[Federal Register Volume 61, Number 18 (Friday, January 26, 1996)]
[Rules and Regulations]
[Pages 2428-2438]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 96-1413]
-----------------------------------------------------------------------
ENVIRONMENTAL PROTECTION AGENCY
40 CFR Part 52
[IL99-2-7003, IN46-2-7004, MI33-2-7005, WI47-2-7006; FRL-5402-8]
Approval of a Section 182(f) Exemption; Illinois, Indiana,
Michigan, and Wisconsin
AGENCY: Environmental Protection Agency (EPA).
ACTION: Final rule.
-----------------------------------------------------------------------
SUMMARY: As requested by the States of Illinois, Indiana, Michigan, and
Wisconsin in a July 13, 1994 submittal pursuant to section 182(f)(3) of
the Clean Air Act (CAA or the Act), the EPA is granting exemptions from
the Reasonably Available Control Technology (RACT) and New Source
Review (NSR) requirements for major stationary sources of Oxides of
Nitrogen (NOX) and from vehicle Inspection/Maintenance (I/M) and
general conformity requirements for NOX for ozone nonattainment
areas within the Lake Michigan Ozone Study (LMOS) modeling domain,
which includes portions of the States of Illinois, Indiana, Michigan,
and Wisconsin. The EPA is also granting exemptions from transportation
conformity requirements for NOX for ozone nonattainment areas
classified as marginal or transitional within the LMOS modeling domain.
The EPA is approving the exemptions based on a demonstration that
additional NOX reductions would not contribute to attainment of
the National Ambient Air Quality Standard (NAAQS) for ozone within the
LMOS modeling domain. The EPA is not taking final action at this time
on the granting of exemptions from the transportation conformity
requirements of the CAA for ozone nonattainment areas classified as
moderate or above in the LMOS modeling domain. The continued approval
of these exemptions is contingent on the results of the final ozone
attainment demonstrations and plans. These plans are expected to be
submitted by mid-1997 and to incorporate the results of the Ozone
Transport Assessment Group process. The attainment plans will supersede
the initial modeling information which is the basis for the waiver EPA
is granting in this document. To the extent the attainment plans
include NOX controls on certain major stationary sources in the
LMOS ozone nonattainment areas, EPA will remove the NOX waiver for
those sources. To the extent the final plans achieve attainment of the
ozone standard without additional NOX reductions from certain
sources, the NOX emissions control exemption would continue for
those sources. EPA's rulemaking action to reconsider the initial
NOX waiver may occur simultaneously with rulemaking action on the
attainment plans.
DATES: This final rule will be effective February 26, 1996.
ADDRESSES: Copies of the exemption request, public comments and EPA's
responses are available for inspection at the following address: United
States Environmental Protection Agency, Region 5, Air and Radiation
Division, 77 West Jackson Boulevard, Chicago, Illinois 60604.
FOR FURTHER INFORMATION CONTACT: Edward Doty, Regulation Development
Section (AR-18J), Regulation Development Branch, Air and Radiation
Division, United States Environmental Protection Agency, Region 5, 77
West Jackson Boulevard, Chicago, Illinois 60604, Telephone Number (312)
886-6057.
SUPPLEMENTARY INFORMATION:
I. Background Information
On July 13, 1994, the States of Illinois, Indiana, Michigan, and
Wisconsin submitted a petition to the EPA requesting that the ozone
nonattainment areas within the LMOS modeling domain be exempted from
requirements to implement NOX controls pursuant to section 182(f)
of the Act. The exemption request is based on modeling demonstrating
that additional NOX emission controls within the nonattainment
areas will not contribute
[[Page 2429]]
to attainment of the ozone NAAQS within the LMOS modeling domain.
On March 6, 1995, EPA published a rulemaking proposing approval of
the NOX exemption petition and specifically identifying the
Counties or areas covered by the exemption. During the 30 day public
comment period, EPA received a number of comments favoring or objecting
to the proposed approval. In addition to these comments, the EPA also
received adverse comments objecting to any NOX control waiver
within the United States, with the commenters requesting that these
comments be addressed in all EPA rulemakings dealing with such emission
control waivers.
II. Public Comments
The following discussion summarizes the comments received regarding
the States' petition and/or EPA's proposed rulemaking and presents
EPA's responses to these comments.
Comment: A number of comments supporting the proposed rulemaking
were received from organizations representing various industrial
groups, local planning organizations, and the States themselves. One
commenter, who generally supported the proposed rulemaking, noted that
the EPA proposed to reverse its decision on the petition if subsequent
modeling results supported such a reversal. The commenter raised a
concern that the EPA should only reverse its decision to approve the
petition if well documented modeling results are available clearly
indicating the need for such a reversal.
Response: The favorable comments support the logic used in the
proposed rulemaking.
With regard to the concern over the quality of the modeling results
needed to reverse this decision, it should be noted that such modeling
results will be well documented and are expected to be based on
validated modeling. The States involved in the LMOS are conducting a
number of additional modeling analyses (subsequent to the preparation
of the NOX waiver request) to assess the impacts of emission
controls on peak ozone concentrations and on ozone concentrations
transported out of the modeling domain (long range ozone transport has
become a significant issue in the development of ozone demonstrations
of attainment in the eastern United States). Additional modeling
analyses are required to support the States' demonstrations of
attainment, which have not been completed. These modeling analyses are
well documented and are now based on a modeling system which has been
accepted by the EPA as being validated for the LMOS modeling domain.
Any conclusion showing the need for NOX controls will be well
supported by the modeling.
It should be noted that the modeling used to support the NOX
waiver petition was not based initially on validated modeling. The
modeling system and its base year inputs were modified to a validated
form subsequent to the submittal of the petition. Nonetheless, the
``signals'' of the modeling results regarding Volatile Organic Compound
(VOC) controls versus NOX controls have not changed with the
validation of the modeling system. The modeling results continue to
show that NOX emission controls in the ozone nonattainment areas
will not contribute to reduction of peak ozone levels within the LMOS
modeling domain, and may actually increase peak ozone levels near the
major urban areas.
Comment: A commenter, who supports the proposed NOX exemption,
considers the exemption, through section 182(f), to also increase the
major source threshold relating to federal operating permit programs
from 25 tons/year (tpy) to 100 tpy (this comment is assumed to apply to
the ozone nonattainment areas classified as severe).
