1
7/
16/
05
RACT
comparison
with
7/
1/
05
version
RACT/
RACM
EXCERPT
8­
HR
O3
NAAQS
IMPLEMENTATION
RULE
PHASE
2
G.
What
requirements
should
apply
for
RACM
and
RACT
for
8­
hour
ozone
nonattainment
areas?
[
Section
VI.
K.
of
June
2,
2003
proposed
rule
(
68
FR
32837);
§
51.912
in
draft
and
final
regulatory
text.]
The
first
subsection
of
this
section
covers
RACT
and
the
second
subsection
covers
RACM.
1.
Reasonably
Available
Control
Technology
(
RACT)
a.
Background
As
described
in
more
detail
in
the
June
2
proposal,
subpart
1
of
part
D
includes
a
requirement
that
an
attainment
plan
provide
for
the
implementation
of
all
RACM
as
expeditiously
as
practicable,
including
such
reductions
that
may
be
obtained
through
RACT.
Under
subpart
2,
marginal
areas
are
required
to
correct
pre­
1990
RACT
requirements
and
new
RACT
requirements
are
specified
for
moderate
and
above
ozone
nonattainment
areas.
Additionally,
States
must
adopt
RACT
for
all
areas
in
an
ozone
transport
region
(
OTR).
The
RACT
requirement
applies
to
both
ozone
precursors
 
NOx
and
VOC.
Since
1990,
we
have
issued
guidance
documents
on
the
RACT
requirements
in
subpart
2.
Prior
to
enactment
of
the
CAA
Amendments
of
1990,
EPA
also
issued
detailed
guidance
documents
on
RACT
for
ozone
nonattainment
area
SIPs.
In
the
period
prior
to
enactment
of
the
1990
CAA
Amendments,
only
the
general
requirements
for
RACM
and
RACT
existed.
Also,
prior
to
the
1990
Amendments,
EPA
had
issued
control
technique
guideline
(
CTG)
documents
to
provide
presumptive
norms
for
RACT
for
VOC
controls
for
States
to
follow
in
adopting
RACT
for
ozone
nonattainment
areas.
In
1990,
Congress
specified
more
detailed
RACT
provisions
and
explicitly
required
RACT
for
NOx
as
well
as
for
VOC.
Also
in
1990,
section
183(
c)
required
EPA
to
issue
technical
documents
which
identify
alternative
control
techniques
(
ACT
documents)
for
large
stationary
sources
of
NOx
and
VOC.
The
EPA
completed
12
ACT
documents
for
VOC
source
categories.
Over
a
5­
year
period,
1991­
94,
EPA
issued
nine
ACT
guideline
documents
for
large
stationary
sources
of
NOx.
In
2000,
updates
to
the
NOx
ACT
documents
were
completed
for
stationary
internal
combustion
engines
and
cement
kilns.
Although
the
ACT
documents
do
not
set
a
presumptive
RACT
level
as
do
CTGs,
they
are
intended
to
help
States
in
making
RACT
determinations.
Section
183(
c)
of
the
CAA
also
requires
EPA
to
"
revise
and
update
such
documents
[
i.
e.,
CTGs
and
ACTs]
as
the
Administrator
determines
necessary."
The
EPA
intends
in
the
future
to
develop
updated
technology
guidance,
especially
with
respect
to
source
categories
emitting
multiple
pollutants
in
large
amounts.
In
addition,
EPA
will
be
considering
related
recommendations
from
2
the
Clean
Air
Act
Advisory
Committee
that
were
sent
to
EPA
in
January
2005
in
response
to
the
recent
National
Research
Council
report
on
Air
Quality
Management
in
the
United
States
(
January
2004).
As
new
or
updated
EPA
RACT
guidance
becomes
available
States
should
consider
the
new
information
in
their
RACT
determinations.
States
should
consider
the
updated
guidance
in
any
RACT
determinations
or
certifications
that
have
not
been
issued
as
of
the
time
such
an
update
becomes
available.
In
addition
to
the
ACTs
and
CTGs,
information
developed
as
part
of
the
process
of
determining
BACT,
BART,
LAER,
and
MACT,
can
be
very
useful
in
supplementing
or,
in
some
cases,
substituting
for
information
to
support
a
RACT
determination.
In
the
June
2,
2003
proposal,
we
proposed
several
requirements
with
respect
to
RACT.
For
subpart
1
areas,
we
proposed
several
options.
We
proposed
in
one
option
to
interpret
the
CAA
in
a
manner
similar
to
that
under
subpart
2
by
requiring
areas
covered
under
subpart
1
to
face
different
RACT
requirements
based
on
the
magnitude
of
the
ozone
problem
in
the
area
(
i.
e.,
the
area's
design
value).
In
another
option,
we
proposed
that
RACT
would
be
met
if
the
area
were
able
to
demonstrate
attainment
of
the
standard
as
expeditiously
as
practicable
with
emission
control
measures
in
the
SIP.
We
also
proposed
as
an
early
attainment
incentive
that
RACT
would
be
met
in
an
area
which
demonstrates
attainment
within
3
years
and
submits
the
demonstration
within
1
year.
For
subpart
2
areas,
we
proposed
to
apply
RACT
as
specified
in
subpart
2.
We
proposed
the
RACT
submittal
dates
for
subpart
1
areas
would
be
either
the
same
as
the
submittal
date
for
the
attainment
plan
(
within
3
years)
or
the
same
as
the
deadline
for
subpart
2
areas
(
within
2
years).
In
addition,
we
proposed
the
date
for
affected
sources
to
implement
RACT
in
subpart
2
areas
would
be
30
months
after
the
required
submittal
date.
We
also
proposed
that
States
may
use
current
EPA
guidance
in
making
RACT
determinations;
consequently,
in
some
cases,
sources
previously
evaluated
under
the
1­
hour
ozone
RACT
requirement
and
sources
subject
to
the
NOx
SIP
Call
cap­
and­
trade
program
could
be
determined
to
meet
the
8­
hour
ozone
RACT
requirement.
b.
Summary
of
final
rule
For
subpart
1
areas
that
do
not
request
an
attainment
date
extension
(
beyond
5
years
after
designation),
RACT
will
be
met
with
control
requirements
sufficient
to
demonstrate
that
the
NAAQS
is
attained
as
expeditiously
as
practicable.
The
RACT
submittal
date
for
these
areas
is
the
same
as
the
submittal
date
for
the
attainment
plan.
This
submission
date
is
no
later
than
3
years
after
designation.
For
subpart
1
areas
that
request
an
attainment
date
extension
(
beyond
5
years
after
designation),
subpart
2
moderate
and
above
areas,
and
areas
within
an
OTR,
RACT
is
required
with
3
the
RACT
submittal
and
is
due
27
months
after
designation,
except
a
subpart
1
area
shall
submit
the
RACT
SIP
with
its
attainment
date
extension
request.
States
must
require
sources
to
implement
RACT
no
later
than
the
first
ozone
season
or
portion
thereof
which
occurs
30
months
after
the
required
submittal
date.
Where
RACT
applies
(
except
certain
subpart
1
areas,
as
described
above,
and
except
certain
sources
subject
to
the
NOx
SIP
Call
or
CAIR,
as
described
below),
State
SIPs
implementing
the
8­
hour
standard
must
assure
that
RACT
is
met,
either
through
a
certification
that
previously
required
RACT
controls
represent
RACT
for
8­
hour
implementation
purposes
or
through
a
new
RACT
determination.
States
may
use
existing
EPA
guidance
in
making
RACT
determinations.
Where
a
State
has
adopted
and
EPA
has
approved
a
control
measure
as
RACT
for
a
specific
major
stationary
source
or
source
category
for
the
1­
hour
ozone
NAAQS,
and
absent
data
indicating
that
the
previous
RACT
determination
is
no
longer
appropriate
in
lieu
of
a
RACT
rule,
the
State
may
submit
a
certification
that
the
source
is
subject
to
RACT.
Such
certification
shall
be
accompanied
by
appropriate
supporting
information,
such
as
consideration
of
information
received
from
public
commenters.
For
purposes
of
meeting
the
NOx
RACT
requirement
in
areas
located
in
a
State
subject
to
the
NOx
SIP
Call
or
CAIR
for
NOx,
the
State
may
conduct
case­
by­
case
RACT
determinations
or
consider
the
RACT
SIP
requirements
met
for
sources
subject
to
the
State's
emission
cap­
and­
trade
program
where
the
cap­
and­
trade
program
has
been
approved
by
EPA
as
part
of
a
SIP
meeting
the
NOx
SIP
Call
or
CAIR
NOx
requirements.
Additionally,
RACT
is
considered
met
for
cement
kilns
and
stationary
internal
combustion
engines
that
are
subject
to
a
SIP­
approved
obligation
to
install
and
operate
controls
that
are
expected
to
achieve
at
least
a
30
percent
and
82
percent
reduction,
respectively,
from
uncontrolled
levels.
States
are
free
to
conduct
case­
by­
case
RACT
determinations
 
or
RACT
determinations
or
certifications
for
groups
of
sources
 
at
their
discretion.
A
State
may
meet
the
NOx
RACT
requirement
by
showing
that
the
weighted
average
emission
rate
from
sources
in
the
nonattainment
area
subject
to
RACT
 
