (2) grandfathering projects 
which have submitted an application, or (3) grandfathering projects 
which have not yet received a permit.
    (b) Major Source. Section 112(g) applies only to major sources as 
defined in section 112(a)(1) of the Act. This definition, already 
included in 40 CFR part 63, subpart A, (the general provisions of part 
63), is as follows:


    The term `major source' means any stationary source or group of 
stationary sources located within a contiguous area and under common 
control that emits or has the potential to emit considering 
controls, in the aggregate, 10 tons per year or more of any 
hazardous air pollutant or 25 tons per year or more of any 
combination of hazardous air pollutants * * *


    The definition also allows the EPA to establish a lesser quantity 
than 10 or 25 tons to define ``major source'' with respect to 
particular HAP where warranted on the basis of potency, persistence, 
and other factors. To date, no such lesser quantities have been 
established.
    As a result of this definition, the section 112(g) requirements do 
not apply if the total emissions from an entire ``contiguous area under 
common control'' (in general, the entire plant site) are less than the 
listed amounts. Once plant-wide emissions exceed this total, then 
certain activities at the plant site are subject to the section 112(g) 
requirements that are outlined in the proposed rule.
    It is necessary to note that neither the proposed Subpart A 
requirements nor this proposed rule contain consideration of Standard 
Industrial Classification Codes (SIC codes) in the definition of 
``major source.'' The EPA considered using the 2-digit SIC code for 
this proposed rule, in a manner similar to that for the proposed 40 CFR 
part 70 operating permits rule. The EPA believes, however, that this 
would be inconsistent with the definition of ``major source'' in 
section 112(a) of the Act, which does not restrict a ``contiguous 
boundary'' to equipment within a 2-digit SIC code. For purposes of 
implementing section 112(g), such a restriction could, in some cases, 
restrict the portion of the plant from which emission offsets could be 
obtained. In other cases, a portion of the plant within a given 2-digit 
SIC code may not be subject to regulation, because that portion would 
not emit enough to be considered a ``major source'' in and of itself. 
The EPA recognizes that the treatment of 2-digit SIC codes under 
section 112 of the Act is not consistent with the 40 CFR part 70 
operating permits regulation. The EPA requests comment on whether the 
operating permits rule should be amended to eliminate this 
inconsistency.
    An important element of the major source definition is the term 
``potential to emit.'' ``Potential to emit'' is based on the source's 
capability to emit hazardous air pollutants with consideration to 
Federally enforceable limitations. Such limitations include 
restrictions on capacity, restrictions on the types of materials used, 
emission limitations, and other types of restrictions. A definition of 
``potential to emit'' is contained in the proposed 40 CFR part 63 
subpart A General Provisions.
    3. 63.40(c)--Exclusion for Steam Generating Units. Paragraph 
63.40(c) of the proposed rule clarifies that electric utility steam 
generating units are not yet subject to the requirements of section 
112(g).
    Section 112(n)(1) requires the EPA to perform a study of the 
hazards to public health associated with HAP emissions from electric 
utility steam generating units. This paragraph states that:

    The Administrator shall regulate electric utility steam 
generating units under this section, if the Administrator finds such 
regulation is appropriate and necessary after considering the 
results of the study required by this paragraph. (emphasis added)

    The EPA reads the phrase ``under this section'' as a broad 
exemption from regulation under section 112, including section 112(g), 
pending the results of the utility health hazards study. The EPA 
requests public comment on this reading.
    4. 63.40(d)--Relationship to State and Local Requirements. Most 
State and local regulatory agencies maintain regulatory programs that 
involve toxic air pollutant reviews for constructed and modified 
sources. Paragraph 63.40(d) clarifies that the requirements of section 
112(g) do not supersede any requirements of these programs that are 
more stringent than the proposed rule.
    5. 63.40(e)--Relationship to Other Standards. The proposed rule 
contains an exemption for sources emitting less than that which is 
regulated by promulgated standards in other subparts of 40 CFR part 63. 
Without this exemption, permitting authorities would be required to 
conduct a case-by-case MACT determination in cases where: (1) The 
emission rate exceeds a de minimis level as defined in the proposed 
rule for purposes of section 112(g) of the Act, but (2) the emitting 
equipment is below an applicability cutoff in a promulgated MACT 
standard. Such standards may describe an equipment size or capacity, or 
a stack concentration below which the requirements may be inapplicable. 
The EPA believes that emissions below such a cutoff are consistent with 
``MACT'' because a MACT evaluation was made in establishing the cutoff.
    An example should serve to clarify this exemption. First, for 
equipment leaks for synthetic organic chemical manufacturing, the EPA 
has proposed regulation of streams in ``VHAP service,'' i.e., where HAP 
contribute more than 5 percent of the stream. (See description of 
proposed 40 CFR part 63, subpart H requirements, 57 FR 62617-62719, 
December 31, 1992.) There may be instances where less than 5 percent of 
such a stream represents an increase that is greater than a de minimis 
emission rate listed in this proposed rule. If the final chemical plant 
standard, when promulgated, contains the 5 percent cutoff, the EPA 
believes that a case-by-case review for pollutants emitted at less than 
5 percent was not intended by Congress in writing section 112(g).
    The last sentence of paragraph 63.40(e) is intended to make clear 
that this paragraph is only intended to address situations involving a 
regulatory cutoff for specifically evaluated emission points. For 
example, the proposed synthetic organic chemical plant standard does 
not address boilers or other combustion equipment. If a physical change 
resulted in a greater than de minimis increase from such equipment, 
then paragraph 63.40(e) should not be interpreted to mean that they are 
exempted from the proposed rule.

B. Section 63.41--Definitions

    1. Terms Defined in the General Provisions. A number of terms used 
in the defined rule have already been proposed for all of 40 CFR part 
63 by the General Provisions contained in subpart A. Readers interested 
in the definitions and rationale for those terms should refer to 
recently promulgated subpart A. Relevant terms defined in the General 
Provisions include:

--Act
--Approved permit program
--Capital expenditure
--Effective date
--Federally enforceable
--Hazardous air pollutant
--Major source
--Permit program
--Potential to emit
--Relevant standard
--Title V Permit

    2. Terms Related to Construction and Reconstruction. As noted above 
in section II of this preamble, the Act imposes more stringent 
requirements for major source ``construction'' and major source 
``reconstruction'' than for major source ``modification.'' There is a 
degree of ambiguity in the statute regarding what must occur in order 
to ``construct'' or ``reconstruct'' a major source. The following terms 
reflect two alternative readings of the statute and are included in 
section 63.41:

--Construct A Major Source
--Reconstruct A Major Source
--Green-field Site
--Emission Unit

A detailed discussion of these definitions and the two approaches is 
included in section II.C of this preamble.
    3. Terms Related to MACT. Definitions for the following terms 
related to levels of control technology are included in section 63.41 
of the proposed rule:

--Available information
--MACT
--Control Technology
--MACT Floor
--MACT Emission Limitation for Existing Sources
--MACT Emission Limitation for New Sources

    The basis for the MACT definitions is statutory language contained 
in section 112(d) of the Act. The term MACT appears only in section 
112(g) of the Act, and does not appear elsewhere in section 112. There 
is, however, considerable legislative history indicating that this term 
refers to the level of control required by section 112(d) emission 
standards. This term was used in this context in the House Bill, H. R. 
3030. For purposes of the definitions in the proposed rule, the EPA 
assumes that is a reference to the ``maximum degree of reduction in 
emissions'' language contained in section 112(d)(3). The minimum 
control technology requirements of section 112(d), often referred to as 
the ``MACT floor'' are cited a number of times in the proposed rule. To 
avoid repeating these requirements each time, the regulation includes a 
definition of ``MACT floor.'' The term ``available information'' is 
used to define the extent of review for permitting authorities and 
applicants for case-by-case MACT determinations.
    4. Terms Affecting Extent of Coverage by MACT. The following terms 
are used to describe equipment subject to a MACT determination:

--Emission point
--MACT-affected emission unit
--List of source categories

    An ``emission point,'' as defined in the regulation, is defined 
narrowly to refer to any individual point of release to the atmosphere. 
As described below, an individual MACT determination will often be made 
at once for a number of emission points. The term ``MACT-affected 
emission unit'' is used to refer to the collection of all emission 
points considered when such a MACT determination is made.
    One purpose of the term ``MACT-affected emission unit'' is to 
clarify that a major source ``construction'' ``reconstruction'' or 
``modification'' project which involves more than one emission point or 
emission unit may require more than one MACT determination. As outlined 
in paragraph (3) of the definition, the EPA believes that MACT 
determinations consistent with section 112(d) of the Act may not 
include combinations of emission points involving more than one 
category on a published list of source categories (57 FR 31576). For 
example, most types of combustion sources appear as individually listed 
categories. As a result, a ``construction'' ``reconstruction'' or 
``modification'' involving boilers and other process equipment must 
make a separate MACT determination for the boilers.
    Another purpose of the term ``MACT-affected emission unit'' is to 
provide owners and operators of modified major source with additional 
flexibility. There are situations, such as that displayed in Figure 4, 
for which there may be overall technologies that would reduce emissions 
more effectively than applying MACT to each emission point being 
changed. Accordingly, paragraph (2)(ii) of the definition gives the 
owner and operator the discretion to include emission points in the 
``MACT-affected emission unit'' in addition to those that are 
``affected by the modification.'' A detailed discussion of ``affected 
by the modification appears below in section III.D of this preamble.
BILLING CODE 6560-50-P

TP01AP94.003


BILLING CODE 6560-50-C
    The EPA considered requiring that the ``MACT-affected emission 
unit'' include additional emission points in cases where it is contrary 
to common practice to consider them in isolation from the emission 
point for which the ``modification'' has occurred. For example, the 
addition of a single pump or valve to a production unit may yield a 
determination to require a leak detection and repair program as the 
control measure. The EPA believes that it would not be good engineering 
practice to suggest a leak detection program for that unique pump or 
valve, or a different leak detection program from that for other 
equipment within the same production unit. By including all equipment 
in the production unit as part of the ``MACT-affected emission unit,'' 
a more reasonable assessment of control alternatives may result. The 
EPA believes, however, that inclusion of Figure 4. Example Illustrating 
Paragraph (2)(ii) in the Definition of ``MACT-affected emission unit'' 
this provision is probably not needed, because permitting authorities 
will generally treat such equipment as a single grouping. In addition, 
this provision may lead to confusion as to which situations would 
require broader MACT coverage. The EPA requests comment on whether this 
provision should be included in the final rule.
    5. De minimis. The definition of de minimis is discussed below in 
section III.E. of this preamble.
    6. Electric Utility Steam Generating Unit. The definition of 
electric utility steam generating unit in the proposed rule is taken 
directly from section 112(a) of the Act.
    7. Source Reduction Project. As discussed below (see discussion 
related to Sec. 63.47 of the proposed rule), the proposed rule provides 
for source reduction projects to be considered in identifying emission 
offsets. The definition of ``source reduction project'' is intended to 
be consistent with the Pollution Prevention Act, Public Law 101-503.
    C. Section 63.42--Requirements for Constructed and Reconstructed 
Major Sources. Section 63.42 (in combination with a number of 
definitions contained in Sec. 63.41) contains the requirements for 
constructed and reconstructed major sources described in section 
112(g)(2)(B) of the Act. Equipment affected by this section must comply 
with a ``new source MACT'' level of control. The EPA requests comment 
on its interpretation of the statutory language pertaining to 
constructed and reconstructed major sources.
    1. ``Green-field'' Facilities. The most straightforward case for 
section 112(g) is for a new plant site emitting (or having the 
potential to emit) more than major amounts of HAP (that is, 10 tons/yr 
individually, 25 tons/yr collectively, or amounts that exceed any 
lesser quantity cutoffs that may be established under subpart C of part 
63). The EPA believes that the statute clearly requires such a new 
plant site to be treated as a ``constructed major source'' subject to a 
``new source MACT'' level of control.
    2. Addition of Equipment at an Existing Plant Site. Another 
important situation to address is the addition of equipment emitting 
major amounts, i.e., greater than 10 tons per year of one HAP, or 
greater than 25 tons per year from all HAP, to an existing major source 
plant site. An example of such an addition is shown in Figure 5.
BILLING CODE 6560-50-P

TP01AP94.004


BILLING CODE 6560-50-C
    The EPA believes that there are two possible readings of the Act 
with respect to this situation, and that both readings are consistent 
with the definition of ``major source'' in section 112(a) of the Act. 
Under the first reading, the addition of equipment at an existing plant 
site would constitute ``construction of a major source'' because the 
addition would entail the construction of ``a stationary source or 
group of stationary sources'' emitting major amounts (that is, 10 tons/
yr individually, 25 tons/yr collectively, or amounts that exceed any 
lesser quantity cutoffs that may be established under subpart C of part 
63). This equipment addition would be subject to a ``new source MACT'' 
level of control which is likely to be more stringent than the 
``existing source MACT'' level of control for ``modifications.'' Also, 
there would be no opportunity to provide for emission offsets in lieu 
of a control technology demonstration.
    Under the second reading, the entire plant site would be treated as 
a ``stationary source or group of stationary sources'' and any such 
addition would be treated as a major source ``modification.'' If 
treated as a possible ``modification,'' then the Act provides the 
opportunity to seek such offsets and, if such offsets are provided, 
then the new equipment could be operated without controls, or with 
controls that are less stringent than MACT, until an applicable 
standard is promulgated under section 112(d). If offsets were not 
provided, the equipment would be controlled with existing source MACT.
    The EPA believes that there are advantages and disadvantages of 
both readings.
    The ``construction'' reading ensures that major-emitting equipment 
additions (that is, those emitting more than 10 tons/year of a HAP, or 
25 tons per year from all HAP, or amounts exceeding a lesser quantity 
cutoff), which generally would represent sizeable investments, would be 
built with state-of-the art control technology. It is generally 
recognized that it is more straightforward to build such a level of 
control technology into the original design, and that it is difficult 
or impossible to retrofit such controls at a later date. A fundamental 
goal of programs such as the new source performance standards (NSPS) 
program under section 111 of the Act and the effluent guidelines 
program under the Clean Water Act is to achieve long-term reductions in 
emissions by requiring ``best'' controls as old production operations 
are replaced with new operations. However, the ``construction'' 
definition precludes offsetting--which in some cases might result in a 
greater reduction in overall hazard by focusing controls on pollutants 
of greater regulatory concern and by reducing the total quantities of 
pollutants.
    For equipment additions occurring after EPA has proposed a section 
112(d) standard, the ``construction'' reading would provide greater 
consistency if the equipment addition meets the definition of ``new 
source'' in that section 112(d) standard. If the equipment is a ``new 
source'' in the proposed standard, any such ``new source'' constructed 
after the proposal date of the standard is required to install ``new 
source MACT'' upon promulgation of the standard. Inconsistencies would 
result if the same equipment, if major-emitting, was treated as a 
``modification'' in today's proposed rule.
    The ``modification'' reading provides sources the flexibility to 
completely offset increased emissions--thereby achieving a greater 
emission reduction than the ``construction'' definition, presumably at 
less cost. However, where sources opt to install existing source MACT 
rather than offset, emission reductions could be less than if the 
source installed new source MACT. In addition, it is difficult to judge 
whether emission reductions being used as offsets would have occurred 
whether or not the plant is being modified. The EPA is not able to 
determine which approach would result in lower net emissions over time.
    Also, the ``modification'' reading may lead to inequities in the 
implementation of the program. As shown in Figure 6, a ``green-field'' 
plant site emitting 10 tons per year would be subject to a new source 
MACT, while addition of identical equipment at an existing plant would 
be subject to existing source MACT (or emissions would have to be 
offset).
BILLING CODE 6560-50-P

TP01AP94.005


BILLING CODE 6560-50-C
    The ``modification'' reading would appear to be the most consistent 
with the approach taken by the EPA in the prevention of significant 
deterioration (PSD) and non-attainment new source review (NSR) programs 
for criteria air pollutants. The PSD program includes a ``netting'' 
provision which takes into account plant-wide emission increases and 
decreases in evaluating whether a ``modification'' has occurred. The 
NSR program also takes plant-wide emission reductions into account in 
determining applicability of the program to equipment additions. The 
ability to consider plant-wide emission reductions was upheld in a 1985 
Supreme Court decision [Chevron USA, Inc. v. Natural Resources Defense 
Counsel, 467 U.S. 837 (1984)]. One difference between the PSD and NSR 
programs and the section 112(g) program in the proposed rule is that 
most equipment additions which avoid ``best'' or ``lowest achievable'' 
controls under these programs must still meet a ``best demonstrated'' 
level of control if NSPS standards have been established. For section 
112(g), however, an equipment addition could avoid controlling HAP 
emissions entirely until a MACT standard is established. Additionally, 
the PSD and NSR programs are focused on preserving or attaining 
national ambient air quality standards. The EPA requests comment on 
whether the lack of such ambient criteria for HAP would suggest a 
greater technology focus under section 112(g).
    The EPA requests public comment on these two interpretations. For 
the proposed rule, both interpretations are presented as separate 
``alternatives.''
    Under Alternative A, addition of a major-emitting ``emission unit'' 
is included within the definition of ``construct a major source.'' The 
definition of emission unit is the same as that used for regulations 
proposed pursuant to section 112(j) of the Act. The definition is 
intended to provide permitting authorities with considerable 
flexibility in determining the ``entity'' which would be treated as 
``construction.'' The EPA requests comment on whether more prescriptive 
language for this term would be desirable.
    Under Alternative B, the only activity that is included within the 
definition of ``construct a major source'' is the addition of major-
emitting equipment at a green-field site. By implication, any such 
activity at a site which is not a green-field site is regulated under 
the modifications provisions of Sec. 63.43 of the proposed rule. The 
term ``green-field site'' generally refers to equipment constructed in 
a previously undeveloped area. There are, however, situations where a 
virtually undeveloped area or small commercial or industrial equipment 
could exist at a site for which a major source is to be constructed. 
Accordingly, the proposed definition of ``green-field site'' includes 
as ``green-field'' a site for which the total emissions of any given 
HAP are less than de minimis. The EPA requests comment on other 
possible definitions of this term.
    The EPA also requests comments on other possible approaches to the 
definition of ``construct a major source.'' One approach suggested to 
the EPA would include any equipment addition at a major source emitting 
more than de minimis quantities within the definition of ``construct a 
major source.'' The EPA believes that this reading is inconsistent with 
the statute. In addition, in structuring standards under section 
112(d), it is unlikely that the EPA will promulgate standards that 
would treat all equipment additions as ``new.'' The EPA requests public 
comment on whether such an interpretation is appropriate.
    3. Reconstruction. Section 112(g) continues the concept of 
``reconstruction'' contained in past regulatory programs. The concept 
of reconstruction is intended to prevent the circumvention of ``new 
source'' requirements by completely overhauling existing equipment. 
Current air pollutant emission standards under previous requirements of 
the Act treat replacement of components as a reconstruction if the 
replacement represents more than 50 percent of an entirely new 
facility.
    For section 112(g), the requirements apply to the reconstruction of 
a ``major source,'' and the proposed rule defines ``reconstruct a major 
source'' as the replacement of components at a major source such that 
the replacement exceeds 50 percent of the capital cost of an entirely 
new major source. Two alternative definitions of ``reconstruct a major 
source'' are included in Sec. 63.41 of the proposed rule; these two 
definitions are intended to coincide with the two alternative 
definitions of ``construct a major source'' discussed previously.
    Under Alternative A, ``reconstruct a major source'' is based upon 
an emission unit. If an emission unit emits major amounts, then the 
replacement of components at that unit would be considered a 
``reconstruction'' if the cost of the replacement exceeds 50 percent of 
the cost of an entirely new unit. In this case, new source MACT would 
be required for the emission unit.
    Under Alternative B, ``reconstruct a major source'' is based upon 
all equipment within the entire contiguous plant site. The definition 
includes only those situations where the replacement of components 
would exceed 50 percent of the cost of the entire plant site. Under 
Alternative B, the probability that a reconstruction would occur is 
substantially decreased. If, however, a reconstruction did occur under 
this definition, it would require the entire plant to install new 
source MACT.
    The EPA requests public comment on the definition of ``reconstruct 
a major source'' in the proposed rule.
    4. Control Technology Review Requirements For Constructed and 
Reconstructed Major Sources. Section 63.42 reflects the statutory 
requirement that an owner or operator who proposes to ``construct or 
reconstruct'' a major source must obtain a determination from ``the 
permitting authority'' that a new source MACT emissions limitation will 
be met. The ``permitting authority'' is defined as the agency 
responsible for the title V permit program. Further discussion on this 
issue, and on other issues related to implementation of section 112(g), 
is contained in section V of this preamble.
    The requirements and procedures for obtaining the MACT 
determinations are contained in Sec. 63.45 of the proposed rule (see 
discussion below).

D. Section 63.43--Requirements for Modified Major Sources

    Section 63.43 of the proposed rule is intended to clarify the 
requirements in sections 112(a) and 112(g) of the Act related to major 
source modifications.
    Section 112(a) of the Act defines the term ``modification'' as:

any physical change in, or change in the method of operation of, a 
major source which increases the actual emissions of any hazardous 
air pollutant emitted by such source by more than a de minimis 
amount or which results in the emissions of any hazardous air 
pollutant not previously emitted by more than a de minimis amount.

    Section 112(g)(2)(A) of the Act states that:

    After the effective date of a permit program under title V in 
any State, no person may modify a major source of hazardous air 
pollutants in such State, unless the Administrator (or the State) 
determines that the maximum achievable control technology emission 
limitation under this section for existing sources will be met. Such 
determination shall be made on a case-by-case basis where no 
applicable emission limitations have been established by the 
Administrator.

    There are a number of questions raised by this statutory language 
for which interpretations are needed. In particular, three questions 
are addressed by Sec. 63.43 of the proposed rule:
    (a) How much of a major source must be controlled to a MACT level 
when a modification occurs?
    (b) What is a physical change or change in the method of operation?
    (c) How should actual emission increases be calculated?
    The EPA requests public comment on the various statutory 
interpretations contained in this section. One important overall 
interpretation is that the process for a modification under section 
112(g) of the Act should follow a similar two-step process as contained 
in previous modifications requirements under the NSPS and PSD programs.
    1. General Requirements for Modifications (Paragraph 63.43(a). 
Paragraph 63.43(a) outlines the overall statutory requirements for 
major source modifications. An owner or operator who wishes to modify a 
major source is required by this paragraph to obtain a determination 
from the permitting authority that ``the MACT emission limitation for 
existing sources'' will be met. The ``permitting authority'' is defined 
in the proposed rule as the agency implementing title V of the Act (see 
further discussion in section V.A of this preamble).
    Paragraph 63.43(a) requires that the MACT determinations be made 
consistent with Sec. 63.45 of the proposed rule (see discussion below). 
A determination is required for ``all emission points affected by the 
modification'' according to paragraph 63.43(b). The phrase all emission 
points is used intentionally rather than each emission point in order 
to provide the flexibility to evaluate control technologies over the 
entire modification. In some cases, a MACT determination made for a 
combination of emission points may yield a more cost-effective strategy 
than controlling each emission point individually.
    Paragraph 63.43(a) also refers to two important exceptions. First, 
certain activities, listed in paragraph 63.43(c), are excluded from 
consideration as ``physical changes'' or ``changes in the method of 
operation.'' Second, paragraph 63.43(e) gives the owner and operator 
the option to provide an offset demonstration.
    2. Paragraph 63.43(b). ``Modification'' and ``Emission Points 
Affected by the Modification.'' Section 112(g)(2)(A) of the Act 
requires that ``the MACT emission limitation for existing sources will 
be met,'' but it does not specify which emitting equipment at the major 
sources is subject to the MACT determination when a modification 
occurs. For the proposed rule the term ``emission points affected by 
the modification'' is used as an approach to clarifying this ambiguous 
phrase.
    Paragraph 63.43(b) in the proposed rule clarifies how the terms 
``modification'' and ``emission points affected by the modification'' 
are used in the proposed rule.
    Paragraph 63.43(b)(1) identifies three different situations which 
could be a ``modification:''

--An emission increase from a single emission point that is greater 
than de minimis,
--Construction of any emission point at a major source that emits 
greater than de minimis amounts, but not enough to be considered 
``construction'' in accordance with Sec. 63.42 of the proposed rule, 
and
--Emission increases from multiple emission points where the sum of the 
emission increases exceeds de minimis amounts for a given modification 
project. (The intent of this latter provision is to ensure that 
modification projects are considered as a whole in evaluating whether 
the increase is greater than de minimis. This provision is not intended 
to require owner or operators to keep a running tally of all emission 
increases and decreases.)

    Paragraph 63.43(b)(2) further clarifies how an ``emission 
increase'' is to be determined for purposes of identifying a 
``modification.'' An ``emission increase'' occurs if a ``physical 
change in or change in the method of operation of'' the major source 
leads to an actual emission increase as calculated in accordance with 
paragraph (e) (see discussion below).
    Paragraph 63.43(b)(3) identifies as ``emission points affected by 
the modification'' those emission points that increase in emissions, as 
determined by paragraph 63.43(b)(2), and that ``contribute'' to a 
greater than de minimis increase in emissions. The word ``contribute'' 
means that emission points are included in cases where that emission 
point in any of itself does not increase emissions by more than a de 
minimis amount, but the modification as a whole does. The EPA 
considered an option which would identify as affected only those 
emission points that ``significantly'' contribute to a greater than de 
minimis increase. In some cases, the total emissions resulting from a 
project may exceed de minimis amounts for a given HAP, but some 
emission points may have very small emission increases of that HAP. For 
such cases, the inclusion of the term ``significant'' could serve to 
exempt such equipment from review. The EPA is concerned, however, that 
it is difficult to define ``significant'' in a way that would be 
reasonable and consistently applied. The EPA believes that the 
inclusion of such sources in other programs, such as BACT reviews in 
the PSD program, has not led to the imposition of unreasonable 
controls. The EPA requests comment on whether the term ``significant'' 
should be included in this paragraph, and, if so, how it should be 
defined.
    The EPA considered alternative approaches to MACT coverage that may 
also be consistent with the Act. One approach would apply MACT plant-
wide when a change to the plant constitutes a modification. This 
approach would interpret the language to mean that a modification of a 
major source requires MACT for the entire major source. While this 
approach would maximize emission reductions, the EPA believes that it 
would greatly complicate the review process. Many plants have hundreds 
of emission points that release HAP to the atmosphere. The EPA does not 
believe that Congress intended for a case-by-case review of all 
emission points any time one emission point was modified, or any time 
an emission point was added to the plant. This would greatly increase 
the review time, would increase the burden on State and local agencies 
to analyze the available control technologies for existing equipment. 
Also, there would be an increase in the costs associated with an 
equipment modification. The EPA does not believe that these results 
were intended.
    Another approach considered would be to subdivide a given major 
source plant site into distinct major-emitting emission units. Such an 
approach would treat each subdivision as a separate ``major source'' in 
and of itself and would apply MACT to all emission points within the 
``major source'' being modified. Under this approach, MACT would not 
apply to the entire plant, but could incorporate additional emission 
points. The EPA believes that such an approach would be very complex to 
administer in that it would be difficult to define the appropriate 
``major source'' subdivisions. The EPA requests comments on whether 
such an approach is consistent with the intent of the statute.
    3. Paragraph 63.43(c) Activities Excluded from the Definition of 
Physical Change or Change in the Method of Operation. (Step 1 of the 2-
Step Process to Identify Modifications). In paragraph 63.43(b), the 
term ``physical change or change in the method of operation'' is used 
frequently. For both the PSD program (40 CFR 52.21), the NSPS program 
(40 CFR part 60), the program for National Emission Standards for 
Hazardous Air Pollutants (NESHAP) prior the 1990 amendments to the Act 
(40 CFR part 61), and the criteria pollutant nonattainment area new 
source review (NSR) program, there are a number of activities that are 
not considered to be a physical change or change in the method of 
operation. For the PSD program (see 40 CFR part 52.21), the following 
activities are excluded:

    (a) Routine maintenance, repair and replacement;
    (b) Use of an alternative fuel or raw material by reason of an 
order under section 2 (a) and (b) of the Energy Supply and 
Environmental Coordination Act of 1974 (or any superseding 
legislation) or by reason of a natural gas curtailment plant 
pursuant to the Federal Power Act;
    (c) Use of an alternative fuel by reason of an order or rule 
under section 125 of the Act;
    (d) Use of an alternative fuel at a steam generating unit to the 
extent that the fuel is generated from municipal solid waste;
    (e) Use of an alternative fuel or raw material by a stationary 
source which:
    (1) The source was capable of accommodating before January 6, 
1975, unless such change would be prohibited under any Federally 
enforceable permit condition which was established after January 6, 
1975 pursuant to 40 CFR 52.21 or under regulations approved pursuant 
to 40 CFR subpart I or 40 CFR 51.166; or
    (2) The source is approved to use under any permit issued under 
40 CFR 52.21 or under regulations approved pursuant to 40 CFR 
51.166;
    (f) An increase in the hours of operation or in the production 
rate, unless such change would be prohibited under any Federally 
enforceable permit condition which was established after January 6, 
1975, pursuant to 40 CFR 52.21 or under regulations approved 
pursuant to 40 CFR subpart I or 40 CFR 51.166.

    Under the NSPS/NESHAP program (see 40 CFR part 60.15 and 61.15), 
the following activities are specifically excluded from the definition 
of a modification:

    (a) Maintenance, repair, and replacement which the Administrator 
determines to be routine for a source category;
    (b) An increase in production rate of an existing facility if 
that increase can be accomplished without a capital expenditure on 
that facility;
    (c) An increase in the hours of operation;
    (d) Use of an alternative fuel or raw material if prior to the 
date of * * * [a particular NSPS or NESHAP], the existing facility 
was designed to accommodate that alternative use. A facility shall 
be considered to be designed to accommodate an alternative fuel or 
raw material if that use could be accomplished under the facility's 
construction specifications as amended prior to the change. 
Conversion to coal required for energy considerations, as specified 
in section 111(a)(8) of the Act, shall not be considered a 
modification.
    (e) The addition or use of any system or device whose primary 
function is the reduction of air pollutants, except when an emission 
control system is removed or is replaced by a system which the 
Administrator determines to be less environmentally beneficial.

