F. Reasonably available control technology and reasonably available
control measures.  

	This section of the preamble discusses the final rule requirements for
RACT and RACM.  In order to explain EPA’s approach in the final rule
more clearly, we first discuss the statutory and regulatory background
for the RACT and RACM requirements, and we then explain the key options
and interpretations upon which we took comment in the proposal. 
Thereafter, we discuss significant comments we received on the proposal
and provide brief responses to those comments. [Additional comments and
responses appear in the RTC for this final rule located in the docket.] 
Most of the comments received on this topic addressed the three options
EPA proposed for the RACT requirement, the relationship between the RACT
requirement and EPA’s Clean Air Interstate Rule (CAIR), and the
control measures to be required or considered for RACT and RACM.

1.  Background on statutory requirements for RACT and RACM tc "General
background " \l 3 	

 Subpart 1 of Part D of the CAA (sections 171 – 179B) applies to all
designated nonattainment areas.  Section 172 of this subpart includes
general requirements for all nonattainment area plans. 

Notably, Congress provided EPA and States a great deal of deference for
determining what measures to include in a nonattainment plan. 
Specifically, Section 172(c)(1) requires that each nonattainment area
plan “provide for the implementation of all reasonably available
control measures as expeditiously as practicable (including such
reductions in emissions from existing sources in the area as may be
obtained through the adoption, at a minimum, of reasonably available
control technology), and shall provide for attainment of the national
primary ambient air quality standards.”  By including language in
Section 172(c)(1) that only “reasonably available” measures be
considered for RACT/RACM, and that implementation of these measures need
be applied only “as expeditiously as practicable,” Congress clearly
intended that the RACT/RACM requirement be driven by an overall
requirement that the measure be “reasonable.”   Thus, the rule of
“reason” drives the decisions on what controls to apply, what should
be controlled, by when emissions must be reduced, and finally, the rigor
required in a State’s RACT/RACM analysis.  For example, we previously
stated that the Act “does not require measures that are absurd,
unenforceable, or impractical” or result in “severely disruptive
socioeconomic impacts” 55 FR 38327.  Moreover, we interpret the term
“reasonably available” to allow States to consider both the costs
and benefits of applying the measure, and whether the measure can be
readily and effectively implemented without undue administrative burden.
 66 FR 26969. 

We also interpret the “reasonably available control measures” in
these provisions as referring to measures of any type that may be
applicable to a wide range of sources, whereas the parenthetical
reference to “reasonably available control technology” refers to
measures applicable to stationary sources. Thus, RACT is a type of RACM
specifically designed for stationary sources.  As noted above, States
are required to implement RACM and RACT “as expeditiously as
practicable” as part of nonattainment area plans designed to attain
the standards.  

Section 172 does not include any specific applicability thresholds to
identify the size of sources that States and EPA must consider in the
RACT and RACM analysis.  Nor, does Section 172 specifically indicate
which pollutant(s) or precursor(s) must be subject to RACM or RACT
measures to attain the NAAQS.  Other pollutant-specific provisions of
the CAA do include applicability thresholds pertaining to nonattainment
area plan requirements for NAAQS and precursor pollutants.  For example,
subpart 2 of part D, which establishes additional requirements for ozone
nonattainment areas, establishes thresholds ranging from 100 to 10 tons
per year for requirements applicable to “major sources” or “major
stationary sources,” depending on the area’s classification or level
of nonattainment.   Subpart 4 of part D, which provides additional plan
requirements for PM-10 nonattainment areas, establishes thresholds of
100 and 70 tons per year for requirements applicable to a “major
source” or “major stationary source.”

Moreover, subpart 1, unlike subparts 2 and 4, does not identify specific
source categories for which EPA must issue control technology documents
or guidelines, or  identify specific source categories for State and EPA
evaluation during nonattainment area plan development.  For ozone,
subpart 2 contains a list of specific requirements for control
techniques guidelines (CTGs) and alternative control techniques (ACT)
documents. For PM10, section 190 of the CAA (in subpart 4) places
particular emphasis on specific sources of area emissions, but does not
identify specific stationary source categories for which RACT guidance
must be issued.  Section 190 requires EPA to develop RACM guidance
documents for residential wood combustion, silvacultural and
agricultural burning, and for urban fugitive dust control. 

2. What is the Overall Approach to Implementing RACT and RACM in the
final rule?

a. Background for RACT. 

	Since the 1970s, EPA has interpreted RACT to mean “the lowest
emissions limitation that a particular source is capable of meeting by
the application of control technology that is reasonably available
considering technological and economic feasibility” as well as other
considerations. In practice, RACT is generally described as the “norm
achievable” by the source category.

	Section 110 of the 1970 Clean Air Act required States to develop SIPs
providing for attainment of the NAAQS by 1975 or 1977.  A number of
areas were having difficulty with developing attainment plans,
particularly for the ozone standard.  In response to the implementation
needs of this time period, EPA introduced the term “RACT” in a 1976
memorandum from Roger Strelow, Assistant Administrator for Air and Waste
Management to Regional Administrators, “Guidance for Determining
Acceptability of SIP Regulations in Non-attainment Areas” (Dec. 9,
1976).  In this early guidance relating to the acceptability of SIP
regulations, we indicated that our overriding concern in approving SIPs
was attaining the particular NAAQS expeditiously as practicable through
reasonably available control technology and other reasonably available
control measures.  “The basis for fully approving state-submitted SIP
regulations continues to be demonstrated attainment and maintenance of
all national ambient air quality standards as expeditiously as
practicable,” the memo stated.

The 1977 Clean Air Act amendments added Part D to Title I of the Act,
and for the first time the Act specifically called for EPA to designate
nonattainment areas and for SIPs to require RACT and RACM in those
nonattainment areas.   In a 1979 Federal Register notice, EPA noted its
view that Congress adopted EPA’s pre-existing conception of RACT in
the 1977 amendments. (44 FR 53782, September 17, 1979).  Also during the
late 1970s, EPA developed a number of new control techniques guideline
(CTG) documents as directed in the 1977 amendments.  These CTGs provided
States with information on controls for a number of categories of
sources emitting VOCs, and recommended a “presumptive norm” for
State RACT determinations based on the control levels achievable by
sources in a given industry.   CTGs reduced the burden on States by
eliminating the need for each State to develop its own technical support
for implementing the RACT requirement.   Since the CTG-recommended
controls were based on general capabilities of an industry, EPA in the
1979 guidance (44 FR 53782) urged States in setting RACT to judge the
feasibility of the recommended controls on particular sources, and to
adjust accordingly.  

	As noted above, EPA’s early guidance related to the RACT requirement
indicated that our overriding concern in approving State RACT
requirements was attaining the particular NAAQS.   We initially required
States to apply RACT to qualify for attainment extensions, and in some
cases, for plans that could not demonstrate attainment.    

During the 1980s, EPA implemented the RACT requirements with a number of
CTGs and guidance documents.   These materials were aimed at addressing
the attainment deadlines of 1982 and 1987 under the 1977 Clean Air Act
amendments.   During this time, EPA, for pollutants other than ozone,
considered RACT to be dependent upon reductions needed for attainment as
expeditiously as practicable.  For ozone, where the State performed
photochemical grid modeling, the approach was the same, but where the
State used less sophisticated tools, we considered RACT to be
independent of whether the controls were needed to reach attainment as
expeditiously as practicable. We took this alternate approach because of
concerns related to the precision of modeling techniques.  In other
words, in those cases, we required that a stationary source of the
requisite type and size be subject to RACT, whether or not such controls
were actually demonstrated to be necessary for the area to attain by its
specified date.  (44 FR 20375-20376, April 4, 1979)  

	Congress followed a similar approach in the 1990 amendments to the CAA
for purposes of the ozone NAAQS in the subpart 2 provisions added at
that time.  For example, section 182(b)(2) requires the imposition of
RACT controls for all VOC source categories covered by a CTG and for all
other major stationary sources of VOC located within certain
nonattainment areas.  Thus, Congress required  these controls without
allowing for an area-specific demonstration by the State that the area
needed the controls for attainment as expeditiously as practicable. 
Extensive discussion of this requirement appeared in the 1992 general
preamble (57 FR 13541), in which EPA provided guidance for
implementation of the ozone NAAQS.  

	Notably, Congress did not significantly amend the generally applicable
provisions for nonattainment areas that appear in subpart 1 of Part D in
1990.  This indicates that Congress intended that the Agency retain the
authority to interpret the generally applicable nonattainment area plan
requirements of section 172(c), including the RACT and RACM
requirements, in the way that is most appropriate for new NAAQS that are
subject to subpart 1.  As discussed below, EPA has determined that an
approach to the RACT requirement in which RACT varies in different
nonattainment areas based on the reductions needed for attainment as
expeditiously as practicable, is appropriate for implementation of the
PM2.5 NAAQS.  We believe that the improved ability to model air quality
impacts of emissions controls allows for this approach.

b. Proposed options for RACT.

	The EPA proposed and requested comment on three alternative approaches
for interpretation of the RACT requirement of section 172(c)(1) for
implementation of the PM2.5 NAAQS.  The EPA proposed these approaches in
order to evaluate which method would best ensure that States consider
and adopt RACT measures for stationary sources in a way that is
consistent with the overarching requirement to attain the standards as
expeditiously as practicable, while providing flexibility for States to
focus regulatory resources on those sources of emissions that contribute
most to local PM2.5 nonattainment.  

	Under the first proposed alternative, EPA would require States to
conduct a RACT analysis and to identify and require reasonably available
controls for all affected stationary sources in the nonattainment area,
comparable to the implementation of RACT provided in subpart 2 governing
implementation of the 1-hour ozone NAAQS.  Under this option, covered
sources would be required to apply reasonable available controls
considering technical and economic feasibility, and there would be no
opportunity for States to excuse stationary sources from control on the
basis that the emissions reductions from those controls would not be
necessary to meet RFP requirements or to reach attainment.  Under this
alternative, EPA proposed to limit the universe of sources for which
States must conduct a RACT analysis and impose RACT controls, by
providing an applicability threshold based upon the amount of emissions
potentially emitted by the sources.  Under this first option, EPA
requested comment on a number of alternative emissions applicability
thresholds.  

Under the second proposed alternative, EPA would require States to
conduct a RACT analysis and to identify reasonably available controls
for all affected stationary sources.  Under this option, however, States
could thereafter determine that RACT does not include controls that
would not otherwise be necessary to meet RFP requirements or to attain
the PM2.5 NAAQS as expeditiously as practicable.  Under this approach,
RACT would be determined as part of the broader RACM analysis and
identification of all measures - for stationary, mobile, and area
sources - that are technically and economically feasible, and that would
collectively contribute to advancing the attainment date.  Because RACT
and RACM are considered together under this alternative, we did not
propose emissions threshold options for evaluation of stationary source
RACT.  In addition, consistent with existing policies, States would be
required to evaluate the combined effect of reasonably available
measures to determine whether application of such measures could advance
the attainment date by at least one year. 

The third proposed alternative, EPA’s preferred option in the
proposal, combined the first two options and is similar to the RACT
approach adopted in the final implementation rule for the 8-hour ozone
program.  Under the third option, EPA would require States to conduct a
RACT analysis and to require reasonably available controls for all
affected stationary sources in nonattainment areas with attainment dates
more than 5 years from the date of designation.  For areas with an
attainment date within 5 years of designation (e.g. by April 2010 for
areas with an effective date for designation of April 2005), EPA would
require RACT as under the second proposed alternative, in which RACT
would be determined as part of the broader RACM analysis.  For these
areas, States could determine that RACT does not include controls that
would not otherwise be necessary to meet RFP requirements or to attain
the PM2.5 NAAQS as expeditiously as practicable.  The same proposed
suboptions with respect to the size of sources for consideration under
the first alternative were also included under this alternative.  

c. Proposed Approach for RACM  

	The EPA proposed and asked for comment on one approach for interpreting
the RACM requirement for PM2.5.  EPA based the proposal on the approach
that we adopted for other NAAQS implementation programs.  Under this
approach, a State provides a demonstration in its SIP that it adopted
all reasonably available measures needed to meet RFP requirements and to
attain the standard as expeditiously as practicable and that no
reasonably available additional measures would advance the advance the
attainment date by at least 1 year or would be necessary to meet the RFP
requirement for the area.

  	Under section 172(a)(2), the state implementation plan must provide
for a nonattainment area to attain as expeditiously as practicable, but
no later than 5 years after the effective date of designation of the
area (e.g., no later than April 2010 for the final designations
effective April 2005).  The statute thus creates a presumption for
attainment within 5 years of designation unless certain statutory
criteria are met for an extension of the attainment date.  Under the
proposed approach to RACM for PM2.5, each State would evaluate available
measures for sources of PM2.5 or its regulatory precursors in the area
to determine if reasonable measures were needed to meet the RFP
requirement or to achieve attainment as expeditiously as practicable. 
If modeling of all RACM and other state, regional and federal measures
indicates that the State will not be able to demonstrate attainment
within 5 years after designation based upon the severity of
nonattainment in that area or the availability or feasibility of
implementing controls in that area, then the State may request an
attainment date extension.  We proposed that under these circumstances,
the EPA could extend the attainment date for a period of 1 to 5 years,
when the State shows that it will implement all RACT and RACM as
expeditiously as practicable, has met its obligation to address
intrastate pollution transport from sources within its jurisdiction, and
still needs additional time to attain.

	 In the proposed rule, the EPA also took comment on the following
overall steps for implementing the statutory requirement for RACM.

(1) Identification of measures.  The State would begin the process of
determining RACM by identifying all available control measures for all
sources of PM2.5 and its precursors in the nonattainment area.  The RACM
can apply to mobile sources, area sources, and stationary sources.  