including
sources
subject
to
the
NOx
SIP
Call
requirements
 
meet
RACT
requirements.
It
should
be
noted
that
for
purposes
of
meeting
the
RACT
obligations
under
section
182(
b)(
2)(
C)
of
the
CAA
for
major
stationary
sources
of
VOCs
and
under
section
182(
f)
of
the
CAA
for
major
stationary
sources
of
NOx,
the
definition
of
major
stationary
source
in
section
302
of
the
CAA,
as
modified
by
the
major
source
definition
in
either
section
182(
b),
(
c),
(
d)
or
(
e)
of
the
CAA
as
applicable
to
the
area's
classification,
applies.
Although
we
drafted
regulatory
language
for
several
aspects
of
the
RACT
program,
we
believe
it
is
sufficient
to
describe
4
EPA's
views
on
the
details
of
the
RACT
program
in
today's
preamble
and
in
other
guidance
(
e.
g.,
the
NOx
Supplement
to
the
General
Preamble,
11­
25­
92).
Thus,
some
detailed
portions
of
the
proposed
regulatory
text
regarding
RACT
were
not
retained
in
the
final
rule.
c.
Comments
and
responses
Comments:
For
subpart
2
ozone
nonattainment
areas,
several
States
expressed
agreement
with
the
proposed
approach
for
implementing
RACT
consistent
with
section
182
of
the
CAA.
Response:
The
EPA
agrees
with
these
comments.
Comments:
For
subpart
1
ozone
nonattainment
areas,
EPA
received
several
comments
for
and
against
the
options
proposed
for
addressing
RACT.
Several
State
and
industry
commenters
supported
EPA's
proposed
approach
that
RACT
would
be
met
if
the
area
is
able
to
demonstrate
attainment
of
the
standard
as
expeditiously
as
practicable
with
emission
control
measures
in
the
SIP.
The
reasons
provided
by
these
commenters
were
generally
as
follows:
States
should
be
able
to
use
their
discretion
in
determining
which
control
strategies
are
the
most
effective
in
addressing
a
particular
area's
air
quality
problem;
flexibility
is
needed
as
areas
differ
in
sensitivity
to
NOx
and
VOC
reductions;
EPA's
regional
modeling
shows
these
requirements
are
unnecessary
in
many
areas;
and
many
of
these
areas
violate
the
ozone
standard
primarily
or
entirely
due
to
transport.
The
EPA
also
received
comments,
primarily
from
several
States
and
environmental
groups,
opposing
the
approach
that
RACT
would
be
met
by
control
measures
that
are
part
of
a
SIP
demonstrating
attainment
of
the
standard
as
expeditiously
as
practicable.
These
commenters
made
the
following
points:
since
section
172(
c)(
1)
of
the
CAA
explicitly
mandates
RACT
"
at
a
minimum"
in
all
nonattainment
areas,
Congress
plainly
intended
to
require
RACT
as
a
floor
level
of
control
technology
in
addition
to
any
measures
needed
to
demonstrate
timely
attainment;
even
where
RACT
does
not
advance
attainment,
it
is
needed
in
order
to
reduce
the
severity
and
number
of
violations;
under
this
approach,
the
statutory
RACT
provisions
add
nothing
to
the
statutory
attainment
mandate
 