Response: The commenter is correct. Based on guidance contained in
40 CFR Part 70.2 (subparagraph (3)(1) under the ``major source''
definition), the major source threshold for federal operating programs
would be revised to 100 tpy, potential to emit, in the areas covered by
the NOX waiver. In addition, for new source considerations, it
should be noted that the waived areas should be considered to be
covered by Prevention of Significant Deterioration requirements, with a
control source size threshold of 250 tpy, potential to emit, for
NOX rather than by nonattainment area new source requirements.
Comment: A commenter notes that, in addition to modeling data
supporting approval of the petition, monitoring data were collected
during the 1991 LMOS field study which also support the approval of the
NOX waiver. The combination of modeling data and monitoring data
meet the requirements for a section 182(f) exemption specified in EPA's
guidance documents titled: ``State Implementation Plan; Nitrogen Oxides
Implementation of Title I of the Clean Air Act Amendments of 1990'' (57
FR 55628, November 25, 1992); and the ``Guideline for Determining
Applicability of Nitrogen Oxide Requirement under Section 182(f)''
(December 1993).
Response: Although the commenter did not specifically reference the
data from which this conclusion was drawn, EPA acknowledges that data,
such as concentrations of non-methane hydrocarbons and NOX and
derived/monitored ozone production potentials of air parcels, collected
for the urban source areas during the 1991 field study support the
approval of the NOX waiver. It is noted, however, that the primary
basis for the approval of the N0x waiver is the modeling results
submitted in support of the waiver. The 1991 field data by themselves
may not be an adequate support for the waiver since these data are
limited in nature and do not present a complete picture of the impacts
of NOX controls on LMOS modeling domain peak ozone concentrations.
Comment: Commenters argue that NOX exemptions are provided for
in two separate parts of the Act, in sections 182(b)(1) and 182(f).
Because the NOX exemption tests in sections 182(b)(1) and
182(f)(1) include language indicating that action on such requests
should take place ``when [EPA] approves a plan or plan revision,''
these commenters conclude that all NOX exemption determinations by
the EPA, including exemption actions taken under the petition process
established by section 182(f)(3), must occur during consideration of an
approvable attainment or maintenance plan, unless the area has been
redesignated as attainment. The commenters also argue that even if the
petition procedures of section 182(f)(3) may be used to relieve areas
of certain NOX requirements, exemptions from the NOX
conformity requirements must follow the process provided in section
182(b)(1), since this is the only provision explicitly referenced by
section 176(c), the Act's conformity provisions.
Response: Section 182(f) contains very few details regarding the
administrative procedures for acting on NOX exemption requests.
The absence of specific guidelines by Congress leaves the EPA with
discretion to establish reasonable procedures consistent with the
requirements of the Administrative Procedure Act (APA).
The EPA disagrees with the commenters regarding the process for
considering NOX exemption requests under section 182(f), and
instead believes that sections 182(f)(1) and 182(f)(3) provide
independent procedures by which the EPA may act on NOX exemption
requests. The language in section 182(f)(1), which indicates that the
EPA should act on NOX exemptions in conjunction with action on a
plan or a plan revision, does not appear in section 182(f)(3). While
section 182(f)(3) references section
[[Page 2430]]
182(f)(1), the EPA believes that this reference encompasses only the
substantive tests in paragraph (1) [and by extension, paragraph (2)],
not the procedural requirement that the EPA act on exemptions only when
acting on State Implementation Plans (SIPs). Additionally, section
182(f)(3) provides that ``person[s]'' [which section 302(e) of the Act
defines to include States] may petition for NOX exemptions ``at
any time,'' and requires the EPA to make its determination within six
months of the petition's submission. These key differences lead EPA to
believe that Congress intended the exemption petition process of
paragraph (3) to be distinct and more expeditious than the longer plan
revision process intended under paragraph (1).
With respect to major stationary sources, section 182(f) requires
States to adopt NOX RACT and NSR rules, unless exempted. These
rules were generally due to be submitted to the EPA by November 15,
1992. Thus, in order to avoid the CAA sanctions, areas seeking a
NOX exemption would have needed to submit this exemption request
for EPA review and rulemaking action several months before November 15,
1992. In contrast, the CAA specifies that the attainment demonstrations
were not due until November 1993 or 1994 (and EPA may take 12 to 18
months to approve or disapprove the demonstrations). For marginal ozone
nonattainment areas (subject to NOX NSR), no attainment
demonstrations are called for in the CAA. For areas seeking
redesignation to attainment of the ozone NAAQS, the CAA does not
specify a deadline for submittal of maintenance demonstrations (in
reality, EPA would generally consider redesignation requests without
accompanying maintenance plans to be unacceptable). Clearly, the CAA
envisions the submittal of and EPA action on NOX exemption
requests, in some cases, prior to submittal of attainment or
maintenance demonstrations.
With respect to the comment that section 182(b)(1) is the
appropriate authority for granting interim-period transportation
conformity NOX exemptions, EPA agrees with the commenters and has
published an interim final rule that changes the transportation
conformity rule's reference to section 182(b)(1) as the correct
authority under the Act for waiving the NOX build/no-build and
less-than-1990 emissions tests for certain areas. See 60 FR 44762 (A
related proposed rule, 60 FR 44790, published on the same day, invited
public comment on how the Agency plans to implement section 182(b)(1)
transportation conformity NOX exemptions. That proposal has been
subsequently finalized. See 60 FR 57179). However, EPA also notes that
section 182(b)(1), by its terms, only applies to moderate and above
ozone nonattainment areas. Consequently, EPA believes that the interim-
reductions requirements of section 176(c)(3)(A)(iii), and hence the
authority provided in section 182(b)(1) to grant relief from those
interim-reduction requirements, apply only with respect to those areas
that are subject to section 182(b)(1). EPA intends to continue to apply
the transportation conformity rule's build/no-build and less-than-1990
emissions tests for purposes of implementing the requirements of
section 176(c)(1), and EPA intends to continue to provide relief from
those requirements under section 182(f). In addition, because general
federal actions are not subject to section 176(c)(3)(A)(iii), which
explicitly references section 182(b)(1), EPA will also continue to
offer relief under section 182(f)(3) from the applicable NOX
requirements of the general conformity rule.