(2) Evaluation of measures.  After the State identifies the universe of
available measures for the sources in the area, the State would evaluate
them to determine whether implementation of such measures is technically
and economically feasible, and whether the measure will contribute to
advancing the attainment date. 

(3) Adoption of measures.  The State would adopt all reasonably
available measures for the area consistent with meeting the applicable
RFP requirements and attaining the NAAQS as expeditiously as
practicable, in accordance with applicable policy and guidance for
attainment demonstrations.  We would then approve or disapprove the
State’s plan through notice and comment rulemaking.  We also noted
that in reviewing the State’s selection of measures for RACM, or
determining that certain measures are not RACM, EPA may independently
supplement the rationale of the State or provide an alternative reason
for reaching the same conclusion as the State.	

c. Final rule. 

	The EPA carefully considered our interpretation of section 172(c)(1)
for the PM2.5 NAAQS.  Because of the variable nature of the PM2.5
problem in different nonattainment areas, which may require States to
develop nonattainment area plans that address widely disparate
circumstances (e.g., different source types and mixes, different
precursors and mixes of precursors, and different meteorological
conditions), we determined that the regulations implementing the PM2.5
NAAQS should provide for a great degree of flexibility with respect to
the RACT and RACM controls.

Selected approach to RACT and RACM.   The final rule reflects EPA’s
decision to select option 2 for RACT and to require a combined approach
to RACT and RACM.  Under this approach, RACT and RACM are those measures
that a State finds are both reasonably available and contribute to
attainment as expeditiously as practical in the specific nonattainment
area.  

By definition, measures that are not necessary either to meet the RFP
requirement or to help the area attain the NAAQS as expeditiously as
practicable, are not required RACT or RACM for such area.  The EPA
believes that this approach provides the greatest flexibility to a State
to tailor its SIP control strategy to the needs of a particular PM2.5
nonattainment area, but it may require the State to conduct a more
detailed analysis to identify the most effective RACT/RACM strategy to
attain the NAAQS.

During the comment period, commenters raised concerns that this approach
may be overly burdensome on States because of the number of potential
control measures a State would need to consider.  Today, we clarify that
although the State must conduct a thorough analysis of reasonably
available measures, States need not analyze every conceivable measure,
as explained in the guidance below.  Instead, “reason” should drive
States identification of potential measures, but States should remain
mindful of the public health risks of PM2.5. As long as a State’s
analysis is sufficiently robust in considering potential measures to
ensure selection of all appropriate RACT and RACM, and the State
provides a reasoned justification for its analytical approach, we will
consider approving that State’s RACT/RACM strategy.  

	Guidance on State analysis to identify RACT, RACM and appropriate
attainment date.  

A State must consider RACT and RACM for all of its nonattainment areas. 
However, EPA believes that if the State projects that an area will
attain the standard within 5 years of designation as a result of
existing national measures (i.e. projected to have a design value of
14.5 or lower), then the State may conduct a limited RACT and RACM
analysis that does not involve additional air quality modeling.  A
limited analysis of this type would involve the review of reasonably
available measures, the estimation of potential emissions reductions,
and the evaluation of the time needed to implement these measures.  If
the State could not achieve significant emissions reductions during 2008
due to time needed to implement the potential measures or other relevant
factors, then the State and EPA could conclude that there are no further
reasonably available control measures for that area that would advance
the attainment date by one year or more relative to the presumptive
outer limit for attainment dates, i.e., 5 years from designation.  In
lieu of conducting air quality modeling to assess the impact of
potential RACT and RACM measures, States may consider existing modeling
information to determine the magnitude of emissions reductions that
could significantly affect air quality and potentially result in
attaining prior to 2010 (e.g. in 2009 based on 2006-8 air quality data).
 If the State, in consultation with EPA, determines from this initial,
limited RACT and RACM analysis that the area may be able to advance its
attainment date through implementation of reasonable measures, then the
State would conduct a more detailed RACT and RACM analysis, including
appropriate air quality modeling analyses, to assess whether it can
advance the attainment date.  

	In general, the combined approach to RACT and RACM in the final rule
includes the following steps:  (1) identification of potential measures
that are reasonable; (2) modeling to identify the attainment date that
is as expeditiously as practicable; and (3) selection of RACT and RACM.

	Identification of potential measures:  The State’s review of
potential measures must be sufficient to identify all appropriate RACT
and RACM.  As stated previously, inherent to RACT/RACM is the basic
requirement that the measure be “reasonable.”  A State need not
evaluate measures in its RACM/RACT analysis that it determines are
unreasonable such as measures that are “absurd, unenforceable, or
impractical” or that would cause “severely disruptive socioeconomic
impacts, (e.g. gas rationing and mandatory source shutdowns); such
measures are not required by the Act.  55 FR 38327  

	As we also stated earlier, a State’s RACT/RACM analysis not only
involves an assessment about what emissions sources to control and to
what level, but also a judgement as to when it is reasonable to require
a sector to comply with a given measure.  Accordingly, if the State or
Federal rules already heavily regulate a given sector, it is reasonable
for the State to first look to unregulated parts of the sector for
RACT/RACM measures, especially, in light of costs already realized by
the regulated sector.  A State may conclude that it is unreasonable to
further regulate the industry, or that it is only reasonable to impose
measures in the latter years of the attainment plan.

	Finally, the State should use reason in the extent of its efforts to
identify potential control measures.  For example, if a review of
monitoring data and modeling studies indicates that reductions in SO2
are much more effective in reducing ambient PM2.5 than reductions in
other pollutants, we expect that the State will more vigorously identify
RACT/RACM measures for SO2 than for other pollutants.  Conversely, if
reductions in a given pollutant, even in large quantities, would have
trivial impacts on PM2.5, less rigor is needed in the State’s
assessment of controls for that pollutant, because such controls could
not contribute to advancing the attainment date.  Likewise, where
reducing emissions of a pollutant is effective in reducing ambient
PM2.5, if the emissions inventory for that pollutant is dominated by a
given type of emissions source, then it would be appropriate to focus
the analysis on measures for that segment of the inventory. No RACT/RACM
analysis is needed for pollutants that are not nonattainment plan
precursors for a particular PM2.5 nonattainment area.

	As supporting information for identification of RACT and RACM, the
State ordinarily provides data on technologically feasible control
measures:

-- A list of all emissions source categories, sources and activities in
the nonattainment area (for multi-State nonattainment areas, this would
include source categories, sources and activities from all states which
make up the area)

-- For each source category, source, or activity, an inventory of direct
PM2.5 and precursor emissions;

-- For each source category, source, or activity, a list of
technologically feasible emission control technologies and/or measures  

-- For each technologically feasible emission control technology or
measure, the State should provide the following information: (1) the
control efficiency by pollutant; (2) the possible emission reductions by
pollutant; (3) the estimated cost per ton of pollutant reduced; and (4)
the date by which the technology or measure could be reasonably
implemented.

	Based on this and other relevant information, the State will identify
the reasonable measures (potential RACT and RACM) to be included in air
quality modeling.  (At its option, the State may prefer not to make a
judgment on whether certain measures are technically and economically
feasible, if it believes they will not contribute to earlier attainment.
 In that case, the State could include those measures in the modeling,
and later exclude them from RACT and RACM by showing that all the
excluded measures together would not advance the attainment date by at
least 1 year.)  As previously mentioned, in determining the attainment
date that is as expeditiously as practicable, the State should consider
impacts on the nonattainment area of intrastate transport of pollution
from sources within its jurisdiction, and potential reasonable measures
to reduce emissions from those sources.  

	Modeling to determine the attainment date that is as expeditiously as
practicable:  Second, for purposes of determining the attainment date
that is as expeditiously as practicable, the State will need to conduct
modeling to show the combined air quality impact of all of the potential
measures identified in the first step with a modeling analysis for the
year 2009.  A base case scenario for the year 2009 would project future
air quality given implementation of existing measures (Federal, State
and local).  If this base case scenario demonstrates attainment by 2010,
then the State must demonstrate why attainment could not be achieved in
an earlier year.   (As noted above, given the April 2008 due date for
SIP submissions, it may be difficult to achieve earlier attainment in
many cases).   

	If the base case scenario does not demonstrate attainment, then a
control case scenario (described below) is needed to examine whether the
reasonable, technically and economically feasible measures identified by
the State would result in attainment in 2009. The control case scenario
would add potential SIP measures -- e.g. potential RACT/RACM, plus any
candidate intrastate transport measures that the State has identified
and would be feasible to implement by that year.  States in multi-State
nonattainment areas are strongly encouraged to collaborate on their
modeling analyses.  This modeling, along with other information known as
weight of evidence considerations, would inform a judgment as to whether
reasonable measures could lead to attainment of the standards within 5
years after designation.  If the analysis does not demonstrate
attainment by April 2010 (2009 analysis year), then the analysis would
serve as the technical basis for the State to seek an extension of the
attainment date for that area.  Further analysis would then be necessary
and is required to identify the specific attainment date. 

	The choice of future years to model beyond 2010 may vary from area to
area.  Often, modeling potential controls in two different future years
may be necessary to support a judgment that a projected attainment year
is as expeditiously as practicable.  If the area is projected to remain
over the standard in the early projection year (e.g., 2009) despite the
emission reductions from the modeled control measures, but is projected
to be well below the standard in the later projection year (e.g., 2012),
interpolation and emission inventory analysis could identify an
intermediate year as the appropriate attainment date.  There may be
cases in which modeling a single year is sufficient because modeling of
all technically and economically feasible controls results in attainment
by a narrow margin in that year.

	For many areas, EPA modeling analysis for CAIR and other modeling
analyses that have been performed suggest a number of nonattainment
areas will have a modest amount (in some cases only a few tenths of a
microgram) of needed reductions in ambient levels after 2010 to reach
attainment.  For any such area, and for areas otherwise expected to
attain relatively soon after 2010 (for example, due to substantial
reductions in a dominant local source), EPA believes that this analysis
should be for a year no later than 2012.  A later date (e.g., 2014) may
be appropriate for areas with very high PM2.5 levels that face
difficulty attaining within 10 years.

The EPA believes that it is not reasonable to require States to model
each and every year between 2009 and 2014 to determine the appropriate
attainment date.  Modeling future year inventories is a time consuming
and resource intensive process.  Multiple models and pre-processors are
needed in order to generate year specific emissions for the various
emissions sectors (e.g. mobile, non-road, non-EGU point, EGU point,
etc.).  Because it is not reasonable to model every year, a logical
choice often may be to model a year in the middle of the period.  As
such, we recommend modeling an emissions year no later than 2012 as the
initial extension date (which translates to a 2013 attainment date).  If
this modeling indicates that the area can reach attainment by 2012, then
the State can further analyze emissions and strategies to determine if
the attainment date can be advanced to an earlier year.  If the modeling
indicates that the area cannot reach attainment by 2012, then the
modeling will serve as further justification for granting a longer
attainment date extension (e.g., attainment date of 2015 with modeling
for 2014). In that case, additional modeling of 2014 with further
emissions controls would be required in order to show attainment. Again,
the State should then further analyze emissions and strategies to
determine if the attainment date can be advanced to an earlier year
between 2012 and 2015.

Additionally, in the discussion of air quality modeling issues in
section II.E above, we discuss the benefits of addressing control
strategies for multiple pollutants. Part of the challenge of
multi-pollutant modeling is coordinating the future modeling years for
different pollutants in order to minimize the number of required future
year model runs. As part of the requirements of the 8-hour ozone
implementation rule, States are currently working on modeling analyses
for 2009 and in some cases for 2012 (serious nonattainment areas).  For
an area that cannot attain the PM2.5 NAAQS by 2010, this may be reason
to select 2012 as the year to model, so that the State could conduct the
modeling for both ozone and PM2.5 in tandem.  This would, in some cases,
allow the pooling of resources (e.g., inventories, model runs, etc.) and
provide for faster development of a PM2.5 attainment demonstration. 

It may also be possible for the State to look at 2009 and 2014 only.  In
this instance, the State may find sufficient data to interpolate results
for the years in between based on estimated changes in emissions. 

	We emphasize that when a State models later years, that this analysis
must take into account potential controls that the State may previously
have determined not to be RACT or RACM.  For example, some measures that
are impractical to implement by 2009 could be reasonable if implemented
by 2010, 2011 or 2012.  Thus, when the State models later years, the
list of potential controls should be expanded to include technically and
economically feasible measures that can be implemented by the analysis
year.  

	Selection of RACT & RACM:  Based on this analysis, the State should
make decisions on RACT, RACM, intrastate measures, and the attainment
date that is as expeditiously as practicable.  Because EPA is defining
RACT and RACM as only those reasonable, technically and economically
feasible measures that are necessary for attainment as expeditiously as
practicable, the State need not adopt all feasible, reasonable measures.
 The State may exclude those reasonable measures that, considered
collectively, would not advance the attainment date. 

Comments and responses.

	Comment:  A number of commenters generally supported EPA’s second
proposed alternative to RACT (option 2).  Most of these commenters
expressed concern that the other options would require the imposition of
controls whether or not they were needed to attain the PM2.5 standards
as expeditiously as practicable. Some State and local commenters also
urged EPA to select option 2 as the best interpretation of the RACT
requirement for PM2.5 because they believe that it will be the most
appropriate approach for designing attainment strategies for their
particular nonattainment area or areas.   

	Response:  The EPA agrees that these two points are important
considerations. After carefully considering the options, we concluded
that Option 2 was the most suitable approach for the PM2.5 NAAQS.
Options 1 and 3 do not reduce the States’ burden to analyze potential
control measures as the States would still be required to look beyond
the mandated RACT for reasonably available measures.  Moreover, Options
1 and 3 could require imposition of controls on some sources that would
not strictly be necessary to attain the NAAQS as expeditiously as
practicable.  Given the nature of the PM2.5 nonattainment problem, EPA
concluded that an interpretation that provides the maximum flexibility
is a better approach.

