DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
1
DAPN
(
a.
k.
a.
"
NSR
III")
Proposed
Rule
Qualitative
Environmental
Analysis
I.
OVERVIEW
In
May
2001,
President
Bush's
National
Energy
Policy
Development
Group
issued
findings
and
key
recommendations
for
a
National
Energy
Policy.
This
document
included
numerous
recommendations
for
action,
including
a
recommendation
that
the
EPA
Administrator,
in
consultation
with
the
Secretary
of
Energy
and
other
relevant
agencies,
review
NSR
regulations,
including
administrative
interpretation
and
implementation.
The
recommendation
requested
that
we
issue
a
report
to
the
President
on
the
impact
of
the
regulations
on
investment
in
new
utility
and
refinery
generation
capacity,
energy
efficiency,
and
environmental
protection.

In
response,
in
June
2001,
we
issued
a
background
paper
giving
an
overview
of
the
NSR
program.
We
solicited
public
comments
on
the
background
paper
and
other
information
relevant
to
the
NSR
90­
day
Review
and
Report
to
the
President.
During
our
review
of
the
NSR
program,
we
met
with
more
than
100
groups,
held
four
public
meetings
around
the
country,
and
received
more
than
130,000
written
comments.
Our
Report
to
the
President
and
our
recommendations
in
response
to
the
energy
policy
were
issued
on
June
13,
2002.

We
previously
finalized
responses
to
the
energy
policy
recommendations
on
December
31,
2002
(
67
FR
80186)
and
October
27,
2003
(
68
FR
61248).
We
are
now
proposing
regulations
for
"
aggregation"
and
"
debottlenecking"
that
are
two
remaining
recommendations.
We
also
are
proposing
a
change
to
our
past
policy
for
project
netting.
This
document
analyzes
the
environmental
impact
of
these
proposals.
As
described
below,
we
have
determined
that
these
proposed
reforms
for
debottlenecking,
aggregation,
and
project
netting
will
provide
greater
regulatory
certainty
without
sacrificing
the
current
level
of
environmental
protection
and
benefit
derived
from
the
NSR
program.
We
believe
that
these
changes
will
facilitate
the
safe,
efficient,
and
reliable
operation
of
affected
facilities.

This
document
is
not
intended
to
be
a
formal
Regulatory
Impact
Analysis
(
RIA)
or
Economic
Impact
Analysis
as
those
terms
are
used
in
Agency
rulemaking.
The
proposed
NSR
DAPN
rule
does
not
require
a
formal
RIA
under
Executive
Order
12866
because
the
rule
does
not
have
an
annual
effect
on
the
economy
of
$
100
million
or
more
or
adversely
affect
in
a
material
way
the
economy,
a
sector
of
the
economy,
productivity,
competition,
jobs,
the
environment,
public
health
or
safety,
or
State,
local,
or
tribal
governments
or
communities.

This
analysis
qualitatively
examines
whether
changes
in
emissions
are
likely
to
occur
as
a
result
of
these
proposed
rule
provisions,
but
does
not
attempt
to
assign
monetary
values
to
any
such
changes.
This
analysis
also
does
not
assign
monetary
values
to
the
other
types
of
benefits
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
2
that
we
expect
to
occur
as
a
result
of
this
rule
proposal,
such
as
the
reduction
in
administrative
costs
from
the
streamlining
of
the
permit
process
and
the
decreased
opportunity
cost
from
delayed
changes.
Quantifying
these
non­
environmental
benefits
is
outside
the
scope
of
this
analysis.

II.
INTRODUCTION
The
purpose
of
this
document
is
to
provide
information
on
the
potential
environmental
effects
of
the
NSR
rules
that
EPA
is
proposing.
This
information
is
intended
to
provide
additional
information
to
the
public
as
a
basis
for
developing
these
rules.
Indeed,
the
EPA's
long
and
detailed
records
support
the
need
for
improving
the
NSR
program,
and,
consistent
with
standard
rulemaking
processes,
we
are
fully
explaining
the
legal
and
policy
basis
for
our
actions.
The
proposed
rules
are
fully
justified
as
a
legal
and
policy
matter,
and
the
soundness
of
EPA's
qualitative
basis
for
the
rule
does
not
depend
on
its
ability
to
specifically
quantify
the
environmental
impact
of
the
rule.

It
is
important
to
note
that
NSR
is
one
of
many
programs
created
by
the
Clean
Air
Act
to
control
or
reduce
emissions
of
air
pollutants
 
particularly
criteria
pollutants
 
that
are
emitted
from
a
wide
variety
of
sources
and
have
an
adverse
impact
on
human
health
and
the
environment.
Other
key
programs
include:
the
title
IV
Acid
Rain
Program,
Maximum
Achievable
Control
Technology
(
MACT)
and
other
air
toxics
standards
for
control
of
Hazardous
Air
Pollutants
(
HAPs),
New
Source
Performance
Standards,
the
Clean
Air
Interstate
Rule
(
CAIR),
the
22­
state
NOx
"
SIP
Call,"
the
Clean
Air
Visibility
Rule
(
CAVR)
and
the
Regional
Haze
program,
numerous
mobile
source
programs,
and
the
basic
state
and
local
air
control
programs
to
attain
and
maintain
the
National
Ambient
Air
Quality
Standards
(
NAAQS).
Together,
these
programs
have
achieved,
and
will
continue
to
achieve,
tens
of
millions
of
tons
per
year
of
reductions
which
are
completely
unaffected
by
this
proposed
rule.