    The EPA believes that Congress intended the EPA to base the 
exclusions for HAP modifications under section 112(g) on these existing 
criteria. According to Senator Lautenberg (136 Cong. Rec. S 17124-5 
(October 26, 1990).):

    With the exception of the allowance for a de minimis increase in 
emissions, the definition is identical to the definition of 
modifications in section 111 of existing law. Under this provision, 
the EPA has issued regulations specifying certain kinds of 
activities which would not constitute a modification. Clearly it is 
intended that such kinds of activities would also be excluded from 
triggering the modification definition under the new section 112. 
(emphasis added)

    The proposed rule incorporates a very similar list of exclusions in 
paragraph 63.43(c). Regarding increases in production rate, the 
proposed rule uses the ``capital expenditure'' language in part 60. The 
definition of a ``capital expenditure'' is given in the proposed 
Subpart A ``General Provisions'' to 40 CFR part 63. A standard 
procedure for the determination of ``capital expenditure,'' using 
methods in an Internal Revenue Service (IRS) publication, is provided 
in this definition. Similar to the PSD program, increases in the hours 
of operation are not considered a physical change under the proposed 
rule unless they are prohibited by an existing Federally enforceable 
requirement. The proposed rule uses the effective date of the title V 
permit program as the date by which an alternative fuel or raw material 
must have been accommodated.
    The EPA requests comment on use of these exclusions in the proposed 
rule. In particular, the EPA requests comment on whether a raw material 
substitution involving a substitution of one raw material with another 
raw material of greater hazard should be automatically excluded from 
consideration as a modification. The proposed rule considers 
substitution with a ``more hazardous'' raw material to constitute a 
possible ``modification'' unless the use of the substitute raw material 
was already allowed by a permit. The proposed rule includes a 
definition of ``operations that the major source is designed to 
accommodate'' which allows for materials accommodated by an existing 
permit to be used without triggering section 112(g) requirements. In 
addition, this definition allows for operational changes to be made in 
cases where they are clearly within the permit. For example, some batch 
chemical reaction trains are allowed by permits to produce a number of 
different chemicals. In switching from production of one chemical to 
another, it may be necessary to make a number of pre-approved equipment 
changes. The EPA requests comment on whether such equipment changes, if 
approved in a permit, issued prior to the effective date of the section 
112(g) rule, should constitute ``operations that the major source is 
designed to accommodate'' and should not trigger a ``modifications'' 
review pursuant to section 112(g).
    The EPA also seeks comment on whether such operational changes, 
contained in permits issued prior to the effective date of the section 
112(g) rule, should be incorporated into a title V permit without 
triggering section 112(g) review.
    The EPA also seeks comment on whether such operational changes, 
contained in permits issued prior to the effective date of the section 
112(g) rule, should be incorporated into a title V permit without 
triggering section 112(g) review.
    In addition, Sec. 63.45(c)(3) of the proposed rule provides that a 
source may seek approval of case-by-case MACT determination for new 
alternate operating scenarios (that were not incorporated in a State 
permit) when obtaining it's title V permit. As a result, the source 
would then be free to activate any such alternative operating scenario 
without having to undergo further section 112(g) review. The EPA 
requests comment on whether the approach contained in proposed 
Sec. 63.45(c)(3) is an appropriate approach to application of section 
112(g) requirements to alternate operating scenarios.
    4. Calculation of Actual Emissions Increase. (Step 2 of the Process 
to Identify a Modification). Once a physical or operational change has 
been identified for a given emission point or set of emission points, 
the next step is to determine whether there has been an increase in 
``actual emissions,'' and to calculate the amount of the increase. If 
such an ``actual emission increase'' is more than a de minimis level 
(per Sec. 63.44 of the proposed rule, see discussion below), then the 
change constitutes a ``modification.''
    Any method for calculating an increase must provide for a 
``before'' case, often referred to as the ``baseline,'' and an 
``after'' case representing the emission after the change.
    When the physical change involves addition of a new emission point, 
the baseline is zero emissions and it is only necessary to define 
anticipated future emissions. For the proposed rule, the ``after'' case 
is considered to be the potential to emit. ``Potential to emit'' is 
defined in subpart A. Physical and operational limitations can be 
considered if the limitations are Federally enforceable.
    When the physical or operational change involves an emission 
increase from already existing equipment, emissions before and after 
the change must be compared.
    In developing an approach to this case for the proposed rule, the 
EPA reviewed two approaches to emission increase calculations which 
have been used in past air pollution programs for criteria air 
pollutants. The first approach is the approach used in the new source 
performance standard (NSPS) program to determine whether ``any emission 
increase'' has occurred due to a physical change or change in the 
method of operation. The second approach is the ``actual emission 
increase'' approach used in the PSD program.
    The approach used for the NSPS program (and the NESHAP program in 
40 CFR part 61, before the 1990 Act amendments) is more straightforward 
than that used for the PSD program. For these regulations, a 
``modification'' occurs if the physical change leads to ``any 
increase'' in emissions. In making this determination, they must follow 
the approach outlined in 40 CFR 60.14(b). This approach considers ``any 
increase'' to occur if the source, operating at its production 
capacity, will release more emissions (on a pound per hour basis) to 
the atmosphere. If the emission factor (that is, the amount of 
emissions per unit of production) increases, or the equipment is 
otherwise inherently more emitting (for example, due to a capital 
expenditure increasing the size or capacity of equipment), an increase 
is considered to have occurred and the equipment is subject to the 
NSPS. In some cases, source tests before and after the change are used 
to demonstrate whether an emission increase has occurred.
    This approach is fairly straightforward to implement and relies on 
immediately available data rather than past records. In addition, 
Senator Lautenberg's belief that the definition of actual emissions is 
``identical to that in section 111'' may suggest that Congressional 
intent was for the NSPS approach to identifying modifications.
    The EPA believes that calculations based upon the NSPS method could 
serve as a reasonable surrogate for an ``actual emissions'' 
calculation. In an ideal sense, a true ``actual emissions'' calculation 
would require perfect knowledge of the level of emissions that actually 
occurred in the past, and perfect knowledge of the emissions that would 
actually occur in the future if the change were to take place. In 
practice, past emissions are difficult to document (and for HAP, 
perhaps impossible if appropriate data have not been collected), and 
future emissions cannot be predicted with certainty. In this context, 
the EPA believes that a policy decision can be made to consider the 
NSPS test as a possible method for actual emissions calculations. 
Although this approach does not attempt to gather ``actual'' data on 
past emission rates, the EPA believes that it can be a reasonable 
surrogate for describing the ``actual'' difference between future and 
past emissions.
    The term ``actual emissions increase'' has been used in the PSD 
program. For PSD, the term ``baseline'' is used to describe emissions 
before a physical or operational change. The ``baseline'' for actual 
emissions for an emission unit as of a particular date is defined as 
the average rate, in tons per year, at which the unit actually emitted 
the pollutant during a 2-year period which precedes the particular date 
and is representative of normal source operation. A different time 
period is allowed if the permitting authority deems that it is more 
representative of normal source operation. Under this approach, 
emissions after the operational change are the potential to emit, in 
tons per year. This approach requires that records be supplied of the 
actual rates of operation during the baseline period.
    The EPA is concerned that the PSD approach may be administratively 
complex. For the PSD program, this approach to applicability has proven 
very complex for criteria pollutants such as volatile organic compounds 
(VOC). Protracted discussions are often required to establish the 
appropriate time period for the actual ``before'' case emissions and to 
approve documentation for the actual rates of production and operation. 
This approach would be more complex for HAP for which VOC totals would 
need to be speciated into individual HAP subtotals. It may be very 
difficult or impossible in many cases to provide for adequate 
documentation of these HAP subtotals.
    The EPA also notes that there is an ongoing project aimed at 
reforming the PSD program. A number of options are being considered. If 
the PSD program is revised to accommodate one of these approaches, that 
approach may be applicable to section 112(g) implementation. The PSD 
reform project is proceeding in parallel with the effort to develop 
this proposed rule. The EPA requests comment on whether any suggested 
applicability approach in the proposed rulemaking for a restructured 
PSD program should be selected as the approach to implementation of 
section 112(g) of the Act. The EPA requests comment on whether any of 
these approaches should be included in the final rule.
    The proposed rule contains, as paragraph 63.43(d), an approach to 
``actual emissions'' that closely resembles the NSPS approach. The EPA 
believes that this approach will yield a more consistently implemented 
program that ensures a technology review when a physical change causes 
increased emissions during the operation of the equipment being 
changed. The approach differs slightly from the NSPS approach for 
pollutants for which the de minimis value listed in Sec. 63.45 of the 
proposed rule is expressed as a tons per year value. For such 
pollutants, paragraph 63.43(d) includes a two-step process. The first 
step is to determine the pounds per hour increase, as would be done for 
the NSPS test. The second step is to convert the pounds per hour value 
to a tons per year value based upon the future hours of operation of 
the equipment. For this conversion, it is assumed that the equipment 
will operate 8760 hours per year unless constrained by a Federally 
enforceable limitation.
    5. Paragraph 63.43(e)--Offsets. If a physical change leads to 
actual emission increases by more than a de minimis amount, that 
increase is not a ``modification'', if, according to section 
112(g)(1)(A) of the Act:

such increase in the quantity of actual emissions of any hazardous 
air pollutant from such source will be offset by an equal or greater 
decrease in the quantity of emissions of another hazardous air 
pollutant (or pollutants) from such source which is deemed more 
hazardous

    Paragraph 63.43(e) of the proposed rule incorporates this offset 
provision. This paragraph allows the owner or operator to submit a 
showing (consistent with Sec. 63.46 and Sec. 63.47, see discussion 
below) to demonstrate the acceptability of the offset. The process for 
determining whether offsets are ``deemed more hazardous'' is contained 
in Sec. 63.48 of the proposed rule (see discussion in section IV of 
this preamble).
    Paragraph 63.43(e) identifies some general restrictions on the 
offsets that are used. (Additional restrictions are contained in 
Secs. 63.46 and 63.47 of the proposed rule.) First, a decrease in 
actual emissions cannot credit any amount of actual emissions that 
exceeds allowable emissions under a Federally enforceable requirement. 
Second, the decrease must be Federally enforceable before operation of 
the physical change being offset. There are a number of mechanisms for 
Federal enforceability including: (1) A Notice of Offset approval under 
Sec. 63.46 or Sec. 63.47, a requirement of a State program approved 
under section 112(l) of the Act, (2) a permit condition contained in a 
permit issued pursuant to 40 CFR part 70 or 40 CFR part 71, (3) a 
Federally enforceable requirement of a PSD or NSR permit, (4) a 
requirement of a Federally approved State Implementation Plan, or (5) a 
Federally enforceable court order. Third, the owner or operator may not 
credit any emission decreases used under the ``early reductions'' 
program to obtain the compliance extensions granted by section 
112(i)(5) of the Act. Any amount exceeding the 90 (or 95) percent 
reduction required by the early reduction program is, however, 
creditable. The EPA considered adding a further restriction on: (1) 
Emission reductions of volatile organic compounds (VOC) that were 
necessary to achieve progress towards attainment of the ozone standard, 
and (2) reductions of VOC and other pollutants previously credited 
under the PSD or NSR programs. The proposed rule would in some cases 
allow such emission reductions to be creditable as offsets. The EPA 
requests comment on this issue.
    6. Paragraph 63.43(f). Increases and Decreases of the Same 
Pollutant. There is some ambiguity in the Act regarding cases where a 
modification leading to an increase in a given pollutant will be 
accompanied by a decrease in the same pollutant elsewhere in the plant. 
Paragraph 63.43(f) is intended to clarify EPA's position on this issue. 
For such cases, the proposed rule requires that the emission decreases 
be documented using the procedures of Sec. 63.46 or 63.47 of the 
proposed rule. If the net emission increase is less than de minimis, 
then a modification has not occurred.
    The EPA considered an alternative that would require that emission 
decreases in such cases to completely offset the increase such that an 
overall decrease would occur. The EPA believes that the requirements of 
proposed Sec. 63.43(f) reflect the most natural reading of the statute. 
The EPA requests comment on this issue.

E. Section 63.44. De Minimis Levels

    As mentioned previously, an emission increase must exceed de 
minimis levels in order to constitute a ``modification'' under section 
112(g) of the Act. The proposed rule includes, as Sec. 63.44, a table 
displaying de minimis emission rates for each of the HAP.
    1. De Minimis. General Principles. The statute gives little 
specific direction on how to establish de minimis quantities. In 
establishing de minimis values, the EPA believes there are general 
principles that have been established. A good discussion of these 
principles is included in the April 20, 1979 Alabama Power decision. 
(Alabama Power v. Costle, 656 F. 2nd 323 (1979). Generally, de minimis 
authority gives regulatory agencies such as the EPA the ability to 
provide exemptions when ``the burdens of regulation yield a gain of 
trivial or no value.'' Further, ``the de minimis exemption must be 
designed with the specific administrative burdens and specific 
regulatory context in mind.'' The overall intent of such exemptions is 
to prevent relatively trivial items from needlessly draining 
administrative resources.
    In keeping with these general principles, the EPA believes the main 
test in establishing de minimis values is to define the emission level 
for HAP for which regulation under section 112(g) would ``yield a gain 
of trivial or no value.''
    It appears that some limited consideration can be given to 
administrative resource implications resulting from a selected de 
minimis emission level. For example, if a selected de minimis cutoff 
yielded very small benefits, but would increase the number of 
applications, reviews, and enforcement resources by an unreasonable 
degree, and a slightly higher cutoff would not create as unreasonable a 
burden, the EPA believes that this could be taken into consideration.
    The EPA recognizes, however, that its authority to provide de 
minimis exemptions is strictly limited. Again citing Alabama Power,

    That implied authority * * * [for de minimis] * * * is not 
available for a situation where the regulatory function does provide 
benefits, in the sense of furthering the regulatory objectives, but 
the agency concludes that the acknowledged benefits are exceeded by 
the costs.

The EPA believes that the interpretation of de minimis detailed herein 
correctly balances the requirement to meet regulatory objectives, while 
alleviating the burdens of regulation which would yield a trivial value 
in this specific regulatory context.
    2. De Minimis Concepts in Section 112 of the Act. The EPA believes 
that Congress has provided guidance in what is considered to be a 
``trivial'' level of a hazardous air pollutant. In section 112(c)(9) of 
the Act, the EPA may delete a source category from possible regulation 
under section 112 if no source in the source category would result in: 
(1) Emission of a carcinogen that could cause a lifetime risk of cancer 
of one in one million to the individual in the populations who is most 
exposed, and (2) emission of a non-carcinogen that would exceed air 
quality levels that would exceed a level adequate to protect public 
health with an ``ample margin of safety'' and would not result in 
adverse environmental impacts. The concepts behind section 112(f) of 
the Act appear similar for ``residual risk'' emission standards to 
address risks remaining after application of technology-based 
standards. The EPA believes that it is reasonable to use the one-per-
million cancer risk and ample margin of safety criteria for 
establishing de minimis levels under section 112(g). The EPA has 
traditionally believed that exposures that cause a risk above one in 
one million are considered important.
    3. Basis for de minimis Values Listed in Sec. 63.44 of the Proposed 
Rule. The table in Sec. 63.44 of the proposed rule lists the de minimis 
values for the 189 HAP listed in section 112(b) of the Act. Readers 
interested in documentation for each de minimis value can refer to a 
technical support document, Background Document. Documentation of De 
Minimis Emission Rates for Proposed 40 CFR part 63, subpart B. (EPA-
453/R-93-035) The following discussion is intended to provide an 
overview of the methods used to develop these values.
    The section 112(b) list includes 172 pollutants that are listed as 
individual chemicals, and 17 pollutants that are listed as chemical 
groups. Where appropriate, the 17 chemical groups were subdivided into 
sub-groupings or individual compounds within the group. The table 
indicates, for each chemical or chemical group, a de minimis emission 
rate and the basis for each de minimis rate.
    The EPA considered expressing the de minimis values as ambient 
concentrations, rather than emission rates. This would require either 
the applicant or the permitting authority to perform a dispersion 
calculation for each proposed release to determine whether a de minimis 
concentration would be exceeded. The EPA believes that this would 
greatly increase the complexity and thus the resources need to 
implement the program. Although the EPA believes that States wishing to 
include this dispersion review as part of an overall section 112(g) 
program should be given the flexibility to do so (see discussion 
below), the EPA believes that most States would prefer de minimis 
values to be expressed as emission rates, rather than concentrations.
    It is important to note that the de minimis values listed in 
Sec. 63.44 were developed specifically for the section 112(g) program, 
and that the values were developed in part based upon the interim 
nature of the time period for which case-by-case MACT determinations 
are required. Such case-by-case MACT determinations are required under 
section 112(g) prior to emission standards promulgated pursuant to 
section 112(d). The EPA does not consider these values to be 
necessarily indicative of the emission rate which may be considered de 
minimis for other programs or decisions, for which the decisions would 
be more long-term in nature. In particular, these values should not be 
considered as precedent-setting for other section 112 issues such as 
the residual risk standard-setting process under section 112(f) or the 
risk criterion established for delisting categories pursuant to section 
112(c) of the Act.
    (a) Values for ``nonthreshold'' HAP which have evidence of 
carcinogenicity. For ``nonthreshold'' HAP which have evidence of 
carcinogenicity (see discussion in section IV.C.3 of this preamble for 
the rationale for identifying such pollutants), the following 
descriptors are used in the ``Basis'' column in the table:

--UR
--UR-CAP
--DEF=1

    The ``UR'' descriptor indicates that the de minimis value was 
calculated based upon a risk-specific dose for the pollutant. The risk-
specific dose is the exposure level associated with a given lifetime 
cancer risk, in this case, a risk management decision of 10-6 
lifetime risk. The risk-specific dose is derived from the unit risk, an 
upper-bound estimate of the excess cancer risk over background 
associated with a continuous lifetime exposure to the pollutant. 
Readers should be aware that there are many uncertainties in the 
derivation of unit risk.
    De minimis emission rates were calculated in four steps. First, 
based upon the unit risk value, the EPA calculated the concentration in 
the ambient air that would yield a lifetime cancer risk of one-per-
million. Using benzene as the example, lifetime continuous exposure to 
1 microgram per cubic meter of benzene is associated with a risk as 
high as 8.3 in one million, and a lifetime risk of one-per million is 
equivalent to 0.12 micrograms per cubic meter (one divided by 8.3).
    The second step in the calculation was to adjust the risk-specific 
dose to account for the expected maximum exposure duration for a major 
source subject to a case-by-case MACT determination under section 
112(g) of the Act. The EPA selected a 7-year exposure period as the 
duration of exposure, rather than the more frequently used 70-year 
lifetime exposure. The 7-year period was selected because emission 
increases avoiding modification requirements under a section 112(g) de 
minimis exemption would be still subject to maximum achievable control 
technology requirements within roughly 7 years under sections 112(j) or 
112(d). The EPA is required to promulgate MACT standards in accordance 
with a schedule in section 112(e) of the Act by November 15, 2000. Such 
standards would require compliance for existing sources by no later 
than the year 2003. Even if the EPA does not meet every deadline in its 
schedule for promulgation of section 112(d) emission standards, States 
are required to develop equivalent emission standards within 18 months 
after the EPA fails to meet a deadline. As a result, the longest time 
for which standards would not be developed is 18 months after November 
15, 2000, i.e., May 2002. Because the section 112(g) program will start 
up in most States in early 1995, (as soon as operating permit programs 
commence) about 7 years (2002 minus 1995), is a reasonable, 
conservative estimate of the time that would elapse before imposition 
of technology requirements for emission increases avoiding 
``modification'' requirements.
    Adjusting for this 7-year exposure period, using benzene as the 
example, a lifetime risk of one-per-million (0.12 micrograms per cubic 
meter over 70 years) is equivalent to the risk associated with exposure 
to 1.2 micrograms per cubic meter over 7 years. The EPA requests 
comment on this adjustment. Other exposure adjustments were considered, 
including: (1) No exposure adjustment, and (2) adjustment by a factor 
less than 70/7. The EPA requests comment on whether these or other 
alternatives better identify emission increases which can be considered 
de minimis for this program.
    As a third step, in order to express the de minimis values as 
emission rates, rather than ambient concentration, the EPA developed a 
``model plant.'' This model plant represents a standard set of 
conditions for the nature of the release and the exposure. The 
following model plant was used:

--stack height: 10 meters
--stack diameter: 1 meter
--distance to nearest exposed individual: 200 meters
--stack temperature: ambient
--exit velocity: 0.1 m/sec
--worst-case down-wash is assumed

    The EPA proposes these conditions as a reasonable set of conditions 
for purposes of setting de minimis values under section 112(g) of the 
Act.
    For this model plant, the EPA performed calculations using 314 sets 
of meteorological data. (A complete description of these calculations 
is contained in the docket for the proposed rule.) The results of these 
calculations indicated, on average, that for each microgram per cubic 
meter of a pollutant added to the atmosphere at the assumed fence-line 
of 200 meters, there would be 2 tons of emissions. This ratio, 2 tons/
yr per 1.0 g/m\3\, annual average, was used as the 
relationship between emission rate and ambient concentration.
    As a fourth step, the EPA used the risk-specific dose at a one-per-
million risk, identified in Step 2 above, in tandem with the 
relationship between emission rate and concentration developed in Step 
3, to calculate a de minimis emission rate. For example, for benzene, 
Step 2 indicated an exposure associated with one-per-million risk of 
1.2 g/m\3\ over the 7-year exposure period. In order to reach 
this exposure level, the model plant would need to emit 1.2  x  2, or 
2.4 tons/year of benzene. For purposes of the proposed rule (i.e., the 
table in Sec. 63.44), each of the values is rounded to one significant 
figure; for benzene, 2.4 tons/year is rounded to 2 tons/year. The EPA 
believes that one significant figure is appropriate, given the 
uncertainties in the unit risk values and exposure assumptions on which 
the values are based.
    The EPA requests comment on the methodology for de minimis values, 
including the appropriateness of the assumptions used to develop the 
model plant. It is recognized that there are other model plant 
assumptions that would result in less dispersion, and that the selected 
model plant does not represent an absolute worst-case. For example, 
less dispersion could be experienced for: (1) Releases for which 
weather conditions represent the worst-case of the 314 stations, rather 
than the median of the 314 stations, (2) releases at ground level, 
rather than the assumed height of 10 meters, (3) releases immediately 
adjacent to residences, which could occur at distances less than the 
assumed 200 meters. The results of the dispersion calculations (which 
are listed in Appendix A of the technical background document) 
indicated that the highest concentration experienced at the 314 
stations was 15.6 g/m3, while the lowest concentration 
was 2.2 g/m3. The median value, 5.0 /m3, 
therefore, could underpredict by a factor of approximately 3, or could 
overpredict by a factor of approximately 2. The EPA also explored the 
sensitivity of the results to stack height and distance to nearest 
receptor. The following table illustrates this sensitivity analysis. 
(Stack release parameters not shown in the table are identical to those 
listed above.) The results indicate that, for a given 10 tons/year 
release, the resulting concentration could be significantly higher 
than, or significantly lower, than that resulting from the selected 
model plant. 

------------------------------------------------------------------------
                                              Distance to      Median   
                                                nearest    concentration
           Stack height (meters)               residence    (g/
                                               (meters)         m3)     
------------------------------------------------------------------------
1...........................................          200         16    
3...........................................          100         34    
3...........................................          500          3.4  
10..........................................          100          6.7  
10..........................................          200         *5.0  
10..........................................          500          2.8  
15..........................................          200          2.5  
50..........................................          200          0.15 
100.........................................          200          0.026 
------------------------------------------------------------------------
*These conditions are the model plant used for the proposed rule.       

    The EPA is considering an approach for which two tables would be 
required to set de minimis values: (1) The table in Sec. 63.44, and (2) 
a second table which would include an adjustment factor for site-
specific conditions. For example, the adjustment factor would lead to 
lower de minimis values for sources with a 1 meter stack height and 50 
meter distance to the receptor, but higher de minimis values for a 
source with a 25 meter stack and 1000 meter distance to the receptor. 
This approach would not require the applicant or the permitting 
authority to perform site-specific dispersion calculations. Rather, the 
table would specify adjustment factors that would apply to given ranges 
of conditions. (For example, there could be an adjustment factor 
applicable to stack heights from 1 to 3 meters, in combination with 
distances 100 meters or less to the receptor). This approach would have 
the advantages of taking site-specific variables into account (the EPA 
also requests comment on whether other variables, such as flow rate and 
temperature could be included). The EPA has two concerns with this 
approach. First, the approach would add administrative complexity to 
the process, in that documentation and enforcement of stack height, 
distance, etc., would be required. Second, the EPA requests comment on 
the policy advantages and disadvantages of an approach that would yield 
different levels of control for similar equipment.
    The ``UR-CAP'' description in the table indicates that the unit 
risk approach yielded an emissions rate greater than 10 tons per year. 
Emissions of 10 tons per year or more of such pollutants from the model 
facility would yield risk levels below the de minimis bench mark. The 
proposed rule ``caps'' de minimis emission rates at 10 tons per year 
because the EPA believes that it would be difficult to assume that 
Congress intended, simultaneously, for an emission rate to be 
considered both ``major'' for identifying major sources and smaller 
than ``trivial'' for emission increases. However, the EPA is concerned 
that capping de minimis rates at 10 tons per year could bring sources 
into the program with modifications that pose a trivial threat to human 
health. The EPA seeks comments on this approach, on the feasibility of 
promulgating de minimis emission rates above 10 tons per year, and on 
the desirability of capping de minimis rates at a level less than 10 
tons per year.
    The ``DEF=1'' descriptor indicates that the pollutant was assigned 
a default value of 1 ton/yr. This default value was assigned for 
pollutants identified as possible, probable or known human carcinogens, 
but for which no unit risk value was available. The choice of 1 ton/yr 
is a policy decision based upon a review of the pollutants with potency 
values. The EPA does not believe that these pollutants should be 
assigned the 10 tons/year cap; if potency values were available, and 
were consistent with the other such pollutants, the value would likely 
be less than 10 tons per year. The EPA requests comment on this default 
value.
    (b) Values for noncancer effects. For the remaining pollutants on 
the section 112(b) list that have not been evaluated for 
carcinogenicity or which have been assigned a weight of evidence 
classification of D or E, the EPA believes that de minimis values 
should be established that would be consistent with concentration 
benchmark that represent an ``ample margin of safety.''
    The descriptor ``RfC'' in the table indicates that the de minimis 
emission rate was calculated based upon the EPA's inhalation reference 
concentrations (RfC's). The RfC is defined as an estimate (with an 
uncertainty spanning perhaps an order of magnitude) of a daily exposure 
to the human population (including sensitive subgroups) that is likely 
to be without appreciable risk of deleterious effects during a long-
term period of exposure. For the proposed rule, the EPA assumes that 
the RfCs themselves represent an ample margin of safety level for 
noncancer effects from long-term exposures. The methodology for 
developing RfCs is discussed in Interim Methods for Development of 
Inhalation Reference Concentrations, EPA/600/8-90-066A. For the 
relatively few chemicals for which RfCs have been established, the EPA 
used a similar method to that described above for carcinogens, with the 
exception that there was no adjustment for the exposure period. No 
adjustment is made here because the RfC is designed to protect against 
chronic exposures, which is interpreted as less than lifetime (i.e., 7 
years in this case).
    There are a number of pollutants that have not been associated with 
cancer, and for which RfCs have not been developed. For these 
pollutants, default de minimis emission rates are derived from 
composite score values. The descriptor ``CS'' indicates when this 
approach was used. The composite score is a chronic toxicity ranking 
system developed for establishing reportable quantities (RQs) under 
section 102 of the Comprehensive Environmental Response, Compensation, 
and Liability Act (CERCLA). The basis for these composite scores is 
described in greater detail in section III of this preamble. For the 
CERCLA section 102 program, RQs are established for chronic noncancer 
effects as follows: 

------------------------------------------------------------------------
                         CS value                            RQ (pounds)
------------------------------------------------------------------------
1-5........................................................         5000
6-20.......................................................         1000
21-40......................................................          100
41-80......................................................           10
81-100.....................................................            1
------------------------------------------------------------------------

The EPA believes that for section 112(g) purposes, pollutants with a 
relatively low CS value, (less than or equal to 20) can be assigned a 
de minimis emission rate at the 10 tons per year maximum. For 
pollutants with greater CS values, the EPA believes that de minimis 
emission rates less than the 10 tons per year ``cap'' should be 
established. For the proposed rule, the EPA has assigned de minimis 
emission rates that mirror the magnitude of the difference in the RQ 
values. That is, pollutants with CS scores between 21 and 40 are 
assigned de minimis emission rates which are a factor of 10 less than 
those for which the CS score is between 6 and 20. Pollutants with CS 
scores greater than 40 are assigned de minimis emission rates that are 
a factor of 10 less than those for which the CS score is between 21 and 
40. (Note that none of the threshold HAP for section 112(g) have a CS 
value greater than 80.) The EPA believes that the assignment of the de 
minimis emission rates in this manner for these ranges of CS values 
represents a reasonable default approach in absence of a more rigorous 
method. The EPA requests comment on this issue and on other approaches 
that could be used. The default assumptions are as follows: 

                                                                        
------------------------------------------------------------------------
     Range of composite score              De minimis emission rate     
------------------------------------------------------------------------
CS = 1 to 20.......................  10 tons/yr.                        
CS = 21 to 40......................  1 ton/yr.                          
CS = 41 and greater................  0.1 ton/yr.                        
------------------------------------------------------------------------

    The descriptor ``DEF=5'' indicates the method used for pollutants 
with neither composite score, reference concentration, weight of 
evidence indicating carcinogenicity, or identified as acutely toxic 
(see section C below). For such pollutants, a default value of 5 tons/
year was used. This value is greater than the 1 ton/year value for the 
pollutants which may be carcinogens. The EPA believes that this is 
reasonable considering the values for the other pollutants. The EPA 
requests comment on the selection of this 5 tons/year default value.
    (c) de minimis values for short-term exposures. As discussed below 
in section IV.C. of this preamble, several pollutants on the HAP list 
produce health effects from short-term exposures. Examples of these 
pollutants include arsine, phosgene, and methyl isocyanate. These 
pollutants are addressed as follows.
    First, a policy decision was made to assign a default annual de 
minimis emission rate of 0.1 tons per year. In this way, these 
pollutants are assigned a de minimis emission rate that is equal to the 
default value assigned for the pollutants of highest concern for 
chronic noncancer health effects.
    Second, for a number of these pollutants, the EPA considered 
establishing and listing in the table short-term de minimis emission 
rates, expressed in pounds/hour. The EPA believes that such short-term 
values would be a better indicator of de minimis for pollutants which 
the primary concern is health effects resulting from short-term 
exposures. At this time, however, the EPA is not proposing these 
pounds/hour de minimis levels for three reasons. First, the EPA has not 
established consistent procedures for establishing RfCs for short-term 
exposures. Ideally, de minimis values for acutely toxic pollutants 
should be based on such short-term RfCs. However, there is currently 
only one short-term RfC which has been developed by the EPA (a value of 
0.3 ppm for developmental toxicity by exposures of 30 minutes or less 
to ethylene oxide). Second, the EPA feels that additional information 
is needed on whether the inclusion of such short-term values would add 
significantly to the overall scope of the section 112(g) program. 
Finally, the EPA has some concern with the potential difficulty of 
collecting or reporting data on short-term emission rates from affected 
facilities.
    The EPA is considering an interim method to establish short-term de 
minimis values based upon Levels of Concern (LOC). The LOC have been 
established for chemicals on the Superfund Amendments and 
Reauthorization Act (SARA) title III section 302 list of ``extremely 
hazardous substances.'' The LOC indicate levels of airborne 
concentrations of chemicals for which no serious irreversible health 
effects occur following a short-term exposure (30 minutes). The LOC are 
by definition one-tenth of ``Immediately Dangerous to Life and Health'' 
levels (IDLH) produced by National Institute for Occupational Safety 
and Health (NIOSH).
    The EPA believes that LOC have some possible merit for use in 
setting short-term de minimis values. The LOC are the only values, of 
which the EPA is aware, which have an extensive data-base and are 
designed to protect from serious effects of short-term or acute 
exposures. The LOCs are intended to protect the general population, 
including sensitive individuals.
    There are, however, several disadvantages for using LOC to set de 
minimis levels; that is, in establishing a level below which public 
health is protected with an ample margin of safety for non-carcinogenic 
effects. Most of the LOC values are based upon animal lethality data. 
Benchmarks derived from such data may not protect against all health 
effects in humans. In addition, the safety factor of 10 which is 
applied to IDLH to protect sensitive individuals of the population and 
for protection against serious health effects may not be adequate. 
There are questions concerning the scientific peer review of the 
rationale for each LOC and supporting data. Finally, it is not known 
what the maximum duration of exposure at the LOC would be for 
protection against adverse effects.
    Despite these serious disadvantages, LOC may be appropriate on an 
interim basis for setting short-term de minimis levels for acutely 
toxic pollutants in the absence of a better methodology and data. The 
EPA requests comment on the methodology and short-term de minimis 
emission rates described below and seeks suggestions concerning other 
methods and supporting data to use in determining such concentrations. 
In addition, the EPA requests comment on the number of major sources of 
HAP whose modifications would cause such de minimis emission rates to 
be exceeded.
    The methodology the EPA is considering is as follows. First, for 
each pollutant of concern for acute exposures, a short-term de minimis 
concentration for each pollutant would be derived by dividing its LOC 
by a safety factor of 1000. This factor of 1000 is a crude estimate of 
the factor needed to convert the LOC, which is based upon mortality or 
very severe effects, into a level that would ensure that no adverse 
human health effects would be observed. Second, a ``reasonable worst 
case'' model plant is developed to describe the relationship between 
the de minimis concentration and a de minimis emission rate. For the 
examples described below, the same model plant was used as that 
described above for setting de minimis levels for long-term exposures:
    Stack height is 10 meters;
    Exit gas velocity is negligible;
    Stack diameter is 1 meter;
    Exit gas temperature is equal to the ambient temperature;
    Worst-case down-wash is assumed;
    The nearest exposed individual is at a distance of 200 meters.
    For this model plant, the ``Tier 1 screening approach'' described 
in A Tiered Modeling Approach for Assessing the Risks Due to Hazardous 
Air Pollutants, EPA-450/4-92-01, is used to describe the relationship 
between the de minimis concentration and a pound/hour de minimis 
emission rate. Use of this approach results in a ratio of maximum off-
site short-term concentration to emission rate of 314 (micrograms/
m\3\)/(lb/hr) or 0.314 (milligrams/m\3\)/(lb/hr). This factor indicates 
that the prototypical facility which emits 1 pound of pollutant in an 
hour will have a maximum short-term concentration off-site which will 
equal to 0.314 milligrams/m\3\.
    The short-term concentration predictions made using the Tier 1 
method are interpreted as 1-hour average concentrations, i.e., they 
account for the dilution due to the general meander of a dispersed 
plume over the course of a 1-hour period. Since the de minimis 
concentration values relate to ``peak'' or very short-term exposure 
levels (maybe on the order of a few seconds), the EPA believes it would 
be desirable to derive peak concentration values from the 1-hour 
predictions. Data taken by the EPA indicate that the concentration 
levels during any few second time interval within the 1-hour period 
will not vary more than a factor of two. Therefore, for purposes of the 
examples described below, a ``peak-to-mean'' ratio of two was used, 
that is, the peak concentration is assumed to be twice that of the 1-
hour average.
    Using the value, [(0.314 milligrams/m\3\)/(lb/hr)], coupled with 
the peak-to-mean ratio of two, the de minimis emission rate, Edm, 
from the de minimis concentration level, Cdm, for each acutely 
toxic pollutant would be calculated as follows:

Edm=[Cdm/(2)]/0.314

    The following lists a number of examples illustrating the LOCs and 
the short-term de minimis emission rates that would result based upon 
this method. (Note: the value for ethylene oxide is derived from a 
short-term RfC of 0.3 parts per million, rather than an adjusted LOC). 
The EPA requests comment on whether the final rule should incorporate 
these values, and on other possible alternative methods that could be 
used to derive short-term de minimis emission rates.
    (d) de minimis values for pollutants having multiple health 
concerns. Some HAP may produce a spectrum of health effects including 
both cancer and effects other than cancer, including acute health 
effects. The de minimis value for these pollutants was the lowest value 
calculated for the cancer and chronic noncancer health effects using 
the procedures described in paragraphs (a) and (b). A short-term 
pounds/hour de minimis emission rate, if developed, would, for some 
pollutants, appear in addition to the annual emission rate.
    4. Proposed de minimis considerations for pollutants of concern 
under EPA's section 112(m) Great Waters program. The descriptor ``GWP'' 
in the table of de minimis values indicates that a value of 0.01 tons 
per year was a ``Great Waters Pollutant'' for which a special de 
minimis value was assigned as a policy decision. The EPA requests 
comment on this selected value and on several alternatives that were 
considered.