	Comment:   Some commenters recommended that EPA modify proposed option
2 to include a tons-per-year threshold.   Under such an approach, the
States and EPA would only require RACT for sources whose emissions were
above the threshold.  Most of these comments recommended a RACT
threshold of 100 tons per year.  These commenters expressed concern that
if option 2 were implemented without such a threshold, States would be
burdened with conducting RACT analyses for very small sources or source
categories with low emissions. 

	Response:  The EPA believes that under the approach chosen for the
final rule in which RACT is considered to be a part of the overall RACM
process, it would be difficult to define a threshold that would apply
for all types of sources and for all types of control measures in all
nonattainment areas.  It has not been common practice under past EPA
policy to establish or use an emissions threshold when considering
sources for possible emission reductions as part of a RACM analysis to
show attainment as expeditiously as practicable.  Indeed, many of the
control technique guidelines for VOC RACT do not recommend an emissions
threshold.   A state needing significant emission reductions to attain
the standards in a given area even by 2015 would likely conclude that
controls should be considered on smaller sources.  In contrast, a State
with an area that exceeds the standard by only a few tenths of a
microgram per cubic meter may not need to consider controls on smaller
source to reach attainment as expeditiously as practicable.  The EPA has
selected option 2 for interpretation of the RACT requirement for PM2.5,
in part, specifically because that approach contemplates that States
will conduct an appropriate analysis of the spectrum of source
categories and potential controls available.  To cut off such analysis
at a set emissions-based cut point for all sources and all areas would
undermine one of the key benefits of the approach.  Accordingly, EPA
disagrees with comments that option 2 should include a
nationally-defined threshold for the size of sources or source
categories that require RACT analyses. 

	Comment:   A number of commenters supported EPA’s first and third
proposed alternative approaches to RACT (option 1 and option 3). 
Commenters supporting these two options used similar reasoning. 
Commenters cited the statutory language in section 172(c)(1) requiring
that the nonattainment plan provide for “at a minimum” the adoption
of RACT.  Accordingly, these commenters argued that RACT is an
independent, minimum requirement of nonattainment plans irrespective of
the attainment demonstration and that option 2, which would not require
the adoption of RACT for all sources, has no policy or legal
justification. Other commenters noted that option 1 would be much easier
to implement, because RACT would be defined according to technical
reasonableness and would not hinge on complicated determinations
involving attainment demonstrations.  Some commenters argued that option
1 provides for greater equity, because similar measures would be
required for similar sources for all nonattainment areas.   Finally,
some commenters believed that it is inherently inconsistent to assert
that plans have met the requirement for attainment “as expeditiously
as practicable” without applying RACT to all major sources. 

	Response:  The EPA disagrees with these comments.  The EPA believes
that option 2 is fully consistent with section 172(c)(1).  Section
172(c)(1) requires that nonattainment plans must provide for the
implementation of RACM as expeditiously as practicable (including such
reductions in emissions from existing sources in the area as may be
obtained through the adoption, at a minimum, of RACT).  Contrary to the
commenters’ assertions, this language does not demonstrate that RACT
is required for all sources, independent of RACM and attainment
demonstrations.   Moreover, this provision does not require RACT whether
or not imposition of technology would advance the attainment date. 
Instead, section 172(c)(1) explicitly provides that RACT is included
within the definition of RACM, and EPA has previously determined that
the CAA only requires such RACM as will provide for attainment as
expeditiously as practicable.  (See 57 FR 13498, 13560).   The courts
have deferred to this interpretation and concluded that EPA interprets
RACM as a collection of reasonable measures that would advance the
attainment date.   See Sierra Club v. EPA, 294 F.3d 155, 162 (D.C. Cir.
2002); see also Sierra Club v. EPA, 314 F.3d 735, 744 (5th Cir. 2002). 
The CAA does not “compel [ ] a State to consider whether any measure
is ‘reasonably available’ without regard to whether it would
expedite attainment in the relevant area.”  Sierra Club v. EPA, 294
F.3d at 162.  The EPA concludes that because section 172(c)(1)
establishes that RACT is a part of RACM, EPA is reasonably applying the
same interpretation to the RACT requirement for PM2.5.  The RACT is a
part of the collection of measures that are necessary to reach
attainment as expeditiously as practicable. It is thus directly related
to what a specific area needs to attain the NAAQS, and States need not
implement reasonably available measures that would not advance the
attainment date as part of the PM2.5 RACT requirement. 

	The EPA also finds that option 2 is consistent with the statutory
language providing that a State must apply RACT to existing sources,
“at a minimum,” to meet its requirement to apply RACM.  We interpret
the “at a minimum” clause to mean that when a State determines that
control of a specified existing stationary source(s) is necessary to
attain, the State must apply RACT to that source.  Further, EPA believes
this requirement for RACT applies to stationary sources as a group, and
not to each stationary source.  

The EPA finds sound policy reasons for choosing option 2.  While an
approach that provided for application of the same controls in all areas
would provide for more equity across areas, EPA emphasizes that equity
is only one of many factors considered by EPA when deciding between
options 1, 2 and 3. The EPA believes that it is also important to ensure
that control strategies focus on the most effective measures with the
greatest possibility for significant air quality improvements.   In
addition, while EPA agrees that options 1 and 3 could provide for
greater ease of implementation, this is also only one of the factors EPA
considered when deciding between the proposed options. Under option 2,
States have a greater burden and responsibility to identify the local
strategy that is tailored to their particular air quality problem.   At
the same time, the States have the ability to identify the sources with
the greatest impact on nonattainment and to identify a sound strategy
that achieves attainment in the most sensible manner.  The EPA believes
that approaching RACT and RACM in this manner is consistent with the
overall philosophy imbedded in the SIP program since its inception in
the late 1960s and early 1970s.

	Comment:   Some commenters believed that the proposed RACM requirement
was too broad.  These commenters believed that the requirement to
analyze the entire “universe” of possible measures was too
burdensome for States.  Commenters felt this was especially true in
light of the lack of federally issued CTG and ACT documents for PM2.5
and its precursors for all potential source categories.   

	Response:  As explained earlier, States should apply “reason” in
identifying measures to evaluate as potential RACM/RACT.  We recognize
that States are implementing the PM2.5 standard for the first time, and
do not have the long history and experience in implementing PM2.5 as
they have in implementing the PM10 and ozone standards.  Accordingly, we
expect that both the States and EPA will expend extra effort in
developing and evaluating nonattainment plans that contain appropriate
controls.  A number of resources exist to provide States with
information on potential control measure costs and emissions reductions.
 We intend to facilitate the sharing of information through a control
measure website and other efforts, and expect that States will develop
screening methods to reduce the burden of analysis.

	Comment:  One commenter asserted that EPA should not require the
analysis for, or implementation of, RACT and RACM for sources throughout
the entire nonattainment area, and should permit States to focus only on
sources located in smaller specific “problem areas” within the
nonattainment area.

	Response:   The EPA designated areas nonattainment based upon analysis
of the geographic area with sources that “contribute” to the
violation of the NAAQS in the area, in accordance with section 107(d).
These designations are based upon, among other things, a network of
monitors that the State and EPA previously agreed represented the
ambient air concentrations throughout the area.  Additional analysis of
information during the designation process indicated those areas that
contributed to the violations at the violating monitor because of, among
other things, the amount of emissions in such adjoining areas. 
Accordingly, the State in which a nonattainment area is located must
evaluate the full range of sources of PM2.5 and its precursors
throughout the designated nonattainment area during the development of
the SIP.  The EPA agrees that there are some nonattainment areas where
one or a few large emissions sources may be causing localized
concentrations at a monitor that are much higher than those within the
remainder of the nonattainment area.   For such areas, the nonattainment
strategy will likely not succeed without addressing those sources.   The
EPA does not, however, believe it is acceptable that the nonattainment
strategy focus only on those sources, because additional reductions
within the nonattainment area would still have the potential to advance
the attainment date.  Exempting portions of the nonattainment area could
expose a portion of the public residing downwind in the area to exposure
to levels of PM2.5 that exceed the NAAQS for longer than necessary, and
the health detriments from such exposure, merely to minimize the impact
of having to impose control strategies on sources upwind.   Moreover, to
the extent that monitoring in one portion of a nonattainment area
indicates violations in multiple portions of the area, a strategy that
solely focused upon the sources in the immediate vicinity of the monitor
might fail to assure that the NAAQS is achieved throughout the area. 
Because NAAQS violations generally reflect a combination of regional
scale, metropolitan scale, and local scale impacts, and all three scales
must be addressed, EPA requires RACT/RACM submittals to address sources
throughout the nonattainment area.

Comment:  Some commenters agreed with EPA’s view   that State’s RACM
analysis must address those measures that a State declines to adopt and
must show  whether the combined measures would cumulatively advance the
attainment date by at least 1 year.  One commenter questioned the legal
basis for EPA’s determination that the only controls necessary to
attain the PM2.5 NAAQS as expeditiously as practicable are those that
would cumulatively advance an area’s projected attainment date by at
least one calendar year.  The commenter suggested that control measures
that would advance attainment by a smaller increment “would meet the
criteria endorsed in Sierra Club [Sierra Club v. EPA, 294 F.3d 155 (D.C.
Cir 2002)] by ‘expedit[ing] attainment in the relevant area.’”

Response:  The EPA has consistently interpreted RACM as a collection of
measures that would advance the attainment date by at least 1 year, and
the courts have determined that the statutory RACM requirement is
ambiguous and deferred to EPA’s interpretation of the requirement. 
See Sierra Club. v. EPA, 314 F.3d 735, 744 (5th Cir. 2002); see also
Sierra Club v. EPA, 294 F.3d, 155 162 (D.C. Cir. 2002).  Contrary to the
commenter’s suggestion, the court in Sierra Club v. EPA, did not
endorse specific criteria for identifying control measures that expedite
attainment, but instead deferred to EPA’s interpretation of an
ambiguous statutory term.  The courts deferred to EPA’s interpretation
after reviewing EPA’s approval of State SIP submissions.  The EPA
conducts such reviews consistent with its determination that the CAA
only requires such RACM as will provide for attainment as expeditiously
as practicable, and its belief that it would be unreasonable to require
implementation of measures that would not in fact advance attainment. 
See 57 FR 13498, 13560 (April 15, 1992); see also 44 FR 20372, 20374
(April 4, 1979).   In considering whether a collection of measures would
advance the attainment date of an area, EPA has previously interpreted
the phrase “advance the attainment date” as meaning that the
attainment date would be advanced by at least 1 year.   See e.g., 66 FR
57160, 57182 (Nov. 14, 2001) (approval of Houston SIP); 66 FR 586 (Jan
3. 2001) (approval of DC area SIP).  Further, EPA’s use of a one-year
increment in determining whether a collection of measures would advance
the attainment date is reasonable and consistent with the fact that all
areas will be designing attainment demonstrations for the annual PM2.5
standard.  Section 172(a)(2)(C) statute uses 1 year as the increment by
which attainment date extensions can be granted. Thus, requiring
evaluation of whether control measures would advance attainment by an
increment of 1 year is a reasonable approach for the PM2.5 NAAQS.

	Comment:   Some commenters recommended that EPA consider not requiring
a RACM analysis for areas projected to attain the standards within 5
years of designation, i.e., by April 2010 for the areas currently
designated nonattainment.  One commenter suggested that practical
considerations would make it impossible for any State projected to
attain by 2010 to advance the attainment date by a year.  This commenter
noted that because measures to provide for attainment by 2010 must be
implemented by the beginning of 2009, and SIPs are not submitted until
April 2008, it would impossible to advance the implementation of
measures by 1 year (that is, the beginning of 2008). 

	Response:  The EPA generally agrees that given the time constraints it
will be difficult for States with areas currently designated
nonattainment to devise, adopt, and implement RACM measures to advance
the attainment date before 2010.   At the same time, however, we note
that nothing precludes States from taking early action and we encourage
States to take actions to reduce PM2.5 concentrations where feasible
even before the SIPs are submitted.  RACM is required by the CAA and
thus EPA cannot waive the requirement for the analysis.  At the same
time, EPA recognizes that a streamlined analysis may be appropriate
given the short time periods involved.

3.   Observations and considerations in determining RACT and RACM tc
"15.What factors should States consider in determining whether control
measures are reasonably available? " \l 3 

a. Background. 

	The preamble to the proposed rule included a discussion of general
considerations for RACT (70 FR 66020 and 66021, latter part of section
III.I.6) and RACM (70 FR 66028, section III.1.15).   The preamble to the
final rule retains this discussion with some modifications and
restructuring to reflect the combined approach to RACT and RACM

b. Final rule.   

	General considerations.  Once the State has identified measures and
technologies that are available for implementation in the nonattainment
area, then it must evaluate those measures to determine whether
implementation of such measures are reasonable, and would collectively
advance attainment.  Many of the factors that the State should take into
consideration in determining whether a measure is “reasonable” are
related to the measure’s technical and economic feasibility. Since
RACM applies to area and mobile sources as well as stationary sources,
the State should consider other factors as well in conducting its RACM
analysis.  For example, in many cases obtaining emissions reductions
from area and mobile sources is achieved not by adding control
technology to a specific emissions source, but by reducing the level of
activity of a fleet of vehicles or by modifying a type of commercial
process.  In these situations, the State should also consider local
circumstances such as infrastructure, population, or workforce and the
time needed to implement the measure in light of the attainment date.