In order to demonstrate conformity, transportation-related federal
actions that are taken in ozone nonattainment areas not subject to
section 182(b)(1) and, hence, not subject to section 176(c)(3)(A)(iii)
must still be consistent with the criteria specified under section
176(c)(1). Specifically, these actions must not, with respect to any
standard, cause or contribute to new violations, increase the frequency
or severity of existing violations, or delay attainment. In addition,
such actions must comply with the relevant requirements and milestones
contained in the applicable state implementation plan, such as
reasonable further progress schedules, assumptions specified in the
attainment or maintenance demonstrations, numerical emission limits, or
prohibitions. EPA believes that the build/no-build and less-than-1990
emissions tests provide an appropriate basis for such areas to
demonstrate compliance with the above criteria.
As noted earlier, EPA intends to continue to offer relief under
section 182(f) from the interim NOX requirements of the conformity
rules that would apply under section 176(c)(1) for the areas not
subject to section 182(b)(1) in the manner described above. EPA
believes this approach is consistent both with the way NOX
requirements in ozone nonattainment areas are treated under the Act
generally, and under section 182(f) in particular. The basic approach
of the Act is that NOX reductions should apply when beneficial to
an area's attainment goals, and should not apply when unhelpful or
counterproductive. Section 182(f) reflects this approach but also
includes specific substantive tests which provide a basis for EPA to
determine when NOX requirements should not apply. There is no
substantive difference between the technical analysis required to make
an assessment of NOX impacts on attainment in a particular area
whether undertaken with respect to mobile source or stationary source
NOX emissions. Moreover, where EPA has determined that NOX
reductions will not benefit attainment or would be counterproductive in
an area, the EPA believes it would be unreasonable to insist on
NOX reductions for purposes of meeting reasonable further progress
or other milestone requirements. Thus, even as to the conformity
requirements of section 176(c)(1), EPA believes it is reasonable and
appropriate, first, to offer relief from the applicable NOX
requirements of the general and transportation conformity rules in
areas where such reductions would not be beneficial and, second, to
rely in doing so based on the exemption tests provided in section
182(f).
For moderate and above ozone nonattainment areas which are relying
on modeling data in petitioning for a transportation conformity
NOX exemption, the proposed change affects the process for
applying for such waivers. Unlike section 182(f)(3), section 182(b)(1)
requires that EPA approve a NOX waiver (i.e., determine that
additional reductions of NOX would not contribute to attainment)
as part of a SIP revision. Thus, under section 182(b)(1), petitions for
transportation conformity NOX waivers for areas subject to that
section must be submitted as formal SIP revisions by the Governor (or
designee) following a public hearing. As explained previously, EPA will
continue to process and approve, under section 182(f)(3), conformity
NOX waivers for areas not subject to section 182(b)(1) without
public hearings or submission by the Governor. Finally, as noted
earlier, the NOX provisions of the general conformity rule would
not be affected by this proposal. A NOX waiver under section
182(f) removes the NOX general conformity requirements entirely
and would continue to do so. The Clean Air Act's provision for
transportation conformity NOX waivers stems from section
176(c)(3)(A)(iii), which addresses only transportation conformity, and
not general conformity. Therefore, the statutory authority for general
conformity NOX waivers is not
[[Page 2431]]
required to be section 182(b) for any areas and may continue to be
section 182(f) for all areas.
It should be noted that EPA is taking no final action on a NOX
exemption for transportation conformity for ozone nonattainment areas
classified as moderate and above in the petition covered by this
rulemaking. The States of Illinois, Indiana, Michigan, and Wisconsin
may seek a transportation conformity NOX exemption for such areas
through formal SIP revisions pursuant to section 182(b)(1) of the Act
(Illinois and Wisconsin have submitted such SIP revisions, which are
currently being reviewed by the EPA).
Comment: Commenters argue that waiver of NOX control
requirements is unlawful if such a waiver would impede attainment and
maintenance of the ozone standard in downwind areas.
Response: As a result of these comments, the EPA reevaluated its
position on this issue and has revised the previously issued guidance.
See Memorandum, ``Section 182(f) Nitrogen Oxides (NOX)
Exemptions--Revised Process and Criteria'' dated February 8, 1995, for
John Seitz's signature. As described in this memorandum, EPA intends to
use its authority under section 110(a)(2)(D) to require a State to
reduce NOX emissions from stationary and/or mobile sources where
there is evidence, such as photochemical grid modeling, showing that
the NOX emissions could contribute significantly to nonattainment
in, or interfere with maintenance by, any other State or in another
nonattainment area within the same State. This action would be
independent of any action taken by EPA on a NOX exemption request
under section 182(f). That is, EPA action to grant or deny a NOX
exemption request under section 182(f) for any area would not shield
that area from EPA action to require NOX emission reductions, if
necessary, under section 110(a)(2)(D).
Significant new modeling analyses are being conducted by the Lake
Michigan Air Directors Consortium (LADCO) (the technical and functional
directors of the Lake Michigan Ozone Study and the Lake Michigan Ozone
Control Program, including representatives of the four LMOS States and
the EPA), EPA and other agencies as part of the Ozone Transport
Assessment Group (OTAG) process. The OTAG process is a consultative
process among the eastern States and EPA. The OTAG process, which ends
at the close of 1996, assesses national and regional emission control
strategies using improved modeling techniques. The goal of the OTAG
process is for EPA and the affected States to reach consensus on the
additional regional and national emission reductions that are needed
for attainment of the ozone standard. Based on the results of the OTAG
process, States are expected to submit by mid-1997 attainment plans
which show attainment of the ozone standard through local, regional,
and national controls.
The OTAG plans to complete additional modeling between now and
September 1996 using emissions data and strategies currently being
developed among OTAG workgroups. These new analyses will improve the
information available on NOX and VOC impacts on ozone
concentrations both in the LMOS area and over the eastern half of the
United States. These analyses will for example, provide more accurate
boundary conditions for the LMOS area analyses; this provides greater
accuracy in both the attainment plan and in the decision regarding
NOX reductions contribution to attainment.