which
violates
basic
canons
of
statutory
interpretation;
RACT
in
nonattainment
areas
will
substantially
reduce
transport
of
ozone
and
ozone
precursors;
for
equity
reasons,
sources
in
similar
areas
should
be
subject
to
the
same
control;
and
RACT
is
a
useful
tool
that
should
not
be
abandoned
through
flexibility
mechanisms.
Response:
The
general
RACT
provision
under
subpart
1
in
the
statute,
is
found
in
section
172(
c)(
1).
It
is
a
portion
of
the
RACM
provision
found
in
that
same
section.
Our
long­
standing
interpretation
of
the
RACM
provision
is
that
areas
need
only
submit
such
RACM
as
will
contribute
to
timely
attainment
and
meet
5
RFP,
and
that
measures
which
might
be
available
but
would
not
advance
attainment
or
contribute
to
RFP
need
not
be
considered
RACM.
This
interpretation
has
been
upheld
in
several
recent
court
cases.
See
Sierra
Club
v.
EPA,
294
F.
39
155,
162
(
DC
Circuit,
2002)
(
concerning
the
Metropolitan
Washington,
D.
C.,
attainment
demonstration)
and
Sierra
Club
v.
EPA,
No.
01­
60537
(
5th
Circuit,
2002)
(
concerning
the
Beaumont
attainment
demonstration).
Since
subpart
1
RACT
is
a
portion
of
RACM,
these
cases
also
support
a
conclusion
that,
where
we
are
dealing
only
with
section
172
RACT,
it
is
reasonable
to
require
only
such
RACT
as
will
meet
RFP
and
advance
attainment.
In
view
of
these
court
cases,
EPA
disagrees
with
the
comments
listed
above
opposing
the
approach
that,
in
subpart
1
areas,
RACT
would
be
met
by
control
measures
in
a
SIP
demonstrating
attainment
of
the
standard
as
expeditiously
as
practicable
and
meeting
RFP.
The
EPA
generally
agrees
with
comments
that
States
should
have
flexibility
to
determine
which
control
strategies
are
the
most
effective
in
reaching
attainment
as
expeditiously
as
practicable
and
providing
for
RFP,
and
the
CAA
gives
primary
authority
to
States
and
local
governments
to
select
the
mix
of
controls
necessary
to
meet
the
NAAQS.
In
addition,
EPA
believes
that
section
172(
c)
is
not
the
appropriate
section
of
the
CAA
to
address
the
transport
of
ozone
and
ozone
precursors;
EPA
has
conducted
and
is
conducting
rulemaking
pursuant
to
sections
110
and
126
for
that
purpose.
Finally,
EPA
agrees
with
commenters
who
suggest,
for
equity
reasons,
that
sources
in
similar
areas
should
be
subject
to
the
same
control.
In
the
proposal,
EPA
suggested
subpart
1
and
2
areas
with
the
8­
hour
ozone
design
values
above
91
ppb
should
all
be
required
to
meet
RACT.
The
EPA
also
proposed
that
RACT
would
be
met
in
an
area
which
demonstrates
attainment
within
3
years
and
submits
the
demonstration
within
1
year.
In
the
final
rule,
EPA
has
addressed
equity
concerns
by
taking
portions
of
these
two
proposals,
such
that
subpart
1
and
2
areas
with
attainment
deadlines
longer
than
5
years
after
designation
must
meet
the
same
RACT
requirements.
We
believe
(
longer
than)
5
years
is
more
appropriate
than
the
3
years
proposed
for
this
requirement
since
this
approximates
the
maximum
attainment
date
for
subpart
2
(
moderate)
areas
subject
to
RACT
and
since
this
approach
is
consistent
with
the
manner
in
which
ROP/
RFP
requirements
are
treated
in
the
final
rule.
Therefore,
in
subpart
1
areas
that
do
not
request
an
extension
beyond
the
initial
5
years
after
designation,
the
final
rule
indicates
that
RACT
would
be
met
by
the
emission
control
measures
in
a
SIP
that
demonstrates
attainment
of
the
standard
as
expeditiously
as
practicable
and
meets
RFP.
In
addition,
the
final
rule
requires
subpart
1
areas
with
maximum
attainment
deadlines
longer
than
5
years
after
designation
to
meet
the
same
6
RACT
requirements
as
subpart
2
areas.
This
approach
minimizes
the
RACT
inequity
with
subpart
2
areas
and
provides
flexibility
for
subpart
1
areas
demonstrating
attainment
within
5
years.
Comment:
One
commenter
believes
that
new
marginal
nonattainment
areas
should
be
subject
to
RACT
under
the
8­
hour
standard
just
as
they
would
have
been
subject
to
RACT
immediately
prior
to
the
CAA
Amendments
of
1990.
Response:
Section
182(
a)
provides
that
marginal
and
higher
classified
areas
for
the
1­
hour
standard
with
pre­
1990
RACT
obligations
had
to
submit
corrections
to
their
RACT
rules
within
6
months
after
classification
under
the
1990
CAA
Amendments.
To
the
extent
that
any
8­
hour
ozone
nonattainment
areas
did
have
this
obligation,
they
already
met
it.
See
footnote
60
in
the
June
2,
2003
proposal.
The
CAA
does
not
require
RACT
for
marginal
areas
other
than
the
obligation
to
"
correct"
pre­
1990
RACT
requirements.
Comment:
The
EPA
received
several
comments
for
and
against
the
proposal
that
States
may
use
a
prior
RACT
determination
with
respect
to
the
1­
hour
ozone
standard
for
purposes
of
meeting
the
RACT
requirements
for
the
8­
hour
ozone
standard.
Further,
EPA
received
comments
on
the
proposal
that
a
new
RACT
determination
is
required
in
cases
where
the
initial
RACT
analysis
under
the
1­
hour
standard
for
a
specific
source
or
source
category
concluded
that
no
additional
controls
were
necessary.
Several
State
and
industry
commenters
supported
EPA's
proposed
approach
that
a
prior
RACT
analysis
under
the
1­
hour
ozone
standard
should
meet
RACT
requirements
under
the
8­
hour
standard
where
major
sources
or
source
categories
were
previously
reviewed
and
controls
applied
to
meet
RACT.
These
commenters
stated
that
RACT
is
not
specific
to
any
particular
ozone
standard,
such
that
once
a
source
has
met
RACT,
it
has
met
RACT,
whether
or
not
the
ozone
standard
is
revised
to
become
more
(
or
less)
stringent;
just
as
with
the
15
percent
VOC
requirement,
the
statute
provides
no
basis
for
duplicative
imposition
of
RACT;
and
there
is
no
basis
in
the
statute
to
read
in
a
new
requirement
for
RACT.
In
addition,
some
industry
commenters
stated
that
electric
generating
units
which
meet
title
IV
NOx
control
requirements
would
also
meet
the
NOx
RACT
requirement.
The
EPA
also
received
comments
from
several
States
opposing
EPA's
proposed
approach.
These
commenters
believe
the
NOx
and
VOC
guidance
is
too
old,
needs
updating
and,
in
the
case
of
NOx
controls,
the
improvement
over
the
last
three
years
has
been
dramatic
with
controls
previously
considered
to
be
BACT
(
and
therefore
generally
considered
at
the
time
to
be
more
stringent
than
RACT)
are
now
considered
to
be
merely
RACT.
In
addition,
one
State
suggested
the
presumptive
RACT
level
should
be
revised
to
at
least
85
percent
control
or
that
NOx
RACT
should
be
defined
as
up
to
$
10,000/
ton
of
pollutant
removed.
7
Two
States
disagreed
with
EPA's
proposal
that
a
new
RACT
determination
should
be
required
in
cases
where
the
initial
RACT
analysis
under
the
l­
hour
NAAQS
found
that
no
additional
controls
were
necessary
for
a
specific
source
or
source
category.
They
indicated
such
re­
analysis
would
be
an
unwise
use
of
resources
because
it
would
not
yield
significant
benefits.
Further,
they
do
not
agree
that
a
RACT
determination
is
warranted
for
major
VOC
or
NOx
sources
not
in
existence
during
the
previous
RACT
determination,
because
new
sources
in
1­
hour
nonattainment
areas
have
been
permitted
pursuant
to
the
requirements
for
NSR
and,
where
applicable,
have
already
been
subject
to
more
stringent
control
requirements.
Response:
In
1992,
EPA
set
presumptive
NOx
RACT
for
boilers
as
combustion
modification,
consistent
with
title
IV
acid
rain
requirements.
For
all
other
NOx
stationary
source
categories,
EPA
guidance
in
1994
indicated
States
should
consider
in
their
RACT
determinations
technologies
that
achieve
30­
50
percent
reduction
within
a
cost
range
of
$
160­
1300
per
ton
of
NOx
removed.
In
the
NOx
SIP
Call
Rule,
we
reviewed
all
major
NOx
source
categories
and
stated
in
the
final
rule
that
the
NOx
SIP
Call
controls,
at
less
than
$
2,000/
ton,
represent
reductions
beyond
those
required
by
RACT.
We
cannot
agree
with
the
suggestion
of
one
State
that
EPA's
RACT
guidance
should
be
revised
to
reflect
85
percent
control
and
$
10,000/
ton
of
pollutant
removed
because
the
comment
lacked
supporting
documentation
that
these
values
represent
feasible
control
levels
for
the
many
source
categories
affected
by
the
RACT
program.
Many
areas
subject
to
the
major
source
RACT
requirement
under
the
8­
hour
ozone
standard
have
previously
addressed
the
RACT
requirement
with
respect
to
the
1­
hour
ozone
standard.
For
example,
major
sources
located
in
States
of
the
OTC
were
subject
to
the
NOx
RACT
requirement
in
the
mid­
1990s.
We
believe
that,
in
many
cases,
a
new
RACT
determination
under
the
8­
hour
standard
would
result
in
the
same
or
similar
control
technology
as
the
initial
RACT
determination
under
the
1­
hour
standard
because
the
fundamental
control
techniques,
as
described
in
the
CTGs
and
ACTs,
are
still
applicable.
In
cases
where
controls
were
applied
due
to
the
1­
hour
ozone
RACT
requirement,
we
expect
the
incremental
emissions
reductions
from
application
of
a
second
round
of
controls
would
be
small
and
would
not
be
cost
effective.
In
such
cases,
EPA
believes
the
cost
per
ton
of
NOx
removed
associated
with
installing
a
second
round
of
RACT
controls
(
and
perhaps
the
removal
of
initial
RACT
controls)
is
likely
to
be
beyond
the
RACT
levels
in
our
current
guidance
($
160­$
1300/
ton).
In
contrast,
a
RACT
analysis
for
uncontrolled
sources
would
be
much
more
likely
to
find
that
RACT
level
controls
are
economically
and
technically
feasible.
The
CTGs
and
ACTs
for
VOC
were
completed
over
a
period
from
8
the
late
1970s
to
mid­
1990s
and
have
not
been
updated.
The
CTGs
are
still
used
to
presumptively
define
VOC
RACT.
The
EPA
issued
NOx
ACT
documents
between
1992
and
1995.
In
September
2000,
updates
to
the
NOx
ACT
documents
were
completed
for
stationary
internal
combustion
engines
and
cement
kilns.
The
NOx
and
VOC
ACTs
describe
available
control
techniques
and
their
cost
effectiveness,
but
do
not
define
presumptive
RACT
levels
as
the
CTGs
do.
Updating
the
ACTs
would
not,
by
itself,
change
EPA's
NOx
or
VOC
RACT
guidance,
but
it
could
provide
information
that
would
lead
to
a
new
conclusion
as
to
which
control
measures
constitute
RACT
for
a
specific
source
or
source
category.
Since
RACT
can
change
over
time
as
new
technology
becomes
available
or
the
cost
of
existing
technology
decreases,
EPA
does
not
agree
with
comments
that
once
a
source
has
met
RACT,
it
has
met
RACT
whether
or
not
the
ozone
standard
is
revised.
We
agree
that
progress
has
been
made
in
improving
the
cost
effectiveness
of
some
NOx
and
VOC
controls.
States
and
other
interested
parties
should
consider
available
information
that
may
supplement
the
CTG
and
ACT
documents.
In
cases
where
additional
information
is
presented,
for
example,
as
part
of
notice­
andcomment
rulemaking
on
a
RACT
SIP
submittal,
States
(
and
EPA)
would
necessarily
consider
the
additional
data
in
reviewing
what
control
obligation
is
consistent
with
RACT.
For
example,
as
part
of
the
1998
NOx
SIP
Call
and
2004
CAIR
rulemakings,
EPA
developed
and
used
control
and
cost
data
for
large
boilers
and
turbines
based
on
newer
information
than
contained
in
the
ACTs.
Similarly,
we
encourage
States
to
use
the
latest
information
available
in
making
RACT
determinations,
whether
it
is
in
CTGs,
ACTs,
or
elsewhere.
The
EPA
agrees
that
it
is
more
efficient
for
EPA
to
broadly
assess
what
is
RACT
for
a
specific
source
category
than
for
States
to
conduct
source­
by­
source
RACT
determinations,
especially
considering
that
States
need
to
initiate
RACT
programs
in
the
near
future
(
as
discussed
in
a
separate
comment/
response).
The
EPA's
current
RACT
guidance
may
be
used
for
purposes
of
the
8­
hour
standard.
At
the
same
time,
we
agree
with
comments
that
many
of
the
CTGs/
ACTs
have
not
been
revised
since
issued
and
thus
may
not
provide
the
most
accurate
picture
of
current
control
options.
Therefore,
we
believe
States
must
consider
new
information
that
has
become
available
and
certify
that
a
1­
hour
ozone
RACT
determination,
even
where
controls
were
required,
still
represents
an
appropriate
RACT
level
of
control
for
the
8­
hour
ozone
program.
In
the
alternative,
the
State
should
revise
the
SIP
to
reflect
a
modified
RACT
requirement
for
specific
sources
or
source
categories.
In
summary,
we
believe
the
current
NOx
and
VOC
RACT
guidance,
including
CTGs
and
ACTs,
may
continue
to
be
used
by
States
in
making
RACT
determinations
with
respect
to
the
8­
hour
9
ozone
standard.
States
should
ensure
that
their
SIPs
accurately
reflect
RACT
based
on
the
current
availability
of
technically
and
economically
feasible
controls.
Therefore,
in
portions
of
8­
hour
ozone
nonattainment
areas
where
major
sources
or
source
categories
were
previously
reviewed
and
controls
subsequently
applied
to
meet
the
RACT
requirement
under
the
1­
hour
standard,
EPA
allows
States
to
review
and,
if
appropriate,
accept
the
initial
RACT
analysis
as
meeting
the
RACT
requirements
for
the
8­
hour
standard.
Absent
data
indicating
that
the
previous
RACT
determination
is
no
longer
appropriate,
the
State
need
not
submit
a
new
RACT
SIP
for
these
sources.
In
such
cases,
the
State
should
submit
a
certification
as
part
of
its
SIP
revision,
with
appropriate
supporting
information,
such
as
consideration
of
new
data
,
that
it
previously
met
the
RACT
requirement
for
these
sources
and
that
the
previous
determination
still
represents
an
appropriate
RACT
level
of
control.
Where
the
initial
RACT
analysis
under
the
1­
hour
standard
for
a
specific
source
or
source
category
concluded
that
no
additional
controls
were
necessary,
EPA
believes
a
new
RACT
determination
is
needed.
The
new
RACT
determination
is
needed
to
consider
whether
more
cost­
effective
control
measures
have
become
available
for
sources
that
were
not
previously
regulated.
A
re­
analysis
may
determine
that
controls
are
now
economically
and
technically
feasible
and
should
be
required
to
meet
RACT.
Furthermore,
in
this
situation,
we
expect
the
incremental
emissions
reductions
to
be
significant,
compared
to
the
uncontrolled
emissions
levels,
thereby
reducing
the
cost
per
ton
of
emissions
controlled
and
therefore
more
reasonably
available.
In
all
cases
where
additional
information
is
presented
as
part
of
notice­
and­
comment
rulemaking,
including
a
RACT
SIP
submittal
for
sources
previously
controlled,
States
(
and
EPA)
must
consider
the
additional
information
as
part
of
that
rulemaking.
We
agree
with
several
State
and
industry
comments
that
RACT
requirements
apply
for
major
sources
in
any
portion
of
the
8­
hour
nonattainment
area
not
subject
to
a
RACT
program
for
the
1­
hour
standard.
Some
commenters
objected
to
EPA's
proposal
that
any
major
VOC
or
NOx
source
that
did
not
exist
during
a
previous
RACT
determination
must
be
subject
to
a
RACT
determination
as
part
of
the
SIP
for
the
8­
hour
ozone
standard.
These
commenters
stated
that
the
BACT
or
LAER
provisions
would
assure
at
least
RACT
level
controls
on
such
sources.
We
agree
this
should
be
true
in
many
cases,
but
not
all.
The
BACT/
LAER
analyses
do
not
automatically
ensure
compliance
with
RACT
since
the
applicability
thresholds
may
differ
and
the
analyses
may
be
conducted
many
years
apart.
States
may,
however,
use
information
gathered
from
prior
BACT
or
LAER
analyses
to
help
complete
a
RACT
analysis.
Comments:
Several
State
and
industry
commenters
supported
10
EPA's
proposed
approach.
These
commenters
stated
that
the
level
of
emissions
reductions
required
by
the
NOx
SIP
Call
is
far
greater
than
the
level
of
reductions
achieved
by
controls
that
have
been
determined
to
be
NOx
RACT.
Other
commenters
pointed
out
that
many
companies
have
employed
averaging
programs
for
NOx
SIP
Call
compliance
and
want
this
option
preserved
under
the
8­
hour
ozone
standard
since
requiring
sources
to
individually
meet
NOx
RACT
requirements
would
greatly
increase
the
costs
of
compliance
at
sources
already
subject
to
the
NOx
cap­
and­
trade
program
without
achieving
greater
emissions
reductions.
One
State
encourages
EPA
to
provide
this
exemption
to
other
areas
subject
to
approved
cap­
and­
trade
programs
in
addition
to
those
areas
affected
by
the
NOx
SIP
Call.
The
EPA
also
received
comments,
primarily
from
several
States
and
environmental
groups,
opposing
the
approach.
These
commenters
stated
that
there
are
no
exceptions
to
the
RACT
mandates
in
either
subpart
1
or
subpart
2
for
sources
subject
to
SIP
call
cap­
and­
trade
programs,
and
EPA
is
without
authority
to
invent
such
an
exception.
Because
the
NOx
SIP
Call's
cap­
andtrade
program
does
not
require
emission
control
technologies
to
be
installed
at
a
particular
source,
some
commenters
conclude
that
RACT
requirements
are
necessary
and
appropriate
to
ensure
that
all
sources
implement
at
least
a
minimum
level
of
control.
One
State
indicated
there
have
been
numerous
cases
where
sources
subject
to
the
NOx
SIP
Call
have
not
had
to
install
controls
comparable
to
RACT.
Commenters
also
suggested
that
RACT
is
intended
to
be
a
benchmark
for
control
technology
at
individual
stationary
sources,
not
a
level
of
regional
reductions.
In
addition,
some
commenters
noted
that
the
NOx
SIP
Call
requirements
are
specific
to
the
ozone
season,
where
RACT
requirements
are
year
round.
Consequently,
these
commenters
recommended
that
EPA
should
also
consider
non­
ozone
related
nitrogen
issues,
including
fine
particles,
visibility,
nitrification
and
acidification
of
watersheds
and
eutrophication
of
coastal
waters
all
of
which
would
be
reduced
with
year­
round
controls.
Response:
As
described
below,
EPA
believes
the
sources
that
comply
with
the
CAIR
or
NOx
SIP
Call
cap­
and­
trade
programs
meet
NOx
RACT
requirements.
Accordingly,
the
State
need
not
perform
a
NOx
RACT
analysis
for
sources
subject
to
the
State's
emission
cap­
and­
trade
program
where
the
cap­
and­
trade
program
has
been
adopted
by
the
State
or
approved
by
EPA
or
a
Federal
implementation
plan
is
in
place
meeting
the
NOx
SIP
Call
or
CAIR
(
without
opt­
in
units)
NOx
requirements.
In
addition,
the
State
need
not
submit
a
new
NOx
RACT
SIP
for
those
sources.
The
EPA
believes
the
RACT
mandate
in
subpart
1
and
subpart
2
applies
in
specific
geographic
areas
but
does
not
necessarily
require
every
major
source
to
install
controls.
For
example,
as
1In
the
1998
NOx
SIP
Call
Rule,
average
costs
of
compliance
were
estimated
at
about
$
1500/
ton
and
average
RACT
level
costs
are
less
than
$
1300/
ton.
Recent
estimates
of
the
projected
cost
of
allowances
are
about
$
2000­
4000/
ton
(
NOx
Budget
Trading
Program,
2003
Progress
and
Compliance
Report,
August
2004,
EPA­
430­
R­
04­
010).
With
respect
to
RACT,
we
would
note
that
in
some
cases,
the
State
may
determine
that
additional
controls
are
not
technically
and
economically
feasible.
Thus,
even
if
the
State
analyzed
RACT
for
each
source
subject
to
the
cap­
and­
trade
program,
that
would
not
guarantee
additional
emissions
reductions
at
each
source
because
the
RACT
determination
for
some
could
conclude
no
additional
controls
are
required.

