250
not
provide
for
exceptions
from
emissions
standards
basedare
not
relying
on
de
minimis
principles
where
a
MACT
floor
exists.'"
(
See
National
Lime
at
640.)
We
recently
reaffirmed
our
position
on
the
unavailability
of
de
minimis
exemptions
from
the
MACT
floor
in
the
final
rule
regulating
organic
liquids
distribution.
(
See
69
FR
5038,
5048­
49
(
February
3,
2004).)

We
see
no
reason
to
revisit
this
fundamental
issue.

Section
112
of
the
CAA
is
replete
with
careful
definitions
of
mass­
or
effect­
based
limitation
on
regulation,

indicating
that
Congress
has
already
defined
what
amounts
of
HAP
emissions
are
too
small
to
warrant
MACT
standards
or
other
controls
under
section
112.
The
requirement
to
adopt
MACT
emission
limitations,
for
example,
applies
without
exception
to
"
each
category
or
subcategory
of
major
sources...
of
[
HAP]."
(
See
CAA
section
112(
d)(
1).)
For
sources
below
the
major
source
threshold,
however,
we
have
discretion
to
require
"
generally
available
control
technologies
or
management
practices."
(
See
CAA
section
112(
d)(
5).)
Congress
has,
thus,
itself
defined
which
sources'
emissions
are
small
enough
not
to
warrant
mandatory
MACT
standards.

Congress
likewise
defined
several
MACT
exceptions
applicable
where
emissions
have
de
minimis
health
effects.

Section
112(
d)(
4)
of
the
CAA
allows
us
to
establish
251
standards
less
stringent
than
MACT
for
HAP
with
an
established
health
threshold,
so
long
as
we
set
a
standard
below
the
health
threshold
with
"
an
ample
margin
of
safety."

Section
112(
b)(
3)(
C)
of
the
CAA
directs
us
to
delist
HAP
 
precluding
section
112(
d)
MACT
standards
 
if
we
determine
that
"
there
is
adequate
data
on
the
health
and
environmental
effects
of
the
substance
to
determine
that
emissions,

ambient
concentrations,
bioaccumulation
or
deposition
of
the
substance
may
not
reasonably
be
anticipated
to
cause
any
adverse
effects
to
the
[
sic]
human
health
or
adverse
environmental
effects."
Section
112(
c)(
9)(
B)(
i)
of
the
CAA
(
discussed
further
below)
lets
us
delete
source
categories
and
subcategories
from
the
category
list
 
the
consequence
again
being
no
MACT
control
 
if
we
determine
that,
for
emissions
of
carcinogenic
HAP,
"
no
source
in
the
category...
emits
such
[
HAP]
in
quantities
which
may
cause
a
lifetime
risk
of
cancer
greater
than
one
in
one
million
to
the
individual
in
the
population
who
is
most
exposed
to
emissions
of
such
pollutant
from
the
source."
For
noncarcinogens,
we
may
delete
source
categories
and
subcategories
if
we
determine
that
"
emissions
from
no
source
in
the
category
or
subcategory...
exceed
a
level
which
is
adequate
to
protect
public
health
with
an
ample
margin
of
safety
and
no
adverse
environmental
effect
will
result
from
emissions
from
any
source."
(
See
CAA
section
252
112(
c)(
9)(
B)(
ii).)
Moreover,
in
defining
which
source
modifications
trigger
additional
regulatory
standards,
CAA
section
112(
g)(
1)(
A)
mentions
a
"
greater
than
de
minimis
increase
in
actual
emission
of
a
[
HAP]."
This
shows
that
Congress
knew
how
to
use
the
de
minimis
concept
when
it
considered
it
appropriate
in
section
112,
and
the
fact
that
Congress
did
not
use
it
in
section
112(
d)(
3)
supports
our
 
and
the
D.
C.
Circuit's
 
conclusion
that
it
is
unavailable
to
support
an
exception
to
a
MACT
floor
that
is
not
otherwise
authorized
under
section
112.

We
do
not
find
persuasive
the
proposition
that
the
overall
purpose
of
CAA
section
112
is
protecting
human
health
and
the
environment,
and
that,
therefore,
as
long
as
this
general
purpose
is
met,
we
may
fashion
de
minimis
exceptions
from
MACT
beyond
those
allowed
under
section
112.

First,
this
position
appears
to
assume
that
the
issue
is
to
be
drawn
on
a
clean
slate,
while
the
D.
C.
Circuit
has
affirmed
our
view
that
section
112(
d)(
3)
provides
no
discretion
to
use
a
de
minimis
rationale
to
avoid
MACT.

Second,
the
commenter
appears
to
give
prominence
to
an
overarching
statutory
goal
over
the
specific
language
of
the
statutory
provisions
themselves,
in
assessing
whether
those
provisions
are
extraordinarily
rigid
regarding
EPA's
otherwise­
inherent
de
minimis
authority;
the
logical
extension
of
such
an
approach
would
be
to
find
that
no
253
single
provision
in
the
CAA
could
restrict
our
de
minimis
authority,
in
light
of
the
CAA's
over­
arching
purpose
"
to
promote
the
public
health
and
welfare."
(
See
CAA
section
101(
b)(
1).)
Third,
the
commenter
does
not
present
any
persuasive
statutory
arguments
to
overcome
those
that
we
presented
to
the
court
 
and
which
the
court
affirmed
 
in
National
Lime.
Fourth,
we
are
unable
to
discern
the
basis
for
the
commenter's
suggestion
that
we
have
in
fact
been
relying
on
de
minimis
authority
in
the
MACT
program
for
several
years
in
establishing
applicability
thresholds,
and
we
are
not
aware
of
any
instance
in
which
we
have
explicitly
justified
an
exception
from
an
applicable
MACT
floor
based
on
a
de
minimis
rationale
that
would
be,
like
the
commenter's
requested
exemption,
in
contravention
of
the
Court's
ruling
in
National
Lime.
Fifth,
notwithstanding
any
asserted
overall
risk­
based
goal
of
CAA
section
112,
CAA
section
112(
d)(
2)
repeatedly
directs
us
to
eliminate
emissions,
where
feasible,
and
CAA
section
112(
d)(
6)
imposes
an
ongoing
obligation
for
us
to
review
and
revise
standards
as
necessary
to
account
for
developments
in
technology,
and
neither
of
these
specific
goals
is
restricted
to
situations
where
health
targets
have
not
yet
been
reached.
Finally,
to
the
extent
the
commenters
believe
such
a
de
minimis
exemption
is
justified
by
the
wish
to
reduce
the
costs
of
the
PCWP
rule
as
proposed,
we
are
not
free
to
grant
a
de
254
minimis
exemption
to
account
for
costs.
Congress
already
took
cost
into
account
in
section
112(
d),
relying
on
prior
business
judgments
by
the
best
performing
sources
to
substitute
for
the
judgment
of
the
rest
of
the
PCWP
industry,
therefore,
denying
us
the
leeway
to
consider
costs
as
a
factor
to
modify
the
MACT
floor.
Only
in
considering
more
stringent
beyond
floor
standards
may
we
consider
costs
in
the
MACT
context.
Therefore,
we
do
not
feel
it
is
appropriate
or
necessary
to
revisit
our
and
the
D.
C.

Circuit's
prior
conclusions
regarding
the
availability
of
the
de
minimis
principle
in
today's
final
PCWP
rule.

for
today's
action,
and
therefore,
do
not
need
to
respond
to
these
comments.

Section
112(
c)(
9)
of
the
CAA
Comment:
Two
commenters
opposed
using
subcategorization
as
a
mechanism
to
exempt
facilities.
One
of
the
commenters
stated
that
subcategorization
is
a
tool
that
should
be
used
in
the
standard
setting
process,
and
using
it
to
exempt
facilities
would
have
a
detrimental
effect
on
the
stringency
of
the
MACT
floor
and
would
generally
degrade
the
standard.

According
to
the
commenter,
the
two­
step
subcategorization
proposal
is
inconsistent
with
how
subcategorization
has
been
done
in
numerous
previous
NESHAP.

The
other
commenter
argued
that
EPA's
subcategorization
theories
are
unlawful.
According
to
the
commenter,
CAA
255
section
112(
c)(
9)
does
not
authorize
EPA
to
separate
identical
pollution
sources
into
subcategories
that
are
regulated
differently
to
weed
out
low­
risk
facilities
or
reduce
the
scope/
cost
of
the
standard.
The
commenter
stated
that
subcategories
based
solely
on
risk
do
not
bear
a
reasonable
relationship
to
Congress'
technology­
based
approach
or
the
statutory
structure
and
purposes
of
CAA
section
112,
and
are
not
authorized
by
the
CAA.
According
to
the
commenter,
categories
and
subcategories
are
required
to
be
consistent
with
the
categories
of
stationary
sources
in
CAA
section
111.
The
commenter
was
not
aware
of
any
instance
in
which
EPA
has
established
categories
or
subcategories
based
on
risk.
The
commenter
stated
that
EPA
routinely
defines
subcategories
based
on
equipment
characteristics
(
e.
g.,
technical
differences
in
emissions
characteristics,
processes,
control
device
applicability,
or
opportunities
for
P2).
According
to
the
commenter,
EPA
has
not
offered
any
explaination
for
why
reinterpreting
the
statute
to
ignore
nearly
12
years
of
settled
practices
and
expectations
under
the
MACT
program
is
reasonable,
nor
why
reducing
the
applicability
of
HAP
emission
standards
serves
Congress's
goals
in
enacting
the
1990
CAA
Amendments.

The
commenter
noted
that
EPA's
discussion
of
the
risk­
based
exemptions
was
contained
in
a
preamble
section
entitled,
"
Can
We
Achieve
the
Goals
of
the
Proposed
Rule
in
256
a
Less
Costly
Manner,"
which
strongly
suggests
that
EPA's
motivation
for
considering
these
risk­
based
approaches
is
consideration
of
cost.
The
commenter
cited
prior
EPA
documentation
and
stated
that
EPA
in
the
past
has
rejected
the
notion
that
cost
should
influence
MACT
determination,

and
this
prior,
consistently
applied
interpretation
better
serves
the
purposes
of
CAA
section
112.
The
commenter
argued
that
subcategorizing
to
set
a
no­
control
MACT
floor
is
the
same
as
refusing
to
set
a
MACT
standard
because
the
benefits
would
be
negligible,
which
is
unlawful.

The
commenter
also
stated
that
CAA
section
112(
c)(
9)(
B)(
i)
does
not
authorize
EPA
to
delist
subcategories.
According
to
the
commenter,
section
112(
c)(
9)(
B)
contains
two
subsections:
subsection
(
i)

refers
only
to
categories,
and
subsection
(
ii)
refers
to
both
categories
and
subcategories.
The
commenter
argued
that
the
absence
of
the
term
"
subcategories"
in
section
112(
c)(
9)(
B)(
i)
indicates
a
Congressional
choice
not
to
permit
the
Administrator
to
delist
subcategories
of
sources
under
section
112(
c)(
9)(
B).
The
commenter
stated
that
this
is
consistent
with
Congress'
decision
to
require
a
higher
standard
to
delist
categories
that
emit
carcinogens.

According
to
the
commenter,
the
section
112(
c)(
9)(
B)(
ii)

requirement
of
less
than
one
in
a
million
lifetime
cancer
risk
for
the
most
exposed
individual
is
a
higher
and
more
257
specific
standard
than
the
standard
for
other
HAP.

To
the
contrary,
two
commenters
stated
that
EPA
has
ample
authority
under
CAA
sections
112(
c)(
1)
and
112(
c)(
9)

to
create
and
delist
low­
risk
categories
or
subcategories.

According
to
the
commenters,
section
112(
c)(
1)
provides
the
Administrator
with
significant
flexibility
to
create
categories
and
subcategories
as
needed
to
implement
CAA
section
112.
One
commenter
stated
that
there
is
nothing
in
the
statute
that
limits
the
criteria
the
Administrator
can
use
in
establishing
categories
and
subcategories.
The
commenter
added
that
there
is
also
nothing
in
the
history
of
EPA's
interpretation
of
section
112(
c)
that
precludes
subcategorization
based
on
risk.
In
addition,
EPA
has
stated
that
emission
characteristics
are
factors
to
be
considered
when
defining
categories.

The
commenter
stated
that
application
of
statutory
authority
to
exclude
sources
from
regulation
under
section
112(
d)(
3)
is
also
supported
by
relevant
case
law,
e.
g.,
in
the
Vinyl
Chloride
case.
(
NRDC
v.
EPA,
824
F.
2D
1126
(
D.
C.

Cir.
1987))
According
to
the
commenter,
the
court
in
that
case
established
a
range
of
acceptable
levels
of
risk
in
establishing
limits
under
prior
language
in
section
112,
and
the
establishment
of
an
acceptable
level
of
risk
could
be
used
to
create
a
low­
risk
subcategory
that
could
be
delisted.
The
commenter
stated
that
technological
or
258
operational
differences
among
sources
may
also
help
discriminate
between
low­
risk
and
high­
risk
sources.
The
commenter
stated
that
effective
use
of
section
112(
c)(
1)

authority
to
create
risk­
based
subcategories
would
significantly
improve
the
cost­
effectiveness
of
the
section
112
program
without
undermining
its
role
in
protecting
public
health
and
the
environment.

Both
commenters
noted
that
CAA
section
112(
c)(
9)(
B)

provides
EPA
with
broad
authority
to
remove
from
MACT
applicability
those
categories
and
subcategories
of
facilities
whose
HAP
emissions
are
sufficiently
low
as
to
demonstrate
a
cancer
risk
less
than
one
in
a
million
to
the
most
exposed
individual
in
the
population
(
for
non­
threshold
carcinogens)
and
no
adverse
environmental
or
public
health
effect
(
for
threshold
HAP).
(
The
commenter
asserted
that
Congress
used
the
terms
category
and
subcategory
interchangeably,
indicating
that
either
one
can
be
delisted.)
One
commenter
suggested
that
sources
able
to
demonstrate
a
basis
for
inclusion
in
the
delisted
category
on
a
case­
by­
case
basis
would
then
be
exempted
from
the
MACT,
subject
to
possible
federally­
enforceable
conditions
designed
by
EPA.
The
commenter
stated
that
the
new
category
could
include
the
following:
all
low­
risk
facilities,

facilities
producing
wood
products
found
to
pose
no
expected
risk
to
human
health
(
i.
e.,
fiberboard,
medium
density
259
fiberboard
and
plywood),
facilities
with
acrolein
emissions
below
a
certain
threshold,
or
facilities
selected
on
the
basis
of
some
other
risk
criterion.
The
commenter
suggested
that
the
low­
risk
category
be
included
in
the
final
rule
and
delisted
within
6
months
following
publication
of
the
final
rule.
The
delisting
notices
would
designate
health
benchmarks
and
facilities
would
be
required
to
submit
evidence
(
e.
g.,
tiered
dispersion
modeling)
demonstrating
that
their
emissions
result
in
exposures
that
fall
below
the
benchmarks.
Following
delisting
of
the
category,
an
affected
source
could
apply
to
EPA
for
a
determination
that
it
qualifies
for
inclusion
in
the
low­
risk
category.
After
evaluating
the
source's
petition,
EPA
would
issue
a
written
determination
of
applicability
based
on
the
petition
that
would
be
binding
on
the
permitting
authority
(
unless
the
petition
was
found
to
contain
significant
errors
or
omissions)
and
appealable
by
the
affected
source
or
interested
parties.
The
EPA
could
require
all
facilities
that
qualify
for
inclusion
in
the
delisted
category
to
comply
with
federally­
enforceable
conditions,
similar
to
the
conditions
established
in
permits
for
synthetic
minor
sources
(
e.
g.,
limits
on
potential
to
emit,
production
limits).

The
commenter
also
responded
to
objections
regarding
the
subcategorization
and
delisting
of
low­
risk
facilities.
The
260
commenter
stated
that
the
contrasting
of
the
terms
category
and
subcategory
offered
a
distinction
that
in
no
way
limited
EPA's
authority
to
delist
low­
risk
facilities.
According
to
the
commenter,
the
argument
that
EPA
cannot
create
subcategories
based
on
risk
is
contradicted
by
the
statutory
language,
which
expressly
states
that
the
categories
and
subcategories
EPA
creates
under
CAA
section
112
need
not
match
those
created
under
CAA
section
111.
Furthermore,

prior
EPA
statements
do
nothing
to
detract
from
EPA's
broad
discretion
to
establish
categories
and
subcategories.
The
subcategorization
factors
previously
discussed
by
EPA
justify
subcategorization
based
on
risk.
The
authority
cited
by
one
commenter
does
not
establish
that
EPA's
discretion
to
alter
subcategorization
is
limited
in
any
way,

and
even
if
it
were,
EPA
is
not
bound
by
any
prior
position.

The
arguments
that
EPA
may
not
delist
subcategories
for
carcinogens
(
or
sources
emitting
carcinogens)
rest
on
a
formalistic
distinction
that
EPA
previously
has
rejected
as
meaningless,
and
that,
at
any
rate,
can
be
remedied
with
a
simple
recasting
of
a
subcategory
as
a
category.
The
commenter
stated
that
doing
so
is
undisputedly
within
EPA's
authority.

Three
commenters
addressed
the
issue
of
subcategorizing
PCWP
facilities
based
on
characteristics
other
than
risk.

One
commenter
stated
that
the
only
option
that
appears
261
consistent
with
the
CAA,
does
not
create
excessive
work
for
State
and
local
agencies,
and
may
be
able
to
be
based
on
science,
is
the
subcategorization
and
delisting
approach.

However,
the
commenter
added
that
the
subcategories
should
be
based
on
equipment
or
fuel
use,
not
risk.
The
commenter
stated
that
a
subcategory
based
on
site­
specific
risk
creates
a
circular
definition
and
does
not
make
sense.
The
commenter
also
stated
that
subcategory
delisting
should
occur
before
the
compliance
date
so
that
facilities
do
not
put
off
compliance
in
the
hope
or
anticipation
of
delisting.

The
second
commenter
stated
that
EPA
requested
comment
on
the
establishment
of
PCWP
subcategories
ostensibly
based
on
physical
and
operational
characteristics,
but
in
reality
based
on
risk.
According
to
the
commenter,
this
indirect
approach
is
just
a
variation
on
the
approach
(
direct
reliance
on
risk)
that
it
claims
EPA
itself
notes
would
disrupt
and
weaken
establishment
of
MACT
floors,
and
is
accordingly
unlawful.
The
commenter
stated
that,
even
if
these
approaches
were
lawful,
to
the
extent
that
EPA's
proposal
could
be
read
to
suggest
that
facilities
could
be
allowed
to
become
part
of
the
allegedly
low­
risk
subcategory
in
the
future
without
additional
EPA
rulemaking,
this
too
would
be
unlawful.
According
to
the
commenter,
CAA
section
112(
c)(
9)
provides
the
EPA
Administrator
alone
the
authority
to
make
delisting
determinations,
and
such
authority
may
not
262
be
delegated
to
other
government
authorities
or
private
parties.
The
commenter
stated
that
EPA's
proposal
suggests
an
approach
entirely
backward
from
the
statute
 
allowing
sources
to
demonstrate
after­
the­
fact
that
they
belong
in
a
subcategory
that
has
been
delisted
under
section
112(
c)(
9),

when
the
statute
requires
that
EPA
determine
that
no
source
in
the
category
emits
cancer­
causing
HAP
above
specified
levels,
or
that
no
source
in
the
category
or
subcategory
emit
non­
carcinogenic
HAP
above
specified
levels,
by
the
time
EPA
establishes
the
standard.
The
commenter
stated
that
EPA
has
provided
no
explanation
of
how
the
suggested
approaches
would
be
lawful
or
workable.

The
third
commenter
indicated
that
low
risk
is
an
adequate
and
appropriate
criterion
for
categorization.
The
commenter
disagreed
that
EPA
should
create
and
delist
categories
on
a
technology
basis
when
the
intent
is
delisting
of
low­
risk
facilities.
The
commenter
believed
that
seeking
a
technology­
based
surrogate
for
risk
is
unnecessary
within
the
statutory
framework.
The
commenter
noted
that
the
Congressional
intent
was
"
to
avoid
regulatory
costs
which
would
be
without
public
health
benefit."
(
S.

Rep.
No.
228,
101st
Cong.,
1st.
Sess.
175­
6
(
1990))

Nevertheless,
the
commenter
described
some
technology­
based
criteria
that
they
believed
could
be
used
to
develop
lowrisk
groups
of
PCWP
facilities.
263
Four
commenters
addressed
the
impact
that
creation
of
a
low­
risk
subcategory
under
CAA
section
112(
c)(
9)
could
have
on
the
establishment
of
MACT
floors
for
the
PCWP
category.

Two
commenters
argued
that
such
subcategorization
would
have
a
negative
effect.
One
commenter
stated
that
this
situation
provided
a
valid
reason
for
EPA
not
to
mix
risk­
based
and
technology­
based
standards
development.
The
commenter
added
that
EPA
also
did
not
address
how
the
"
once
in,
always
in"

policy
would
apply
in
such
a
situation.
The
other
commenter
stated
that
this
situation
was
another
compelling
reason
why
the
suggested
section
112(
c)(
9)
subcategorization
approach
was
unlawful
and
arbitrary.
The
commenter
stated
that
the
flaw
was
so
obvious,
inherent,
and
contrary
to
the
MACT
floor
provisions
of
CAA
section
112
and
its
legislative
history,
that
it
proves
the
undoing
of
the
suggested
section
112(
c)(
9)
exemption.
According
to
the
commenter,
EPA
cannot
simultaneously
exercise
its
source
category
delisting
authority
consistent
with
section
112(
c)(
9),
establish
appropriate
MACT
floors
under
CAA
section
112(
d),
and
establish
subcategory
exemptions
in
the
manner
suggested
by
EPA,
because
the
latter
approach
contravenes
both
section
112(
c)(
9)
and
the
section
112(
d)
floor­
setting
process.
The
commenter
stated
that
CAA
section
112'
s
major
source
thresholds
and
statutory
deadlines
make
clear
that
sources
meeting
MACT
by
the
time
EPA
is
required
to
issue
MACT
264
standards
must
install
MACT
controls
and
may
not
subsequently
throw
them
off
or
be
relieved
from
meeting
the
MACT­
level
standards.
While
the
CAA
section
112(
f)
residual
risk
process
allows
EPA
to
establish
more
stringent
emissions
standards,
there
is
nothing
in
the
CAA
that
suggests
EPA
possesses
authority
to
relax
promulgated
MACT
standards.

The
third
commenter
indicated
that
dilution
of
the
MACT
floor
would
not
occur
if
low­
risk
category
delisting
occurred
as
follows:
(
1)
propose
low­
risk
category
with
final
PCWP
rule,
(
2)
promulgate
low­
risk
category
6
months
after
proposal,
and
(
3)
delist
facilities
prior
to
MACT
compliance
deadline.
If
EPA
issued
the
final
PCWP
rule
 
thereby
setting
the
MACT
floor
 
before
it
allowed
affected
sources
to
apply
for
inclusion
in
the
low­
risk
category
to
be
delisted,
then
every
affected
source
would
be
considered
in
the
establishment
of
the
MACT
floor.
Thus,
as
a
result
of
this
timing,
the
MACT
floor
could
not
be
diluted
because
no
sources
would
be
exempted
from
MACT
before
the
MACT
floor
is
set.

The
fourth
commenter
believed
that
a
MACT
floor
reevaluation
would
be
appropriate
and
would
further
ensure
that
only
facilities
posing
significant
risk
are
required
to
install
expensive
controls.

