"
Leiter,
Amanda"
<
aleiter@
nrdc.
org>
11/
11/
2004
10:
54
AM
To
Michael
Leavitt/
DC/
USEPA/
US@
EPA,
Jeff
Holmstead/
DC/
USEPA/
US@
EPA,
Bill
Wehrum/
DC/
USEPA/
US@
EPA,
Mary
Kissell/
RTP/
USEPA/
US@
EPA,
Patricia
Embrey/
DC/
USEPA/
US@
EPA,
Jim
Eddinger/
RTP/
USEPA/
US@
EPA,
Group
A­
AND­
R­
DOCKET@
EPA,
a­
r­
docket@
epamail.
epa.
gov
Cc
"
Leiter,
Amanda"
<
aleiter@
nrdc.
org>

Subject
OAR­
2002­
0058
(
A­
96­
47)
Reconsideration
Petition
Amanda
Leiter
Natural
Resources
Defense
Council
tel:
202
289
2398
fax:
202
289
1060
2
November
11,
2004
Michael
O.
Leavitt
Administrator,
Environmental
Protection
Agency
1101A
EPA
Headquarters
Ariel
Rios
Building
1200
Pennsylvania
Ave.,
NW
Washington,
DC
20460
BY
ELECTRONIC
MAIL
Dear
Administrator
Leavitt:

This
is
a
petition
under
Clean
Air
Act
§
307(
d)(
7)(
B),
42
U.
S.
C.
§
7607(
d)(
7)(
B),
submitted
by
the
Natural
Resources
Defense
Council,
1200
New
York
Avenue,
NW,
Ste.
400,
Washington,
DC
20005,
and
the
Environmental
Integrity
Project,
919
18th
Street,
NW,
Ste.
975,
Washington,
DC
20006.
By
this
petition,
we
request
that
you
reconsider
certain
aspects
of
the
final
action
taken
at
69
Fed.
Reg.
55,218
et
seq.
(
September
13,
2004)
and
entitled
National
Emission
Standards
for
Hazardous
Air
Pollutants
for
Industrial,
Commercial,
and
Institutional
Boilers
and
Process
Heaters.

Sincerely,

Amanda
Leiter
Eric
Schaeffer
Natural
Resources
Defense
Council
Environmental
Integrity
Project
3
BEFORE
THE
ADMINISTRATOR
UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
__________________________________________
)
In
the
Matter
of
the
Final
Rule:
)
)
OAR­
2002­
0058
National
Emission
Standards
for
Hazardous
Air
)
A­
96­
47
Pollutants
for
Industrial,
Commercial,
and
)
Institutional
Boilers
and
Process
Heaters
)
__________________________________________)

PETITION
FOR
RECONSIDERATION
Pursuant
to
Section
307(
d)(
7)(
B)
of
the
Clean
Air
Act,
1
the
Natural
Resources
Defense
Council
and
the
Environmental
Integrity
Project
hereby
petition
the
Administrator
of
the
Environmental
Protection
Agency
(
the
Administrator,
EPA,
or
the
Agency)
to
reconsider
the
National
Emission
Standards
for
Hazardous
Air
Pollutants
(
NESHAP)
captioned
above
and
published
at
69
Fed.
Reg.
55,218
(
September
13,
2004).

The
Clean
Air
Act
(
CAA)
directs
EPA
to
establish
emissions
standards
for
each
category
of
major
sources
of
hazardous
air
pollutants
(
HAPs).
These
standards,
which
must
reflect
the
maximum
achievable
control
technology
(
MACT),
are
designed
to
obligate
each
source
in
a
category
to
limit
its
emissions
to
a
level
commensurate
with
the
best
performers
in
the
industry.
In
this
rule,
however,
EPA
sets
only
25
of
the
72
MACT
standards
that
the
Agency
itself
identifies
as
necessary
to
control
emissions
from
industrial,
commercial,
and
institutional
(
ICI)
boilers
and
process
heaters.
2
That
is,
the
1
42
U.
S.
C.
§
7607(
d)(
7)(
B).
2
These
72
standards
correspond
to
emissions
of
(
1)
non­
mercury
metallic
HAP,
mercury,
inorganic
HAP,
or
organic
HAP
from
(
2)
existing
or
new­
or­
reconstructed
units
(
3)
that
are
large,
small,
or
limited­
use,
and
(
4)
that
burn
solid,
liquid,
or
gaseous
fuels.
4
agency
entirely
fails
to
adopt
47
of
the
72
necessary
control
standards.
3
Moreover,
EPA
further
attempts
to
evade
the
MACT
requirement
by
creating
"
health­
based
compliance
alternatives"
for
two
significant
HAPs,
hydrogen
chloride
(
HCl)
and
manganese
(
Mn),