11
discussed
in
a
separate
comment/
response,
where
we
are
dealing
only
with
subpart
1
RACT,
we
only
require
such
RACT
as
will
advance
attainment
or
meet
RFP.
Thus,
EPA
does
not
agree
with
commenters
who
conclude
that
RACT
requirements
are
necessary
and
appropriate
to
ensure
that
all
sources
implement
at
least
a
minimum
level
of
control
or
that
RACT
is
intended
to
be
a
benchmark
for
control
technology
at
all
individual
stationary
sources.
Some
commenters
pointed
out
that
the
NOx
SIP
Call
requirements
are
specific
to
the
ozone
season,
yet
RACT
requirements
are
year
round.
While
there
are
some
exceptions,
EPA
agrees
that
RACT
usually
is
an
application
of
controls
year
round.
With
respect
to
CAIR
and
the
NOx
SIP
Call,
EPA
believes
that
the
vast
majority
of
sources
in
the
cap­
and­
trade
program
will
install
combustion
controls
to
decrease
NOx
emissions
because
it
is
almost
always
the
most
cost­
effective
strategy.
Since
combustion
controls
meet
EPA's
NOx
RACT
guidance
for
level
of
control
and
since
these
controls
would
be
operated
year
round,
EPA
believes
the
RACT
requirement
would
be
met
year
round
in
the
vast
majority
of
cases.
Further,
because
many
affected
sources
also
installed
post­
combustion
controls
to
meet
the
NOx
SIP
Call
requirements,
these
additional
emissions
reductions
provide
further
assurance
that
RACT
equivalent
emissions
reductions
will
occur,
on
average,
at
sources
affected
by
the
cap­
and­
trade
program.
We
agree
that
there
would
be
non­
ozone­
related
nitrogen
benefits,
including
fine
particles,
visibility,
nitrification
and
acidification
of
watersheds
and
eutrophication
of
coastal
waters
due
to
year­
round
controls.
We
expect
the
vast
majority
of
sources
subject
to
RACT
and
rules
implemented
by
the
State
under
CAIR
or
the
NOx
SIP
Call
would
choose
to
install
RACT
level
control
equipment
rather
than
meet
the
NOx
SIP
Call
requirements
solely
by
purchasing
allowances.
Most
likely
the
cost
for
purchasing
allowances
will
be
higher
than
the
cost
for
achieving
a
RACT
level
of
control.
1
2The
EPA's
1992
NOx
RACT
guidance
sets
title
IV
equal
to
RACT
and
specifies
emission
rates
three
times
larger
than
the
rates
later
used
for
coal­
fired
units
in
the
NOx
SIP
Call
(
0.45­
0.50
lb/
mmBtu
versus
0.15).