Response:
We
feel
that
establishing
a
low­
risk
PCWP
265
subcategory
under
CAA
section
112(
c)(
1)
and
deleting
that
subcategory
under
CAA
section
112(
c)(
9)
best
balances
Congress'
dual
concerns
that
categories
and
subcategories
of
major
sources
of
HAP
be
subject
to
technology­
based
(
and
possible
future
risk­
based)
emission
standards,
but
that
undue
burdens
not
be
placed
on
groups
of
sources
within
the
PCWP
source
category
whose
HAP
emissions
are
demonstrated
to
present
little
risk
to
public
health
and
the
environment.

We
do
not
contend
that
the
CAA
specifically
directs
us
to
establish
categories
and
subcategories
of
HAP
sources
based
on
risk,
and
we
recognize
that,
at
the
time
of
the
1990
CAA
Amendments,
Congress
may
have
assumed
that
we
would
generally
base
categories
and
subcategories
on
the
traditional
technological,
process,
output,
and
product
factors
that
had
been
considered
under
CAA
section
111.

However,
when
properly
considered,
it
becomes
apparent
that
Congress
did
not
intend
the
unduly
restrictive
 
and
consequently
over­
regulatory
 
reading
of
the
CAA
that
some
commenters
urge
regarding
low­
risk
PCWP
facilities.

Numerous
CAA
section
112
provisions
evidence
Congress'

intent
that
we
be
able
to
find
that
sources,
such
as
those
in
the
PCWP
category
whose
HAP
emissions
are
below
identified
risk
levels,
should
not
necessarily
be
subject
to
MACT.
These
provisions,
together
with
other
indications
of
Congressional
intent
regarding
the
goals
of
section
112,
266
must
all
be
considered
in
determining
whether
we
may
base
a
PCWP
subcategory
on
risk
and
delist
that
group
of
sources,

without
requiring
additional
HAP
regulation
that
would
be
redundant
for
purposes
of
meeting
Congress'
risk­
based
goals.

While
it
is
true
that
CAA
section
112(
c)(
1)
provides
that
"[
t]
o
the
extent
practicable,
the
categories
and
subcategories
listed
under
this
subsection
shall
be
consistent
with
the
list
of
source
categories
established
pursuant
to
section
111
and
part
C[,]"
the
provision
also
states
that
"[
n]
othing
in
the
preceding
sentence
limits
the
Administrator's
authority
to
establish
subcategories
under
this
section,
as
appropriate."
Therefore,
by
its
plain
terms,
section
112(
c)(
1)
does
not
preclude
basing
subcategories
on
criteria
other
than
those
traditionally
used
under
section
111
before
1990,
or
those
used
after
1990
for
sections
111
and
112.
Moreover,
while
after
1990
we
have
principally
used
the
traditional
criteria
to
define
categories
and
subcategories,
such
use
in
general
does
not
restrict
how
we
may
define
a
subcategory
in
a
specific
case,

"
as
appropriate,"
since
each
HAP­
emitting
industry
presents
its
own
unique
situation
and
factors
to
be
considered.

(
See,
e.
g.,
Sierra
Club
v.
EPA,
D.
C.
Cir.
No.
02­
1253,
2004
U.
S.
App.
LEXIS
348
(
decided
Jan.
13,
2004).)

Even
assuming
for
argument
that
the
language
of
section
267
112(
c)(
1)
may
initially
appear
to
restrict
our
authority
to
define
subcategories,
section
112(
c)(
1)
cannot
be
read
in
isolation.
A
broad
review
of
the
entire
text,
structure,

and
purpose
of
the
statute,
as
well
as
Congressional
intent
shows
that,
applied
within
the
context
of
CAA
section
112(
c)(
9),
our
approach
of
defining
a
low­
risk
subcategory
of
PCWP
facilities
is
reasonable,
at
the
very
least
as
a
way
to
reconcile
the
possible
tension
between
the
arguably
restrictive
language
of
section
112(
c)(
1)
and
the
Congressional
intent
behind
section
112(
c)(
9).
(
See,
e.
g.,

Virginia
v.
Browner,
80
F.
3d
869,
879
(
4th
Cir.
1996).)

Alternatively,
even
if
the
language
is
clear
on
its
face
in
restricting
our
ability
to
define
subcategories,
we
feel
that,
as
a
matter
of
historical
fact,
Congress
could
not
have
meant
what
the
commenter
asserts
it
appears
to
have
said,
and
that
as
a
matter
of
logic
and
statutory
structure,

it
almost
surely
could
not
have
meant
it.
(
See,
e.
g.,

Engine
Mfrs.
Ass'n
v.
EPA,
88
F.
3d
1075,
1089
(
D.
C.
Cir.

1996).)

Our
interpretation
of
the
CAA
is
a
reasonable
accommodation
of
the
statutory
language
and
Congressional
intent
regarding
the
relationship
of
the
statutory
categorization
and
subcategorization,
delisting,
MACT
and
residual
risk
provisions
that
apply
to
the
PCWP
category.

This
becomes
clear
in
light
of
the
issue
addressed
by
268
commenters,
which
is
whether
we
may
delist
a
subcategory
of
low­
risk
PCWP
facilities
only
if
such
a
group
of
sources
is
defined
by
criteria
we
have
traditionally
used
to
define
categories
and
subcategories
for
regulatory,
rather
than
delisting
purposes.
Our
approach
implements
Congressional
intent
to
avoid
the
over­
regulatory
result
that
flows
from
an
overly
rigid
reading
of
the
CAA.
When
the
CAA
is
read
as
a
whole,
it
is
apparent
that
Congress
 
which
in
1990
likely
did
not
fully
anticipate
the
policy
considerations
that
come
into
play
in
regulating
HAP
emissions
from
PCWP
facilities
 
has
not
spoken
clearly
on
the
precise
issue.
Our
interpretation
is
necessary
to
fill
this
statutory
gap
and
prevent
the
thwarting
of
Congressional
intent
not
to
unnecessarily
burden
low­
risk
PCWP
facilities
by
forcing
them
to
meet
stringent
MACT
controls
when
they
already
meet
the
risk­
based
goals
of
section
112.
Our
interpretation
thus
lends
symmetry
and
coherence
to
the
statutory
scheme.

While
we
do
not
feel
that
CAA
section
112(
c)(
1)
actually
restricts
our
authority
to
establish
a
low­
risk
PCWP
subcategory,
even
if
the
language
is
so
restrictive,
it
must
be
read
within
the
context
of
Congress'
purpose
in
allowing
us
to
delist
categories
and
subcategories
of
low­
risk
sources
that
are
defined
according
to
the
traditional
criteria
under
CAA
section
111.
It
is
beyond
dispute
that
Congress
determined
that
certain
identifiable
groups
or
sets
269
of
sources
may
be
delisted
if,
as
a
group
and
without
a
single
constituent
source's
exception,
they
are
below
the
enumerated
eligibility
criteria
of
CAA
section
112(
c)(
9).

There
is
no
apparent
reason
why
such
a
group
or
set
of
sources
must
be
limited
to
those
defined
by
traditional
categorization
or
subcategorization
criteria.
This
is
because,
first,
Congress
in
section
112(
c)(
1)
clearly
did
not
absolutely
prohibit
us
from
basing
categories
and
subcategories
on
other
criteria
generally;
and,
second,
the
underlying
characteristic
of
an
eligible
set
or
group
of
sources
under
section
112(
c)(
9)
 
that
no
source
in
the
set
or
group
presents
risks
above
the
enumerated
levels
 
can
be
applied
under
several
approaches
to
defining
categories
and
subcategories
and
is
not
dependent
upon
such
set
or
group
being
traditionally
defined
in
order
to
implement
the
purpose
of
section
112(
c)(
9).
Put
another
way,
there
is
nothing
apparent
in
the
statute
that
precludes
us
from
delisting
a
discernible
set
of
low­
risk
PCWP
facilities
just
because
that
set
cannot
also
be
defined
according
to
other
traditional
criteria
that
have
nothing
to
do
with
the
question
of
whether
each
of
the
constituent
PCWP
facilities
is
low
risk.
As
a
matter
of
logic
and
statutory
structure,

Congress
almost
surely
could
not
have
meant
to
require
that
every
identifiable
group
of
low­
risk
PCWP
facilities,
no
matter
how
large
in
number
or
in
percentage
with
respect
to
270
higher­
risk
facilities
in
the
PCWP
category,
must
remain
subject
to
CAA
section
112,
simply
because
that
group
could
not
be
subcategorized
as
separate
from
the
higher
risk
PCWP
facilities
by
application
of
traditional
subcategorization
criteria.

Where
Congress
squarely
confronted
the
issue,
it
explicitly
provided
relief
for
categories
and
subcategories,

defined
by
traditional
criteria,
that
also
happen
to
present
little
risk.
(
See
CAA
sections
112(
d)(
4),
112(
c)(
9),
and
112(
f)(
2).)
These
CAA
provisions
addressing
risk­
based
relief
from,
or
thresholds
for,
HAP
emissions
regulation
evidence
Congressional
concern
that
the
effects
of
such
pollution
be
taken
into
account,
where
appropriate,
in
determining
whether
regulation
under
CAA
section
112
is
necessary.
At
the
time
of
the
1990
Amendments,
Congress
did
not
consider
it
necessary
to
provide
express
relief
for
additional
groups
such
as
low­
risk
PCWP
facilities,
beyond
those
defined
by
traditional
category
and
subcategory
criteria,
because
it
assumed
we
could
implement
a
comprehensive
regulatory
scheme
for
air
toxics
that
would
both
address
situations
where
technology­
based
standards
were
needed
to
reduce
source
HAP
emissions
to
levels
closer
to
the
risk­
based
goals
of
section
112,
and
avoid
unnecessary
imposition
of
technology­
based
requirements
on
groups
of
sources
that
were
already
meeting
those
goals.
271
Congress
enacted
or
revised
various
CAA
air
toxics
provisions
 
including
sections
112(
c),
(
d)
and
(
f)
 
to
that
end.
Had
events
unfolded
in
that
anticipated
fashion,

in
the
case
of
each
industrial
category
and
subcategory,

there
would
have
been
a
perfect
correlation
between
the
traditional
criteria
for
defining
categories
and
subcategories
and
the
facts
showing
whether
those
groups
are
either
high­
or
low­
risk
HAP
sources.

This
context
turned
out
to
be
more
complex
than
Congress
anticipated,
and
in
the
case
of
PCWP
facilities
there
is
no
clear
differentiation
between
high­
versus
low­
risk
sources
that
corresponds
to
our
traditional
approach
for
identifying
source
categories
and
subcategories.
Nevertheless,
as
in
the
case
of
a
low­
risk
source
group
defined
by
traditional
category
or
subcategory
criteria,
for
the
PCWP
industry,
we
are
able
to
identify
a
significant
group
of
sources
whose
HAP
emissions
pose
little
risk
to
public
health
and
the
environment,
applying
the
same
section
112(
c)(
9)
delisting
criteria
that
would
apply
to
any
traditionally­
defined
source
group.
We
feel
it
is
reasonable
to
conclude
that
Congress
would
not
have
intended
to
over­
regulate
the
lowrisk
PCWP
facilities
due
to
the
inability
to
define
such
a
group
by
traditional
criteria
and
thereby
frustrate
the
coherent
scheme
Congress
set
forth
of
ensuring
that
HAP
sources
ultimately
meet
common
risk­
based
goals
under
272
section
112.

The
commenter's
assertion
that
we
are
inappropriately
altering
our
interpretation
of
the
applicable
statutory
provisions
and
departing
from
the
traditional
categorization
and
subcategorization
criteria
in
addressing
low­
risk
PCWP
facilities
is
thus
unfounded.
As
explained
above,
the
complexity
of
the
air
toxics
problem
and
the
relationship
between
the
traditional
criteria
and
what
might
be
groups
of
low­
risk
sources,
a
context
not
fully
understood
by
either
Congress
or
EPA
at
the
time
of
the
1990
Amendments,
provides
adequate
justification
for
any
unique
applications
of
the
our
approach
for
low­
risk
PCWP
facilities.

Our
approach
does
not
equate
to
one
that
Congress
considered
and
rejected
that
would
have
allowed
source­

bysource
exemptions
from
MACT
based
on
individualized
demonstrations
that
such
sources
are
low
risk.
This
is
because,
contrary
to
that
approach,
we
rely
upon
the
application
of
specific
eligibility
criteria
that
are
defined
in
advance
of
any
source's
application
to
be
included
in
the
low­
risk
PCWP
subcategory,
in
much
the
same
way
as
any
other
applicability
determination
process
works.

Moreover,
in
response
to
the
assertion
that
our
approach
nevertheless
conflicts
with
legislative
history
rejecting
a
similar
(
but
not
identical)
approach
Congress
considered
under
CAA
section
112,
this
legislative
history
is
not
273
substantive
legislative
history
demonstrating
that
Congress
voted
against
relief
from
MACT
in
this
situation
 
there
is
no
such
history.
The
commenters
point
to
a
provision
in
the
House
bill
that
was
not
enacted
but
that
would
have
provided
in
certain
situations
for
case­
by­
case
exemptions
for
lowrisk
sources.
There
is
no
evidence
that
this
provision
was
ever
debated,
considered,
or
voted
upon,
so
its
not
being
enacted
is
not
probative
of
congressional
intent
concerning
our
ability
to
identify
and
delist
a
group
of
low­
risk
PCWP
facilities.
Instead,
it
is
reasonable
to
assume
that,
had
Congress
been
aware
in
1990
of
the
possibility
that
an
identifiable
group
of
PCWP
facilities
is
low
risk,
while
that
group
does
not
correspond
to
traditional
criteria
differentiating
categories
and
subcategories,
Congress
would
have
expressly,
rather
than
implicitly,
authorized
our
action
here.

Moreover,
the
commenters
are
unable
to
cite
any
provision
in
CAA
section
112
that
would
prevent
us
from
being
able
to
add
individual
or
additional
groups
of
lowrisk
PCWP
facilities
to
the
group
we
initially
identify
in
our
final
delisting
action,
as
those
additional
low­
risk
PCWP
facilities
prove
their
eligibility
for
inclusion
in
the
delisted
group
over
time.
In
fact,
the
approach
we
are
taking
for
identifying
additional
low­
risk
PCWP
facilities
is
fully
consistent
with
the
approach
we
have
long
taken
in
274
identifying,
on
a
case­
by­
case
basis
and
subject
to
appropriate
review,
whether
individual
sources
are
members
of
a
category
or
subcategory
subject
to
standards
adopted
under
CAA
sections
111
and
112.

Regarding
the
comment
that
Congress
did
not
expressly
provide
relief
for
carcinogen­
emitting
low­
risk
groups
of
sources
within
the
PCWP
category
other
than
as
an
entire
category,
we
construe
the
provisions
of
CAA
section
112(
c)(
9)
to
apply
to
listed
subcategories
as
well
as
to
categories.
This
construction
is
logical
in
the
context
of
the
general
regulatory
scheme
established
by
the
statute,

and
it
is
the
most
reasonable
one
because
section
112(
c)(
9)(
B)(
ii)
expressly
refers
to
subcategories.
Under
a
literal
reading
of
section
112(
c)(
9)(
B),
no
subcategory
could
ever
be
delisted,
notwithstanding
the
explicit
reference
to
subcategories,
since
the
introductory
language
of
section
112(
c)(
9)(
B)
provides
explicit
authority
to
only
delist
categories.
Such
a
reading
makes
no
sense,
at
the
very
least
because
Congress
plainly
assumed
we
might
also
delist
another
collection
of
sources
besides
either
categories
or
subcategories,
even
in
the
case
of
sources
of
carcinogens.
Both
sections
112(
c)(
9)(
B)(
i)
and
(
ii)
refer
additionally
to
groups
of
sources
in
the
case
of
area
sources
as
being
eligible
for
delisting,
even
though
only
a
category
of
sources
is
specifically
identified
as
eligible
275
for
delisting
in
the
introductory
language
of
section
112(
c)(
9)(
B).
In
light
of
the
broader
congressional
purpose
behind
the
delisting
authority,
we
interpret
the
absence
of
explicit
references
to
subcategories
in
this
introductory
language
and
in
section
112(
c)(
9)(
B)(
i)
as
representing
nothing
more
than
a
drafting
error.

Regarding
the
comments
about
establishing
PCWP
subcategories
based
on
characteristics
other
than
risk,
the
criteria
for
the
low­
risk
subcategory
we
are
delisting
are
based
solely
on
risk
and
not
on
technological
differences
in
equipment
or
emissions.
We
performed
an
analysis
to
determine
which
major
source
PCWP
facilities
may
be
low­
risk
facilities.
Whether
facilities
are
low
risk
or
not
depends
on
the
facility
HAP
emissions;
and
facility
HAP
emissions
are
a
function
of
the
type
and
amount
of
product(
s)

produced,
the
type
of
process
units
(
e.
g.,
direct­
fired
versus
indirect­
fired
dryers)
used
to
produce
the
product,

and
the
emission
control
systems
in
place.
Our
analysis
indicates
that
the
facilities
which
show
low
risk
could
include
facilities
producing
various
products
such
as
particleboard,
molded
particleboard,
medium
density
fiberboard,
softwood
plywood,
softwood
veneer,
fiberboard,

engineered
wood
products,
hardboard,
and
oriented
strandboard.
However,
there
are
also
major
sources
that
produce
these
products
that
are
not
low
risk,
and,
276
therefore,
product
type
cannot
be
used
to
define
the
lowrisk
subcategory.
There
is
no
correlation
between
production
rate
and
low­
risk
facilities
(
e.
g.,
when
facilities
are
sorted
by
production
rate
for
their
product,

the
low­
risk
facilities
are
not
always
at
the
lower
end
of
the
production
rate
range),
so
production
rate
cannot
be
used
as
criteria
for
defining
the
low­
risk
subcategory.
The
low­
risk
facilities
use
a
variety
of
process
equipment
(
e.
g.,
veneer
dryers
at
softwood
plywood
plants
and
tube
dryer
at
MDF
plants).
This
same
equipment
is
used
at
PCWP
plants
that
are
not
low
risk,
and,
therefore,
there
is
no
process
unit
type
distinction
that
can
be
used
to
define
the
low­
risk
subcategory.
The
pollutant
that
drives
the
risk
estimate
can
vary
from
facility
to
facility
because
of
the
different
types
of
process
units
at
each
facility.
There
is
no
clear
distinction
among
low­
risk
and
non­
low­
risk
facilities
when
ranked
by
emissions
of
individual
pollutants
because
of
other
factors
that
contribute
to
facility
risk
such
as
presence
of
a
co­
located
PCWP
facility
or
variability
in
the
pollutants
emitted.
Thus,
there
is
no
emissions
distinction
that
can
be
used
to
define
the
lowrisk
subcategory.
There
is
no
technological
basis
for
creating
a
subcategory
of
PCWP
facilities
that
are
low
risk.

The
commonality
between
all
of
the
low­
risk
PCWP
facilities
is
that
they
are
low
risk,
and,
therefore,
we
have
277
established
the
low­
risk
subcategory
based
on
risk.

We
do
not
agree
with
the
commenters'
assertions
that
our
approach
for
the
low­
risk
PCWP
subcategory
undermines
our
ability
to
identify
the
MACT
floor
for
the
larger
PCWP
category,
either
in
today's
final
PCWP
rule
or
in
any
future
consideration
of
technological
development
under
CAA
section
112(
d)(
6).
This
is
because,
while
low­
risk
PCWP
facilities
will
literally
be
part
of
a
separate
subcategory,
there
is
nothing
in
the
CAA
that
prevents
us
from
including
them
in
any
consideration
of
what
represents
the
best
controlled
similar
source
in
the
new
source
MACT
floor
context,
and
because
it
is
not
unprecedented
for
us
to
look
outside
the
relevant
category
or
subcategory
in
identifying
the
average
emission
limitation
achieved
by
the
best
controlled
existing
sources
if
doing
so
enables
us
to
best
estimate
what
the
relevant
existing
sources
have
achieved.
In
fact,
EPA
has
taken
this
very
approach
in
the
Industrial
Boilers
MACT
rulemaking,
in
order
to
identify
the
MACT
floor
for
mercury
emissions.
Moreover,
the
unique
issues
presented
by
the
low­
risk
PCWP
subcategory
show
that
it
would
be
unreasonable
to
exclude
any
better­
performing
low­
risk
PCWP
sources
from
the
MACT
floor
pool
for
the
larger
PCWP
category.

Traditionally,
EPA
has
based
categories
and
subcategories
partly
on
determinations
of
what
pollution
control
measures
can
be
applied
to
the
relevant
groups
of
sources
in
order
to
278
effectively
and
achievably
reduce
HAP.
In
other
words,
EPA
has
identified
subcategories
for
purposes
of
identifying
the
MACT
floor
in
a
way
that
accounts
for
the
differences
of
sources
types
in
their
abilities
to
control
HAP
emissions.

But
whether
a
PCWP
source
is
a
low­
risk
source
does
not
necessarily
turn
on
such
a
distinction
 
two
sources
might
have
identical
abilities
to
control
HAP
emissions,
but
the
unique
circumstances
of
one
source
regarding
the
impacts
of
its
HAP
emissions
will
determine
whether
or
not
it
is
a
lowrisk
PCWP
source.
(
In
fact,
it
is
theoretically
possible
that
between
two
sources
the
better
performing
source
will
be
a
high­
risk
source,
and
the
worse­
performing
source
will
be
a
low­
risk
source,
based
on
circumstances
that
are
unrelated
to
the
question
of
what
abilities
the
sources
have
to
control
HAP
emissions
through
application
of
MACT,
such
as
the
sources'
locations
vis
a
vis
exposed
human
populations.)
Therefore,
EPA
feels
that
not
only
is
it
appropriate
to
include
any
better­
performing
low
risk
PCWP
sources
in
the
MACT
floor
determinations
for
the
larger
PCWP
category,
but
that
excluding
such
sources
simply
based
on
the
unique
facts
of
the
impacts
of
their
emissions,
with
there
being
no
difference
in
the
abilities
of
high­
risk
and
low­
risk
sources
to
apply
HAP
emission
control
measures,

would
becould
result
in
an
unlawfulundesirable
weakening
of
the
MACT
floor
for
the
larger
PCWP
category.
To
that
end,
279
the
MACT
floors
established
for
PCWP
process
units
today
are
in
no
way
affected
by
our
establishment
of
the
low­
risk
PCWP
subcategory.

Finally,
we
disagree
with
the
argument
by
one
commenter
that
the
low­
risk
PCWP
subcategory
approach
represents
an
impermissible
cost­
based
exemption
from
MACT
or
factor
in
determining
MACT.
Certainly
it
is
true
that
costs
may
not
be
considered
in
setting
the
MACT
floor.
However,
there
is
nothing
in
the
CAA
that
prevents
us
from
noting
the
cost
impacts,
beneficial
or
adverse,
of
our
actions
in
setting
MACT
floors,
assessing
possible
beyond­
the­
floor
measures,

or
conducting
risk­
based
actions
under
CAA
section
112.
In
fact,
we
routinely
evaluate
the
costs
of
our
regulatory
actions,
even
when
cost
factors
may
not
be
used
to
influence
the
regulatory
decision
itself,
in
order
to
comply
with
applicable
Executive
Order
and
statutory
administrative
review
requirements.
Simply
because
there
is
a
cost
benefit
to
some
members
of
the
PCWP
category
in
our
establishing
a
low­
risk
PCWP
subcategory
does
not
make
that
action
impermissible,
provided
that
our
subcategorization
and
delisting
are
otherwise
properly
based
on
the
appropriate
risk­
based
criteria
under
CAA
section
112(
c)(
9).
Section
112
by
its
own
terms
does
not
forbid
the
goal
of
achieving
environmental
protection
in
a
less
costly
manner.