and
allowing
individual
large,
solid­
fuel
sources
to
choose
whether
to
meet
the
categorywide
MACT
standard
for
these
HAPs
or
instead
to
submit
data
demonstrating
compliance
with
the
more
lenient
health­
based
alternatives.
By
declining
to
set
most
of
the
necessary
standards,
and
by
creating
individualized,
risk­
based
MACT
exemptions4
for
large,

solidfuel
sources
of
HCl
and
Mn,
EPA
fundamentally
subverts
the
CAA's
MACT
program,

which
depends
for
its
success
on
the
promulgation
of
a
comprehensive
set
of
easilyenforced
category­
wide,
technology­
based
HAP
standards.

We
seek
reconsideration
of
the
rule
on
seven
principal
grounds:
(
1)
EPA
has
no
legal
justification
for
its
refusal
to
adopt
almost
two­
thirds
of
the
standards
necessary
to
regulate
HAP
emissions
from
ICI
boilers
and
process
heaters;
(
2)
the
CAA
does
not
authorize
plant­
by­
plant,
risk­
based
exemptions
from
otherwise
applicable,

categorywide
emissions
standards;
(
3)
even
if
such
exemptions
were
otherwise
lawful,
EPA
may
not
adopt
them
for
pollutants
like
HCl
and
Mn,
for
which
the
necessary
health
thresholds
are
not
established;
(
4)
even
if
such
thresholds
were
established,
they
would
likely
be
exceeded
under
this
rule,
because
the
risk­
based
exemptions
fail
to
account
for
alternative
sources
and
background
levels
of
HCl
and
Mn;
(
5)
compounding
the
two
preceding
problems,
EPA
uses
HCl
as
a
surrogate
for
other
(
unidentified)
non­
metallic,
inorganic
pollutants,
for
which
health
thresholds
may
or
may
not
be
established;
(
6)
the
Agency
has
3
We
do
not
concede
that
EPA
is
correct
to
identify
only
72
standards
as
necessary
to
control
emissions
from
ICI
boilers
and
process
heaters.
Rather,
we
argue
that
once
the
Agency
has
identified
those
72
standards,
it
cannot
evade
its
statutory
responsibility
to
set
numerical
emissions
limits
for
each
one.
4
EPA
itself
calls
the
health­
based
alternatives
"
exemptions"
in
various
places
in
the
rule.
See,
e.
g.,
69
Fed.
Reg.
at
55,240­
41.
5
structured
the
Mn
exemption
in
such
a
way
that
it
effectively
permits
plants
with
low
Mn
emissions
to
avoid
controlling
emissions
of
other
non­
mercury
metals;
and
finally,
(
7)
the
described
procedures
for
demonstrating
compliance
with
the
§
112(
d)(
4)
risk­
based
exemptions
are
significantly
flawed.