	The EPA believes that while areas projected to attain within 5 years of
designation as a result of existing national measures should still be
required to conduct a RACT and RACM analysis, such areas may be able to
conduct a limited RACT and RACM analysis that does not involve
additional air quality modeling.  A limited analysis of this type could
involve the review of available reasonable measures, the estimation of
potential emissions reductions, and the evaluation of the time needed to
implement these measures.  If the State could not achieve significant
emissions reductions by the beginning of 2008 due to time needed to
implement reasonable measures or other factors, then it could be
concluded that reasonably available local measures would not advance the
attainment date.  In lieu of conducting air quality modeling to assess
the impact of potential RACT and RACM measures, existing modeling
information could be considered in determining the magnitude of
emissions reductions that could significantly affect air quality and
potentially result in earlier attainment.  If the State, in consultation
with EPA, determines from this initial, more limited RACT and RACM
analysis that the area may be able to advance its attainment date
through implementation of reasonable measures, then the State would
conduct a more detailed RACT and RACM analysis.

	Observations on control opportunities.  The implementation of the PM2.5
NAAQS is in its initial stages, and many of the designated PM2.5
nonattainment areas are not current or former PM10 nonattainment areas. 
Thus, some existing stationary sources in these areas may currently be
uncontrolled or undercontrolled for PM2.5 or PM2.5 precursors.  Further,
to this point in time, emissions controls for existing sources in these
areas may have focused primarily on particulate matter that is
filterable at stack temperatures and thus may not adequately control
condensable emissions.  In addition, States should bear in mind that the
controlled sources may have installed emission controls 15 years ago or
more, and there may now be cost-effective opportunities available to
reduce emissions further through more comprehensive and improved
emissions control technologies, or through production process changes
that are inherently lower in emissions.  

	Moreover, improved monitoring methods may enhance the ability of
sources to maintain the effectiveness of installed emissions controls
and to reduce emissions by detecting equipment failures more quickly. 
For example, State imposition of requirements for more frequent
monitoring (e.g., continuous opacity monitors, PM continuous emissions
monitors, etc.) may provide greater assurance of source compliance and
quicker correction of inadvertent upset emissions conditions than
existing approaches. 

	Even in former or current PM10 nonattainment areas, existing
requirements for controlling direct PM emissions (e.g., with a baghouse
or electrostatic precipitator) may not have been revised significantly
since the 1970's.  When EPA established the PM10 standards in 1987, we
stated in the  preamble that it was reasonable to assume that control
technology that represented RACT and RACM for total suspended
particulates (TSP) should satisfy the requirement for RACT and RACM for
PM10.  52 FR 24672 (July 1, 1987).   The  basis for EPA’s belief was
that controls for PM10 and TSP would both focus on reducing coarse
particulate matter, and specifically that fraction of particulate matter
that is solid (rather than gaseous or condensable) at typical stack
temperatures.  However, emission controls to capture coarse particles in
some cases may be less effective in controlling PM2.5.  For this reason,
there may be significant opportunities for sources to upgrade existing
control technologies and compliance monitoring methods to address direct
PM emissions contributing to fine particulate matter levels with
technologies that have advanced significantly over the past 15 years.  

	Precursor Controls.  It will be important for States to conduct RACT
and RACM determinations for stationary sources of PM2.5 precursors as
well as direct PM2.5 emissions although, as noted above, the known
atmospheric chemistry of the area may dictate the necessary rigor of
this analysis.  A significant fraction of PM2.5 mass in most areas
violating the standards is attributed to secondarily-formed components
such as sulfate, nitrate, and some organic PM, and EPA believes that
certain stationary sources of precursors of these components in
nonattainment areas currently may be poorly controlled.  Accordingly, to
address these precursors, States should review existing sources for
emission controls or process changes that could be reasonably
implemented to reduce emissions from activities such as fuel combustion,
industrial processes, and solvent usage. 

Multi-State nonattainment areas.  States in multi-State nonattainment
areas will need to consult with each other on appropriate level of RACT
and RACM for that area. We anticipate that States may decide upon RACT
and RACM controls that differ from State to State, based upon the
State’s determination of the most effective strategies given the
relevant mixture of sources and potential controls in the relevant
nonattainment areas.  So long as each State can adequately demonstrate
that its chosen RACT and RACM approach will provide for meeting RFP
requirements and for attainment of the NAAQS as expeditiously as
practicable for the nonattainment area at issue, we anticipate approving
plans that may elect to control a somewhat different mix of sources or
to implement somewhat different controls as RACT and RACM. 
Nevertheless, States should consider RACT and RACM measures developed
for other areas or other States. EPA may consider such measures in
assessing the approvability of a State’s SIP. 

c. Comments and Responses

	Comment:  In the proposed rule, EPA indicated that States could
consider the “social acceptability” of measures as a factor in the
determination of what constitutes RACM in a given area.   A number of
commenters recommended that EPA eliminate use of this factor.  Some
commenters questioned whether States or EPA had the legal authority to
exclude measures from consideration based on social acceptability or
popularity, if the measures are technically and economically available,
and are needed to attain the NAAQS for protection of public health.  
Others expressed concerns that inclusion of such a factor would
inevitably result in the elimination of controls for area and mobile
sources and for this reason would unfairly focus emissions reduction
strategies on industrial sources of PM2.5 and precursors.

	Response:  The EPA believes that in developing RACM measures, it is
important that States not rely unduly on measures that would be very
difficult to enforce in practice.  We discourage States from relying on
measures that on paper may seem reasonably available but in practice
might fail to achieve benefits due to resistance to implementing the
measures.  However, we recognize that the CAA does not identify
“social acceptability” as a factor in the definition of what may
constitute RACT or RACM, and more generally the CAA does not establish a
preference for measures that affect industrial sources instead of the
general public and are therefore more likely to be “socially
acceptable.”  Therefore, given the concerns raised by commenters that
establishment of “social acceptability” as a factor in the RACM
analysis is without basis in the CAA and might result in inappropriate
skewing of control strategies, we have removed this term from the final
rule. We reiterate, however, that capability of effective implementation
and enforcement are relevant considerations in the RACM analysis, even
though public “unpopularity” is not.  Moreover, in assessing the
efficacy of measures and the credit they should be given in the context
of attainment demonstrations or RFP calculations, EPA believes that such
considerations are important.  

4.	What factors should States consider in determining whether an
available control technology or measure is technically feasible?

a. Background.

	 The following provides guidance for States to consider in determining
whether an available control technology is technologically feasible.  

b.  Final Rule. 

	The technological feasibility of applying an emission reduction method
to a particular source should consider factors such as the source’s
process and operating procedures, raw materials, physical plant layout,
and any other environmental impacts such as water pollution, waste
disposal, and energy requirements.  For example, the process, operating
procedures, and raw materials used by a source can affect the
feasibility of implementing process changes that reduce emissions and
the selection of add-on emission control equipment.  The operation and
longevity of control equipment can be significantly influenced by the
raw materials used and the process to which it is applied.  The
feasibility of modifying processes or applying control equipment also
can be influenced by the physical layout of the particular plant.  The
space available in which to implement such changes may limit the choices
and will also affect the costs of control.  

	Reducing air emissions may not justify adversely affecting other
resources by increasing the pollution of bodies of water, creating
additional solid waste disposal problems or creating excessive energy
demands.  An otherwise available control technology may not be
reasonable if these other environmental impacts cannot reasonably be
mitigated.  For analytic purposes, a State may consider a PM2.5 control
measure technologically infeasible if, considering the availability (and
cost) of mitigating adverse impacts of that control on other pollution
media, the control would not, in the State's reasoned judgment, provide
a net benefit to public health and the environment.  However, in many
past situations, States and owners of existing sources have adopted
PM2.5 control technologies with known energy penalties and some adverse
effects on other media, based on the reasoned judgment that installation
of such technology would result in a net benefit to public health and
the environment.  States should consider this in determining technical
feasibility.  The costs of preventing adverse water, solid waste and
energy impacts should be included in assessing the economic feasibility
of the PM2.5 control technology.

	One particular cross-media issue relates to concentrated animal feeding
operations (CAFOs).  Since 2003, EPA and many stakeholders have been
interested in developing a framework to enable CAFOs to pursue superior
environmental performance across all media.  We are aware that today
some CAFOs voluntarily conduct whole-farm audits to evaluate releases of
pollutants to all media through Environmental Management Systems,
self-assessment tools, performance track, ISO 14001 certification, and
State-approved trade offs in meeting regulatory thresholds between air
and water that accomplish the best overall level of environmental
protection given State and local conditions.  The EPA continues to
believe the development of new and emerging technologies offers the
potential to achieve equivalent or greater pollutant reductions than
achieved solely by effluent guidelines and standards.  Many of these are
superior from a multimedia perspective, and EPA would like to encourage
superior multimedia solutions.  SIPs which need to address ammonia may
provide a unique opportunity to encourage multimedia approaches at
CAFOs.  For example, the addition of animal by-products provides a
valuable source of nutrients for crops, improves soil structure which
enhances soil permeability, and adds valuable organic matter that
improves soil health.  However, inappropriate application can lead to
air and water quality concerns or the improvement of one media at the
cost of another.  Optimal application technologies and rates reduce
potential air and water quality standards violations.  The EPA does not
want to discourage approaches that are superior from a cross media
perspective.

	The EPA recommends that States evaluate alternative approaches to
reducing emissions of particulate matter by reviewing existing EPA
guidance and other sources of control technology information.  The
EPA’s 1998 guidance presents information on topics such as the design,
operation and maintenance of general particulate matter control systems
such as electrostatic precipitators, fabric filters, and wet scrubbers. 
The filterable particulate matter collection efficiency of each system
is discussed as a function of particle size.  The guidance document also
provides information concerning other relevant considerations such as
energy and environmental considerations, procedures for estimating costs
of particulate matter control equipment, and evaluation of secondary
environmental impacts.  Because control technologies and monitoring
approaches are constantly being improved, the State should also consider
more updated or advanced technologies not referenced in this 1998
guidance when conducting a RACT determination.  Emissions reductions may
also be achieved through the application of monitoring and maintenance
programs that use critical process and control parameters to verify that
emission controls are operated and maintained so that they more
continuously achieve the level of control that they were designed to
achieve.

c. Comments and Responses

	Comment:  One commenter noted that the guidance for “technical
feasibility” implies that States look at individual sources with a
BACT-like case-by-case analysis.  The commenter recommended that source
owners conduct such a site-specific analysis and submit the analysis to
the State through the permitting process.

	Response:    While the analytical analysis to identify RACT is similar
to BACT, as noted above, EPA in the past has issued CTGs that describe
the presumptive norm for RACT controls for a given industry, but that
allow for case-by-case considerations for a given source.  Where States
wished to require source owners to conduct such a site-specific analysis
as part of the control technology review, EPA supports this type of
process.  On the other hand, EPA does not believe it would be
appropriate to require all RACT-eligible sources to conduct such an
analysis, given that States have the primary responsibility for
identifying and analyzing measures for such sources.

5.  What factors should States consider in determining whether an
available control technology or measure is economically feasible? tc
"7.What factors should States consider in determining whether an
available control technology is economically feasible? " \l 3 

a. Background.

	The follow provides guidance for States to consider in determining
whether an available control technology is economically feasible for
purposes of identifying reasonably available control measures.  This
guidance is slightly modified from our proposal. 

b.   Final Rule.

	 Economic feasibility encompasses considerations such as whether the
cost of a potential measure is reasonable considering attainment needs
of the area and the costs of other measures, and whether the cost of a
measure is reasonable for the regulated entity to bear, in light of
benefits.

	While many States generally establish RACT requirements for a category
of sources, the Act does not require the same level of control on all
sources in a category, nor does the Act require that each source be
controlled individually.  Similar sources may have different marginal
costs, profit margins and abilities to pass costs through to the
consumer.  These factors are appropriate to consider in determining
whether a given level of control is appropriate for an individual source
or category of sources.  Accordingly, there is no presumption that a
given source must bear a cost similar to any other source.  

	States should consider the capital costs, annualized costs, cost
effectiveness of an emissions reduction technology, and effects on the
local economy in determining whether a potential control measure is
reasonable for an area or State.  One available reference for
calculating costs is the EPA Air Pollution Control Cost Manual, which
describes the procedures EPA uses for determining these costs for
stationary sources. The above costs should be determined for all
technologically feasible emission reduction options if such measure is
inherently “reasonably available” (e.g., not absurd or clearly
impractical).   States may give substantial weight to cost effectiveness
in evaluating the economic feasibility of an emission reduction
technology.  The cost effectiveness of a technology is its annualized
cost ($/year) divided by the emissions reduced (i.e., tons/year) which
yields a cost per amount of emission reduction ($/ton).  Cost
effectiveness provides a value for each emission reduction option that
is comparable with other options and other facilities.  Where multiple
control options exist for a given source or source category, States
should consider both the cost effectiveness (dollars per ton) of each
option, and the incremental cost effectiveness per ton between the
options (incremental increase in cost between options divided by the
incremental tons reduced).

In determining whether a given measure is reasonable, States may
consider costs per ton of other measures previously employed to reduce
that pollutant, but similar costs are not conclusive.  As discussed
above, States may evaluate equity considerations in weighing the
economic feasibility of imposing a measure on a given source or source
category.   

We anticipate that States may decide upon RACT and RACM controls that
differ from State to State, based on the State’s determination of the
most effective strategies given the relevant mixture of sources and
potential controls in the relevant nonattainment areas, and differences
in the difficulty of reaching attainment.  

	In considering what level of control is reasonable, EPA is not
proposing a fixed dollar per ton cost threshold for RACT, consistent
with the views of multiple commenters.  Areas with more serious air
quality problems typically will need to obtain greater levels of
emissions reductions from local sources than areas with less serious
problems.  Where essential reductions are more difficult to achieve
(e.g., because many sources are already controlled), the cost per ton of
control may necessarily be higher.  