In light of the modeling completed thus far and considering the
importance of the OTAG process and attainment plan modeling efforts,
EPA is granting this waiver on a contingent basis. As the OTAG modeling
results and control recommendations are completed in 1996, this
information will be incorporated into the attainment plans being
developed by the LADCO States. When these attainment plans are
submitted to EPA in mid-1997, these new modeling analyses will be
reviewed to determine if the NOX waiver should be continued,
altered, or removed.
The attainment plans will supersede the initial modelling results
which are the basis for the waiver which EPA is granting in this rule.
To the extent the attainment plans include NOX controls on certain
major stationary sources in the LMOS ozone nonattainment areas, EPA
will remove the NOX waiver for those sources. To the extent the
plans achieve attainment without additional NOX reductions from
certain sources, the NOX reductions would be considered excess
reductions and, thus, the exemption would continue for those sources.
EPA's rulemaking action to reconsider the initial NOX waiver may
occur simultaneously with rulemaking action on the attainment plans.
Comment: Comments were received regarding the scope of exemption of
areas from the NOX requirements of the conformity rules. The
commenters argue that such exemptions waive only the requirements of
section 182(b)(1) to contribute to specific annual reductions; not the
requirement that conformity SIPs contain information showing the
maximum amount of motor vehicle NOX emissions allowed under the
transportation conformity rules and, similarly, the maximum allowable
amounts of any such NOX emissions under the general conformity
rules. The commenters admit that, in prior guidance, EPA has
acknowledged the need to amend a drafting error in the existing
transportation conformity rules to ensure consistency with motor
vehicle emissions budgets for NOX, but want EPA, in actions on
NOX exemptions, to explicitly affirm this obligation and to also
avoid granting waivers until a budget controlling future NOX
increases is in place.
Response: As explained previously, EPA's transportation conformity
rule originally provided for a NOX waiver if an area received a
section 182(f) exemption. The EPA published amendments to the
transportation conformity rule in a final rule on November 14, 1995 (60
FR 57179) which addresses the issue of conformity to NOX budgets
in SIPs when a NOX waiver for transportation conformity has been
approved. The final rule is based on an August 29, 1995 (60 FR 44790)
proposed rule and comments which were received regarding that proposal.
The final rule requires consistency with NOX motor vehicle
emissions budgets in control strategy SIPs regardless of whether a
NOX waiver has been granted. The NOX build/no-build tests and
less-than-1990 tests, however, no longer apply to ozone nonattainment
areas receiving a NOX waiver. Furthermore, some flexibility is
possible for areas that have been issued a NOX waiver based on air
quality modeling data. This flexibility is described in the notice of
final rulemaking (60 FR 57183). The NOX emission budget provisions
of the revised rules will be effective 90 days after the date of the
final rule (November 14, 1995).
Comment: Commenters argue that the Act does not authorize any
waiver of the NOX reduction requirements until conclusive evidence
exists that such reductions are counterproductive.
Response: EPA does not agree with this comment since it ignores the
Congressional intent as evidenced by the plain language of section
182(f), the structure of the Title I ozone subpart as a whole, and
relevant legislative history. By contrast, in developing and
implementing its NOX exemption policies, EPA has sought an
approach that reasonably accords with that intent. Section 182(f), in
addition to imposing control requirements on major stationary sources
of NOX similar to those that apply for sources of VOC, also
provides for an exemption (or limitation) from application of these
requirements if, under one of several
[[Page 2432]]
tests, EPA determines that in certain areas NOX reductions would
generally not be beneficial towards attainment of the ozone standard.
In section 182(f)(1), Congress explicitly conditioned action on
NOX exemptions on the results of an ozone precursor study required
under section 185B of the Act. Because of the possibility that reducing
NOX in an area may either not contribute to ozone attainment or
may cause the ozone problem to worsen, Congress included attenuating
language, not just in section 182(f), but throughout Title I of the
Act, to avoid requiring NOX reductions where such would not be
beneficial or would be counterproductive. In describing these various
ozone provisions, including section 182(f), the House Conference
Committee Report states in the pertinent part: ``[T]he Committee
included a separate NOX/VOC study provision in section [185B] to
serve as the basis for the various findings contemplated in the
NOX provisions. The Committee does not intend NOX reduction
for reduction's sake, but rather as a measure scaled to the value of
NOX reductions for achieving attainment in the particular ozone
nonattainment area.'' H.R. Rep. No. 490, 101st Cong., 2d Sess. 257-258
(1990).
As noted in response to an earlier comment, the command in section
182(f)(1) that EPA ``shall consider'' the 185B report taken together
with the timeframe the Act provides for completion of the report and
for acting on NOX exemption petitions clearly demonstrate that
Congress believed the information in the completed section 185B report
would provide a sufficient basis for EPA to act on NOX exemption
requests, even absent the additional information that would be included
in affected areas' attainment or maintenance demonstrations.
While there is no specific requirement in the Act that EPA actions
granting NOX exemption requests must await ``conclusive
evidence,'' as the commenters argue, there is also nothing in the Act
to prevent EPA from revisiting an approved NOX exemption if
warranted by additional, current information.
In addition, the EPA believes, as described in EPA's December 1993
guidance, that section 182(f)(1) of the Act provides that the new
NOX requirements shall not apply (or may be limited to the extent
necessary to avoid excess reductions) if the Administrator determines
that any one of the following tests is met:
(1) in any area, the net air quality benefits are greater in the
absence of NOX reductions from the sources concerned;
(2) in nonattainment areas not within an ozone transport region,
additional NOX reductions would not contribute to ozone attainment
in the area; or
(3) in nonattainment areas within an ozone transport region,
additional NOX reductions would not produce net ozone air quality
benefits in the transport region. Based on the plain language of
section 182(f), EPA believes that each test provides an independent
basis for a full or limited NOX exemption.
Only the first test listed above is based on a showing that
NOX reductions are ``counter productive.'' If one of the tests is
met (even if another test is failed or not applied), the section 182(f)
NOX requirements would not apply or, under the excess reductions
provision, a portion of these requirements would not apply.
Comment: Commenters argue that, while NOX controls may be less
beneficial than VOC-only controls in reducing ozone concentrations in
some areas of the Lake Michigan region on some days, the States have
not demonstrated that VOC-only controls will sufficiently reduce ozone
concentrations for the majority of episodes, particularly in areas
farther downwind.