While
the
variety
of
programs
mentioned
above
will
continue
to
play
the
dominant
role
in
reducing
emissions
of
air
pollution,
the
NSR
program
will
continue
to
serve
its
intended
role
of
assuring
that
new
and
significantly
modified
sources,
when
they
increase
their
emissions,
are
wellcontrolled
and
permitted
consistent
with
these
overall
air
quality
management
programs.
NSR
is
a
broad
program
that
covers
new
and
modified
sources
across
a
wide
range
of
source
categories,
and
only
a
small
portion
of
the
sources
covered
by
NSR 
specifically,
those
major
sources
[
FOOTNOTE]
that
undergo
physical
changes
or
changes
in
the
method
of
operation
 
would
be
affected
by
these
proposed
regulatory
revisions.
The
NSR
recommendations
by
the
Energy
Policy
Task
Force
provide
a
targeted
set
of
rule
changes
that
focus
on
issues
related
to
modifications
to
existing
emissions
units.
EPA
estimates
that
more
than
80
percent
of
the
NSR's
benefits
come
from
regulating
new
sources
and
new
units
at
existing
sources.
[
FOOTNOTE]

Having
set
forth
the
overall
context
of
this
analysis,
there
are
fundamental
limitations
on
the
ability
to
do
a
full
quantitative
analysis
of
the
environmental
benefits
of
these
proposed
NSR
reforms.
In
many
EPA
air
rules,
it
is
possible
to
do
a
comprehensive
quantitative
analysis
of
the
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
1
U.
S.
EPA.
"
Suppplemental
Analysis
of
the
Environmental
Impact
of
the
2002
Final
NSR
Improvement
Rules."
November
21,
2002.

3
health
and
environmental
benefits
of
the
regulation.
These
types
of
analyses
rely
upon
the
ability
to
estimate
the
effects
that
the
regulation
is
expected
to
have
on
emissions
over
time.
If
the
locations
of
the
projected
emissions
changes
are
reasonably
well­
known,
models
can
be
used
to
estimate
air
quality
impacts,
and
this
information
can
be
used
to
estimate
resulting
health
and
environmental
benefits.
Thus,
where
one
can
reasonably
quantify
the
projected
emissions
impacts
of
a
particular
rule,
it
is
possible
to
estimate
that
rule's
impact
on
public
health.
However,
for
reasons
explained
below,
the
EPA
cannot
quantify
with
specificity
the
emissions
changes
for
a
given
pollutant
or
pollutants,
if
any,
that
could
result
from
these
proposed
NSR
rule
changes,
nor
can
we
reliably
determine
the
anticipated
locations
of
any
emissions
changes.
The
following
list
illustrates
the
reasons
why
a
more
detailed
provision­
by­
provision
health
analysis
of
the
proposed
NSR
revisions
is
not
possible:

°
New
vs.
Modified
sources.
EPA's
previously
reported
estimates
of
the
magnitude
of
NSR
benefits
are
calculated
based
on
permitting
data
for
all
types
of
sources
 
new
"
greenfield"
facilities,
new
units
at
existing
facilities,
and
modifications
to
existing
units.
1
However,
these
proposed
NSR
reforms
apply
primarily
to
modifications
at
existing
units.
As
noted
above,
less
than
one­
fifth
of
the
1997­
99
PSD
benefits
were
from
projects
that
involved
modifications
to
existing
units.
However,
the
data
available
do
not
allow
identification
of
modified
units
with
sufficient
specificity
(
e.
g.,
location,
number,
size
of
affected
units)
to
develop
any
more
detailed
estimate
of
the
possible
benefits
of
these
proposed
rules.

°
Difficult
to
link
permits
to
environmental
results.
The
proposed
NSR
rule
changes
relate
primarily
to
the
provisions
governing
whether
a
source
must
obtain
an
NSR
permit.
Although
we
believe
that
slightly
fewer
sources
will
undergo
the
full
NSR
permitting
process
as
a
result
of
these
proposed
rule
changes,
many
of
these
sources
will
still
be
subject
to
alternate
NSR
provisions
that
the
EPA
finalized
in
2002
(
e.
g.,
the
requirement
to
cap
emissions).
There
is
not
a
straightforward
relationship
between
the
changes
in
number
of
permits
and
the
real
changes
in
emissions
resulting
from
the
collection
of
these
proposed
reforms
and
the
previously
finalized
reforms.
Also,
the
number
of
NSR
permits
is
a
poor
indicator
for
the
program's
environmental
benefit
for
several
reasons:

°
The
emissions
benefits
that
result
from
an
NSR
permit
process
vary
widely.
They
may
be
quite
large
in
cases
such
as
a
new
greenfield
source,
but
can
be
negligible
(
or
zero)
in
such
cases
as
a
modification
to
an
already
well­
controlled
unit.