It is not appropriate to assume that the same cost per ton range is
reasonable for direct PM2.5 and different precursors, because an equal
amount of emission reduction in different pollutants has a different
impact on PM2.5 ambient levels.  For example, in a given nonattainment
area, reductions of direct PM2.5 emissions may prove more expensive than
reductions of Nox emissions, but the resulting benefits of reductions of
direct PM2.5 might warrant the higher costs.  A State should consider
this differential impact on ambient PM2.5 in considering the cost per
ton that is reasonable for controlling different pollutants.  During the
SIP process, States and regional planning organizations typically
conduct sensitivity modeling that can provide this information.  Also,
the PM NAAQS RIA provides information on the differential impact of
PM2.5 and PM precursor reductions on ambient PM2.5 levels in various
areas.

One of the factors that could affect estimated compliance costs of an
emission reduction measure is the timing of its implementation. 
Hypothetically, if a short compliance period were contemplated for a set
of sources, and if the short compliance period resulted in high demand
for a limited supply of labor or other resources, compliance costs could
be higher than if the same measure were implemented by a later
compliance date.  In such a case it may be reasonable for the State to
find that the measure is reasonable only if implemented by the later
date. 

	If a source contends that a source-specific RACT level should be
established because it cannot afford the technology that appears to be
RACT for other sources in its source category, the source can support
its claim with such information as:

-- fixed and variable production costs ($/unit)

-- product supply and demand elasticity,

-- product prices (cost absorption vs. cost pass-through),

-- expected costs incurred by competitors,

-- company profits once the technology or measure is in operation
(considering the annualized costs and the marginal costs of alternative
technologies and measures), , 

-- employment costs, and

-- any other unique factor(s) particular to the individual source.

	Finally, the EPA clarifies that if the State demonstrates through
economic analysis that the imposition of the measure would cause
unacceptable economic disruption for the local economy, that is, a plant
shutdown or a severe curtailment in plant employment or output, a State
may reject the measure as not reasonable to reach attainment as
expeditiously as practicable.

c. Comments and Responses. 

	Comment:   Some commenters agreed with EPA’s proposal not to
establish presumptive cost-effectiveness thresholds. 

	Response:  The EPA agrees with the commenters.

	Comment:  A number of commenters expressed concerns over the references
to health benefits as a consideration in whether measures are
technically or economically available.   Some commenters believed this
is a consideration not authorized by the CAA.  Others believed that
consideration of benefits, in combination with EPA’s estimates of
benefits per ton, would have the effect of converting RACT to more
stringent LAER levels.  Some commenters expressed concerns whether
States had the resources or expertise to conduct cost-benefit analyses
for this purpose.  

	Response:   The EPA wishes to clarify that the reference to health
benefits does not mean that a cost-benefit, or a detailed health
benefits assessment, is a necessary part of a control strategy
demonstration.   We also wish to clarify that EPA is not requiring that
the costs of all technologies and measures for PM2.5 and precursors be
deemed acceptable at any dollar/ton levels at or below the calculated
monetized benefits per ton of reduction.   We do, however, continue to
believe that the significant benefits associated with PM2.5 ambient
reductions is a relevant consideration in control strategy development. 
We believe that the general level of benefits is something States and
the public should consider when making decisions on whether to pursue
control measures and technologies.   The EPA disagrees that this limited
consideration of benefits would convert the RACT process to the
equivalent of LAER.  Because RACT is linked to what is necessary to
attain as expeditiously as practicable, we disagree that any
consideration of benefits would convert RACT to LAER.

	Comment:  One commenter objected to EPA’s proposed requirement that
States consider competitive factors such as production costs, demand
elasticity, product prices, and cost incurred by competitors in the
determination of RACT. The commenter believed that this information is
generally not accessible to States or industrial facility owners, and is
not necessary for a RACT determination.  

	Response:   The EPA generally disagrees that this type of information
is unavailable.  For example, EPA calculates or reviews this type of
data on a regular basis as part of our work on MACT, NSPS, and other
emissions standards.   A document that describes these types of analyses
and the data used to prepare them is the OAQPS Economic Resource Manual
found at     HYPERLINK "http://www.epa.gov/ttn/ecas/analguid.html" 
http://www.epa.gov/ttn/ecas/analguid.html .    EPA believes that this
issue is most relevant to category-wide RACT rules where a source seeks
a case-by-case exemption.   Further, EPA believes most RACT
determinations will be developed through case-by-case analyses rather
than rules affecting entire source categories.  Accordingly, this
analysis likely will be relevant in few cases.   

6.  What specific source categories and control measures should a State
evaluate when determining RACT and RACM for a nonattainment area? tc
"16.What specific source categories and control measures should a State
evaluate when determining RACM for a nonattainment area? " \l 3 

a.  Background.

	 Section 172 does not provide a specific list of source categories and
control measures that must be evaluated for RACT and RACM for PM2.5. 
However, section 172(c)(3) indicates that the nonattainment area plan
must include a “comprehensive, accurate, current, inventory of actual
emissions from all sources of the relevant pollutant.”  This indicates
that States should look broadly at the different types of sources in the
nonattainment area.  We recognize that PM2.5 is a new NAAQS without a
long history of implementation as with ozone.  Therefore, we included a
list of potential RACM measures in the preamble to the proposed rule,
based upon a review of information about the contribution of various
sources to emissions inventories and a review of potential control
measures for such sources.  We requested comment on the specific sources
and potential control measures recommended for RACM analysis on this
list.   Based on comments received and additional information available
to EPA since the proposal, we have made some changes to the list. We
also refer to this list of potential “RACT and RACM” measures for
the combined approach to RACT and RACM in the final rule.

	In the preamble to the proposed rule, EPA indicated that due to the
short time available, it does not plan to develop new control techniques
guidance (CTG) or ACT documents specifically for purposes of PM2.5
implementation.  The EPA indicated that other information was available
on control technologies, and EPA also indicated its intention to
maintain an updated list of references for new PM2.5 control technology
information.  

b. Final rule.

 Emission reduction measures constituting RACM should be determined on
an area-by-area basis.  We believe that a State should consider each of
the measures listed in this section to determine if each measure is
reasonably available in the applicable nonattainment area.  However, we
do not presume that each of these measures is reasonably available in
each nonattainment area. 

	We recommend that each State use the list of source categories in this
section as a starting point for identifying potentially available
control strategies (regulatory and voluntary) for a nonattainment area. 
States are encouraged and expected to add other potentially available
measures to the list based on its knowledge of the particular universe
of emissions sources in the area and comments from the general public. 
We expect that, depending on the potential measure being analyzed, the
State’s degree of evaluation will vary as appropriate.  Detailed
information on emission control technologies is available from a number
of sources.   EPA intends to maintain a website with links to sources of
information for controlling emissions of direct particulate matter and
PM precursors.  

	As discussed in section II.J.5. above, EPA recognizes that control
technology guidance for certain source categories has not been updated
for many years.  Section 183(c) of the CAA, which addresses control
technologies to address ozone nonattainment problems, requires EPA to
“revise and update such documents as the Administrator determines
necessary.”  As new or updated information becomes available States
should consider the new information in their RACT determinations.  A
State should consider the new information in any RACT determinations or
certifications that have not been issued by the State as of the time
such updated information becomes available.

Stationary source measures

- Stationary diesel engine retrofit, rebuild or replacement, with
catalyzed particle filter

- New or upgraded emission control requirements for direct PM2.5
emissions at stationary sources (e.g., installation or improved
performance of control devices such as a baghouse or electrostatic
precipitator; revised opacity standard; improved compliance monitoring
methods)

- Improved capture of particulate emissions to increase the amount of
PM2.5 ducted to control devices, and to minimize the amount of PM2.5
emitted to the atmosphere, for example, through roof monitors

- New or upgraded emission controls for PM2.5 precursors at stationary
sources (e.g., SO2 controls such as wet or dry scrubbers, or reduced
sulfur content in fuel; desulfurization of coke oven gas at coke ovens;
improved sulfur recovery at refineries; increasing the recovery
efficiency at sulfuric acid plants)

- Energy efficiency measures to reduce fuel consumption and associated
pollutant emissions (either from local sources or distant power
providers)

- Measures to reduce fugitive dust from industrial sites

Mobile source measures

- Onroad diesel engine retrofits for school buses,11 trucks and transit
buses using EPA-verified technologies

- Nonroad diesel engine retrofit, rebuild or replacement, with catalyzed
particle filter12

- Diesel idling programs for trucks, locomotive, and other mobile
sources13

- Transportation control measures (including those listed in section
108(f) of the CAA as well as other TCMs), as well as other
transportation demand management and transportation systems management
strategies14

- Programs to reduce emissions or accelerate retirement of high emitting
vehicles, boats, and lawn and garden equipment

- Emissions testing and repair/maintenance programs for onroad vehicles

- Emissions testing and repair/maintenance programs for nonroad
heavy-duty vehicles and equipment15

- Programs to expand use of clean burning fuels

- Prohibitions on the sale and use of diesel fuel that exceeds a high
sulfur content

- Low emissions specifications for equipment or fuel used for large
construction contracts, industrial facilities, ship yards, airports, and
public or private vehicle fleets

- Opacity or other emissions standards for “gross-emitting” diesel
equipment or vessels

- Reduce dust from paved and unpaved roads

Area source measures					

- New open burning regulations and/or measures to improve program
effectiveness such as programs to reduce or eliminate burning of land
clearing vegetation

- Smoke management programs to minimize emissions from forest and
agricultural burning activities

- Programs to reduce emissions from woodstoves and fireplaces including
outreach programs, curtailments during days with expected high ambient
levels of PM2.5, and programs to encourage replacement of woodstoves
when houses are sold

- Controls on emissions from charbroiling or other commercial cooking
operations

- Reduced solvent usage or solvent substitution (particularly for
organic compounds with 7 carbon atoms or more, such as toluene, xylene,
and trimethyl benzene)

- Reduce dust from construction activities and vacant disturbed areas

Category- Specific Guidelines on innovative approaches.  The EPA has
issued a number of category specific guidelines on approaches to taking
into account innovative approaches to emissions reductions for purposes
of SIPs.   Categories currently covered by these guidelines include: (1)
Electric-sector Energy Efficiency and Renewable Energy Measures (2) Long
Duration Switch Yard Locomotive Idling (3) Long Duration Truck Idling
(4) Clean Diesel Combustion Technology (5) Commuter Choice Programs.  
See http://www.epa.gov/ttn/airinnovations/measure_specific.html

c.  Comments and Responses

Comment:  Some commenters recommended that EPA provide new CTGs or other
control technology review documents for purposes of assisting States to
address PM2.5 and its precursors, because the information in some
current documents is out-dated.

Response: The EPA recognizes that issuance of new or updated CTGs
specifically tailored for PM2.5 would be useful.  Unfortunately,
limitations on time and resources preclude EPA from developing such CTGs
in advance of the SIP submission date.  The EPA cannot delay the
statutorily specified outer date for SIP submission.  However, EPA
believes that there are already many sources of information and guidance
on key source categories.  To the extent that States need to examine
potential control measures for sources never addressed before in any
area or other context for a previous NAAQS, EPA anticipates that it will
work closely with States during the process of plan development and
approval to ensure an appropriate approach.   

Comment:  A number of commenters expressed concerns with references to
the STAPPA and ALAPCO Menu of Options document.  Some commenters
believed that this document must be subject to formal review and comment
to ensure appropriate stakeholder input.     

Response:  The language in the final preamble has been changed to refer
to a website EPA maintains that provides access to a variety of
information sources regarding control technologies that may be useful to
States to consider in developing their PM2.5 SIPs.  These links include
evaluations developed by government and nongovernment organizations. 
One such source with potentially useful information is the STAPPA and
ALAPCO Menu of Options.  However, EPA is not specifically recommending
that States adhere to this list of available sources of information
during their evaluation nor is EPA endorsing any of the specific
evaluations as being appropriate in any specific situation.  Rather, we
think documents such as the Menu of Options provide potentially useful
ideas.  Specifically, States would need to assess which items on the
menu are applicable in their areas, and will have to assess the costs of
applying controls locally.   Accordingly, there would be ample
opportunity for public review of the State’s analysis of the local
cost and air quality impacts of any measure listed in the document which
is included in a State’s SIP.  The EPA is not requiring that States
adhere to the list of measures in the Menu of Options.  The EPA does not
in any way mean to imply that the measures in the Menu of Options are
presumed to be RACM, merely that they are potential controls for areas
to consider.  The Menu of Options has no regulatory significance and
thus need not be issued through notice-and-comment rulemaking.  The EPA
notes, however, that the Menu of Options does provide a comprehensive
list of potential sources and measures that can help inform States in
the development of their plans.  Similarly, our own list of potential
measures is not intended to be a categorical list of measures which
States must adopt, rather it is intended to provide guidance about the
types of sources and measures that States can consider in constructing
their nonattainment area plans.  The EPA emphasizes that whether a
source category or potential measure is or is not on this list is simply
not conclusive as to whether a given measure is appropriate to consider
in the RACT and RACM analysis.  That can be determined only through the
State’s development of the nonattainment plan, and EPA’s evaluation
of such plan.

Comment:  A commenter representing the paper industry interpreted the
proposed rule as requiring electrostatic precipitator and tighter
sulfur-in-fuel requirements for the forest products industry.  The
commenter believed that  EPA was creating limits for such without
adequate rulemaking process.

Response:  The EPA disagrees that the listing of control technologies in
the table in the rule creates a “rebuttable presumption.”   Rather,
the table identifies potential opportunities for emissions reductions
which should be reviewed in light of technical and economic feasibility,
and which a State should consider in a list of possible RACT and RACM
measures for purposes of attaining the standards as expeditiously as
practicable.  The EPA is currently conducting a sector-based approach to
the paper industry.  One of the goals of the sector initiative on pulp
and paper is to work with the industry to identify reductions in SO2 and
PM2.5 that will assist us in meeting the NAAQS, considering facility
locations, magnitude of emissions, emission stream characteristics, and
cost effectiveness of controls.  

Comment:  A number of commenters believed that EPA should develop not
only a list of measures to consider for RACM, but should develop a list
of mandatory measures that States should include, particularly for areas
with attainment dates more than 5 years after designation.