Reconsideration
of
these
seven
issues
is
appropriate
because
they
are
of
central
relevance
to
the
outcome
of
the
rule,
and
most
could
not
practicably
have
been
raised
during
the
public
comment
period.
The
Administrator
must
therefore
"
convene
a
proceeding
for
reconsideration
of
the
rule
and
provide
the
same
procedural
rights
as
would
have
been
afforded
had
the
information
been
available
at
the
time
the
rule
was
proposed."
5
In
addition,
we
ask
that
EPA
stay
the
effectiveness
of
the
rule's
risk­
based
exemptions.
A
stay
is
warranted
because
under
the
final
rule,
existing
sources
of
HCl
and
Mn
that
wish
to
take
advantage
of
the
exemptions
must
submit
the
relevant
data
to
EPA
"
one
year
prior"
6
to
the
rule's
fixed
compliance
date
for
such
sources
(
September
13,
20077),
and
new
or
reconstructed
sources
that
begin
operation
after
their
compliance
date
(
November
12,
20048)
must
take
similar
action
"
within
180
days"
of
startup.
9
EPA
should
not
encourage
members
of
the
regulated
community
to
waste
time
and
resources
demonstrating
that
they
meet
unlawful
risk­
based
exemptions
when
those
exemptions
are
legally
unjustified.
Nor
should
the
Agency
permit
sources
to
delay
compliance
with
the
binding
MACT
standards
for
HCl
and
Mn
based
on
a
false
hope
that
the
exemptions
will
5
42
U.
S.
C.
§
7607(
d)(
7)(
B).
Simultaneously
with
this
petition,
we
are
filing
a
petition
for
review
of
this
rule
in
the
U.
S.
Court
of
Appeals
for
the
D.
C.
Circuit.
6
69
Fed.
Reg.
at
55,284.
7
Id.
at
55,254.
8
Id.
9
Id.
at
55,284.
6
withstand
legal
challenge.
Thus,
EPA
should
exercise
its
authority
under
section
307(
d)(
7)(
B)
of
the
CAA10
to
stay
the
effectiveness
of
the
exemptions
during
the
course
of
the
reconsideration
proceedings.

Indeed,
given
the
likely
success
of
this
petition
for
reconsideration,
EPA
should
not
only
grant
a
limited
stay
under
section
307(
d)(
7)(
B)
but
also
take
the
necessary
administrative
steps
to
stay
the
risk­
based
exemptions
indefinitely
pending
judicial
review.
Such
exemptions
are
completely
at
odds
with
the
basic
purpose
and
structure
of
the
MACT
requirements.
EPA
should
not
ask
sources
to
embark
on
a
fool's
errand
of
demonstrating
compliance
with
these
unlawful
alternative
standards,
nor
should
the
Agency
waste
limited
federal
resources
reviewing
and
approving
sources'
compliance
data.

1.
EPA
Must
Reconsider
Its
Adoption
of
Numerous
"
No
Control"
Standards.

a.
EPA
has
a
clear
statutory
obligation
to
adopt
MACT
standards
for
each
HAP
emitted
by
a
major
source
in
a
listed
category
or
subcategory.

The
language
of
the
CAA
is
unequivocal:
Under
subsections
112(
c)(
2)
and
(
d)(
1),

EPA
"
shall
promulgate
regulations
establishing
emission
standards
for
each
category
or
subcategory
of
major
sources
and
area
sources
of
hazardous
air
pollutants
listed
for
regulation."
11
As
the
D.
C.
Circuit
has
observed,
this
amounts
to
a
"
clear
statutory
obligation
to
set
emission
standards
for
each
listed
HAP."
12
Moreover,
the
CAA
strictly
cabins
the
Agency's
discretion
in
setting
these
standards.
For
new
HAP
sources,
EPA
must
adopt
a
standard
no
"
less
stringent
than
the
emission
control
that
is
achieved
in
10
42
U.
S.
C.
§
7607(
d)(
7)(
B)
("
The
effectiveness
of
the
rule
may
be
stayed
during
 
reconsideration
 
by
the
Administrator
 
for
a
period
not
to
exceed
three
months.").
11
Id.
§
7412(
d)(
1)
(
emphasis
added).
12
National
Lime
Assn.
v.
EPA,
233
F.
3d
625,
634
(
D.
C.
Cir.
2000)
(
National
Lime).
7
practice
by
the
best
controlled
similar
source,"
13
and
for
existing
sources,
it
must
adopt
a
standard
no
"
less
stringent
 
than
 
the
average
emission
limitation
achieved
by
the
best
performing
12
percent
of
the
existing
sources
 
in
the
category
or
subcategory."
14
b.
The
final
rule
falls
far
short
of
meeting
that
obligation.