°
The
avoidance
of
a
permitting
process
does
not
necessarily
mean
that
emissions
will
increase.
As
discussed
in
our
June
2002
report
to
the
President,
we
concluded
that
the
NSR
program,
in
a
variety
of
ways,
has
impeded
or
resulted
in
the
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
4
cancellation
of
projects
that
would
have
maintained
and
improved
the
reliability,
efficiency,
or
safety
of
existing
energy
capacity.
The
impediments
exist
for
nonenergy
sources
as
well.
In
short,
we
believe
that
our
past
policies,
some
of
which
are
simply
unclear,
can
have
the
effect
of
leading
sources
to
artificially
constrain
production
and
have
discouraged
plant
owners
or
operators
from
engaging
in
projects
that
are
important
to
restoring,
maintaining
and
improving
plant
safety,
reliability,
and
efficiency.

°
The
type
of
benefit,
if
any,
that
results
from
a
permit
process
depends
on
the
type
of
source,
the
pollutants
it
will
emit,
and
the
air
quality
in
the
area
where
it
locates.
We
cannot
resolve
these
differences
as
they
may
relate
to
these
proposed
reforms,
because
it
is
not
possible
to
model
source
behavior
with
sufficient
specificity.
This
not
only
affects
the
ability
to
generate
emissions
estimates,
but
also
severely
hinders
any
ability
to
relate
this
to
health
benefits.

°
The
NSR
program
can
lead
to
changes
in
source
behavior
that
have
environmental
effects
even
for
sources
that
do
not
get
an
NSR
permit.
Many
of
these
effects
may
be
such
that
elimination
of
a
permit
requirement
actually
yields
an
environmental
benefit.
Similarly,
the
rule
removes
barriers
and
creates
incentives
for
more
energy
efficient
or
lower­
emitting
processes,
but
does
so
without
requiring
a
full
NSR
permit
process.
Permitting
data
tell
us
little
about
these
effects,
and
it
is
very
difficult
to
model
source
behavior
to
predict
these
effects
some
other
way.

°
The
NSR
program
allows
for
emissions
increases
to
occur,
as
long
as
they
are
well
controlled.
From
an
overall
air
quality
standpoint,
the
facilities
that
go
through
NSR
do
not
necessarily
get
emissions
reductions
needed
to
attain
and
maintain
air
quality
standards,
and
may,
in
fact,
increase
emissions.
Absent
other
air
quality
management
programs,
emissions
avoided
by
NSR
may
simply
appear
elsewhere
in
the
same
airshed.
Thus,
a
higher
level
of
NSR
permitting
in
an
area
does
not
necessarily
indicate
cleaner
air.

°
Detailed
records
not
usually
kept.
These
proposed
reforms
relate
mainly
to
the
NSR
applicability
process,
which
is
an
element
of
the
NSR
program
for
which
there
is
only
limited
quantitative
information
available.
While
sources
may
do
detailed
applicability
calculations
to
determine
whether
to
apply
for
a
permit,
EPA
does
not
have
records
of
these
calculations
because
(
1)
when
a
source
determines
not
to
apply
for
an
NSR
permit,
EPA
is
usually
not
notified,
and
(
2)
when
a
source
does
apply
for
an
NSR
permit,
it
is
not
required
to
prove
to
EPA
or
the
State
permitting
agency
that
it
is
subject.

Because
of
these
and
other
difficulties,
it
is
very
difficult
to
model
the
likely
changes
in
emissions
or
air
quality
that
will
occur
as
a
result
of
these
proposed
provisions
and,
thus,
to
carry
forward
a
health
analysis
of
these
changes.
Clearly,
the
following
assessment
is
qualitative,
not
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
2
For
an
explanation
of
the
term
"
debottlenecked"
emissions
units,
or
"
debottlenecking,"
see
preamble
section
III.
A.
2
of
the
proposed
rule.

5
quantitative.
Nonetheless,
the
EPA
understands
that,
where
available,
quantitative
information
can
provide
a
useful
supplement
to
these
conclusions.
For
that
reason,
we
are
specifically
requesting
data
through
this
rule
proposal
on
the
potential
environmental
impacts
of
these
reforms.
If
we
are
presented
with
usable
data
to
analyze,
we
will
include
that
information
in
the
analysis
for
the
final
rule.

III.
PRELIMINARY
FINDINGS
A.
Debottlenecking
Our
current
regulations
define
a
"
major
modification"
as
one
in
which
a
physical
change
or
a
change
in
the
method
of
operation
of
a
major
stationary
source
results
in
a
significant
emissions
increase
of
a
regulated
NSR
pollutant
("
Step
1"
of
NSR
applicability)
and
a
significant
net
emissions
increase
of
that
pollutant
at
the
source
("
Step
2").
The
total
increase
in
emissions
that
must
be
included
in
determining
if
there
will
be
a
post­
change
significant
emissions
increase
includes:
(
1)
increases
occurring
at
all
new
or
modified
units,
and
(
2)
any
other
increases
at
existing
emissions
units
not
being
modified
which
could
experience
emissions
increases
related
to
the
change.
Thus,
for
a
new
project
at
an
existing
unit,
the
emission
increase
associated
with
the
project
is
based
on
the
"
actual­
to­
projected­
actual"
test
and
includes
increases
projected
not
only
from
the
unit(
s)
undergoing
the
change
but
also
from
all
other
unit
at
the
major
stationary
source
that
result
from
the
change,
such
as
units
that
are
debottlenecked.
2
When
an
emissions
increase
is
projected
at
a
debottlenecked
emissions
unit
as
a
result
of
a
change
occurring
at
another
unit,
the
increase
must
be
added
to
the
increase
projected
at
the
changed
unit,
along
with
the
sum
of
all
contemporaneous
emissions
increases
and
decreases,
to
determine
whether
NSR
applies
to
the
source.
Consequently,
even
when
a
project
increases
emissions
by
less
than
a
significant
amount
at
the
changed
unit,
the
project
would
trigger
major
NSR
if:
(
1)
it
debottlenecks
another
unit
at
the
source;
(
2)
the
emissions
increase
(
of
that
same
pollutant)
is
large
enough
at
the
debottlenecked
unit
that
there
is
a
significant
emissions
increase;
and
(
3)
there
are
too
few
contemporaneous
emissions
decreases
to
avoid
having
a
significant
net
emissions
increase.
If
NSR
applies,
then
the
source
goes
through
permitting,
the
changed
unit
undergoes
a
Best
Available
Control
Technology
(
BACT)
or
Lowest
Achievable
Emissions
Rate
(
LAER)
analysis,
and
the
emissions
increase
is
accounted
for
in
the
air
quality
analysis.