12
In
addition,
many
units
subject
to
CAIR
or
the
NOx
SIP
Call
will
install
beyond­
RACT
controls,
as
discussed
below.
We
believe
the
combination
of
RACT­
level
controls
on
the
vast
majority
of
sources
and
the
installation
of
beyond­
RACT
controls
on
many
units
will
achieve
at
least
a
RACT
level
of
control,
on
average,
among
sources
subject
to
RACT
in
each
nonattainment
area.
Meeting
RACT
in
this
way
is
consistent
with
EPA
guidance
allowing
RACT
to
be
met
in
the
aggregate
within
a
nonattainment
area,
rather
than
by
each
individual
source.
In
addition,
through
the
NOx
SIP
Call
or
other
programs,
(
including
a
SIP
approved
cap­
and­
trade
program
outside
the
NOx
SIP
Call
area),
States
may
have
adopted
control
measures
for
specific
NOx
sources
that
equal
or
exceed
RACT
requirements.
For
these
sources,
States
may
choose
to
submit,
as
part
of
its
NOx
RACT
SIP
revision,
documentation
that
the
previously
adopted
control
measure
meets
the
RACT
requirement.
The
EPA's
NOx
RACT
guidance
(
NOx
General
Preamble
at
57
FR
55625)
encourages
States
to
develop
RACT
programs
that
are
based
on
"
areawide
average
emission
rates."
Thus,
States
can
submit
a
demonstration
as
part
of
their
RACT
submittal
showing
that
the
weighted
average
emission
rate
from
sources
in
the
nonattainment
area
subject
to
RACT
 
including
sources
reducing
emissions
to
meet
the
NOx
SIP
Call
requirements
 
meet
RACT
requirements.
We
agree
with
comments
indicating
the
overall
emission
reduction
due
to
the
NOx
SIP
Call
would
be
greater
overall
than
with
a
RACT
program
alone,
and
we
believe
they
would
be
greater
within
each
nonattainment
area.
All
States
submitting
SIP
revisions
to
meet
the
NOx
SIP
Call
(
October
27,
1998,
63
FR
57356)
elected
to
require
large
boilers
and
turbines
to
comply
with
an
emissions
cap­
and­
trade
program
consistent
with
EPA's
model
cap­
and­
trade
rule.
As
a
result,
all
these
sources
are
already
subject
to
stringent
control
requirements.
As
described
in
the
June
2,
2003
proposal,
these
sources
collectively
achieve
more
emissions
reductions
than
would
be
required
by
application
of
RACT
requirements
to
each
source.
In
the
NOx
SIP
Call
control
case,
electricity
generating
units
(
EGUs)
achieve
a
64
percent
reduction
beyond
the
base
case
requirements.
2
The
non­
EGUs
subject
to
the
States'
cap­
and­
trade
program
achieve
a
60
percent
reduction
from
uncontrolled
levels.
These
EGU
and
non­
EGU
reductions
are
clearly
beyond
the
30­
50
percent
expected
from
a
3
Memorandum
of
March
16,
1994,
from
D.
Kent
Berry
re:
"
Cost­
Effective
Nitrogen
Oxides
(
NOx)
Reasonably
Available
Control
Technology
(
RACT)."
U.
S.
Environmental
Protection
Agency,
Research
Triangle
Park,
North
Carolina.

13
RACT
program.
3
We
stated
in
the
final
rule
that
the
NOx
SIP
Call
controls
"
represent
reductions
beyond
those
required
by
Title
IV
or
Title
I
RACT."
In
addition,
because
the
cap­
and­
trade
program
covers
units
serving
a
25
megawatt
generator,
it
requires
emission
reductions
at
many
units
emitting
at
less
than
the
general
RACT
threshold
of
100
tpy.
In
some
cases,
a
facility
might
choose
to
meet
NOx
RACT
through
emissions
averaging;
e.
g.,
over
controlling
one
large
unit
and
not
controlling
other
smaller
units
because
the
smaller
units
may
be
retired
in
the
near
future
or
labor
or
capital
is
not
immediately
available.
We
agree
with
comments
that
emission
averaging
programs
such
as
the
NOx
SIP
Call
Rule
achieve
emission
reductions
at
lower
costs.
The
EPA's
NOx
RACT
guidance,
published
on
November
25,
1992
(
57
FR
55625),
was,
in
part,
for
the
purpose
of
"
enhancing
the
ability
of
States
to
adopt
marketbased
trading
systems
for
NOx"
and
to
encourage
States
to
"
structure
their
RACT
requirements
to
inherently
incorporate
an
emissions
averaging
concept
(
i.
e.,
installing
more
stringent
controls
on
some
units
in
exchange
for
lesser
control
on
others)."
Since
the
NOx
SIP
Call
requires
greater
reductions,
on
average,
than
RACT,
(
see
proposal
notice
at
68
FR
32839
and
discussion
below)
we
believe
the
average
emissions
at
the
facility
in
this
example
would
be
less
than
its
emissions
under
application
of
a
RACT
program
alone.
We
would
also
expect
this
scenario
to
be
a
short­
term
case
since
the
facility
may
choose
later
to
install
controls
on
the
smaller
units
and
sell
any
excess
allowances,
depending
on
the
projected
cost
of
allowances.
Further,
to
help
determine
if
sources
meeting
the
NOx
SIP
Call
requirements
and
subject
to
NOx
RACT
requirements
for
the
8­
hour
ozone
standard
will
be
operating
RACT­
level
combustion
controls
by
2010,
EPA
analyzed
data
from
base
case
IPM
runs
for
the
Clean
Air
Interstate
final
rule
(
CAIR).
The
data,
"
IPM
Base
Case
2010
Parsed,"
are
available
in
the
CAIR
docket
(
at
http://
www.
epa.
gov/
airmarkets/
epa­
ipm/
iaqr.
html)
and
show
which
electricity
generating
units
(
EGUs)
are
projected
to
have
NOx
controls
in
2010
before
consideration
of
the
CAIR
requirements.
As
described
below,
the
results
of
the
analysis
support
the
conclusion
that
the
vast
majority
of
sources
meeting
the
NOx
SIP
Call
requirements
and
subject
to
NOx
RACT
requirements
for
the
8­
hour
ozone
standard
would
be
operating
RACT­
level
combustion
controls
by
2010.
The
data
base
included
8,517
EGUs
in
the
CAIR
2010
emission
inventory.
Our
analysis
considered
the
738681
EGUs
in
the
data
14
base
which
are
projected
to
emit
more
than
1
ton
per
day
of
NOx
in
2010.
This
level
of
emissions
approximates
the
cutoff
level
for
"
large"
EGUs
in
the
NOx
SIP
Call
Rule.
Of
the
738
EGUs,
566
are
identified
as
having
NOx
controls
(
in
the
2010
base
case)
and
172
units
do
not
indicate
any
NOx
controls.
Of
the
172
units
without
NOx
controls,
130
units
are
not
in
areas
subject
to
8­
hr
NOx
RACT;
i.
e.,
they
are
located
in
attainment,
marginal
or
subpart
1
areas
that
are
not
expected
to
need
an
attainment
date
beyond
5
years
after
designation
(
these
subpart
1
areas
meet
RACT
through
their
attainment
demonstrations
under
the
rule
published
today).
The
initial
analysis
indicates
42
NOx
SIP
Call
EGUs
may
not
have
NOx
controls
in
2010
and
may
be
subject
to
8­
hr
NOx
RACT.
These
42
are
located
as
follows:
5­
IL
and
3­
IN
(
all
in
Chicago
metropolitan
area);
3
MO
(
all
in
St.
Louis
area);
4
OH
(
all
Cleveland);
and
27
in
the
Ozone
Transport
Region
(
2
CT;
1
DE;
8
MA;
1
MD;
2
NJ;
8
PA;
and
5
NY).
Further
analysis
indicates
30
of
the
42
EGUs
were
previously
subject
to
1­
hr
NOx
RACT
determinations
(
all
3
in
St.
Louis
and
all
27
in
the
Ozone
Transport
Region).
Because
the
1­
hour
NOx
RACT
determination
may
be
adequate
for
the
8­
hour
RACT
determination
for
these
30
EGUs,
only
12
of
the
EGUs
may
be
subject
to
a
(
new)
8­
hour
NOx
RACT
determination.
In
summary,
of
738
NOx
SIP
Call681
EGUs,
as
few
as
4­
123
­
11
are
projected
to
be
without
NOx
controls
in
2010
CAIR
base
case
and
subject
to
an
8­
hour
ozone
NOx
RACT
determination.
It
should
be
noted
that
8
of
these
12
EGUs
are
located
in
the
Chicago
area
which
was
granted
a
waiver
from
NOx
RACT
under
the
1­
hour
ozone
standard
based
on
a
dispersion
modeling
analysis.
Since
the
area
may
request
a
waiver
from
NOx
RACT
under
the
8­
hour
standard,
these
8
EGUs
might
not
be
subject
to
a
NOx
RACT
determination
for
ozone.
This
could
leave
only
4
NOx
SIP
Call
EGUs
subject
to
a
(
new)
8­
hour
ozone
NOx
RACT
determination.
Because
The
remaining
3
EGUs
are
located
in
the
Cleveland
area;
because
there
are
7
additional
large
EGUs
in
the
same
area,
all
of
which
are
expected
to
have
NOx
controls
(
low­
NOx
burners,
selective
non­
catalytic
reduction
(
SNCR)
and/
or
selective
catalytic
reduction
(
SCR)),
the
area
may
meet
RACT
by
averaging.
Additionally,
because
installation
of
combustion
controls
to
decrease
NOx
emissions
is
almost
always
the
most
cost­
effective
strategy
and
since
the
2010
CAIR
base
case
did
not
take
into
consideration
that
both
areas
are
subject
to
RACT
and
attainment
demonstration
requirements
for
both
ozone
and
PM2.5
(
the
highest
particulate
nitrate
concentrations
in
the
East
tend
to
occur
in
cooler
regions,
xxx
add
cite
to
CAIR
preamble)
EPA
expects
sources
subject
to
the
CAIR
or
NOx
SIP
Call
cap
and
trade
programs
in
these
areas
will
install
at
least
RACT
level
controls.
15
In
a
similar
manner,
the
same
2010
CAIR
base
case
data
was
used
to
determine
if
EGUs
emitting
more
than
one
ton/
day
of
NOx
will
be
operating
RACT­
level
combustion
controls
by
2010
in
the
following
scenario:
(
1)
subject
to
CAIR
for
ozone,
(
2)
not
subject
to
the
NOx
SIP
Call
requirements
and
(
3)
subject
to
NOx
RACT
requirements
for
the
8­
hour
ozone
standard.
That
is,
how
many
EGUs
outside
the
NOx
SIP
Call
area
may
meet
8­
hour
ozone
RACT
through
CAIR
and
what
controls
are
expected
to
be
in
place
at
those
EGUs
in
the
2010
base
case?
The
following
6
States
are
subject
to
CAIR
for
ozone
but
are
not
included
in
the
NOx
SIP
Call:
Arkansas,
Florida,
Iowa,
Louisiana
,
Mississippi
and
Wisconsin.
The
following
States
do
not
contain
areas
subject
to
NOx
RACT
for
the
8­
hour
ozone
standard:
Florida,
Iowa,
Louisiana
(
Baton
Rouge
is
marginal),
and
Mississippi.
Crittenden
County
in
Arkansas
is
subject
to
NOx
RACT
for
the
8­
hour
ozone
standard;
however,
there
are
no
large
EGUs
in
the
data
base
for
this
area.
The
Milwaukee
and
Sheboygan
areas
in
Wisconsin
are
subject
to
NOx
RACT
for
the
8­
hour
ozone
standard;
in
these
areas
11
of
13
large
EGUs
have
combustion
controls
for
NOx
in
the
2010
base
case.
The
2
without
NOx
controls
are
both
located
in
Milwaukee
County
(
South
Oak
Creek).
Thus,
there
may
be
only
2
EGUs
at
one
site
which
are
subject
to
CAIR
and
NOx
RACT
for
ozone,
are
outside
the
NOx
SIP
Call
area,
and
are
expected
to
be
uncontrolled
for
NOx
in
the
2010
CAIR
base
case.
Because
there
are
6
additional
EGUs
in
Milwaukee
County
with
NOx
controls
in
the
2010
base
case,
the
area
may
be
able
to
demonstrate
RACT
by
averaging
emissions
over
all
the
affected
sources.
Further,
because
installation
of
combustion
controls
to
decrease
NOx
emissions
is
almost
always
the
most
cost­
effective
strategy
and
since
the
2010
CAIR
base
case
did
not
take
into
consideration
that
the
area
is
subject
to
RACT
and
attainment
demonstration
requirements
for
both
ozone
and
PM2.5
(
the
highest
particulate
nitrate
concentrations
in
the
East
tend
to
occur
in
cooler
regions,
xxx
cite
to
CAIR
preamble),
EPA
expects
sources
subject
to
the
CAIR
or
NOx
SIP
Call
cap
and
trade
programs
in
Milwaukee
will
install
at
least
RACT
level
controls.
Under
CAIR,
a
State
may
elect
to
meet
its
state
caps
for
NOx
emissions
by
allowing
large
NOx
sources
that
are
not
electric
generating
units
to
opt­
in
to
the
program.
If
only
part
of
the
CAIR
reductions
are
required
from
EGUs,
and
the
balance
of
the
reductions
obtained
from
non­
EGU
sources,
then
the
stringency
of
CAIR
EGU
control
would
be
diminished
to
some
extent
(
an
amount
that
cannot
currently
be
determined).
In
that
event,
States
would
need
to
conduct
individual
RACT
analyses
for
EGUs
to
determine
whether
or
not
the
lesser
EGU
reductions
satisfied
RACT.
As
explained
above,
we
believe
that
the
vast
majority
of
sources
subject
to
NOx
RACT
and
rules
implemented
by
the
State
4As
described
in
the
April
21,
2004
rule
(
69
FR
21608).