Despite
the
clarity
of
these
statutory
obligations,
the
ICI
boiler
and
process
heater
rule
fails
to
establish
numerical
standards
for
almost
two­
thirds
of
the
boiler­
size/

fueltype
HAP
groupings
listed
in
the
rule.
EPA
first
indicates
that
it
plans
to
regulate
(
1)

existing
and
new­
or­
reconstructed
units,
(
2)
that
are
large,
small,
or
limited­
use,
(
3)
that
burn
solid,
liquid,
or
gaseous
fuel,
and
(
4)
that
emit
non­
mercury
metallic
HAPs
(
arsenic,

beryllium,
cadmium,
chromium,
lead,
manganese,
nickel,
and
selenium15),
mercury,

inorganic
HAPs
(
predominantly
HCl16),
or
organic
HAPs
(
predominantly
formaldehyde,

benzene,
and
acetaldehyde17).
To
regulate
each
of
these
groupings,
EPA
would
need
to
adopt
72
MACT
standards,
ranging
from
a
standard
for
non­
mercury
metallic
HAP
emissions
from
existing,
large,
solid­
fuel
units
to
a
standard
for
organic
HAP
emissions
from
new,
limited­
use,
gaseous
units.
Yet
EPA
sets
numerical
standards
for
only
25
of
these
groupings,
adopting
a
"
no
control"
standard
for
the
remaining
47.18
Thus,
the
rule
violates
the
Act.

c.
Reconsideration
is
warranted
because
the
proposed
rule
failed
to
identify
or
explain
the
supposed
legal
basis
for
adoption
of
"
no
control"
standards.

13
42
U.
S.
C.
§
7412(
d)(
3).
14
Id.
§
7412(
d)(
3)(
A).
For
subcategories
with
fewer
than
30
sources,
the
Agency
must
instead
adopt
a
standard
"
no
less
stringent
 
than
 
the
average
emission
limitation
achieved
by
the
best
performing
5
sources."
Id.
§
7412(
d)(
3)(
B).
15
69
Fed.
Reg.
at
55,220.
16
Id.
17
Id.
18
See
69
Fed.
Reg.
at
55269­
270,
Table
1
to
Subpart
DDDD
of
Part
63;
Memorandum
from
Roy
Oommen,
Eastern
Research
Group,
to
Jim
Eddinger,
U.
S.
EPA,
"
MACT
Floor
Analysis
for
the
Industrial,
Commercial,
and
Institutional
Boilers
and
Process
Heaters
National
Emission
Standards
for
Hazardous
Air
Pollutants"
(
October
2002),
OAR­
2002­
0058­
0028,
at
table
8­
1
(
summarizing
MACT
floor
emission
limits
for
boiler­
size/
fuel­
type/
HAP
groupings).
8
In
the
proposed
ICI
boiler
and
process
heater
rule,
19
the
Agency
signaled
its
intention
to
set
"
no
control"
standards
for
many
of
the
identified
boiler­
size/

fueltype
HAP
groupings.
The
proposed
rule
proffered
various
justifications
for
this
dereliction
of
duty.
For
example,
for
new,
small,
solid­
fuel­
fired
units,
EPA
indicated
that
"
no
control
technology
being
used
in
the
existing
population
of
boilers
and
process
heaters
 
consistently
achieved
lower
emission
rates
than
uncontrolled
levels,
such
that
a
best
controlled
similar
source
for
organic
HAP
could
be
identified,"
so
"
the
MACT
floor
for
new
sources
in
this
subcategory
is
no
emissions
reductions
for
organic
HAP."
20
For
similar
reasons,
EPA
proposed
adopting
a
"
no
control"
standard
for
mercury
emissions
from
new,
liquid­
fired
units,
21
and
for
all
emissions
except
organic
HAPs
from
new,
gas­
fired
units.
22
For
existing
units,
the
Agency
proposed
adopting
"
no
control"

standards
for
all
but
4
of
the
36
boiler­
size/
fuel­
type/
HAP
groupings.
For
some
groupings,
the
Agency
justified
this
omission
by
noting
that
"
fewer
than
6
percent
of
units
in
[
the]
grouping
used
controls
or
limited
emissions,
[
so]
the
median
unit
for
[
the]

grouping
reflect[
ed]
no
emissions
reduction."
23
For
other
groupings,
EPA
observed
that
"
uncontrolled
units
(
or
units
with
low
efficiency
add­
on
controls)
were
 
identified
as
being
among
the
best
performing
12
percent
of
sources"
due
solely
to
"
characteristics
of
the
fuel
that
they
burn"
rather
than
the
efficacy
of
their
control
technologies.
24
Anticipating
the
obvious
rejoinder
to
this
observation
 