In
this
analysis,
we
compare
our
proposed
debottlenecking
provisions
to
the
current
rule
provisions.
The
proposed
provisions
would
apply
the
same
actual­
to­
projected
actual
test
to
debottlenecked
units
as
do
the
current
rules,
but
only
in
cases
where
"
causation"
is
established.
Thus,
when
a
project
at
an
emissions
unit
debottlenecks
an
upstream
or
downstream
unit,
the
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
6
proposed
rules
do
not
automatically
assume
that
any
emissions
change
at
the
unchanged
unit
resulted
from
the
project
unless
the
emissions
increase
could
not
have
been
accommodated
prior
to
the
change.
This
approach
is
consistent
with
how
our
"
actual­
to­
projected­
actual"
emissions
test
is
structured,
allowing
a
source
to
subtract
from
its
post­
project
emissions
those
emissions
that
the
unit
could
have
accommodated
during
the
baseline
period
and
that
are
unrelated
to
the
change
(
referred
to
as
the
"
demand
growth
exclusion").
That
is,
the
source
can
emit
up
to
its
current
maximum
capacity
without
triggering
major
NSR
under
the
actual­
to­
projected­
actual
test,
as
long
as
the
increase
is
unrelated
to
the
physical
or
operational
change.
We
are
proposing
that
causation
can
take
any
of
the
following
three
forms:
°
Legal
(
i.
e.,
previously
authorized
through
a
permit
at
the
unchanged
unit);
°
Physical
(
i.
e.,
the
unit
previously
able
to
accommodate
the
increased
productive
capacity
from
demand
at
the
other
parts
of
the
plant
or
off­
site
of
the
plant);
and
°
Economic
(
i.
e.,
same
as
physical
but
considers
how
economically
feasible
it
would
have
been
to
previously
operate
at
higher
levels).

We
expect
that
any
of
the
three
causation
principles
presented
above
would
better
identify
projects
for
which
major
NSR
should
apply
than
did
our
prior
debottlenecking
policies.
Major
NSR
will
continue
to
apply
when
projects
cause
an
emissions
increase
greater
than
the
significance
levels.
Thus,
we
believe
the
proposed
approaches
are
sound
interpretations
of
the
statute
and
strike
a
better
balance
between
Congress'
desire
to
promote
economic
growth
and
the
need
for
environmental
protection
than
does
the
current
approach.

We
are
assessing
whether
environmental
protection
benefits
resulted
from
the
current
debottlenecking
rule
and
policies
that
would
not
occur
under
the
proposed
new
approach.
We
recognize
that
the
proposed
emissions
test
for
debottlenecked
units,
when
final,
could,
theoretically,
result
in
fewer
projects
undergoing
major
NSR
than
would
the
current
rules.
However,
there
are
a
number
of
reasons
why,
in
practice,
this
rule
will
not
impact
many
projects
at
all,
including:

°
Since
we
do
not
require
unchanged
units
to
install
controls,
we
are
only
concerned
with
those
projects
that
would
apply
controls
at
the
changed
unit
under
the
old
rules
that
would
not
apply
the
same
level
of
control
under
the
proposed
debottlenecking
provisions.
Specifically,
the
affected
types
of
projects
are
limited
to
those
that
involve
changes
to
emissions
units
that
themselves
result
in
de
minimis
increases
but
would
have
triggered
NSR
under
only
the
current
rules
(
and
not
the
proposed
rules)
due
to
emissions
increases
at
the
debottlenecked
unit(
s)
at
the
source.
This
is
described
in
more
detail
below
in
the
description
of
how
the
proposed
rule
would
impact
different
scenarios.

°
The
universe
of
emissions
units
that
are
now
potentially
subject
to
the
debottlenecking
provisions
of
our
rules
has
been
reduced
as
a
result
of
recently
finalized
NSR
rule
provisions,
such
as
"
Actuals
PALs"
and
the
"
actual­
to­
projected­
actual"
emissions
test
for
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
7
nonutility
sources.