Response: Several modeling and data analyses were performed by the
States and LADCO to examine the relative benefits of VOC versus
NOX emission controls. The modeling analyses included emissions
sensitivity tests for several different basecase scenarios, including:
(1) an original base period emissions inventory; (2) increased VOC
emissions in the base period inventory (higher VOC/NOX ratios);
(3) increased base period VOC/NOX ratios through either increased
VOC emissions or decreased NOX emissions; and (4) differences in
photochemistry photolysis rates as applied in the Urban Airshed Model--
Version IV (UAM-IV) (the photochemical dispersion model generally
accepted and supported by the EPA) and in UAM-V (the photochemical
dispersion model approved by the EPA for use in the LMOS).
Despite differences in the absolute and relative amounts of VOC and
NOX emissions in the sensitivity analyses, the analyses found that
the modeled domain-wide peak ozone concentration, the areal coverage of
modeled ozone concentrations exceeding 120 parts per billion (ppb), and
the number of hours with modeled ozone concentrations exceeding 120 ppb
decreased in response to VOC emission reductions and increased in
response to NOX emission reductions (up to more than 60 percent
controls for some episode analysis days) for all modeled episodes.
VOC and NOX emission reductions were found to produce
different impacts spatially. In and downwind of major urban areas,
within the ozone nonattainment areas, VOC reductions were effective in
lowering peak ozone concentrations, while NOX emission reductions
resulted in increased peak ozone concentrations. Farther downwind,
within attainment areas, VOC emissions reductions became less effective
for reducing ozone concentrations, while NOX emission reductions
were effective in lowering ozone concentrations. It must be noted,
however, that the magnitude of ozone decreases farther downwind due to
NOX emission reductions was less than the magnitude of ozone
increases in the ozone nonattainment areas as a result of the same
NOX emission reductions.
Analyses of ambient data by LMOS contractors provided results which
corroborated the modeling results. These analyses identified areas of
VOC- and NOX-limited conditions (VOC-limited conditions would
imply a greater sensitivity of ozone concentrations to changes in VOC
emissions. The reverse would be true for NOX-limited conditions.)
and tracked the ozone and ozone precursor concentrations in the urban
plumes as they moved downwind. The analyses indicated VOC-limited
conditions in the Chicago/Northwest Indiana and Milwaukee areas and
NOX-limited conditions further downwind. These results imply that
VOC controls in the Chicago/Northwest Indiana and Milwaukee areas would
be more effective at reducing peak ozone concentrations within the
severe ozone nonattainment areas.
The consistency between the modeling results and the ambient data
analysis results for all episodes with joint data supports the view
that the UAM-V modeling system developed in the LMOS may be used to
investigate the relative merits of VOC versus NOX emission
controls. The UAM-V results for all modeled episodes point to the
benefits of VOC controls versus NOX controls in reducing the
modeled domain peak ozone concentrations.
Comment: Commenters argue that the UAM-V modeling system is
experimental and untested and has not yet undergone extensive peer
review by independent experts, unlike the Regional Oxidant Model (ROM)
supported by the EPA. The EPA should review the ROM results for the
episodes modeled in the LMOS to show
[[Page 2433]]
consistency between the ROM results and those for UAM-V.
Response: Even though the UAM-V modeling system is relatively new,
it has undergone external review. LADCO supported an external review of
the computer code used in the modeling system and an external
evaluation of model performance in the Lake Michigan region. Modeling
results show that the system, as it is currently being used for control
strategy analyses, produces ozone concentrations which meet EPA-
established criteria for adequate model performance.
Direct comparisons of ROM and UAM-V results must be conducted with
caution and may produce conflicting results even though both modeling
systems are performing adequately. The UAM-V modeling system is
theoretically more complete and incorporates improved scientific
principles and more area-specific input data. ROM, on the other hand,
is a simpler modeling system with lower spatial resolution, more
uncertain emission estimates, and no special treatment of
meteorological phenomena, such as lake-breeze effects (critical factors
in the Lake Michigan area), and individual source plumes for large
sources. These differences in model formulation and data input
resolution as well as differences in output resolution may preclude
direct comparisons of the two models. It should be noted, that such a
comparison may be attempted in the near future because UAM-V may be
applied to a larger domain to assess the impacts of long range
transport of ozone and ozone precursors.
Comment: Commenters state that the EPA must rely on the recent
National Academy of Sciences (NAS) report in its review of NOX
waivers. The commenters pointed out that the NAS report found that to
reduce transported ozone, NOX reductions are needed.
Response: The NAS report and EPA's companion report both support
the conclusion that, as a general matter for ozone nonattainment areas
across the country, NOX reductions in addition to VOC reductions
will be needed to achieve attainment. This general conclusion, however,
must be assessed in the context of the more detailed analysis provided
in those same reports. For example, the NAS report notes that NOX
reductions can have either a beneficial or detrimental effect on ozone
concentrations, depending on the locations and emission rates of VOC
and NOX sources in a region. The effect of NOX reductions
depends on the local VOC/NOX ratio and a variety of other factors.
In its report issued pursuant to section 185B of the Act, EPA stated
that ``[a]pplication of gridded photochemical models on a case by case
basis is required to determine the efficacy of NOX controls,
because the ozone response to precursor reductions is area specific.''
The analyses performed in the Lake Michigan region demonstrate a
local disbenefit from NOX control in the urban nonattainment
areas. Those same analyses suggest there would be ozone benefits
experienced farther downwind from NOX control in the urban
nonattainment areas. LADCO acknowledges that NOX controls in the
LMOS modeling domain may be needed ultimately to reduce ozone transport
in the eastern United States. Nonetheless, the modeling results show
that, due to the ozone reduction disbenefits associated with NOX
reductions for the ozone nonattainment areas in the LMOS domain, these
areas meet the test under section 182(f)(1)(A) of the Act required to
support a waiver from the NOX requirements of section 182(f).
Comment: Commenters believe that NOX emission reductions will
not only reduce transported ozone, but will also improve visibility,
especially in downwind Class I areas.
Response: The NOX control waiver request was submitted in
conjunction with the preparation of a four-State ozone control plan. To
this end, the focus is on the local ozone problem in the Lake Michigan
region. Other air pollution problems will be dealt with as part of
separate regulatory activities.