Response:  See discussion in section II.D.3 regarding rule requirements
for attainment date extensions and the issue of whether certain measures
should be mandatory in order for an area to receive an extension. 

Comment:   Some commenters believed that the list of possible measures
was deficient in not including sources of PM2.5 and PM2.5 precursors
from agricultural sources.  One commenter believed the list is
incomplete without identifying the contribution of ammonia emissions
associated with livestock, poultry, and crop fertilizers.   

Response:   As we indicated in the proposal, we included a list of
potential RACM measures in the preamble to the proposed rule, based upon
a review of information about the contribution of various sources to the
emissions inventories and a review of potential control measures for
such sources.  We did not identify emissions from agricultural sources
in this review.  Because ammonia is not presumed to be a PM2.5 precursor
unless identified for a specific area by the State or EPA, regulation of
ammonia emissions from agricultural sources may not be necessary. 

We also note that the agricultural industry presents unique challenges
to regulators given the nature of the emissions source. Moreover, we
currently lack good methods to quantify agricultural emissions, and we
do not fully understand their contribution to nonattainment problems. 
We have entered into an agreement with several animal producer sectors
to monitor animal feeding operations to develop better tools to assess
emissions from this industry.  Hopefully, these tools will enhance our
knowledge of agricultural emissions and their contribution to
nonattainment problems, but until emissions from these sources are
better understood, States should apply appropriate deference in
determining whether any specific measure is RACT/RACM for this industry.
 

	Comment:   Some commenters raised concerns about a statement in the
proposal that “[i]n addressing a nonattainment area having military
training, testing and operational activities occurring within it, the
State should not need to target these activities for emission
reductions.”  Some commenters interpreted this statement as an
exemption from any emission reduction requirements for military sources.
 

	Response:  The statement in the proposal was not intended as an
exemption for all military activities.  Emissions potentially
contributing to PM2.5 concentrations at military installations originate
from a variety of sources:  basic operational activities (such as power
generation, other fuel combustion, and transportation to and from
residences, offices, and schools); and from field training and testing
activities (such as personnel training, obscurants used in training,
operation of nonroad vehicles and equipment, and related prescribed
burning operations).  EPA believes that in evaluating emissions for a
specific nonattainment area having military activities occurring within
it, the State should consult with DOD for information on the nature of
these activities and their associated emissions. 

	With regard to military training activities specifically, such
activities are periodic in nature, and when they do occur, the principal
type of emissions generated by these activities is dust (e.g. inorganic
direct PM emissions) from field operations.  Other pollutants may be
emitted to a lesser degree from certain onroad and nonroad motor
vehicles.  While military training activities may contribute some degree
of primary PM2.5 emissions to certain nonattainment area inventories,
the fugitive dust generated from military training activities is
predominantly comprised of coarse PM rather than fine PM.  

	Based on data from the PM2.5 speciation monitoring network operated by
EPA and the States, the contribution of inorganic dust to total PM2.5
mass on an annual average basis is relatively low in most nonattainment
areas, on the order of 0.5 to 1.5 micrograms per cubic meter (generally
10% or less of total PM2.5 mass).  Dust from military training
activities would be a subset of these levels.  Depending on the
available information and specific circumstances for a particular area,
a State could find in its SIP development analyses that direct PM2.5
emissions from military training activities do not significantly
contribute to PM2.5 concentrations in the nonattainment area, and
therefore would not need to target military training activities for
emission reductions in its nonattainment plan.  

  

7.  How should States consider EGU reductions for CAIR in meeting
RACT/RACM requirements?  

a.  Background.

	 In section III.I.11 of the preamble to the proposed rule, we discussed
the nature of the SO2 and NOx RACT obligations of electric generating
unit (EGU) sources in states subject to the CAIR emission reduction
requirements.  The CAIR rulemaking was finalized in March 2005 and
published at 70 FR 25221 (May 12, 2005).  CAIR requires 28 states and
the District of Columbia to significantly reduce emissions of SO2 and/or
NOx.  The 26 jurisdictions in the CAIR PM2.5 region are required to
reduce annual emissions of SO2 and NOx, and the 26 jurisdictions in the
CAIR ozone region are required to reduce seasonal emissions of NOx.  
These jurisdictions also have the option of participating in
EPA-administered annual SO2, annual NOx, and seasonal NOx cap-and-trade
programs (the CAIR trading programs) to meet these emission reduction
requirements.  In addition, in March 2006, EPA promulgated a Federal
implementation plan (FIP) to implement CAIR in these jurisdictions until
they have EPA approved CAIR SIPs in place (71 FR 25328, April 28, 2006).
 The FIP adopts, as the control measure, the CAIR trading programs
slightly modified to allow for Federal instead of State implementation.
When fully implemented, CAIR will reduce SO2 emissions in these
jurisdictions by over 70 percent and NOx emissions by over 60 percent
from 2003 levels. This will result in $85 to $100 billion in health
benefits and nearly $2 billion in visibility benefits per year by 2015
and will substantially reduce premature mortality in the eastern United
States. The benefits will continue to grow each year with further
implementation.

	Sources subject to cap-and-trade programs such as the CAIR trading
programs generally have the option of installing emissions control
technology, adopting some other strategy to reduce emissions, or
purchasing emissions allowances and thereby effectively paying other
sources covered by the cap to reduce emissions. In the proposal, we
noted that a number of EGUs expected to be covered by the CAIR trading
programs are located in nonattainment areas. Based on emissions
projections for 2010 and 2015 using the Integrated Planning Model (IPM),
some of these EGUs are expected to comply with CAIR by purchasing
allowances under the trading program and some are expected to comply by
installing emission controls.  

	The proposal also described our past experience with the implementation
of the NOx SIP Call and our belief that many power companies will
develop their strategies for complying with CAIR based, in part, on
consultations with State and local air quality officials in order to
address local PM2.5 and ozone attainment planning needs.  The EPA
suggested that consultations on location of CAIR controls would be
timely during State development of the CAIR SIP, which is due in 2006,
prior to the April 2008 deadline for submitting PM2.5 nonattainment area
SIPs.

	The EPA proposed a determination that in states that fulfill their CAIR
SO2 emission reductions entirely through EGU emission reductions (i.e.
without reductions from non-EGU sources or allowing non-EGU sources to
opt-in to the CAIR SO2 trading program), participation in the CAIR SO2
trading program would satisfy the SO2 RACT requirement for the EGU
sources.  The EPA also proposed that in states that fulfill their CAIR
NOx emission reductions entirely through EGU emission reductions, CAIR
would satisfy NOx RACT for the EGU sources, provided that those sources
with existing selective catalytic reduction (SCR) emission control
technology installed on their boilers operate that technology on a
year-round basis beginning in 2009.  Note that direct PM2.5 emissions
are not addressed by the CAIR program, and EPA did not propose any
determination that compliance with CAIR would satisfy RACT for direct
PM2.5 emissions.  The proposal included a discussion of the rationale
for these proposed determinations for SO2 and NOx, and requested
comments on the issue.  

b.  Final rule.

As discussed in section II.F.2 on our overall policy for RACT and RACM,
we consider an area’s obligation to implement RACT to be part of the
area’s overall RACM obligation – to adopt those reasonably available
measures needed to reach PM2.5 attainment as expeditiously as
practicable.  The final rule also reflects this combined RACT/RACM
approach regarding EGU control obligations under CAIR and the extent to
which meeting CAIR also satisfies a source’s RACT and RACM
requirements for attainment.  

Specifically, the final rule includes a presumption that in States that
fulfill their CAIR SO2 emission reduction requirements entirely through
EGU emission reductions (i.e. without reductions from non-EGU sources or
allowing non-EGU sources to opt in to the CAIR SO2 trading program),
compliance by EGU sources with an EPA-approved CAIR SIP or a CAIR FIP
would satisfy their SO2 RACT/RACM requirements for attaining the fine
particle NAAQS.  This section also includes a presumption that in States
that are subject to CAIR annual NOx emission reduction requirements and
fulfill these requirements entirely through EGU emission reductions
(i.e. without reductions from non-EGU sources or allowing non-EGU
sources to opt in to the CAIR annual NOx trading program), compliance by
EGU sources with an EPA-approved CAIR SIP or a CAIR FIP would satisfy
the NOx RACT/RACM requirement for the PM2.5 NAAQS, provided that the
sources with existing selective catalytic reduction (SCR) emission
control technology installed on their boilers operate that technology on
a year-round basis beginning in 2009.  This final position is based on a
number of factors identified in the proposal and discussed below.  

Many PM2.5 nonattainment areas are projected to achieve significant SO2
and NOx reductions under the CAIR program.  We do not believe that
requiring source-specific RACT/RACM controls on specified EGUs in
nonattainment areas would reduce total SO2 and NOx emissions from
sources covered by CAIR below the regionwide levels that will be
achieved under CAIR alone.  Nor do we believe that “beyond CAIR” EGU
controls for SO2 and NOx are “reasonably available” control measures
for most areas within the CAIR Region.  Accordingly, most States need
not evaluate additional control measures on EGUs to satisfy RACT/RACM
requirements as explained above.

As discussed previously, we are not requiring that States impose RACT on
any specific size or type of source.  Instead, States must conduct a
RACT/RACM analysis considering measures that are “reasonably
available” to meet the overarching requirement to attain the standards
as expeditiously as practicable.  Thus, today’s final rule imposes no
specific requirement on States to impose RACT/RACM on EGUs.  

Nonetheless, in evaluating RACT/RACM for EGUs, EPA believes it is
appropriate for States (states that achieve all reductions from EGUs) to
consider the special attributes of that group of facilities including
the unique interrelated nature of the power supply network, and their
participation in the CAIR program.  For EGUs in the CAIR region, based
upon the presumption explained here, States may define RACT/RACM as the
CAIR level of control on the collective group of sources in the region
rather than impose a specific level of control on an individual source. 
This approach is similar to the Agency’s past “bubble” policy, as
discussed in section (c) addressing comments on the proposal.   

As discussed more fully in the CAIR final rulemaking notice, EPA has set
the 2009 and 2010 CAIR caps for SO2 and NOx at a level that will require
EGUs to install emission controls on the maximum total capacity on which
it is feasible to install emission controls by those dates.  The EPA
concluded that the CAIR compliance dates represent an aggressive
schedule that reflects the limitations of the labor pool, and
equipment/vendor availability, and need for electrical generation
reliability for installation of emission controls. 

	Although the actual SO2 cap does not become effective until 2010, we
designed banking provisions in CAIR so that covered EGUs will begin to
reduce their SO2 emissions almost immediately after CAIR is finalized,
and will continue steadily to reduce their emissions in anticipation of
the 2010 cap and the more stringent cap that becomes effective in 2015. 
The 2015 SO2 and NOx caps are specifically designed to eliminate all SO2
and NOx emissions from EGUs that are highly cost effective to control
(the first caps represent an interim step toward that end).  

Moreover, we predicted that the majority of large coal-fired utilities
will install advanced control technologies under CAIR because the larger
and higher emitting source offer an opportunity to obtain more
cost-effective emissions reductions.  We expect that the
largest-emitting sources will be the first to install SO2 and NOx
control technology and that such control technology will gradually be
installed on progressively smaller-emitting sources until the ultimate
cap is reached.  As a result, few, if any coal-fired units with greater
than 600 MW of operating capacity should operate in PM2.5 nonattainment
areas without advanced control after full implementation of CAIR. Of the
remaining units operating without advanced pollution controls, a great
many of these units will have operating capacities below 300 MW.  We
predict that these units “will be utilized less often,” and
“typically have baghouses and electrostatic precipitators for
particulate control, have combustion controls for NOx control, and burn
low-sulfur coal.”  See “Contributions of CAIR/CAMR/CAVR to NAAQS
Attainment: Focus on Control Technologies and Emission Reductions in the
Electric Power Sector,” Office of Air and Radiation, U.S.
Environmental Protection Agency, April 18, 2006 (available at
http://www.epa.gov/airmarkets/cair/analyses/naaqsattainment.pdf).  In
light of these expected results, we generally believe that the cost to
install additional controls on these smaller units would be
unreasonable. 

We are also concerned that if States require specific EGUs to install
advanced pollution control measures, it could interfere with the
market-based incentives inherent in the cap and trade program.  This
could increase the cost of compliance and shift the location of the
units that would otherwise opt to install advanced emissions controls. 
Such a result may be counterproductive to that State’s attainment
efforts, as the State may forego a larger quantity of more beneficial
reductions in transported pollutants, in exchange for a smaller quantity
and less beneficial reduction in local emissions.  Moreover, it may
reduce the benefits expected in other nonattainment areas as well. 
Accordingly, even if a State found the cost to control an individual
unit acceptable on a cost per ton basis, the potential overall
disbenefit of control may nonetheless make imposition of the control not
“reasonably available.”  

	EPA finds that the control installations projected to result from CAIR
NOx and SO2 caps in 2009 and 2010 are as much as feasible from EGUS
across the CAIR Region by those dates.  In fact, if states chose to
require smaller-emitting sources in nonattainment areas to meet
source-specific RACT requirements by 2009, they would likely use labor
and other resources that would otherwise be used for emission controls
on larger sources.  Because of economies of scale, more boiler-makers
may be required per megawatt of power generation for smaller units than
larger units.  In this case, the imposition of source-specific RACT/RACM
on smaller emitting sources by 2009 could actually reduce the amount of
banking that would otherwise occur and result in higher SO2 emissions in
2009 as compared to the level that would result from implementation of
CAIR alone.