°
The
current
debottlenecking
emissions
methodology
provides
an
inherent
incentive
to
keep
actual
emissions
high,
since
the
baseline
for
determining
emissions
increases
at
modified
units
is
based
on
a
source's
actual
emissions.
Through
removing
incentives
to
keep
pollution
high,
and
through
removing
barriers
to
emissions­
reducing
changes,
EPA
believes
there
could
be
environmental
benefits
that
result
from
switching
to
the
proposed
debottlecking
emissions
test.
However,
it
is
difficult
to
model
the
behavior
of
individual
sources
in
sufficient
detail
to
quantify
these
benefits,
either
locally
or
nationally.

°
By
implementing
this
proposed
debottlenecking
approach,
facilities
will
no
longer
be
discouraged
from
undertaking
more
energy
efficient
projects
and
lower
emitting
processes
than
in
the
past,
which
will
result
in
decreased
emissions
from
the
changed
emissions
units
per
unit
of
production
while
maintaining
the
debottlenecked
unit
within
its
permitted
allowable
emissions.
On
the
other
hand,
if
a
proposed
project
is
determined
to
be
a
major
modification,
the
source
is
still
required
to
perform
an
air
quality
impact
analysis
(
and
obtain
offsets
if
located
in
a
nonattainment
area).
Some
reviewing
authorities
may
also
require
that
sources
with
projects
that
qualify
as
minor
NSR
perform
modeling
to
ensure
protection
of
the
NAAQS
and
PSD
increments.

Given
the
above
described
conditions
and
factors,
we
examined
the
range
of
scenarios
that
could
occur
as
a
result
of
implementing
the
proposed
debottlenecking
provisions.
There
are
three
principal
emissions
unit
scenarios,
some
of
which
require
a
more
detailed
examination
into
the
range
of
sub­
scenarios.
At
issue
is
whether,
in
any
of
the
scenarios,
controls
could
be
required
for
the
changed
unit
under
the
current
debottlenecking
rules
(
i.
e.,
major
NSR
is
triggered),
but
would
not
be
required
under
the
proposed
rule.
For
simplicity,
we
base
our
conclusions
on
a
legal
causation
test,
since
that
option
appears
to
be
the
easiest
to
implement.
Also,
we
assume
in
each
scenario
that
the
debottlenecked
unit's
emissions
are
higher
than
historical
emissions
levels.
Since
those
units
are
not
physically
or
operationally
changed
by
the
project,
to
increase
their
operation
or
utilization
would
logically
mean
they
would
emit
more
than
when
bottlenecked.

First
principle
scenario
is
where
a
plant
is
making
a
change
at
a
nonemitting
unit
(
e.
g.,
steam
turbine)
and,
in
doing
so,
it
debottlenecks
a
unit
upstream
or
downstream
of
it.
For
nonemitting
units
that
undergo
a
change,
since
they
have
no
emissions,
we
do
not
require
application
of
BACT
or
LAER
at
the
unit.
Thus,
for
changes
at
nonemitting
units,
these
proposed
provisions
do
not
change
the
BACT
outcome.

The
second
principle
scenario
is
where
a
well­
controlled
unit
(
e.
g.,
has
BACT
or
MACT
controls)
is
changed
and
debottlenecks
a
unit
upstream
or
downstream
of
it.
This
scenario
has
a
few
different
tiers
of
analysis
to
it.

°
For
emissions
units
that
have
a
greater
than
significant
emissions
increase
of
a
pollutant
at
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
8
the
changed
unit,
there
is
no
change
in
whether
emissions
controls
would
be
required
at
the
changed
unit
under
the
current
or
proposed
debottlenecking
provisions.

°
For
changed
units
that
have
a
less
than
significant
(
i.
e.,
de
minimis)
emissions
increase
of
a
pollutant,
and
the
debottlenecked
unit
has
an
emission
increase,
but
of
a
different
pollutant,
then
there
is
also
no
difference
in
the
control
requirements
between
the
proposed
and
current
rules.
This
is
because
our
rules
apply
on
a
pollutant­
by­
pollutant
specific
basis;
thus,
BACT
and
LAER
is
only
triggered
at
a
changed
emissions
unit
when
the
the
pollutant
that
has
a
significant
net
emissions
increase
is
emitted
by
that
unit.
For
example,
it
is
common
for
recovery
boilers
at
a
pulp
and
paper
mill
to
be
initially
contructed
with
additional
capacity
to
handle
plant
expansion
projects.
Being
oversized,
they
are
underutilized
and,
therefore,
bottlenecked
until
the
plant
expansion
projects
are
implemented.
When
the
plant
decides
to
install
a
larger
digester
to
increase
black
liquor
production,
the
recovery
boiler
is
debottlenecked
and
will
increase
emissions
due
to
additional
operation.
However,
the
boilers
emit
NOx,
SO2,
PM,
and
CO,
while
the
digester's
blow
tank
emits
a
variety
of
volatile
organic
compounds
(
VOC)
and
perhaps
some
total
reduced
sulfur
(
TRS).
Thus,
the
increased
NOx
emissions
from
the
boiler
would
not
count
toward
triggering
the
major
NSR
threshold
at
the
changed
unit,
since
NOx
is
not
emitted
by
the
changed
unit.

°
For
changed
units
with
a
de
minimis
emissions
increase
of
a
pollutant
and
the
debottlenecked
unit
has
an
emissions
increase
of
that
same
pollutant,
but
the
debottlenecked
unit
does
not
have
a
practically
enforceable
permit.
Therefore,
the
current
debottlenecking
emissions
test
(
i.
e.,
actual­
to­
future­
actual
with
presumed
causation)
and
thus
there
is
no
change
in
whether
emissions
controls
would
be
required
at
the
changed
unit
under
the
current
or
proposed
debottlenecking
provisions.