16
under
CAIR
or
the
NOx
SIP
call
cap
will
install
combustion
controls
at
a
minimum,
which
will
meet
our
existing
NOx
RACT
guidance.
We
expectalso
believe
sources
could
meet
NOx
RACT
through
emissions
averaging,
recognizing
that
CAIR
and/
or
the
NOx
SIP
Call
towill
achieve
more
emissions
reductions
within
each
nonattainment
area
than
would
be
achieved
through
a
RACT
program
alone.
As
proposed,
in
cases
where
States
have
adopted
controls
for
cement
kilns
consistent
with
the
NOx
SIP
Call
(
i.
e.,
30
percent
reduction),
the
State
may
choose
to
accept
the
NOx
SIP
Call
requirements
as
meeting
the
NOx
RACT
requirements
for
the
8­
hour
standard
and
need
not
submit
a
new
NOx
RACT
SIP
for
those
sources.
In
its
RACT
SIP
submission,
the
State
should
identify
the
cement
plants
that
are
subject
to
NOx
SIP
Call
controls
and
that,
therefore,
already
meet
RACT.
The
EPA
received
comments
from
States
supporting
the
proposal.
For
the
same
reasons,
EPA
believes
a
State
may
choose
to
accept
the
Phase
II
NOx
SIP
Call
control
level
for
stationary
internal
combustion
engines4
as
meeting
the
NOx
RACT
requirements.
It
should
also
be
noted
that
this
action
in
no
way
limits
States'
discretion
to
require
beyond­
RACT
NOx
reductions
from
any
source
(
including
CAIR
or
NOx
SIP
Call
sources)
in
a
plan
to
demonstrate
attainment
of
the
health­
based
ozone
standards.
In
certain
areas,
States
may
choose
to
require
NOx
controls
based
on
more
advanced
control
technologies
to
provide
for
attainment
of
the
ozone
standards.
Although
it
is
not
the
basis
for
the
decision
in
the
ozone
implementation
rule,
we
intend
to
propose
in
the
PM2.5
implementation
rulemaking
an
alternative
interpretation
that
would
also
support
a
finding
that
in
states
which
fulfill
their
CAIR
emission
reductions
entirely
through
emission
reductions
from
EGUs,
CAIR
would
satisfy
RACT
requirements
for
EGU
sources.
This
alternative
is
outlined
below.
As
discussed
more
fully
in
the
CAIR
final
rulemaking
notice,
EPA
has
set
the
2009
CAIR
NOx
cap
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
2015
NOx
cap
is
specifically
designed
to
eliminate
all
NOx
emissions
from
EGUs
that
are
highly
cost
effective
to
control
(
the
first
cap
represents
an
interim
step
toward
that
end).
In
general,
we
expect
that
the
largest­
emitting
sources
will
be
the
first
to
install
NOx
control
technology
and
that
such
control
technology
will
gradually
be
installed
on
progressively
smaller­
emitting
sources
until
the
ultimate
cap
is
reached.
We
do
not
believe
that
requiring
source­
specific
RACT
controls
on
EGUs
in
nonattainment
areas
will
reduce
total
NOx
17
emissions
from
sources
covered
by
CAIR
below
the
levels
that
would
be
achieved
under
CAIR
alone.
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
on
smaller
emitting
sources
by
2009
could
actually
reduce
the
amount
of
"
banking"
that
would
otherwise
occur
and
result
in
higher
emissions
in
2009
as
compared
to
the
level
that
would
result
from
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
emissions
from
sources
subject
to
the
program.
Under
a
cap­
and­
trade
program
such
as
CAIR,
there
is
a
given
number
of
allowances
that
equals
a
given
emission
level.
Source­
specific
control
requirements
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
it
does
not
affect
total
emissions.
If
sourcespecific
requirements
were
targeted
at
the
units
that
can
be
controlled
most
cost­
effectively,
then
the
imposition
of
sourcespecific
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
capand
trade
program
alone.
Thus,
the
imposition
of
source­
specific
RACT
on
EGUs
covered
by
CAIR
would
not
reduce
total
emissions,
but
would
likely
achieve
the
same
total
emission
reductions
in
a
more
costly
way.
Comments:
Several
States
expressed
support
for
the
proposed
RACT
submittal
date
of
2
years
after
designation
for
subpart
1
and
2
areas.
Other
commenters
suggested
the
RACT
submittal
date
for
subpart
1
areas
should
be
3
years
after
designation
in
order
to
coincide
with
the
attainment
demonstration
submittal
deadline
and
to
allow
a
more
efficient
use
of
resources.
In
addition,
comments
from
industry
suggested
a
48­
60
month
period
is
needed
for
installation
of
controls,
rather
than
the
30
month
period
proposed.
Response:
As
described
in
an
earlier
comment/
response,
in
subpart
1
areas
that
do
not
request
an
extension
of
their
attainment
date,
RACT
is
met
with
the
control
requirements
associated
with
a
demonstration
that
the
NAAQS
is
attained
as
expeditiously
as
practicable.
The
EPA
agrees
with
commenters
that
it
would
be
more
efficient,
in
these
areas,
if
the
date
for
submittal
of
the
RACT
rules
were
to
coincide
with
submittal
of
5This
assumes
the
ozone
season
in
this
example
begins
May
1.
The
maximum
statutory
timeframe
for
subpart
1
areas
that
do
not
seek
an
extension
is
June
15,
2009.
Since
the
CAA
requires
attainment
as
expeditiously
as
practicable,
some
subpart
1
nonattainment
areas
may
have
an
attainment
date
earlier
than
June
15,
2009.