Similarly,
it
is
appropriate
for
EPA
to
note
the
beneficial
280
air
pollution­
related
impacts
of
not
requiring
low­
risk
PCWP
sources
to,
for
example,
install
criteria
pollutant
emission­
producing
RTOs.
While
it
is
true
that
such
air
quality­
related
impacts
could
not
constitute
non­
air
quality
health
and
environmental
impacts
that
EPA
must
consider
when
setting
MACT
under
CAA
section
112(
d)(
2),
nothing
in
the
CAA
prevents
EPA
from
taking
account
of
such
impacts
in
developing
its
policy
regarding
whether
it
is
appropriate
to
delist
a
subcategory
under
section
112(
c)(
9)
when
that
subcategory
otherwise
meets
the
statutory
criteria
for
delisting.
Therefore,
EPA
does
not
agree
with
commenters
who
claim
that
its
approach
to
delisting
the
low
risk
PCWP
subcategory
conflicts
with
how
it
has
argued
issues
regarding
either
de
minimis
authority,
cost­
based
exemptions
from
MACT,
or
the
treatment
of
non­
air
quality
impacts
and
the
consideration
of
risk
in
setting
the
actual
MACT
standard
before
the
U.
S.
Court
of
Appeals
for
the
D.
C.

Circuit.
Nor
does
our
approach
contravene
any
of
that
Court's
rulings
on
these
issues.

3.
Criteria
for
demonstrating
low
risk
Dose­
response
values
Comment:
Two
commenters
suggested
that
EPA
incorporate
into
the
PCWP
rule
the
findings
of
the
nationwide
wood
products
risk
assessment,
which
they
claim
demonstrates
that
the
vast
majority
of
wood
products
sources
cause
no
281
meaningful
risk
to
human
health
or
the
environment
at
current
emission
levels.
The
commenters
stated
that
the
risk
assessment
used
existing
air
dispersion
modeling
studies
of
34
wood
products
facilities
throughout
the
U.
S.

to
estimate
the
maximum
annual
off­
site
HAP
concentrations
at
wood
products
facilities
nationwide.
According
to
the
commenters,
the
risk
assessment
indicates
that
large
subgroups
of
facilities
that
are
affected
sources
under
the
PCWP
rule
as
proposed
(
i.
e.,
fiberboard,
medium
density
fiberboard,
and
plywood
facilities)
generally
are
expected
to
pose
insignificant
risks
to
human
health,
based
on
a
comparison
of
predicted
off­
site
concentrations
with
applicable
health
benchmarks.
One
of
the
commenters
stated
that
many
of
the
facilities
with
low
off­
site
concentrations
will
likely
be
smaller
plants
that
would
not
be
able
to
justify
installation
of
(
additional)
emission
controls
and
may
face
closure
without
a
risk­
based
compliance
option.

The
other
commenter
stated
that
a
comparison
of
off­
site
concentrations
of
formaldehyde
and
acetaldehyde
with
benchmarks
reflecting
the
latest
toxicological
evidence
indicates
that
exposures
to
those
HAP
are
well
below
levels
of
concern.
Acrolein
was
the
only
HAP
with
potential
exposures
at
some
facilities
(
i.
e.,
subset
of
fiberboard,

medium
density
fiberboard
and
plywood
facilities)
that
exceeded
the
health
benchmark.
However,
the
commenter
282
stated
that
the
acrolein
findings
may
not
represent
an
actual
risk
to
human
health
because
exceedences
of
the
benchmark
may
be
attributable
to
EPA
averaging
a
large
number
of
non­
detects
at
one­
half
the
detection
limit,

thereby
artificially
increasing
predicted
acrolein
emissions.
Based
on
these
overall
findings,
the
commenter
concluded
that
the
wood
products
risk
assessment
indicates
that
incinerator
control
is
not
warranted
on
the
basis
of
human
health
concerns
for
a
large
number
of
facilities.

Response:
We
acknowledge
receipt
of
the
industrysponsored
nationwide
wood
products
MACT
risk
assessment
submitted
by
the
commenter.
However,
we
conducted
our
own
risk
analysis
to
evaluate
the
merits
of
including
and
delisting
a
low­
risk
subcategory
in
today's
final
PCWP
rule.

The
methodology
used
in
our
risk
analysis
differed
widely
from
the
methodology
used
in
industry's
risk
assessment.

For
example,
industry's
risk
assessment
was
based
on
previously
conducted
air
dispersion
modeling
studies
for
34
PCWP
facilities,
while
our
analysis
used
emission
estimates
developed
for
each
PCWP
facility
expected
to
be
a
major
source
of
HAP.
We
used
different
(
generally
more
protective)
human
health
benchmarks
in
our
risk
assessment
than
were
used
in
industry's
risk
assessment.
We
also
considered
all
HAP
(
including
metal
HAP)
in
our
risk
analysis,
whereas
industry's
risk
assessment
considered
only
283
methanol,
formaldehyde,
acetaldehyde,
acrolein,
phenol,
and
propionaldehyde.

Based
on
our
risk
analysis,
we
conclude
that
HAP
emissions
from
some
PCWP
facilities
pose
little
risk
to
human
health
and
the
environment.
Therefore,
we
have
included
a
subcategory
of
low­
risk
PCWP
facilities
in
today's
final
PCWP
rule,
and
are
delisting
that
subcatgeorysubcategory.
Appendix
B
to
subpart
DDDD
of
40
CFR
part
63
includes
procedures
that
facilities
may
use
to
demonstrate
that
they
are
part
of
the
delisted
low­
risk
subcategory,
and,
therefore,
are
not
subject
to
the
compliance
options
included
in
today's
final
PCWP
MACT
rule.

To
demonstrate
eligibility
for
the
low­
risk
subcategory,

facilities
must
first
conduct
emissions
testing
for
up
to
13
HAP
(
five
organic
HAP
from
all
process
units,
seven
metal
HAP
from
direct­
fired
process
units,
and
MDI
from
presses
processing
product
containing
MDI
resin).
The
rationale
for
selection
of
these
13
HAP
is
described
elsewhere
in
this
section
and
in
the
supporting
documentation
for
the
final
rule.
Facilities
must
use
the
results
from
emissions
testing
to
preliminarily
demonstrate,
subject
to
EPA
approval,
that
they
are
part
of
the
low­
risk
subcategory
using
either
a
look­
up
table
analysis
(
based
on
the
look­
up
tables
included
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63)
or
site­
specific
risk
assessment
methodology
(
described
284
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
and
other
analytical
tools,
such
as
the
"
Air
Toxics
Risk
Assessment
Reference
Library"
if
appropriate
for
the
specific
source)

and
risk
benchmarks
(
described
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63).

Regarding
acrolein,
the
commenter
is
correct
in
that,

when
developing
AP­
42
emission
factors,
we
used
a
value
of
one­
half
the
detection
limit
for
all
non­
detect
sample
runs
if
acrolein
was
detected
in
any
sample
runs
from
the
applicable
source
category.
Acrolein
has
been
detected
in
process
unit
emissions
from
all
sectors
of
the
PCWP
industry,
except
for
hardwood
plywood
manufacturing.
When
using
emission
factors
to
estimate
emissions
from
PCWP
facilities,
we
did
not
estimate
emissions
of
a
pollutant
when
all
of
the
emissions
test
runs
were
non­
detect.

However,
we
did
use
emission
factors
that
included
a
mixture
of
detectable
values
and
values
based
on
one­
half
of
the
method
detection
limit
(
MDL)
when
acrolein
was
detected
at
least
once
for
a
particular
type
of
process
unit.
We
maintain
that
this
approach
to
handling
non­
detects
is
appropriate
for
the
purposes
that
we
used
the
emissions
data.
Facilities
will
conduct
emissions
tests
instead
of
using
emission
factors
to
demonstrate
eligibility
for
the
low­
risk
subcategory.
To
prevent
facilities
from
including
HAP
that
are
not
detected
in
their
low­
risk
demonstrations,
285
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
states
that
facilities
may
use
zero
for
non­
detects
when
all
of
the
emission
test
runs
are
below
the
MDL,
provided
that
certain
criteria
are
met
to
ensure
that
emissions
testing
and
analysis
procedures
are
adequate
to
detect
low
concentrations
of
HAP.

Comment:
One
commenter
stated
that
CAA
section
112(
d)(
4)
is
particularly
ill­
suited
to
the
PCWP
and
industrial
boiler
source
categories.
The
commenter
stated
that,
even
if
EPA
had
authority
to
create
individualized
MACT
exemptions
based
on
health
thresholds,
it
could
not
do
so
if
there
is
insufficient
evidence
on
the
pollutants
emitted
to
establish
a
NOEL.
According
to
the
commenter,

section
112(
d)(
4)
does
not
apply
for
chemicals
that
do
not
have
a
well­
defined
threshold
based
on
reliable
science.

The
commenter
stated
that
available
evidence
does
not
establish
a
no­
effect
threshold
for
acetaldehyde,
acrolein,

benzene,
carbon
tetrachloride,
chloroform,
formaldehyde,

manganese,
methylene
chloride,
and
phenol.
As
rationale,

the
commenter
presented
a
summary
of
the
available
health
effects
data
for
each
of
these
pollutants.

Response:
As
stated
elsewhere
in
this
preamble,
we
are
not
pursuing
establishment
of
a
threshold
emission
rate
for
the
PCWP
source
category
under
CAA
section
112(
d)(
4)
because
PCWP
facilities
emit
non­
threshold
pollutants.
Therefore,
286
this
comment
is
irrelevant
in
the
context
of
the
PCWP
rule.

Comments
pertaining
to
the
Industrial/
Commercial/
Institutional
Boilers
and
Process
Heaters
NESHAP
are
addressed
in
the
comment­
response
document
for
that
rule.
(
See
Docket
ID
No.
OAR­
2002­
0058.)

Comment:
Two
commenters
expressed
concern
about
the
health
benchmark
data
sources
that
EPA
used.
The
first
commenter
argued
that
the
proposal
inappropriately
used
draft
guidelines
and
toxicity
profiles
that
had
not
been
subject
to
public
review
and/
or
were
not
publicly
available.

The
commenter
was
particularly
concerned
with
the
use
of
non­
linear
carcinogenic
risk
values
and
toxicity
profiles
(
for
HAP)
that
have
not
been
finalized
and
are
not
available
for
review
by
the
public.

The
second
commenter
argued
that
EPA
should
not
rely
solely
on
the
health
benchmarks
in
its
Integrated
Risk
Information
System
(
IRIS)
database.
The
commenter
stated
that
IRIS,
while
useful
for
obtaining
information
about
the
health
effects
of
chemicals,
is
far
from
definitive,
as
EPA
resource
constraints
have
resulted
in
many
chemical
summaries
that
are
significantly
outdated
and
do
not
reflect
the
most
recent
scientific
developments.
Moreover,
the
commenter
stated
that
the
IRIS
database
is
a
non­
statutory,

in­
house
EPA
activity,
and
IRIS
entries
are
not
subject
to
formal
notice
and
comment.
The
commenter
noted
that
EPA
287
management
has
repeatedly
emphasized
in
directives
that
other
information
must
be
considered,
in
addition
to
the
IRIS
database,
when
evaluating
the
health
effects
of
chemicals
in
a
regulatory
context.
The
commenter
concluded
that
EPA
must
use
a
scientifically
appropriate
health
benchmark
based
on
a
consideration
of
all
relevant
information
to
ensure
that
the
health
benchmark
is
up­
to­
date
and
scientifically
credible,
even
if
that
means
departing
from
the
value
in
IRIS.

A
third
commenter
agreed
with
EPA's
choice
to
derive
their
data
from
IRIS,
California
EPA
(
CalEPA),
and
Agency
for
Toxic
Substances
and
Disease
Registry
(
ATSDR)
for
its
documentation
for
establishing
risk
based
threshold
and
nonthreshold
values.
The
commenter
added
that
almost
all
HAP
are
being
reviewed
and
reevaluated
on
a
regular
basis,
and
it
would
be
inappropriate
to
single
out
formaldehyde
and
acetaldehyde
at
this
time.
The
commenter
stated
that
EPA
can
only
rely
on
what
is
currently
published
and
has
underdone
either
peer
review
or
EPA
review.
According
to
the
commenter,
the
issue
of
changing
health­
based
guideline
values
will
always
be
a
concern
once
health­
based
regulations
are
promulgated.

Response:
We
agree
with
the
first
two
commenters
that
we
should
use
the
best
available
sources
of
health
effects
information
for
risk
or
hazard
determinations.
As
we
have
288
1U.
S.
Environmental
Protection
Agency.
1999.
Residual
Risk
Report
to
Congress.
Office
of
Air
Quality
Planning
and
Standards,
Research
Triangle
Park,
NC
27711,
March
1999,
EPA­
453/
R­
99­
001;
available
at
http://
www.
epa.
gov/
ttn/
oarpg/
t3/
meta/
m8690.
html.
(
EPA
1999)
stated
previously,
we
will
not
be
relying
exclusively
on
IRIS
values,
but
will
be
considering
all
credible
and
readily
available
assessments.
1
For
air
toxics
risk
assessments,
we
identify
pertinent
toxicity
or
dose­
response
values
using
a
default
hierarchy
of
sources,
with
IRIS
being
the
preferred
source,
to
assist
us
in
identifying
the
most
scientifically
appropriate
benchmarks
for
our
analyses
and
decisions.
The
IRIS
process
contains
internal
and
external
peer
review
steps
and
represent
EPA
consensus
values.
When
adequate
toxicity
information
is
not
available
in
IRIS,
we
consult
other
sources
in
a
default
hierarchy
that
recognizes
the
desirability
of
these
qualities
in
ensuring
that
we
have
consistent
and
scientifically
sound
assessments.

Furthermore,
where
the
IRIS
assessment
substantially
lags
the
current
scientific
knowledge,
we
have
committed
to
consider
alternative
credible
and
readily
available
assessments.
For
our
use,
these
alternatives
need
to
be
grounded
in
publicly
available,
peer­
reviewed
information.

Formaldehyde
is
an
example
of
this
situation.
We
are
not
using
information
that
does
not
meet
these
requirements.
We
also
agree
with
the
third
commenter
that
the
issue
of
289
changing
health­
based
guideline
values
is
a
general
challenge
in
setting
health­
based
regulations.
However,
we
are
committed
to
setting
such
regulations
that
reflect
current
scientific
understanding,
to
the
extent
feasible.

Facilities
conducting
low­
risk
demonstrations
should
refer
to
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
and
other
analytical
tools,
such
as
the
"
Air
Toxics
Risk
Assessment
Guidance
Reference
Library"
(
if
appropriate
for
the
specific
source)
for
guidance
on
choosing
appropriate
dose­
response
values.

Comment:
With
the
support
of
several
others,
one
commenter
pointed
out
that
the
science
with
respect
to
formaldehyde
and
acetaldehyde
has
changed
since
EPA's
initial
IRIS
entries
for
those
pollutants
were
completed.

Consequently,
the
commenter
believed
it
would
be
inappropriate
for
EPA
to
rely
on
the
unit
risk
factors
for
those
pollutants
in
the
IRIS
database
in
establishing
a
property
line
concentration
threshold
in
the
PCWP
rule
as
proposed.
The
commenter
supported
EPA's
efforts
in
revising
its
formaldehyde
and
acetaldehyde
IRIS
assessment
and
noted
that
both
revisions
are
expected
to
be
finalized
before
the
final
PCWP
rule
is
published
in
2004.
Regarding
formaldehyde,
the
commenter
noted
that
EPA
plans
on
using
the
model
from
the
Chemical
Industry
Institute
of
Technology
(
CIIT)
to
revise
its
formaldehyde
IRIS
assessment
and
290
encouraged
this
action.
The
commenter
pointed
out
that
the
CIIT
model
has
been
recognized
by
several
authoritative
bodies
(
e.
g.,
Health
Canada/
Environment
Canada,
Organization
for
Economic
Coordination
and
Development,
and
World
Health
Organization)
as
providing
the
most
scientifically
defensible
analysis
of
formaldehyde.
(
Another
commenter
added
that
the
IRIS
risk
criteria
for
formaldehyde
clearly
cause
formaldehyde
risk
estimates
to
be
overstated
but
argued
that,
even
using
the
very
conservative
IRIS
numbers,

risks
are
still
low.
A
third
commenter
urged
EPA
not
to
use
the
formaldehyde
values
in
ATSDR,
stating
that
they
are
fundamentally
flawed,
as
detailed
in
their
comment.)

Regarding
acetaldehyde,
the
commenter
recommended
that
EPA
use
a
health
benchmark
between
27
and
390
micrograms
per
cubic
meter
(
µ
g/
m3)
and
included
their
rationale
in
an
attachment
to
their
comment.
If
EPA
is
unable
to
complete
its
reassessments
before
the
PCWP
rule
is
finalized,
the
commenter
encouraged
EPA
not
to
revert
to
the
original
IRIS
unit
risk
factors
for
formaldehyde
and
acetaldehyde.

Instead,
the
commenter
recommended
that
EPA
use
the
CIIT
model
(
or
alternatively
defer
to
Health
Canada/
Environment
Canada)
for
formaldehyde
and,
at
a
minimum,
use
the
IRIS
reference
concentration
(
RfC)
of
9
µ
g/
m3
for
acetaldehyde.

Response:
With
the
exception
of
formaldehyde,
we
are
using
the
human
health
values
currently
used
by
EPA's
air
291
toxics
program
and
available
at:

http://
www.
epa.
gov/
ttn/
atw/
toxsource/
summary.
html.
These
dose
response
values
come
from
several
sources
including
EPA's
IRIS,
the
Centers
for
Disease
Control's
ATSDR,
and
California
EPA.
See
the
supporting
information
for
this
rulemaking
for
a
summary
of
the
human
health
values
we
used
in
our
assessment.

For
formaldehyde,
we
do
not
use
the
dose­
response
value
reported
in
IRIS.
The
dose­
response
value
in
IRIS
is
based
on
a
1987
study,
and
no
longer
represents
the
best
available
science
in
the
peer­
reviewed
literature.
Since
that
time,

significant
new
data
and
analysis
have
become
available.
We
based
the
dose­
response
value
we
used
for
formaldehyde
on
work
conducted
by
the
CIIT
Centers
for
Health
Research
(
formerly,
the
Chemical
Industry
Institute
of
Toxicology).

In
1999,
the
CIIT
published
a
risk
assessment
which
incorporated
mechanistic
and
dosimetric
information
on
formaldehyde
that
had
been
accumulated
over
the
past
decade.

The
risk
assessment
analyzed
carcinogenic
risk
from
inhaled
formaldehyde
using
approaches
that
are
consistent
with
EPA's
draft
guidelines
for
carcinogenic
risk
assessment.
The
CIIT
model
is
based
on
computational
fluid
dynamics
(
CFD)
models
of
airflow
and
formaldehyde
delivery
to
the
relevant
parts
of
the
rat
and
human
respiratory
tract,
which
are
then
coupled
to
a
biologically­
motivated
two­
staged
clonal
growth
292
model
that
allows
for
incorporation
of
different
biological
effects.
These
biological
effects,
such
as
interaction
with
DNA
and
cell
proliferation,
are
processes
by
which
formaldehyde
may
contribute
to
development
of
cancer
at
sites
exposed
at
the
portal
of
entry
(
e.
g.,
respiratory
tract).
The
two­
staged
model
is
a
much
more
advanced
approach
for
examining
the
relevance
of
tumors
seen
in
animal
models
for
human
populations.

We
believe
that
the
CIIT
modeling
effort
represents
the
best
available
application
of
the
available
mechanistic
and
dosimetric
science
on
the
dose­
response
for
portal
of
entry
cancers
due
to
formaldehyde
exposures.
We
note
here
that
other
organizations,
including
Health
Canada,
have
adopted
this
approach.
Accordingly,
we
have
used
risk
estimates
based
on
the
CIIT
airflow
model
coupled
to
a
two­
staged
clonal
growth
model
as
the
basis
for
the
dose­
response
values
for
this
analysis.
This
model
incorporates
state­

ofthe
art
analyses
for
species­
specific
dosimetry,
and
encompasses
more
of
the
available
biological
data
than
any
other
currently
available
model.
As
with
any
model,

uncertainties
exist,
and
this
model
is
sensitive
to
the
inputs,
but
we
believe
it
represents
the
best
available
approach
for
assessing
the
risk
of
portal­
of­
entry
cancers
due
to
formaldehyde
exposures.

We
are
aware
of
recently
completed
epidemiologic
293
analyses
that
update
existing
cohorts
of
workers
exposed
to
formaldehyde.
Not
all
these
results
have
been
published
in
peer
reviewed
journals
at
this
point,
and
the
new
data
are
not
fully
available.
These
studies
examine
the
potential
for
formaldehyde
to
cause
cancer
in
other
organs.
These
new
results
in
combination
with
the
previously
available
literature
regarding
the
potential
for
systemic
effects
indicates
the
pollutant
may
be
taken
into
the
body
and
distributed
to
extra
respiratory
organs,
potentially
causing
adverse
effects
that
are
not
picked
up
by
the
CIIT
model.

As
these
new
studies
are
published
in
peer
reviewed
journals
and
data
sets
are
made
available
for
further
analysis,
we
will
evaluate
them
in
the
context
of
the
full
body
of
literature
to
determine
the
relevance
to
our
conclusions
with
regard
to
risks
from
exposure
to
formaldehyde.

Currently,
the
CIIT
information
and
other
recent
information,
including
recently
published
epidemiological
studies,
are
being
reviewed
and
considered
in
the
reassessment
of
our
formaldehyde
unit
risk
estimate
(
URE).

We
plan
to
bring
this
reassessment
to
the
Science
Advisory
Board
in
the
summer
of
2004.
The
feasibility
of
delisting
a
subgroup
of
facilities
based
on
risk
is
not
compromised
by
the
existing
formaldehyde
dose­
response
value
because
some
facilities
would
qualify
for
delisting
based
on
this
current
value.
We
are
moving
forward
with
the
final
PCWP
rule
at
294
this
time
because
there
is
a
court­
ordered
deadline,
and
we
are
including
the
low­
risk
PCWP
subcategory
delisting
and
basing
our
review
of
sources's
eligibility
on
the
CIIT
model
for
formaldehyde
because,
at
this
point
in
time,
that
reflects
the
best
available
source
of
health
effects
information.
We
disagree
with
the
statement
by
one
of
the
commenters
that
risks
are
still
low
using
the
current
IRIS
number
for
formaldehyde.
Our
analysis
has
demonstrated
that
not
all
PCWP
facilities
can
be
considered
low
risk
when
either
the
current
IRIS
or
CIIT
URE
for
formaldehyde
is
employed.

While
we
recognize
the
similarities
between
acetaldehyde
and
formaldehyde
with
regard
to
suggested
modes
of
action,

the
reassessment
of
acetaldehyde
is
lagging
behind
that
of
formaldehyde.
The
formaldehyde
reassessment
is
further
along
because
of
the
preponderance
of
data
specific
to
formaldehyde
and
the
potentially
greater
impact
of
a
change
in
potency
to
our
regulatory
decisions.
Unlike
for
formaldehyde,
an
alternative,
peer­
reviewed,
publicly
available
assessment
does
not
currently
exist
for
acetaldehyde,
leaving
us
with
the
current
IRIS
assessment.