°
For
changed
units
with
a
de
minimis
emissions
increase
of
a
pollutant
and
the
debottlenecked
unit
has
an
emissions
increase
in
that
same
pollutant,
but
when
applying
the
current
debottlenecking
rules
(
i.
e.,
actual­
to­
future­
actual
with
presumed
causation)
the
collective
emissions
of
the
change
remains
beneath
the
de
minimis
threshold,
then
there
is
no
change
in
whether
emissions
controls
would
be
required
at
the
changed
unit
under
the
current
or
proposed
debottlenecking
provisions.

°
For
changed
units
with
a
de
minimis
emissions
increase
of
a
pollutant
and
the
debottlenecked
unit
has
an
emissions
increase
of
that
same
pollutant.
The
debottlenecked
unit
has
a
practically
enforceable
permit,
but
its
post­
change
emissions
do
not
exceed
its
permit.
Collectively,
the
emissions
from
the
changed
and
unchanged
unit
exceed
the
significance
threshold
for
that
pollutant,
when
applying
the
current
rules
for
calculating
emissions
from
debottlenecked
units.
In
this
case,
the
proposed
rule
changes
would
mean
the
difference
between
triggering
major
NSR
requirements
and,
thus,
requiring
a
BACT
or
LAER
assessment
at
the
changed
emissions
unit.
However,
we
expect
that
in
most
of
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
3
See
40
CFR
51.165(
a)(
3).

9
these
cases,
if
the
changed
unit
is
well­
controlled,
then
an
BACT
or
LAER
review
would
not
result
in
additional
emissions
reductions
since,
for
BACT,
the
cost
effectiveness
of
applying
additional
controls
on
an
already
well­
controlled
process
unit
can
be
very
high.
In
other
words,
the
incremental
amount
of
emissions
that
will
be
reduced
by
applying
additional
controls
is
unreasonable
considering
the
capital
and
operating
cost
of
new
control
equipment.
Thus,
no
additional
controls
would
be
required
in
attainment
areas.
This
is
particularly
true
for
surface
coating
operations,
most
of
which
have
been
required
to
control
to
MACT
levels
and
is
unlikely
that
a
BACT
analysis
would
achieve
higher
control
than
is
required
under
MACT.
Likewise,
if
the
source
has
already
undergone
a
recent
LAER
review,
it
is
difficult
to
imagine
that
controls
would
have
advanced
so
much
in
a
short
period
of
time
that
would
result
in
additional
controls.
Thus,
no
additional
controls
would
be
required
in
nonattainment
areas.

The
third
principle
scenario
is
where
an
uncontrolled,
or
a
poorly
controlled,
emissions
unit
is
changed
and
debottlenecks
a
unit
upstream
or
downstream
of
it.
All
of
the
sub­
scenarios
for
this
scenario
mirror
the
sub­
scenarios
for
a
well­
controlled
unit
undergoing
a
change,
except
for
the
last
one
described
above.
It
is
conceivable
that
if
an
uncontrolled
emissions
unit
experiences
a
physical
change
and
has
a
de
minimis
emissions
increase
of
a
pollutant
and
the
debottlenecked
unit
has
a
"
within
its
permitted"
emissions
increase
of
that
same
pollutant,
and
their
collective
emissions
exceed
the
significance
threshold
for
that
pollutant,
then
a
BACT
or
LAER
assessment
could
result
in
control
requirements
under
the
current
rules,
while
the
proposed
new
rules
would
not
cause
major
NSR
to
be
triggered.
However,
we
strongly
expect
that
many
of
these
changes
to
poorly
controlled
emissions
equipment
will
actually
cause
emissions
to
decrease,
rather
than
increase,
at
the
uncontrolled
unit.
This
is
because
many
of
the
poorly
controlled
units
are
older
and
are
being
replaced
or
modified
with
cleaner
technology,
even
without
applying
controls.
Also,
minor
NSR
requirements
may
require
controls
on
even
de
minimis
emissions
increases.
This
means
that
there
are
very
few
units
that
would
actually
see
a
significant
emissions
increase,
unless
the
debottlenecked
units
experience
a
very
large
emissions
increase.
This
is
certainly
possible,
but
the
debottlenecked
units'
emissions
increase
would
have
to
exceed
the
emissions
decrease
at
the
changed
unit,
plus
the
significance
level,
to
cause
major
NSR
to
be
triggered,
which
is
difficult
to
envision.
Finally,
in
the
few
cases
that
remain
where
major
NSR
is
triggered,
we
expect
that
the
source
will
examine
a
number
of
options
on
how
to
proceed,
but
is
likely
to
avoid
major
NSR
 
e.
g.,
accepting
permit
limits,
cancelling
the
project,
etc.
 
after
weighing
the
additional
production
against
the
cost
of
expensive
controls.