Comment: Commenters argue that the burden of proof is on the States
and LADCO to demonstrate that NOX reductions will not be
beneficial over the entire Lake Michigan region. It was the explicit
intent of Congress that NOX reductions are to be presumed to be
beneficial unless demonstrated otherwise.
Response: Modeling and data analyses addressed in the States'
NOX waiver request demonstrate the positive benefits of VOC
control in the major urban areas and downwind in the areas of highest
ozone concentrations. These analyses also show the negative effects of
NOX control in these same ozone nonattainment areas, and suggest
positive benefits from NOX control farther downwind in attainment
areas. In other words, the benefits resulting from NOX control are
modelled to occur in areas that experience, based on modeling and
monitoring data, ozone concentrations well below the ozone standard
even prior to the implementation of emission controls. Consequently, as
required under section 182(f), the States have demonstrated the
disbenefits of implementing NOX emission controls in terms of
greater domain-wide peak ozone concentrations throughout the LMOS
modeling domain. Since these States are relying on the section
182(f)(1)(4) ``contribute to attainment'' test, they do not also need
to demonstrate NOX reduction benefits over the entire Lake
Michigan region as the commenters claim.
As noted above, the EPA believes, as described in EPA's December
1993 guidance, that section 182(f)(1) of the Act provides that the new
NOX requirements shall not apply if the Administrator determines
that any one of the following tests is met:
(1) in any area, the net air quality benefits are greater in the
absence of NOX reductions from the sources concerned;
(2) in nonattainment areas not within an ozone transport region,
additional NOX reductions would not contribute to ozone attainment
in the area; or
(3) in nonattainment areas within an ozone transport region,
additional NOX reductions would not produce net ozone air quality
benefits in the transport region. Based on the plain language of
section 182(f) and the modeling results supplied with the LMOS States'
NOX waiver request, the EPA believes these States have met the
requirements of test (2) above since the States have demonstrated that
across-the-board NOX controls in the LMOS ozone nonattainment
areas will interfere with the attainment of the ozone standard in these
nonattainment areas. Based on the scheme provided by Congress under the
Act, it is not necessary for the States to also demonstrate the lack of
ozone benefits from NOX controls everywhere within the entire Lake
Michigan region.
As a separate matter and as noted above, the States, LADCO, and the
EPA are conducting additional studies on the impact of ozone precursor
(including NOX) emission reductions in areas outside of the LMOS
ozone nonattainment areas on downwind ozone concentrations. These
studies, in part, will consider the LMOS nonattainment areas as
downwind areas to assess the impact of upwind emissions controls on
ozone and ozone precursor transport into these areas.
Comment: Commenters argue that LADCO's statistical comparisons
provide an incomplete evaluation of model performance and do not assess
the model's ability to accurately predict the impact of VOC versus
NOX control.
Response: LADCO, through a September 1994 model evaluation report,
has documented a thorough evaluation of the modeling system
performance. The model evaluation, which is based on an ideal model
[[Page 2434]]
evaluation process proposed by a number of technical experts 1,
includes the following elements:
\1\ ``A Conceptual Framework for Evaluating the Performance of
Grid-Based Photochemical Air Quality Simulation Models'', Roth,
Reynolds, Tesche, and Dennis (1991).
---------------------------------------------------------------------------
(1) Evaluations of the scientific formulation of the model;
(2) Assessment of the fidelity of the computer codes to scientific
formulation, governing equations, and numerical solution procedures;
(3) evaluation of the predictive performance of the individual
process modules and preprocessor modules;
(4) evaluation of the full model's predictive performance;
(5) application of sensitivity tests to assure conformance of the
model with known or expected behavior;
(6) application of comparative modeling; and
(7) implementation of quality control/quality assurance activities.
The September 1994 model evaluation report addressed all of these
elements for the modeling system used in the LMOS. In addition, the
report also discussed several analyses which were performed
specifically to assess the reliability of the model's response to VOC
and NOX emission reductions (see response to comment above
concerning the response of the model to VOC-only controls).
The model evaluation conducted for the LMOS modeling system
examined performance over as wide a range of emission densities as
possible (both spatial and temporal ranges were considered), considered
topographic and land use uncertainties, and evaluated the impacts of
variations in meteorology. Demonstration of acceptable model
performance over this range of conditions reflects correct
representation of the governing chemical and physical processes. It is,
therefore, reasonable to assume that the model will respond
realistically irrespective of emission strengths of VOC versus
NOX.
Comment: Commenters argue that LADCO has failed to conduct
additional analyses of model performance which provide a better test of
VOC-NOX sensitivity [e.g., analyses of afternoon concentrations of
total reactive nitrogen (NOy)]. An examination of ambient NOX
concentrations over Lake Michigan clearly show NOX-limited
conditions (i.e., NOX control should be beneficial for reducing
ozone concentrations) and, further, that the modeled NOX
concentrations are overestimated, which would cause the model to
incorrectly identify VOC control as being preferential to NOX
control. NOX concentrations, as predicted by the model to occur
over Lake Michigan (i.e., 90 parts per billion), are unlikely to occur
anywhere other than in urban centers.
Response: The September 1994 model evaluation report submitted by
LADCO does include the type of analysis suggested by the commenters.
This analysis of predicted and measured NO2 concentrations
(NOy concentrations were not measured making evaluation of modeled
results for NOy impossible. NOX is assumed to be primarily
NO2 at the peak ozone times and locations.) at the time and
location of maximum ozone concentration for each day shows no
discernible bias in the model predictions.
The September 1994 model evaluation report also includes a general
assessment of model performance for NO2. Rather than focusing on
just one or two days, as was done by the commenters, the evaluation
considered all of the modeled high ozone days. The results for all high
ozone days demonstrate that model performance overall for NO2 is
good.
The magnitude of the predicted NOX concentrations over Lake
Michigan, as cited by the commenters, is not correct. The model
predicted NO2 concentrations over Lake Michigan on the order of 50
parts per billion or less. NOX measurements by the LMOS aircraft
over Lake Michigan were on the order of 30--50 parts per billion, in
good agreement with the model's predicted concentrations (NOX over
Lake Michigan is primarily NO2).