                                                                        
------------------------------------------------------------------------
                                                             Short-term 
                                                             de minimis 
      CAS #              Pollutant          LOC (mg/m\3\)    value (lbs/
                                                                 hr)    
------------------------------------------------------------------------
107028...........  Acrolein.............  1.15............       0.00183
7783702..........  Antimony               2.70............       0.00430
                    pentafluoride.                                      
1303282..........  Arsenic pentoxide....  8.00............        0.0127
1377533..........  Arsenic oxide........  1.40............       0.00223
7784421..........  Arsine...............  1.90............       0.00302
94077............  Benzotrichloride.....  0.700...........       0.00111
100447...........  Benzyl chloride......  5.18............       0.00824
57578............  beta-Propiolactone...  1.50............       0.00239
1366190..........  Cadmium oxide........  4.00............       0.00637
7782505..........  Chlorine.............  7.25............        0.0115
79118............  Chloroacetic acid....  1.80............       0.00286
107302...........  Chloromethyl methyl    1.82............       0.00290
                    ether.                                              
10025737.........  Chromic Chloride.....  0.0500..........     0.0000795
10210681.........  Cobalt carbonyl......  0.270...........      0.000430
77781............  Dimethyl sulfate.....  5.00............       0.00800
534521...........  4,6-Dinitro-0-cresol   0.500...........      0.000800
                    and salts.                                          
151564...........  Ethyleneimine........  4.00............       0.00636
75218............  Ethylene oxide\1\....  0.3ppm..........  ............
62207765.........  Fluomine.............  3.00............       0.00477
77474............  Hexachlorocyclopentad  0.0195..........     0.0000310
                    iene.                                               
7664393..........  Hydrogen fluoride....  1.64............       0.00261
7783075..........  Hydrogen selenide....  0.660...........       0.00105
12108133.........  Methylcyclopentadieny  0.600...........      0.000955
                    lmanganese.                                         
60344............  Methyl hydrazine.....  0.940...........       0.00150
624839...........  Methyl isocyanate....  4.70............       0.00748
13463393.........  Nickel carbonyl......  0.350...........      0.000557
56382............  Parathion............  2.00............       0.00318
75445............  Phosgene.............  0.800...........       0.00127
7723140..........  Phosphorous..........  3.00............       0.00477
151508...........  Potassium cyanide....  5.00............       0.00796
143339...........  Sodium cyanide.......  5.00............       0.00796
13410010.........  Sodium selenate......  1.60............       0.00255
10102188.........  Sodium selenite......  2.30............       0.00366
78002............  Tetraethyllead.......  4.00............       0.00637
75741............  Tetramethyllead......  4.00............       0.00637
7550450..........  Titanium               1.00............       0.00159
                    tetrachloride.                                      
584849...........  Toluene diisocyanate.  7.00............        0.0111
------------------------------------------------------------------------
\1\This is not a LOC but a short-term RFC for a 30-minute exposure and  
  is in ppm rather than mg/m\3\. The value in mg/m\3\ is 0.54.          

    The EPA believes that de minimis values under section 112(g) can 
take into account a hazardous air pollutant's potential for causing 
non-air quality health and environmental impacts. For example, 
deposited pollutants which are persistent and bioaccumulate are of 
special concern to the living resources in the ecosystem into which 
they are deposited. The EPA is required by section 112(m) of the Act to 
investigate the potential for adverse impacts of atmospheric deposition 
to the Great Lakes, Chesapeake Bay, Lake Champlain and Coastal Waters 
(collectively referred to as the ``Great Waters.'' Interim results of 
these investigations indicate the following 13 HAP appear to be of the 
greatest concern: lead and lead compounds, polycyclic organic matter 
(POM), hexachlorobenzene, mercury, polychlorinated Biphenyls (PCBs), 
chlorinated dioxins, chlorinated furans, toxaphene, chlordane, DDE, 
D(lchloro)D(lphebyl)T(Richloroethane) (DDT), lindane, a-hexachlorcyclo-
hexane, and cadmium. Ref: Swain et al., Exposure and Effects of 
Airborne Contamination for the Great Waters Program Report. December 
22, 1992.
    For these pollutants, the EPA does not believe that methods are 
currently available to quantify the relationship between emission rates 
and exposures for these pollutants. Accordingly, the EPA does not 
believe that a quantitative method for developing de minimis values yet 
exists. However, since these reports identify these specific pollutants 
as posing a potentially serious environmental risk, the EPA believes 
that it is appropriate to place greater emphasis by assigning 
relatively low de minimis values to these pollutants.
    For the proposed rule, a ``cap'' of 0.01 tons per year was used. 
This value represents 10 percent of the lowest value assigned based 
upon chronic toxicity (i.e., 10 percent of the value assigned to 
pollutants with a composite score greater than 40). If the value based 
upon other considerations (described above) yielded a value greater 
than 0.01 tons per year, the 0.01 tons per cap was assigned. For 
example, for mercury compounds, the health-based and default criteria 
yielded values of 0.1, 0.6, and 5 tons per year, depending on the 
specific compound involved. For each of these mercury compounds, the 
proposed rule lowers the value by assigning the 0.01 tons per year 
``cap.'' On the other hand, the value for dioxin was already well below 
0.01 tons per year, so the 0.01 tons per year ``cap'' was not the 
limiting consideration.
    Other policy approaches were considered. One approach would be to 
select an alternative ``cap'' such as 0.1 tons per year. Another 
possible approach might be to lower the de minimis values to one-tenth 
that of the default or health-based values. The EPA requests comment on 
whether special consideration is needed to account for atmospheric 
deposition to water bodies, and on other alternatives that could be 
considered.
    For POM, the EPA requests comment on the appropriate method for 
determining whether POM emissions exceed a de minimis amount. POM is a 
general term referring to a complex mixture of thousands of polycyclic 
aromatic compounds, including many diverse classes of hydrocarbons 
(e.g., polycyclic aromatic hydrocarbons or PAH), substituted aromatic 
hydrocarbons/e.g., nitrated PAH), and heterocyclic aromatic compounds 
(e.g., aza-arenes). Combustion sources using any of a variety of fuels 
are a major source of POM and routinely emit a large number of 
different POM compounds with the level and composition of POM emissions 
generally dependent on the extent of incomplete combustion. Important 
combustion sources of POM include diesel and gasoline engines, heaters, 
burners, and incinerators. Other sources include coke ovens, petroleum 
refineries, primary aluminum smelters, carbon black production, asphalt 
roofing manufacturing, hot asphalt processing plants, wood charcoal 
production, secondary lead smelting and ferroalloy production.
    Because there is no widely accepted method for measuring or 
assessing risks from all POM emissions, the EPA is soliciting comment 
on a preferred approach for determining POM emissions for the purposes 
of today's proposed rule. Various approaches have been used in past 
studies wherein a single POM, such as benzo-a-pyrene (B(a)P), or the 
sum of representative or particularly toxic PAH species, have been used 
as surrogates for POM. [ref: Cancer Risk from Outdoor Exposure to Air 
Toxics, Volume II, EPA-450/1-90-004b; Roussel, et al., Atmospheric 
Polycylic Aromatic Hydrocarbons at a Point Source of Emissions, J. Air 
Waste Manag. Assoc. 42:1609-1613; Assessing Multiple Pollutant Multiple 
Source Cancer Risks from Urban Toxics, EPA-450/2-89-010.] 
Alternatively, the EPA's Office of Research and Development has been 
developing an approach using the extractable organic matter (EOM) 
content of particulate matter as an appropriate measure of complex POM 
mixtures. The EOM is believed to contain the PAH and substituted-PAH 
compounds that predict cancer risk better than any individual PAH or 
any sum of PAH species. [Lewtas, Environmental Health Perspective, 100: 
211-218 (1993)]
    All of these methods for estimating POM emissions in the context of 
this proposal contain some inherent advantages and disadvantages. Using 
B(a)P alone is not thought to represent adequately either the total 
mass of POM emissions or the related cancer risks. However, a 
reasonable data base exists for determining B(a)P emission from a wide 
variety of sources. Using a ``sum-of-individual PAH species'' approach, 
while perhaps better than using B(a)P alone, may still not represent 
adequately the cancer-related risks from some sources that emit 
significant levels of substituted-PAH compounds. However, a reasonably 
extensive, data base exists for speciated PAH emissions. Unfortunately, 
there is little consistency as to what particular PAH compounds have 
traditionally been measured (L&E for POM]) and moreover, the widely 
varying toxicities of various PAH compounds further complicates the 
determination of a single POM de minimis level based on the sum of PAH 
species. Regarding the use of EOM as a measure of POM, the EPA is 
evolving a data base of EOM emissions from a variety of sources and has 
evaluated the toxicity of a number of EOM mixtures (Lewtas). It may be 
possible to list differing toxicity-weighted de minimis emission rates 
for EOM for a number of combustion and industrial categories of 
sources. This approach may be the most consistent with evolving the EPA 
health evaluations of POM mixtures. This approach, however, would 
require applicants to calculate an expected EOM emission rate which 
would be compared to the de minimis value(s). Because test data for EOM 
may not be as widely available as for B(a)P, it may be more difficult 
for some applicants to make these calculations.
    The EPA requests comment on this issue not only in terms of the 
section 112(g) program, but also in terms of the appropriate treatment 
of POM in other section 112 programs such as the urban area source 
effort under section 112(k) and the specific pollutant program under 
section 112(c)(6).
    5. State Option for Case-by-Case Dispersion Calculations. The 
definition of de minimis in section 63.41 of the proposed rule allows a 
State the option of establishing de minimis values on a case-by-case 
basis. Such case-by-case values, established according to subparagraph 
(2) of the definition of de minimis in Sec. 63.41 of the proposed rule, 
would supersede any de minimis values contained in the table in 
Sec. 63.44 of the proposed rule. States wishing to use the option are 
required to obtain approval from the EPA at the time EPA reviews a 
submittal from the State for delegation of authority under section 
112(l) of the Act for implementation of the section 112(g) program.
    The EPA has developed draft guidelines for use in this delegation 
review process. These draft guidelines are contained in the docket to 
the proposed rule. The guidelines contain procedures for identifying 
air quality benchmark concentrations and procedures for dispersion 
calculations for use in identifying case-by-case de minimis values. In 
no case would EPA approve a benchmark (e.g., risk-specific doses 
associated with one-per-million, RfCs) less stringent than the EPA's. 
State programs that use benchmarks at least as stringent as those 
contained in the guidelines would be approved.
    Further, any case-by-case de minimis values developed pursuant to a 
delegated program may not exceed 10 tons per year. As stated 
previously, the EPA believes that a ``cap'' of 10 tons per year is 
suggested by the major source cutoff in section 112(a) of the Act.
    6. Other Alternatives Considered. Other alternatives for setting de 
minimis values were considered. One suggested alternative would presume 
that the 10 tons per year major source cutoff is adequate for purposes 
of establishing de minimis values. The EPA believes that many 
pollutants on the list would present a substantial health concern at 
values considerably less than 10 tons per year (depending on exposure 
scenarios), and that use of the 10 tons per year criterion would not 
adequately provide for a level that could be considered ``trivial.''
    A second alternative would calculate de minimis emission rates for 
each pollutant based upon a selected fraction of the major source 
cutoff. For example, under this approach the EPA might establish 1 
percent or 10 percent of the 10 tons/yr major source cutoff as a de 
minimis value. The EPA favors the approach in the proposed rule, 
because it explicitly takes into account differences in toxicity.
    The EPA also requests comment on whether the de minimis values 
listed in Sec. 63.44 should be rounded to one significant figure, as is 
the case in the proposed rule, or ``binned'' into order-of-magnitude 
groupings (for example, assign values of 0.01, 0.1, 1, 10, etc.).
    The EPA also requests comment on whether pollutants should be 
considered as less than de minimis if they are present in less than a 
specified concentration in a product or mixture. For example, for 
reporting to the toxic release inventory (TRI) required under section 
313 of the Emergency Planning and Community Right to Know Act (EPCRA), 
chemicals do not need to be reported if they are present at 
concentrations less than 1 percent (0.1 percent for carcinogens). The 
proposed rule does not provide for such an exemption. The EPA is 
concerned that pollutants present as small percentages of the total 
emissions could be of substantial concern in evaluating whether a 
modification should be required to install emission controls. For 
example, pollutants such as dioxin and hexavalent chromium are 
hazardous at such small concentrations that emission increases present 
at quantities well below 0.1 percent in a stream may be of concern. In 
addition, the EPA believes that the 10 tons per year de minimis values 
for many commonly-used HAP will ensure that in most cases there will 
not be an unreasonable analytical burden on solvent mixtures. For such 
mixtures, in most cases the EPA believes that 10 tons per year will not 
involve undetectable quantities of the HAP. The EPA has, however, not 
collected data to confirm this judgment and comment is requested.
    7. Pollutant Mixtures. For the proposed rule, pollutants are 
evaluated individually with respect to the de minimis values in 
Sec. 63.44. For pollutant mixtures, the emission increase for each 
pollutant in the stream is compared individually to the de minimis 
value for that pollutant. There may be situations for which no one 
pollutant exceeds the de minimis rate, but several pollutants are 
approaching the de minimis rate. The EPA requests comment on an 
alternative that would create a de minimis ``index'' for which the 
contributions toward de minimis are treated as additive. The EPA 
believes that this alternative would increase the complexity of the 
program and would probably not greatly affect the scope of the program.
    8. Updates to De Minimis Table. The EPA intends to provide periodic 
updates to the de minimis table contained in Sec. 63.44. Such updates 
will be appropriate when the health data used as the basis for the 
tables are revised, or if new health studies become available for 
pollutants with ``default'' values in the table.
    9. De Minimis Values for Radionuclides. One of the 189 listed HAP, 
for which de minimis values must be established, is ``radionuclides.'' 
This grouping comprises a large number of different radionuclides. For 
today's proposed rule, the EPA relies on previous efforts to evaluate 
cancer risks from radionuclide exposures. The definition of de minimis 
for radionuclides is listed as paragraph 3 of the definition.
    For radionuclides, the EPA believes that an effective dose 
equivalent of 0.3 millirem per year for a 7 year exposure period would 
result in a cancer risk consistent with the one-per-million criterion 
used for other nonthreshold pollutants on the HAP list. Accordingly, 
this 0.3 millirem level serves as the basis for a de minimis 
evaluation. Techniques for evaluating the level of radionuclide 
emissions that would result in a 0.3 millirem dose are contained in 
subpart B and I, and Appendix E of 40 CFR part 61. These techniques are 
available for a large number of individual radionuclides, including 
those that would be expected to be emitted by major sources of HAP 
subject to section 112(g) of the Act. The EPA requests comment on the 
proposed de minimis definition for radionuclides, including comment on 
the types of emitting sources that may exceed such de minimis values.

F. Section 63.45. MACT Determinations

    As discussed previously, Secs. 63.42 and 63.43 require permitting 
authorities to make MACT determinations for an owner or operator who 
constructs, reconstructs, or modifies a major source of HAP. This 
section of the preamble discusses the EPA's proposed procedures for 
making these MACT determinations. These procedures include technical 
review procedures needed to establish a MACT emission limitation and a 
corresponding MACT control technology, and, (where appropriate), 
administrative procedures for submitting and reviewing applications for 
MACT determinations. In the proposed rule, the overall process for MACT 
determinations is outlined in Sec. 63.45. In addition to the proposed 
rule, EPA is making available for public comment a draft document 
entitled Guidelines for MACT Determinations under Section 112(g) (EPA-
450/3-92-007b). This document contains more details on the procedures 
and examples illustrating how they could be implemented.
    1. Overall Process for MACT Determinations. The overall process for 
MACT determinations contained in Sec. 63.45 of the proposed rule is 
shown in Figure 7. The primary emphasis, for the MACT requirements in 
Sec. 63.45 of the proposed rule and in the MACT guidelines, is on the 
procedures for case-by-case MACT determinations when no applicable MACT 
standard has been promulgated by the EPA. The procedures for 
determinations after MACT standards have been promulgated are more 
straightforward.
BILLING CODE 6560-50-P

TP01AP94.006


BILLING CODE 6560-50-C
    When a MACT standard has been promulgated for a given category, 
section 112(g) does not require a case-by-case determination of a MACT 
emission limitation. Rather, section 112(g) requires that ``the 
Administrator (or the State) determine'' that a MACT emission 
limitation will be met. For existing equipment being modified, there 
may be some such modifications for which no change in control 
technology will be required to meet the MACT emission limitation, even 
though the emissions may increase above section 112(g) de minimis 
levels. For such modifications, the EPA believes that a notification to 
the permitting authority prior to operation should suffice for purposes 
of the ``determination.'' For the proposed rule, such a notification is 
required before operation of the modified equipment. The EPA requests 
comment on whether this notification should be required prior to 
startup, for example, 30 or 60 days before startup. Requirements for 
what such a notification should contain are included as Sec. 63.45(f) 
of the proposed rule.
    In other cases where a MACT standard has been promulgated, the MACT 
standard itself will contain administrative procedures for 
modifications affecting the level of control. For example, in the 
proposed standard for synthetic organic chemical plants (57 FR 62608-
62808, December 31, 1992), there are administrative procedures 
dictating the review when ``Type 2'' equipment (requiring a lesser 
degree of control) is modified to become ``Type 1'' equipment 
(requiring a greater degree of control). Where such administrative 
provisions exist in the standard, the EPA believes that such provisions 
would suffice for the ``determination'' requirements for any such 
changes that would be considered ``modifications'' under section 
112(g).
    Where no MACT standard has been promulgated, section 112(g) 
requires a case-by-case determination of the MACT emission limitation. 
The EPA believes that the ``determination'' could take two broad 
options: (1) A revision to a Part 70 permit, or (2) a ``Notice of MACT 
Approval.'' These two options are described in Sec. 63.45(c) of the 
proposed rule. Under either approach, the process for review would be 
conceptually similar.
    The process begins with a MACT analysis by the owner and operator. 
This MACT analysis must be consistent with the Guidelines for MACT 
Determinations, including general principles described in paragraph 
63.45(d). The owner or operator provides an application for a MACT 
determination to the permitting authority. Requirements for the 
contents of this application are listed in paragraph 63.45(e). (The EPA 
wishes to clarify that the requirement in Sec. 63.45(e)(2)(vi) to list 
emission rates is intended as background information to enable the 
permitting authority to identify the pollutants requiring MACT 
controls. The EPA recognizes that there is often a significant effort 
required to obtain precise estimates of HAP emission rates and 
specifications. The EPA does not intend in this paragraph to require a 
greater level of detail than is necessary for evaluating applicability 
and emission control issues). This application for a MACT determination 
is then reviewed by the permitting authority according to either: (1) 
The administrative procedures outlined in 40 CFR part 70, where this 
option is selected, or (2) the administrative procedures described in 
paragraph 63.45(g). If approvable, the permitting authority would 
either: (1) Revise the Part 70 permit, or (2) issue a Notice of MACT 
Approval. In either case, the owner or operator would be required to 
comply with requirements described in paragraph 63.45(h). Provisions 
dealing with compliance with the requirements of the Notice of MACT 
Approval are described in paragraphs 63.45(j), (k), (l) and (m).
    Where EPA determines that the MACT determination made by the 
permitting authority fails to meet any of the requirements of paragraph 
63.45, EPA may take one of two actions to address the deficient MACT 
determination.
    (a) Where the MACT determination is made part of a source's part 70 
permit, EPA may veto issuance of the permit in accordance with the 
provisions of 40 CFR 70.8(c). The EPA may also use the veto process 
outlined in 40 CFR 70.8(c) where the State has ``enhanced'' its section 
112(g) process to incorporate the part 70 procedures.
    (b) Where the MACT determination is made through a Notice of MACT 
Approval before the source obtains or revises its part 70 permit, EPA 
may exercise the authority authorized under section 113(a)(5) of the 
Act to prohibit construction or modification, issue an administrative 
penalty order or bring a civil action against the source upon finding 
that the State has not acted in compliance with any requirement or 
prohibition relating to the construction of new sources or the 
modification of existing sources.
    2. Requirement for Preconstruction Determination. Section 63.45 
requires the MACT determination before construction, reconstruction or 
modification of the major source. The requirement is based upon the 
language in section 112(g)(2) (A) and (B) requiring that the 
Administrator (or the State) determine that MACT ``will be met.'' The 
EPA believes that the future tense suggests an up-front determination.
    Commentors to the EPA have suggested that the future tense does not 
suggest a preconstruction review. These comments assert that the phrase 
``unless'' the Administrator (or the State) determines that MACT will 
be met does not impose the same requirement that would be imposed if 
the language were to read ``until'' the Administrator (or the State) 
determines that MACT will be met. Moreover, these commentors suggest 
that Congress intended to avoid preconstruction reviews for 
modifications, and that this is the reason for the language in section 
112(g)(3) requiring ``reasonable procedures'' for assuring that 
modification requirements are reflected in the major source's operating 
permit. Although the EPA currently believes that a requirement for 
preconstruction review reflects the better reading of the Act, the EPA 
requests comment on the alternative suggested by these commentors.
    Specifically, the EPA requests comment on an alternative approach 
that would incorporate a similar administrative process to the proposed 
rule, except that the review, and the associated terms and conditions, 
would have to be completed prior to commencement of operation, rather 
than construction. Under such an approach, the source owner would be 
allowed to construct at its own risk pending the outcome of the review. 
If the permitting authority during its review were to determine that 
the increase would fail to meet the control technology requirements of 
this rule, the source would be liable for violating the requirement to 
apply case-by-case MACT and would not be allowed to operate the 
equipment until MACT was installed. The risk of violating the MACT 
requirement would fall entirely on the source making the election to 
bypass the pre-approval process.
    The EPA believes that the risk of such a retrofit would create an 
incentive for the source to ensure that the selected control technology 
in fact complies with the requirements of this rule. Moreover, sources 
may adopt an even greater level of control under this approach in order 
to ensure that its MACT demonstration complies with the requirements of 
this rule. Under this circumstance, it is possible that there could be 
greater emission reductions than would be legally required, as well as 
the economic benefits of providing a process that allows source to 
avoid the delays associated with a pre-construction approval process.
    The EPA believes that there are substantial implementation 
disadvantages for any program that would allow equipment to be 
constructed before a determination is made. The EPA's past experience 
in enforcing air quality regulations suggests strongly that it would be 
very difficult to require substantial changes in the design of 
equipment once it is in place. The EPA feels that fairness or equity 
arguments, based on investments already made and the costs of retrofit 
and shutdown, could be made by a source seeking to begin operation 
under these circumstances. Under the alternative approach described 
above, the EPA believes that such arguments are not valid. The EPA 
requests comment on the practical viability of preventing operation 
under this approach.
    The EPA is sensitive to the concern that the program should not 
lead to unreasonable delays for small changes to equipment. The EPA 
believes that the treatment of ``physical change'' and ``actual 
emissions'' in Sec. 63.43, and the provision for de minimis values, as 
well as the opportunity to offset emission increases, should ensure 
that very small changes in operations, particularly those changes 
within existing allowable operating scenarios do not require a review. 
In addition, Sec. 63.45 contains streamlined administrative procedures 
which should ensure that the preconstruction review is completed 
quickly. The EPA recognizes that with a requirement for a 
``preconstruction'' review there remains some ambiguity with respect to 
activities that are prohibited without a review. For some rules, the 
term ``commence'' construction is used. This term prohibits an owner 
from entering into binding contracts prior to the review. 
Alternatively, the term ``begin actual construction'' has been used to 
describe the actual on-site fabrication of equipment. For the proposed 
rule, language very similar to a ``begin actual construction'' 
definition is used. As proposed, the owner and operator must obtain the 
determination before they are allowed to ``fabricate (on-site), erect, 
or install'' regulated emission points. The EPA requests comment on 
this issue.
    3. General Principles for MACT Determinations. Paragraph 63.45(d) 
reviews a number of general principles that would govern MACT 
determinations under the proposed rule. When a MACT standard has been 
promulgated, the control technology selected by the owner or operator 
must be capable of achieving the emission standards and requirements of 
the standard. When a MACT standard has not been promulgated, a case-by-
case MACT determination is needed.
    In general, the purpose of a case-by-case MACT determination is to 
develop technology-based limitations for HAP emissions that the 
Administrator (or a permitting agency to whom authority has been 
delegated) approves as equivalent to the emission limitations that 
would be required for the source category if promulgated MACT standards 
were in effect under section 112(d) or section 112(h) of the Act.
    When no MACT standard has been promulgated, today's proposed rule 
requires a case-by-case determination by the permitting authority that 
the technology selected by the owner or operator is consistent with 
what would have been required under section 112(d) of the Act.
    Section 112(d)(2) of the Act describes the general considerations 
for a MACT determination. A MACT level of control is ``the maximum 
degree of reduction in emissions of the hazardous air pollutants that 
the Administrator, taking into consideration the cost of achieving such 
emission reduction, and any non-air quality health and environmental 
impacts and energy requirements, determines is achievable for new and 
existing sources in the category or subcategory * * *'' This paragraph 
of the Act continues to describe a number of items that might be 
considered in designing MACT standards such as material substitutions, 
enclosure of processes, capture and control of emissions, design and 
work practice standards, and operational standards. In the proposed 
rule, this list of items is included in the definition of ``control 
technology'' in Sec. 63.41 of the proposed rule.
    Section 112(d) also imposes certain minimum requirements on the 
determination of ``maximum achievable control technology.'' 
Collectively, these minimum requirements are defined in the proposed 
rule as the ``MACT floor.''
    For constructed and reconstructed major sources, the MACT floor for 
a case-by-case MACT determination, consistent with section 112(d), is 
the level of control that is achieved in practice by the best 
controlled similar source. The definition of MACT floor for new source 
MACT in the proposed rule does not require consideration of sources 
outside the United States; the EPA requests comment on this issue.
    For existing sources, the MACT floor for the case-by-case 
determination, consistent with section 112(d) of the Act, is an 
emission limitation equal to the average emission limitation achieved 
by the best performing 12 percent of existing sources for categories or 
subcategories with 30 or more sources, or the average emission 
limitation achieved by the best five sources for categories with fewer 
than 30 sources. The MACT floor for existing sources also takes into 
account sources achieving the ``lowest achievable emission rate'' as 
defined for the criteria pollutant new source review program under 
section 171 of the Act. The EPA interprets the ``best performing 12 
percent'' to mean the best performing 12 percent of sources in the 
United States. The phrase ``in the United States'' is added to the 
existing source MACT floor definition in order to clarify that 
territories and possessions of the United States are included.
    In rules currently under development, the EPA is considering two 
interpretations of the statutory language concerning the MACT floor for 
existing sources. One interpretation groups the words ``average 
emission limitation achieved by'' the best performing 12 percent. This 
interpretation places the emphasis on ``average.'' It would correspond 
to first identifying the best performing 12 percent of the existing 
sources, then determining the average emission limitation achieved by 
these sources as a group. Another interpretation groups the words 
``average emission limitation'' into a single phrase and asks what 
``average emission limitation'' is ``achieved by'' all members of the 
best performing 12 percent. In this case, the ``average emission 
limitation'' might be interpreted as the average reduction across the 
HAP emitted by an emission point over time. Under this interpretation, 
the EPA would look at the average emission limits achieved by each of 
the best performing 12 percent of existing sources, and take the 
lowest. This interpretation would correspond to the level of control 
achieved by the source at the 88th percentile if all sources were 
ranked from the most controlled (100th percentile) to the least 
controlled (1st percentile).
    The EPA is proposing to adopt the first interpretation and solicits 
comment on its interpretation of ``the average emission limitation 
achieved by the best performing 12 percent of existing sources'' 
(section 112(d)(3)(A) of the Act). The draft MACT Guidelines (EPA-450/
3-92-007b), reflect the first interpretation. The EPA is also 
soliciting comment on these two interpretations in a separate Federal 
Register notice, which is a reopening of the comment period for the 
hazardous organic national emission standard. Persons wanting to 
comment on this issue are asked to submit comments to docket A-90-19. 
However, comments specific to this issue as it relates to section 
112(g) should be submitted to the section 112(g) docket.
    The EPA recognizes that when information is available to define a 
MACT floor, the Act clearly requires that the case-by-case MACT 
determination must take that information into account. The EPA is 
working to develop data bases and other approaches which could 
facilitate transfer of information on available technologies and 
calculations of the MACT floor. In the proposed rule, Sec. 63.45(c)(3) 
requires that the owner or operator must determine whether a MACT floor 
exists, based upon the available information. The MACT Guidelines (EPA-
450/3-92-007b) contain several methods that could be used to provide 
the documentation for a MACT floor calculation. The EPA requests 
comment on the treatment of the MACT floor in the proposed rule.
    A key element in the floor finding is to review the ``available 
information.'' In some instances, such information sources are readily 
apparent. For example, if a Federal MACT standard has been proposed, 
but not yet promulgated, the EPA expects that a MACT floor 
determination will strongly consider that proposal. (Other information 
may be available in some cases, for example, based upon public comment 
on the MACT proposal, but such data would need to be adequate to refute 
the floor finding in the proposal). In other cases, the EPA will have 
generated background documents summarizing MACT floor findings which 
should be readily available.
    In addition to these background documents, the EPA currently 
maintains a number of data bases that may be useful as a resource for 
information on available control technologies or to obtain data to 
calculate the MACT floor. These data bases include the National Air 
Toxics Information Clearinghouse (NATICH), the Best Available Control 
Technology/Lowest Achievable Emission Rate (BACT/LAER) Clearinghouse, 
and the Aerometric Information Retrieval System (AIRS)/AIRS Facility 
Subsystem (AFL). These data systems are included within the definition 
of ``available information'' as sources of data to explore in 
determining whether a MACT floor exists.
    The EPA is also designing a data management system to support case-
by-case MACT determinations. The data base under development for this 
purpose is called the MACT data base. The EPA is intending to use AIRS/
AFL for States to store and retrieve information in the data base. The 
EPA is making changes to AIRS/AFL for the MACT Data Base to better meet 
the needs of States, industry and other interested users in accordance 
with comments and input received over the last 18 months of its 
development. The EPA is developing guidance documents on how to use the 
MACT Data Base and how its correct use assists States in determining 
MACT on a case-by-case basis.
    Under the current design plans, EPA will make information available 
on source categories currently under study for the development of MACT 
standards pursuant to section 112(d). States will be asked to submit 
source category-specific data to the data management system in 
accordance with a pre-determined schedule for the remaining source 
categories. This schedule will be coordinated with the Draft Schedule 
for the Promulgation of Emission Standards, (57 FR 44147). In addition 
to this data collection effort, EPA intends to require States to report 
all case-by-case MACT determinations that are made to the MACT Data 
base. This overall approach avoids repetitive data collection efforts, 
and provides States, industry and environmental groups access to 
information on section 112 major sources and pollutants in order to 
develop case-by-case MACT determinations that are consistent on a 
nationwide basis. The EPA requests comments on the design and use of 
this data base for case-by-case MACT determinations. In particular, EPA 
requests comments from users on whether other existing data bases, such 
as a modified BACT/LAER Information System (BLIS), might be a better 
repository for part or all of the information collected.
    Several comments received indicated that both States and industry 
would favor the MACT Data Base as the sole source of available 
information for making a MACT floor finding, if a proposed standard and 
background information document are not available. The EPA is 
requesting comment on whether use of this data base alone would 
constitute a sufficient effort for making a MACT floor finding. While 
the EPA agrees that a centralized location for available information 
for MACT floor determinations is highly desirable, the EPA has some 
concerns with absolute reliance on such a system as the sole source of 
MACT floor information. First, it is likely that in some instances 
there may be readily available industry or EPA-supported studies which 
may provide useful information with respect to the application of given 
technologies. In addition, reliance on the data base may suggest the 
need for a mandatory reporting requirement that States submit source-
category-specific information for certain source categories. The EPA 
requests comment, particularly from those supporting reliance on the 
MACT data base as the sole source of MACT floor information, on the 
feasibility and need for such a mandatory reporting requirement on the 
part of States.
    The EPA requests comment on the cutoff date that should be 
incorporated into the definition of ``available information.'' For the 
proposed rule, information is considered to be ``available'' if it is 
available as of the permitting authority's final determination, i.e., 
the date the permitting authority makes the final determination after 
receiving all comments. The EPA requests comment on other alternatives 
including: (1) The date of a complete application, (2) the date of a 
preliminary determination, and (3) the deadline for comments from the 
public and the EPA.
    When a MACT floor exists based upon the ``available information,'' 
the proposed rule requires that the control technology selected by the 
owner or operator achieve an equal or greater level of control than 
that MACT floor. The owner or operator should consider, in determining 
whether to select a control technology achieving a level of control 
greater than the floor, the cost, non-air quality health and 
environmental impacts and energy requirements of achieving that level 
of control. (See section 112(d)(2) of the Act.)
    When a MACT floor cannot be determined, the proposed rule requires 
a maximum degree of reduction in emissions with consideration to the 
cost, non-air quality health and environmental impacts and energy 
requirements. The MACT Guidelines discuss procedures for establishing a 
case-by-case MACT emission limitation under these circumstances. These 
procedures are conceptually similar to the procedures for establishing 
BACT requirements under criteria pollutant permitting programs.
    4. General Issues with Regard to the MACT Floor Determinations. For 
both ``new source MACT'' and ``existing source MACT,'' there are 
general issues for which the EPA requests public comment. For ``new 
source MACT'' the EPA requests comment on the criteria for identifying 
the ``best controlled similar source.'' For ``existing source MACT,'' 
the EPA requests comment on the degree of subcategorization which 
should be permitted in determining the level of control associated with 
the best performing 12 percent of sources.
    For constructed and reconstructed major sources, section 112(g) of 
the Act requires an emission limitation consistent with a ``new source 
MACT'' level of control. The Act states, ``the maximum degree of 
reduction that is deemed achievable for new sources in a category or 
subcategory shall not be less stringent than the emission control that 
is achieved in practice by the best controlled similar source, as 
defined by the Administrator.'' The Act does not specifically define 
the term ``best controlled similar source.'' In addition, unlike for 
existing sources for which the Act states, ``the average emission 
limitation achieved by the best performing 12 percent of the existing 
sources * * * in the category or subcategory for categories or 
subcategories with 30 or more sources,'' the Act does not specifically 
indicate that the determination of the best controlled similar source 
should be limited to from within the source category.
    For the purposes of section 112(g), EPA is proposing to use two 
criteria to determine if a source is similar: (1) Whether the two 
sources have similar emission types, and (2) whether the sources can be 
controlled with the same type of control technology. The EPA has 
developed a draft emission classification system to help determine 
emission types for case-by-case MACT determinations. When comparing 
emission types under this classification scheme, consideration should 
be given to the concentration and constituents of a gas stream. The 
five types within this draft system are: (1) Process vent or stack 
discharges, (2) equipment leaks, (3) evaporation and breathing losses, 
(4) transfer losses, and (5) operational losses. The draft guidance 
document MACT Determinations under Section 112(g) provides more 
detailed descriptions of each of these emission types. EPA request 
comments on the usefulness of this classification scheme in 
distinguishing similar sources.
    The EPA believes that because the Act specifically indicates that 
existing source MACT should be determined from within the source 
category and does not make this distinction for new source MACT that 
Congress intends for transfer technologies to be considered when 
establishing the minimum criteria for new sources. EPA believes that 
the use of the word ``similar'' provides support for this 
interpretation. The EPA believes that Congress could have explicitly 
restricted the minimum level of control for new sources, but did not. 
The use of the term ``best controlled similar source'' rather than 
``best controlled source within the source category'' suggests that the 
intent is to require a consideration of transfer technologies when 
appropriate.
    The EPA believes that there will be cases when such technology 
transfers are entirely reasonable. For example, suppose that the best 
controlled tank within a source category did not have state-of-the-art 
controls. Yet, tanks from outside the source category storing similar 
organic liquids use state-of-the-art controls vented to an emission 
control device. EPA believes that such tanks are clearly ``similar'' 
within the language of Section 112(d). The EPA also believes that the 
Act does not compel such technology transfers in all cases, and that 
emission types and the ability to install such controls are strong 
factors in determining when sources should be considered similar. For 
example, within source category X, spray booths tend to be uncontrolled 
due to gas streams with low concentrations and relatively high 
airflows. The EPA does not believe that controls from another category 
should be considered in determining the best controlled similar source 
where emissions from spray booths are of high concentration and low 
airflow. The emissions from these sources are clearly not similar. 
However, if it is technologically feasible, these same controls could 
be considered in establishing the new source level of control if 
consideration is given to cost, non-air quality health and 
environmental impacts and energy requirements. The EPA requests comment 
on language that could serve to clarify the meaning of ``similar'' for 
cases involving technology transfer from other source categories.
    A general problem that must be addressed, in determining the MACT 
``floor'' for existing sources, is the identification of the universe 
of equipment that must be considered in establishing that floor. The 
list of source categories established in section 112(c)(1) of the Act 
(see 57 FR 31579) provides some guidance in this regard. The EPA 
believes, however that additional guidance may be needed regarding the 
degree of subcategorization of these categories that might be 
appropriate for case-by-case evaluations.
    When the notice of initial list of categories of sources under 
section 112(c)(1) of the Act was published (57 FR 31576), the EPA 
listed broad categories of major and area sources rather than narrowly 
defined categories. The EPA chose to establish broad source categories 
at the time the source category list was developed because there was 
too little information to identify technically distinct groupings 
within these broad categories. During the standard-setting process, EPA 
may find it appropriate to further subcategorize to distinguish among 
classes, types and sizes of sources.
    This lack of subcategorization may pose some difficulty to owners 
and operators, and to reviewing agencies, in establishing a case-by-
case emission limitation. The source category list contains categories 
that will regulate more than one process type. Within these different 
processes different types of materials may be used or similar emission 
units may be used for different applications. In both instances, the 
emission potential of the emission unit may vary. For example, there 
are several different methods for applications of inks within the 
printing and publishing (surface coating) source category. One method 
may have an inherently lower potential to emit hazardous air 
pollutants, or it may not emit hazardous air pollutants at all. When 
the EPA develops a MACT standard for this industry, after gathering 
information on the source category, a decision will be needed on 
whether it is appropriate to: (1) Consider all process and emission 
units as one source when determining the MACT floor level of control 
or, (2) subcategorize the category in technically distinct groupings. 
The EPA believes that similar exercises may be needed in setting a 
``MACT floor'' for case-by-case MACT determinations.
    The EPA has discussed several possible approaches to dealing with 
subcategorization for the purposes of case-by-case MACT determinations. 
These discussions involved environmental interest groups, industry, and 
State agencies through meetings with the Clean Air Act Advisory 
Committee. Under the first approach, EPA would further subcategorize 
the source category list into subcategories for the purposes of case-
by-case MACT determinations. While this option may provide for the 
greatest consistency in MACT determinations from all permitting 
authorities, this approach may not be feasible. As noted, the EPA 
lacked information to properly characterize each source category at the 
time the source category list was developed under section 112(c)(1). 
Doing so for implementation of section 112(g) would be expensive and 
time consuming.
    Under a second option, EPA would allow the applicant to submit a 
suggested method for subcategorizing the source category. The EPA would 
review the proposed subcategorization scheme. If the method is an 
acceptable method for the purposes of case-by-case MACT determinations, 
a notice would be issued in the Federal Register. This option has merit 
in that members of a source category would be given the opportunity to 
demonstrate, using their knowledge about the differences in emission 
units, that technically distinct subcategories exist within the 
category. On the other hand, having EPA issue a Federal Register notice 
may make this option too time consuming to be practical.
    As a third option, the EPA could also delegate authority for 
subcategorization of categories to individual permitting authorities, 
and each permitting authority could address the problem on a case-by-
case basis. This would allow permitting authorities the greatest 
flexibility in case-by-case MACT determinations. However, some 
permitting authorities have indicated that reviewing agencies may not 
have the resources to address this subcategorization issue. Also, this 
option would not promote nationwide uniformity in MACT determinations.
    Finally, EPA could disallow subcategorization for the purposes of 
making a MACT floor finding for a case-by-case MACT determinations. 
This policy would be the easiest to establish and enforce. However, as 
discussed above, this could lead to inequitable MACT floor findings.
    At this time, EPA is seeking additional comments on these or other 
approaches to dealing with subcategorization of categories for case-by-
case MACT determinations. If subcategorization is allowed, EPA is also 
seeking comment on the criteria for which subcategorization would be 
allowed. Possible criteria might include technically distinct processes 
or operations (including differences between batch and continuous 
operation) fundamental differences in emission characteristics or 
control device applicability, cost differences, differences in safety 
considerations, and the appropriate consideration of opportunities for 
pollution prevention.
    5. Application for a MACT Determination.
    Paragraph 63.45(e) of the proposed rule describes the information 
the owner or operator is required to provide with an application for a 
MACT determination or in a part 70 application for which a MACT 
determination is requested. These information requirements are designed 
to identify the ``MACT-affected emission units'' and to demonstrate 
that the selected control technology for those units is consistent with 
or exceed the requirements of the statute. Further information on the 
uses of this information are described in the MACT Guidelines.
    Paragraph 63.45(f) of the proposed rule describes the notification 
requirements for an application when a MACT standard for a MACT-
affected emission unit has been promulgated. The requirements are much 
more straightforward for such cases, because the technical 
justification for a case-by-case determination is not required. This 
paragraph is designed to identify any possible situations for which the 
existing control technology in place may affect the ability of the 
emission unit to continue to comply with the promulgated standard.
    6. Review Process. Analysis of the relationship of section 112(g) 
to the operating permits program. The proposed rule, in paragraphs 
63.45(g),(h),(i) and (j) establishes an Administrative process for 
reviewing a request by an owner or operator for a MACT determination. 
As discussed previously, the EPA believes that section 112(g) of the 
Act requires such a determination to be made before constructing, 
reconstructing, or modifying a major source.
    In order for commenters to understand the structure of the proposed 
Sec. 63.45, it is necessary to discuss the EPA's reading of the Act 
regarding relationship of the MACT review process required by section 
112(g) to the operating program requirements pursuant to title V of the 
Act. The requirements for State title V permit programs, contained in 
40 CFR part 70, were published on July 21, 1992 (57 FR 32250). One 
approach to establishing an administrative process for determinations 
under section 112(g) of the Act would be to rely on the part 70 or part 
71 review process as the sole mechanism for establishing MACT 
requirements. The EPA believes that, while in some cases this may be a 
viable approach, the section 112(g) program cannot rely solely upon 
this process. First, the part 70 requirements clearly do not require a 
new green-field plant to apply for an operating permit until 1 year 
after the plant begins operation. Because the part 70 permit must be 
issued within 18 months of the application, it could be up to 30 months 
after operation before section 112(g) requirements would be 
incorporated into the permit. The EPA believes that an alternative 
federally enforceable mechanism is needed in the interim for such 
cases. Second, even for modifications at already permitted facilities, 
the part 70 requirements do not ensure that a MACT determination will 
be conducted before construction. As noted above, the EPA believes that 
section 112(g) requires a determination be made before construction. 
While in some cases, States with part 70 programs may require 
preconstruction reviews as part of the operating permit process, this 
will not always be the case. Third, there is an important time period 
for section 112(g) implementation for which the title V permit process 
is not equipped to handle section 112(g) determinations. Section 112(g) 
determinations are required at the beginning of the title V permit 
program, upon the ``effective date'' of the program. According to part 
70, sources subject to the permitting requirements are required to 
submit permit applications within 1 year of the effective date, and 
there is a 3-year period under which states can issue the initial 
permits. As a result, there is a potentially lengthy transition period 
under which the title V process is not designed to handle section 
112(g) determinations.
    The question of when a part 70 permit must be revised to reflect a 
case-by-case determination of MACT for a modification is presently the 
subject of some uncertainty. It has come to EPA's attention that 
certain provisions of the part 70 regulations can be interpreted as 
yielding conflicting results on this issue. In particular, 
Sec. 70.5(a)(1)(ii), addressing the timeliness of applications, 
provides:

    Part 70 sources required to meet the requirements under section 
112(g) of the Act, or to have a permit under the preconstruction 
review program approved into the applicable implementation plan 
under part C or D of title I of the Act, shall file a complete 
application to obtain the part 70 permit or permit revision within 
12 months after commencing operation or on such earlier date as the 
permitting authority may establish. Where an existing part 70 permit 
would prohibit such construction or change in operation, the source 
must obtain a permit revision before commencing operation.

    This provision would allow a source required to obtain a section 
112(g) case-by-case determination to apply for a part 70 permit 
revision up to 12 months after commencing operation of the 
modification, unless the change would conflict with the terms of an 
existing permit.
    Section 70.4(b)(15), on the other hand, requires that title V 
program submissions must contain:

provisions prohibiting sources from making, without a permit 
revision, changes that are not addressed or prohibited by the part 
70 permit, if such changes are subject to any requirements under 
title IV of the Act or are modifications under any provision of 
Title I of the Act.

    This provision requires that an approvable State program must 
prohibit changes that are modifications under title I from taking place 
at a permitted facility without a prior permit revision. Because 
section 112(g) case-by-case determinations are modifications under 
title I, Sec. 70.4(b)(15) could be read as prohibiting these changes 
from commencing operation prior to title V review, as authorized by 
Sec. 70.5(a)(1)(ii).
    The EPA believes that the approach set out in Sec. 70.5(a)(1)(ii) 
should govern the incorporation of section 112(g) determinations into 
title V permits. Accordingly, the EPA is proposing in today's 
rulemaking to revise the part 70 regulations to more consistently 
reflect this approach. The proposed revision to Sec. 70.4(b)(15) allows 
that section 112(g) case-by-case determinations, as well as 
modifications under part C and D of title I, need not submit a part 70 
permit application until up to 12 months after commencing operation. 
For the reasons discussed below, EPA is proposing to retain the 
Sec. 70.4(b)(15) prohibition that modifications under section 111 of 
the Act not be allowed to commence operation at a permitted facility 
until the permit is revised.
    It is a key underlying principle of title V that changes that are 
not expressly prohibited by the permit may occur without a prior permit 
revision. These changes are referred to in the part 70 rule as ``off-
permit'' changes. As stated in the preamble to the operating permits 
rule:

    The Agency continues to believe that section 502(a) allows 
certain facility changes at a permitted facility that need not be 
incorporated into the permit until renewal. Section 502(a) prohibits 
a source from operating any of certain listed types of sources 
``except in compliance with a permit * * *'' EPA's view is that it 
does not violate this prohibition for a source to operate in ways 
that are neither addressed nor prohibited by the permit.

57 Fed. Reg. 32269.

    The language of the section 502(a) prohibition does not suggest 
differential application for sources subject to review as title I 
modifications. To the contrary, section 502(a) expressly lists sources 
``required to have a permit under parts C or D of title I'' as being 
subject to the prohibition not to operate except in compliance with a 
permit. Since section 502(a) is the underpinning for the off-permit 
authorization, this language provides a textual basis for concluding 
that title I modifications may occur without a prior title V permit 
revision.
    The EPA believes it is reasonable to interpret section 502(a) such 
that title I modifications not addressed or prohibited by the permit 
can occur without prior review under title V. Changes triggering a 
title I modification by their nature merit a higher degree of review 
than other less significant source changes. However, in structuring the 
part 70 requirements for these types of changes, the EPA believes it is 
appropriate to consider not only the review requirements imposed by 
title V, but also those imposed under title I itself. The EPA 
regulations for modifications under parts C or D of title I already 
require a very substantial review of the proposed change prior to 
construction, involving public review and notice to EPA. This 
preconstruction review establishes the applicable requirements for that 
modification, and results in a Federally enforceable preconstruction 
permit incorporating the emissions limitations and other requirements 
resulting from that review.
    One possible reading of title V would allow EPA discretion to 
require that title V review occur pre-operation for all title I 
modifications at a permitted source. However, in the case of part C or 
D title I modifications, pre-operation review may entail a needless 
expenditure of administrative resources. Once pre-construction review 
has occurred, there is little if any gain to be had from imposing 
further operational requirements prior to the source's commencing 
operation. If title V review is to result in more effective regulation 
of the source's operation, for instance by the addition of compliance 
requirements provided for in Sec. 70.6, it may be more appropriate to 
conduct this review after a period of operation during which the 
effectiveness of the preconstruction permit requirements can be 
measured. The EPA believes this scheme of post-operation incorporation 
of the preconstruction permit requirements may yield a more rational 
integration of the title I and title V review requirements, and, 
depending on the characteristics of the particular State program, may 
result in additional improvements in air quality than a scheme which 
required the operating permit review to occur prior to operation. The 
EPA solicits comment on this issue.
    The reasoning set forth above regarding title I modifications under 
part C and D applies with equal force in the context of section 112(g) 
modifications. As proposed here, changes qualifying as modifications 
under section 112(g) must undergo pre-construction review, including 
public and EPA review of the proposed case-by-case determination. As 
with modifications under part C and D of title I, this process should 
obviate the need for additional review prior to operation.
    That this treatment of section 112(g) modifications is consistent 
with the intent of the Act is also evidenced by the language of section 
112(g)(3), which requires EPA (or the State) to establish ``reasonable 
procedures for assuring that the requirements for applying to 
modifications under this section are reflected in the permit.'' The 
requirement that these determinations be merely ``reflected'' in the 
permit suggests that Congress viewed pre-operation title V review as 
unnecessary. The legislative history supports this interpretation of 
section 112(g)(3). 136 Cong. Rec. S 17124-5 (October 26, 1990).
    As noted above, EPA believes this treatment of title I major 
modifications is consistent with the general structure of the Act and 
the purpose of title V to function as an operating permits program. The 
EPA has therefore reconsidered the policy position taken in the 
preamble to the operating permits rule, where it stated that ``it is 
not reasonable to allow [title I] modifications to be made outside the 
title V permit system'' 57 FR 32269.
    The EPA has also reconsidered the legal rationale stated in the 
same preamble discussion. The preamble stated that, because section 
502(b)(10) explicitly excludes title I major modifications from the 
class of changes that can be made without a permit revision, it would 
be anomalous to read section 502(a) as authorizing this same 
flexibility. As a textual matter, however, EPA has read section 
502(b)(10) to allow certain changes that contravene existing permit 
terms. It is not anomalous to strictly limit changes which contravene 
an express determination by the permitting authority by, among other 
things, excluding title I modifications, while allowing title I 
modifications to occur off-permit where determinations made in the 
title V permit remained unaltered. Moreover, the fact that title I 
modifications have already been subject to examination in the pre-
construction review process further supports the conclusion that this 
reading of the Act does not produce anomalous results.
    Although this proposal would allow both title I modifications and 
other types of source changes not addressed or prohibited by the permit 
to occur without a prior permit revision, there is an important 
difference in the treatment of these categories of source changes. For 
source changes that are not title I modifications, Sec. 70.4(b)(14) 
mandates that a State may not, except as a matter of State law, require 
off-permit activities to undergo title V review prior to operation. 
Under Sec. 70.5(a)(1)(ii), however, a State may require sources subject 
to the requirements of part C or D, or section 112(g) of title I to 
submit applications any time earlier than 12 months after commencing 
operation. This would allow a State, at its option, to require title V 
review to occur prior to operation, or even contemporaneously with the 
preconstruction review. The primary reason for this differential 
treatment is that it would allow a State to integrate its 
preconstruction and operating permit programs. The EPA recognizes that, 
depending on the particulars of a State's preconstruction and operating 
permits program, this approach to integration may be the most efficient 
from an air quality control and administrative standpoint.
    The proposed revision to part 70 should not result in any State 
having to revise its operating permits program to gain approval under 
part 70. A State program that followed the alternative reading of 
Sec. 70.4(b)(15) and required the part 70 permit to be revised prior to 
operation of a title I modification would be consistent with the 
requirements of Sec. 70.5(a)(1)(ii), and thus would also be consistent 
with this proposed regulatory revision.
    7. Implications. The EPA believes, based upon the above 
considerations, that there will be cases when the title V permit 
process will be used for section 112(g) reviews, and there will be 
cases when it will not be used and MACT determinations will be 
incorporated into the permit after commencement of operation. Section 
63.45(c) of the proposed rule states that when the title V procedures 
are used, this process would be sufficient. When the title V process 
does not occur until after construction, reconstruction, or 
modification of a major source requiring a case-by-case MACT 
determination, the proposed rule requires that the owner or operator 
follow an administrative review process contained in paragraphs 63.45 
(g), (h), (i), and (j). Where the change that is subject to section 
112(g) review is addressed or prohibited by an existing title V permit, 
the change would of course need to be processed as a revision to the 
title V operating permit prior to commencing operation.
    Regardless of the timing for incorporation of section 112(g) 
determinations into the operating permit, there are certain part 70 
requirements that apply. The title V permit must be revised or issued 
according to procedures set forth in Secs. 70.7 and 70.8. In addition, 
the permit must incorporate compliance provisions of Sec. 70.6. If, 
during the EPA's review of the section 112(g) determination, it becomes 
apparent that the determination is not in compliance with the Act, then 
EPA must object to the issuance or revision of that permit.
    These requirements are obviously satisfied either if part 70 
requires revision to an existing title V permit prior to operation, or 
if the permitting authority otherwise requires incorporation into a 
title V permit as a step in the section 112(g) determination process. 
However, even where there is no formal incorporation into a title V 
permit prior to operation, subsequent title V review may effectively be 
avoided if the State's section 112(g) process is ``enhanced'' to 
include the required title V procedures, thereby allowing for later 
incorporation into the title V permit by administrative amendment.
    Section 70.7(d) of the operating permits rule defines an 
``administrative amendment'' to include a revision that ``incorporates 
into the part 70 permit the requirements from preconstruction review 
permits authorized under an EPA-approved program, provided that such a 
program meets procedural requirements substantially equivalent to those 
contained in Secs. 70.7 and 70.8 of this part and compliance 
requirements substantially equivalent to those contained in Sec. 70.6 
of this part.'' This process of ``enhancement'' of preconstruction 
procedures was discussed in the preamble to the operating permits rule 
in the context of existing State new source review programs (see 57 FR 
32289), but was not discussed in relation to section 112(g) because the 
procedures associated with section 112(g) determinations had not been 
articulated. However, the language of Sec. 70.7(d)(v) would allow for 
use of administrative amendments for an enhanced section 112(g) 
process, and the EPA believes such use is clearly within the intent of 
that provision.
    Enhancement of the section 112(g) process may be partial only, 
incorporating some elements of the required part 70 review or 
compliance provisions in the section 112(g) process itself, with the 
remaining elements occurring during the title V process. For instance 
public review of the section 112(g) determination that meets the 
requirements of Sec. 70.7(h) need not be repeated at the time of 
incorporation into the title V permit. However, for the administrative 
amendment procedures to be available for determinations that have been 
through an enhanced process, the public, EPA and affected States must 
have had the opportunity to review all aspects of the section 112(g) 
determination, including any compliance provisions required under 
Sec. 70.6. Thus, public review during the preconstruction section 
112(g) process would not suffice for purposes of title V if the 
preconstruction process did not specify the application of compliance 
provisions substantially equivalent to those in Sec. 70.6, including 
monitoring, reporting, record-keeping, and compliance certification.
    Paragraph 63.45(f)(4) of the proposed rule clarifies that 
notification of the permitting authority is sufficient for 
modifications when no case-by-case review is needed because an 
applicable MACT standard has been promulgated and the MACT-affected 
emission unit does not require a change in control technology. When the 
modification requires a change in control technology in order to 
continue to comply with the MACT standard, the EPA believes that a 
preconstruction review is appropriate. Readers should note that section 
112(i)(1) of the Act requires a preconstruction review for major 
sources subject to new source MACT under a promulgated MACT standard.
    8. Streamlined Administrative Process. Paragraphs 63.45 (g), (h), 
(i) and (j) of the proposed rule establish an administrative review 
process for case-by-case MACT determinations. This process, patterned 
after the existing administrative process for reviewing proposed 
equipment subject to emission standards under 40 CFR part 61, is 
displayed in Figures 8 and 9. The process begins with a 30-day 
completeness determination. Once a complete application is received, 
approval or an intent to disapprove the application is required. If an 
intent to disapprove is issued, the owner or operator is given the 
opportunity to provide further information. The proposed decision to 
either approve or disapprove the application is then subject to public 
review. This proposed rule would provide for public review through 
issuance of a notice containing all the relevant background information 
about the application and allowing 45 days for the public to comment on 
whether the application should or should not be granted. In order to 
expedite approval of noncontroversial case-by-case MACT determinations 
the proposed rule would allow such determinations to go final following 
the close of the comment period if no adverse comments have been 
received. If adverse comments are received, a final notice must be 
published either approving or disapproving the application and 
addressing the comments.
    This proposal requiring public review prior to approval of case-by-
case MACT determinations is consistent with current EPA practice in 
other Clean Air Act programs where Federal enforceability is required. 
For example, 40 CFR 51.161 requires a 30 day public comment period for 
review of an agency's proposed approval or disapproval of a minor new 
source permit. Similarly, in a 1989 rulemaking EPA enumerated five 
criteria that must be met before a State issued operating permit can 
become Federally enforceable. One of those criteria is that the permit 
must be subject to public review before issuance. This criteria was 
described in the notice as being consistent with the EPA's current 
practice for construction permits codified at 40 CFR 51.161. (See 54 FR 
27283 (June 28, 1989).
    Thus, the EPA's current practice is to require public review of 
decisions required to be Federally enforceable. Without a compelling 
reason to deviate from this established practice the EPA must continue 
to follow it. As stated by the Supreme Court in Motor Vehicle Mfrs. 
Ass'n v. State Farm Mut. Automobile Insurance Co. et al., 463 U.S. 27, 
43 (1983), ``an agency changing its course is obligated to supply a 
reasoned analysis for the change beyond that which may be required when 
an agency does not act in the first instance.'' In this case there is 
an established practice of requiring public review as a prerequisite to 
federal enforceability. The EPA proposes to follow that practice in 
this case unless a compelling reason can be provided for either 
changing that practice or deviating from it in this case.
    As discussed above, the EPA is proposing to require public review 
as a prerequisite to Federal enforceability of case-by-case MACT 
determinations. Comment is specifically requested concerning whether 
public review should be a prerequisite to Federal enforceability of 
case-by-case MACT determinations, and if it should not, what 
justification there would be for deviating from established practice by 
the EPA in this area.
    The EPA recognizes that there are cases for which sources would 
prefer to minimize delays in the process, particularly for operations 
which change relatively frequently, and where the owner or operator is 
willing to control emissions from those changes with technologies that 
could be recognized as best available controls. The EPA is exploring 
suggestions that the general permit procedures, outlined in 40 CFR 
70.6(d), be available for such situations. The general permit may have 
application for section 112(g) determinations where the permitting 
authority is able to make a presumptive determination of MACT for a 
given type of source. The general permit would have to set forth the 
controls required by part 70. Once the general Figure 8: Timeline for 
S. 112(g) Administrative Process--CBC MACT permit is issued, subsequent 
application of the MACT determination at a particular source would 
involve merely a determination that the source falls within the source 
category covered by the general permit. Sources in that category may 
then apply for authorization under the general permit as the 
modifications occur.

BILLING CODE 6560-50-P

TP01AP94.007


TP01AP94.008


BILLING CODE 6560-50-C
    As discussed in the preamble to the operating permit regulations, 
general permits may be issued to cover discrete emissions units at 
permitted facilities, 57 FR 32279. While a general permit cannot be 
used to modify the terms of an existing title V permit, it may be 
issued to any change at an existing plant that otherwise would be 
eligible to apply for a new individual permit. In that case, the 
requirements of the general permit could be incorporated into the 
permit for the facility at renewal.
    The EPA also recognizes that some State programs may have well-
established administrative procedures that are similar to, but not 
identical to the administrative process in Sec. 63.45. For example, 
some States may require determinations in a slightly shorter or longer 
time period than the 60 day period in the proposed rule. The EPA 
believes that if such programs are substantially equivalent to the 
process in Sec. 63.45, and include all steps needed to ensure Federal 
enforceability, then EPA could approve such alternatives in accordance 
with section 112(l) of the Act. The EPA requests comment on this point, 
which is included in the proposed rule as paragraph 63.45(g)(7).
    9. Notice of MACT Approval. The end result of the proposed 
administrative review process is a determination set forth in a 
document that is termed a ``Notice of MACT Approval.'' Requirements for 
this Notice are provided in paragraph 63.45(h) of the proposed rule. 
This Notice is required to contain the emission limitations, 
notification, operating and maintenance, performance testing, 
monitoring, reporting, record-keeping and any other requirements needed 
to ensure that the case-by-case MACT emission limitation will be met.
    The Notice of MACT Approval serves to provide a mechanism for 
Federal enforceability of these conditions in the interim time period 
between initial operation of the constructed, reconstructed, or 
modified major source and the time the conditions are added to the part 
70 or part 71 permit.
    The EPA is considering adding a provision under which a Notice of 
MACT approval would expire if construction does not begin within a 
specified time frame. Specifically, the EPA requests comment on whether 
the Notice should expire if construction has not begun within 18 
months. Such an 18 month period is included in criteria pollutant 
preconstruction review programs.
    10. Compliance. The Notice of MACT Approval must establish 
compliance dates for MACT. For constructed and reconstructed major 
sources subject to a ``new source MACT'' level of control, compliance 
upon startup is required. For modifications, compliance upon startup is 
required unless the owner or operator demonstrates that the time needed 
to install the control technology exceeds the time needed to accomplish 
the modification.
    The EPA considered two other compliance date requirements for 
modifications. The first alternative would require compliance upon 
startup in all cases, consistent with the current approach under other 
major source modification programs such as the PSD program. The EPA 
believes that such an approach would be inconsistent with the 
provisions in section 112(i) for promulgated MACT standards, which 
allow for up to 3 years for sources subject to these standards to 
comply. The second alternative would be to allow the full 3 year period 
in all cases. The EPA believes that this approach would also be 
inconsistent with section 112(i) in that MACT standards must ensure 
compliance ``as expeditiously as practical, not to exceed three 
years.''
    Another important issue with respect to compliance is how to define 
the types of requirements that are needed for the Notice of MACT 
Approval in order to ensure that the MACT emission limitation is 
Federally enforceable. As noted above (see discussion of the definition 
of ``Federally enforceable''), the EPA believes that it is necessary 
that emission limitations include terms and conditions necessary to 
ensure that the limitation is practically enforceable.
    To ensure Federal enforceability, the proposed rule requires that 
the Notice of MACT approval contain, at a minimum, monitoring, 
recordkeeping and reporting requirements sufficient to document the 
source's compliance. Because major sources obtaining MACT 
determinations will incorporate that determination into a title V 
permit, the proposed rule includes a requirement that the monitoring, 
recordkeeping, and reporting requirements required for a case-by-case 
MACT determination be consistent with the compliance requirements 
contained in part 70.
    In addition to part 70 compliance requirements, additional 
requirements may need to be considered at the time of the MACT 
determination. Under section 114(a)(3) of the Act, the EPA is required, 
for major sources, to incorporate enhanced monitoring into all new 
rules promulgated after the 1990 Act amendments. The goal of these 
monitoring requirements is to assure that owners or operators are 
accountable for their emissions and compliance status on a continuous 
basis. In this way, the EPA is assured that the emissions reductions 
intended by regulations are in fact achieved.
    It is important to distinguish between continuous compliance and 
continuous monitoring. Under section 112 of the Act, to demonstrate 
continuous compliance, a source may not be required to record emissions 
data on a continuous, instantaneous basis such as with a continuous 
emission monitor. Depending on the type of standard, regular parameter 
monitoring, equipment inspections, and/or maintenance of raw material 
records, etc., may be sufficient to demonstrate continuous compliance. 
For all standards, monitoring frequency must be based on the averaging 
time of the applicable limitation or standard, and the likely 
variability of potential emissions from a particular emissions unit. If 
the potential variability is high, monitoring must be done frequently. 
If the potential variability is low, monitoring may be conducted less 
frequently at regular intervals.
    For examples of enhanced monitoring, source owners or operators may 
refer to EPA's Enhanced Monitoring Reference Document being developed 
in conjunction with the 40 CFR part 64 regulation. For more information 
on this document, contact Keith Brown of the EPA at (703) 308-8676. 
This document is intended to provide a constantly evolving compendium 
of monitoring systems and procedures that can be used to satisfy 
enhanced monitoring requirements.
    Where the Notice of MACT approval fails to meet any requirement of 
paragraph 63.45, EPA may exercise its authority under section 113(a)(5) 
of the 1990 Amendments to prohibit construction or modification, issue 
an administrative penalty order or bring a civil action against the 
source upon finding that the State has not acted in compliance with any 
requirement or prohibition relating to the construction of new sources 
or the modification of existing sources.
    11. Reporting to National Data Base. Section 63.45(n) requires 
permitting authorities to provide EPA with information on all MACT 
determinations. The intent of this paragraph is to use EPA's MACT data 
base to store data on well-controlled sources and on previous MACT 
determinations to help facilitate the MACT determination process.