We
do
not
feel
it
is
necessary
to
wait
for
our
acetaldehyde
reassessment
to
be
completed,
due
to
the
court­
ordered
deadline
for
the
final
PCWP
MACT
rule,
and
due
to
the
fact
that
until
otherwise
concluded
the
IRIS
values
for
295
acetaldehyde
reflect
the
best
available
source
of
health
effects
information.
Therefore,
we
are
relying
on
the
IRIS
values
for
acetaldehyde
in
both
cancer
and
non­
cancer
risk
assessments
for
the
final
rule.

Facilities
conducting
low­
risk
demonstrations
should
refer
to
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
and
other
analytical
tools,
such
as
the
"
Air
Toxics
Risk
Assessment
Reference
Library"
(
if
appropriate
for
the
specific
source)
for
guidance
on
choosing
appropriate
doseresponse
values.

Comment:
One
commenter
stated
that
EPA
should
consider
formaldehyde
and
acetaldehyde
as
carcinogens
unless
a
reassessment
classifies
them
as
threshold
pollutants.
A
second
commenter
argued
that
formaldehyde
and
acetaldehyde
are
properly
treated
as
threshold
pollutants.
This
commenter
contended
that
the
legislative
history
of
the
CAA
makes
clear
that
Congress
considered
"
threshold
pollutants"

to
be
those
for
which
a
"
no
observed
effect
level"
can
be
established.
(
See,
e.
g.,
S.
Rep.
No.
228,
101st
Cong.,
1st
Sess.
175­
176
(
1990)).
By
contrast,
a
non­
threshold
pollutant
is
one
for
which
a
no
observed
effect
level
cannot
be
identified,
i.
e.,
a
pollutant
for
which
adverse
effects
may
be
seen
at
any
dose
level
above
zero.
The
commenter
noted
that
EPA
has
historically
assumed
that
all
carcinogens
are
non­
threshold
pollutants
that
may
trigger
a
carcinogenic
296
effect
at
any
exposure
level,
no
matter
how
small.
However,

as
mechanistic
data
on
the
mode
of
action
of
carcinogenesis
advances,
that
conservative
assumption
may
prove
not
to
be
accurate
for
certain
pollutants.
The
commenter
stated
that
the
available
science
strongly
suggest
that
these
pollutants
act
as
threshold
carcinogens.
The
commenter
contended
that
there
is
a
no
observed
effect
level
for
formaldehyde
below
which
the
carcinogenic
risk
either
does
not
exist
or
cannot
be
measured,
as
documented
in
an
attachment
to
their
comment.
The
commenter
stated
that
acetaldehyde
should
be
viewed
similarly
because
acetaldehyde
is
similar
to
formaldehyde
structurally
and
toxicologically,
and
is
expected
to
behave
similarly
mechanistically.
Because
acetaldehyde
is
a
less
potent
carcinogen
than
formaldehyde
(
by
an
order
of
magnitude),
non­
cancer
health
effects
(
which
clearly
are
threshold
health
effects)
are
the
likely
risk
driver
for
that
pollutant.
Finally,
the
commenter
noted
that
EPA's
recently
issued
Draft
Final
Guidelines
for
Carcinogenic
Risk
Assessment
provide
that,
for
non­
linear
carcinogens,
EPA
will
calculate
a
reference
dose
(
RfD)
or
RfC,
which
are
safe
lifetime
doses
(
i.
e.,
doses
below
which
adverse
effects
will
not
occur).
The
commenter
stated
that
this
is
exactly
what
a
threshold
pollutant
is.
Thus,
EPA's
revised
guidelines
support
the
conclusion
that
formaldehyde
and
acetaldehyde
should
be
treated
as
threshold
pollutants.
297
Response:
We
agree
that
we
should
consider
formaldehyde
and
acetaldehyde
as
carcinogens
unless
a
reassessment
classifies
them
as
threshold
pollutants.
Currently,

formaldehyde
and
acetaldehyde
are
considered
probable
human
carcinogens.
Both
are
under
review,
and
their
dose­
response
values
for
carcinogenicity
are
likely
to
change.
For
the
final
rule,
we
are
using
an
alternative
dose­
response
value
for
formaldehyde
based
on
a
peer­
reviewed,
publicly
available
assessment.
However,
we
do
not
have
comparable
quantitative
information
for
acetaldehyde.
Therefore,
we
will
use
the
current
IRIS
value.
Facilities
conducting
lowrisk
demonstrations
should
refer
to
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
(
and/
or
the
"
Air
Toxics
Risk
Assessment
Reference
Library")
for
guidance
on
choosing
appropriate
dose­
response
values.

Comment:
One
commenter
expressed
concern
about
some
of
the
health
benchmarks
that
EPA
plans
to
publish.
The
commenter
reviewed
various
health
studies
for
each
pollutant
and
recommended
several
RfC
values.
The
commenter
noted
that,
because
IRIS
does
not
have
an
RfC
for
methanol,
EPA
has
indicated
it
plans
to
determine
a
de
minimis
threshold
for
methanol
using
a
value
of
4.0
milloigrams
per
cubic
meter
(
mg/
m3)
as
an
RfC.
The
commenter
noted
that
this
value
is
the
noncancer
chronic
reference
exposure
level
(
REL)
derived
by
CalEPA.
The
commenter
stated
that
CalEPA's
298
derivation
of
that
REL
contains
some
errors
and
inaccurate
assumptions.
According
to
the
commenter,
a
more
accurate
estimate
of
a
human
safe
level
for
chronic
exposure
to
methanol
by
inhalation,
derived
from
the
same
mouse
study
data,
is
171
mg/
m3,
which
is
discussed
further
in
their
comments.
The
commenter
stated
that
their
discussion
presents
new
analyses
not
previously
reviewed
by
EPA
and
a
ground­
breaking
new
approach
to
a
hazard
assessment
for
methanol.
The
commenter
noted
that
EPA
is
currently
revising
its
assessment
for
acrolein
and
has
provided
for
public
information
a
draft
toxicological
review
and
draft
IRIS
summary
for
acrolein.
The
draft
IRIS
document
states
that
the
proposed
new
RfC
of
0.03
µ
g/
m3
replaces
the
previous
RfC
of
0.02
µ
g/
m3,
and
that
this
new
RfC
is
based
on
a
more
recent
interpretation
of
the
database.
The
commenter
noted
the
basis
for
the
revised
acrolein
RfC
(
Feron
et
al,
1978)
and
argued
that
EPA's
interpretation
of
this
study
is
overly
conservative.
The
commenter
stated
that
EPA
has
used
the
maximum
uncertainty
factors
that
could
reasonably
be
justifiable
and
thereby
developed
an
RfC
that
almost
certainly
goes
beyond
what
is
needed
to
protect
human
health.
The
commenter
suggested
that
EPA
should
instead
use
the
more
realistic
reference
exposure
level
developed
by
CalEPA,
which
is
more
conservative
than
the
Health
Canada
Tolerable
Concentration.
299
The
commenter
noted
that
EPA
has
not
published
a
health
benchmark
for
phenol.
The
commenter
agreed
with
EPA's
proposal
to
use
the
CalEPA
REL
of
200
µ
g/
m3
for
phenol
in
implementing
the
risk­
based
approach
for
wood
products
facilities.
According
to
the
commenter,
the
REL
is
intended
to
serve
the
same
goal
as
an
RfC.

The
commenter
supported
using
a
health
benchmark
of
110
µ
g/
m3
for
propionaldehyde
and
believed
that
this
value
would
protect
human
health
with
an
ample
margin
of
safety.
The
commenter
described
how
the
110
µ
g/
m3
value
was
derived
based
on
the
threshold
limit
value
(
TLV)
for
propionaldehyde
identified
by
the
American
Conference
of
Governmental
Industrial
Hygienists
(
ACGIH).
The
commenter
explained
that
this
benchmark
is
consistent
with
values
developed
by
other
organizations.

Response:
We
are
currently
developing
an
IRIS
assessment
for
methanol,
and
any
new
information
that
exists
that
has
undergone
peer
review
will
be
considered
in
this
re­
evaluation.
We
publish
yearly
in
the
Federal
Register
a
list
of
all
chemicals
for
which
we
are
planning
IRIS
assessment
activity.
This
action
further
requests
submission
of
pertinent
data
for
these
chemicals.
In
lieu
of
the
pending
IRIS
assessment,
we
will
continue
to
draw
on
other
sources
identified
by
our
established
default
hierarchy
of
data
sources,
which
have
as
part
of
their
300
development
processes
external
or
peer
review,
in
addition
to
extensive
internal
reviews.

A
reassessment
of
acrolein
was
completed
in
June
of
2003.
The
RfC
resulting
from
that
reassessment
(
i.
e.,
an
RfC
of
0.02

g/
m3,
with
an
uncertainty
factor
of
1,000)
is
what
is
currently
on
IRIS.
As
with
all
announced
IRIS
reassessments,
time
was
provided
for
new
data
or
relevant
information
to
be
submitted.
In
addition,
each
assessment
undergoes
extensive
internal
review
as
well
as
external
peer
review
to
ensure
that
the
data
used
are
scientifically
sound.
We
feel
that
we
have
developed
the
most
scientifically
sound
RfC
that
will
ensure
that
risk
assessments
using
this
number
are
health­
protective.

Facilities
conducting
low­
risk
demonstrations
should
refer
to
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
(
and/
or
the
"
Air
Toxics
Risk
Assessment
Reference
Library")
for
guidance
on
choosing
appropriate
dose­
response
values.

We
do
not
currently
have
plans
to
develop
an
IRIS
assessment
for
phenol.
We
will
continue
to
rely
on
our
hierarchy
of
other
sources
when
IRIS
values
are
not
available.

We
do
not
have
an
IRIS
file
for
proprionaldehyde,
and
an
assessment
is
not
available
from
the
alternative
sources
in
our
default
hierarchy.
The
hierarchy
sources
do
not
include
ACGIH,
as
that
organization
develops
reference
values
for
301
use
in
occupational
exposure
settings,
as
opposed
to
the
ambient
air
exposures
that
are
the
focus
of
this
action.

Development
of
an
IRIS
assessment
for
propionaldehyde
is
currently
underway.
Once
available,
it
will
be
used
in
future
risk
analyses.
In
the
meantime,
this
HAP
was
not
included
in
the
assessment
conducted
for
PCWP
facilities.

Comment:
One
commenter
stated
that
comparison
of
modeled
exposures
to
the
RfC
or
similarly­
derived
health
benchmark
is
highly
protective
and
meets
the
CAA's
"
ample
margin
of
safety"
requirement.
Although
the
commenter
claims
the
CAA
does
not
explicitly
define
"
ample
margin
of
safety,"
in
the
Vinyl
Chloride
case,
the
D.
C.
Circuit
Court
of
Appeals
articulated
the
purpose
of
the
ample
margin
of
safety
determination
as
obtaining
a
"
reasonable
degree
of
protection"
in
light
of
scientific
uncertainties
and
information
gaps.
(
Natural
Res.
Def.
Council
v.
EPA,
824
F.
2D
1146,
1152­
53
(
D.
C.
Cir.
1987)).
The
commenter
stated
that,
in
regulatory
practice,
the
ample
margin
of
safety
analysis
consists
of
a
consideration
of
the
NOEL
for
a
pollutant
and
the
subsequent
application
of
factors
to
account
for
scientific
uncertainty
surrounding
that
safe
level
of
exposure.
According
to
the
commenter,
this
is
the
approach
called
for
by
the
Senate
Report
accompanying
the
1990
CAA
Amendments
(
S.
Rep.
No.
228,
101st
Cong.
Sess.
171
(
1990)),
and
this
is
exactly
what
is
done
in
deriving
an
RfC
302
or
similar
inhalation
health
benchmark.
The
commenter
stated
that
EPA's
derivation
of
the
RfC
contains
multiple
layers
of
conservatism
to
account
for
scientific
uncertainty.
The
commenter
believed
that
RfC
values
and
similar
inhalation
health
benchmarks
already
incorporate
sufficient
uncertainty
factors
to
fulfill
or
exceed
the
ample
margin
of
safety
mandate
of
CAA
sections
112(
d)(
4)
and
112(
c)(
9).

Response:
Today's
final
PCWP
rule
will
utilize
CAA
section
112(
c)(
9)
rather
than
CAA
section
112(
d)(
4).
We
agree
that
the
CAA
does
not
define
"
ample
margin
of
safety"

explicitly.
The
CAA
does,
however,
in
section
112(
f)

explicitly
recognize
our
Federal
Register
notice
of
September
14,
1989,
which
described
our
interpretation
of
ample
margin
of
safety
in
the
case
of
linear
carcinogens,

and
our
approach
to
implementing
that
interpretation.
While
the
first
step
identifies
the
presumptive
limit
on
maximum
individual
risk,
the
second
step
of
that
2­
step
approach
describes
the
setting
of
the
risk­
based
standard
at
a
level
that
provides
an
ample
margin
of
safety,
in
consideration
of
a
number
of
factors.
As
we
noted
in
the
1989
notice,
the
objective
in
protecting
public
health
with
an
ample
margin
of
safety
under
CAA
section
112
is
to
ensure
an
individual
lifetime
risk
level
no
higher
than
one
in
a
million
to
the
greatest
number
of
persons
possible,
and
to
limit
to
no
303
higher
than
one
in
ten
thousand
the
estimated
risk
for
a
person
living
near
a
plant
if
they
were
exposed
for
70
years.

In
assessing
risk
or
hazard
of
nonlinear
effects,
we
use
the
RfC
or
comparable
value.
This
value
represents
an
estimate
(
with
uncertainty
spanning
perhaps
an
order
of
magnitude)
of
a
continuous
inhalation
exposure
to
the
human
population
(
including
sensitive
subgroups)
that
is
likely
to
be
without
an
appreciable
risk
of
deleterious
non­
cancer
effects
during
a
lifetime.
The
RfC
values
and
comparable
values
are
derived
from
assessments
of
pertinent
toxicological
information
to
identify
the
lowest
point
of
departure
(
in
human
equivalent
terms)
from
the
experimental
data
that
is
also
representative
of
the
threshold
region
(
the
region
where
toxicity
is
apparent
from
the
available
data)
for
the
array
of
toxicity
data
for
that
chemical.
The
objective
is
to
select
a
prominent
toxic
effect
that
is
pertinent
to
the
chemical's
key
mechanism
or
mode
of
action.

This
approach
is
based,
in
part,
on
the
assumption
that
if
the
critical
toxic
effect
is
prevented,
then
all
toxic
effects
are
prevented.
The
RfC
is
derived
from
the
point
of
departure
(
POD)
(
in
terms
of
human
equivalent
exposure)
for
the
critical
effect
by
consistent
application
of
uncertainty
factors,
which
are
to
account
for
recognized
uncertainties
in
the
extrapolations
from
the
experimental
data
conditions
304
2U.
S.
Environmental
Protection
Agency.
1994.
Methods
for
Derivation
of
Inhalation
Reference
Concentrations
and
Application
of
Inhalation
Dosimetry.
Office
of
Research
and
Development.
EPA/
600/
8­
90/
066F.
(
EPA
1994)
to
an
estimate
appropriate
to
the
assumed
human
scenario.
2
In
considering
the
extrapolation
of
the
ample
margin
of
safety
objective
described
for
linear
cancer
risk
to
the
management
of
risk
for
nonlinear
effects
under
CAA
section
112(
c)(
9)
(
i.
e.,
in
decisions
to
delist
a
subcategory
from
any
further
regulatory
action),
we
consider
exposures
relative
to
the
RfC
or
comparable
values
for
all
of
the
emitted
HAP,
with
specific
attention
to
those
affecting
a
similar
physiological
target
organ
or
system,
and
the
contribution
of
co­
located
sources
of
the
same
HAP.

Comment:
One
commenter
stated
that
the
uncertainty
factors
used
in
deriving
the
wood
products
HAP
health
benchmarks
are
particularly
large.
The
unit
risk
factors
for
acetaldehyde
and
formaldehyde
were
calculated
using
the
linear
multi­
stage
model,
which
assumes
a
linear
relationship
between
cancer
incidence
and
exposure
to
the
pollutant
at
low
doses.
According
to
the
commenter,
the
available
data
on
acetaldehyde
and
formaldehyde
strongly
suggest
that
this
assumption
is
incorrect
and
overly
conservative.

The
commenter
pointed
out
that
EPA's
health
assessment
of
acrolein
is
two
to
three
times
more
conservative
than
305
CalEPA's,
even
though
both
are
intended
to
protect
sensitive
individuals
from
any
adverse
effects
following
a
lifetime
of
exposure.
The
commenter
stated
that
EPA
has
developed
an
extremely
conservative
RfC
for
acrolein.
The
commenter
argued
that
adopting
a
HI
of
0.2
would
add
another
five­
fold
safety
factor
to
this
already
extremely
conservative
RfC.

The
commenter
noted
that
acrolein
is
the
HAP
of
greatest
importance
in
determining
risk
from
PCWP
facilities.

Response:
The
dose­
response
values
used
to
determine
the
criteria
for
defining
the
low­
risk
subcategory
are
drawn
from
IRIS,
as
well
as
from
certain
alternative
sources.
The
IRIS
process
contains
internal
and
external
peer
review
steps
and
represents
EPA
consensus
values.
When
adequate
toxicity
information
is
not
available
in
IRIS,
we
consult
other
sources
in
a
default
hierarchy
that
recognizes
the
desirability
of
these
qualities
in
ensuring
that
we
have
consistent
and
scientifically
sound
assessments.
In
the
case
of
acrolein,
specifically
mentioned
by
the
commenter,

consultation
of
other
sources
was
not
necessary
because
the
acrolein
assessment
was
completed
within
the
past
9
months
and
represents
current
scientific
knowledge.
In
those
cases
(
e.
g.,
formaldehyde),
where
the
IRIS
assessment
substantially
lags
the
current
scientific
knowledge,
we
consider
alternative
credible
and
readily
available
assessments.
As
pointed
out
elsewhere
in
this
section,
the
306
3Ibid.
RfC
values
or
comparable
values
have
been
derived
with
the
incorporation
of
uncertainty
factors.
The
uncertainty
factors
are
to
account
for
recognized
uncertainties
in
the
extrapolations
from
the
experimental
data
conditions
pertaining
to
the
chemical's
particular
toxicological
data
set
to
an
estimate
appropriate
to
the
assumed
human
scenario.
3
The
size
variation
of
the
uncertainty
factors
across
RfC
values
reflects
the
size
variation
of
the
uncertainties
associated
with
that
extrapolation.

Comment:
One
commenter
stated
that
the
combination
of
conservative
air
dispersion
modeling
techniques
and
a
conservative
human
health
benchmark
ensure
that,
where
a
source
meets
the
requirements
for
a
risk­
based
compliance
option,
human
health
will
be
protected
with
an
ample
margin
of
safety.
The
commenter
pointed
out
that,
for
most
individuals
in
the
general
population,
actual
exposures
likely
are
one
or
more
orders
of
magnitude
below
the
maximum
exposures
predicted
by
the
tiered
modeling
approach.
The
commenter
noted
that
EPA's
tiered
modeling
methodology
is
designed
to
identify
the
highest
annual
property
line
or
off­
site
concentrations
that
might
occur
around
each
facility
(
as
opposed
to
actual
population
exposure).
The
tiered
approach
models
exposures
of
a
maximally
exposed
individual
(
MEI)
and
incorporates
a
number
of
conservative
307
assumptions.
According
to
the
commenter,
actual
average
concentrations
are
likely
to
be
much
lower.
The
commenter
argued
that,
even
if
the
modeled
concentrations
were
reflective
of
continuous
average
concentrations,
it
is
highly
unlikely
that
any
individual
would
actually
be
exposed
to
such
concentrations
for
a
lifetime.
The
commenter
noted
that
the
Presidential/
Congressional
Commission
on
Risk
Assessment
and
Risk
Management
concluded
that
the
conservatism
inherent
in
use
of
the
MEI
was
often
so
unrealistic
that
its
use
impaired
the
scientific
credibility
of
health
risk
assessment.

Response:
We
discussed
a
tiered
analytical
approach
in
the
preamble
to
the
proposed
rule,
beginning
with
relatively
simple
lookup
tables
and
followed
by
increasingly
more
sitespecific
but
more
resource
intensive
tiers
of
analysis,
with
each
tier
being
more
refined
and
less
conservative.
In
today's
final
rule,
we
are
setting
forth
two
options,
as
specified
in
Appendix
B
to
subpart
DDDD.
In
the
first
option,
facilities
can
qualify
for
inclusion
in
the
delisted
subcategory
by
using
site­
specific
emissions
test
data
and
look­
up
tables
that
were
developed
using
health­
protective
input
parameters.
As
a
second
option,
facilities
may
choose
to
use
a
more
refined
site­
specific
risk
assessment.
A
more
refined
analysis
requires
more
effort,
but
produces
results
that
are
less
likely
to
overestimate
risk.
308
Comment:
One
commenter
noted
that
the
regulatory
requirements
in
the
proposed
rule
focused
on
six
HAP
that
are
emitted
from
PCWP
facilities:
acrolein,
acetaldehyde,

formaldehyde,
methanol,
phenol,
and
propionaldehyde.
Those
HAP
represent
96
percent
of
the
emissions
from
PCWP
facilities.
The
commenter
believes
that
any
risk­
based
compliance
mechanisms
may
reasonably
be
limited
to
consideration
of
the
risks
from
these
six
HAP.
The
commenter
noted
that
EPA's
preliminary
risk
analysis
conducted
prior
to
proposal
narrowed
the
list
of
HAP
emitted
from
PCWP
facilities
to
include
the
following:
acrolein,

acetaldehyde,
formaldehyde,
methanol,
phenol,
benzene,

methylene
chloride,
and
manganese.
The
commenter
referred
to
the
results
of
their
sensitivity
analysis,
which
was
conducted
based
on
the
data
used
in
EPA's
pre­
proposal
risk
analysis.
The
analysis
evaluated
the
impact
of
increasing
or
decreasing
facility
emissions
by
30
percent,
using
different
health
benchmarks
than
those
identified
in
EPA's
analysis,
and
conducting
the
risk
assessment
with
the
six
HAP
targeted
in
the
proposed
rule
versus
the
additional
HAP
identified
by
EPA.
The
commenter's
sensitivity
analysis
showed
that
formaldehyde
and
acetaldehyde
made
up
the
bulk
of
the
cancer
risk,
while
benzene
and
methylene
chloride
had
little
or
no
influence
on
cancer
risk,
depending
on
the
scenario
considered.
Under
all
scenarios,
acrolein
309
contributed
the
most
non­
cancer
risk.
The
remainder
of
the
non­
cancer
risk
was
divided
between
acetaldehyde,

formaldehyde
and
manganese,
with
manganese
contributing
between
5.6
and
12.2
percent
of
the
non­
cancer
risk,

depending
on
the
scenario.
Under
all
scenarios,
methanol,

benzene,
methylene
chloride
and
phenol
did
not
contribute
at
all
to
the
non­
cancer
risk
from
wood
products
facilities
(
with
one
exception,
where
the
phenol
risk
contribution
was
0.1
percent).
Based
on
these
results,
the
commenter
stated
that
there
appeared
to
be
little
reason
to
include
evaluation
of
methylene
chloride
or
benzene
in
the
risk­
based
compliance
option.
However,
the
commenter
stated
that
it
may
be
reasonable
to
take
an
extremely
conservative
approach
and
include
evaluation
of
manganese
in
the
risk­
based
compliance
mechanisms.