the
Agency
should
require
fuel
switching
as
part
of
the
MACT
standard
 
EPA
stated
only
that
while
such
switching
19
68
Fed.
Reg.
1660
(
Jan.
13,
2003).
20
Id.
at
1682.
21
Id.
at
1683­
84.
22
Id.
at
1684.
23
E.
g.,
id.
at
1676.
24
68
Fed.
Reg.
at
1672.
9
would
decrease
some
HAP
emissions,
it
might
increase
others,
and
"
assess[
ing]
the
relative
risk
associated
with
each
HAP
emitted,
and
determin[
ing]
whether
requiring
[
fuel
switching]
would
result
in
overall
lower
risk,"
would
require
"
analysis
 
not
appropriate
at
this
stage
in
the
regulatory
process."
25
Various
commenters
to
the
proposed
rule
identified
the
central
flaw
in
these
justifications:
As
the
D.
C.
Circuit
has
repeatedly
recognized,
26
the
CAA
requires
promulgation
of
a
numerical
standard
for
HAP
emissions
from
each
identified
major
source
category
or
subcategory,
even
if
existing
plants
in
the
category
or
subcategory
do
not
provide
a
ready
example
of
a
control
technology
adequate
to
achieve
that
standard.

That
is,
EPA's
sole
legal
obligation
in
this
respect
is
to
adopt
numerical
HAP
standards
that
reflect
the
emissions
levels
achieved
by
the
best
performing
plants
in
the
category
or
subcategory;
27
the
Agency
need
not
consider
how
individual
plants
will
achieve
that
standard
(
i.
e.,
by
adopting
control
technologies,
switching
fuels,
implementing
process
changes,
or
some
combination
of
the
above),
and
it
certainly
may
not
absolve
itself
of
the
obligation
to
set
MACT
standards
simply
because
those
standards
will
pose
a
greater
challenge
for
some
plants
than
for
others.

Making
this
point
in
its
comments
to
the
proposed
rule,
Earthjustice
observed:

EPA
argues
that
it
could
not
set
floors
at
the
emission
levels
achieved
by
the
best
performing
sources
because
these
sources
did
not
necessarily
have
the
best
add­
on
control
equipment.
68
Fed.
Reg.
at
1672.
The
agency
goes
on
to
claim
that
the
best
performing
units
are
not
necessarily
the
"
best­
controlled"
because
they
may
have
lower
efficiency
end­
of­
stack
control
equipment.
Id.
EPA
completely
misreads
the
mandate
of
§
112'
s
25
Id.
26
See,
e.
g.,
Cement
Kiln
Recycling
Coalition
v.
EPA,
255
F.
3d
855,
861,
865
(
D.
C.
Cir.
2001)
(
CKRC);
National
Lime,
233
F.
3d
at
633­
34.
27
See
CKRC,
255
F.
3d
at
865
("
Section
7412(
d)(
3)
requires
only
that
EPA
set
floors
at
the
emission
level
achieved
by
the
best­
performing
sources.
If
EPA
cannot
meet
this
requirement
using
the
MACT
methodology,
it
must
devise
a
different
approach
capable
of
producing
floors
that
satisfy
the
[
CAA].").
10
floor
requirement
 
a
remarkable
achievement
given
that
the
D.
C.
Circuit
has
now
explained
that
requirement
three
times.
Floors
under
§
112
must
reflect
the
emission
levels
actually
achieved
by
the
"
best
performing"
sources
 
i.
e.,
those
with
the
lowest
emission
levels
 
not
the
ones
that
EPA
deems
to
be
the
"
best
controlled."
[
Cement
Kiln
Recycling
Coalition
v.
EPA,
255
F.
3d
855,
861,
865
(
D.
C.
Cir.
2001)
(
CKRC)].
It
does
not
matter
how
these
sources
achieve
their
superior
emission
levels;
they
may
do
so
by
burning
a
cleaner
fuel,
by
being
better
designed
or
newer,
by
being
better
maintained
or
operated,
by
using
better
end­
of­
stack
control
technologies,
or
by
using
more
than
one
control
technology.
EPA's
simple
but
mandatory
task
is
to
identify
the
relevant
best
performing
sources
 