G. Section 63.46 and 63.47--Offset Showing

    As mentioned previously in section III.D.4 of this preamble, 
section 112(g)(1)(A) of the Act requires owners and operators to submit 
a ``showing'' to the permitting authority if they choose to provide 
offsets to avoid MACT requirements. Sections 63.46 and 63.47 of the 
proposed rule provide the requirements for the offset showing. These 
sections provide two alternative approaches an owner or operator could 
use.
    1. Statutory Language. Section 112(g)(1)(A) of the Act states that 
a physical change otherwise meeting the definition of a 
``modification'' in section 112(a) shall not be considered a 
modification if the emissions ``will be offset by a decrease * * *'' 
The EPA believes that the phrase ``will be offset'' can be interpreted 
in several ways. One interpretation would be that the offsetting 
decreases must occur no earlier than the corresponding increases in 
emissions. Alternatively, the phrase could be interpreted to mean that, 
by the time the physical change or change in the method of operation 
has occurred there must have been corresponding decreases in emissions. 
Such decreases could have occurred before the increases that they are 
offsetting. The EPA believes that the statute does not mandate choosing 
one reading over the other.
2. Proposed Approaches. a. Approach under the PSD program. One overall 
approach to an offset demonstration would be to design the program to 
allow for a ``contemporaneous'' period similar to that in the PSD 
program.
    The PSD program is a stationary source permitting program that was 
established in the 1970's in reaction to general language in the Act 
directing the EPA to ensure that air pollution regulations serve to 
protect and enhance air quality. The program was born as the result of 
a major court decision, and it has been formally included in the Act 
since 1977 as part C, section 160 through 169a. The goal of the PSD 
program is to ensure that, for major sources of pollutants contributing 
to increases in ambient levels of ``criteria air pollutants'' 
(particulate matter, sulfur dioxide, nitrogen oxides, volatile organic 
compounds, etc.), that emission increases from these sources do not 
``significantly degrade'' nearby air quality.
    One aspect of the PSD program, generally referred to as the 
``netting'' procedure, is used to define which types of increases in 
emissions are subject to review. (For more information on PSD 
``netting,'' see New Source Review Workshop Manual. Prevention of 
Significant Deterioration and Non-attainment Area Permitting. Draft, 
October 1990. This document is available from the New Source Review 
section of EPA's Office of Air Quality Planning and Standards, Mail 
Drop 15, Research Triangle Park, NC 27711). The ``netting'' procedure 
involves the accounting of the overall plant-wide emission patterns for 
a given pollutant. For a given physical change, the ``netting'' process 
involves the following steps. First, the emissions increase from a 
given physical change are evaluated. If they exceed ``significant'' 
levels then they would constitute a modification. If, however, the 
emission increases from the change, when added to all other ``net 
increases and decreases'' at the plant, would not lead to a 
``significant'' increase, then the physical change is not subject to 
the BACT and air quality impacts analysis requirements of the program.
    Under the ``netting'' process, the owner and operator must account 
for all increases and decreases at the plant within the recent past. 
The PSD program uses the term ``contemporaneous'' to refer to a 5-year 
time period for which the net increases and decreases must be tallied.
    One approach to the treatment of ``offsets'' for the section 112(g) 
program would view the emission ``offsets'' under the program as 
analogous to the ``netting credits'' for the PSD program. As such, the 
program would contain a similar accounting system and would require 
establishment of a ``contemporaneous'' period. Under this approach, the 
overall goal would be to ensure that the overall hazard of plant-wide 
emissions is not increased in the interim period between the onset of 
the title V program (which triggers section 112(g) requirements) and 
the issuance of MACT standards.
    The EPA has a number of concerns with this approach. First, the 
approach presumes that a reference condition exists that can be defined 
as acceptable. For example, the PSD program is intended to preserve air 
quality levels that are considered acceptable. So long as overall 
emissions do not increase, then this acceptable air quality is not 
``degraded.'' For purposes of section 112(g), use of the 
contemporaneous period would presume, in essence, that a reference 
point in time at the beginning of the contemporaneous period 
constitutes a condition of ``acceptable'' hazardous air pollutant 
emissions. In light of the general intent of section 112 to provide for 
steady reductions in hazardous air pollutant emissions, the EPA is 
uncertain whether this approach is reasonable for offsetting in 
accordance with section 112(g).
    Another problem with the contemporaneous approach is that it 
imposes an additional administrative burden. Owners and operators would 
be required to document all increases and decreases over the 
contemporaneous period. For large and complex facilities, this might 
involve considerable documentation and recordkeeping. In addition, it 
may be more difficult to provide adequate documentation of hazardous 
air pollutant increases and decreases than it has been for pollutants 
addressed in the criteria pollutant program such as VOC and sulfur 
dioxide. While total VOC emissions may be adequately documented, it may 
be more difficult to provide for reliable estimates of the emission 
rates of individual HAP, for which a reliable accounting of speciated 
VOC totals or speciated particulate matter totals may not be available.
    b. Prospective Approach. Another approach to offset demonstrations 
was considered. Under this approach, the demonstration would not 
require an accounting of all recent increases and decreases. The offset 
demonstration, under this approach, would involve only prospective 
reductions and would require accounting of: (1) Increases from the 
emission points for the proposed change and (2) decreases from one 
emission point, or a few emission points which could achieve ``extra'' 
reductions that would were not otherwise required.
    This approach would make it less likely that emission reductions 
used as offsets would represent ``windfalls'' from emission reduction 
activities that would likely have occurred even in the absence of the 
MACT requirement. As a result, this approach would likely increase the 
number of modifications under the program, increase the number of MACT 
determinations, and possibly lead to greater overall emission 
reductions.
    This approach, however, does have potentially serious 
disadvantages. If offsetting reductions occurring before the increase 
were absolutely prohibited, then owners and operators would be 
discouraged from providing early reductions in HAP emissions or from 
completing pollution prevention projects. Where feasible, there would 
be an incentive to preserve emission reduction credits until the time 
of the increase.
    c. Proposed Approach. In the proposed rule, the EPA allows for both 
types of offset demonstrations. The owner or operator of a modification 
who is seeking offsets would have the option of providing either type 
of demonstration.
    Section 63.46 allows the owner and owner and operator to provide 
for a ``contemporaneous'' demonstration resembling a ``netting'' 
demonstration under the PSD program. Section 63.47 allows for companies 
to opt for a more simplified demonstration in cases where additional 
control measures can be undertaken.
    The contemporaneous offset procedure borrows a number of terms from 
the PSD program. The proposed rule retains the PSD definitions of what 
increases and decreases are ``creditable.'' Emission decreases must be 
Federally enforceable prior to the change being offset. Emission 
changes that were taken into account in issuing another permit under 
the section 112(g) program cannot be ``double-counted'' in a subsequent 
permit action. In addition, emission reductions used for a MACT 
extension under the section 112(i)(5) program, which according to that 
program must be ``permanent,'' are not creditable as emission decreases 
under the proposed rule. (The proposed rule does, however, note that 
any amount exceeding the 90 or 95 percent criteria in the section 
112(i)(5) program are creditable.)
    Section 63.47 provides for a ``simplified'' offset demonstration. 
Under this approach, as described in paragraph 63.47(b), only limited 
activities would be considered creditable as decreases. The control 
measure would need to constitute a new emission reduction that would be 
accomplished after the date of the application but before the startup 
of operation of the physical change being offset. The control measure 
could not be a shutdown or curtailment. Section 63.47 does, however, 
provide for special consideration for source reduction activities. The 
EPA believes that it is important to provide incentives for such 
activities. The EPA requests comment on this ``simplified'' offset 
approach, particularly the prohibition on the use of shutdowns or 
curtailments and the special consideration for source reduction 
activities.
    The EPA requests comment on both of the approaches to offsetting 
(the ``contemporaneous'' and ``simplified'' approaches) including 
whether they are feasible alternatives, and whether one of these 
approaches ought not to be included in the rule.
3. Administrative Process for Offsets
    Section 112(g)(1)(A) of the Act requires that the owner or operator 
seeking offsets in lieu of modification must submit a ``showing'' to 
the permitting authority that the increase ``has been offset.'' 
Although this language is not entirely clear, the proposed rule 
represents the EPA's policy decision that the program should require 
review of the offset demonstration before the emission increase being 
offset commences operation. The EPA believes that this is the better 
approach because of a concern that the enforceability of the program 
would be adversely affected if the modifications could occur before an 
offset demonstration is approved.
    In addition to the approach the EPA is proposing today, the EPA is 
considering an alternative mechanism for approving offsets. The EPA is 
considering this alternative approach because of concerns which have 
been expressed about the costs of delay which industries may experience 
as a result of the offset pre-approval process. Under the alternative 
approach, a source owner or operator would reduce emissions sufficient 
to offset the planned emission increase. The source would submit its 
offset demonstration to the permitting authority at the time it begins 
operation of the equipment causing the increase. The permitting 
authority would review the offset demonstration upon submittal--that 
is, after the equipment has begun operation. This review process could 
be structured according to the administrative process in the proposed 
rule, or an alternative process established by the permit authority. If 
the permitting authority during its review were to determine that the 
offset failed to meet the offset requirements of this rule, the source 
would be liable for violating the requirement to apply case-by-case 
MACT to the equipment causing the increase and would be subject to the 
full range of enforcement activities and penalties available under the 
Act. The risk of violating the MACT requirement would fall entirely on 
the source making the election to bypass the pre-approval process.
    The EPA is mindful that this approach puts not only the source at 
risk, but potentially could result in greater (or more hazardous) 
emissions for the period during which the source operated without an 
appropriate offset. The EPA believes that the penalties faced by 
sources under the Act could create an incentive for the source to 
ensure that the offset in fact complies with the requirements of this 
rule. Moreover, the severity of the penalties under the Act are such 
that sources may adopt an even greater margin of safety under this 
approach in order to ensure that its offset demonstration complies with 
the requirements of this rule. Under this circumstance, it is possible 
that there could be greater emission reductions than would be legally 
required, as well as the economic benefits of providing a process that 
allows source to avoid the delays associated with a pre-operation 
approval process.
    Consequently, the EPA believes that this approach deserves 
consideration and seeks comment on whether such an approach would 
provide a significant benefit to industry by reducing delays and 
whether it would be likely to create a risk to human health or would 
yield human health benefits. In addition, the EPA is interested in 
comments on whether this approach would create obstacles to 
enforcement, and the nature of those obstacles, by the federal or State 
governments or citizens should it be determined that the source's 
offset was inadequate. In weighing this option, the EPA also requests 
that commenters consider other provisions of the rule, including de 
minimis levels and applicable State and local review processes, that 
also affect the speed with which a source can proceed with construction 
when it intends to rely on an offset. In addition, the EPA is also 
interested in comments with specific examples on the extent to which 
alternative post-operation offset provisions might be appropriate in 
those cases where no additional new construction would be involved.
    Sections 63.46 and 63.47 establish an administrative process for 
the offset demonstration. The processes are virtually identical. The 
process is very similar to the administrative process described above 
for MACT demonstrations under Sec. 63.45. An overview of the process is 
displayed in Figure 10.

BILLING CODE 6560-50-P

TP01AP94.009


BILLING CODE 6560-50-C
    The process begins with the submittal of an application containing 
the documentation of the emission increases and decreases. The contents 
of the application are described in paragraphs 63.46(d) and 63.47(c). 
The application must summarize and document each emission increase and 
decrease, and must document that the emissions meet the ``more 
hazardous'' requirement contained in Sec. 63.48 of the proposed rule. 
(For more information on the ``more hazardous'' test, there is an 
extensive discussion in section IV of this preamble).
    A streamlined administrative process for review of the application 
is provided in paragraphs 63.46 (e) and (f) and in paragraphs 63.47 (d) 
and (e). The permitting authority must review a ``contemporaneous'' 
demonstration within 45 days of receipt of a complete application, and 
must review a ``simplified'' demonstration with 30 days of receipt of a 
complete application. If approvable, the permitting authority issues a 
``Notice of Offset Approval.'' This Notice establishes emission 
limitations needed to ensure Federal enforceability of the offsets. The 
Notice is required to contain sufficient recordkeeping, monitoring and 
reporting requirements to demonstrate continuing compliance with 
Federally enforceable emission limits. Where EPA determines that the 
offset determination made by the permitting authority fails to meet any 
of the requirements of Sec. 63.46 or Sec. 63.47, EPA may take one of 
two actions to address the deficient offset determination: (1) Where 
offset determination is made as part of a source's part 70 permit, EPA 
may veto the issuance of the permit in accordance with the provisions 
of 40 CFR 70.8(c). The EPA may also use the veto process outlined in 40 
CFR 70.8(c) where the State has ``enhanced'' its section 112(g) process 
to incorporate the part 70 procedures. (2) Where the offset 
determination is made through a Notice of Offset Approval before the 
source obtains or revises their part 70 permit, EPA may exercise the 
authority authorized under section 113(a)(5) of the Act to prohibit 
construction or modification, issue an administrative penalty order or 
bring a civil action against the source upon finding that the State has 
not acted in compliance with any requirement or prohibition relating to 
the construction of new sources or the modification of existing 
sources. The EPA requests comment on whether the offset determinations 
should be required to adhere to the compliance provisions in 40 CFR 
part 70, or any enhanced monitoring provisions pursuant to section 114 
of the Act.
    Unless and until a Notice of Offset Approval is issued, the 
physical change or change in the method of operation increasing 
emissions is considered a ``modification.'' Operation of the change 
without a MACT demonstration or an offset approval would constitute 
noncompliance with section 112(g) of the Act. Because the 
``modification'' requirements in Sec. 63.43 prohibit the ``fabrication 
(on site), erection, or installation'' of the physical change or change 
in the method of operation unless MACT is established, the owner or 
operator will be out of compliance with Sec. 63.43 if any offset 
demonstration is disapproved subsequent to such ``construction'' date. 
Accordingly, applicants are strongly encouraged to obtain offset 
approvals before construction of the physical change.
    The EPA wishes to clarify with paragraphs 63.46(e)(8) and 
63.47(d)(8) that the administrative process described in the proposed 
rule is not the only possible review process a permitting authority 
could use. Other administrative review processes would be acceptable if 
they ensure that the source provides federally enforceable emission 
reductions before operation of the increase being offset, and that 
those reductions have been approved by the permitting authority as 
meeting the emission quantity and offsetting restrictions outlined 
elsewhere in the proposed rule.
    The EPA believes that the proposed procedures provide the 
opportunity for applicants to develop and document a continuing 
emission reduction program that can provide offsets for future 
activities. The EPA believes, in particular, that the proposed rule 
provides the opportunity for applicants to develop Federally 
enforceable source reduction programs which could serve to reduce 
emissions from a given production area while avoiding additional MACT 
requirements. The EPA requests comment on additional language that 
would serve to facilitate such programs.
H. Requirements for Emission Units Subject to a Subsequently 
Promulgated MACT Standard or MACT Requirement
    Emission units for which a case-by-case MACT determination is 
obtained in accordance with Sec. 63.45 of the proposed rule will be 
subject to future MACT requirements. For some emission units, at a 
future date, the emission unit may be a ``source'' in a subsequently 
promulgated MACT standard pursuant to section 112(d) of the Act. Also, 
some units subject to case-by-case MACT under section 112(g) may be in 
existence upon the date for which an ``equivalent emission limitation 
by permit'' is required pursuant to section 112(j) of the Act.
    Section 112(g) of the Act does not explicitly address how such an 
emission unit complying with a case-by-case MACT emission limit 
pursuant to section 112(g) should be treated under MACT standards or 
section 112(j).
    The EPA believes that there are cases where it would be reasonable 
to give emission units additional time to comply with Federally 
promulgated MACT standards, and that there are additional cases where 
compliance with the MACT standard should not be delayed. For example, 
if the case-by-case MACT standard required installation of a costly 
emission control device (for example, a state-of-the-art electrostatic 
precipitator), and if the subsequently promulgated MACT standard 
necessitated use of a similarly costly, but different device (for 
example, a dry scrubber followed by a fabric filter), then additional 
time would appear reasonable. On the other hand, if the only difference 
between the case-by-case standard and the promulgated standard was the 
use of a readily available alternate raw material, then little or no 
relief would appear reasonable.
    The EPA believes that section 112(j) provides a reasonable policy 
precedent for treatment of section 112(g) determinations relative to 
subsequent MACT standards. Accordingly, Sec. 63.49 of the proposed rule 
requires that emission units comply with subsequent MACT standards ``as 
expeditiously as practicable'' but afford the owner or operator the 
opportunity to justify a delay of up to 8 years. Such an approach is 
consistent with the statutory scheme of section 112(j), which is 
analogous to section 112(g) in this area. It is also a reasonable 
approach developed by the EPA to fill gaps left by Congress in section 
112(g).
    The EPA considered expanding to Sec. 63.49 to clarify the EPA's 
proposed reading of section 112(j) of the Act for situations where a 
case-by-case MACT determination is already in existence under section 
112(g). The EPA believes that if ``case-by-case MACT'' has been 
established under section 112(g), then it is not necessary to revisit 
this determination under section 112(j) for the portion of the major 
source subject to section 112(g). The EPA believes that this position 
is already taken in the proposed rule implementing section 112(j) of 
the Act, in proposed Sec. 63.52 and proposed Sec. 63.54. In Secs. 63.52 
and 63.54, if an owner or operator already has a permit requiring 
``compliance with a limit that would meet the requirements of section 
112(j) of the Act,'' then the owner or operator is not subject to any 
substantive review process. The EPA believes that a permit requiring a 
MACT emission limitation pursuant to a section 112(g) requirement would 
be one that ``meets the requirements of section 112(j) of the Act.'' Of 
course, where section 112(g) has required MACT to be applied to part of 
a major source, section 112(j) may require more extensive coverage. The 
EPA requests comment on this reading, and on whether explicit mention 
should be made of this reading in the section 112(g) regulation. In 
addition, comment is requested concerning whether a time limit should 
apply regarding whether compliance with a section 112(g) case-by-case 
MACT limit is sufficient to meet the requirements of section 112(j). 
For example, if an emission unit receives a permit containing a case-
by-case limit under section 112(g) in 1994, but the section 112(j) 
hammer does not fall until 1999, should the unit retain its 1994 limit, 
or be required to update it?
    Paragraph 63.49(c) addresses situations where a case-by-case MACT 
determination pursuant to section 112(g) is more stringent than a 
subsequently promulgated standard pursuant to section 112(d) of the 
Act. This paragraph clarifies that the permitting authority is not 
required to relax the emission limitation for such situations. However, 
the EPA also wishes to clarify that the permitting authority is not 
required by the proposed rule to maintain the more stringent case-by-
case emission limitation, and that it has the option to relax the 
emission limit to the 112(d) standard. The EPA requests comment on 
whether, if the emission limitation is not relaxed to the level of the 
112(d) standard, the facility should be allowed to use the additional 
reduction as a credit for offset or trading purposes.

IV. Proposed Approach for Demonstrating That Offsets Are ``More 
Hazardous'' (Sec. 63.48). Summary and Rationale

    The previous section of the preamble describes the overall 
requirements for the offset demonstrations pursuant to section 
112(g)(1)(A) of the Act. One important aspect of these requirements is 
that the offsetting emission decreases must be deemed ``more 
hazardous'' than the emission increases. This section of the preamble 
describes the rationale for EPA's proposed approach for making such a 
``more hazardous'' finding. The EPA requests comment on this approach. 
This discussion is supported by a technical background document which 
describes the rationale for the procedures in greater detail, and which 
provides documentation of the data used to evaluate the hazard of each 
listed HAP. (Draft Technical Background Document to Support Rulemaking 
Pursuant to the Clean Air Act section 112(g). Ranking of Pollutants 
with Respect to Hazard to Human Health, EPA 450/3-92-010).

A. Statutory Requirements for a ``More Hazardous'' Finding

    As discussed in previous sections of this preamble, section 112(g) 
provides the owner and operator the option of providing emission 
``offsets.'' If the owner or operator proposes to make a physical 
change or change in the method of operation of a major source that 
would constitute a ``modification,'' the control technology 
requirements for modifications can be avoided if emission offsets are 
provided.
    1. Offset Provision. Alternative Readings of ``More Hazardous.'' 
The offset provision in section 112(g)(1)(A) of the Act reads as 
follows:

    A physical change in, or change in the method of operation of, a 
major source which results in a greater than de minimis increase in 
actual emissions of a hazardous air pollutant shall not be 
considered a modification, if such increase in the quantity of 
actual emissions of any hazardous air pollutant from such source 
will be offset by an equal or greater decrease in the quantity of 
emissions of another hazardous air pollutant (or pollutants) from 
such source which is deemed more hazardous, pursuant to guidance 
issued by the Administrator * * *

    The EPA believes that this language is ambiguous, because it is not 
entirely clear whether ``pollutant'' or ``quantity'' is being modified 
by the phrase ``which is deemed more hazardous.'' The EPA believes this 
language would support multiple interpretations.
    Under one interpretation, each HAP increasing emissions above a de 
minimis level would require an offsetting decrease by another HAP which 
must be deemed ``more hazardous.'' For example, if emissions of 
pollutant A increased by 10 tons per year, a reduction of 10 or more 
tons per year of pollutant B could be allowed as an offset if pollutant 
B is considered to be more hazardous than pollutant A.
    Under an alternative interpretation, the increased emissions of 
each HAP would be required to be offset by a ``quantity of emissions of 
another HAP'' which is deemed more hazardous. Under this approach, if 
emissions of pollutant A were increased by 10 tons, pollutant B would 
not necessarily have to be a ``more hazardous pollutant'' so long as 
the decreased emissions from pollutant B represented a ``more hazardous 
quantity.'' Consequently, under this interpretation, the requirement 
that offsets must decrease ``hazard'' may be satisfied either by 
decreasing an equal or greater quantity of a ``more hazardous 
pollutant,'' or by decreasing a ``more hazardous quantity'' of another 
pollutant.
    2. Requirement for Guidance. Section 63.48 of the proposed rule is 
intended to satisfy the requirement in section 112(g)(1)(B) of the Act 
that EPA publish guidance that includes a relative ranking of the 
pollutants according to their ``hazard to human health.'' The statute 
reads as follows:

    The Administrator shall, after notice and opportunity for 
comment and not later than * * * [May 15, 1992] * * * publish 
guidance with respect to * * * [section 112(g) of the Act] * * * 
Such guidance shall include an identification, to the extent 
practicable, of the relative hazard to human health resulting from 
emissions to the ambient air of each of the pollutants * * * [listed 
as HAP under section 112(b) of the Act] * * * sufficient to 
facilitate the offset showing * * * Such guidance shall not 
authorize offsets between pollutants where the increased pollutant 
(or more than one pollutant in a stream of pollutants) causes 
adverse effects to human health for which no safety threshold for 
exposure can be determined unless there are corresponding decreases 
in such types of pollutant(s).

    The EPA believes the phrase ``to the extent practicable'' indicates 
that Congress recognized the limitations and uncertainties in data and 
methodologies for evaluating relative hazard of the 189 listed HAP. 
This phrase gives the EPA discretion to identify pollutants or health 
effects for which a relative hazard ranking is ``not practicable.''
    The EPA consulted an independent panel of scientific experts for 
input into the considerations that should be made in identifying the 
``practicable'' limitations in methodologies and data for the relative 
hazard ranking. This panel of the EPA's SAB was apprised of the EPA's 
draft outline for hazard ranking in a public meeting held on October 28 
and 29, 1991. The consultation meeting provided members of the SAB an 
opportunity to provide verbal feedback on several approaches. Summaries 
of a few of the comments by SAB panel members are included in this 
pollutant ranking discussion.
    The EPA interprets the phrase ``relative hazard to human health'' 
to mean that only health effects relevant to humans should be 
considered in the ranking, and not other adverse effects to the 
environment such as to wildlife, aquatic life and other natural 
resources. These latter effects are, however, addressed elsewhere in 
section 112 of the Act.

B. Overview of the Issues Involved in Establishing Procedures for a 
``More Hazardous'' Finding

    The requirement to identify the relative hazard of the 189 
pollutants, and the requirement to provide guidance for determining 
whether any decrease in one of the pollutants is ``more hazardous,'' 
both present a formidable challenge. Although air quality permit 
programs have, for years, allowed for emission offsets and ``netting 
credits,'' none of these programs have ever allowed emission offsetting 
between different hazardous air pollutants. The task is made 
particularly difficult by the magnitude of the list (189 pollutants, 17 
of these representing multi-pollutant groupings), the varying degrees 
of knowledge about the health effects caused by these chemicals, and 
the uncertainty in comparing the hazard potential of different health 
effects.
    In developing an approach to the ``more hazardous'' finding, legal, 
policy, scientific, and practical judgments must be made. From a legal 
standpoint, the approach must be consistent with the statutory 
language. As discussed previously, the EPA believes that the statute 
would support alternative readings. From a scientific standpoint, the 
approach should be defensible to the scientific community and should be 
consistent with the EPA's overall goal of incorporating the best 
scientific information available for decision-making. (For further 
information on the role of science in EPA decision-making, see 
Safeguarding the Future: Credible Science, Credible Decisions. EPA/600-
9-91-050. March 1992.) From a policy standpoint, any approach must: (1) 
Ensure that offsets are unlikely to increase the overall hazard to 
public health and (2) ensure consistency with the EPA's overall goal of 
providing regulated facility owners with the flexibility and incentives 
to seek emission reduction alternatives that are environmentally 
beneficial and cost-effective. From a practical standpoint, the 
approach must be implementable by applicants and by State and local 
permitting authorities and therefore must not be overly complex. In 
evaluating whether an emissions decrease is ``more hazardous'', the EPA 
believes that there is a tension between these overall objectives. The 
EPA requests public input as to whether the approach presented in the 
proposed rule strikes an appropriate balance.
    Commentors on this proposed rule should note that in developing a 
pollutant ranking ``to the extent practicable,'' that it was necessary 
to rely on current EPA methodologies and approaches. The EPA continues 
active efforts to review its approaches to assessing pollutant hazard. 
For example, the EPA is currently reviewing the implications of the 
recent NAS report entitled Science and Judgement in Risk Assessment, 
pending a detailed review, EPA may determine that changes to the 
section 112(g) hazard ranking are appropriate. In addition, section 
112(f)(1) requires that the EPA issue a report, by November 5, 1996, on 
methods to calculate risk and section 112(o)(7) requires that EPA 
publish revised Cancer Risk Assessment that incorporates results of the 
NAS study, or explains why the EPA elected not to make changes. 
Finally, section 303(f) requires that the Risk Assessment and 
Management Commission issue a report by May 15, 1994, on risk 
assessment methods and risk management practices. All three of these 
allow opportunity for public comment. The EPA encourages the public to 
take advantage of opportunities, such as those described above, to 
provide input into its decision making process. For the proposed rule, 
the EPA encourages commentors to address the application of its current 
methodologies (e.g. Reportable Quantities Composite Score system for 
determinations of relative chronic toxicity) to the hazard ranking for 
section 112(g), rather than on fundamental issues with the approaches 
themselves.
    The EPA realizes that no system of hazard ranking is going to be 
error free. Consequently, policy-based restrictions in this offset 
proposal that because of data limitation, etc., are intended to 
prohibit offsets that may increase hazard, may also restrict offsets 
which reduce hazard. The current proposal attempts to balance the 
concern with a potential increase in health hazards from offsets with 
the need for flexibility. That is, the proposal seeks to provide as 
wide a choice of offsetting possibilities from affected sources as is 
consistent with the need for a system which assures that the public 
health is protected, given the currently existing methodologies and 
data available from which to construct a hazard ranking system. The EPA 
requests comment on this approach and on specific changes that would 
increase the flexibility afforded by the offset program while assuring 
the protection of public health. The EPA also asks for comment on 
specific changes that would increase protection of public health while 
assuring that the program retains the flexibility necessary to an 
offset program. In either instance, changes suggested to the current 
approach should not increase the complexity of the system so as to 
render it too difficult to use or implement.

C. The Establishment of Relative Hazard Between Categories of 
Pollutants

    Section 112(g) requires that the EPA distinguish pollutants for 
which ``no safety threshold for exposure can be determined'' from the 
listed pollutants for the purposes of offsetting. Consequently the 
pollutants must be, at a minimum, categorized as either 
``nonthreshold'' or ``threshold.'' The EPA proposes that a third 
category also be established for pollutants which may be of ``high-
concern'' from either short term exposure (acute) or highly ranked for 
extremely severe chronic toxicity. Furthermore, the language in section 
112(g) directs the EPA to relatively rank the hazardous air pollutants 
``to the extent practicable.'' Such language recognizes that it may not 
be possible to relatively rank some of the listed pollutants and thus a 
fourth category of pollutants is created in which ``unrankable'' 
pollutants are placed.
    The first step in the relative ranking of the pollutants is to 
assign the pollutants to the four categories and to establish the 
relative hazard between these categories. The EPA's proposed approach 
to assign the pollutants to these categories and to relatively rank 
them is given below.
    1. Criteria for Assignment of Each Pollutant to One of The Four 
Categories. Figure 11 illustrates the four categories of pollutants 
used in the hazard ranking. As a first step, pollutants were 
categorized as either ``threshold'' or ``nonthreshold.'' As noted 
previously, the Act requires special consideration of ``nonthreshold'' 
pollutants.
    The types of health effects considered to be ``nonthreshold'' and 
the sources of information for identifying pollutants causing such 
``nonthreshold'' effects are discussed below. Pollutants which were not 
identified specifically as ``nonthreshold'' pollutants are categorized 
as ``threshold'' pollutants.

BILLING CODE 6560-50-P

TP01AP94.010


BILLING CODE 6560-50-C
    As a second step, the EPA created a list of pollutants called 
``high-concern'' pollutants to be separated from the rest of the 
``threshold'' and ``nonthreshold'' pollutants. This list is not 
explicitly required by the Act. The EPA believes, however, that 
pollutants whose primary concern is the potential for causing health 
effects from short-term exposures, and pollutants with very high 
toxicity from chronic exposures, require special consideration. The EPA 
requests comment on the creation of the ``high-concern'' category of 
pollutants. The data used for identifying these ``high-concern'' 
pollutants is discussed further below. ``Nonthreshold'' and 
``threshold'' pollutants which meet the criteria as a ``high-concern'' 
pollutant are listed in Table III. Pollutants which do not meet such 
criteria remain in the original two categories. Pollutants which do not 
have ``sufficient data'' to be placed in either ``nonthreshold,'' 
(Table I), ``threshold,'' (Table II), or ``high-concern'' (Table III) 
categories are considered to be ``unrankable'' and are placed in Table 
IV. The criteria for ``sufficient data'' are discussed below.
    2. Relative Hazard Between Categories. The determination of 
relative hazard between categories is described in Figure 12. 
Pollutants in the ``nonthreshold'' table (Table I) and the ``high-
concern'' table (Table III) are considered to be ``more hazardous'' 
than pollutants in the ``threshold'' table (Table II) and decreases in 
these pollutants are available as offsets for increases in emissions of 
``threshold'' pollutants. The EPA requests comment on these criteria.

BILLING CODE 6560-50-P

TP01AP94.011


BILLING CODE 6560-50-C
    The EPA considers it ``not practicable'' to establish the relative 
hazard between the ``high-concern'' and ``nonthreshold'' categories of 
pollutants. Emissions of a pollutant in Table I cannot, therefore, be 
used to offset emission increases of a pollutant listed in Table III. 
Similarly, emission reductions of a pollutant in Table III cannot be 
used to offset increases in the emissions of a pollutant listed in 
Table I.
    The EPA also believes it is ``not practicable'' to establish the 
relative hazard between those ``unrankable'' pollutants listed in Table 
IV and those pollutants belonging to the other three categories. As a 
result, emission reductions of pollutants in Table IV are not available 
as offsets for emission increases of any HAP. In addition, emission 
increases of pollutants in Table IV cannot be offset by decreases in 
emissions of any other HAP.
    3. Types of Toxicity Data Used To Assign Hazard To Pollutants In 
Each Category. Basis For Hazard Ranking.
    a. Identification of ``Nonthreshold'' Pollutants. Although the Act 
does not provide specific direction for identifying pollutants that 
have ``no safety threshold of exposure,'' Senator Lautenberg's 
testimony (136 Cong. Rec. S 17124) suggests that Congress intended for 
the EPA, at a minimum, to include pollutants with evidence of 
carcinogenicity in this category. The EPA presumes for cancer that any 
exposure is associated with some risk. Therefore, the consideration of 
carcinogens as having, in the absence of adequate evidence to the 
contrary, ``no safety threshold'' for exposure is consistent with the 
EPA's ``Guidelines for Carcinogen Risk Assessment'' (in The Risk 
Assessment Guidelines of 1986, U.S. EPA/ORD/OHEA, EPA 600/8-87-045), 
the Office of Science and Technology Policy (50 FR 10372-10442, March 
14, 1985), and the National Research Council (Risk Assessment in the 
Federal Government: Managing the Process. National Research Council. 
National Academy Press, Washington, DC, 1983).
    For the proposed rule for section 112(g), the EPA identified 
carcinogens based upon weight of evidence classifications described in 
the EPA's ``Guidelines for Carcinogen Risk Assessment.'' The proposed 
rule treats as ``nonthreshold'' pollutants, those pollutants with a 
weight of evidence classification under the guidelines of either A 
(Human Carcinogen), B (Probable Human Carcinogen), or C (Possible Human 
Carcinogen), or the equivalent thereof.2 The EPA believes that 
inclusion of all three categories for the proposed rule is consistent 
with section 112(f) of the Act which requires the EPA to promulgate 
emission standards to protect health and environment from ``known, 
probable, and possible'' human carcinogens. In addition, the EPA 
identified several pollutants which have been classified by the 
International Agency for Research on Cancer (IARC), but which have not 
been formally reviewed by the EPA. These pollutants are categorized by 
IARC as Group 1 (agent carcinogenic to humans), Group 2A (probable 
human carcinogen), and Group 2B (possible human carcinogen). These 
pollutants are also treated as ``nonthreshold'' pollutants under the 
proposed rule.
---------------------------------------------------------------------------

    \2\The EPA is in the process of revising its cancer guidelines. 
These classifications may change in any future revisions of the 
guidelines, and hence may impact the hazard ranking as outlined in 
this proposal. The current guidelines are contained in 51 FR 3392; 
September 24, 1986 and the EPA publication number is EPA/600/8-87/
045.
---------------------------------------------------------------------------