Response:
We
agree
that
it
is
appropriate
to
limit
the
number
of
HAP
that
must
be
included
in
PCWP
facility
lowrisk
demonstrations
to
only
those
HAP
that
may
possibly
result
in
meaningful
contributions
to
the
facility
risk.

However,
we
disagree
that
limiting
the
HAP
included
in
the
low­
risk
demonstration
to
the
six
HAP
defined
as
total
HAP
in
subpart
DDDD
of
40
CFR
part
63
(
acrolein,
acetaldehyde,

formaldehyde,
methanol,
phenol,
and
propionaldehyde)
is
appropriate.
We
identified
the
most
prevalent
HAP
based
on
mass
emitted
for
purposes
of
developing
MACT
compliance
310
options
because
MACT
is
technology­
based
(
i.
e.,
the
same
technology
that
reduces
emissions
of
the
six
HAP
also
reduces
emissions
of
other
organic
HAP).
As
discussed
earlier
in
this
preamble,
the
six
HAP
defined
as
total
HAP
in
subpart
DDDD
of
40
CFR
part
63
are
the
HAP
that
are
most
often
emitted
in
detectable
amounts
from
the
most
PCWP
process
units,
and
these
HAP
make
up
96
percent
of
the
mass
of
nationwide
HAP
emissions
from
the
PCWP
industry.

However,
the
risk
associated
with
emissions
of
HAP
are
dependent
on
the
mass
emitted
and
the
relative
toxicity
of
each
HAP.
Thus,
the
HAP
emitted
in
the
greatest
mass
may
not
result
in
the
most
risk
because
the
HAP
may
not
be
as
potent
as
other
HAP
emitted
in
lower
mass.
For
example,

methanol
is
the
HAP
emitted
from
the
PCWP
industry
in
the
greatest
mass,
but
because
methanol
is
not
as
toxic
as
other
HAP
emitted
(
e.
g,
formaldehyde,
certain
HAP
metals),
it
does
not
result
in
as
much
risk
as
do
other
HAP.
To
ensure
protection
of
public
health,
all
HAP
must
be
considered
when
determining
which
facilities
are
low
risk.
Simply
importing
the
surrogate
pollutants
that
are
reasonably
used
for
MACT
purposes
into
the
risk
assessment
context
is
not
appropriate,
as
surrogacy
for
MACT
is
based
on
factors
and
considerations
relating
to
technological
control
capabilities
and
not
on
how
surrogate
pollutants
might
indicate
how
non­
surrogates
affect
risks
to
human
health
and
311
the
environment.
For
example,
just
because
in
many
cases
particulate
matter
is
a
useful
surrogate
for
measuring
the
control
efficiency
of
devices
used
to
capture
non­
mercury
HAP
metals,
that
fact
is
unrelated
to
what
risks
the
HAP
metals
may
present
individually
or
collectively,
as
HAP
metals
apart
from
the
risks
they
pose
as
being
particulates.

The
commenter
is
correct
in
that
our
preliminary
risk
analysis
conducted
prior
to
proposal
narrowed
the
list
of
HAP
emitted
from
PCWP
facilities.
We
acknowledge
receipt
of
the
commenter's
sensitivity
analysis
based
on
the
data
used
in
our
pre­
proposal
risk
analysis.
Following
proposal,
we
conducted
a
more
detailed
risk
analysis
to
evaluate
the
merits
of
including
a
low­
risk
subcategory
in
the
final
PCWP
rule.
This
post­
proposal
analysis
considered
emissions
of
more
than
30
HAP
emitted
from
the
PCWP
source
category.

Many
of
these
HAP
are
only
emitted
in
minute
amounts
that
have
been
detected
from
a
small
number
of
PCWP
process
units.
Nevertheless,
we
included
them
in
our
risk
analysis
to
determine
their
contribution
to
PCWP
facility
risk.
We
reviewed
the
toxicity
values
for
each
HAP
and
the
mass
of
each
emitted
from
PCWP
facilities
to
determine
if
it
would
be
appropriate
to
narrow
the
list
of
HAP
that
PCWP
facilities
must
consider
in
their
low­
risk
demonstrations.

Based
on
our
review,
we
determined
that
95
percent
of
the
cancer
risk
at
PCWP
facilities
is
accounted
for
by
the
312
following
HAP:
acetaldehyde,
benzene,
arsenic,
beryllium,

cadmium,
hexavalent
chromium,
lead,
nickel
subsulfide,
and
formaldehyde.
We
also
determined
that
95
percent
of
the
non­
cancer
risk
at
PCWP
facilities
is
accounted
for
by
the
following
HAP:
acetaldehyde,
acrolein,
formaldehyde,

phenol,
MDI,
arsenic,
cadmium,
and
manganese.
We
feel
that
inclusion
of
these
HAP
in
a
demonstration
of
eligibility
of
the
low­
risk
PCWP
subcategory
is
appropriate.
Limiting
the
list
of
HAP
that
must
be
included
in
the
low­
risk
demonstration
to
13
HAP
minimizes
emissions
testing
costs,

while
ensuring
that
the
HAP
that
drive
the
risk
at
PCWP
facilities
are
accounted
for
on
a
site­
specific
basis.

Background,
multipathway,
and
ecological
exposures
Comment:
Two
commenters
argued
that
multipathway
exposures
should
not
be
considered
for
PCWP
facilities.
One
commenter
stated
that,
because
the
HAP
emitted
from
the
PCWP
source
category
are
not
bioaccumulative,
it
is
unnecessary
to
consider
multipathway
exposures.
The
other
commenter
stated
that
there
is
no
policy
basis
for
considering
multipathway
exposures
because
U.
S.
Government
surveys
and
regulatory
actions
demonstrate
that
non­
inhalation
exposure
to
the
six
HAP
emitted
by
wood
products
facilities
is
insignificant.
The
commenter
provided
rationale
for
the
conclusion
that
dietary
and
drinking
water
exposures
to
the
six
HAP
are
not
significant.
Because
the
six
HAP
primarily
313
emitted
from
the
PCWP
source
category
(
acetaldehyde,

acrolein,
and
formaldehyde,
methanol,
phenol,
and
propionaldehyde)
do
not
exhibit
bioaccumulative
characteristics,
the
commenter
considered
it
unnecessary
to
consider
multipathway
exposures.

Three
commenters
argued
that
multipathway
exposures
should
be
considered
for
PCWP
facilities.
One
commenter
stated
that,
when
persistent
biological
toxicant
or
metal
emissions
are
significant,
ingestion
and
other
pathways
should
be
considered
in
the
risk
screening.
Another
commenter
stated
that
the
concentration­
based
applicability
threshold
approach
in
the
proposed
PCWP
rule
does
not
address
non­
inhalation
exposures
or
adverse
effects
on
the
environment.
The
third
commenter
stated
that
CAA
section
112(
d)(
4)
requires
EPA
to
consider
all
possible
ways
that
a
pollutant
could
affect
human
health
or
the
environment
because
it
refers
to
pollutants
"
for
which
a
health
threshold
has
been
established,"
i.
e.,
pollutants
that
have
no
adverse
health
or
environmental
effects.
(
See
5
Legislative
History
at
8511.)
According
to
the
commenter,

EPA
has
recognized
repeatedly
in
the
past
that
many
of
the
pollutants
emitted
by
the
source
category
are
re­
deposited
from
the
atmosphere
and
then
contaminate
soil
and
water
for
long
periods
of
time.
The
commenter
added
that
these
pollutants
bioaccumulate
in
wildlife
and
food
sources,
314
poisoning
people
and
animals
alike.
(
See,
e.
g.,
64
FR
52828,
53014
(
September
30,
1999);
64
FR
31898,
31908­
31909
(
June
14,
1999);
63
FR
14182,
14193
(
March
28,
1998);
61
FR
17358,
17478
(
April
19,
1996)
(
due
to
bioaccumulation,

mercury
levels
may
be
10,000,000
times
higher
in
fish
than
in
water
those
fish
inhabit).)
The
commenter
concluded
that,
to
evaluate
whether
a
pollutant
is
a
threshold
pollutant
and
what
its
health
threshold
and
ample
margin
of
safety
must
be,
EPA
must
consider
all
the
potential
health
and
environmental
effects
of
deposition,
persistence
and
bioaccumulation
of
that
pollutant.
The
commenter
argued
that
EPA
would
contravene
section
112(
d)(
4)
by
considering
only
health
effects
caused
by
inhalation.

Response:
We
agree
with
the
last
commenter
that
CAA
section
112(
d)(
4)
requires
us
to
evaluate
ecological
and
multimedia
human
exposures
when
determining
the
applicability
of
section
112(
d)(
4).
This
rule
is
relying
not
on
CAA
section
112(
d)(
4),
but
on
section
112(
c)(
9),

which
also
hasstates
that
as
similar
requirements
for
determining
risk
applicability.
Although
the
commenter
refers
to
CAA
section
112(
d)(
4),
their
concern
is
addressed
in
section
112(
c)(
9)
also.
We
agree
that
potential
ecological
effects
and
multimedia
human
exposures
need
to
be
considered.
We
have
conducted
an
ecological
assessment
and
a
multipathway
exposure
assessment
on
those
HAP
emitted
from
315
PCWP
facilities
(
including
HAP
not
among
the
six
mentioned
by
one
commenter)
that
we
have
identified
as
having
the
potential
for
persisting
and
bioaccumulating
in
the
environment.
From
this
analysis
we
determined
that
adverse
ecological
effects
and/
or
multimedia
health
effects
are
unlikely
from
PCWP
facilities.
Therefore,
PCWP
facilities
attempting
to
demonstrate
their
low­
risk
status
will
not
be
required
to
include
an
ecological
assessment
or
a
multimedia
assessment.

Comment:
Several
commenters
stated
that
there
is
no
legal
or
policy
basis
for
EPA
to
consider
background
or
multipathway
(
non­
inhalation)
exposures.
The
commenters
claimed
that
CAA
section
112(
d)
requires
that
MACT
standards
be
based
only
on
emissions
from
the
MACT­
regulated
portion
of
the
facility;
it
does
not
give
EPA
the
authority
to
consider
existing
background
levels.
One
commenter
asserted
that
CAA
section
112
can
be
distinguished
from
other
statutory
provisions,
both
in
the
CAA
and
in
other
environmental
legislation,
where
EPA
has
clearly
been
given
authority
to
consider
background
sources.

Another
commenter
argued
that
the
CAA's
legislative
history
does
not
support
a
requirement
to
consider
other
exposures.
The
commenter
also
claimed
that
the
statutory
provisions
on
which
EPA
would
rely
to
implement
the
risk­
based
mechanisms
(
i.
e.,
CAA
section
112(
d)(
4),
CAA
section
112(
c)(
9)(
B),
or
EPA's
de
minimis
authority)
316
exclusively
focus
on
the
emissions
from
the
source
in
making
regulatory
decisions.
According
to
the
commenter,
EPA
has
existing
regulatory
programs
(
e.
g.,
for
mobile
and
area
sources
(
Urban
Air
Toxics
Strategy))
in
place
to
address
HAP
emissions
from
other
sources.

The
commenter
argued
that
over­
control
of
PCWP
facilities
is
unjustified
because
PCWP
facilities
account
for
very
small
proportions
of
HAP
emissions
nationwideBless
than
1.75
percent
of
acetaldehyde,
1.7
percent
of
acrolein,

and
1
percent
of
formaldehyde
emissions,
according
to
their
industry­
sponsored
risk
assessment.
Given
these
results,

the
commenter
concluded
that
PCWP
facilities
cannot
reasonably
be
considered
to
contribute
meaningfully
to
background
concentrations.

The
commenter
stated
that
CAA
section
112(
d)(
4)
does
not
require
the
complete
absence
of
a
hazard
and
noted
that
all
substances
have
adverse
effects
on
health
and
the
environment
at
sufficiently
high
doses.
Accordingly,

section
112(
d)(
4)
incorporates
the
concept
of
exposure,
and
the
Senate
Report
made
clear
that
the
provision
was
applicable
in
instances
where
"
it
is
possible
to
establish
a
NOEL
below
which
human
exposure
is
presumably
safe."
In
addition,
delisting
criteria
and
the
so­
called
trigger
component
of
the
residual
risk
provision
focus
exclusively
on
emissions
and
whether
the
risk
posed
by
any
source
in
the
category,
by
itself,
exceeds
one
in
a
million
cancer
risk,
317
the
commenter
claimed.

Two
commenters
opposed
the
use
of
available
data
on
background
concentrations
and
facility­
specific
measurement
of
background
concentrations
to
determine
the
extent
of
exposures
from
other
sources,
arguing
that
the
CAA
and
sound
public
policy
warrant
a
focus
exclusively
on
the
emissions
from
the
source
category
at
hand
when
evaluating
the
applicability
of
a
risk­
based
compliance
option.
Because
a
HI
of
1.0
(
or
higher)
is
amply
protective
of
public
health
and
is
warranted
under
EPA's
statutory
mandate,
the
commenters
stated
that
consideration
of
background
concentration
is
not
appropriate.

Response:
We
disagree
that
we
do
not
have
the
authority
to
consider
background
andare
not
considering
background
emissions
in
this
section
112(
c)(
9)
delisting
of
the
lowrisk
PCWP
subcategory.
Rather,
as
we
indicated
in
the
Residual
Risk
Report
to
Congress,
the
Agency
intends
to
assess
such
facility­
wide
emissions
of
HAP
in
future
residual
risk
actions
under
section
112(
f)(
2),
as
appropriate
and
to
the
extent
it
is
reasonable
and
appropriate
to
do
so.
See
also,
54
Fed.
Reg.
at
38,059
(
Sept.
14,
1989)
(
benzene
NESHAP).

Regarding
multipathway
exposures.
Section
112(
f)(
1)(
C)
of
the
CAA
explicitly
directed
EPA
to
investigate
and
report
on
risks
presented
by
background
concentrations
of
HAP
in
its
residual
risk
report
to
318
Congress,
which
presumably
it
would
not
have
done
if
it
assumed
EPA
was
to
ignore
such
concentrations
in
the
subsequent
risk­
based
regulatory
process.
In
fact,
EPA
explained
in
the
Report
that
we
generally
would
consider
background
concentrations
as
part
of
residual
risk
decisionmaking
(
see
Report
at
pp.
92­
93).
In
our
1989
benzene
NESHAP
rulemaking,
which
Congress
explicitly
endorsed
in
CAA
section
112(
f)(
2)(
B),
we
stated
that
we
agree
that
exposures
to
background
concentrations
and
multiple
sources
of
a
pollutant
may
be
considered
to
the
extent
it
is
practical
and
reasonable
to
do
so.
See
54
FR
at
38059
(
Sept.
14,

1989).
Moreover,
the
legislative
history
to
the
1990
CAA
Amendments
explicitly
states
that
all
risk
from
colocated
sources
at
major
sources
subject
to
MACT
must
be
considered,

albeit
not
necessarily
at
the
same
time.
See
1
Legis.
Hist.

at
868­
69
(
statement
of
Sen.
Durenberger,
floor
manager
of
air
toxics
provisions,
during
the
debate
on
the
Conference
Report).
Finally,
we
disagree
with
assertions
that
background
levels
are
already
accounted
for
by
uncertainty
factors
built
into
RfC/
RfDs,
as
having
no
factual
basis.

Commenters
presumed
the
use
of
CAA
section
112(
d)(
4)
for
the
PCWP
delisting.
However,
we
are
using
CAA
section
112(
c)(
9)

and
not
CAA
section
112(
d)(
4).
For
delisting
purposes,
we
are
required
to
make
not
only
a
determination
of
a
level
which
is
safe
to
the
individual
exposed
to
the
maximum
concentration
but
also
a
determination
whether
that
319
individual
and
the
general
population
are
protected
to
a
level
that
provides
a
clear
ample
margin
of
safety
in
order
to
have
the
confidence
that
a
decision
to
exclude
these
facilities
from
regulation
(
including
residual
risk)
may
be
made.
These
are
the
same
requirements
that
are
found
as
the
trigger
in
the
residual
risk
program
because
all
decisions
of
this
type
under
section
112
of
the
CAA
(
sections
112(
c)(
9),
d(
4),
and
f(
2))
should
be
consistent
and
based
on
the
same
risk
approaches.

We
agree
that
imposition
of
costly
regulatory
requirements
on
sources
that
do
not
contribute
meaningfully
to
background
exposures
is
not
beneficial.
However,

regarding
the
results
of
the
industry­
sponsored
wood
products
MACT
risk
assessment,
we
note
that
CAA
section
112(
c)(
9)
requires
the
evaluation
of
all
facility
level
emission
points,
all
HAP,
and
all
pathways
in
order
to
satisfy
that
the
emissions
from
these
facilities
provide
protection
of
the
population
with
an
ample
margin
of
safety.

The,
the
industry's
wood
products
MACT
risk
assessment
does
not
address
HAP
emitted
from
PCWP
facilitiesaffected
sources
that
have
the
potential
to
bioaccumulate
and
persist
in
the
environment
(
e.
g.,
lead,
cadmium,
and
mercury).
We
conducted
an
exposure
assessment
for
these
HAP
to
determine
exposure
from
ingestion
as
well
as
inhalation.
The
maximum
multipathway
risks
were
considerably
lower
than
the
predicted
maximum
inhalation
risks
from
the
PCWP
source
320
category.
Therefore,
PCWP
facilitiesaffected
sources
are
not
required
to
conduct
site­
specific
multipathway
risk
assessments
as
part
of
their
low­
risklow
risk
demonstrations.
The
look­
up
tables
included
in
appendix
B
to
subpart
DDDD
were
developed
using
conservative
input
parameters
to
ensure
that
facilitiesaffected
sources
qualifying
for
the
low­
risk
subcategory
based
on
the
look­
up
tables
would
not
pose
a
risk
to
multimedia
endpointsvia
multipathway
exposures.

As
discussed
elsewhere
in
this
preamble,
for
today's
final
PCWP
rule,
we
consider
that
an
HI
limit
of
1.0
provides
an
ample
margin
of
safety
for
protecting
public
health
under
CAA
section
112(
c)(
9)
for
this
delisting
of
low­
risk
PCWP
affected
sources.
The
RfCs
that
are
used
to
calculate
the
HI
are
developed
to
protect
sensitive
subgroups
and
to
account
for
scientific
uncertainties
and
their
inclusion
helps
ensure,
ensuring
that
usingthe
use
of
an
HI
limit
of
1.0
provides
an
ample
margin
of
safety.
We
believeconclude
that
an
HI
limit
of
1.0
is
appropriate
for
the
section
112(
c)(
9)
demonstrations
for
the
PCWP
source
category
that
are
described
in
today''
s
action.
In
future
risk­­
based
actions
for
this
and
other
source
categories
(
e.
g.,
residual
risk
rulemakings
under
CAA
section
112(
f))

we
may
identify
factors
on
a
case­
by­
casecase­
by­
case
basis
that
would
lead
us
to
conclude
that
HI
limits
other
than
1.0
would
be
more
appropriate
for
those
other
actions.
321
The
look­
up
tables
included
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
were
developed
based
on
an
HI
of
1.0.

For
site­
specific
chronic
inhalation
risk
assessments,

facilitiesaffected
sources
are
required
to
ensure
that
the
TOSHI
is
less
thantheir
TOSHI
(
or,
alternately,
a
sitespecific
set
of
hazard
indices
based
on
mechanistic
data
or
dose­
response
data
for
their
HAP
mixture)
are
less
than
or
equal
to
a
value
of
1.0.
Facilities
are
required
to
determine
TOSHI
forThese
assessments
focus
on
respiratory
effects
and
CNS
effects,
because
based
on
our
analysis,

noncancer
impacts
were
dominated
primarily
by
impacts
on
these
systems.
Other
target
organs
or
systems
were
found
to
be
negligibly
impacted.

Comment:
One
commenter
stated
that
EPA
had
provided
inadequate
discussion
of
how
environmental
risks
would
be
evaluated.
The
commenter
added
that
the
CAA
requires
EPA
consider
the
environment
as
well
as
public
health,
and
that,

at
a
minimum,
a
facility
would
be
required
to
conduct
an
assessment
based
on
EPA's
1998
Guidelines
for
Ecosystem
Assessment.
The
commenter
referred
EPA
to
appendix
A
of
"
Generic
Assessment
for
Endpoints
for
Ecological
Risk
Assessment"
for
a
detailed
discussion
on
the
legal
basis
from
"
such
statutes
as
the
CAA...
that
require
EPA
to
consider
and
protect
organism­
level
attributes
or
various
taxa
including
fish,
birds,
and
plants
and
more
generally,
322
animals,
wildlife,
aquatic
life,
and
living
things."

Another
commenter
cited
an
analysis
they
commissioned
that
showed
it
to
be
highly
unlikely
that
emissions
from
PCWP
facilities
would
pose
a
hazard
to
ecological
receptors
at
levels
that
are
protective
of
human
health.
Thus,

concern
over
ecological
receptors
would
not
provide
a
valid
basis
for
reducing
the
HI
below
1.0.

Response:
An
ecological
assessment
is
required
under
sections
112(
d)(
4),
(
c)(
9),
and
(
f)(
2)
of
the
CAA
regarding
the
presence
or
absence
of
"
adverse
environmental
effects"

as
that
term
is
defined
in
CAA
section
112(
a)(
7).

Therefore,
delisting
under
section
112(
c)(
9)
requires
consideration
of
ecological
effects.
The
look­
up
tables
developed
for
today's
final
PCWP
rule
are
intended
to
accommodate
enough
conservatism
that
any
facility
qualifying
for
inclusion
in
the
delisted
subcategory
using
them
will
qualify
based
on
all
endpoints,
including
ecological
endpoints.
Based
on
our
analysis
of
ecological
effects
(
in
the
supporting
information
for
the
final
rule),
we
feel
it
is
unlikely
that
PCWP
facilities
would
pose
any
significant
ecological
risks
to
any
actual
ecosystem
or
ecosystems
nearby.
We
also
conclude,
given
the
low
impacts
from
the
hypothetical
worst­
case
scenario
investigated,
that
it
is
unlikely
that
any
potentially­
exposed
threatened
or
endangered
species
would
be
adversely
affected
by
HAP
323
emissions
from
these
facilities.
Therefore,
PCWP
facilities
are
not
required
to
conduct
site­
specific
ecological
risk
assessments
as
part
of
their
low­
risk
demonstration.

Assuming
the
assessment
referenced
by
the
first
commenter
included
only
the
six
HAP
listed
in
subpart
DDDD
of
40
CFR
part
63,
we
disagree
that
these
six
HAP
should
be
the
sole
focus
of
an
ecological
assessment.
It
is
not
clear
from
the
comment
whether
the
commenter
is
suggesting
that
we
might
consider
lowering
the
human
health
HI
values
to
below
1.0
in
order
to
reflect
ecological
concerns
or
whether
they
are
suggesting
that
an
ecological
HI
value
should
not
be
reduced
below
1.0.
In
the
former
case,
that
is
not
done.

Human
health
and
ecological
assessments
are
independent
assessments
with
their
own
risk
management
criteria.

Hazard
index
Comment:
Two
commenters
stated
that
hazard
quotients
(
HQ)
for
chemical
mixes
should
not
be
summed
to
determine
the
HI
unless
the
primary
effects
are
on
the
same
organ
by
the
same
mechanism;
otherwise
the
risk
would
be
overestimated.
One
commenter
stated
that
CAA
section
112(
d)(
4)
refers
to
threshold
pollutants,
with
each
health
threshold
augmented
by
an
ample
margin
of
safety.
These
ample
margin
of
safety
values
are
already
incorporated
into
RfC
values.
The
risk
criteria
applied
are
confined
to
the
effects
upon
which
the
RfC
is
based,
which
reflect
the
most
324
sensitive
target
organ.
According
to
the
commenter,
a
decision
to
add
risk
posed
by
chemicals
that
affect
the
same
target
organ
but
have
unknown
mechanisms
of
action
represents
an
unnecessarily
conservative
assumption
that
would
tend
to
inflate
the
final
risk
estimate.