6In
the
1990
CAA
Amendments,
Congress
specifically
added
RACT
requirements
for
major
sources
in
section
182.
Section
182
required
the
RACT
rules
to
be
implemented
"
as
expeditiously
as
practicable"
but
no
later
than
30
months
after
the
submittal
deadline.

18
the
attainment
demonstration
since
RACT
is
closely
tied
to
the
attainment
demonstration.
Therefore,
in
the
final
rule,
the
RACT
submittal
date
for
these
areas
is
the
same
as
the
submittal
date
for
the
attainment
plan,
which
is
3
years
after
designation
(
June
2007).
Although
EPA
is
not
setting
a
specific
RACT
rule
implementation
deadline
for
these
areas,
as
provided
in
the
Phase
1
rule,
controls
must
be
implemented
by
the
beginning
of
the
attainment
year
ozone
season.
Thus,
for
example,
States
would
need
to
require
implementation
no
later
than
May
1,
2008
in
order
to
help
areas
attain
the
ozone
standard
by
a
June
15,
2009
attainment
date.
5
In
some
cases,
the
time
from
State
rule
adoption
to
installation
of
controls
by
sources
may
be
relatively
short;
in
other
cases,
sources
may
need
more
time.
Therefore,
EPA
encourages
States
to
adopt
any
necessary
rules
expeditiously
(
prior
to
the
June
2007
deadline,
where
possible)
so
that
sources
have
more
than
sufficient
time
to
install
the
controls
prior
to
the
start
of
the
attainment
year
ozone
season.
For
subpart
2
moderate
and
above
areas
and
areas
within
an
OTR,
the
final
rule
is
similar
to
provisions
in
section
182
of
the
CAA
which
require
States
to
submit
RACT
rules
for
these
areas
within
24
months
after
the
designation.
Several
commenters
supported
this
approach.
Since
some
States
may
rely
on
submittal
of
SIP
revisions
meeting
CAIR
to
also
satisfy
RACT
for
some
sources,
the
final
rule
extends
the
proposed
RACT
submittal
date
of
24
months
to
27
months
after
designation
(
September
15,
2006),
to
be
consistent
with
the
date
for
submittal
of
the
CAIR
provisions
(
September
10,
2006).
For
areas
subject
to
the
27­
month
RACT
submittal
date,
EPA
believes
the
proposed
30­
month
period
for
installation
of
controls
is
reasonable,
given
that
this
is
the
statutorilyprescribed
period6
(
for
the
areas
covered
under
subpart
2)
and
based
on
our
prior
experience
with
States
adopting
and
implementing
RACT
requirements.
Subsequent
to
the
NOx
RACT
SIP
revisions
for
the
1­
hour
standard
subject
to
the
30­
month
CAA
period,
EPA
approved
NOx
RACT
SIP
submittals
in
some
areas
which
7Note,
since
the
CAA
requires
attainment
as
expeditiously
as
practicable,
some
moderate
nonattainment
areas
may
have
an
attainment
date
earlier
than
June
15,
2010.

19
had
been
exempt
from
the
requirements,
including
the
Dallas
and
Houston
areas,
which
required
implementation
within
2
years
from
the
State
adoption
date.
The
EPA
recently
determined
that
a
24­
month
period
is
adequate
for
stationary
internal
combustion
engines
to
install
low
emission
combustion
controls
(
April
21,
2004;
69
FR
21633).
Furthermore,
EPA
believes
30
months
is
adequate
since
available
data
support
that
one
of
the
most
challenging
retrofit
installations
 
application
of
selective
catalytic
reduction
(
SCR)
as
described
in
the
January
2004
CAIR
proposal
 
can
be
completed,
on
average,
in
21
months
(
69
FR
4617).
The
60­
month
period
(
June
15,
2011)
for
installation
of
controls
suggested
by
some
commenters
was
not
adequately
supported,
i.
e.,
there
was
a
lack
of
detailed
justification
that
more
time
is
necessary.
In
addition,
as
described
in
an
earlier
comment/
response,
EPA
anticipates
that
many
sources
which
applied
controls
due
to
RACT
requirements
with
the
1­
hour
ozone
standard
will
not
need
to
install
new
controls
for
the
8­
hour
standard.
Thus,
because
fewer
sources
will
be
subject
to
a
new
RACT
requirement,
there
will
be
less
demand
for
control
equipment.
States
and
many
sources
face
a
reduced
burden
compared
to
the
same
CAA
requirement
in
the
1990s.
Since
the
ozone
season
(
40
CFR
part
58,
appendix
D)
does
not
begin
for
many
areas
until
May
1,
however,
the
final
rule
allows
sources
until
the
beginning
of
the
area's
2009
ozone
season
(
generally
May
1,
2009)
rather
than
March
15,
2009
to
install
controls.
Installation
of
controls
before
the
2009
ozone
season
is
sufficient
to
provide
the
benefits
for
timely
attainment
of
the
ozone
standard
in
areas
with
a
2010
or
later
attainment
date.
7
And
the
short
delay
(
generally
between
March
15,
2009
and
May
1,
2009)
will
cause
no
harm
since
it
is
prior
to
the
ozone
season,
which
is
when
ozone
levels
are
most
likely
to
be
at
harmful
levels.
Sources
meeting
NOx
RACT
through
compliance
with
CAIR
would
implement
controls
by
January
1,
2009,
as
required
under
CAIR.
For
subpart
1
areas
that
request
an
attainment
date
extension
(
beyond
5
years
after
designation),
the
final
rule
sets
the
RACT
submittal
and
implementation
dates
the
same
as
required
for
subpart
2
moderate
and
above
areas,
except
subpart
1
areas
shall
submit
the
RACT
SIP
with
its
attainment
date
extension
request.
The
EPA
believes
this
approach
is
reasonable
and
provides
equity
for
subpart
1
and
2
areas
with
attainment
dates
beyond
5
years
after
designation.
8"
State
Implementation
Plans;
General
Preamble
for
the
Implementation
of
Title
I
of
the
Clean
Air
Act
Amendments
of
1990;
Proposed
Rule."
57
FR
13498
at
13560
(
April
16,
1992).
"
Guidance
on
the
Reasonably
Available
Control
Measures
(
RACM)
Requirement
and
Attainment
Demonstration
Submissions
for
Ozone
Nonattainment
Areas."
John
S.
Seitz,
Director,
Office
of
Air
Quality
Planning
and
Standards.
November
30,
1999.
Web
site:
www.
epa.
gov/
ttn/
oarpg/
t1pgm.
html.
Memorandum
of
December
14,
2000,
from
John
S.
Seitz,
Director,
Office
of
Air
Quality
Planning
and
Standards,
re:
"
Additional
Submission
on
RACM
from
States
with
Severe
One­
Hour
Ozone
Nonattainment
Area
SIPs."

20
2.
Reasonably
Available
Control
Measures
(
RACM)
a.
Background
As
noted
in
the
June
2,
2003
proposed
rule,
subpart
1
of
part
D
includes
general
requirements
for
all
designated
nonattainment
areas,
including
a
requirement
that
a
nonattainment
plan
provide
for
the
implementation
of
all
RACM
as
expeditiously
as
practicable,
including
such
reductions
that
may
be
obtained
through
RACT.
We
have
also
issued
guidance
for
implementing
the
RACM
provisions
of
the
CAA
that
interprets
those
provisions
to
require
a
demonstration
that
the
State
has
adopted
all
reasonable
measures
to
meet
RFP
and
attainment
as
expeditiously
as
practicable
and
thus
that
no
additional
measures
that
are
reasonably
available
will
advance
the
attainment
date
or
contribute
to
RFP
for
the
area.
8
The
RACM
requirement,
which
is
set
forth
in
section
172(
c)(
1)
of
the
CAA,
applies
to
all
nonattainment
areas
that
are
required
to
submit
an
attainment
demonstration,
whether
covered
under
only
subpart
1
or
also
subpart
2.
The
June
2,
2003
proposal
implied
that
the
current
policies
and
procedures
related
to
RACM
would
continue
to
apply
for
the
8­
hour
ozone
standard.
The
draft
regulatory
text
provided
that
for
each
nonattainment
area
required
to
submit
an
attainment
demonstration
under
§
51.908,
the
State
would
have
to
submit
with
the
attainment
demonstration
a
SIP
revision
demonstrating
that
it
has
adopted
all
control
measures
necessary
to
demonstrate
attainment
as
expeditiously
as
practicable
and
to
meet
any
RFP
requirements.
b.
Summary
of
final
rule
Section
51.912(
d)
of
the
final
rule
reflects
our
proposal
and
draft
regulatory
text.
For
each
nonattainment
area
required
to
submit
an
attainment
demonstration
under
§
51.908,
the
State
must
submit
with
the
attainment
demonstration
a
SIP
revision
demonstrating
that
it
has
adopted
all
control
measures
necessary
to
demonstrate
attainment
as
expeditiously
as
practicable
and
to
meet
any
RFP
requirements.
9"
State
Implementation
Plans;
General
Preamble
for
the
Implementation
of
Title
I
of
the
Clean
Air
Act
Amendments
of
1990;
Proposed
Rule."
57
FR
13498
at
13560
(
April
16,
1992).