Comment: Commenters argue that VOC emissions are likely
underestimated in the emission inventory used for the LMOS modeling,
which would cause a bias in the model towards favoring VOC control.
Also, LADCO's finding that its VOC inventory may be low by only 30
percent conflicts with studies elsewhere which suggest a high degree of
underestimation.
Response: Several methods were used by LADCO to evaluate the LMOS
emissions inventory, including comparisons of ambient to emissions-
based nonmethane organic compound to NOX (NMOC:NOX) ratios;
comparisons of ambient to emissions-based carbon monoxide to NOX
ratios; receptor modeling; and comparisons of ambient to model-based
NMOC:NOX ratios. These analyses for an initial emissions inventory
suggested a significant underestimation of VOC emissions,
overestimation of NOX emissions, or some combination of these two.
Consequently, LADCO conducted an extensive re-evaluation of the
emissions inventory and made several modifications. The resulting,
final emissions inventory was found to compare more closely to the
ambient NMOC:NOX ratios (the ambient NMOC:NOX ratios are only
about 1.0--1.5 times greater than the emissions inventory-based
NMOC:NOX ratios).
To assess the effect of the emissions uncertainty on the model's
response to VOC and NOX reductions, sensitivity tests were
performed with a higher VOC:NOX ratio. The results of this
modeling were qualitatively the same (NOX disbenefits were
demonstrated for attainment of the ozone standard) as those found for
the unadjusted emissions inventory. Consequently, any possible
underestimation of VOC emissions does not affect the conclusions drawn
concerning VOC versus NOX controls.
With regard to the results of other emissions studies, it should
first be noted that a certain degree of variability of emissions ratios
(NMOC:NOX) exists depending of the locations of the studies and
the sources sampled. Application of the results of these studies to the
LMOS source areas is not straight forward and must be viewed to have a
high degree of uncertainty. The LMOS results leading to the adjustment
of emissions and the favorable comparison of modeled and monitored
results lends some credibility to the emissions used in the LMOS.
Secondly, the LMOS States and LADCO, based on the studies of mobile
source emissions conducted previously in other areas, recognized the
potential for the underestimation of mobile source VOC emissions. This
recognition was part of the basis for the comparison of monitored and
modeled emissions and the modeling sensitivity studies considering
alternate NMOC:NOX ratios. As indicated above, increased
NMOC:NOX ratios lead to the same conclusions regarding the impacts
of VOC versus NOX emissions controls.
Comment: A commenter notes that the problems with the LMOS modeling
are not ``routine'' model errors. The LMOS model results, as presented
in a February 1994 report by Alpine Geophysics, showed large errors in
comparison with measurements for certain pollutant species and these
errors suggest a bias in favor of VOC control and against NOX
control. The finding that the model systematically overestimates
NOy also suggests that the model is biased in favor of VOC
control.
Response: The commenter has chosen to rely on outdated results from
a preliminary February 1994 model evaluation report. Since then, as
documented, for example, in the
[[Page 2435]]
September 1994 model evaluation report submitted to the EPA by LADCO,
significant improvements have been made in the modeling system and in
its inputs. (See also the discussions in response to other comments
regarding the model's performance.) The improved modeling system and
its results make moot the concerns of the commenter.
Comment: A commenter is concerned about the quality of the multi-
species evaluation contained in the September 1994 model evaluation
report. The commenter notes that an interim report indicated that the
model performed poorly in modeling the concentrations of paraffins,
frequently erring by a factor of two or more. Such an error implies
that the model may be biased in favor of VOC controls. The commenter
further notes that the September 1994 model evaluation report fails to
include a significant discussion of multi-species evaluations,
particularly a discussion of modeled versus measured paraffin
concentrations.
Response: The September 1994 report does discuss the fact that
multi-species analyses were performed for the updated modeling system
and updated input data. As noted above, the updated model performed
acceptably for the prediction of species such as ozone, NO2,
NOX, and VOC:NOX. The report did fail to discuss most other
species addressed in the model. LADCO has acknowledged this failure,
and has offered to supply any data requested by the EPA. LADCO,
however, has indicated, in its own responses to the comments on the
proposed approval of the NOX waiver, that the multi-species
performance of the model has significantly improved from past versions
of the modeling system and input data. It is not clear how the
performance of the model regarding prediction of paraffin
concentrations has changed.
Comment: A commenter notes that the emissions inventory used in the
modeling underestimates emissions of both anthropogenic and biogenic
VOC emissions. A particular deficiency is the lack of any biogenic
isoprene emissions in the Chicago area. In addition, the failure to
evaluate model performance for isoprene is especially important. Models
that recommend VOC-based control strategies should be required to
demonstrate that they have not underestimated ambient concentrations of
isoprene.
Response: As noted in a response to a comment above, the current
version of the emissions inventory used in the modeling reasonably
agrees with the ambient data. Although the current LMOS emissions
inventory does not contain biogenic isoprene emissions, calculations
made by LADCO, as discussed in LADCO's response to this comment,
indicate that this does not result in a significant change in the VOC
inventory (addition of biogenic isoprene emissions would only increase
the regional VOC inventory by 1 percent or less). Ambient VOC
measurements also reflect negligible isoprene concentrations in the
Chicago, Gary, and Milwaukee urban areas. The lack of an evaluation of
isoprene concentrations should not detract from the overall assessment
of model performance.
LADCO has noted that the EPA-recommended emission factors for
biogenic isoprene are under review nationally. LADCO has committed to
revise the emissions inventory if these emission factors are changed
significantly, particularly if they are significantly increased.
LADCO has noted that the UAM-V modeling system has been thoroughly
evaluated. In fact this evaluation significantly exceeds the
requirements of the EPA and exceeds the evaluations employed for UAM in
most other ozone nonattainment areas in the United States.
Comment: A commenter notes that the September 1994 model evaluation
report fails to include modeled versus measured NO2 concentrations
from locations that represent maximum measured ozone concentrations. It
is also noted that two-thirds of the modeled-measured data pairs that
were documented in the model evaluation report lie outside of the
factor-of-two range implying poor agreement between modeled and
measured concentrations.
Response: LADCO notes that the modeled versus measured NO2
data were included in the final model evaluation report (October 1994).