In
the
rare
case
that
a
project
avoids
triggering
major
NSR
under
the
proposed
new
debottlenecking
rules,
but
its
emissions
increase
would
have
triggered
major
NSR
under
the
current
rules,
the
project
could
cause
an
emissions
increase
in
nonattainment
areas
as
a
result
of
our
emissions
offset
regulations.
3
However,
CAA
§
173(
a)(
1)(
A)
requires
States
to
monitor
new
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
10
sources
and
minor
source
growth
through
a
nonattainment
area
plan
that
achieves
reasonable
further
progress
toward
attaining
the
NAAQS
based
on
allowable
emissions;
thus,
States
must
account
for
these
new
emissions
in
their
inventory
and
attainment
demonstrations,
even
when
major
NSR
is
not
triggered.
Similarly,
the
CAA
§
163
requires
States
to
address
overall
emissions
in
attainment
areas
by
monitoring
and
tracking
PSD
increment
consumption
based
on
allowable
emissions
of
specified
pollutants.
Thus,
the
CAA
authorizes
States
to
address
and
mitigate
allowable
emissions
increases
through
SIP
planning,
and
we
expect
that
States
will
use
this
mechanism
to
properly
mitigate
increased
emissions
caused
by
debottlenecking
at
sources.
We
are
asking
for
comment
on
how
our
proposed
provisions
will
affect
air
quality
in
light
of
our
emissions
offset
requirements
and
reconcilation
of
attainment
demonstrations.

For
these
reasons,
we
expect
the
environmental
impact
of
requiring
our
proposed
emissions
test
for
debottlenecked
units
will
be
negligible.
As
discussed
in
the
scenarios
above,
we
expect
the
vast
majority
of
sources
will
be
unaffected
by
this
change.

B.
Aggregation
The
term
"
aggregation"
is
relevant
to
Step
1
of
the
NSR
applicability
test,
and
describes
the
process
of
grouping
together
multiple
projects
(
i.
e.,
physical
changes
or
changes
in
the
method
of
operation)
and
summing
their
emissions
changes
for
purposes
of
determining
whether
a
significant
emissions
increase
occurs
at
the
major
stationary
source.
Specifically,
when
undertaking
multiple
projects,
the
source
must
consider
whether
NSR
applicability
should
be
determined
collectively
or
whether
the
emissions
from
each
of
the
projects
should
separately
undergo
a
Step
1
analysis.

We
are
proposing
to
add
our
aggregation
policy
to
our
NSR
regulations
to
achieve
greater
national
consistency
and
provide
further
clarity
in
aggregation
determinations.
This
will
codify
the
provisions
of
our
existing
policy
and
provides
specific
circumstances
where
emissions
should
be
aggregated
for
purposes
of
NSR
applicability.
EPA
proposes
to
revise
the
regulations
to
state
that
a
source
must
aggregate
emissions
from
projects
that
are
technically
or
economically
dependent.
This
will
reduce
the
uncertainty
that
industry
and
regulators
have
faced
in
the
past
with
the
lack
of
documentation
of
our
aggregation
policy.

The
impact
of
this
rule
proposal
is
environmentally
neutral
because
we
are
simply
proposing
to
codify
our
existing
aggregation
policy.

C.
Project
Netting
This
proposal
proposes
to
revise
and
change
the
current
rules
with
respect
to
projects
that
involve
both
increases
and
decreases
in
emissions.
We
are
concerned
with
inconsistent
implementation
of
our
past
policy
that
considers
only
emissions
increases
in
Step
1
of
the
NSR
applicability
test.
We
are
proposing
that
all
emissions
changes
(
i.
e.,
both
increases
and
decreases)
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
11
that
occur
within
the
scope
of
a
project
be
counted
in
Step
1
of
the
NSR
applicability
test.
This
proposed
change
would
streamline
permitting
by
moving
the
consideration
of
creditable
decreases
associated
with
the
project
from
Step
2
of
the
NSR
test
(
i.
e.,
the
contemporaneous
netting
step)
to
Step
1.

As
with
our
debottlenecking
analysis,
we
expect
the
emission
consequences
of
this
proposed
action
are
best
analyzed
by
reviewing
the
range
of
scenarios.
There
are
four
scenarios
to
consider.
At
issue
is
whether,
for
any
of
the
scenarios,
major
NSR
(
and,
thus,
a
control
technology
review)
would
be
triggered
under
the
current
rules
that
do
not
allow
for
project
netting,
but
would
not
be
triggered
under
the
proposed
rule
provisions.
Each
of
these
scenarios
is
shown
in
Table
1
for
illustrative
purposes
and
described
below.

Scenario
Rule
Version
projected
increase
projected
decrease
contemp.
increase
contemp.
decrease
Step
1
Step
2
Major
NSR?

1
old
rules
60
30
60
30
No
proposed
rules
60
30
30
N/
A
No
2
old
rules
60
30
+
5
60
25
No
proposed
rules
60
30
5
35
N/
A
No
3
old
rules
60
15
30
60
45
Yes
proposed
rules
60
30
15
30
N/
A
No
4
old
rules
60
15
30
60
45
Yes
proposed
rules
60
30
15
30
N/
A
No
Table
1.
Project
Netting
Scenarios
Under
the
first
scenario,
the
source
has
an
emissions
increase
(
60
tpy)
and
decrease
(
30tpy)
associated
with
the
project,
and
has
no
other
increases
or
decreases
that
have
occurred
within
a
contemporaneous
period
(
i.
e.,
5
years).
As
the
table
shows,
allowing
for
project
netting
in
this
scenario
does
not
change
the
outcome
of
the
overall
NSR
applicability
test.
It
simply
means
that
the
decreases
are
counted
in
Step
1
rather
than
Step
2.
Since
there
are
no
other
increases
or
decreases
to
consider,
the
outcome
of
NSR
applicability
is
the
same
under
either
rule.
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
12
Under
the
second
scenario,
the
source
has
contemporaneous
decreases,
but
not
increases.
As
in
the
first
scenario,
counting
the
decrease
in
the
Step
1
rather
than
Step
2
makes
no
difference
in
the
outcome
of
NSR
applicability,
since
there
are
only
other
decreases
(
in
this
case,
5
tpy).