The
commenters
also
noted
that,
according
to
the
National
Research
Council
and
the
Presidential/
Congressional
Commission
on
Risk
Assessment
and
Risk
Management,

additivity
at
low
doses
is
more
likely
to
overestimate
than
to
underestimate
total
risk.
As
stated
in
the
Commission's
1997
Final
Report:
"
When
the
individual
components
of
a
chemical
mixture
exhibit
different
kinds
of
toxicity
or
have
different
biological
mechanisms
of
toxicity,
they
do
not
interact
 
they
act
independently
at
low
doses.
In
that
case,

the
dose­
response
relationships
for
each
chemical
should
be
considered
independently...
[
By
contrast]
studies
in
which
similar
chemicals
with
similar
mechanisms
and
target
were
administered
simultaneously
indicate
that
antagonism
is
the
usual
outcome..."(
Falk
and
Kotin
1964,
Schmal
et
al.
1977)

Response:
Our
recommended
approach
for
assessing
risks
from
exposure
to
a
mixture
of
pollutants
is
to
utilize
a
325
4U.
S.
Environmental
Protection
Agency.
1986.
Guidance
for
Conducting
Health
Risk
Assessment
of
Chemical
Mixtures.
Risk
Assessment
Forum,
Washington,
DC.
EPA/
630/
R­
98/
002;
available
at
http://
cfpub.
epa.
gov/
ncea/
raf/
recordisplay.
cfm?
deid=
20533.
(
EPA
1986)

5U.
S.
Environmental
Protection
Agency.
2000.
Supplementary
Guidance
for
Conducting
Health
Risk
Assessment
of
Chemical
Mixtures.
Office
of
Research
and
Development.
EPA/
630/
R­
00/
002
(
EPA
2000)

6Ibid.
dose­
response
assessment
developed
for
that
mixture.
4,5
There
are
few
mixtures
(
e.
g.,
coke
oven
emissions),
however,

for
which
such
assessments
are
available.
When
mixturespecific
dose­
response
assessments
are
not
available,
a
component­
by­
component
approach
is
recommended.
The
method
for
component
data
depends
on
a
judgment
of
toxicologic
similarity
among
components.
The
specific
term
toxicologic
similarity
represents
a
general
knowledge
about
the
action
of
a
chemical
or
a
mixture
and
can
be
expressed
in
broad
terms
such
as
at
the
target
organ
level
in
the
body.
In
our
guidance,
assumptions
about
toxicologic
similarity
are
made
in
order
to
choose
among
risk
assessment
methods.
In
general,
we
assume
a
similar
mode
of
action
across
mixtures
or
mixture
components
and,
in
some
cases,
this
requirement
may
be
relaxed
to
require
that
these
chemicals
act
only
on
the
same
target
organ.
6
The
primary
method
for
component­
based
risk
assessment
of
toxicologically
similar
chemicals
is
the
HI,
which
is
derived
from
dose
addition.
In
our
guidance,
dose
addition
326
7Ibid.
is
interpreted
as
simple
similar
action,
where
the
component
chemicals
act
as
if
they
are
dilutions
or
concentrations
of
each
other
differing
only
in
relative
toxicity.
Dose
additivity
may
not
hold
for
all
toxic
effects.
Furthermore,

the
relative
toxic
potency
between
chemicals
may
differ
from
different
types
of
toxicity
or
toxicity
by
different
routes.

To
reflect
these
differences,
the
HI
is
then
usually
developed
for
each
exposure
route
of
interest,
and
for
a
single
specific
toxic
effect
of
toxicity
to
a
single
target
organ.
A
mixture
may
then
be
assessed
by
several
HI,
each
representing
one
route
and
one
toxic
effect
or
target
organ.
7
To
assess
the
cumulative
risk
or
hazard
associated
with
nonlinear
effects
of
HAP
in
our
analysis
of
PCWP
facilities,

HAP
hazard
quotients
pertaining
to
the
same
target
organs
or
systems
are
summed
to
generate
TOSHI.
While
it
may
be
preferable
to
focus
on
the
addition
of
HAP
HQ
that
involve
the
same
mechanism
or
mode
of
action,
that
level
of
information
is
not
generally
available
for
all
of
the
HAP
on
which
we
are
focusing.
Pending
the
availability
of
such
data
for
the
HAP
components
of
the
mixture
being
assessed,

the
method
employed
under
CAA
section
112(
c)(
9)
is
to
aggregate
HAP
HQ
by
target
organ
to
generate
a
TOSHI.

Comment:
Two
commenters
supported
a
HI
of
1.0
(
or
327
greater)
as
an
appropriate
benchmark
for
comparing
exposures
attributable
to
facility
emissions,
which
should
fully
provide
for
the
statutory
mandate
of
an
ample
margin
of
safety.
The
commenters
referred
to
the
1997
Final
Report
of
the
Presidential/
Congressional
Commission
on
Risk
Assessment
and
Risk
Management
in
Regulatory
Decision­
Making
as
support
for
their
position.
Specifically,
the
Commission
supported
a
noncancer
HI
of
10.0,
stating
that
there
are
few
HAP
with
RfC
values
within
a
factor
of
10
of
their
no
observable
adverse
effects
level
(
NOAEL).
Because
RfC
values
are
typically
one­
thousandth
of
a
NOAEL,
a
noncancer
HI
of
10.0
in
those
cases
would
still
leave
a
margin
of
exposure
of
100.
The
Commission
recommended
that
EPA
should,
on
the
basis
of
screening
assessments
of
source
categories,
do
further
risk
assessment
and
analysis
of
categories
where
the
noncancer
HI
exceeds
10.0.
Where
more
detailed
risk
assessments
yield
noncancer
hazard
indices
less
than
1.0,

the
Commission
recommended
that
no
further
action
be
required.
The
commenters
agreed
that
sources
should
not
be
required
to
go
below
that
level
(
e.
g.,
to
an
arbitrary
level
such
as
0.2),
arguing
that
EPA
has
neither
a
legal
mandate
nor
a
rational
basis
for
limiting
the
HI
to
less
than
1.0.

One
of
the
commenters
stated
that
the
comparison
of
RfC
or
similarly­
derived
health
benchmarks
to
modeled
maximum
annual
average
concentrations
is
extremely
health­
protective
328
and
meets
the
ample
margin
of
safety
requirement
of
the
statute.
Given
this
high
degree
of
conservatism,
the
commenter
stated
that
neither
the
CAA
nor
sound
policy
requires
that
background
and
multipathway
exposures
be
incorporated
into
an
evaluation
of
the
degree
of
risk
posed
by
affected
sources.
Under
these
circumstances,
the
commenter
argued,
the
mere
possibility
of
exposure
from
multiple
sources,
or
multiple
HAP
from
a
single
source,
does
not
justify
a
uniform
adjustment
to
all
RfC
values
or
similarly­
derived
health
benchmarks
for
all
facilities.

Similarly,
the
commenter
believed
that
EPA
should
not
mandate
modeling
risks
from
the
entire
facility,
but
rather
only
from
the
portions
of
the
facility
that
are
within
the
source
category.

Two
other
commenters
objected
to
a
noncancer
HI
of
1.0
(
or
greater).
The
first
commenter
stated
that,
while
the
HI
is
useful
in
evaluating
site­
specific
impacts,
choosing
a
generic
HI
(
some
multiple
of
1.0)
for
application
to
a
wide
range
of
sites
is
inappropriate.
The
commenter
added
that
selection
of
an
arbitrary
multiple
of
1.0
is
not
science,

does
not
conform
with
CAA
section
112(
d)(
4),
and
does
not
protect
public
health.
The
commenter
stated
that
the
selection
of
a
HI
of
0.2
as
a
rough
screening
tool
seemed
reasonable,
although
it
was
unsupported
by
any
analysis.

The
commenter
added
that
if
a
default
HI
is
used,
then
EPA
329
should
include
a
provision
that
would
disallow
its
use
to
exclude
a
facility
from
MACT
if
better
background
information
is
available
suggesting
the
default
HI
does
not
protect
public
health.
However,
the
commenter
believed
that
the
CAA
does
not
support
an
interpretation
that
includes
the
use
of
such
a
default
to
allow
exemptions
for
individual
sources.
The
commenter
believed
that
the
expansion
of
the
interpretation
to
include
non­
threshold
pollutants
is
in
direct
conflict
with
section
112(
d)(
4).

The
second
commenter
evaluated
the
four
potential
options
that
EPA
proposed
to
ensure
that
a
risk
analysis
under
CAA
section
112(
d)(
4)
considered
the
total
ambient
air
concentrations
of
all
the
HAP
to
which
the
public
is
exposed.
Option
1,
which
requires
that
the
HI
for
all
pollutants
be
no
greater
than
1.0,
does
not
consider
additional
sources
or
background
and
is
unacceptable,

according
to
the
commenter.
Option
3,
which
uses
existing
data
such
as
NATA
to
determine
background
and
requires
that
the
HI
be
no
greater
than
1.0,
is
also
unacceptable,

according
to
the
commenter.
The
commenter
pointed
out
that
EPA
has
clearly
stated
at
public
meetings
that
the
NATA
is
not
to
be
used
to
make
regulatory
decisions.
(
As
the
first
commenter
noted,
NATA
information
includes
warnings
that
the
information
is
useful
for
large­
scale
planning
purposes
and
not
for
local
area
assessment.)
The
commenter
added
that
330
NATA
relies
on
data
submitted
to
EPA
voluntarily
and
has
been
reported
to
consistently
underestimate
measured
concentrations.
Until
EPA
requires
that
HAP
inventories
be
submitted
as
proposed
in
the
Consolidated
Emissions
Reporting
Rule
(
CERR),
and
the
NATA
conducts
refined
modeling
around
stationary
sources,
the
commenter
argued
that
NATA
should
not
be
considered
for
estimating
background
concentrations.
Option
4,
which
allows
individual
facilities
to
monitor
the
HAP
backgrounds
for
use
in
their
own
analysis,
requires
oversight
and
evaluation
by
the
States
to
ensure
proper
site
selections
and
analytical
methods
and
should
not
be
considered,
according
to
the
commenter.
The
commenter
believed
Option
2,
which
requires
that
the
HI
be
no
greater
than
0.2,
would
be
the
only
viable
option
at
this
time
using
a
conservative
risk
screening
analysis.
However,
the
commenter
did
not
endorse
using
any
of
the
proposed
threshold
limit
applicability
methods
to
exempt
process
sources
from
NESHAP
requirements.

Two
other
commenters
raised
additional
objections
to
EPA's
proposed
methodologies
for
determining
the
contribution
of
other
sources
to
the
overall
hazard.
The
first
commenter
stated
that
EPA
had
not
discussed
the
need
to
assess
cumulative
risks,
aggregate
exposures,
and
health
impacts
associated
with
exposure
to
chemical
mixtures
emitted
from
facilities
within
the
source
categories.
The
331
commenter
referred
EPA
to
the
extensive
progress
that
has
been
made
in
more
completely
addressing
risks
from
exposure
to
air
pollution
and
integrated
decisionmaking
in
such
areas
as
children's
risk
issues,
cumulative
exposure,
and
chemical
mixtures.
The
commenter
requested
that
the
recent
advancements
be
incorporated
into
the
risk
assessment
methods
and
overall
cost
estimates
associated
with
riskbased
exemptions
in
the
proposed
rules.

The
second
commenter
stated
that
EPA's
proposed
alternative
methodologies
for
determining
the
contribution
of
other
sources
to
cumulative
risk
are
untenable
and
deeply
flawed.
According
to
the
commenter,
the
first
and
second
approaches
(
HI
of
1.0
and
HI
of
0.2)
would
allow
exemptions
based
on
blanket
assumptions
about
exposure,
but
EPA
provided
no
basis
for
making
any
assumption.
The
commenter
noted
that
the
third
option
suggests
relying
on
existing
estimates
of
background
levels
of
certain
HAP,
but
argued
that
these
information
sources
(
e.
g.,
NATA,
ATSDR)
are
neither
designed
nor
adequately
precise
to
be
used
as
the
basis
of
regulatory
applicability
determinations.
According
to
the
commenter,
EPA
has
cautioned
that
NATA
emission
estimates
"
cannot
be
used
to
identify
exposures
and
risks
for
specific
individuals,
or
even
to
identify
exposures
and
risks
in
small
geographic
regions
such
as
a
specific
census
tract."
(
U.
S.
EPA,
Limitations
in
the
1996
National­
Scale
332
Air
Toxics
Assessment)
The
commenter
pointed
out
that
NATA
does
not
estimate
exposure
to
a
number
of
HAP,
(
e.
g.,

hydrogen
fluoride
(
HF),
HCl),
and
the
ATSDR
profiles
offer
generalized
assessments,
but
are
not
specific
enough
to
establish
as
baseline
for
a
given
facility.

Response:
For
today's
final
PCWP
rule,
we
are
considering
an
HI
limit
of
1.0
to
provide
an
ample
margin
of
safety
for
protecting
public
health
under
CAA
section
112(
c)(
9).
However,
we
do
not
feel
that
increasing
the
HI
limit
above
1.0
is
justified
by
currently
available
science.

Safety
factors
are
included
in
the
dose­
response
values
used
to
calculate
the
HI
to
account
for
scientific
uncertainties,

and
their
inclusion
helps
ensure
that
using
a
HI
limit
of
1.0
provides
an
ample
margin
of
safety.
The
TOSHI
approach
required
for
site­
specific
risk
assessment
in
today's
final
PCWP
rule
assumes
additivity
in
mixtures
of
chemicals
that
target
the
same
organ
system.
For
their
site­
specific
risk
assessments,
facilities
are
requiredencouraged
to
determine
TOSHI
for
respiratory
and
CNS
effects
to
simplify
analysis.

More
detailed
analysis
of
mixture
additivity,
incorporating
mechanistic
data
and
uncertainty
and
including
dose­
response
data
for
specific
mixtures,
where
available,
may
also
be
included
in
site­
specific
analyses
using
scientificallyaccepted
peer­
reviewed
methodologies.
Based
on
our
analysis,
noncancer
impacts
were
dominated
primarily
by
333
impacts
on
these
systems
and
other
target
organ
systems
were
found
to
be
negligibly
impacted.
We
are
not
using
background
concentrations
from
NATA
in
today's
final
PCWP
rule.
Several
commenters
presumed
the
use
of
CAA
section
112(
d)(
4)
for
the
PCWP
rule
as
proposed.
However,
we
are
using
CAA
section
112(
c)(
9)
and
not
section
112(
d)(
4).

Discussion
of
our
authority
to
consider
background
and
multipathway
exposures
is
provided
elsewhere
in
this
section.

Tiered
approach
Comment:
Several
commenters
supported
EPA's
proposed
tiered
modeling
approach,
which
begins
with
simple
look­
up
tables
and
progresses
to
more
refined
facility­
specific
risk
assessments.
One
commenter
noted
that
the
State
of
Wisconsin
uses
a
tiered
approach
similar
to
the
approach
proposed
by
EPA,
and
in
general,
this
approach
has
worked
well.
The
approach
first
allows
sources
to
demonstrate
compliance
if
their
potential
emissions,
stack
height,
and
exhaust
direction
are
within
the
ranges
provided
in
conservative
look­
up
tables.
The
second
tier
allows
facilities
to
provide
site­
specific
modeling
to
demonstrate
compliance
with
ambient
air
standards
at
the
property
line.

Another
commenter
added
that
EPA
should
be
flexible
in
accepting
evolving
improvements
in
exposure
assessment
and
risk
modeling,
and
should
take
into
account
the
inherent
334
strengths
and
weaknesses
of
the
types
of
modeling
used.
A
third
commenter
noted
that
most
sources
would
use
the
tiered
modeling
approach
but
believed
that
facilities
should
be
allowed
to
use
any
EPA­
approved
modeling
technique
to
demonstrate
that
their
emissions
are
below
the
applicable
health
benchmark.
The
commenter
also
recommended
that,
for
the
final
PCWP
rule,
EPA
adopt
the
model
regulatory
text
that
they
provided
for
the
risk­
based
framework.

One
commenter
opposed
EPA's
proposed
tiered
modeling
approach,
stating
that
if
EPA
decided
to
pursue
a
generic
risk
screening
approach
under
section
112(
d)(
4),
it
would
need
to
be
conservative.
According
to
the
commenter,
the
use
of
a
(
non­
tiered)
conservative
approach
would
represent
the
least
cost
to
the
regulated
community
and
would
be
the
least
time­
consuming
for
States
reviewing
the
facility's
application.

Response:
We
acknowledge
the
model
regulatory
text
submitted
by
one
of
the
commenters.
However,
as
discussed
elsewhere,
we
developed
our
own
regulatory
text
to
specify
how
facilities
must
demonstrate
that
they
are
part
of
the
low­
risk
subcategory.
We
discussed
a
tiered
analytical
approach
in
the
preamble
to
the
proposed
rule,
beginning
with
relatively
simple
lookup
tables
and
followed
by
increasingly
more
site­
specific
but
more
resource
intensive
tiers
of
analysis,
with
each
tier
being
more
refined
and
335
less
conservative.
We
are
specifying
a
similar
approach
for
today's
final
rule
in
Appendix
B
to
subpart
DDDD.

Facilities
can
qualify
for
inclusion
in
the
delisted
subcategory
by
using
site­
specific
emissions
test
data
and
look­
up
tables
that
were
developed
using
conservative
input
parameters.
These
look­
up
tables
are
included
in
Appendix
B
to
subpart
DDDD.
Facilities
that
cannot
demonstrate
eligibility
for
the
low­
risk
subcategory
based
on
the
conservative
screening
assessment
(
i.
e.,
look­
up
tables)
may
used
more
refined
site­
specific
risk
assessments
as
described
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
and
in
other
analytical
tools
such
as
the
"
Air
Toxics
Risk
Assessment
Reference
Library,"
(
which
may
be
appropriate
for
specific
sources).
A
more
refined
analysis
requires
more
effort,
but
produces
results
that
are
less
uncertain
and
less
conservative
(
i.
e.,
less
likely
to
overestimate
risk).

Appendix
B
to
subpart
DDDD
of
40
CFR
part
63
describes
a
process
based
on
this
air
toxics
approach
that
must
be
used
by
facilities
attempting
to
demonstrate
eligibility
for
the
low­
risk
subcategory
of
PCWP
facilities.
Facilities
may
demonstrate
eligibility
for
the
low­
risk
subcategory
using
either
the
look­
up
table
analysis
(
described
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63)
or
a
tiered
site­
specific
analysis
(
described
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
and
other
analytical
tools
which
may
be
appropriate
336
for
a
specific
source,
such
as
the
"
Air
Toxics
Risk
Assessment
Reference
Library.")
As
discussed
elsewhere
in
this
section,
through
low­
risk
demonstrations.
Also,
we
will
be
reviewing
the
low­
risk
demonstrations
submitted
by
PCWP
facilities
to
remove
the
burden
of
reviewing
risk
assessments
from
States.

We
will
review
all
risk
assessments
performed
in
support
of
a
demonstration
of
eligibility
for
the
low­
risk
subcategory
with
regard
to
a
variety
of
aspects,
including
the
consistency
of
the
methodology
and
modeling
techniques
with
those
currently
accepted
by
the
scientific
community
and
EPA.
However,
we
will
consider
assessments
that
use
risk
methodology
and
modeling
techniques
in
addition
to
or
in
lieu
of
those
described
in
EPA's
"
Air
Toxics
Risk
Assessment
Reference
Library,"
as
appropriate,
provided
they
have
undergone
scientific
peer
review
pertinent
to
their
use
in
the
submitted
assessment.

Comment:
One
commenter
stated
that,
for
EPA
to
conduct
an
up­
front
risk
analysis,
the
procedure
would
need
to
be
conducted
using
the
most
conservative
stack
parameters,
with
a
hypothetical
facility
fence
line
to
satisfy
the
many
impact
scenarios
that
could
occur.

Response:
We
conducted
a
rough
risk
assessment
to
estimate
the
number
of
PCWP
facilities
that
might
qualify
for
the
delisted
low­
risk
subcategory.
The
data
used
in
our
337
rough
risk
assessment
were
a
combination
of
facilityspecific
data
(
e.
g.,
process
unit
throughput)
and
industry
average
data
(
e.
g.,
industry
average
stack
parameters,

average
emission
factors
for
estimating
emissions).

Facilities
do
not
qualify
for
the
low­
risk
subcategory
based
on
our
rough
risk
assessment,
with
the
exception
of
eight
facilities
who
were
determined
to
pose
very
low
risk
based
on
our
analysis
(
i.
e.
g.,
with
TOSHI
less
than
0.1,
and
a
cancer
risk
of
less
than
0.1
in
1
million).
However,

facilities
can
qualify
for
inclusion
in
the
delisted
subcategory
by
using
site­
specific
emissions
test
data
and
the
look­
up
tables
or
by
conducting
a
low­
risk
demonstration,
as
described
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
and
in
other
analytical
tools
such
as
the
"
Air
Toxics
Risk
Assessment
Reference
Library,"
(
which
may
be
appropriate
for
specific
sources).
Look­
up
tables
were
developed
using
the
conservativehealth­
protective
air
dispersion
model
SCREEN3.
Stack
height
and
fenceline
distance
vary
in
the
tables,
so
facilities
will
choose
the
most
appropriate
combination
of
these
parameters.
Invariant
facility
parameters
built
into
the
look­
up
tables
are
either
average
values
or
biased
towards
conservativehealthprotective
values,
based
on
available
data.
Thus,
we
believe
the
look­
up
tables
are
appropriately
conservativehealth­
protective
to
accommodate
the
many
impact
338
scenarios
that
could
occur.

Risk
assessment
guidance
Comment:
Several
commenters
stated
that
EPA
neglected
to
follow
its
own
guidelines
and
science
policies
in
its
proposal
for
risk­
based
exemptions.
One
commenter
argued
that
EPA
had
proposed
a
disorganized
and
cursory
approach
to
implement
risk­
based
exemptions
that
fell
far
below
the
quality
of
risk
analysis
typically
required
by
EPA
across
its
other
programs.
According
to
the
commenter,
the
proposal
did
not
adhere
to
EPA's
established
guidelines
for
characterizing
human
health
and
ecological
risks,
did
not
incorporate
risk
assessment
guidelines
for
conducting
multipathway
risk
assessments,
and
did
not
reference
EPA
guidelines
for
cumulative
risk
assessment
that
specifically
require
consideration
of
non­
inhalation
pathways.
The
commenter
noted
that
EPA's
March
1995
Risk
Characterization
Policy
set
goals
of
transparency,
clarity,
consistency,
and
reasonableness
which
apply
to
risk
assessment
practices
across
EPA.
The
commenter
argued
that
the
inconsistencies
between
EPA's
proposal
to
provide
risk­
based
exemptions
in
the
MACT
standard
process
and
its
risk
assessment
guidelines
would
undermine
many
regulatory
programs
throughout
EPA.