    The EPA will consider as ``nonthreshold'' pollutants, HAP with a 
weight of evidence of ``known, probable, and possible'' human 
carcinogens since in the absence of evidence to the contrary, hazard 
may be associated with any exposure level. Knowledge of the mechanism 
of action can raise questions regarding threshold assumptions (Thyroid 
Follicular Cell Carcinogenesis: Mechanistics and Science Policy 
Considerations, U. S. EPA, Office of Research and Development, Draft 
document December 1987 (edited 1988), EPA 265/3-88-014A). Currently 
there is no HAP for which the EPA has found such considerations to 
indicate convincingly for the entire weight of evidence so that the 
default assumptions are not applicable.
    The EPA's cancer guidelines include two additional weight of 
evidence categories: D (Not Classifiable, because there is insufficient 
data) and E (Evidence of Non-carcinogenicity). The proposed rule 
considers pollutants in either of these categories to be ``threshold'' 
pollutants pending further information.
    The EPA recognizes that there may be additional health endpoints 
besides cancer for which a ``nonthreshold'' assumption would be 
reasonable. At this time, the EPA has not identified any pollutants 
that would be added to the ``no safety threshold'' list based upon such 
nonthreshold noncancer effects. One class of pollutants, lead 
compounds, which has been suggested as having an increasing dose-
response relationship at current exposure levels for neurobehavioral 
effects, (Letter from R.C. Loehr and A. Upton to William K. Reilly. 
Subject: Science Advisory Board's Review of the Draft Assessment 
Document ``Review of the Carcinogenic Potential of Lead Associated with 
Oral Exposure.'' November 21, 1989) is also identified as a 
``nonthreshold'' pollutant based on it being considered probably 
carcinogenic to humans (B2). Based on these noncancer concerns, lead 
compounds are treated as ``high-concern'' hazardous air pollutants. The 
EPA requests comment on this approach and on other chemicals and health 
endpoints which should be considered as ``nonthreshold.''
    4. Data Base for Ranking Pollutants in Table I (``Nonthreshold'' 
Pollutants). Because all of the ``nonthreshold'' pollutants in the 
proposed rule have associated carcinogenic effects, the assignment of 
relative hazard of such pollutants was based on current EPA procedures 
for describing the hazard of carcinogens.
    Consistent with the EPA's current cancer guidelines, a description 
of carcinogenic hazard has two parts, a qualitative characterization 
(the weight of evidence that a substance causes cancer in humans) and a 
quantitative characterization (the assessment of the dose-response 
relationship). The EPA believes that both of these parts are important 
in the ranking of carcinogens for section 112(g).
    a. Weight of evidence. For the qualitative characterization, the 
EPA uses the previously identified weight of evidence classifications. 
Group A chemicals, relatively few in number, are those for which there 
is sufficient human evidence of carcinogenicity, often from 
epidemiologic studies of occupational exposure. Group B chemicals, 
which are more numerous, include chemicals for which there is 
``sufficient'' animal evidence, but for which there is limited (Group 
B1) or ``inadequate'' (Group B2) evidence in human studies. Group C 
chemicals are those with ``limited'' evidence in animal studies (which 
includes data from a single species, strain, or from an exposure route 
not directly relevant to humans), and ``inadequate'' evidence in human 
studies.
    The EPA used the following hierarchy of data sources in 
characterizing the weight of evidence of chemicals in the proposed 
cancer ranking: (The references are listed in order of preference, and 
are discussed in further detail in the Technical Background Document)
    i. Weight of evidence characterizations in the EPA's Integrated 
Risk Information System (IRIS).
    ii. For pollutants not yet reviewed by the EPA's carcinogen risk 
assessment verification endeavor (CRAVE) work group, weight of evidence 
characterizations in documents prepared by the EPA's Office of Research 
and Development (ORD), including various health assessment documents 
and profiles, and including documents prepared for Evaluating Potential 
Carcinogens in Support of Reportable Quantities of pursuant to CERCLA 
section 102. In addition, sources other than ORD documents were used 
(for example, SAB comments for perchloroethylene and 
trichloroethylene).
    iii. For pollutants which have not been reviewed by the EPA, 
characterizations by IARC were used.
    b. Dose-response. The characterization of the dose-response 
relationship is useful for making inferences about response associated 
with a particular level of exposure and for making relative comparisons 
between chemicals based on potency. The dose associated with a 10 
percent increase over background in cancer incidence (ED10) is chosen 
as the parameter for which to compare relative potencies across 
``nonthreshold'' HAP for several reasons. First, the ED10 is considered 
to be within the experimental data; issues related to the shape of the 
dose-response curve beyond the observable range are not relevant. 
Second, the ED10 is a statistically stable estimate and is relatively 
insensitive to the choice of the dose-response model.
    The EPA has used the ED10 as a hazard ranking tool for adjusting 
statutory Reportable Quantities under section 102 of CERCLA 
(Methodology for Evaluating Potential Carcinogenicity in Support of 
Reportable Quantity Adjustment Pursuant to CERCLA section 102, EPA/600/
8-89/053, June 1988). In addition, the same data which supports an 
estimate of the unit risk (as identified in IRIS or another EPA 
document) support the estimate of the ED10. The EPA requests comments 
on the use of the ED10 for ranking hazard.
    There are a number of chemicals identified as ``nonthreshold'' in 
Table I, that are not supported by data sufficient to develop a potency 
estimate. There are others for which the potency has not yet been 
evaluated by the EPA. Such pollutants are included in Table I but 
cannot be relatively ranked within the category. (A similar problem 
would exist if any ``nonthreshold'' pollutants were identified for 
effects other than cancer.) For EPA's proposed approach, this type of 
pollutant is generally considered to be more hazardous than the 
``threshold'' pollutants listed in Table II. However, because no 
potency value (1/ED10) is available, such ``nonthreshold'' pollutants 
cannot be ranked among the other ``nonthreshold'' pollutants. 
Consequently, EPA's proposed approach for identifying a ``more 
hazardous'' decrease in emissions in the proposed rule does not allow 
them as offsets for or to be offset by ``nonthreshold'' pollutants 
having potency estimates. Similarly, pollutants without potency 
estimates are not allowed as offsets for each other.
    The EPA requests comment on another option which is to assign 
default values to ``nonthreshold'' pollutants with no potency 
determination. Although such an option would provide a greater pool of 
pollutants available for offsets, the number of offsets that could 
potentially increase the human health hazard would also increase. 
Specifically, the EPA requests input on the default values that would 
be selected, and the data or policy assumptions that would be used to 
support such default values. The EPA also asks for comments on the use 
of structure-activity relationship analysis as one possible method for 
deriving quantitative potency estimates.
    c. Uncertainties. Several uncertainties arise in developing a 
relative ranking of hazard. First, in the absence of human data, an 
assumption is made that human sensitivity may be as great as the most 
sensitive responding animal species. Exceptions to the assumption of 
human sensitivity may be expected, however. For example, recent 
research shows that the development of kidney tumors through proximal 
tubule damage, resulting from accumulation of alpha 2 micro-globin in 
hyaline droplets, appears specific for the male rat (U. S. EPA, 
1991;Alpha 2 micro-Globin: Association with Chemically Induced Renal 
Toxicity and Neoplasia in the Male Rat, EPA/625-3-91/019F). For 
pollutants for which the primary concern is kidney tumors in the male 
rat, such tumors may have diminished relevance in evaluating the 
potential for human health effects. In this and in similar situations, 
humans may be qualitatively different than animals. In addition, it is 
assumed that the human response is quantitatively similar to the most 
sensitive animal species. That is, humans will have the same shape of 
the dose-response curve as animals. Differences in pharmacokinetics, 
metabolic, and pharmacodynamic processes that influence dose-response 
have only been addressed in a limited number of cases.
    Third, in many cases, the ranking is based on data from the oral 
exposure route since inhalation data are absent. Questions exist 
regarding the applicability of these data to identify an inhalation 
hazard since first-pass and dose-rate effects may be important. When 
route extrapolations have been made (e.g., when inhalation risks in 
IRIS are based on oral data), in almost all cases, an assumption of 100 
percent absorption from all exposure routes was used. Only in one case 
(bromoform) was a difference (an arbitrary judgment) in absorption via 
an inhalation exposure accounted for in the estimate of the ED10. Data 
are sparse in which to gauge the magnitude of error introduced into the 
ranking from the use of oral data. Pelpelko (1991; Effects of Exposure 
Route on Potency of Carcinogens. Regulatory Toxicology and Pharmacology 
13:3-17), in a limited comparison, observed, for systemic tumors, 
differences less than an order of magnitude between oral and inhalation 
routes for doses associated with either a 1 percent or 25 percent 
additional risk of cancer. It can be asserted from this comparison, 
that for HAP's which engender systemic carcinogenic hazards with both 
oral and inhalation exposures, the absence of inhalation data most 
likely does not lead to a large misclassification of HAP's in a 
relative ranking. No insight may be gained for HAP's which engender an 
oral hazard, but no inhalation hazard, or HAP's which elicit portal-of-
entry effects.
    5. Types of Toxicity Data Used for Relative Ranking of 
``Threshold'' Pollutants. Under the proposed rule, pollutants which are 
not listed in Tables I, II, or IV are considered to have ``safety 
thresholds of exposure'' and are subsequently labeled ``threshold'' 
pollutants.
    a. Data Base Selected for Use--Reportable Quantities Data Base. 
Under section 104 of CERCLA, the EPA developed a chronic toxicity 
scoring system as one of the methods used under CERCLA in establishing 
reportable quantities. (Technical Background Document to Support 
Rulemaking Pursuant to CERCLA section 102. Volume 2. Report to EPA/ORD 
and EPA/OSWER. August 1986.) This methodology explicitly takes both the 
dose and severity of effect into account, for chronic exposure, to 
determine the relative hazard of each pollutant. The hazard potential 
of each pollutant is determined by calculating a composite score (CS) 
which is the product of a dose rating (RVd) and a severity of 
effect rating (RVe). The RVd is based upon the human minimum 
effective dose (MED) for a given endpoint, often derived from animal 
data. A log-based algorithm in the scoring system is used to translate 
the MED into an RVd value between 1 and 10. Effects which occur at 
a low MED are assigned a relatively high RVd value. The severity 
rating, RVe, is a number between 1 and 10 which assigns a 
numerical score to the severity of a given health endpoint. The system 
used to assign the RVe values is included in this preamble as 
Table 1 as it appears in the technical support document for CERCLA 
section 102 volume 2. The resulting composite score, the product of the 
dose and severity ratings, is therefore a number between 1 and 100. 
Using this method, pollutants which elicit severe effects at relatively 
low doses are assigned a high composite score and those which produce 
relatively minor effects at high doses are given a low composite score.
    There are a number of advantages to using the RQ approach as a 
relative ranking approach for ``threshold'' pollutants. Because the 
approach has been used by the EPA for a number of years in the RQ 
process, data are available for the majority of ``threshold'' 
pollutants on the HAP list (the section 112(b) list). Additionally, the 
EPA believes it is appropriate for the purposes of the proposed rule to 
consider severity of health effects in the ranking. The EPA recognizes 
in this proposal that the severity scores are somewhat subjective, but 
believes that this scoring system nonetheless represents the best 
available tool for relatively ranking the large number of HAP on the 
section 112(b) list.
    The EPA used a number of data sources in evaluating the hazard of 
threshold pollutants. Three types of documents were available from 
EPA's Environmental Criteria and Assessment Office in the Office of 
Research and Development, including Reportable Quantity documents, 
Health and Environmental Effects Documents (HEEDs) and Health and 
Environmental Effects Profiles (HEEPs).
    In addition to the above data sources, for which composite scores 
could be obtained directly, there were a number of additional 
pollutants for which composite scores could be calculated based upon 
information used to develop RfCs. Documentation of each composite score 
used in the ``nonthreshold'' pollutant ranking is contained in the 
Technical Support Document (EPA-450/3-02-010).

  Table 1.--Severity of Effect Rating Values Used for Derivation of the 
                            Composite Score                             
------------------------------------------------------------------------
  Rating                                                                
   (RVe)                               Effect                           
------------------------------------------------------------------------
1.........  Enzyme induction or other biochemical change with no        
             pathologic changes and no change in organ weights.         
2.........  Enzyme induction and subcellular proliferation or other     
             changes in organelles but no other apparent effects.       
3.........  Hyperplasia, hypertrophy, or atrophy but no change in organ 
             weights.                                                   
4.........  Hyperplasia, hypertrophy, or atrophy with changes in organ  
             weights.                                                   
5.........  Reversible cellular changes: cloudy swelling, hydropic      
             change or fatty changes.                                   
6.........  Necrosis, or metaplasia with no apparent decrement of organ 
             function. Any neuropathy without apparent behavioral,      
             sensory, or physiologic change.                            
7.........  Necrosis, atrophy, hypertrophy, or metaplasia with a        
             detectable decrement of organ functions. Any neuropathy    
             with a measurable change in behavior, sensory, or          
             physiologic activity.                                      
8.........  Necrosis, atrophy, hypertrophy, or metaplasia with          
             definitive organ dysfunction. Any neuropathy with gross    
             changes in behavior, sensory, or motor performance. Any    
             decrease in reproductive capacity. Any evidence of         
             fetotoxicity.                                              
9.........  Pronounced pathologic changes with severe organ dysfunction.
             Any neuropathy with loss of behavioral or motor control or 
             loss of sensory ability. Reproductive dysfunction. Any     
             teratogenic* effect with maternal toxicity.                
10........  Death or pronounced life shortening. Any teratogenic effect 
             without signs of maternal toxicity.                        
------------------------------------------------------------------------
*Because this table is taken directly from the guidance document for the
  development of Composite Scores, the term ``teratogenic'' appears     
  here. The EPA now prefers to use the term ``developmental'' to refer  
  to these effects.                                                     

    b. Other alternatives considered for ``threshold'' pollutant 
ranking. Another approach considered by the EPA is to rank 
``threshold'' pollutants using (RfCs and Oral Reference Doses (RfDs). 
The RfC is defined as an estimate (with uncertainty spanning perhaps an 
order of magnitude) of a daily exposure to the human population 
(including sensitive subgroups) that is likely to be without 
appreciable risk of deleterious effects during a lifetime. The RfD is a 
similar type of estimate for oral exposures.
    For purposes of the proposed rule, the EPA prefers the composite 
score of the RQ scoring system to an RfC/RfD system. If RfCs were 
available for more chemicals, the RfC may be an appropriate ranking 
tool for ``threshold'' pollutants. However, as of the time of this 
proposal, RfCs are available for only a relatively small number of the 
threshold pollutants on the section 112(b) list.
    At this time, the EPA is reluctant to use RfDs, which are based on 
oral exposures, for ranking chemicals under a program for which 
inhalation exposures are the primary concern. Factors such as portal of 
entry effects (needed for metals, irritants, sensitizes) and liver 
first-pass effects limit the use of oral studies as meaningful 
indicators of toxicity from exposure by inhalation. The EPA requests 
comment on the relative merits of alternative approaches and their 
inherent uncertainties, and accompanying data sets in relatively 
ranking ``threshold'' pollutants and on chemical-specific information 
relevant to the hazard evaluation of each chemical.
    Another approach to the ranking, suggested by members of EPA's SAB, 
would make use of an information matrix. Several members of the SAB 
expressed interest in having the EPA present more data about the 
pollutants than was represented by a single composite score for each 
pollutant. Rather than using one composite score to describe the 
endpoint of concern for a pollutant, it was suggested that, in essence, 
multiple scores be developed for the many effects a pollutant may 
elicit. Also, it was suggested that the ``limiting effect'' or effect 
of most concern for each pollutant be chosen from the matrix and that 
offsetting only be allowed between pollutants with the same endpoint of 
concern. Specifically, only pollutants with the same target organ and 
endpoint would be allowed as offsets. Selection criteria would have to 
be developed to determine the effect of most concern for each 
pollutant. If a pollutant was of concern for more than one endpoint, 
the classification of the pollutant and the offsetting restrictions 
that apply to each pollutant would have to be determined. This is 
further complicated due to the fact that few, if any, HAP have been 
tested on all organ systems.
    The EPA considers this concept to be valid, but believes that the 
implementation of these suggestions would require more data on each 
pollutant than are currently available and would introduce another 
level of complexity for both sources and for the reviewing authorities. 
Also, the approach would greatly restrict the universe of pollutants 
available for offsets. The EPA invites comment on the information 
matrix approach suggested by SAB members for ranking threshold 
pollutants.
6. Criteria and rationale for identification of ``high-concern'' 
pollutants (Table III). The hazard ranking, which is used for 
determining a ``more hazardous'' decrease in emissions, categorizes 
pollutants as either ``threshold'' or ``nonthreshold'' in accordance 
with the requirements of the Act. A situation which must be addressed 
is the determination of a ``more hazardous'' decrease in emissions when 
the offsetting pollutant is a ``nonthreshold'' pollutant, but the 
pollutant being increased is a ``threshold'' pollutant.
    a. Statutory need. As noted above, the language in section 
112(g)(1)(B) of the Act specifically prohibits increases in emissions 
of pollutants having no ``safety threshold for exposure'' to be offset 
by decreases in ``threshold'' pollutants. However, the converse is not 
prohibited, and increases in emissions of ``threshold'' pollutants may 
be offset by decreases in ``nonthreshold'' pollutants if that decrease 
meets the ``more hazardous'' test. As stated previously, the 
``nonthreshold'' pollutants are considered to be more hazardous than 
the ``threshold'' pollutants. However, because it may be impossible to 
determine the relative hazard between some ``threshold'' pollutants and 
certain ``nonthreshold'' pollutants, a problem exists in having only 
two categories of pollutants.
    b. Recommended approach. The EPA recognizes the difficulty in 
evaluating pollutants for relative toxicity when there are two or more 
different types of effects (cancer and noncancer endpoints). 
Ultimately, in developing an approach to addressing this problem, there 
is no ``scientific'' solution, and policy judgements must be made. One 
possible approach would be to treat any ``nonthreshold'' pollutant as 
more hazardous than any of the ``threshold'' pollutants. While the EPA 
believes that this is appropriate in many cases, some weakly potent 
carcinogens on the ``nonthreshold'' list may represent a lesser human 
health hazard than certain highly toxic ``threshold'' pollutants. In 
order to begin to account for such cases, the proposed rule identifies 
a special category of pollutants which are referred to as ``high-
concern'' pollutants.
    The ``high-concern'' pollutants are listed in Table III of the 
proposed rule. Pollutants in this table are either: (1) Pollutants with 
composite scores above 20 (potent chronic toxicants), (2) pollutants 
whose primary toxicity is manifested from short exposures or peak 
releases at relatively low concentrations, or (3) pollutants for which 
concern from chronic toxicity may outweigh that of carcinogenicity 
(e.g., lead).
    Threshold pollutants which are not identified as ``high-concern 
pollutants'' in Table III are deemed ``threshold'' pollutants and are 
listed in Table II.
    The selection of the cutoff of 20 is a policy decision that is 
designed to be consistent with the approach taken for reportable 
quantities under CERCLA. For CERCLA, all carcinogens are assigned a 
reportable quantity of 100 pounds or less. For chronic noncancer 
effects, the CERCLA program assigns a reportable quantity of 100 pounds 
or less only to those pollutants with a composite score greater than 
20.
    The list of ``high-concern'' pollutants also includes a number of 
pollutants of concern for toxicity from short-term exposures. The EPA 
believes that toxicity from short-term exposures is not qualitatively 
or quantitatively comparable to carcinogenesis or chronic toxicity and 
therefore proposes that acutely toxic (and extremely potent) pollutants 
be identified and placed in the category of ``high-concern'' 
pollutants. For the proposed rule, the EPA has identified a number of 
pollutants of concern from short-term exposure using the ``LOC'' for 
``extremely hazardous substances'' pursuant to section 302 of the SARA. 
Pollutants are treated as high concern from short-term exposure if the 
LOC value is less than 8.0 mg/m3 or if an RFC exists for short-
term exposure exists and is below the 8.0 mg/m3 cutoff. The 
determination of this value is a policy decision based on the 
distribution of pollutants having LOC among the 189 HAP. Documentation 
of the LOC values for the HAP on the section 112(b) is provided in the 
draft technical background document (EPA 450/3-92-010). Although there 
are many caveats regarding the use of LOCs as a health safety benchmark 
or as ``safe breathing levels,'' the EPA believes that their use for 
the proposed rule as a tool to identify acutely toxic pollutants avoids 
some of the drawbacks for their use.
    The ``high-concern'' list in Table III includes a few carcinogens 
which are also of concern from short-term exposures. Where a LOC was 
available for a carcinogen, and that LOC is less than 8 milligrams per 
cubic meter, the carcinogen was placed in Table III instead of Table I, 
because of the implications for greater restrictions on offsetting.
    The EPA believes that the LOC approach is a reasonable first step 
in identifying pollutants for which toxicity from short-term exposures 
would be of high concern. The LOC values are indicative of the relative 
concentrations at which the pollutants create an immediate danger of 
death or irreversible health effects. Although owners and operators of 
hazardous air pollutant sources regulated under section 112(g) are 
unlikely to propose emitting these ``acutely'' toxic substances at 
levels that would cause the LOC values to be exceeded, the EPA believes 
that these values can be used as a tool to identify pollutants of 
concern for toxicity from short-term exposures.
    At the same time, the EPA recognizes that this is a very imperfect 
tool. The LOC values, which rank pollutants essentially by their 
relative ability to cause lethality, are probably an inadequate 
indicator of the relative likelihood that short-term exposures could 
cause effects such as neurological, developmental, and reproductive 
effects. The EPA requests comment on whether the ``high-concern'' 
category should be created, and on the policy decisions that were made 
in identifying the pollutants for this category. The EPA is 
particularly interested in comments concerning the merits of using 
Composite Scores and LOC values and the implications of its proposed 
cutoffs. In this regard, the EPA seeks comment concerning examples of 
offsets that reduce hazard which would be precluded by the criteria in 
the proposed rule. Similarly, the EPA asks for examples where the 
current proposal is not restrictive enough to prevent offsets from 
occurring which cause an increase in hazard. The EPA is attempting to 
identify other methods which could complement the LOC screening 
approach to identifying pollutants of concern for short-term exposures. 
The EPA requests comments on methodologies that could be used to 
address this specific issue.
7. Table IV. Unrankable pollutants. If a pollutant does not meet the 
criteria to be included as a ``nonthreshold'' pollutant in Table I, has 
insufficient chronic toxicity data for purposes of Table II, and does 
not meet the Table III criteria for a ``high-concern'' pollutant, then 
the EPA considers this pollutant as not ``practicable'' to rank at this 
time. As more information becomes available on such pollutants, the EPA 
will amend the hazard ranking to include the pollutants as is 
appropriate. Alternatively, a panel of experts may be convened by the 
EPA to rank such pollutants as necessary. The EPA asks for comment on 
how best to address these pollutants.
8. Treatment of chemical groups. There are 17 HAP listed under section 
112(b) which are chemical groups having no unique chemical abstract 
service (CAS) numbers. These groups are as follows: antimony compounds, 
arsenic compounds (inorganic including arsine), beryllium compounds, 
cadmium compounds, chromium compounds, cobalt compounds, coke oven 
emissions, cyanide compounds, glycol ethers, lead compounds, manganese 
compounds, mercury compounds, fine mineral fibers, nickel compounds, 
polycyclic organic matter, radionuclides (including radon), and 
selenium compounds.
    The pollutants within each chemical group do not always have 
homogeneous toxicological profiles. In some instances, the EPA believes 
it is reasonable to assume that all chemical species within a group are 
equally toxic. However, in order to provide a hazard ranking that meets 
the ``extent practicable'' test, the EPA believes that it is sometimes 
necessary to identify appropriate subgroupings or, in some cases, 
individual pollutants within these 17 broad classes with distinct 
toxicological properties.
    One frequently used approach for chemical groups is to treat the 
group as a class according to most toxic pollutant or subgrouping 
within the class. Unlike other programs for which such an approach is 
health-protective, for the section 112(g) program this would not 
necessarily be the case. For example, if a group of chemicals was 
proposed as offsets based upon the most toxic chemical in the group, 
and if the pollutant actually being reduced was the least toxic 
chemical in the group, then an improper hazard comparison would result 
and an increased risk to the public would be allowed.
    As a starting point in identifying appropriate subgroupings, EPA 
staff identified the chemicals and chemical classes within the 17 
groups for which reportable quantities have been established under 
CERCLA. This served to identify the chemicals within the groups for 
which ED10s and composite scores were available, and therefore could be 
included within the ranking. The information supporting currently 
available RQ was complemented with some more recent information (for 
example, for cresols and for several glycol ether compounds). In some 
cases, hazard is inferred for a chemical class (e.g. inorganic arsenic 
compounds). In other cases, available data indicated a significantly 
different hazard potential (e.g., beryllium salts vs. other beryllium 
compounds). The rationale and documentation concerning the treatment of 
the 17 chemical groups is presented in the draft technical support 
document (EPA 450/3-92-010). The EPA requests public comment on the 
selection of the pollutants and subgroupings that were included in 
hazard ranking, and on the methodology for ranking them.

D. The Determination of Relative Hazard Within Categories of Pollutants

Requirements of the Proposed Rule
    Having established the four categories of pollutants and the 
relative hazard between them, the next step is to assess the relative 
hazard of pollutants within each category. The proposed rule contains 
methods to define a ``more hazardous'' and ``equally hazardous'' 
pollutant within the categories.
1. ``Nonthreshold'' Pollutant Increases
    Paragraph 63.48(a). a. Approach in Proposed Rule. As discussed 
above, all of the ``nonthreshold'' pollutants are included in Table I 
based upon their potential to cause cancer. For comparing the relative 
hazard of two ``nonthreshold'' pollutants, Table I lists both 
quantitative hazard information and weight of evidence information. 
This information is used in determining whether one ``nonthreshold'' 
pollutant is ``more hazardous'' than another. In order for a carcinogen 
to be deemed ``more hazardous,'' three conditions must be satisfied.
    First, the pollutant being decreased must be another pollutant in 
Table I. As discussed above, section 112(g)(1)(B) does not allow 
decreases in ``threshold pollutants'' as offsets for increases in 
``nonthreshold'' pollutants.''
    Second, the pollutant must have an equal or greater weight-of-
evidence classification. Pollutants with a weight of evidence of ``A'' 
or ``B'' according to the EPA's Guidelines for Carcinogen Risk 
Assessment (U.S. EPA, 1986) are determined to be ``more hazardous'' 
than pollutants with a weight of evidence of ``C'' (possibly 
carcinogenic to humans). For the purposes of the proposed rule, 
chemicals categorized as having weight of evidence of ``Group A'' or 
``Group B'' are treated as having greater hazard than a weight of 
evidence of ``Group C.'' The EPA feels that the evidence of 
carcinogenicity in either animals or humans which is defined as 
``sufficient'' under EPA's cancer guidelines (U.S. EPA, 1986) provides 
a compelling case for a greater hazard concern than evidence defined as 
less than ``sufficient'', i.e., ``limited'' evidence in animals.
    Third, the ``more hazardous'' pollutant must have a potency (1/
ED10) value, listed in Table I of the proposed rule, which is greater 
than the ``less hazardous'' pollutant by at least a factor of 3. Any 
given potency value is uncertain and the EPA proposes, as a policy 
decision, that this uncertainty spans approximately one order of 
magnitude. This uncertainty is assumed, again as a policy decision, to 
effectively bracket each ED10 value by an amount equal to a factor of 3 
(approximately the square root of one order of magnitude) both above 
and below the ED10 value. The EPA requests comment on this treatment of 
uncertainty in determining ``more hazardous nonthreshold pollutants.''
    Three of the 189 hazardous air pollutants are considered to have 
weight of evidence which falls on the spectrum between ``Group 
B2'' and ``Group C.'' These are: perchloroethylene, 
trichloroethylene, and lindane. For the purposes of ranking hazard, 
these pollutants are considered by the proposed rule to be more like 
``B'' than ``C.''
    For ``nonthreshold'' pollutants listed in Table I, two pollutants 
are considered ``equally hazardous'' if the pollutants have potency (1/
ED10) estimates that do not differ by more than a factor of 3, and the 
offsetting pollutant has the same (treating A and B carcinogens as one 
group) or a greater weight of evidence as the pollutant whose emissions 
are being increased.
    The EPA requests comment on the merits of these three proposed 
criteria. In particular, the EPA is interested in the way carcinogens 
are grouped in the proposed rule. The EPA is also interested in comment 
on whether the proposed factor of three adequately takes into account 
uncertainty surrounding the potency estimate for the carcinogens. In 
this regard, the EPA seeks comment concerning examples of offsets that 
reduce hazard which would be precluded by the criteria in the proposed 
rule. Similarly, the EPA seeks examples where the current proposal is 
not restrictive enough to prevent offsets from occurring which cause an 
increase in hazard.
    b. Other Alternatives Considered for Identifying ``More Hazardous'' 
Decrease in Emissions of Carcinogens. The EPA reviewed several 
alternatives to the approach outlined in Paragraph 63.48(a) of the 
proposed rule.
    One alternative approach for determining the relative hazard 
between pollutants is to develop an ordinal ranking of potency 
estimates. Such a ranking would treat each potency estimate as a 
discrete value and would ignore the uncertainty of the estimate. For 
example, a potency value of 10 would indicate a greater hazard than a 
potency value of 9.5. The EPA believes that for the purposes of the 
ranking such fine scale distinctions should not be made when the 
uncertainty in the data is taken into account. Additionally, this 
approach could prompt frequent reordering of the ranking as new 
scientific data become available and the potency estimates change.
    Another approach that EPA considered would subdivide the potency 
estimates into groupings or ``bins.'' This approach increases the 
stability of the ranking, because for any given pollutant, small 
changes in the potency value would not cause a change in the bin 
assignment. This approach may also have advantages in the treatment of 
multiple-pollutant streams. (If a group of pollutants were in the same 
bin, an emission total for that bin could in some cases be more 
straightforward to evaluate than treating each pollutant individually). 
However, this approach does not adequately reflect the differences in 
hazard for pollutants immediately adjacent to the borderline of the 
bins. For example, using bins of 1-10, 10-100 and 100-1000, a pollutant 
with a value of 101 would be treated as more hazardous than a pollutant 
with a value of 99, while a pollutant with a value of 99 would be 
treated as equally hazardous as another pollutant with a value of 11. A 
comparison between the approach in the proposed rule and to the fixed 
bin approach is displayed in Tables 2 and 3. The EPA also considered 
alternatives that are based solely on quantitative estimates of 
potency, and which would not consider weight of evidence as a factor in 
the hazard comparison. Ignoring weight of evidence classification, for 
example, could allow increases of a known (Group A) human carcinogen of 
low potency to be offset by decreases in a moderately or highly potent 
possible (Group C) human carcinogen. The EPA requests comment on 
whether such an offset should be considered as a decrease in hazard. 
The EPA believes that weight of evidence is an important consideration 
in describing the hazard associated with carcinogens.
    The potency (1/ED10) values for the cancer ranking were obtained 
using the data base for RQ under CERCLA as a starting point. In 
compiling these values, the EPA recognized that the data base generated 
for ranking air pollutants under section 112(g) of the Act would 
require a greater emphasis on inhalation. For the proposed rule, the 
EPA reviewed the ED10 values in the CERCLA data base and made 
adjustments where deemed appropriate (oral values adjusted to 
inhalation values). Adjustments were also performed to ensure that the 
data base used in the determination of the ED10 was consistent with 
that used to develop unit risk estimates of carcinogenic potency for 
pollutants listed in the EPA's IRIS. The documentation of the specific 
ED10s used for the proposed cancer ranking is discussed in the 
technical background document.

                        Table 2.--Example of Proposed ``Nonthreshold'' Pollutant Ranking                        
----------------------------------------------------------------------------------------------------------------
                                                                                                     Weight of  
    CAS No.                              Example pollutant                           Potency (1/     evidence   
                                                                                      ED10)\1\    classification
----------------------------------------------------------------------------------------------------------------
75558...........  1,2-Propylenimine...............................................         260    B             
118741..........  Hexachlorobenzene...............................................          13    B             
91941...........  3,3-Dichlorobenzidine...........................................           7.5  B             
95534...........  o-Toluidine.....................................................           1.6  B             
79469...........  Vinylidene chloride.............................................           1.2  C             
79469...........  2-Nitropropane..................................................       (\2\)    B             
----------------------------------------------------------------------------------------------------------------
\1\The greater the potency value, the more hazardous the pollutant.                                             
\2\Not available.                                                                                               

    Example comparisons:
    1,2-Propylenimine is more hazardous than Hexachlorobenzene because 
the potency value is more than three times greater (260 is more than 
3 x 13) and because the weight of evidence classification is equal 
(both are Group B).
    3,3-Dichlorobenzidine is ``equally hazardous'' than 
Hexachlorobenzene because the potency value is within a factor of three 
(13 is less than 3 times 7.5) and the weight of evidence classification 
is equal (both are Group B).
    Vinylidene chloride cannot be considered more hazardous than any of 
the other pollutants because it does not have an equal or greater 
weight of evidence classification (Group C class is lower than the 
others which are either Group A or Group B).
    2-Nitropropane (no potency value available) would not be allowed to 
offset, or be offset by other ``nonthreshold'' pollutants.

                          Table 3.--Example Illustrating Fixed-Bin Alternative Approach                         
----------------------------------------------------------------------------------------------------------------
                                                                         Potency (1/                            
     CAS No.                        Example pollutant                       ED10)          WOE         Bin No.  
----------------------------------------------------------------------------------------------------------------
75558...........  1,2-Propylenimine...................................         260    B                        1
118741..........  Hexachlorobenzene...................................          13    B                        2
91941...........  3,3-Dichlorobenzidine...............................           7.5  B                        3
95934...........  o-Toluidine.........................................           1.6  B                         
75014...........  Vinyl chloride......................................           1.6  A                         
75354...........  Vinylidene chloride.................................           1.2  C                         
79469...........  2-Nitropropane......................................       (\1\)    B                        ?
----------------------------------------------------------------------------------------------------------------
\1\Not available.                                                                                               

    Example comparisons: 1, 2-Propylenimine and Hexachlorobenzene are 
more hazardous than pollutants 3, 3 Dichlorobenzidine, o-Toluidine, 
Vinyl chloride and Vinylidene chloride because ``Bin 1'' and ``Bin 2'' 
pollutants are considered to be more hazardous than ``Bin 3'' 
pollutants.
2. ``Threshold'' Pollutant Increases
    Paragraph 63.48(b). As discussed above, Table II contains 
pollutants that are considered to be ``threshold pollutants.'' Two 
types of pollutants are considered ``more hazardous'' than pollutants 
listed in Table II.
    First, consistent with the overall approach described in section 
IV.C. above, any ``nonthreshold'' pollutant in Table I is considered a 
``more hazardous pollutant'' than any pollutant in Table II.
    Second, any ``high-concern'' pollutant in Table III is considered 
to be a more hazardous pollutant than those in Table II.
    A more hazardous pollutant from Table II can be defined by the 
following. In order to determine the relative hazard between pollutants 
listed in Table II, a ``more hazardous pollutant'' is defined as a 
pollutant whose composite score exceeds that of another pollutant by a 
sufficient amount. Similar to the approach described above for ranking 
carcinogens, the EPA believes that the uncertainty in the data should 
be considered in determining whether one threshold pollutant is ``more 
hazardous'' than another. For the proposed rule, a ``threshold'' 
pollutant is assumed to be more hazardous than another if its composite 
score exceeds that of the other pollutant by at least four composite 
score units. This value of four reflects a policy judgment by the EPA. 
A discussion of how this value was obtained is discussed in the draft 
technical support document (EPA 450/3-92-010). For the purposes of this 
rule, a ``threshold'' pollutant is considered to be ``equally 
hazardous'' to other ``threshold'' pollutants whose composite score is 
less than four composite score units. The EPA asks for public comment 
on whether the uncertainty in composite scores should be considered 
and, if so, how it should be considered for the section 112(b) 
pollutants.
    The EPA recognizes that the policy decision in the proposed rule 
provides different treatment of pollutants which differ by 4 or more 
composite score units than those that are within 4 composite score 
units. In this regard, the EPA seeks comment concerning examples of 
offsets that reduce hazard which would be precluded by the criteria in 
the proposed rule. Similarly, the EPA asks for examples where the 
current proposal is not restrictive enough to prevent offsets from 
occurring which cause an increase in hazard.
3. ``High-concern'' Pollutant Increases
    Paragraph 63.48(c). ``High-concern'' pollutants are listed in Table 
III. For some pollutants in Table III, a composite score is listed, 
while for other pollutants, the descriptor ``A'' is given. An asterisk 
indicates that the pollutant is also treated as a carcinogen.
    Pollutants which list a composite score are included in the table 
on the basis of severe chronic toxicity. For these pollutants, a ``more 
hazardous pollutant'' is another pollutant in Table III whose composite 
score is at least four composite score units greater than the pollutant 
being increased.
    Pollutants having the descriptor ``A'' are included on the list on 
the basis of severe effects from short-term exposures to relatively low 
concentrations. The EPA believes that it is not ``practicable'' to 
determine a ``more hazardous pollutant'' for pollutants of concern for 
``acute'' toxicity. Therefore, pollutants appearing in the ``high-
concern'' category with the designation of A are not allowed as offsets 
or to be offset with each other. Furthermore, they are not allowed to 
offset or be offset with pollutants in the category assigned composite 
scores.
    The EPA requests comment on the merits of the proposed data and 
methodologies for identifying the ``high-concern'' category. In this 
regard, the EPA seeks comment concerning examples of offsets that 
reduce hazard which would be precluded by the criteria in the proposed 
rule. Similarly, the EPA asks for examples where the current proposal 
is not restrictive enough to prevent offsets from occurring which cause 
an increase in hazard.
4. ``Unrankable'' Pollutants
    Paragraph 63.48(d). As discussed previously in section IV.C. of 
this preamble, the ``unrankable'' pollutants in Table IV cannot be used 
as offsets and cannot be offset by other HAP. The EPA requests comment 
on the merits of not allowing offsetting for ``unrankable'' pollutants. 
In this regard, the EPA seeks comment concerning examples of offsets 
that reduce hazard which would be precluded by the criteria in the 
proposed rule. Similarly, the EPA asks for examples where the current 
proposal is not restrictive enough to prevent offsets from occurring 
which cause an increase in hazard.
5. Treatment of Pollutant Mixtures
    The EPA recognizes that not all proposed offsets will involve 
comparison of a single pollutant being increased with a single 
pollutant being decreased. The ``more hazardous'' finding must also 
address emission streams containing mixtures of pollutants being 
increased and decreased.
    For the proposed rule, components of such pollutant mixtures are 
treated individually. The first step required, when an emission 
increase involves a mixture of pollutants, is to identify the HAP in 
the mixture which are being emitted in greater than de minimis 
quantities. A source owner seeking to offset the emission increases 
would need only to offset those pollutants which are increased above de 
minimis levels. Each pollutant in the stream is then categorized as 
being either ``nonthreshold'' (if listed in Table I), ``threshold'' (if 
listed in Table II), ``high-concern'' (if listed in Table III) or ``not 
practicable'' to rank (if listed in Table IV). Appropriate offsetting 
decreases must then be identified, depending on which of these 
categories the pollutant falls into. An example of the offsetting 
process for a stream of pollutants is given in Tables 4, 5, and 6 
below.
    If an appropriate offset cannot be identified for a given pollutant 
in the stream, then that stream is subject to the control technology 
requirements in section 63.35 of the proposed rule. 