Under
the
third
scenario,
the
source
has
contemporaneous
decreases
and
increases.
In
this
case,
it
depends
on
whether
a
source
would
have
netted
out
under
the
current
rules.
If
netting
under
the
current
rules
(
i.
e.,
only
contemporaneous
netting)
would
cause
the
source
to
not
trigger
major
NSR,
then
there
is
no
difference
in
whether
project
netting
is
allowed.
However,
if
the
netting
under
the
current
rules
would
have
still
triggered
major
NSR,
then
it
is
possible
that
the
new
provisions
for
project
netting
would
allow
the
source
to
not
trigger
major
NSR.
For
example,
a
source
expects
a
project
to
increase
emissions
by
60
tpy
at
an
emissions
unit
and
decrease
emissions
by
30
tpy
at
another
unit
at
the
source.
The
source
has
contemporaneous
increases
and
decreases
totalling
15
tpy.
Under
the
current
rules,
project
netting
is
not
allowed,
so
a
Step
1
analysis
would
reveal
a
60
tpy
emissions
increase,
and
the
Step
2
analysis
would
equal
45
tpy
(
i.
e.,
60
­
30
+
15),
which
is
a
significant
net
emissions
increase,
thereby
triggering
major
NSR.
However,
under
the
proposed
rule
provisions,
project
netting
will
be
allowed,
which
means
that
the
Step
1
analysis
would
reveal
a
30
tpy
increase
(
i.
e.,
60
­
30),
which
is
not
a
significant
emissions
increase
and
therefore
major
NSR
is
not
triggered
(
i.
e.,
no
requirement
to
proceed
to
Step
2).
However,
in
these
cases,
rather
than
assuming
that
NSR
will
be
triggered,
it
is
important
review
the
types
of
project
and
anticipate
whether
the
owner
would
still
proceed
in
doing
the
project
in
light
of
major
NSR
being
triggered.
For
example,
if
the
source
is
removing
older,
rundown
equipment
and
replacing
it
with
new
highly
efficient
unit
that
pollutes
less
even
without
control,
then
there
is
a
potential
emissions
benefit
under
proposed
rule.

Under
the
fourth
scenario,
the
source
has
contemporaneous
increases
only.
This
scenario
mirrors
closely
the
third
scenario
 
i.
e.,
there
may
be
cases,
depending
on
the
amount
of
contemporaneous
emission
increase
and
the
project
decrease,
that
NSR
could
be
triggered
under
the
current
rules
and
not
triggered
under
the
proposed
rules.
However,
if
this
happens,
it
is
likely
that
a
source
could
weigh
the
benefits
of
the
project
and
may
opt
to
avoid
NSR
by
either
taking
a
permit
limit
or
cancelling
the
project.
In
cases
where
these
NSR­
triggered
projects
would
cause
a
net
environmental
benefit,
this
would
be
an
unfortunate
outcome
of
not
proceeded
with
finalizing
the
proposed
rules.
And,
for
well­
controlled
units,
it
is
likely
that
an
additional
BACT
or
LAER
review
would
result
in
no
added
control.

We
do
not
have
enough
information
to
determine
whether,
or
how
much
of,
an
emissions
increase
could
result
from
allowing
for
project
netting
in
Step
1
of
the
NSR
applicability
test.
For
the
reasons
cited
in
this
section,
many
projects
will
be
unaffected
by
this
change.
However,
in
cases
where
allowing
project
netting
theoretically
could
cause
a
permit
to
not
trigger
major
NSR,
it
is
possible,
and
perhaps
very
likely,
that
owner
or
operator
of
the
source
would
choose
to
forego
the
project
simply
to
avoid
the
expense
and
time
necessary
with
major
NSR.
Also,
it
should
be
noted
that
under
no
scenario
would
the
proposed
project
netting
rule
change
cause
a
new
increase
to
the
environment
that
exceeds
the
significant
emissions
threshold
 
i.
e.,
the
DRAFT:
Deliberative
and
Privileged;
do
not
cite,
quote,
or
distribute
13
triggering
of
major
NSR
would
be
entirely
a
result
of
past
changes
at
the
source
that
previously
avoided
major
NSR.

IV.
PRELIMINARY
CONCLUSIONS
This
document
illustrates
the
serious
limitations
on
EPA's
current
ability
to
quantify
the
environmental
impact
of
these
proposed
regulations.
Nonetheless,
we
expect
that
these
reforms
will
have
little,
if
any,
environmentally
impact
compared
to
the
current
program.
Through
our
rule
proposal,
we
will
seek
comment
and
data
to
enable
a
more
quantitative
assessment
of
the
rule
impacts.
We
will
update
this
environmental
analysis
to
incorporate
new
information
and
our
final
conclusions
upon
finalizing
the
rule.