In any event, the imposition of source-specific control requirements on
a limited number of sources also covered by a cap-and-trade program
would not reduce the total regionwide emissions from sources subject to
the program.  Under a cap-and-trade program such as CAIR, a given number
of allowances are issued in order to achieve a given emission level. 
Source-specific control requirements within the CAIR program may affect
the temporal distribution of emissions (by reducing banking and thus
delaying early reductions) or the spatial distribution of emissions (by
moving them around from one place to another), but they would not affect
total regional emissions under the program.  If source-specific
requirements were targeted at the units that could be controlled most
cost-effectively, then the imposition of source-specific controls would
likely achieve the same result as the cap-and-trade program.  If not,
however, the imposition of source-specific requirements would make any
given level of emission reduction more costly than it would be under the
cap-and-trade program alone.  Thus, the imposition of source-specific
RACT on EGUs covered by CAIR would not reduce total regionwide
emissions, but would likely achieve the same total emission reductions
in a more costly way.  

  Given the considerations described above, we think that in many areas
additional controls on EGUs generally would not be “reasonably
available.”  Notwithstanding these conclusions, we recognize that
States are in the best position to determine how best to achieve
attainment with the PM2.5 NAAQS in light of local needs and conditions. 
As we acknowledged in our proposed rule, power plant operators typically
have ongoing relationships with the state and local officials involved
in air quality planning.  We expect that power plants will continue to
collaborate with State officials to determine how best to address
multiple air quality goals, and which plant locations to control under
CAIR, considering local PM2.5 and ozone attainment needs.

EPA expects States and local air agencies to identify reasonably
available control measures that are necessary and reasonable to attain
the standards as expeditiously as practicable; and that after consulting
with power companies, the State may conclude that establishing
additional “beyond CAIR” emission control requirements on specific
sources in nonattainment areas is warranted to provide for attainment as
expeditiously as practicable.  Nevertheless, in preparing the overall
attainment demonstration, States should be aware of the expected
benefits of the market-based incentives of the CAIR program, the cost
effectiveness of control, feasibility of implementation, and any
disbenefits that would result from requiring “beyond CAIR” controls
on any specific EGU before concluding that addition controls on EGUs are
“reasonably available” and necessary to satisfy RACT/RACM
requirements.  

	Year-round NOx controls.  In the CAIR final rulemaking notice, EPA
found that the operation of existing SCRs on a year-round basis, instead
of operating them only during the ozone season, could achieve NOx
reductions at low cost relative to other available NOx controls.  The
EPA projected that power generators would employ this control measure to
comply with CAIR SIPs.  Based on this control opportunity, EPA estimated
the average cost of non-ozone-season NOx control at $500/ton.  These
considerations support a finding that RACT should include year-round
operation of existing SCRs that are located in PM2.5 nonattainment
areas.  Because all PM2.5 nonattainment areas violate the annual form of
the PM2.5 standard and public health can be affected by high PM2.5
levels in the winter as well as the summer, we believe that year-round
operation of existing SCR that are located in nonattainment areas where
NOx is a nonattainment plan precursor will provide additional health
benefits for relatively low dollar cost per ton of pollutant reduced.

In the proposal notice, EPA proposed to define “existing” SCRs as
those units that were in place by the date of the proposed rule
(November 1, 2005).  We selected this date rather than the final date to
avoid creating an incentive to delay installation of new SCR.  Today, we
finalize our proposed approach with one clarification.  To avoid
confusion over the proper interpretation of the phrase “in place,”
we are clarifying that an existing SCR is one which is fully installed
and capable of operation by November 1, 2005.

We also proposed that these existing SCR begin year-round operations no
later than January 1, 2009 to qualify as RACT/RACM under our presumptive
approach. We noted that year round operation of existing SCR involves
little to no alteration of existing equipment, and that EGUs could
conduct any required work during normal outages.  Today, after taking
these factors into account, we finalize our proposed rule.  The
year-round operation requirement, however, will not be federally
enforceable to individual EGUs until EPA approves a State’s  SIP
including the requirement.

c.  Comments and responses

	Comment:  Some commenters supported the proposed determination
described in section (a) that in States that fulfill their CAIR SO2
emission reduction requirements entirely through EGU emission reductions
(i.e. without reductions from non-EGU sources or allowing non-EGU
sources to opt in to the CAIR SO2 trading program), compliance by EGU
sources with an EPA-approved CAIR SIP or a CAIR FIP would satisfy the
SO2 RACT requirement for the sources; and in States that are subject to
CAIR annual NOx emission reduction requirements and fulfill these
requirements entirely through EGU emission reductions (i.e. without
reductions from non-EGU sources or allowing non-EGU sources to opt in to
the CAIR annual NOx trading program), compliance by EGU sources with an
EPA-approved CAIR SIP or a CAIR FIP would satisfy the NOx RACT
requirement for the sources, provided that the sources with existing
selective catalytic reduction (SCR) emission control technology
installed on their boilers operate that technology on a year-round basis
beginning in 2009.  One commenter supported EPA’s approach so long as
states may pursue additional reductions from EGUs if needed for
attainment as expeditiously as practicable.  A number of other
commenters opposed the proposed determination regarding RACT for EGUs
based on a number of issues.

	Response:  Based on the rationale described in the sections above, the
final rule includes a presumption that compliance with CAIR satisfies
SO2 and NOx RACT/RACM requirements for EGUs in many areas.  Nonetheless,
States can require “beyond CAIR” EGU controls if a State determines
that it is a necessary and reasonable means to attain the PM2.5
standards.  Comments opposing this approach are addressed in more detail
below.

	

	Comment:  A number of commenters objected to the proposed
determination, arguing that it would result in greater control
requirements and economic burden on non-EGU sources located in
nonattainment areas.  These commenters urged EPA to adopt a final rule
that provides for implementing the most cost-effective controls
necessary to attain the standard.  They assert that with the proposed
finding that compliance with CAIR satisfies RACT for EGUs, the proposed
rule would not provide for the most cost-effective approach to
attainment.  They argue EPA and States should develop cost-effectiveness
guidance that includes all stationary source control measures and they
should develop SIPs based on the most economic means to attain the
standard.  They make several arguments to support this position.  The
commenters asserted that if an EGU control is more cost-effective than a
non-EGU control, the EGU should be subject to “beyond-CAIR”
controls.  They also asserted that if EPA chooses to consider the CAIR
rule as satisfying SO2 and NOx RACT for EGUs, then other sources should
not be subjected to control costs greater than those found reasonable
under CAIR (i.e., $800/ton).  They believe it would be inequitable to
require smaller sources to pay a higher cost for emissions reductions
than larger sources, which are a more significant contributor to the
problem and which may be able to make more cost-effective emission
reductions.  One commenter also suggested that EPA should authorize a
presumption that emissions reductions required on electric utilities
under the CAIR will be equivalent to RACT only if a particular source in
a CAIR State has installed controls that achieve the average level of
control that EPA has projected will occur for the particular pollutant
under the CAIR requirements.

Response:  The EPA has determined that implementation of the CAIR
trading program represents highly cost-effective controls that will
achieve widespread regional SO2 and NOx emissions reductions from EGUs
and will provide significant air quality benefits for ozone and PM2.5
nonattainment areas. In developing attainment SIPs and identifying RACM,
States will need to consider additional cost-effective and reasonable
controls to reach attainment as expeditiously as practicable.  EPA does
not agree with the commenter’s argument that controls on non-EGUs
should be no more than the projected cost of EGU controls under CAIR. 
EPA expects that in order to achieve attainment as expeditiously as
practicable, some states may need to adopt control measures for some
sources which cost more per ton but which still are considered to be
reasonable and cost-effective.

In addition, States must consider the economic feasibility of
implementing a given control measure.  Because of facility-specific
factors, EPA believes it would be inappropriate to establish a threshold
of control effectiveness (e.g. dollars per ton) based on control of EGUs
and apply this threshold to all source categories.  The ability of a
source to cost-effectively reduce emissions is dependent on
case-specific factors, including the ability of the given source to
sustain the cost of control, and prevailing costs in the specific
geographical location.  A direct correlation between the size of an
emissions source and the economic feasibility of controls for that
source and location does not necessarily exist.

We also disagree with the commenter who suggests that RACT requirements
should only be satisfied if a source achieves an average level of
control that EPA projects to occur under CAIR.  EPA maintains that the
presumption that CAIR satisfies SO2 and NOx RACT/RACM for EGUs in most
areas is an appropriate policy.  As discussed further below, we  have
always recognized that States could determine RACT for a single source
or group of sources.   

Comment:  A number of commenters opposed the proposed determination that
CAIR would satisfy the SO2 and NOx RACT requirement for EGU’s.  The
commenters argued that this determination is unlawful, it does not
comply with section 172(c)(1) of the CAA which requires RACT (i.e.
controls that are technologically and economically feasible) “at a
minimum” for all existing sources in the nonattainment area, it would
allow very large stationary sources to escape cost-effective controls
entirely, and it is largely based on the legally-irrelevant contention
that CAIR will reduce emissions more cost-effectively than RACT.  They
claim that EPA has no authority to displace the Congressionally-mandated
RACT requirement, that CAIR was designed to address regional pollution
transport (not to be an attainment strategy), and that EPA should remove
these proposed provisions in the final rule.  Commenters claim that the
EPA(s proposed approach to allow EGU emissions to be addressed solely
through CAIR would undermine states( efforts to meet the Federal PM2.5
health standard, particularly when EGU sources are among the most
cost-effective to control.  Another commenter claimed that EPA’s
proposal allowing states that choose to fulfill their CAIR requirements
entirely through emission reductions from EGUs to also use CAIR to
satisfy their SO2 and NOx PM2.5 RACT requirements, thereby equating
these two requirements for the EGU sector, is flawed. This commenter
argued that allowing a cap-and-trade program, such as the CAIR, to
substitute for the RACT requirement undermines the effectiveness of the
controls by allowing facilities to use allowances to offset emissions,
rather than control them at the source.  The purchase of allowances,
they assert, does not satisfy RACT requirements.  

	Response:  The EPA disagrees with these comments.  Today’s final rule
does not displace the RACT requirement for any sources.  Instead, EPA is
exercising its authority to interpret the section 172 RACT and RACM
requirements for the purposes of implementing the 1997 PM2.5 standards. 
For the reasons described in section (b) above, we believe that states
can rely on EPA’s presumption that compliance with a CAIR SIP or FIP,
meeting certain requirements, will satisfy the RACT/RACM requirement for
certain EGU sources. EPA historically issued control technology
guidelines setting forth presumptive levels of emissions control that
satisfy the RACT requirement for a given industry.  Today’s final rule
is similar to this practice in establishing a presumption that SO2 and
NOx reductions under the CAIR program satisfy the RACT/RACM requirement
for EGUs in CAIR states.  In identifying reasonably available control
measures to ensure attainment as expeditiously as practicable, States
will need to take CAIR reductions into account as well as any additional
cost-effective reductions that are technologically and reasonably
available.  

We further find that the attempt by many commenters to characterize CAIR
as a strategy to address only regional pollution transport and not an
attainment strategy as overly simplistic.  EPA analyses for CAIR show
that there are significant air quality benefits projected for individual
nonattainment areas as a result of SO2 and NOx reductions across the
multistate CAIR region.  The Act does not prevent States from properly
crediting measures that achieve multiple objectives (e.g. regional
transport or local nonattainment).  Moreover, Section 110(a)(2)(D)
requires SIPs to contain adequate provisions to assure that sources in
the State do not contribute significantly to nonattainment in any other
State.  The CAIR rule is an integral element in meeting the States’
Section 110 attainment obligations.  Accordingly, it is reasonable to
incorporate this consideration in determining what measures qualify as
RACT/RACM.

Finally, EPA does not interpret the provisions of Section 172(c)(1)
related to the RACT requirement as precluding States’ use of a cap and
trade approach as a means of regulating existing sources and achieving
RACT/RACM reductions, especially in light of Congresses’ expressed
authorization to auction emission rights in Section 172(c)(6). 

EPA has long recognized that RACT need not apply to individual sources. 
As stated earlier, our early guidance on RACT requirements stated that
States could establish RACT for an “individual sources or a group of
sources.” (emphasis added)  See Memo. Strelow (Dec. 1976) and 44 FR
71779.  Importantly, Congress ratified the early interpretations of RACT
and RACM when it enacted the 1990 Amendments. See 42 U.S.C. Section 7515
(Clean Air Act section 193).  Our 1986 emissions trading policy also
recognized a number of advantages offered through application of a
“bubble” approach including faster compliance with RACT limits and
earlier reductions. Moreover, Courts have upheld EPA’s approval of
States’ use of “bubbling” multiple units to meet RACT
requirements.  See e.g. Natural Resources Defense Council v. EPA, 941
F.2d 1207 (finding that EPA need not adhere to a source specific RACT
determination to satisfy RACT requirements and acknowledging EPA’s
special knowledge and expertise in the area.) 

	Comment:  The EPA’s proposal to allow EGU emissions to be addressed
solely through CAIR undermines prospectively States( efforts to meet the
Federal PM2.5 health standard.  EGU sources are among the most
cost-effective to control.

	Response:  For the reasons described in section (b) above, EPA believes
that states can rely on EPA’s presumption that compliance with a CAIR
SIP or FIP, meeting certain requirements, satisifies the SO2 and NOx
RACT/RACM requirement for certain EGU sources.  Areas can require
“beyond CAIR” EGU controls if a State determines that it is a
necessary and reasonable means to attain as expeditiously as
practicable.  

	Comment:  CAIR fails to address the need for short-term reductions in
PM2.5 and precursor emissions on high pollution days.  While RACT
restricts emissions over a 1-hour to 24-hour period, CAIR only provides
for an annual or seasonal cap. Reliance on CAIR therefore fails to
recognize the importance of reducing short-term emissions, which was
recently highlighted by the EPA(s own proposal to tighten the 24-hour
PM2.5 health standard.  Local and short-term adverse air quality effects
of PM2.5, must he addressed in the final rule by requiring RACT for all
major facilities in addition to CAIR.  