21
In
performing
their
RACM
analysis,
States
should
be
aware
of
EPA's
conclusions
concerning
compliance
dates
in
the
CAIR
rulemaking
(
May
12,
2005
(
70
FR
at
25221
et
seq.))
that
represents
an
agressive
schedule
that
reflects
the
limitations
of
the
labor
pool,
and
equipment/
vendor
availability,
and
need
for
electrical
generation
reliability
for
installation
of
NOx
emission
controls.
States
should
be
aware
of
these
limitations
in
developing
their
own
compliance
schedules
for
NOx
emission
controls
in
meeting
their
CAIR
and
RACT
responsibilities
and
use
EPA's
schedule
as
a
guide.
We
believe
CAIR
constitutes
control
as
expeditious
as
practicable
for
EGUs
for
the
area
covered
by
the
CAIR.
c.
Comments
and
responses
Comment:
One
commenter
asked
that
we
clarify
whether
old
SIP
measures
become
RACM.
Response:
Under
EPA's
policy
concerning
RACM,
there
are
no
measures
that
are
automatically
deemed
RACM.
The
determination
of
whether
a
SIP
contains
all
RACM
requires
an
area­
specific
analysis
that
there
are
no
additional
economically
and
technologically
feasible
control
measures
(
alone
or
in
conjunction
with
others)
that
will
advance
the
attainment
date.
9
Comment:
One
commenter
recommended
that
EPA
revise
its
policy
permitting
SIPs
to
exclude
otherwise
feasible
and
potentially
RACM
that
achieve
emissions
reductions
in
increments
less
than
the
amount
necessary
to
advance
the
attainment
date
by
a
full
year.
The
commenter
believed
this
was
an
onerous
standard
that
has
stymied
development
of
new
control
measures,
particularly
transportation
control
measures.
The
commenter
believed
EPA's
RACM
standard
is
especially
harmful
to
the
ability
to
provide
SIP
credit
for
Smart
Growth
land
use,
due
to
the
long
timeframe
over
which
land
is
developed
and
redeveloped.
The
commenter
believes
that
ever­
increasing
suburbanization
of
our
nation
inflates
the
growth
rate
in
VMT,
thereby
neutralizing
improvements
in
vehicle
emissions.
The
commenter
claimed
that
a
significant
air
quality
improvement
strategy
for
the
21st
Century
is
compact
mixed
use
pedestrian­
friendly
development
near
frequent
transit
and
believed
that
changing
land
use
plans
in
this
direction
will
benefit
air
quality
by
reducing
the
rate
of
growth
in
VMT
and
emissions.
The
commenter
recommended
that
EPA
be
aware
of
this
and
revise
its
RACM
standard
to
encourage
local
governments
to
alter
their
land
use
plans
by
providing
a
mechanism
to
give
credit
for
air
quality
beneficial
land
use
10Improving
Air
Quality
Through
Land
Use
Activities;
Transportation
and
Regional
Programs
Division,
Office
of
Transportation
and
Air
Quality,
U.
S.
Environmental
Protection
Agency.
EPA420­
R­
01­
001.
January
2001.

22
changes.
Response:
We
do
not
believe
our
RACM
policy
has
"
stymied"
development
of
new
control
technologies.
New
emission
reduction
technologies
have
surfaced
and
continue
to
surface
to
meet
market
demands
resulting
in
part
from
CAA
requirements,
which
include
the
requirements
to
demonstrate
attainment
as
expeditiously
as
practicable
and
to
make
RFP
toward
attainment.
In
addition,
viable
control
measures
that
produce
emissions
reductions
can
be
approved
into
SIPs
whether
or
not
such
measures
meet
the
definition
of
RACM.
Our
RACM
policy
merely
interprets
the
CAA
as
not
mandating
measures
that
do
not
contribute
to
expeditious
attainment
and
timely
RFP.
The
policy
does
not
limit
the
potential
for
States
to
develop
any
control
measures
they
wish,
including
land
use
measures.
In
fact,
we
have
prepared
a
separate
guidance
document
on
how
areas
can
develop
and
receive
SIP
credit
for
land
use
control
measures.
10
We
conclude,
however,
that
to
require
areas
to
adopt
and
implement
as
RACM
every
control
technology
or
measure
that
obtains
a
small
amount
of
emissions
reductions
even
if
such
measure
would
not
advance
the
attainment
date
or
contribute
to
RFP
is
not
justified.
Such
a
policy
would
be
extremely
burdensome
to
planning
agencies,
would
detract
from
the
effort
to
develop
more
reasonable
and
effective
controls
to
meet
the
NAAQS,
and
would
not
be
necessary
to
meet
the
statutory
goal
of
expediting
attainment.
For
these
reasons,
and
because
such
requirement
is
not
mandated
by
the
statute,
we
are
not
adopting
such
a
policy.
Comment:
One
commenter
believed
that
the
RACM
requirements
for
subpart
1
areas
should
be
designed
so
as
to
not
require
extensive
and
unneeded
control
due
to
the
fact
that
in
most
or
all
cases
these
controls
will
not
be
needed
for
the
area
to
attain.
Response:
We
believe
the
current
RACM
guidance,
which
applies
to
both
subpart
1
and
2
areas,
works
to
avoid
extensive
and
unneeded
controls,
while
ensuring
that
areas
meet
the
healthbased
NAAQS
as
expeditiously
as
practicable.
Comment:
One
commenter
believed
our
RACM
guidance
provides
only
minimum
requirements
to
ensure
attainment
as
expeditiously
as
practicable
and
believes
that
every
nonattainment
area
must
be
required
to
consider
adoption
of
measures
that
have
been
implemented
in
other
areas,
including
the
South
Coast
of
California,
so
as
to
achieve
progress
and
attainment
as
expeditiously
as
practicable.
An
area
should
be
allowed
to
11In
"
State
Implementation
Plans;
General
Preamble
for
the
Implementation
of
Title
I
of
the
Clean
Air
Act
Amendments
of
1990;
Proposed
Rule,"
we
noted
in
the
discussion
of
the
RACM
requirement
that
"
In
addition,
any
measure
that
a
commenter
indicates
during
the
public
commenter
period
is
reasonably
available
for
a
given
area
should
be
closely
reviewed
by
the
planning
agency
to
determine
if
it
is
in
fact
reasonably
available
for
implementation
in
the
area
in
light
of
local
circumstances."
The
discussion
of
RACM
in
that
document
contains
other
relevant
history
concerning
the
RACM
requirement.
reject
such
measures
only
upon
a
showing
that
they
are
not
practicable
due
to
specified
unique
circumstances.
The
commenter
urged
that
given
the
importance
of
this
issue
to
fair,
expeditious
and
lawful
implementation
of
the
8­
hour
standard,
EPA's
final
8­
hour
standard
implementation
rule
must
explicitly
require
compliance
with
this
guidance.
Response:
To
meet
the
RACM
provision
of
the
CAA,
the
State
must
determine
as
part
of
its
attainment
demonstration
whether
there
are
additional
measures
that
are
feasible
that
would
expedite
attainment.
In
addition,
EPA's
RACM
policy
indicates
that
areas
should
consider
all
candidate
measures
that
are
potentially
available,
including
any
that
have
been
suggested
for
the
particular
nonattainment
area.
11
Although
areas
should
consider
all
available
measures,
including
those
being
implemented
in
other
areas
such
as
California,
areas
need
adopt
measures
only
if
they
are
both
economically
and
technologically
feasible
and
will
advance
the
attainment
date.
This
interpretation
of
the
section
172
requirements
has
recently
been
upheld
by
several
courts.
See,
e.
g.,
Sierra
Club
v.
EPA,
et
al.,
294
F.
3d
155
(
D.
C.
Circuit,
2002).
Comment:
Several
commenters
agreed
with
our
proposal
to
require
that
the
RACM
analysis
and
measures
be
submitted
within
3
years
after
the
effective
date
of
designation
for
the
8­
hour
NAAQS.
Response:
We
acknowledge
the
support
of
the
comments
on
the
submission
timing
of
the
RACM
requirements.