regardless
of
how
they
are
achieving
their
superior
emission
levels
 
and
set
standards
reflecting
the
average
emission
level
these
sources
are
achieving.
Thus,
EPA's
contention
that
floors
reflecting
the
emission
levels
achieved
by
the
best
performers
may
not
reflect
what
is
achievable
for
all
sources
through
using
a
chosen
control
technology
is
irrelevant.
Further,
EPA's
contention
that
floors
must
reflect
what
is
achievable
through
using
a
chosen
control
technology
is
flatly
unlawful;
indeed,
that
contention
has
already
been
rejected
twice
by
the
D.
C.
Circuit.
CKRC,
255
F.
3d
at
861,
865;
National
Lime
Assn.
[
v.
EPA,
233
F.
3d
625,
633
(
D.
C.
Cir.
2000)
(
National
Lime).]
28
Nowhere
in
the
proposed
rule
did
EPA
address
this
central
legal
defect
in
its
chosen
approach.
Indeed,
in
direct
contravention
of
CAA
§
307(
d)(
3)(
C),
which
requires
that
any
proposed
rule
"
include
a
summary"
of
"
major
 
underlying
 
legal
interpretations,"
29
the
proposed
rule
does
not
mention
the
two
key
D.
C.
Circuit
decisions,

CKRC
and
National
Lime,
even
though
the
Agency's
proposed
course
of
action
departs
from
the
holdings
in
those
cases
in
a
manner
that
demands
explanation.

In
the
final
rule,
the
Agency
tries
to
remedy
this
glaring
omission
by
offering
a
novel
 
and
insupportable
 
interpretation
of
National
Lime
(
though
the
Agency
again
fails
to
grapple
with
CKRC).
In
National
Lime,
the
court
considered
challenges
to
EPA
regulations
setting
emissions
standards
for
cement
manufacturing
facilities.
The
regulations
"
established
emission
floors
of
`
no
control'
for
HCl,
mercury,
and
total
28
Comments
of
Earthjustice
to
the
Proposed
NESHAP
for
ICI
Boilers
and
Process
Heaters,
OAR­
2002­
0058­
0451,
7­
8.
29
42
U.
S.
C.
§
7607(
d)(
3)(
C).
11
hydrocarbons
 
because
the
Agency
found
no
cement
plants
using
control
technologies
for
these
pollutants."
30
Observing
that
"[
n]
othing
in
the
[
CAA]
even
suggests
that
EPA
may
set
emission
levels
only
for
those
listed
HAPs
controlled
with
technology,"
the
court
found
the
"
no
control"
standards
"
contrary
to
the
[
CAA's]
plain
language"
and
remanded
the
regulations
"
for
EPA
to
 
set[]
emission
standards
for
those
pollutants."
31
EPA
attempts
to
distinguish
the
"
no
control"
standards
repudiated
in
National
Lime
from
those
adopted
in
the
final
ICI
boiler
and
process
heater
rule
as
follows:

In
the
National
Lime
case,
the
court
threw
out
EPA's
determination
of
a
no
control
floor
because
it
was
based
only
on
a
control
technology
approach.
The
court
stated
that
EPA
must
look
at
what
the
best
performers
[
in
the
category
or
subcategory]
achieve,
regardless
of
how
they
achieve
it.
Therefore,
our
determination
that
the
MACT
floor
for
certain
subcategories
or
HAP
is
`
no
emissions
reduction'
is
lawful
because
we
determined
that
the
best­
performing
sources
were
not
achieving
emissions
reductions
through
the
use
of
an
emission
control
system
and
there
were
no
other
appropriate
methods
by
which
boilers
and
process
heaters
could
reduce
HAP
emissions.
32
That
is,
apparently,
EPA
feels
it
is
justified
in
ignoring
§
112'
s
mandate
in
the
ICI
boiler
and
process
heater
context
because
the
Agency's
determination
of
"
no
control"
floors
in
this
instance
is
not
"
based
only
on
a
control
technology"
but
also
on
an
assessment
 
however
arbitrary,
self­
serving,
and
lacking
in
record
support
 
that
other
methods
of
achieving
emissions
reductions
are
not
"
appropriate"
(
a
term
the
Agency
never
defines).