These data show that there is no discernible bias in the model
predictions. Furthermore, only a few data pairs reflect an
overprediction by more than a factor of two. Most of the data pairs lie
either within the factor-of-two range, or reflect underprediction by
more than a factor-of-two (underprediction of NO2 would favor
NOX control over VOC control in reducing ozone concentrations).
Despite the possible underprediction of some NO2 concentrations,
the model continues to show that NOX control provides disbenefits
for attainment of the ozone standard in the LMOS domain.
Comment: It is noted that Table 7 in the September 1994 model
evaluation report contains NOX data which differ from those in a
February 1994 model evaluation report. It is also noted that the
September 1994 model evaluation report also fails to include data for a
critical site (the Mid-Lake Boat) on July 18, 1991.
Response: The NOX values contained in the February 1994 report
did not reflect the final quality-assured data for the boat-based
monitors used in the 1991 LMOS field study. The final data were
addressed in the September 1994 model evaluation report. Nevertheless,
no firm conclusions should be based on the NOX data from the boats
because these data were found to be suspect.
Table 7 in the September 1994 report did not include the Mid-Lake
Boat data for July 18, 1991 because the Boat stopped collecting data on
this day after 1600 Central Daylight Time (CDT). The modeling domain-
wide peak observed and modeled concentrations, as noted in Table 7,
occurred after 1600 CDT. In any case, the peak ozone concentration at
the Mid-Lake Boat on this day was 158 parts per billion (1400 CDT). The
magnitude of the NOX concentration for this hour was still fairly
high (13 parts per billion), indicating that the air mass may still
have been VOC-limited, which favors VOC control of upwind sources over
NOX control for the reduction of ozone levels.
Comment: The February 1994 model evaluation report contains
evidence that the mixing algorithm in UAM-V has serious problems. In
particular, the model is overestimating ambient NOX concentrations
by a factor of three or more during the mid-July 1991 episode.
Response: The September 1994 model evaluation report shows that the
model performance statistics for NO2 (as noted above, NOX
over Lake Michigan is primarily NO2 with little NO) during the
mid-July episode are reasonable. The spatial concentration plots for
NO2 show that the predicted values are highest in the Chicago
downtown area and decrease downwind over Lake Michigan. The latest
baseline model input data set (Basecase C) produces significantly lower
peak NO2 concentrations than did the earlier baseline model input
data set considered by the commenter. The new input data lead to
results similar to concentrations measured in aircraft over Lake
Michigan during the 1991 field study.
Comment: A commenter claims that the September 1994 model
evaluation report erroneously claims that 1991 field study NOy
measurements were not available and that most local afternoon NOy
is expected to be NO2.
Response: Contrary to the commenter's claims, NOy data were
not collected during the 1991 field program. The only nitrogen species
for which
[[Page 2436]]
ambient data were collected were NO, NO2, NOX, and
peroxyacetlnitrate (PAN) (collected at only a few sites). LADCO
responds and EPA agrees that, while NOy reflects many nitrogen
compounds, NO2 is a reasonable surrogate for these analyses.
Comment: Commenters note that LADCO has requested and received EPA
approval to assume a future modeling domain boundary peak ozone
concentration of 60 parts per billion. An analysis of this assumption
leads the commenters to conclude that NOX transported into the
modeling domain would have to be reduced by approximately 66 percent
from current emission levels. Given the policy established in the
approval of the NOX exemption petition, the commenters question
the feasibility of this boundary condition assumption.
Response: It is true that the EPA has approved the assumption of a
future modeling domain boundary ozone concentration not exceeding 60
parts per billion. It should be noted, however, that this is a
temporary assumption to be used only in the initial phase of ozone
modeling needed to develop the areas' final ozone demonstrations of
attainment. Regional modeling over a larger domain will be conducted to
better assess the level of ozone transport in the Eastern United
States. This regional modeling will also assess the impacts of possible
national emission control efforts to generally lower ozone precursor
emissions throughout this area. The final phase of local ozone modeling
will use ozone boundary conditions based on the regional modeling.
It should also be noted that the EPA, under section 110(a)(2)(D) of
the Act, may require additional NOX emission controls in the areas
exempted from specific NOX control requirements under section
182(f) of the Act. The NOX emission reduction requirements under
section 110(a)(2)(D) may exceed those under section 182(f) if the
regional modeling supports the need for such emission reductions. The
boundary ozone concentration that will ultimately be used in the final
demonstrations of attainment will be backed by adequate ozone precursor
emission reductions.
Comment: Commenters argue that the NOX exemption petition
ignores the LMOS States' contribution to their own boundary conditions.
Insufficient analyses have been presented that consider the benefits in
lowered boundary ozone levels that could be achieved during episodes
when locally generated ozone and ozone precursors are transported out
of and back into the modeling domain. Exceedances observed on June 18,
1994 are of note in this regard. On this day, it appears that the
Chicago/Gary ``plume'' actually moved north-northeast only to later
reimpose itself on the metropolitan area. The benefits for NOX
control are not presented for this meteorological phenomenon.
Response: Modeling for LMOS considered all high ozone episodes in
1991. Modeling for these episodes will form the basis for the ultimate
ozone demonstrations of attainment to be completed in 1997 under
current EPA policy. The NOX exemption petition is based on
modeling for all of these high ozone episodes, and, as such, meets the
modeling requirements in the December 1993 EPA guidance. It should be
noted that the episodes considered cover a significant range of
meteorological phenomena, including ozone transport and recirculation
within the LMOS domain. A more complete picture of ozone transport out
of and back into the modeling domain will not be available until after
the completion of the regional modeling discussed in the response to
the previous comment.
Comment: A commenter argues that incorporating the Michigan
Counties of Saginaw, Bay, Genessee, Shiawasse, Midland, Ingham,
Jackson, Lenawee, and Calhoun is an attempt to factitiously expand the
domain of LADCO's NOX disbenefit analysis. It is also noted that
the EPA has included the fictional Michigan County of Hillside. The
commenter argues that, if EPA had intended to exempt Hillsdale County
rather than ``Hillside County,'' the EPA should publish a correction
notice amending the proposed rulemaking notice.
Response: When LADCO conducted the modeling analysis of NOX
control impacts, NO