The
commenter
stated
that
the
risk­
based
scheme
was
based
on
a
fundamental
misunderstanding
of
the
use
of
public
health
and
ecological
risk
assessments
in
the
regulatory
339
process.
The
commenter
added
that
the
Federal
risk
assessment
guidelines
require
EPA
to
conduct
risk
assessments
consistently
across
all
Federal
environmental
programs.
According
to
the
commenter,
the
approaches
outlined
by
industry's
white
papers
neglected
to
include
risk
characterization,
which
provides
needed
and
appropriate
information
to
decision
makers.
The
approaches
also
did
not
incorporate
the
critical
recommendation
of
the
Commission
of
Risk
Assessment
and
Risk
Management
to
establish
a
framework
for
stakeholder­
based
risk
management
decision
making.
The
commenter
stated
that
these
omissions
in
the
proposal
would
prevent
regulatory
agencies
from
demonstrating
to
the
public
that
public
health
and
the
environment
are
adequately
protected.

Several
commenters
stated
that
EPA
also
needed
to
be
consistent
with
residual
risk
guidelines
currently
under
development.
One
commenter
stated
that
the
tools
needed
to
identify
sources
eligible
for
the
risk­
based
exemption
would
be
the
same
tools
necessary
for
a
CAA
section
112(
f)

residual
risk
assessment,
which
the
commenter
understood
were
not
yet
ready
for
general
use.
Another
commenter
noted
that
the
cancer
risk
guidelines
are
currently
undergoing
public
review.

A
third
commenter
stated
they
had
serious
reservations
about
EPA's
apparent
attempt
to
conduct
an
ad­
hoc
risk
340
analysis
for
specific
source
categories
by
seeking
comments
on
the
specific
elements
to
be
included
in
the
risk
analysis.
The
commenter
did
not
believe
these
rulemakings
were
an
adequate
forum
to
develop
this
risk
analysis
process.
The
commenter
indicated
that
any
risk
analysis
conducted
by
the
EPA
must
adhere
to
the
risk
assessment
principles
outlined
in
the
Residual
Risk
Report
to
Congress.

One
commenter
argued
that
the
proposal
is
consistent
with
EPA
risk
assessment
guidelines
and
policies
and
believed
that
others'
technical
objections
were
without
merit.
The
commenter
added
that
the
contemplated
risk­
based
applicability
criteria
were
not
in
conflict
with
the
classification
of
carcinogens
and
noncarcinogens.

Response:
We
discussed
a
tiered
analytical
approach
in
the
preamble
to
the
proposed
rule,
beginning
with
relatively
simple
lookup
tables
and
followed
by
increasingly
more
sitespecific
but
more
resource
intensive
tiers
of
analysis,
with
each
tier
being
more
refined
and
less
conservative.
We
are
specifying
a
similar
approach
forIn
today's
final
rule
in
Appendix
B
to
subpart
DDDD.
Facilities
can
qualify
for
inclusion
in
the
delisted
subcategory
by
using
site­
specific
emissions
test
data
and
look­
up
tables
that
were
developed
using
conservative
input
parameters.
These
look­
up
tables
are
included
in
Appendix
B
to
subpart
DDDD.
Facilities
that
cannot
demonstrate
eligibility
for
the
low­
risk
subcategory
341
based
on
the
conservative
screening
assessment
(
i.
e.,

lookup
tables)
may
used
more
refined
site­
specific
risk
assessments
as
described
in
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
and
in
other
analytical
tools
such
as
the
"
Air
Toxics
Risk
Assessment
Reference
Library,"
(
which
may
be
appropriate
for
specific
sources).
A
more
refined
analysis
requires
more
effort,
but
produces
results
that
are
less
uncertain
and
less
conservative
(
i.
e.,
less
likely
to
overestimate
risk).
While
appendix
B
to
subpart
DDDD
of
40
CFR
part
63
describes
a
process
based
on
this
air
toxics
approach
particular
to
risk
assessment
for
112(
c)(
9),
the,

we
are
adopting
a
somewhat
different
approach
for
meeting
the
requirements
of
CAA
section
112(
c)(
9),
as
discussed
elsewhere
in
this
preamble.
The
basis
for
this
processapproach
stems
from
the
general
air
toxics
assessment
approach
presented
in
the
Residual
Risk
Report
to
Congress,

which
was
developed
with
full
consideration
of
EPA
risk
assessment
policy,
guidance,
and
methodology.

Section
112(
c)(
9)
of
the
CAA
requires
us
to
determine
whether
the
public
and
the
environment
are
protected.
Any
analyses
we
did
to
establish
the
feasibility
of
the
riskbased
approach
or
to
develop
health
­
protective
look­
up
tables
included
consideration
of
human
health
as
well
as
ecological
criteria.
The
supporting
information
to
the
final
rule
details
the
assessment
we
conducted
to
determine
342
the
feasibility
of
delisting
a
low­
risk
subcategory
and
the
look­
up
tables
we
developed
to
be
used
by
facilities
in
their
demonstrations,
thereby
providing
a
public
demonstration
of
the
method
employed
to
ensure
protection
of
the
public
health
and
environment
in
decisions
associated
with
this
rule.
Additionally,
protection
against
the
potential
for
exposures
via
non­
inhalation
pathways
(
e.
g.,

ingestion)
is
also
inherent
in
the
values
for
persistent,

bioaccumulative
HAP
is
also
inherent
in
the
values
in
the
look­
up
tables.
As
discussed
previously,
and
in
the
supporting
information
for
the
final
rule,
we
conducted
a
screening
assessment
of
multipathway
and
ecological
effects
for
the
PCWP
source
category.
We
concluded
that
multipathway
risks
are
considerably
lower
than
predicted
maximum
inhalation
risks
and
that
it
is
unlikely
that
PCWP
facilities
would
pose
any
significant
risk
to
nearby
ecosystems.
Therefore,
facilities
are
not
required
to
conduct
site­
specific
multipathway
and
ecological
risk
assessments
as
part
of
their
low­
risk
demonstrations.

We
agree
that
the
tools
needed
to
identify
sources
eligible
for
the
delisted
low­
risk
subcategory
of
PCWP
facilities
are
the
same
tools
necessary
for
a
CAA
section
112(
f)
residual
risk
assessment.
And,
as
stated
in
the
Residual
Risk
Report
to
Congress,
we
intend
to
rely
on
the
general
methodology
and
process
illustrated
by
the
framework
343
presented
in
that
report
in
our
risk
assessment
activities
throughout
the
air
toxics
program.
Facilities
must
demonstrate
eligibility
for
the
delisted
low­
risk
subcategory
using
either
a
look­
up
table
analysis
(
based
on
the
look­
up
tables
included
in
appendix
B
to
subpart
DDDD
of
this
part)
or
using
the
suggested
site­
specific
methodology
described
together
with
the
criteria
in
appendix
B
to
subpart
DDDD
of
this
part.
The
"
Air
Toxics
Risk
Assessment
Guidance
Library"
is
the
sameLibrary,"
developed
specifically
for
EPA's
Residual
Risk
program,
is
provided
as
an
example
of
one
document
that
could
be
used,
where
appropriate,
for
these
facility­
specific
residual
risk
assessments
required
under
CAA
section
112(
f).
This
document
has
been
peer­
reviewed
and
was
developed
according
to
the
principles,
tools
and
methods
outlined
in
the
Residual
Risk
Report
to
Congress.
However,
it
may
not
be
appropriate
for
all
sources,
and
for
that
reason
sources
and
EPA
may
consider
alternative
analytical
tools
for
these
risk
assessments.

The
comment
that
the
new
cancer
guidelines
are
still
under
review
is
correct
but,
as
stated
in
the
November
29,

2001
Federal
Register
notice
(
66
FR
59593),
these
1999
draft
344
8U.
S.
EPA.
1999.
Guidelines
for
Carcinogen
Risk
Assessment.
NCEA­
F­
0644.
Risk
Assessment
Forum,
Washington,
DC
guidelines
are
to
be
considered
the
interim
guidance.
8
4.
Implementation
State
and
local
resources
Comment:
Several
commenters
contended
that
the
proposal
would
place
a
very
intensive
resource
demand
on
State
and
local
agencies
(
e.
g.,
permitting
authorities)
to
review
sources'
risk
assessments.
State
and
local
agencies
may
not
have
expertise
in
risk
assessment
methodology
or
the
resources
needed
to
verify
information
submitted
with
each
risk
assessment.
The
commenters
argued
that,
if
EPA
intends
to
have
the
affected
industries
conduct
the
analysis,
then
EPA
must
consider
the
cost
incurred
by
States,
which
may
lack
the
necessary
expertise
to
evaluate
and
review
these
analyses.

One
commenter
pointed
out
that
the
proposal
only
considered
cost
for
the
regulated
source
category,
and
not
for
regulatory
agencies.
According
to
the
commenter,
EPA
did
not
consider
the
cost
and
resources
associated
with
the
following:
(
1)
the
public
process
required
in
reviewing
and
approving
the
proposed
approaches
and,
if
approved,
making
substantial
changes
to
existing
regulations;
(
2)
the
development
of
methods
and
guidance
for
human
health
and
345
ecological
risk
assessments
of
affected
sources;
(
3)
the
review
by
already
budgetarily
constrained
State
agencies
of
the
assessments
and
assurance
of
adequate
public
participation
in
the
process;
and
(
4)
the
collection
and
verification
of
source­
specific
data
needed
for
conducting
risk
assessments
(
e.
g.,
emissions
data
and
stack
parameters).
The
commenter
added
that
the
proposal
did
not
address
the
critical
need
for
qualified
risk
assessors
to
evaluate
the
scientific
and
technical
basis
for
exempting
facilities
from
regulation
on
a
case­
by­
case
basis.
The
commenter
estimated
that
if
one
additional
full­
time
employee
(
FTE)
were
required
per
State
to
review
risk­
based
exemptions,
then
the
cost
would
be
an
additional
$
7.5
million
annually.

Another
commenter
pointed
out
that
the
ongoing
assurance
that
low­
risk
facilities
remain
low
risk
would
also
increase
the
burden
for
the
State
and
local
agencies.
The
commenter
also
stated
that
diverting
State
and
local
resources
to
focus
on
presumably
insignificant
sources
would
detract
from
efforts
associated
with
significant
sources.

A
third
commenter
stated
that,
since
States
generally
do
not
have
the
right
staff
or
resources
to
hire
additional
staff
to
review
lengthy
and
complex
risk
analyses,
they
may
refuse
delegation
of
the
PCWP
rule,
which
would
shift
the
burden
to
EPA
in
a
time
of
tight
budgets.
According
to
the
346
commenter,
large
expenditures
are
not
justified
when
only
a
small
number
of
facilities
may
end
up
qualifying
for
an
exemption.

By
contrast,
several
commenters
stated
that
a
risk­
based
program
approaches
could
be
structured
and
implemented
in
a
manner
that
would
not
impose
a
substantial
cost
or
resource
burden
on
States.
One
commenter
stated
that
assuring
compliance
with
risk­
based
applicability
criteria
would
be
straightforward
and
would
not
entail
an
added
resource
burden.
Another
commenter
suggested
that
EPA
work
closely
with
States
and
industry
to
implement
the
risk­
based
approach
in
a
non­
burdensome
manner.
Two
commenters
stated
that
the
risk­
based
approaches,
like
other
MACT
standards,

would
simply
be
incorporated
into
each
State's
existing
title
V
program.
Because
the
title
V
framework
already
exists,
the
addition
of
a
risk­
based
MACT
standard
would
not
require
States
to
overhaul
existing
permitting
programs.

One
commenter
stated
that
the
risk­
based
approach
would
not
increase
the
number
of
sources
regulated
by
each
State.
The
commenter
believed
that
the
final
MACT
rule
itself
should
set
forth
the
applicability
criteria,
including
the
threshold
levels
of
exposure,
that
sources
must
meet
to
qualify
for
a
risk­
based
determination.
Each
source
would
have
the
burden
of
demonstrating
that
its
exposures
are
below
this
limit,
and,
therefore,
the
States
would
not
be
347
required
to
develop
their
own
risk
assessment
guidance
or
to
conduct
source­
specific
risk
assessments.
One
commenter
stated
that
the
risk
assessment
guidance
to
be
issued
by
EPA
within
the
next
several
months
would
streamline
the
risk­
based
approach
and
further
reduce
any
burden
on
the
States.
Three
commenters
supported
having
States
charge
reasonable
increased
fees
(
as
a
component
of
annual
operating
permit
fees
or
other
fees)
to
cover
any
significant
additional
workload
demands
associated
with
reviewing
more­
detailed
tier
2/
3
modeling.

Response:
We
acknowledge
that
review
of
the
eligibility
demonstrations
for
the
delisted
low­
risk
subcategory
will
require
resources
for
verification
of
information
and
may
require
expertise
in
risk
assessment
methodology
that
is
not
yet
available
in
some
States.
We
also
acknowledge
that
States
may
choose
to
reject
delegation
of
the
final
PCWP
rule.
To
alleviate
these
concerns
and
to
ensure
consistency
in
the
applicability
determinations
for
the
delisted
lowrisk
subcategory
from
State­
to­
State,
we
will
review
and
approve/
disapprove
the
low­
risk
subcategory
eligibility
demonstrations
submitted
by
PCWP
facilities.
As
mentioned
previously
in
this
preamble,
we
encourage
facilities
to
submit
their
assessments
for
review
early
to
facilitate
a
timely
review
process.

We
have
considered
the
above
comments
in
developing
the
348
criteria
defining
the
delisted
low­
risk
subcategory
of
PCWP
facilities,
and
we
feel
that
the
approach
that
is
included
in
today's
final
PCWP
rule
provides
clear,
flexible
requirements
and
enforceable
compliance
parameters.
Today's
final
PCWP
rule
provides
two
ways
that
a
facility
may
demonstrate
that
it
is
part
of
the
delisted
low­
risk
subcategory
of
PCWP
facilities.
First,
look­
up
tables,

which
are
included
in
appendix
B
to
subpart
DDDD
of
this
part,
allow
facilities
to
determine,
using
a
limited
number
of
site­
specific
input
parameters,
whether
emissions
from
their
sources
might
cause
an
HI
limit
to
be
exceeded.

Finally,
a
site­
specific
modeling
approach
can
be
used
by
those
facilities
that
cannot
demonstrate
that
they
are
part
of
the
delisted
low­
risk
subcategory
using
the
look­
up
tables.
With
respect
to
guidance
for
performing
low­
risk
demonstrations,
one
possible
available
set
of
procedures
for
performing
risk
assessments
is
discussed
in
EPA's
"
Air
Toxics
Risk
Assessment
Guidance
Reference
Library,"
and
may
be
used,
where
appropriate.

Only
a
portion
of
the
223
PCWP
major
sources
will
submit
eligibility
demonstrations
for
low­
risk
subcategory.
Of
this
portion
of
major
sources,
we
feel
that
most
will
find
themselves
in
the
low­
risk
subcategory
based
on
screening
analyses
(
e.
g.,
look­
up
table).
However,
it
is
likely
that
some
facilities
will
submit
more
detailed
risk
modeling
349
results.
We
are
experienced
in
reviewing
emission
test
results
and
site­
specific
risk
assessments
and
will
allocate
resources
for
completion
of
these
tasks.
We
will
review
and
approve/
disapprove
low­
risk
subcategory
eligibility
demonstrations
based
on
look­
up
table
analyses
and
low­
risk
demonstrations.
Following
review
of
each
low­
risk
subcategory
eligibility
demonstration
for
a
facility,
we
will
issue
a
letter
of
approval/
disapproval
to
the
facility
and
will
send
a
carbon
copy
to
the
facility's
title
V
permitting
authority
to
be
used
to
develop
source­
specific
permit
terms
and
conditions
that
will
ensure
that
the
source
remains
eligible
for
the
low
risk
subcategory.
The
letter
of
notification
regarding
approval/
disapproval
of
a
facility's
low
risk
demonstration
will
also
be
sent
to
any
other
interested
stakeholders.
The
criteria
for
low­
risk
subcategory
delisting
are
clearly
spelled
out
in
today`
s
final
PCWP
rule,
along
with
criteria
needed
to
ensure
that
facilities
in
the
low­
risk
subcategory
remain
low
risk.

Because
these
requirements
are
clearly
spelled
out
in
today's
final
PCWP
rule
and
because
any
standards
or
requirements
created
under
CAA
section
112
are
considered
applicable
requirements
under
40
CFR
part
70,
the
terms
and
conditions
demonstrating
eligibility
for
membership
in
the
delisted
low­
risk
subcategory
would
be
incorporated
into
title
V
permits,
pursuant
to
State's
existing
permitting
350
programs.

With
respect
to
the
burden
associated
with
ongoing
assurance
that
facilities
remain
low
risk,
the
burden
to
States
of
assuring
that
facilities
continue
to
be
low
risk
will
be
no
more
than
the
burden
associated
with
ongoing
title
V
enforcement
because
the
parameters
that
rendered
a
facility
low
risk
will
be
reflected
in
terms
and
conditions
to
be
incorporated
into
the
title
V
permit.
We
have
developed
continuous
compliance
requirements
for
facilities
that
initially
qualify
as
low
risk,
and
the
facilities
will
be
responsible
for
demonstrating
that
they
continue
to
be
low
risk
if
changes
are
made
to
the
facilities'
operations
that
would
affect
the
risk
that
the
facilities
pose
to
human
health
and
the
environment.
We
will
review
and
approve/
disapprove
revised
low­
risk
demonstrations.

With
respect
to
our
consideration
of
the
public
process
required
in
reviewing/
approving
the
proposed
approaches
and
making
substantial
changes
to
existing
regulations,
our
inclusion
of
a
risk­
based
compliance
option
in
today's
final
PCWP
rule
applies
only
to
the
PCWP
rule
and
does
not
directly
impact
other
regulations.
Furthermore,
the
PCWP
proposal
provided
the
public
with
the
opportunity
to
comment
on
the
consideration
of
risk
in
the
final
PCWP
rule.

Regarding
the
assurance
of
adequate
public
participation
in
the
process
of
reviewing
the
risk
analyses,
the
risk­
351
based
compliance
options
are
part
of
a
rule
that
was
subject
to
public
comment.
The
supporting
information
to
the
final
rule
details
the
assessment
we
conducted
to
determine
the
feasibility
of
delisting
a
low­
risk
subcategory
and
the
look­
up
tables
we
developed
to
be
used
by
facilities
in
their
demonstrations,
thereby
providing
a
public
demonstration
of
the
method
employed
to
ensure
protection
of
the
public
health
and
environment
in
decisions
associated
with
the
final
rule.
We
will
be
responsible
for
reviewing
the
low­
risk
demonstrations,
but,
similar
to
facilities
requesting
applicability
determinations
regarding
promulgated
standards,
individual
low­
risk
demonstrations
will
not
be
subject
to
public
review
and
comment.
We
will,

however,
periodically
publish
updating
notices
in
the
Federal
Register
identifying
any
additional
members
of
the
low
risk
PCWP
subcategory
(
or
deletions
therefrom),
again,

similarly
to
how
we
update
notices
regarding
applicability
determinations.
These
actions
will
represent
final
agency
actions
for
purposes
of
judicial
review
under
CAA
section
307(
b)(
1).
However,
the
parameters
that
rendered
a
facility
of
the
low­
risk
subcategory
will
be
incorporated
into
a
title
V
permit
and
subject
to
the
public
review
process
through
title
V.

Comment:
One
commenter
stated
that
if
EPA
intends
to
have
the
affected
industries
conduct
the
analysis,
then
EPA
352
must
consider
the
additional
cost
incurred
by
smaller
sources
to
do
the
analysis.

Response:
As
mentioned
previously,
there
are
two
ways
that
a
PCWP
facility
may
demonstrate
eligibility
for
the
delisted
low­
risk
subcategory:
(
1)
look­
up
tables,
and
(
2)

a
site­
specific
modeling
approach
that
can
be
used
by
facilities
that
cannot
demonstrate
eligibility
for
the
delisted
low­
risk
subcategory
using
the
look­
up
tables.
The
look­
up
tables
included
in
appendix
B
to
subpart
DDDD
of
this
part
allow
facilities
to
determine,
using
a
limited
number
of
site­
specific
input
parameters,
whether
they
are
eligible
for
the
low­
risk
subcategory.
Attempting
to
demonstrate
eligibility
for
the
delisted
low­
risk
subcategory
is
completely
voluntary.
Facilities
that
are
not
eligible
for
the
delisted
low­
risk
subcategory
based
on
look­
up
tables
are
not
required
to
pursue
a
site­
specific
analysis
(
which
can
be
increasingly
complex
and
expensive
as
it
becomes
more
refined).
Each
facility
must
weigh
the
costs
of
making
a
low­
risk
demonstration
with
the
costs
of
MACT
compliance.
We
feel
that
in
general
the
costs
associated
with
demonstrating
eligibility
for
the
low­
risk
subcategory
will
be
lower
than
the
costs
associated
with
complying
with
MACT
for
many
facilities,
particularly
smaller
facilities
and
other
facilities
that
have
not
already
otherwise
installed
pollution
controls.
The
353
majority
of
the
cost
associated
with
demonstrating
eligibility
for
the
delisted
low­
risk
subcategory
will
be
emissions
testing
costs.
Smaller
facilities
have
fewer
process
units
to
be
tested,
and,
because
of
their
lower
production
rates
relative
to
larger
facilities,
they
will
also
likely
have
lower
emissions.
Thus,
smaller
PCWP
facilities
may
be
more
likely
than
their
larger
counterparts
to
fall
into
the
delisted
low­
risk
subcategory.

Successfully
demonstrating
eligibility
for
the
low­
risk
provisions
will
result
in
cost­
savings
for
smaller
facilities
because
these
facilities
will
not
have
to
expend
the
costs
(
e.
g.,
the
costs
of
installing
operating,
and
maintaining
emission
controls)
for
MACT
compliance.

The
cost
and
economic
analyses
developed
as
part
of
the
MACT
rulemaking
were
based
on
the
costs
to
install
controls
and
comply
with
the
MACT
requirements.
The
costs
associated
with
voluntarily
conducting
risk
analyses
were
not
estimated.
Therefore,
our
estimate
of
costs
associated
with
today's
final
PCWP
rule
are
conservative,
because
the
control
costs
are
significantly
higher
than
the
costs
of
conducting
emissions
tests
and
risk
analyses.

Title
V
Comment:
Two
commenters
opposed
implementing
the
riskbased
approaches
through
the
States'
existing
title
V
programs.
One
commenter
stated
that
risk­
based
exemptions
354
are
such
an
implausible
interpretation
of
the
CAA
that
States
do
not
even
have
the
authority
to
grant
them
under
their
title
V
permit
programs.
The
commenter
was
not
aware
of
any
approach
to
ensure
that
emissions
remain
below
specified
levels.
According
to
the
commenter,
MACT
standard
applicability
is
the
gate­
keeper
for
being
subject
to
a
title
V
operating
permit.
Once
a
source
is
exempt
from
a
MACT
standard,
it
would
be
exempt
from
the
monitoring,

reporting
and
recordkeeping
requirements
needed
to
demonstrate
compliance.

The
other
commenter
stated
that
implementing
the
CAA
section
112(
d)(
4)
exemption
interpretation
through
title
V
would
be
unlawful
and
unworkable.
The
commenter
stated
that
Congress
knew
how
to
authorize
States
to
establish
case­

bycase
emission
standards
and
implement
them
using
postrulemaking
title
V
permits
because
it
did
so
in
CAA
section
112(
j).
However,
it
did
not
do
so
in
section
112(
d)(
4).