	Response:  The CAIR program is oriented toward reducing SO2 and NOx
emissions in order to reduce air quality concentrations on an annual and
seasonal basis.  Because all PM2.5 nonattainment areas were designated
due to violations of the annual standard (and the two designated areas
in California also violated the 24-hour standard), the focus of this
implementation rule is attainment of the annual standard.  CAIR is
projected to provide significant air quality benefits in 2010 and 2015
for eastern PM2.5 nonattainment areas on both an annual basis and on a
98th percentile 24-hour basis.  Since the PM2.5 24-hour standard was
revised in September 2006 from 65 micrograms per cubic meter to 35
micrograms per cubic meter, EPA recommends that as States develop
implementation plans to address the annual standard, they consider
adopting strategies that also will provide benefits toward attaining the
revised 24-hour standard.  

	Comment:  The proposal is silent on the issue of whether EGUs are
subject to direct PM2.5 emissions RACT requirements.  It is critical
that RACT be required for all facilities with respect direct PM2.5
emissions, regardless of a facility(s participation in CAIR. 

	Response:  In the final rule and preamble, EPA has clarified that all
EGUs in nonattainment areas are subject to RACT/RACM for direct PM2.5
emissions.  The determination described above applies only to SO2 and
NOx RACT/RACM, not RACT/RACM for direct PM2.5 emissions from EGUs.  

	Comment:  The EPA fails to consider the geographical distributional
impacts of the emission reductions.  Equating CAIR with RACT fails to
take into account the substantial contribution that emissions from EGUs
within a nonattainment area may make toward that area(s PM2.5
nonattainment problem.  The EPA does not attempt to explain how such a
generalized determination satisfies RACT for PM2.5.

	Response:  The establishment of recommended levels for RACT/RACM is an
area Congress delegated to the specific expertise of the Agency.  Based
on our analysis, we conclude that the CAIR emissions caps presumptively
represent the level of emissions control achievable through application
of “reasonably available” control technologies.  Nonetheless, in
developing attainment plans, each State will evaluate the impact of
stationary sources located within the nonattainment area in developing
its attainment strategies for the local area.

	Comment:  A few commenters stated that EPA should explain how this
proposal would be implemented for states that request an extension of an
attainment date because attaining in 5 years or less is impracticable;
i.e., whether EPA would still hold to its interpretation that CAIR
equals RACT for EGUs and not require additional reductions from EGUs
even if an area cannot attain in 5 years and controls on EGUs could lead
it to attain more expeditiously.  These commenters argue that, in
considering if additional RACT is needed in states that obtain
extensions of the attainment deadline after 2010, EPA cannot ignore
potential RACT for electric generating units any more than they would be
allowed legally to avoid consideration of any other RACT candidates. 
One commenter is particularly concerned that States would not include
EGUs in their RACT determinations and instead require smaller industrial
boilers or process heaters to control emissions.

	Response:  EPA’s determination regarding CAIR and RACT is not limited
to areas attaining within five years.  The Agency’s rationale is
presented in the “final rule” section above.  We disagree that the
CAIR-RACT presumptions necessarily shift emission control burdens from
EGUs to smaller industry boilers and process heaters because, in
implementing the RACM requirement, the State may include an evaluation
of control options on those sources as part of their RACT/RACM analyses.
 As stated above, EPA concluded that the CAIR compliance dates represent
an aggressive schedule that reflects the limitations of the labor pool,
and equipment/vendor availability, and need for electrical generation
reliability for installation of emission controls.  Accordingly,
additional controls on EGUs may not be a reasonably available control
measure that can be effectively implemented in a manner that advances an
area’s attainment date.  Nonetheless, States should consider the
economic and technologic feasibility of implementing controls on all
types of sources (stationary, mobile, area)located in the nonattainment
area, given local conditions, in developing attainment plans.  

	Comment:  The EPA designated many partial counties nonattainment for
PM2.5 solely because the areas contained EGU emission sources thought to
cause or contribute to violations of the NAAQS.  In implementing
attainment plans, it makes sense to consider further control of these
sources, and because they are located in nonattainment areas, the
ability to do so is provided for and legal under the CAA.

	Response:  The EPA designated PM2.5 nonattainment counties because they
either had a violating monitor or they contributed to a nearby air
quality problem.  Importantly, EPA designated these areas without
considering the air quality benefits expected in the future from CAIR. 
Accordingly, the fact that an EGU is located in a partial county and we
included the partial county in the nonattainment area because we believe
that the EGU was causing or contributing to the nonattainment
violations, does not equate with a finding that more than CAIR is
required to remedy the nonattainment problem.  Nonetheless,   EPA
believes that States should evaluate the impact of stationary sources in
all designated counties, including those partial counties noted by the
commenter, in its assessment of reasonably available control strategies
to ensure attainment as kly as practicable.  

	Comment:  The EPA should adopt the Ozone Transport Commission(s (OTC(s)
approach to cap-and-trade programs. When the OTC developed its NOx
Budget Program (which was the basis for EPA(s NOx SIP call and
subsequently CAIR), it assumed that RACT was applied first. Thus the
cap-and-trade program operated in an environment that assumed RACT was
in force, not in lieu of RACT.

	Response:  Under the ozone national ambient air quality standards, NOx
and VOC RACT have been implemented progressively for the past 30 years
or more, prior to development of the NOx SIP call regional control
program.  In contrast, the PM2.5 implementation program is the first
instance in which we have required RACT/RACM specifically for fine
particle pollution.  For this reason, the CAIR program is not operating
with SO2 and NOx RACT limits already in place for attainment of the
PM2.5 standards.  Nonetheless, as discussed above, EPA believes that
implementation of the CAIR requirements will provide for substantial
progress in attaining the PM2.5 standards and that States may presume
that RACT/RACM requirements are equal to the CAIR level of control.

	Comment: A few commenters stated that EPA should clarify and modify the
part of its proposal that explains why a State cannot rely on EPA’s
determination that CAIR can satisify the NOx RACT requirement for PM2.5
if the State (elect[s] to allow non-EGU sources to voluntarily enter the
EPA-administered CAIR trading program through an opt-in provision in the
CAIR model rule.( (70 FR 66025 col. 3). These commenters believe that
this part of the proposal might be construed to preclude states subject
to both the NOx SIP Call and included in the CAIR region for ozone from
relying on the NOx RACT determination for PM2.5 if the states choose (to
bring their non-CAIR [including non-EGU] NOx SIP Call trading sources
into the CAIR ozone season NOx cap and trade program.  (70 FR 49708,
49728 col. 3)  (August 24, 2005). The commenters assert that EPA gave
states the option of bringing non-EGU NOx SIP Call sources into the CAIR
seasonal NOx trading program to ensure that non-CAIR sources, including
non- EGUs, that are subject to the NOx SIP Call rule would not be
(stranded,( starting in 2009, by being left in an ozone season NOx
control program with no EGU trading partners.  The commenters argued
that “EGUs should not be penalized, in the form of denial of CAIR =
RACT treatment, as a result of states exercising their option to avoid
financial and compliance difficulties for non-EGUs that otherwise would
be left without allowance trading partners in the EGU sector after the
NOx SIP Call trading program ends in 2008.”  These commenters point to
EPA’s determination in the final Phase 2 ozone implementation rule,
that participation in the CAIR trading programs can satisify NOx RACT
for ozone even if a State brings non-EGUs in the NOx SIP Call trading
program into the trading program after 2008, see 70 FR 71657 col. 2,
provided the State retains an (EGU [emission] budget under CAIR that is
at least as restrictive as the EGU budget that was set in the State(s
NOx SIP call SIP,( id. At 71658 col. 1. These commenters argue that EPA
should make a similar determination here regarding NOx RACT for purposes
of PM2.5 NAAQS implementation.

Response:  All states with EPA approved CAIR SIPs or subject to a CAIR
FIP implementing the annual NOx emission reduction requirements, and
obtaining those reductions solely from EGUs may rely on EPA’s
determination that CAIR presumptively satisifies NOx RACT/RACM for PM2.5
for these sources.  This determination is unaffected by whether or not a
State permits NOx SIP Call non-EGUs to participate in the CAIR ozone
season trading program.  In the final rule, we have included the
presumption that NOx RACT/RACM for PM2.5 is satisfied for EGUs complying
with a CAIR SIP or CAIR FIP implementing the annual CAIR NOx emission
reduction requirements (provided the State implementation of the CAIR
NOx annual trading program includes EGUs only). 

	In the final ozone implementation rule, EPA addressed numerous issues
relating to the transition from the NOx SIP Call to the CAIR ozone
season trading program, including the impact of bringing NOx SIP Call
non-EGUs into the CAIR ozone season trading program.  Commenters
suggestion that these determinations are relevant to this PM2.5
implementation rule ignores the fact that both the NOx SIP Call and the
CAIR ozone season trading program are seasonal, not annual, trading
programs.  The NOx SIP Call EGU and non-EGU budgets are seasonal NOx
budgets and do not address annual NOx emissions.  As discussed above,
PM2.5 levels year-round contribute to an area’s annual average
concentration, and NOx emissions during non-summer months contribute to
nitrate concentrations, which are typically highest in cooler
temperatures.  For these reasons, EPA believes it would be inappropriate
to accept commenters’ suggestion.  

 Under the Tribal Air Rule (TAR), requirements for RACT and RACM may be
considered to be severable elements of implementation plan requirements
for Tribes. ADVANCE \d 12 

 See, 44 FR 53782, September 17, 1979, and 1976 memorandum from Roger
Strelow, Assistant Administrator for Air and Waste Management to
Regional Administrators, “Guidance for Determining Acceptability of
SIP Regulations in Non-attainment Areas” (Dec. 9, 1976)

 See e.g. Workshop on Requirements for Non-atttainment Area
Plans-Compilation of Presentations (OAQPS No. 1.2-103, revised edition
April 1978)

 Under the Tribal Air Rule (TAR), requirements for RACT and RACM may be
considered to be severable elements of implementation plan requirements
for Tribes. ADVANCE \d 12 

 In Sierra Club v. EPA, 294 F.3d 155 (D.C. Cir. 2002), the court stated
in upholding EPA’s statutory interpretation of RACM that the Act does
not compel a state to consider a measure without regard to whether it
would expedite attainment.  

 In this notice, where we use the shorthand phrase “advance the
attainment date,” it means “advance the attainment date by one year
or more.”

 In the context of the PM10 NAAQS, EPA has concluded that “advancement
of the attainment date” should mean an advancement of at least one
calendar year.  See State Implementation Plans; General Preamble for the
Implementation of Title I of the CAA Amendments of 1990, 57 FR 12498
(April 16, 1992).  See also Sierra Club v. EPA, 294 F.3d 155 (D.C. Cir.
2002). ADVANCE \d 12 

 The EPA believes that it is not necessary to identify every possible
variation of every type of control measure, or all possible combinations
of technologies and measures that would apply to a given source or
activity if the State has properly characterized the potentially
available emissions reductions and their costs.   For example, EPA
believes that the State can conduct a thorough analysis of VMT reduction
measures without including every possible level or stringency of
implementation of certain possible measures or combinations of measures
for reducing VMT, so long as those measures would not affect the overall
assessment of VMT reduction capabilities and the associated costs.  

 For example, see past EPA guidance on PM2.5 control technologies:
Stationary Source Control Techniques Document for Fine Particulate
Matter (EPA-452/R-97-001), EPA Office of Air Quality Planning and
Standards, October 1998.

 Stationary Source Control Techniques Document for Fine Particulate
Matter (EPA-452/R-97-001), EPA Office of Air Quality Planning and
Standards, October 1998. See also:  Controlling SO2 Emissions: A Review
of Technologies (EPA/600/R-00/093), EPA Office of Research and
Development, November 2000.

 See EPA’s website for more information: 
http://www.epa.gov/ttn/emc/monitor.html.

 EPA Air Pollution Control Cost Manual - Sixth Edition (EPA
452/B-02-001), EPA Office of Air Quality Planning and Standards,
Research Triangle Park, NC, Jan 2002. ADVANCE \d 12 

  There are a number of sources of information on technologies for
reducing emissions of PM2.5 and its precursors.  Links are provided to a
number of national, state and local air quality agency sites from
EPA’s PM2.5 website: http://www.epa.gov/pm/measures.html 

12See Clean School Bus USA program at
http://www.epa.gov/cleanschoolbus/.  See also:  AWhat You Should Know
About Diesel Exhaust and School Bus Idling", (June 2003,
EPA420-F-03-021) at
http://www.epa.gov/otaq/retrofit/documents/f03021.pdf. ADVANCE \d 12 

12 See EPA’s voluntary diesel retrofit program web site at
http://www.epa.gov/otaq/retrofit/overfleetowner.htm. ADVANCE \d 12 

13 See EPA’s voluntary diesel retrofit program web site at
http://www.epa.gov/otaq/retrofit/idling.htm. ADVANCE \d 12 

14 See EPA’s website on transportation control measures at
http://www.epa.gov/otaq/transp/traqtcms.htm. ADVANCE \d 12 

15 See EPA’s web site on nonroad engines, equipment, and vehicles at
http://www.epa.gov/otaq/nonroad.htm. ADVANCE \d 12 

 See the regulatory impact analysis chapter on air quality for the 2006
PM NAAQS review at
http://www.epa.gov/ttn/ecas/regdata/RIAs/Chapter%204--Air%20Quality.pdf.

 EPA’s CAIR-RACT presumption also would not apply if a State required
sources other than EGUs to achieve a portion of the reductions required
by CAIR (e.g., the State’s CAIR SIP achieved some reductions from EGUs
but took credit for non-EGU reductions achieved under new, more
stringent requirements implemented to meet NOx SIP call caps).  Under
the CAIR rule such a State would not be eligible to participate in the
EPA-administered CAIR trading system.

 PAGE   

 PAGE   106 