                     Table 4.--Example Emission Increase Involving a Multi-Pollutant Stream                     
 [The following table illustrates an example emission increase from a multi-pollutant stream. The example stream
 contains several pollutants. For each pollutant, the example identifies where in the ranking (i.e., Sec. 63.48 
of the proposed rule) the pollutant can be found. Finally, the table displays the type of data pertinent to that
                                         section of the proposed rule.]                                         
----------------------------------------------------------------------------------------------------------------
                                                                                            Weight              
     CAS No.        Pollutant being increased      Ranking category (which    Potency (1/     of      Composite 
                                                      section of rule?)          ED10)     evidence     score   
----------------------------------------------------------------------------------------------------------------
91941...........  3,3 Dichlorobenzidine.......  ``Nonthreshold'' Table I....          7.5  B                    
75014...........  Vinyl chloride..............  ``Nonthreshold'' Table I....          1.6  A                    
748794..........  Mercuric chloride...........  ``High-concern'' Table III..  ...........  ........           40
126998..........  Toluene.....................  ``Threshold'' Table II......  ...........  ........            7
----------------------------------------------------------------------------------------------------------------


  Table 5.--Pollutants Considered ``Equally Hazardous'' Under the EPA's 
                            Proposed Approach                           
------------------------------------------------------------------------
                                                      Which offsetting  
                                                    pollutants would be 
    CAS No.               Stream pollutant          considered ``equally
                                                        hazardous''     
------------------------------------------------------------------------
91941...........  3,3-Dichlorobenzidine..........  Any ``nonthreshold'' 
                                                    pollutant, i.e.,    
                                                    found in Table I,   
                                                    if:                 
                                                   Weight of evidence is
                                                    A or B and the      
                                                    potency (1/ED10) is 
                                                    in following range: 
                                                   Greater than 2.5     
                                                    (i.e., 7.5/3) and   
                                                    less than 23 (i.e., 
                                                    7.5 x 3).           
75014...........  Vinyl chloride.................  Any ``nonthreshold'' 
                                                    pollutant if:       
                                                   Weight of evidence is
                                                    A or B and the.     
                                                   Potency is in        
                                                    following range:    
                                                   Greater than 0.53    
                                                    (i.e., 1.6/3) and   
                                                    less than 4.8 (i.e, 
                                                    1.6 x 3).           
748794..........  Mercuric chloride..............  Any ``high-concern'' 
                                                    pollutant, i.e.,    
                                                    Table III, if:      
                                                   The composite score  
                                                    is greater than 36  
                                                    (i.e., 40 minus 4)  
                                                    and less than 44    
                                                    (i.e., 40 plus 4).  
126998..........  Toluene........................  Any ``threshold''    
                                                    pollutant, i.e., in 
                                                    Table II if:        
                                                   The composite score  
                                                    is greater than 3   
                                                    (i.e., 7 minus 4)   
                                                    and less than 11    
                                                    (i.e., 7 plus 4).   
------------------------------------------------------------------------

    Pollutants in Table I are equally hazardous if the potency (1/ED10) 
value varies by less than a factor of 3 and weight of evidence 
restrictions are observed.
    Pollutants in Table II are equally hazardous if the potency 
(composite score) value varies by less than + or -4 composite score 
units.

   Table 6.--Pollutants Considered ``More Hazardous'' Under the EPA's   
                           Proposed Approach                            
------------------------------------------------------------------------
                                                      Which offsetting  
                                                    pollutants would be 
    CAS No.               Stream pollutant             deemed ``more    
                                                        hazardous''     
------------------------------------------------------------------------
91941...........  3,3-Dichlorobenzidine..........  Must be another      
                                                    ``nonthreshold''    
                                                    pollutant (Table I):
                                                    must pass potency   
                                                    and wt of evidence  
                                                    tests.              
                                                   Potency must be > 23 
                                                    (i.e., 3 x 7.5).    
                                                   Weight of evidence   
                                                    must be A or B.     
75014...........  Vinyl chloride.................  Must be another      
                                                    ``nonthreshold''    
                                                    pollutant (Table I).
                                                   Potency must be > 4.8
                                                    (i.e., 3 x 1.6).    
                                                   Weight of evidence   
                                                    must be A or B.     
748794..........  Mercuric chloride..............  Must be another      
                                                    ``high-concern''    
                                                    pollutant (Table    
                                                    III).               
                                                   Composite score must 
                                                    be > 44 (4 + 40).   
                                                   ``Nonthreshold''     
                                                    pollutants (Table I)
                                                    are not considered  
                                                    ``more hazardous''. 
126998..........  Toluene........................  Can be another       
                                                    ``threshold''       
                                                    pollutant in Table  
                                                    II, if composite    
                                                    score is > or = 11. 
                                                   Any ``high-concern'' 
                                                    pollutant in Table  
                                                    III is considered   
                                                    ``more hazardous''. 
                                                   Any ``nonthreshold'' 
                                                    pollutant (including
                                                    potency A, B, or C  
                                                    is ``more           
                                                    hazardous'').       
------------------------------------------------------------------------

E. Determination of a ``More Hazardous'' Decrease in Emissions

    As mentioned previously, there are two possible interpretations of 
the language in section 112(g) describing an emissions decrease ``which 
is deemed more hazardous.'' Under the approach based upon a ``more 
hazardous pollutant'' reading of the statute, a pollutant with 
increased emissions must be offset by an equal or greater quantity of 
emissions from another HAP considered to be ``more hazardous.''
    The approach proposed by the EPA in this rule attempts to determine 
the quantity of emissions decrease needed to constitute a ``more 
hazardous quantity.'' Under this approach, the EPA does not attempt to 
establish the magnitude of the difference between pollutants.
    Under EPA's proposed approach, a ``more hazardous quantity'' 
consists of two possibilities: (1) A greater or equal decrease in a 
pollutant that is deemed ``more hazardous'' according to the scheme 
described above, and (2) more (the proposed rule requires 1.25 times as 
much) of a pollutant that is deemed ``equally hazardous.''
    The EPA's proposed approach is similar to a ``more hazardous 
pollutant'' approach in all aspects except for one. Pollutants which 
have similar enough hazard estimates so that a determination cannot be 
made with certainty that they are different in hazard are considered to 
be ``equally hazardous.'' While ``equally'' hazardous pollutants are 
not allowed as offsets under the ``more hazardous pollutant'' approach, 
they are allowed as offsets under the EPA's proposed approach if a 
``more hazardous quantity'' of an ``equally hazardous'' pollutant is 
used as an offset. The EPA's proposed approach does not attempt to 
establish the magnitude in difference in ``hazard'' between pollutants. 
The establishment of a ``more hazardous quantity'' of an ``equally 
hazardous'' pollutant is a fixed percentage increase and is a policy 
decision. The EPA proposes to set that percentage to be 25 percent more 
than the increase in emissions and recognizes that, due to uncertainty 
in the hazard estimates, it may not accurately reflect the actual 
differences in hazard between two pollutants. The EPA solicits comment 
on the merits of establishing a 25 percent minimum increase in 
reductions as the definition of a ``more hazardous quantity'' of 
equally hazardous pollutants. In this regard, the EPA seeks comment 
concerning examples of offsets that reduce hazard which would be 
precluded by the criteria in the proposed rule. Similarly, the EPA asks 
for examples where the current proposal is not restrictive enough to 
prevent offsets from occurring which cause an increase in hazard.

F. Miscellaneous Hazard Ranking Issues

    The above discussion outlines the overall methods for ranking the 
pollutants under the EPA's proposed approach. Within this overall 
framework, rest a few important issues for which the EPA seeks comment.
1. Consideration of Non-Inhalation Hazard
    A potentially important consideration in the hazard ranking is the 
potential of a given pollutant for non-inhalation routes of human 
exposure. This non-inhalation exposure potential could be particularly 
important for particulate HAP which could deposit in the vicinity of 
the release point, and which are persistent and/or bioconcentrate. Such 
pollutants could create a greater hazard for exposure by non-inhalation 
pathways such as soil, plant or fish ingestion, than would other 
pollutants which do not deposit, are not persistent, or do not tend to 
bioconcentrate. All other properties being equal, it would appear that 
a pollutant which has a high potential for such non-inhalation 
exposures should be considered more hazardous than another pollutant 
with a low potential.
    The EPA believes that there are a number of unaddressed questions 
which need further discussion before non-inhalation exposure potential 
is explicitly accounted for in the ranking: (1) How would a qualitative 
criterion such as non-inhalation exposure potential be incorporated 
into the ranking methodologies proposed? (2) Should non-inhalation 
exposure potential be given the same weight as potency and weight of 
evidence for carcinogens and composite score for non-carcinogens? (3) 
Should the toxicity of certain pollutants be downgraded if non-
inhalation exposure is not of great concern? and (4) Which dose should 
be used for consideration of effect; the doses received at the ``fence 
line'' immediately or the dose which is accumulated after a specified 
length of time?
    For this proposal, there is no explicit consideration given for 
non-inhalation exposure potential in the hazard ranking. Currently, the 
EPA does not believe there is an adequate quantitative procedure for 
such a consideration.
    The EPA considered possible approaches for addressing non-
inhalation exposure potential in the ranking. Under section 112(i), 
Early Reductions, an additional weighting factor of 10 was added to 
pollutants of concern for bioaccumulation. As noted previously in this 
preamble in the discussion regarding de minimis values, in conjunction 
with the Great Waters study pursuant to section 112(m) of the Act, the 
EPA has identified a list of pollutants of greatest concern. One 
possible approach would be to make adjustments in the hazard ranking 
for the pollutants on this list. However, it would be difficult to 
select the amount of adjustment that should be made. Another option 
would be to use information on bioconcentration and persistence to 
place threshold pollutants onto the high-concern threshold pollutant 
list and further restrict the pollutants available for offsetting. The 
EPA requests comment on how this information could be used to provide 
for explicit consideration of non-inhalation exposure potential in the 
ranking.
2. Half-Life
    Aside from the non-inhalation exposure considerations, another 
possible consideration in the ranking is the persistence of the 
pollutant in the atmosphere. The EPA considered accounting in the 
ranking for the fact that some pollutants (for example, carbon 
tetrachloride) have very long half-lives in the atmosphere, while other 
pollutants are highly reactive, with half-lives on the order of a few 
hours. Pollutants with relatively long half-lives would tend to expose 
greater numbers of people to very low concentrations. Pollutants with 
relatively short half-lives would tend to have a similar impact to 
nearby residents, but would have a lesser exposure potential further 
downwind.
    The EPA proposes that half-life not be taken into consideration in 
the pollutant ranking. The EPA believes that the focus of the program 
should be the potential for ambient exposures to individuals exposed to 
potentially high concentrations. The EPA requests comment on this 
issue.
    A corollary to this issue is the potential for reactive pollutants 
to form transformation products which may be more hazardous than the 
parent compounds. At present, the EPA has not identified a method for 
taking this into account in the ranking. The EPA requests comment on 
this issue.
3. Appeal Process; Consideration of Exposure Data and Other Mitigating 
Factors
    The goal of the hazard ranking is to ensure that the offsetting 
emissions will provide for public health improvements relative to the 
emission increases that will occur. During the SAB consultation meeting 
and in other discussions, commentors have questioned whether a hazard 
ranking system can be developed that will never make a mistake, i.e., 
allow an increased risk, due to the complexity of the task and the 
sparseness of the scientific data. They identified a number of 
mitigating factors which have not been incorporated into the ranking 
which could affect the health impact. The SAB members and other 
commentors have advised that EPA provide for the opportunity to do an 
exposure analysis and to consider other physical parameters such as 
chemical properties and atmospheric formation or degradation potential 
in determining offsetting restrictions. The hazard ranking in the 
proposed rule, in effect, uses emission rates as a surrogate for 
exposure assessment. While a comprehensive site-specific risk 
assessment, which could include detailed information on exposure, would 
conceivably be more health protective, the EPA is concerned that the 
time required to perform and review such an assessment would seriously 
burden both affected industries and reviewing authorities. Such an 
extensive review time seems contrary to Congressional intent for 
minimizing the delays associated with these requirements.
    In addition to the exposure issue, there are probably offsets 
allowed or prohibited by the hazard ranking that would be treated 
differently based upon a detailed case-by-case review of the health 
effects data base for the chemicals in question. The EPA recognizes 
that any general scheme for ranking the hazard of pollutants as diverse 
as those listed in section 112(b) will not be error-free. Such 
situations were of particular concern to participants of the SAB 
consultation meeting, who advised that any hazard ranking include an 
appeal process to account for such errors.
    The EPA requests public comment on whether the final rule should 
contain such an appeal process, either by the plant owner, by the 
permitting authority, or by the public. While recognizing the overall 
merits of such a process, the EPA is concerned that such an appeal 
process could substantially increase the time needed for review and 
subsequently the decision on whether a source could avoid a case-by-
case MACT determination. Additionally, many of the permitting 
authorities, which EPA envisions will be implementing section 112(g), 
do not have staff with expertise in toxicology and exposure assessment. 
One option would be to centralize the appeal process at the Federal 
level, but this could substantially delay the review process.
4. Treatment of Noncancer Effects of Carcinogens
    The EPA recognizes that ``nonthreshold'' pollutants may produce a 
variety of health effects in addition to cancer, including noncancer 
toxicity from ``short-term,'' subchronic, and chronic exposures. For 
example, health effects data for some pollutants indicate that exposure 
may produce noncancer effects such as respiratory irritation, 
neurotoxicity, or developmental toxicity in addition to cancer. The 
EPA's proposed approach currently ranks carcinogens primarily by their 
carcinogenic potency. Potential human carcinogens which are acutely 
toxic, or manifest toxicity from relatively low concentrations at short 
durations of exposure, are unavailable as offsets to be offset by other 
carcinogens. Such pollutants are also not available as offsets with 
each other but are considered to be ``more hazardous'' than the 
``threshold'' pollutants.
    A question which the present approach does not address is how to 
treat carcinogens which may also be of concern for chronic toxicity. 
Two options for evaluation of noncancer toxicity of carcinogens have 
been developed by the EPA. In option 1, ``noncancer'' pollutants with 
sufficient evidence and extent of noncancer toxicity are identified as 
``high-concern'' pollutants. In option 2, ``nonthreshold'' pollutants 
are evaluated for noncancer health effects in addition to cancer 
concerns by application of expanded offset rules. The EPA asks for 
comment on two basic approaches which may be considered for 
appropriately taking into account the non-carcinogenic effects of 
carcinogens. Both approaches would have problems with implementation 
due to inadequate data and would increase the complexity of the current 
scheme.
    In the first approach, ``nonthreshold'' pollutants which may have 
sufficiently high chronic noncancer toxicity as the endpoint of concern 
rather than carcinogenicity are identified as ``high-concern 
pollutants.'' The first step would be to assign appropriate composite 
scores to all ``nonthreshold'' pollutants, identify those which would 
be of concern for chronic toxicity (i.e., have high composite scores). 
Then those pollutants which have composite scores greater than 20 would 
be put into the ``high-concern'' category. A variation of that approach 
is to put a subset, only those pollutants with either no or a low 
potency estimate (1/ED10 less than 0.1) and a Composite Score greater 
than 20, into the ``high-concern'' category. Once placed in the ``high-
concern'' category, these pollutants would not be available as offsets 
for other potential human carcinogens, other pollutants in the ``high-
concern'' category with composite scores, or pollutants identified as 
being acutely toxic. These pollutants would still be considered to be 
more hazardous than ``threshold'' pollutants. This group of compounds 
would have a status similar to lead compounds under the current EPA 
approach.
    In the second approach, the spectrum of health effects associated 
with pollutants are considered in evaluating pollutant offsets. The 
health effects that may result from exposure to pollutants are 
generally identified as: (1) carcinogenicity (equated in this guidance 
with ``nonthreshold''), (2) noncancer effects from chronic exposures, 
and (3) noncancer effects from ``short-term'' exposures. In option 2, 
both carcinogenicity and noncancer toxicity from chronic exposures are 
explicitly considered in evaluating offsets, while noncancer toxicity 
from acute exposures is evaluated insofar as identifying pollutants 
that meet the ``high-concern'' list criteria. A sequential evaluation 
of carcinogenicity and noncancer effects from chronic exposures occurs 
in this option. First, cancer is evaluated according to the procedures 
in Sec. 63.48 of the proposed rule. Then, emissions of the pollutant 
being increased are compared to de minimis values to determine if 
noncancer toxicity from chronic exposures should be evaluated. If the 
emissions are above a de minimis value for noncancer effects from 
chronic exposures, then the composite scores are evaluated consistent 
with the determination of relative hazard for threshold pollutants (as 
discussed in Sec. 63.48). To implement this approach requires both de 
minimis values for noncancer effects from chronic exposures as well as 
composite scores. Public comments are requested on these options, on 
the availability of data to implement this approach, and on the 
scientific and policy options the EPA should consider when little or no 
health effects data needed to implement the approach are available.
    The EPA requests comments on the overall concepts of Option 2, and 
on the practicality and availability of data of each of the proposed 
options.
5. Weighted Offsets
    Another critical decision in the determination of ``a more 
hazardous'' decrease in emissions is whether or not to allow for 
``weighted offsets.''
    Under the proposed rule, a ``more hazardous'' decrease in emissions 
may be determined in two ways: (1) A greater or equal decrease in 
emissions of a pollutant that is deemed ``more hazardous'' according to 
the scheme described above, or (2) more (the proposed rule requires 
1.25 as much) of a pollutant that is deemed ``equally hazardous.''
    Under a weighting system, a policy judgment would be made that 
would assign a weighting factor for each pollutant. For each pollutant 
being increased, the emission rate would be multiplied by the weighting 
factor to yield a weighted emission increase. The sum, over all 
pollutants, of the weighted emission increases would yield an overall 
weighted emission increase for the modification. Similarly, for each 
pollutant being proposed as offsets, the emission decrease would be 
multiplied by the weighting factor to yield a weighted emission 
decrease. The sum over all pollutants being decreased would yield an 
overall weighted decrease. If the sum of the weighted decreases 
exceeded the sum of the weighted increases by a specified amount or 
ratio, then the offsets would be approved. Such a simple system of 
weighting factors for the most toxic HAP has been developed by the EPA 
for use in another Clean Air Act program for early reductions pursuant 
to section 112(i)(5).
    A number of commentors to the EPA have supported the concept of 
weighted trading as providing more flexibility for sources to find 
acceptable offsets. Such a system could create more opportunities for 
offsets than the proposed rule. For example, a system of weighted 
offsets would allow offsets of a ``more hazardous'' quantity of ``less 
hazardous'' pollutants as well as less than 1 to 1 offsets of a ``more 
hazardous'' pollutant if the exact quantity of that pollutant which is 
``more hazardous'' can be determined.
    Another commenter to the EPA has suggested that the ``more 
hazardous pollutant'' language in section 112(g) does not provide for 
weighted offsets. Consequently only an equal or greater quantity of a 
``more hazardous'' pollutant will satisfy the requirements for a ``more 
hazardous emissions decrease.''
    The EPA attempted to develop a system of weighting factors for use 
in the section 112(g) rule and a draft approach was circulated 
internally within the EPA and to members of the Clean Air Act Advisory 
Committee. During the review process, EPA scientists emphasized their 
belief that is not possible to develop a generalized system of 
weighting factors for such a large and diverse number of pollutants 
that is scientifically credible. This conclusion is based upon 
limitations in the scientific data base on health effects for the 
various HAP, and in particular the uncertainty in methods for 
quantifying the difference in hazard between pollutants with varying 
endpoints of concern. In order to provide a basis for public dialogue 
on this issue, the EPA is providing in the public docket for the 
proposed rule: (1) A description of the EPA draft weighting factor 
system, and (2) a critique on the issue of weighting factors by the 
EPA's Office of Research and Development. The EPA requests comment on 
the scientific defensibility of this approach, and other possible 
weighting factor approaches, relative to the approach in the proposed 
rule.
    The EPA believes that there are important differences between the 
needs of the section 112(g) program and those of the section 112(i)(5) 
``early reductions'' program. The approach used in the early reductions 
program is described in detail in the Federal Register notice for the 
program (56 FR 27738, June 13, 1991). For the early reductions program, 
the Act requires the EPA to ``limit'' the offsetting of ``high-risk'' 
pollutants. Accordingly, weighting factors were selected for a 
relatively small subset of the list of 189 HAP. For section 112(g), an 
approach is needed which addresses the entire list of 189 HAP. For the 
early reductions program, it was felt that emissions of an extremely 
small (trace) amount of a very hazardous pollutant could prevent a 
company from participating in the program (they may not be able to 
reduce such an amount by 90 percent); the weighting factors were 
developed principly to allow companies to reduce hazard by controlling 
emissions which could be reduced instead of the trace amount, thus, 
avoiding this trace emission problem. For section 112(g), the EPA 
believes that trace emission increases of HAP will not prevent 
beneficial results, because trace emission increases likely will not be 
greater than the de minimis emission rates established in the proposed 
rule.
    Another important difference between the two programs is that the 
early reductions program will always involve an environmental benefit, 
i.e., the 90 or 95 percent reduction in HAP emissions. If weighting 
factors are inaccurate under the early reductions program, the 
environmental benefits would be somewhat minimized, but benefits would 
nonetheless occur. On the other hand, if a weighting system is used for 
the section 112(g) rule and is inaccurate, offsets could lead to an 
increase in hazard or the health risk.
    The EPA believes that it is important to recognize that the early 
reductions weighting approach was not intended to serve as a precedent 
for other programs. In the June 13, 1991 Federal Register preamble, the 
EPA emphasized this point as follows:

    The selection of today's approach for purposes of section 
112(i)(5)(E) is not intended to establish a precedent for the other 
provisions affected by hazard ranking or preclude the consideration 
of other alternatives * * * (See 56 FR 27361).

    In particular, the weighting factors of 1 and 10 for non-
carcinogens, which were based upon a broad policy decision for the 
early reductions program, are inadequate for describing the differences 
in potency or severity between pollutants for purposes of offset 
comparisons under section 112(g). The actual differences in potency 
between the non-carcinogens could span many orders of magnitude.
    The EPA recognizes the need to provide updated information related 
to potency estimates and weight of evidence as new scientific data 
become available. The EPA requests comment on two approaches to 
providing such updates. The first approach would be to provide for 
periodic updates to the ranking, which would involve periodic revisions 
to the proposed rule. A second approach would be to provide for a data 
base which could reflect automatic updates to the regulation. The 
latter approach could provide for a more expeditious process for 
updating the ranking. The primary disadvantage of this approach is that 
it might require the data base to be consulted frequently to ensure an 
up-to-date assessment.

V. Discussion of the Relationship of the Proposed Requirements to Other 
Requirements of the Act

    The previous sections of this preamble discuss the requirements of 
the proposed rule in defining the requirements of section 112(g) of the 
Act as it relates to constructed, modified or reconstructed major 
sources of HAP. In addition, there are a number of issues concerning 
the relationship between the requirements of section 112(g) and other 
requirements of the Act that are relevant to the implementation of the 
requirements of the proposed rule. These issues are important in 
defining the overall responsibilities of States and the EPA in carrying 
out the requirements of section 112(g), and in understanding how 
section 112(g) requirements relate to other important requirements of 
the Act. The purpose of this section of the preamble is to present, and 
to take comment on, a number of regulatory and statutory 
interpretations related to these implementation issues.

A. Relationship of Section 112(g) Implementation to Title V Program 
Approval

    Title V of the Act and the part 70 regulations provide that a State 
seeking to obtain or retain approval of a title V program must have 
authority to assure compliance with all applicable requirements through 
the title V permit. Section 502(b)(5)(A); 40 CFR 70.4(b)(3)(i). The 
preamble to the proposed operating permits rule explains that, in the 
context of section 112, the permitting authority must have authority 
to, ``develop and enforce case-by-case determinations of MACT for new, 
reconstructed, or modified sources where no applicable emissions 
limitations have been yet established.'' (56 FR 21723 (May 10, 1991)). 
This is reaffirmed in the preamble to the final operating permits rule. 
(See 57 FR 32260 (July 21, 1992).) The final rule and preambles to the 
proposed and final rule thus make clear that, to ``assure compliance'' 
with section 112(g), the State must be able to make case-by-case 
determinations of MACT.
    This rule and preamble language represent what EPA considers to be 
the most natural reading of section 112(g). The EPA reads the reference 
in section 112(g)(2) to case-by-case determinations made by ``the 
Administrator (or the State)'' to mean that these determinations must 
be made by the title V permitting authority. This reading is consistent 
with the reference in section 112(g)(2) to the effective date of the 
title V program as the date on which the requirements of section 112(g) 
become applicable, and with the title V requirement that major sources 
of HAP submit applications for title V permits regardless of whether 
they are subject to a MACT standard. It is also consistent with the 
reference in section 112(j) to ``the Administrator (or the State)'' as 
the entity that must make case-by-case determinations of MACT and issue 
permits incorporating these determinations.
    As noted above, the authority to implement all aspects of section 
112(g) is a prerequisite to obtaining and retaining title V approval. 
However, since section 112(g)(2) does not take effect until the 
effective date of a title V program, EPA believes it is not necessary 
for the State to have this authority upon the date of submittal of the 
title V program. For purposes of the permit program submittal, it is 
sufficient for the State to demonstrate that it has the broad 
legislative authority needed to implement all aspects of section 
112(g), to include a commitment that it will have the necessary 
additional authority and resources to implement section 112(g) upon the 
effective date of the title V program. In practical terms, this means 
the State must be able to demonstrate the adequacy of its authority and 
resource capabilities with respect to section 112(g) prior to the 
approval of the title V program. However, the EPA does not intend to 
require a formal demonstration of adequate authority and resources at 
that time unless it believes there is reason to question the State's 
ability to implement section 112(g).

B. Relationship to the Section 112(l) Delegation Process

    Under section 112(l) of the Act, States have the option of 
developing and submitting to the Administrator a program for 
implementing the requirements of section 112, including section 112(g). 
The EPA promulgated a rule for the implementation of section 112(l) on 
November 26, 1993 (58 FR 62262). This rulemaking added Secs. 63.90 
through 63.96 to 40 CFR part 63.
    The EPA proposes that the approval processes provided under section 
112(l) be used to facilitate the implementation of section 112(g) by 
States in a way that minimizes disruption of existing State and local 
toxic air pollutant permit programs. During the mid to late 1980's, 
most States adopted regulations or procedures to review toxic air 
pollutant emissions from new and modified sources. In some cases, those 
programs already regulate some or all of the equipment covered by 
section 112(g). In other cases, the programs affect fewer pollutants or 
do not require as stringent control requirements. The section 112(l) 
process can be used by States to preserve existing requirements, while 
incorporating the requirements of this proposed rule, into an overall 
program that meets the requirements of the Act.
    The section 112(l) rule provides a State with several options for 
gaining EPA approval of alternative State requirements. Two of these 
options are applicable to section 112(g). Under the first option, 
addressed by 40 CFR 63.92 of the section 112(l) rule, a State may 
submit a program implementing section 112(g) with changes that clearly 
make their program no less stringent than the Federal rule (i.e., the 
final version of 40 CFR part 63, subpart B). These changes are referred 
to in the section 112(l) rule as ``adjustments'' and are listed in 40 
CFR 63.92. An example of such an adjustment for section 112(g) would be 
lower de minimis values. Another option which is available to a State 
wishing to make broader changes, or when it is less clear that 
differences in the State's program make them as stringent as the 
Federal rule, is to submit a detailed demonstration in accordance with 
40 CFR 63.93. States wishing to use this approach must provide a 
detailed demonstration ensuring that for all affected sources, the 
State rule is no less stringent than the Federal rule.
    Some States may choose to adopt the Federal section 112(g) program 
with no changes. Where this is the case, the EPA does not believe that 
a formal review process is required under section 112(l). For such 
States, the EPA believes that the review and approval process involved 
in its review of the part 70 permit program submission is adequate to 
ensure that the section 112(l)(5) criteria for approval of the State's 
section 112(g) program are met.
    There are two important issues related to section 112(l) program 
submittals as they relate to the section 112(g) requirements. The EPA 
requests comment on these two issues, for purposes of providing 
guidance to EPA Regional Offices and States regarding the section 
112(l) approval process for section 112(g) programs.
    First, a number of State programs contain technology requirements 
(sometimes referred to as Toxic Best Available Control Technology, or 
T-BACT). An important issue with regard to approval of section 112(l) 
programs is whether such T-BACT programs need to include consideration 
of the ``MACT floor'' described in section 112(d) of the Act. The EPA 
believes that the MACT floor is a fundamental requirement of section 
112(g) determination, and hence, the ``MACT floor'' must be considered 
to obtain approval of section 112(g) authority. As discussed previously 
in this preamble, the EPA is working with State agencies to develop 
technology transfer methods to ensure that the consideration of the 
MACT floor is not unduly burdensome.
    Second, a number of States have expressed interest in implementing 
the section 112(g) program without the provision for emission offsets. 
The EPA requests comment on whether the deletion of offsets could be 
considered an ``adjustment'' which could be added to the list of 
adjustments in 40 CFR 63.92. For existing State law or programs which 
do not provide for offsets, the EPA believes that the inclusion of 
emission offsets provided by the Federal section 112(g) rule would be 
of little practical use. Accordingly, if a State chose to continue 
their program, a choice which is specifically allowed by section 
112(l), and that program does not provide offsets, then EPA does not 
believe that an insistence on offsets as a condition of a section 
112(l) approval is necessary. Furthermore, the EPA believes that States 
have broad authority to make changes to specific offset requirements 
that are imposed under the Federal rule (for example, a State 
requirement for a risk assessment in addition to the requirements in 
the section 112(g) rule), so long as those changes ensure that the 
State program is no less stringent than any requirement imposed by the 
Federal rule. The EPA requests comment on the above discussion with 
respect to State programs without offsets, including any specific 
changes to the proposed rule which could serve to clarify this issue.

C. Section 112(i)(5) Early Reductions Program

    Section 112(i)(5) allows owners and operators, that provide early 
reductions in hazardous air pollutant emissions, to be granted a 6-year 
extension of any compliance date for emission standards issued under 
section 112(d). In order to participate in the section 112(i)(5) 
program, the owner or operator defines a ``source'' at a plant site for 
which a 90 or 95 percent reduction in emissions can be accomplished 
before the proposal date of the emission standard. There are a number 
of items of clarification on the relationship between the section 
112(i)(5) requirements and section 112(g).
    First, the extension granted by section 112(i)(5) applies only to 
that equipment incorporated within the ``source'' for which the 90 or 
95 percent reduction was accomplished. Other equipment at a plant-site 
not included within that ``source'' definition are subject to section 
112(g) requirements if they make changes that would be considered to be 
construction, reconstruction, or modification of a major source 
pursuant to the proposed rule.
    On the other hand, equipment within the ``source'' definition for 
which there is an approved early reductions submittal are not subject 
to further control technology requirements under section 112(g). 
Section 112(g) requires case-by-case MACT where no ``applicable 
emission limitation'' exist. The EPA proposes that the ``alternative 
emission limitation'' under section 112(i)(5) be considered an 
``applicable emissions limitation'' for purposes of section 112(g), 
such that compliance with such alternative emissions limitation shields 
a source from having to comply with section 112(g).