The
commenter
argued
that
EPA
lacks
the
authority
to
delegate
section
112(
d)(
4)
to
the
States
and
may
not
implement
any
section
112(
d)(
4)
applicability
cutoff
through
a
post­
rulemaking
mechanism
such
as
a
title
V
permit.
With
the
exception
of
carefully
delineated
compliance
monitoring,

reporting,
and
certification
provisions
in
the
statute,

title
V
permits
may
not
create
applicable
requirements
or
exemptions
from
applicable
requirements.
The
commenter
355
added
that,
even
if
this
approach
is
legal,
it
is
still
unworkable
because
of
the
resource
challenges
faced
by
States
and
the
widespread
delays
in
issuing
title
V
permits.

The
commenter
noted
that
State
permit
engineers
and
officials
that
prepare
and
issue
title
V
permits
generally
are
not
experts
in
risk
assessment
or
air
dispersion
modeling.
According
to
the
commenter,
States
and
the
public
would
be
confronted
with
more
self­
serving
facility
arguments
and
data
than
could
be
adequately
scrutinized,

which
could
cause
important
health
and
risk
determinations
to
be
rubber
stamped
or
cause
the
permit
process
to
grind
to
a
halt.
The
commenter
added
that
most
State
title
V
permit
programs
are
already
behind
the
statute's
permit
issuance
deadlines,
and
implementation
of
EPA's
risk­
based
approach
would
exacerbate
this
unlawful
situation
further.

Several
commenters
supported
implementing
the
risk­
based
approaches
in
the
PCWP
rule
as
proposed
through
the
States'

existing
title
V
programs.
One
commenter
suggested
that
States
which
qualify
and
choose
to
do
so
should
be
delegated
the
authority
to
implement
the
risk­
based
alternatives.
The
commenter
added
that
this
would
allow
States
to
coordinate
between
the
MACT
alternatives
and
State
air
toxics
requirements.

A
second
commenter
stated
that
implementing
the
CAA
section
112(
d)(
4)
risk­
based
approach
though
title
V
would
356
be
lawful
and
workable.
According
to
the
commenter,
no
facility­
specific
post­
rulemaking
mechanisms
nor
expansion
of
the
scope
of
title
V
permit
process
would
be
necessary,

just
the
incorporation
of
the
NESHAP's
risk­
based
compliance
option,
which
would
contain
the
criteria
for
showing
what
the
source
would
have
to
meet
to
qualify
for
the
risk­
based
approach.
The
commenter
stated
that
the
objections
from
other
commenters
to
the
risk­
based
criteria
were
invalid,

arguing
that
their
objections
were
in
tension
with
the
conclusions
of
a
CAAAC
Workgroup
on
State/
Local/
Tribal
air
toxics
issues
and
that
their
comments
provided
no
basis
for
concluding
that
States
lack
the
legal
authority
to
implement
the
risk­
based
approach.

A
third
commenter
noted
that
title
V
permits
could
provide
enforceable
limitations,
appropriate
recordkeeping
requirements,
and
periodic
review
upon
renewal.
The
commenter
added
that,
since
the
PCWP
rule
would
apply
only
to
major
sources,
title
V
permits
already
are
required
and
would
not
be
an
added
burden;
title
V
could
also
be
used
to
implement
applicability
cutoffs.
However,
the
workload
involved
with
the
options
requiring
modeling,
ambient
monitoring,
or
other
means
to
establish
background
concentrations
would
be
a
hindrance
to
any
implementation
mechanism.
The
commenter
stated
that,
with
respect
to
potential
risk­
based
provisions,
monitoring
is
more
useful
357
for
demonstrating
non­
compliance
than
compliance
because
the
regulation
would
apply
to
potential
emissions
under
any
weather
conditions,
whereas
monitoring
reflects
current
weather
and
emission
conditions.

A
fourth
commenter
suggested
changes
to
the
§
63.2240
of
the
proposed
rule
that
would
incorporate
permitting
procedures
similar
to
those
under
40
CFR
part
70,
which
would
allow
facilities
that
pose
little
risk
in
their
respective
airsheds
to
apply
for
a
risk
determination
to
be
incorporated
into
their
title
V
permits.
Each
source
applying
to
be
permitted
as
a
subcategorized
toxic
emitter
with
an
acceptable
risk
determination
would
be
required
to
perform
detailed
risk
analyses
for
review
by
the
public
at
large,
local
citizens,
State
agencies,
and
Federal
authorities.
This
permitting
exercise
would
allow
managers
of
the
airshed
to
develop
custom­
fit
compliance
plans
that
address
source­
specific
risks
and
would
allow
the
most
flexibility
for
forest
producers
to
reduce
their
identified
risks.

Response:
As
discussed
previously,
we
have
determined
that
a
CAA
section
112(
d)(
4)
risk­
based
exemption
would
not
be
appropriate
for
the
PCWP
source
category.
Instead,
using
our
discretion
in
establishing
subcategories
of
sources
based
on
size,
type,
class,
or
other
appropriate
criteria
under
CAA
sections
112(
d)(
1)
and
(
c)(
1),
we
have
established
358
a
low­
risk
subcategory
of
PCWP
facilities,
and
delisted
that
subcategory
under
CAA
section
112(
c)(
9)(
B).
The
requirements
for
qualifying
for
and
remaining
in
the
delisted
low­
risk
subcategory
are
clearly
spelled
out
in
appendix
B
to
subpart
DDDD
of
this
part,
and
any
standards
or
requirements
created
under
CAA
section
112
are
considered
applicable
requirements
under
40
CFR
part
70.
Unless
a
PCWP
source
meets
these
conditions,
it
will
remain
subject
to
the
PCWP
MACT
rules.
Therefore,
the
parameters
used
to
demonstrate
that
facilities
are
part
of
the
delisted
lowrisk
subcategory
would
be
incorporated
into
title
V
permits
as
federally
enforceable
permit
terms,
and
States
would
not
have
to
overhaul
existing
permitting
programs.
We
note
that
our
rules
implementing
title
V
of
the
CAA
specifically
provide
for
situations
such
as
this.
For
example,
in
its
provisions
governing
what
types
of
permit
revisions
may
proceed
through
the
abbreviated
"
minor
permit
modification"

process,
our
rules
state
that
such
procedures
may
not
be
used
"
to
establish
or
change
a
permit
term
or
condition
for
which
there
is
no
corresponding
underlying
applicable
requirement
and
that
the
source
has
assumed
to
avoid
an
applicable
requirement
to
which
the
source
would
otherwise
be
subject."
40
CFR
70.7(
e)(
2)(
i)(
A)(
4);
40
CFR
71.7(
e)(
1)(
i)(
A)(
4).
We
feel
that
permit
terms
reflecting
a
low
risk
PCWP
source's
eligibility
clearly
represent
such
359
terms,
and
are,
therefore,
allowed
under
title
V.
Also,

such
terms
would
be
required
to
be
added
or
revised
through
the
more
formal
"
significant
modification"
procedures
of
40
CFR
70.7(
e)(
4)
and
40
CFR
71.7(
e)(
3).

Facilities
that
qualify
as
part
of
the
delisted
low­
risk
subcategory
will
initially
demonstrate
that
they
are
lowrisk
using
either
the
look­
up
tables
or
site­
specific
monitoring.
They
will
demonstrate
that
risk
does
not
increase
by
documenting
that
parameters
that
impact
the
risk
analysis
do
not
change
in
a
way
that
increases
risk.

Facilities
will
not
be
required
to
perform
detailed
risk
analyses
for
public
review,
although
the
public
will
have
an
opportunity
to
comment
on
draft
permit
terms
and
conditions
that
reflect
low
risk
demonstrations,
and
to
judicially
challenge
final
EPA
approvals
of
eligibility
demonstrations
under
CAA
section
307(
b)(
1).

We
acknowledge
the
resource
challenges
faced
by
States,

and,
therefore,
we
will
retain
the
authority
to
review
and
approve/
disapprove
the
low­
risk
subcategory
eligibility
demonstrations
submitted
by
PCWP
facilities.

With
regard
to
the
title
V
permit
programs
being
behind
the
statute's
permit
issuance
deadlines,
the
incorporation
of
the
NESHAP
requirements
is
a
necessary
step
that
will
require
some
resources.
Inclusion
of
the
low­
risk
subcategory
delisting
should
be
a
straightforward
part
of
360
the
process
and
should
not
cause
significant
delay.

V.
Statutory
and
Executive
Order
Reviews
A.
Executive
Order
12866:
Regulatory
Planning
and
Review
Under
Executive
Order
12866
(
58
FR
51735
(
October
4,

1993)),
the
Agency
must
determine
whether
the
regulatory
action
is
"
significant"
and,
therefore,
subject
to
OMB
review
and
the
requirements
of
the
Executive
Order.
The
Order
defines
"
significant
regulatory
action"
as
one
that
is
likely
to
result
in
a
rule
that
may:

(
1)
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;

(
2)
create
a
serious
inconsistency
or
otherwise
interfere
with
an
action
taken
or
planned
by
another
agency;

(
3)
materially
alter
the
budgetary
impact
of
entitlements,
grants,
user
fees,
or
loan
programs,
or
the
rights
and
obligation
of
recipients
thereof;
or
(
4)
raise
novel
legal
or
policy
issues
arising
out
of
legal
mandates,
the
President's
priorities,
or
the
principles
set
forth
in
the
Executive
Order.

Pursuant
to
the
terms
of
Executive
Order
12866,
it
has
been
determined
that
the
final
rule
is
a
"
significant
regulatory
action"
because
the
annual
costs
of
complying
361
with
the
final
rule
are
expected
to
exceed
$
100
million.
As
such,
this
action
was
submitted
to
OMB
for
EO
12866
review.

Changes
made
in
response
to
OMB
suggestions
or
recommendations
are
documented
in
the
public
record
(
see
ADDRESSEES
section
of
this
preamble).

We
did
not
estimate
health
and
welfare
benefits
associated
with
changes
in
emissions
of
HAP,
CO,
VOC,
PM,

NOX
and
SO2
for
the
final
rule.

B.
Paperwork
Reduction
Act
The
information
collection
requirements
in
the
final
rule
have
been
submitted
for
approval
to
the
OMB
under
the
Paperwork
Reduction
Act,
44
U.
S.
C.
3501
et
seq.
(
ICR
1984.02)
The
information
collection
requirements
are
not
enforceable
until
OMB
approves
them.

The
information
requirements
are
based
on
notification,

recordkeeping,
and
reporting
requirements
in
the
NESHAP
General
Provisions
(
40
CFR
part
63,
subpart
A),
which
are
mandatory
for
all
operators
subject
to
national
emission
standards.
These
recordkeeping
and
reporting
requirements
are
specifically
authorized
by
section
114
of
the
CAA
(
42
U.
S.
C.
7414).
All
information
submitted
to
the
EPA
pursuant
to
the
recordkeeping
and
reporting
requirements
for
which
a
claim
of
confidentiality
is
made
is
safeguarded
according
to
Agency
policies
set
forth
in
40
CFR
part
2,

subpart
B.
362
Today's
final
rule
will
require
maintenance
inspections
of
the
control
devices
but
will
not
require
any
notifications
or
reports
beyond
those
required
by
the
NESHAP
General
Provisions.
The
recordkeeping
requirements
require
only
the
specific
information
needed
to
assure
compliance.

The
annual
monitoring,
reporting,
and
recordkeeping
burden
for
this
collection
(
averaged
over
the
first
3
years
after
the
effective
date
of
the
rule)
is
estimated
to
be
4,692
labor
hours
per
year,
at
a
total
annual
cost
of
$
250,528.
This
estimate
includes
notifications
that
facilities
are
subject
to
the
rule;
notifications
of
performance
tests;
notifications
of
compliance
status,

including
the
results
of
performance
tests
and
other
initial
compliance
demonstrations
that
do
not
include
performance
tests;
SSM
reports;
semiannual
compliance
reports;
and
recordkeeping.
In
addition
to
the
requirements
of
40
CFR
part
63,
subpart
A,
facilities
that
wish
to
implement
emissions
averaging
provisions
must
submit
an
EAP.

Facilities
may
also
submit
a
request
for
a
routine
control
device
maintenance
exemption
to
justify
the
need
for
routine
maintenance
on
the
control
device
and
to
show
how
the
facilities
plan
to
minimize
emissions
to
the
greatest
extent
possible
during
the
maintenance.
The
average
number
of
respondents
during
the
3­
year
period
after
the
effective
date
of
the
rule
is
220,
and
the
average
number
of
responses
363
estimated
to
be
submitted
is
197.
The
resulting
estimated
burden
per
response
is
24
hours.
Total
capital/
startup
costs
associated
with
the
testing,
monitoring,
reporting,

and
recordkeeping
requirements
over
the
3­
year
period
of
the
ICR
are
estimated
to
be
$
122,040,
with
operation
and
maintenance
costs
of
$
5,178.

Burden
means
the
total
time,
effort,
or
financial
resources
expended
by
persons
to
generate,
maintain,
retain,

or
disclose
or
provide
information
to
or
for
a
Federal
agency.
This
includes
the
time
needed
to
review
instructions;
develop,
acquire,
install,
and
utilize
technology
and
systems
for
the
purposes
of
collecting,

validating,
and
verifying
information,
processing
and
maintaining
information,
and
disclosing
and
providing
information;
adjust
the
existing
ways
to
comply
with
any
previously
applicable
instructions
and
requirements;
train
personnel
to
be
able
to
respond
to
a
collection
of
information;
search
data
sources;
complete
and
review
the
collection
of
information;
and
transmit
or
otherwise
disclose
the
information.

An
agency
may
not
conduct
or
sponsor,
and
a
person
is
not
required
to
respond
to,
a
collection
of
information
unless
it
displays
a
currently
valid
OMB
control
number.

The
OMB
control
numbers
for
EPA's
regulations
are
listed
in
40
CFR
part
9
and
48
CFR
chapter
15.
The
OMB
control
364
numbers
for
the
information
collection
requirements
in
the
final
rule
will
be
listed
in
an
amendment
to
40
CFR
part
9
in
a
subsequent
Federal
Register
document
after
OMB
approves
the
ICR.

C.
Regulatory
Flexibility
Analysis
The
EPA
has
determined
that
it
is
not
necessary
to
prepare
a
regulatory
flexibility
analysis
in
connection
with
this
final
rule.
The
EPA
has
also
determined
that
the
final
rule
will
not
have
a
significant
economic
impact
on
a
substantial
number
of
small
entities.

For
purposes
of
assessing
the
impacts
of
today's
final
rule
on
small
entities,
small
entity
is
defined
as:
(
1)
a
small
business
ranging
from
500
to
750
employees
depending
on
the
businesses
NAICS
code;
(
2)
a
small
governmental
jurisdiction
that
is
a
government
of
a
city,
county,
town,

school
district
or
special
district
with
a
population
of
less
than
50,000;
and
(
3)
a
small
organization
that
is
any
not­
for­
profit
enterprise
which
is
independently
owned
and
operated
and
is
not
dominant
in
its
field.

After
considering
the
economic
impact
of
today's
final
rule
on
small
entities,
EPA
has
concluded
that
this
action
will
not
have
a
significant
economic
impact
on
a
substantial
number
of
small
entities.
We
have
determined
that,
based
on
SBA
size
definitions
for
the
affected
industries
and
reported
sales
and
employment
data,
17
of
the
52
companies,
365
or
32
percent,
owning
affected
facilities
are
small
businesses.
Although
small
businesses
represent
32
percent
of
the
companies
within
the
source
category,
they
are
expected
to
incur
8
percent
of
the
total
industry
compliance
costs
of
$
142
million.
There
are
three
small
firms
with
compliance
costs
equal
to
or
greater
than
3
percent
of
their
sales.
In
addition,
there
are
seven
small
firms
with
costto
sales
ratios
between
1
and
3
percent.

We
performed
an
economic
impact
analysis
to
estimate
the
changes
in
product
price
and
production
quantities
for
the
firms
affected
by
this
rule.
The
analysis
shows
that
of
the
32
facilities
owned
by
affected
small
firms,
one
small
firm
would
be
expected
to
shut
down
rather
than
incur
the
cost
of
compliance
with
the
rule.
Although
any
facility
closure
is
cause
for
concern,
it
should
be
noted
that
the
baseline
economic
condition
of
the
facilities
predicted
to
close
affects
the
closure
estimate
provided
by
the
economic
model.

Facilities
which
are
already
experiencing
adverse
economic
conditions
for
reasons
unconnected
to
this
rule
are
more
vulnerable
to
the
impact
of
any
new
costs
than
those
that
are
not.

The
analysis
indicates
that
the
final
rule
should
not
generate
a
significant
economic
impact
on
a
substantial
number
of
small
entities
for
the
PCWP
manufacturing
source
category
for
the
following
reasons.
First,
of
the
ten
small
366
firms
that
have
compliance
costs
greater
than
1
percent
of
sales,
three
small
firms
have
compliance
costs
of
greater
than
3
percent
of
sales.
Second,
the
results
of
the
economic
impact
analysis
show
that
one
facility
owned
by
a
small
firm
out
of
the
32
facilities
owned
by
affected
small
firms
may
close
due
to
the
implementation
of
the
final
rule.

The
facility
that
may
close
rather
than
incur
the
cost
of
compliance
appears
to
have
low
profitability
levels
currently.
It
also
should
be
noted
that
the
estimate
of
compliance
costs
for
this
facility
is
likely
to
be
an
overestimate
due
to
the
lack
of
facility­
specific
data
available
to
assign
a
precise
control
cost
in
this
case.

Although
the
final
rule
will
not
have
a
significant
economic
impact
on
a
substantial
number
of
small
entities,

EPA
nonetheless
has
tried
to
reduce
the
impact
of
the
rule
on
small
entities.
First,
we
considered
subcategorization
based
on
production
and
throughput
level
to
determine
whether
smaller
process
units
would
have
a
different
MACT
floor
than
larger
process
units.
Our
data
show
that
subcategorization
based
on
size
would
not
result
in
a
less
stringent
level
of
control
for
the
smaller
process
units.

Second,
we
chose
to
set
the
control
requirements
at
the
MACT
floor
control
level
and
not
at
a
control
level
more
stringent.
Thus,
the
control
level
specified
in
the
final
PCWP
rule
is
the
least
stringent
allowed
by
the
CAA.
Third,
367
the
final
rule
contains
multiple
compliance
options
to
provide
facilities
with
the
flexibility
to
comply
in
the
least
costly
manner
while
maintaining
a
workable
and
enforceable
rule.
The
compliance
options
include
emissions
averaging
and
PBCO
which
allow
inherently
low­
emitting
process
units
to
comply
without
installing
add­
on
control
devices
and
facilities
to
use
innovative
technology
and
P2
methods.
Fourth,
the
final
rule
includes
multiple
test
method
options
for
measuring
methanol,
formaldehyde,
and
total
HAP.
Fifth,
the
final
rule
allows
PCWP
facilities
to
demonstrate
eligibility
for
the
delisted
low­
risk
subcategory
and
thereby
avoid
MACT
altogether.
In
addition,

we
worked
with
various
trade
associations
during
the
development
of
the
final
rule.

As
discussed
in
earlier
sections
of
this
preamble,
we
present
the
impacts
of
the
rule
associated
with
allowing
PCWP
facilities
to
demonstrate
eligibility
for
the
delisted
low­
risk
subcategory
and
thereby
avoid
MACT
altogether.
The
number
of
small
businesses
impacted
is
reduced
to
seven
from
the
original
17,
and
the
total
number
of
businesses
impacted
is
reduced
to
42,
down
from
the
original
52.
Small
businesses
represent
17
percent
of
the
companies
within
the
source
category,
which
is
down
from
the
32
percent
estimate
for
the
final
rule.
These
small
businesses
are
expected
to
incur
4
percent
of
the
total
industry
compliance
costs
of
368
$
75
million
(
the
costs
considering
inclusion
of
the
delisted
low­
risk
subcategory).
There
are
no
small
firms
with
compliance
costs
equal
to
or
greater
than
3
percent
of
their
sales
as
compared
to
three
for
the
final
rule.
In
addition,

there
are
four
small
firms
with
cost­
to­
sales
ratios
between
1
and
3
percent,
which
is
down
from
seven
for
the
final
rule.

D.
Unfunded
Mandates
Reform
Act
Title
II
of
the
Unfunded
Mandates
Reform
Act
of
1995
(
UMRA),
Public
Law
104­
4,
establishes
requirements
for
Federal
agencies
to
assess
the
effects
of
their
regulatory
actions
on
State,
local,
and
tribal
governments
and
the
private
sector.
Under
section
202
of
the
UMRA,
EPA
generally
must
prepare
a
written
statement,
including
a
cost­
benefit
analysis,
for
proposed
and
final
rules
with
"
Federal
mandates"
that
may
result
in
expenditures
by
State,

local,
and
Tribal
governments,
in
aggregate,
or
by
the
private
sector,
of
$
100
million
or
more
in
any
1
year.

Before
promulgating
an
EPA
rule
for
which
a
written
statement
is
needed,
section
205
of
the
UMRA
generally
requires
EPA
to
identify
and
consider
a
reasonable
number
of
regulatory
alternatives
and
adopt
the
least­
costly,
most
cost­
effective,
or
least­
burdensome
alternative
that
achieves
the
objectives
of
the
rule.
The
provisions
of
section
205
do
not
apply
when
they
are
inconsistent
with
369
applicable
law.
Moreover,
section
205
allows
EPA
to
adopt
an
alternative
other
than
the
least­
costly,
most
cost­
effective,
or
least­
burdensome
alternative
if
the
Administrator
publishes
with
the
final
rule
an
explanation
why
that
alternative
was
not
adopted.
Before
EPA
establishes
any
regulatory
requirements
that
may
significantly
or
uniquely
affect
small
governments,

including
tribal
governments,
it
must
have
developed
under
section
203
of
the
UMRA
a
small
government
agency
plan.
The
plan
must
provide
for
notifying
potentially
affected
small
governments,
enabling
officials
of
affected
small
governments
to
have
meaningful
and
timely
input
in
the
development
of
EPA
regulatory
proposals
with
significant
Federal
intergovernmental
mandates,
and
informing,

educating,
and
advising
small
governments
on
compliance
with
the
regulatory
requirements.

Since
the
final
rule
is
estimated
to
impose
costs
to
the
private
sector
in
excess
of
$
100
million
per
year,
it
is
considered
a
significant
regulatory
action.
Therefore,
we
have
prepared
the
following
statement
with
respect
to
sections
202
through
205
of
the
UMRA.
370
1.
Statutory
Authority
This
final
rule
establishes
control
requirements
for
existing
and
new
PCWP
sources
pursuant
to
section
112
of
the
CAA.
The
CAA
requires
NESHAP
to
reflect
the
maximum
degree
of
reduction
in
emissions
of
HAP
that
is
achievable.
This
is
commonly
referred
to
as
MACT.
Section
112(
d)(
3)
of
the
CAA
further
defines
a
minimum
level
of
control
that
can
be
considered
for
MACT
standards,
commonly
referred
to
as
the
MACT
floor,
which
for
new
sources
is
the
level
of
control
achieved
by
the
best
controlled
similar
source,
and
for
existing
sources
is
the
level
of
control
achieved
by
the
average
of
the
best
performing
12
percent
of
sources
in
the
category
(
or
the
best­
performing
five
sources
for
categories
with
fewer
than
30
sources).

Control
technologies
and
their
performance
are
discussed
in
the
background
information
document
for
this
proposal
(
Docket
numbers
A­
98­
44
and
OAR­
2003­
0048).
We
considered
emission
reductions,
costs,
environmental
impacts,
and
energy
impacts
in
selecting
the
MACT
standards.
The
final