D. Section 112(j) ``Hammer'' Provision

    Section 112(j) of the Act contains case-by-case MACT requirements 
for sources for which EPA has not promulgated emission standards 
according to the schedule contained in section 112(e) of the Act. If 
EPA has not promulgated an emission standard for a category, then, 
within 18 months of the deadline for that standard, the owner or 
operator must submit a permit application to obtain a case-by-case 
emission limitation judged by the permitting authority to be equivalent 
to a Federal MACT emission limitation. The EPA has proposed a rule that 
would implement the section 112(j) requirements. (58 FR 37778, July 13 
1993.)
    States and sources implementing the requirements of section 112 of 
the Clean Air Act need to understand the potentially complex 
relationships among the requirements of section 112(d), (g) and (j).
    The EPA's primary goal is to create as seamless a web as possible 
between case-by-case MACT determinations under section 112(g) and 
implementation of subsequent section 112(d) standards for those same 
source categories. In addition, the EPA desires to rationalize the 
section 112(g) provisions with the section 112(j) provisions requiring 
case-by-case MACT determinations for constructed, reconstructed, and 
modified major sources. While under the Act some of the specific 
substantive requirements of section 112(g) differ under the Act from 
the substantive requirements of sections 112(j) and 112(d), the EPA 
intends to ensure the greatest possible operational consistency among 
section 112(d), (g), and (j) provisions.
    One fundamental principal guiding the design of all three programs 
is that substantive control requirements under section 112(g) hold only 
until the requirements of a sections 112(j) or 112(d) standard become 
effective. In other words, after the effective date of a section 112(j) 
case-by-case MACT determination or a section 112(d) MACT standard, the 
control requirements of section 112(j) or section 112(d) supersede the 
control requirements of section 112(g).
    The EPA considered an alternative approach, i.e., the finding that 
section 112(g) governs all changes and additions of new emission units 
at existing sources whether or not a section 112(d) or section (j) 
standard exists. The EPA rejected this approach for reasons enumerated 
below. Nevertheless the EPA requests comment on both approaches.
    One reason for rejecting the approach that section 112(g) control 
extends to sources covered by section 112(d) or section 112(j) 
standards is that it leads to the conclusion that many new sources 
within the section 112(a)(4) definition of new source would forever 
escape having to apply a new source MACT level of control. Such an 
interpretation is in conflict with the requirements of section 112(d).
    Section 112(a)(4) defines a new source as ``a stationary source the 
construction or reconstruction of which is commenced after the 
Administrator first proposes regulations under this section 
establishing an emission standard applicable to such source.'' Thus, 
once a standard has been set under section 112(d), any new source will 
be subject to new source MACT. Moreover, under section 112(a), a 
``stationary source'' can be ``major'' section (112(a)(1)) or ``area'' 
section (112(a)(2)). The MACT standard will define the portion of a 
facility that is considered a ``source'' for the purposes of the 
particular standard.
    Section 112(g) applies to construction, reconstruction, or 
modification of major sources, and in many cases will have an effect on 
sources earlier than sections 112 (d) or (j) standards. However, 
section 112(g) only requires new source MACT on new major sources, and 
considers any other new emission unit to be a modification of an 
existing major source. As a ``modification,'' such a new emission unit 
will be required to apply for existing source case-by-case MACT 
determination under section 112(g). Therefore, if section 112(g) were 
to constrain the application of a subsequent section 112(j) or section 
112(d) standard, many new emission units under the section 112(a)(4) 
definition of ``new source'' would never be required to comply with new 
source MACT.
    In addition, under section 112(g) a new emission unit might not 
even be required to meet an existing source MACT level of control. 
Section 112(g) allows for modifications to either: (1) Comply with a 
case-by-case ``existing source'' MACT determination under section 
112(g); (2) offset emissions increases in lieu of applying section 
112(g) existing source MACT requirements; or (3) if its emissions were 
below section 112(g) de minimis levels, not be subject to any control 
requirements at all. The EPA believes that section 112(g) thus provides 
major sources with a great deal of needed flexibility before section 
112 (d) or (j) standards are set; but that once those standards are in 
place the Act intends that these sources must comply with the specific 
requirements of those standards.
    Finally, the interpretation that section 112(g) governs the 
addition of new equipment at major sources to which section 112 (d) or 
(j) standards already apply has some anomalous implications. One 
example would be a new emission unit whose emissions are below section 
112(g) de minimis levels for a particular hazardous air pollutant. If 
that emission unit were added to a major source, it would be exempt 
from the requirements of section 112(g), but would be required to apply 
new source MACT control under section 112(j). However, if that emission 
unit were not below section 112(g) de minimis levels, it would be 
required to comply with section 112(g). If section 112(g) requirements 
limit the application of section 112(j), then the source would be 
required to apply existing source MACT. In this instance, a smaller 
emission unit would be required to control more stringently than a 
larger emission unit.
    Another example of anomalies resulting from this reading of the 
statute would be a section 112(d) standard that sets new source MACT 
for new area sources in a source category. Under this reading, major 
sources adding new sources could avoid new source MACT, but any new 
area source would have to meet new source MACT. Again, a smaller unit 
would be required to control more stringently than a larger emission 
unit.
    Therefore EPA believes that the substantive control requirements of 
section 112(g) are pre-empted by the requirements of a relevant section 
112(j) or section 112(d) standard, in cases where the construction, 
reconstruction or modification occurs after the date section 112(j) or 
section 112(d) standards apply.
    However, as noted above in the discussion on Sec. 63.49 of the 
proposed rule, the EPA believes that an emission unit already complying 
with a case-by-case determination under section 112(g) should be 
assumed to comply with the requirements of section 112(j).

E. Subpart A ``General Provisions''

    As mentioned previously, the EPA has proposed ``general 
provisions'' to the MACT program as proposed Subpart A to 40 CFR part 
63 (58 FR 42760, August 11, 1993). These general provisions contain a 
number of definitions and provisions that generally affect the subparts 
of part 63 that follow, including Subpart B discussed here. Another 
purpose of the general provisions is to provide general descriptions of 
requirements that are intended to direct the reader to appropriate 
subparts. The proposed general provisions envisioned that a subsequent 
rulemaking would add language to cover general aspects of the section 
112(g) requirements. In today's proposed rule, the EPA proposes a 
number of additions to subpart A to accomplish this goal. These 
additions are not intended as substantive interpretations of section 
112(g) of the Act, but are intended to direct a reader of the general 
provisions to the substantive requirements in Subpart B. The EPA 
requests comment on these general provisions additions.

F. Section 112(g) Implementation During the Transition Period

    Section 112(g)(2)(A) requires that after the effective date of a 
part 70 permit program to implement title V of the Act in any State, no 
person may modify, construct or reconstruct a major source of HAP in 
such State, unless the Administrator (or the State) determines that the 
MACT limitation in this section has been met or that sufficient offsets 
have been provided. The EPA interprets the statute to require that 
States must implement section 112(g), including development of case-by-
case MACT determinations, in order to obtain and retain approval of a 
part 70 permit program. The EPA also believes that the prohibition on 
construction, reconstruction, or modification takes effect on the 
approval date of the part 70 program. That is, subject sources may not 
construct, reconstruct, or modify unless the permitting authority has 
approved either a case-by-case MACT determination or an offset showing.
    Under section 502(a), the EPA must approve or disapprove within 12 
months a State submittal to implement title V which is due November 15, 
1993. As a result, the final EPA rule defining technical and procedural 
requirements for source changes subject to section 112(g) likely will 
not be published until or after the effective date of some State title 
V programs. Under this scenario, these States would be faced with 
implementing section 112(g) without necessarily having sufficient time 
to adopt rules at the State level that specifically implement section 
112(g) and, in some cases, before promulgation of the final section 
112(g) rules. In an effort to avoid unnecessary disruption during this 
period, EPA is clarifying how States can implement and sources may 
comply with all aspects of section 112(g) during the transition period 
occurring as the State adopts the final EPA section 112(g) rule as 
needed or has an alternate program approved under section 112(l). 
Furthermore, the EPA proposes that in order to maintain full approval 
of this part 70 program, States must complete any rulemaking needed to 
implement EPA's final section 112(g) rules on or before 18 months from 
the date of their promulgation.
    Any interim mechanism operating in the transition period must 
produce source requirements that are federally enforceable and 
consistent with those that otherwise would result from implementing the 
promulgated section 112(g) rule (or, prior to promulgation, reasonably 
consistent with the proposed rule). The EPA believes that any 
transition program must provide results consistent with the final rule 
(in substance and process) in order for sources to have reasonable 
certainty regarding their ability to comply with section 112(g) and its 
implementing regulations. Finally, the EPA believes that the transition 
program must be carried out through existing mechanisms which do not 
require significant new investments in time and infrastructure 
development by the State to implement or by sources in order to comply. 
The EPA prefers that these limited resources be instead focused on 
incorporating the requirements of the promulgated rule into existing 
State programs. The EPA believes that there are several potential 
mechanisms available to implement the section 112(g) requirements 
consistent with these criteria.
    First, sources proposing changes that would be subject to section 
112(g) could be issued a part 70 permit which would apply to all 
requirements applicable to the source under the Clean Air Act, 
including those under section 112(g) which apply to the HAP emissions 
from the constructed, reconstructed, or modified unit(s) of a facility 
and the additional facility emission units used in an offset 
demonstration, if any. Using this option could involve some additional 
procedural steps beyond those required in the section 112(g) rules and 
certainly would require more comprehensive permit conditions to be met 
than just those of section 112(g). However, this procedure would not 
necessarily result in large new delays because the Act does require 
expeditious processing of any part 70 permit involving a construction 
activity. In particular, section 503(c) of the Act in discussing the 
required schedule for State processing of operating permit applications 
necessitates:

    Such authority shall establish reasonable procedures to 
prioritized such approval or disapproval actions in the case of 
applications for construction or modification under applicable 
requirements of this Act.

    The EPA interprets this language to require that States prioritized 
operating permit review for all facilities which have been or become 
subject to the requirements of section 112(g). The EPA expects that 
this priority schedule will further ensure timely compliance with the 
promulgated rule.
    The EPA also takes comment on the possibility for States to issue 
specialty title V permits specific to section 112(g) actions. The State 
would determine terms and conditions for such a permit based on the 
promulgated EPA rule for section 112(g). Although the Act in general 
requires that any title V permit must address all applicable 
requirements, EPA believes that exceptions to this rule may be 
possible, but only in exceptional circumstances where the issuance of a 
short-lived single-purpose operating permit is necessary in order to 
implement clear statutory objectives. The EPA takes comment on whether 
this is the case here.
    As a second basic option, the EPA proposes that all State 
Implementation Plans (SIP) approved new source review (NSR) programs 
are authorized to establish Federally enforceable conditions for HAP, 
as well as for criteria pollutants. Clearly, States can develop HAP 
emission limits meeting section 112(g) in the NSR program approved by 
EPA pursuant to section 110(a)(2)(C) as part of the State's overall 
plan to attain and maintain the national ambient air quality standards 
(NAAQS). Moreover, these existing programs either already require or 
can be enhanced in any specific permit issuance situation to meet a 
process equivalent to that which is likely to be required in the final 
section 112(g) rules. Extending this recognition directly to all HAPs 
would be based in part on the general authority of section 112(l) to 
recognize State toxic control programs to establish Federally 
enforceable requirements, and in part on EPA's general authority under 
section 301 to ``prescribe such regulations as are necessary to carry 
out its functions'' under the Act. The EPA believes that such 
recognition is consistent with both the need for a workable transition 
program for section 112(g) and the reasonable need for a mechanism to 
allow sources to develop conditions limiting their potential to emit 
for HAP and their potential applicability to section 112.
    Under this approach, the EPA would presumptively approve a State 
preconstruction program pursuant to section 112(l) to the extent that 
it would be used to meet criteria equivalent to those provided for in 
this proposal for notice of case-by-case MACT approvals. That is, EPA 
would, under section 112(l) authority, presume acceptance and use of 
this mechanism in conjunction with the approval of the part 70 program 
for a State, unless the State requests otherwise. Accordingly, in most 
States, no formal rulemaking pursuant to proposed Subpart E would be 
required. States working with this approach could then use the State's 
NSR permitting mechanism to develop appropriate limits under adequate 
procedures to meet the EPA's final rules to implement section 112(g).
    The EPA requests comments on all aspects of the proposed approach 
to establishing workable section 112(g) procedures in the interim 
period before promulgation of the final section 112(g) rule and, in 
some cases, before requisite new authorities are obtained by State 
authorities. In particular, the EPA requests comments on the need for 
and appropriate maximum length of this interim period.
    Finally, the EPA also requests comment on an alternative approach 
to facilitate implementation of section 112(g) requirements during this 
interim period. Under this approach, during this interim period only, 
the EPA would recognize any applicable limits that are State-
enforceable as adequate to meet the requirements of the statute 
concerning section 112(g).

VI. Administrative Requirements

A. Executive Order 12866

    A regulatory impact analysis (RIA) was prepared for the proposed 
regulation. The RIA was prepared under the guidelines outlined in 
Executive Order 12866 and submitted to the Office of Management and 
Budget (OMB) even though the proposed regulation is not expected to be 
``significant'' as defined in the Order. The regulation is not expected 
to have an annual effect on the economy of $100 million or more; it is 
not expected to cause a major increase in costs or prices to society; 
and it is not expected to cause significant adverse effects on 
competition. The objective of the RIA is to evaluate the costs and 
benefits associated with the proposed regulation. A further objective 
of the RIA is to show that the proposed regulation will maximize net 
benefits to society.
    It is difficult to address with precision the cost impacts of 
section 112(g) of the Act. The calculation of future impacts requires: 
(1) Data on the number of major sources of hazardous air pollutants in 
the United States, (2) predictions of the number of changes that would 
occur at these major sources that would trigger section 112(g), and (3) 
estimates of the average impacts each such change will experience if 
section 112(g) were to require early MACT controls and (4) estimates of 
the degree to which a typical construction, reconstruction or 
modification subject to section 112(g) would already be subject to a 
promulgated MACT standard or State and local requirement. Because great 
uncertainty exists in each of these areas, the RIA estimates for the 
various scenarios of section 112(g) approaches are intended to 
``bracket'' the range of possible impacts, rather than to predict with 
precision the differences that may occur under various policy options.
    The impacts (cost and emission reductions) of the section 112(g) 
program are assumed to begin in 1993 and increase as additional state 
agencies are subject to the program. Impacts are expected to extend to 
the year 2002.
    The annual cost impacts are expected to reach a peak cost between 
1995 and 1996 and decrease as more major sources become subject to 
Federal MACT standards issued under section 112(d). Scenario 1, which 
is used to illustrate a program with wide applicability and which 
creates little opportunity for obtaining offsets, would result in the 
greatest cost. Scenario 1 is estimated to result in an average annual 
cost of $96 million. Scenarios 2 and 3, involving intermediate 
coverage, would result in approximate annual control costs of $25 to 
$28 million, respectively. Scenario 4, which is used to illustrate a 
program with minimal coverage, would result in costs of about $9 
million annually. These cost estimates do not include estimates of the 
opportunity cost associated with requiring approval of a proposed 
offset prior to operation and the pre-construction review requirements 
for case-by-case MACT determination. The EPA plans to develop a rough 
estimate of these costs in developing the final RIA, and solicits 
comments and quantitative information on the nature and magnitude of 
these costs. The scope and extent of analysis on these issues will be 
constrained by the EPA's budgetary limitations.
    The absence of valuation and sufficient exposure-response 
information precludes a quantitative benefits analysis. However, there 
was an attempt made to determine the benefits that would be needed to 
justify general program directions. The EPA believes it is probable 
that the benefits of the proposed regulation are generally consistent 
with the predicted impacts.
    The proposal identifies at several places key alternative options 
for implementing section 112(g). In developing a final rule, the EPA 
intends to analyze the costs and emission reductions associated with a 
few of the most important variables associated with these options. 
Examples of possible issues include costs associated with the 
alternative definitions of ``construct,'' and opportunity costs of up-
front reviews and offset restrictions. The EPA solicits comments on 
which such issues should be included, as well as comments and data on 
these issues that would support this analysis. The scope and extent of 
analysis on these issues will be constrained by the EPA's budgetary 
limitations.

B. Regulatory Flexibility Act

    The Regulatory Flexibility Act (5 U.S.C. 601 et seq.) requires the 
EPA to consider potential impacts of proposed regulations on small 
entities. If a preliminary analysis indicates that a proposed 
regulation would have a significant economic impact on a substantial 
number of small entities, then a regulatory flexibility analysis must 
be prepared.
    The Regulatory Act guidelines applicable to the proposed regulation 
indicate that an economic impact should be considered significant if it 
meets one of the following criteria: (1) Compliance increases annual 
production costs by more than 5 percent, assuming costs are passed on 
to consumers; (2) compliance costs as a percentage of sales for small 
entities are at least 10 percent more than compliance costs as a 
percentage of sales for large entities; (3) capital costs of compliance 
represent a ``significant'' portion of capital available to small 
entities, considering internal cash flow plus external financial 
capabilities; or (4) regulatory requirements are likely to result in 
closures of small entities. A substantial number of small entities is 
generally considered to be more than 20 percent of the small entities 
in the affected industry. An analysis was conducted to assess the 
economic impacts associated with the proposed regulation. Although the 
number of facilities affected by section 112(g) can be identified, 
information on the sizes of these facilities is unavailable. However, 
it is expected that the control cost each facility may face will be 
small compared to the average annual revenues in the industries 
affected by section 112(g). Therefore, it is not expected that the 
impacts of section 112(g) will be considered significant as defined 
above.
    This regulation does not affect a significant number of small 
businesses, small governmental jurisdictions, or small institutions. 
Pursuant to the provisions of 5 U.S.C. 605(b), I hereby certify that 
this proposed rule, if promulgated, will not have a significant 
economic impact on a substantial number of small entities.

C. Paperwork Reduction Act

    The information collection requirements in this proposal have been 
submitted for approval to OMB under the Paperwork Reduction Act, 44 
U.S.C. 3501 et seq. An information collection request (ICR) document 
has been prepared by the EPA (ICR No. 1658.01) and a copy may be 
obtained from Sandy Farmer, Information Policy Branch (PM-2234), U.S. 
Environmental Protection Agency, 401 M Street SW., Washington, DC 20460 
or by calling (202) 260-2740.
    The EPA prepared estimates of the average annual burden hours 
needed to collect and prepare information required under section 112(g) 
for each of the regulatory scenarios examined in preparing the 
Regulatory Impact Analysis (see discussion above). The burden estimates 
presented below are an accumulation of the estimated annual burden 
hours that would be experienced by industry respondents, state and 
local agencies, and EPA under the various regulatory scenarios.
    The average annual burden-hours that would be required under 
Scenario 1, which was used to illustrate a wide-reaching program, is 
approximately 1,823,000 hours. Scenario 2 and 3, used to illustrate a 
program with intermediate coverage, would result in an average annual 
burden of approximately 460,000 and 390,000 hours, respectively. 
Scenario 4, a program with minimal coverage, would result in an average 
burden of approximately 228,000 hours. The estimates for Scenario 2 and 
3 are probably the most consistent with the provisions of the proposed 
rule.
    Send comments regarding these burden estimates or any other aspect 
of this collection of information, including suggestions for reducing 
the burden to Chief, Information Policy Branch (PM-223Y); U.S. EPA, 401 
M Street, SW., Washington, DC 20460; and to the Office of Information 
and Regulatory Affairs, Office of Management and Budget, Washington, DC 
20503, marked ``Attention: Desk Officer for EPA.'' The final rule will 
respond to any OMB or public comments on the information collection 
requirements contained in this proposal.

VII. Suggest Format for Comments

    There are a relatively large number of issues for which the EPA is 
requesting comment on the proposed rule. Accordingly, the EPA expects a 
significant effort will be required in responding to those comments. 
The EPA would appreciate efforts by commenters to follow the general 
outline contained in Table 7. Additionally, the EPA would appreciate, 
if possible, that comments generated using a word processing software 
should be sent on a clearly labeled 3.5 inch IBM-compatible diskette. 
Comments formatted in WordPerfect 5.0 or 5.1 may be submitted as is; 
comments prepared by other word processing software should be submitted 
in an ``unformatted'' mode. Comments should refer to page numbers and 
columns whenever possible, and should cross-reference the issue number 
in Table 7.

Statutory Authority

    The statutory authority for this proposal is provided by sections 
101, 112, 114, 116, and 301 of the Clean Air Act, as amended; 42 
U.S.C., 7401, 7412, 7414, 7416, and 7601.

                  Table 7.--Suggested Format for Comments on the Proposed Section 112(g) Rule                   
----------------------------------------------------------------------------------------------------------------
                                      Citation in         Citation in                                           
      Topic and Issue No.             regulation           preamble                       Issues                
----------------------------------------------------------------------------------------------------------------
I. General Applicability                                                                                        
 Issues:                                                                                                        
    1..........................  ....................  II.B.3            Requirements for guidance under section
                                                                          112(g).                               
    2..........................  63.40(b)............  III.A.2.a         Grandfathering of projects already     
                                                                          underway.                             
    3..........................  63.40(b)............  III.A.2.b         SIC codes inclusion in definition of   
                                                                          major source.                         
    4..........................  63.40(c)............  III.A.3           Exclusion for Steam generating units.  
II. Requirements for                                                                                            
 Constructed and Reconstructed                                                                                  
 Sources:                                                                                                       
    5..........................   63.42..............  III.C.2           Definitions of constructed and         
                                                                          reconstructed sources.                
    6..........................  63.42...............   III.C.2          Alternative readings of the Act        
                                                                          regarding new equipment at an existing
                                                                          plant site. Preference for Alternative
                                                                          A vs. Alternative B.                  
    7..........................  63.42...............  III.C.2           Whether definition of emission unit    
                                                                          should be more prescriptive under     
                                                                          Alternative A.                        
    8..........................  63.42...............  III.C.2           Definition of ``greenfield site''.     
    9..........................  63.42...............  III.C.2           Whether Act would support a reading    
                                                                          treating all new equipment at a major 
                                                                          source plant site as ``construction''.
    10.........................  63.42...............  III.C.2           Other possible definitions of          
                                                                          ``construct''.                        
    11.........................  63.42...............  III.C.3           Definition of reconstruct a major      
                                                                          source.                               
III. Requirements for Modified                                                                                  
 Major Sources:                                                                                                 
    12.........................  63.43(b)(3).........  III.D.2           ``Modification'' and emission points   
                                                                          affected by the modification.         
    13.........................  63.43(b)(3).........  III.D.2           Other approaches to MACT coverage,     
                                                                          including plant-wide coverage or      
                                                                          subdividing plant site into major-    
                                                                          emitting emission units.              
    14.........................  63.43(c)(1) through   III.D.3           Exclusion issues (except for           
                                  63.43(c)(4) and                         alternative raw materials and         
                                  63.43(c)(7).                            operations).                          
    15.........................  63.41, 63.43(c)(5)    III.D.3           Definition of ``Operations that the    
                                  and (6).                                major source is designed to           
                                                                          accommodate.'' Permit-allowable       
                                                                          framework for raw material            
                                                                          substitutions/alternative operating   
                                                                          scenarios.                            
    16.........................  63.43(c)(6).........  III.D.3           Whether operational changes allowed in 
                                                                          State permits can be incorporated into
                                                                          Title V permits without triggering    
                                                                          section 112(g).                       
    17.........................  63.43(d)............  III.D.4           Proposed method for determining actual 
                                                                          emissions. NSPS vs. PSD approach.     
    18.........................  63.43(e), 63.46       III.D.4           Whether offsetting should be restricted
                                  63.47.                                  for VOC reductions used for control of
                                                                          ozone or credited under PSD/NSR.      
    19.........................  63.43(f)............  III.D.5           Proposed treatment of increases and    
                                                                          decreases of the same pollutant.      
IV. De Minimis Values:                                                                                          
    20.........................  63.44...............  III.E.3(a)        Adjustment for 7-year exposure period, 
                                                                          including specific alternatives.      
    21.........................  63.44...............  III.E.3(a)        Model plant assumptions.               
    22.........................  63.44...............  III.E.3(a)        Alternative approach involving 2       
                                                                          tables, to account for site-specific  
                                                                          conditions.                           
    23.........................  63.44...............  III.E.3(a)        10 ton cap on de minimis values.       
    24.........................  63.44...............  III.E.3(a)        Default value of 1 ton/year de minimis 
                                                                          for carcinogens with no potency       
                                                                          estimate.                             
    25.........................  63.44...............  III.E.3(b)        Composite score approach to setting de 
                                                                          minimis levels for threshold          
                                                                          pollutants with no RfC.               
    26.........................  63.44...............  III.E.3(b)        5 tons/year default de minimis levels  
                                                                          for pollutants lacking data.          
    27.........................  63.44...............  III.E.3(c)        Methodology and values described in    
                                                                          preamble for setting de minimis rates 
                                                                          for pollutants to reflect concerns    
                                                                          regarding short-term exposure.        
    28.........................  63.44...............  III.E.4           Cap of 0.01 tons per year for Great    
                                                                          Waters pollutants; other alternatives 
                                                                          such as 0.1 tons per year or 10% of   
                                                                          health-based values.                  
    29.........................  63.44...............  III.E.4           Proposed approach to de minimis level  
                                                                          determination for polycyclic organic  
                                                                          matter (POM).                         
    30.........................  63.44...............  III.E.4           Comment on treatment of POM under      
                                                                          section 112(k) and section 112(c)(6). 
    31.........................  63.44...............  III.E.6           One significant figures approach vs.   
                                                                          alternative ``binning'' of de minimis 
                                                                          values.                               
    32.........................  63.44...............  III.E.6           Whether the de minimis definition      
                                                                          should consider a specified           
                                                                          concentration in a product or mixture.
    33.........................  63.44...............  III.E.7           Concept of de minimis index for        
                                                                          addition of assessing contribution of 
                                                                          multiple pollutants toward a de       
                                                                          minimis level.                        
    34.........................  63.44...............  III.E.9           De minimis levels for radionuclides.   
V. MACT Determinations:                                                                                         
    35.........................  63.41...............  III.B.8           Inclusion of additional emission points
                                                                          in the definition of ``MACT-affected  
                                                                          emission unit''.                      
    36.........................  63.45(b)(2) and       III.F.1           Process outlined in Figure 7 pertaining
                                  63.45(f).                               to 112(g) requirements after MACT     
                                                                          standard is issued. Notification      
                                                                          requirements; alternative             
                                                                          requirements, e.g. 30-60 days before  
                                                                          startup.                              
    37.........................  63.45...............  III.F.2           Legal reading of the act regarding     
                                                                          preconstruction review of MACT        
                                                                          determinations.                       
    38.........................  63.45...............  III.F.2           Alternative approach requiring pre-    
                                                                          operation review.                     
    39.........................  63.45...............  III.F.3           MACT floor determination. Inclusion of 
                                                                          sources outside the U.S.              
    40.........................  63.45, MACT           III.F.3           Treatment of MACT floor in the         
                                  Guidelines.                             regulation, including suggested       
                                                                          methods in MACT guidelines document.  
    41.........................  63.45...............  III.F.3           Design and use of MACT Database;       
                                                                          feasibility of reliance on data base  
                                                                          for MACT floor information.           
    42.........................  63.45...............  III.F.3           Cut-off date for ``available           
                                                                          information.''                        
    43.........................  63.45...............  III.F.4           Approaches to subcategorization.       
    44.........................  63.45...............  III.F.6           Relationship between Title III and     
                                                                          Title V, particularly whether section 
                                                                          112(g) requirements can be added to   
                                                                          the permit post-operation.            
    45.........................  63.45(i)............  III.F.8           Public review as a prerequisite to     
                                                                          federal enforceability of case-by-case
                                                                          MACT determinations.                  
    46.........................  63.45(g)(7).........  III.F.8           Use of section 112(l) to approve       
                                                                          alternative State administrative      
                                                                          requirements for review of case-by-   
                                                                          case MACT determinations.             
    47.........................  63.45(h)............  III.F.9           Whether MACT approval should expire if 
                                                                          failure to construct occurs after 18  
                                                                          months.                               
VII. Offset Demonstrations:                                                                                     
    48.........................  63.47...............  III.G.2           Restriction on shutdowns and           
                                                                          curtailments in simplified offset     
                                                                          demonstration.                        
    49.........................  63.46, 63.47........  III.G.2           Requirement for pre-operation review of
                                                                          offsets. Alternative approaching      
                                                                          allowing for review post-operation.   
    50.........................  63.46, 63.47........  III.G.2(c)        General comments on the two approaches 
                                                                          ``contemporaneous'' and ``simplified''
                                                                          including whether one ought not be    
                                                                          included in the final rule.           
    51.........................  63.46, 63.47........  III.G.3           Whether offset determinations should   
                                                                          comply with certain provisions in 40  
                                                                          CFR part 70 or enhanced monitoring    
                                                                          provisions pursuant to section 114 of 
                                                                          the Act.                              
    52.........................  63.46, 63.47........  III.G.3           Additional language to facilitate      
                                                                          consideration of source reduction     
                                                                          programs.                             
    53.........................  63.49...............  III.H             Compliance extension for emission units
                                                                          subject to subsequent MACT standards. 
                                                                          Also, whether section 112(g) MACT is  
                                                                          sufficient for subsequent section     
                                                                          112(j) requirements.                  
VIII. Proposed Approach for                                                                                     
 Demonstrating That Offsets Are                                                                                 
 ``More Hazardous'':                                                                                            
    54.........................  63.48...............  IV.A              Two alternative readings of section    
                                                                          112(g)(1)(A). More hazardous          
                                                                          ``pollutant'' vs. more hazardous      
                                                                          ``quantity.''                         
    55.........................  63.48...............  IV.B              Whether overall approach to offsetting 
                                                                          restrictions strikes an appropriate   
                                                                          balance given legal, policy,          
                                                                          scientific and practical              
                                                                          considerations.                       
    56.........................  63.48...............  IV.C.1            Creation of the ``high-concern''       
                                                                          category of pollutants.               
    57.........................  63.48...............  IV.C.2            Relative hazard between categories of  
                                                                          pollutants.                           
    58.........................  63.48...............  IV.C.3(a)         Other endpoints and pollutants which   
                                                                          qualify a pollutant to be considered  
                                                                          ``non-threshold''.                    
    59.........................  63.48...............  IV.C.4(b)         Use of ED10 as a potency estimate for  
                                                                          carcinogens in the hazard ranking.    
    60.........................  63.48...............  IV.C.4(b)         Treatment of carcinogens without       
                                                                          potency estimates with regard to      
                                                                          relative hazard ranking.              
    61.........................  63.48...............  1V.C.5.(b)        Matrix approach to the hazard ranking. 
    62.........................  63.48...............  IV.C.6.(b)        Methodologies that could be used to    
                                                                          identify and rank pollutants of       
                                                                          concern for severe toxicity from short-
                                                                          term exposures.                       
    63.........................  63.48...............  IV.C.7            Treatment of ``unrankable'' pollutants.
    64.........................  63.48...............  IV.C.8            Treatment of the 17 HAPs listed under  
                                                                          section 112(b) as chemical groupings. 
                                                                          Selection of subgroupings included in 
                                                                          the ranking. Methodology for ranking  
                                                                          them.                                 
    65.........................  63.48...............  IV.D.1(a)         Relative hazard determination between  
                                                                          carcinogens: treatment of uncertainty 
                                                                          of potency estimates.                 
    66.........................  63.48...............  IV.D.1(b)         Potential hazard reduction when a group
                                                                          A carcinogen is offset by a group C   
                                                                          carcinogen of higher potency.         
    67.........................  63.48(b)............  IV.D.2            Treatment and consideration of         
                                                                          uncertainty in composite score        
                                                                          determination of relative hazard.     
IX. Other Hazard Ranking                                                                                        
 Issues:                                                                                                        
    68.........................  ....................  IV.G.1            Consideration of non-inhalation hazard 
                                                                          from hazardous air pollutants.        
    69.........................  ....................  IV.G.1            How to consider explicitly information 
                                                                          on bioconcentration and persistence to
                                                                          place threshold pollutants into the   
                                                                          ``high-concern'' pollutant list and   
                                                                          apply further offsetting restrictions.
    70.........................  ....................  IV.G.2            Half-life consideration in the hazard  
                                                                          ranking.                              
    71.........................  ....................  IV.G.2            Treatment of reactive transformation   
                                                                          products of hazardous air pollutants. 
    72.........................  ....................  IV.G.3            Appeal process to address errors in a  
                                                                          potential offset.                     
    73.........................  ....................  IV.G.4            Treatment of non-cancer effects of     
                                                                          carcinogens: possible approaches.     
    74.........................  ....................  IV.G.4            Comment on taking the spectrum of      
                                                                          health effects directly into account  
                                                                          for offset determinations and the     
                                                                          policy and scientific options         
                                                                          considered when little or no health   
                                                                          effects data are available.           
    75.........................  ....................  IV.G.5            Development of scientifically          
                                                                          defensible system of weighted offsets.
    76.........................  ....................  IV.G.5            Approaches for providing updates to the
                                                                          ranking.                              
X. Discussion of the                                                                                            
 Relationship of the Proposed                                                                                   
 Requirements to Other Parts of                                                                                 
 the Act:                                                                                                       
    77.........................  ....................  V.A               Relationship of Section 112(g)         
                                                                          implementation to Title V program     
                                                                          approval.                             
    78.........................  ....................  V.B               Whether a State program deleting       
                                                                          offsets could be considered an        
                                                                          adjustment under section 112(l) of the
                                                                          act and added to 40 CFR 63.92.        
    79.........................  ....................  V.B               Whether, in States with existing State 
                                                                          rules which would make offsets        
                                                                          unusable, that program can be approved
                                                                          under section 112(l) without offsets. 
    80.........................  ....................  V.D               Relationship between 112(g), 112(d) and
                                                                          112(j).                               
    81.........................  ....................  V.E               General provisions: additions to them  
                                                                          in 112(g) rule.                       
    82.........................  ....................  V.F               Possibility for States to issue        
                                                                          specialty title V permits specific to 
                                                                          section 112(g) actions.               
    83.........................  ....................  V.F               112(g) procedure in the interim period 
                                                                          before promulgation of the final      
                                                                          section 112(g) rule: need for and     
                                                                          appropriate maximum length of the     
                                                                          interim period.                       
    84.........................  ....................  V.F               Feasibility of approaches described for
                                                                          implementation of section 112(g)      
                                                                          requirements during interim period.   
XI. Analysis Relative to                                                                                        
 Federal Administrative                                                                                         
 Requirements:                                                                                                  
    85.........................  ....................  VI.A              Comments on regulatory impact analysis,
                                                                          including any additional analysis that
                                                                          could be prepared for the final rule. 
    86.........................  ....................  VI.C              Comments relative to paperwork         
                                                                          reduction act, including comments on  
                                                                          burden analysis.                      
----------------------------------------------------------------------------------------------------------------

List of Subjects

40 CFR Part 63

    Environmental protection, Administrative practice and procedure, 
Air pollution control, Hazardous substances, Intergovernmental 
relations, Reporting and recordkeeping requirements.

40 CFR Part 70

    Administrative practice and procedure, Air pollution control, 
Intergovernmental relations.


    Dated: February 28, 1994.
Carol M. Browner,
Administrator.
[FR Doc. 94-7241 Filed 3-31-94; 8:45 am]
BILLING CODE 6560-50-P