This
cramped
reading
of
the
D.
C.
Circuit's
broad
language
is
untenable.
The
court
states
unequivocally
that
EPA
has
a
"
clear
statutory
obligation
to
set
emission
standards
for
each
listed
HAP."
33
The
court
does
observe
that
EPA
must
set
standards
30
233
F.
3d
at
632.
31
Id.
at
628,
633­
34.
32
69
Fed.
Reg.
at
55,233.
33
233
F.
3d
at
634.
12
even
"
for
HAPs
not
controlled
with
technology,"
34
but
to
read
this
requirement
(
set
standards
even
if
no
technology
is
readily
available
to
achieve
that
standard)
to
permit
EPA
not
to
set
standards
when
achieving
the
standards
might
require
a
combination
of
several
approaches
(
such
as
fuel
switching,
and/
or
process
changes,
and/
or
installation
of
technological
controls)
is
patently
absurd.

EPA
does
not
and
cannot
claim
that
the
best
performing
ICI
boilers
and
process
heaters
in
each
subcategory
perform
no
better
than
worse
performing
units.
The
Agency
says
only
(
1)
that
the
best
performing
plants
do
not
consistently
use
control
technologies
different
from
worse
performing
plants,
(
2)
that
process
changes
or
work
practices
"
would
be
ineffective
in
reducing
[
some]
fuel
related
HAP
emissions,"
35
and
(
3)
that
EPA
cannot
be
bothered
to
do
the
analysis
necessary
to
determine
whether
fuel
switching
would
produce
net
reductions
in
overall
HAP
emissions.
36
National
Lime
unequivocally
holds
that
such
excuses
do
not
absolve
the
Agency
of
its
statutory
obligation
to
set
MACT
standards
for
HAP
emissions
from
each
and
every
identified
category
or
subcategory
of
major
source.
37
Moreover,
CKRC
resolves
any
remaining
doubt
on
the
subject.
In
that
case,
the
court
held
that
in
setting
MACT
standards,
EPA
must
"
set
floors
at
the
emission
level
achieved
by
the
best­
performing
sources,"
whether
the
Agency
does
so
using
"
MACT
methodology"
or
by
"
consider[
ing]

factors
such
as
`
process
changes,
substitution
of
materials
or
other
modifications
 

34
Id.
35
69
Fed.
Reg.
at
55,233.
36
See
68
Fed.
Reg.
at
1672
(
claiming,
with
no
justification,
that
analysis
of
the
HAP­
related
consequences
of
fuel
switching
is
"
not
appropriate
at
this
stage
in
the
regulatory
process").
37
See
generally
233
F.
3d
at
633­
34.
13
design,
equipment,
work
practice,
or
operational
standards
 
[
or]
a
combination
of
the
above.'"
38
In
short,
EPA's
legal
justification
for
its
adoption
of
47
"
no
control"
standards
is
tortured
and
wholly
inadequate.
Because
 
contrary
to
§
307(
d)(
3)(
C)
39
 
the
Agency
first
stated
that
justification
in
the
final
rule,
we
were
unable
to
identify
the
specific
flaws
in
the
legal
analysis
until
after
the
comment
period
closed.
The
flaws
exposed
by
our
objections
go
to
the
heart
of
the
final
rule,
casting
significant
legal
doubt
on
the
"
standards"
adopted
for
47
of
the
72
boiler­
size/
fuel­
type/
HAP
groupings
identified
in
the
rule.
Accordingly,
the
Agency
must
"
convene
a
proceeding
for
reconsideration
of
[
those
47
standards]
and
provide
the
same
procedural
rights
as
would
have
been
afforded
had
the
information
been
available
at
the
time
the
rule
was
proposed."
40
2.
EPA
Must
Reconsider
the
Final
Rule's
Risk­
Based
Exemptions
for
HCl
and
Mn
Emissions.

a.
The
proposed
rule's
description
of
the
risk­
based
exemptions
was
inadequate
to
provide
public
notice
of
the
likely
contours
of
the
program.

The
final
ICI
boiler
and
process
heater
rule
creates
risk­
based
exemptions
for
large,
solid­
fuel
sources
of
two
important
HAPs,
HCl
and
Mn.
The
rule
lays
out
the
methodology
and
criteria
individual
plants
may
use
to
demonstrate
eligibility
for
these
exemptions
in
a
detailed,
five­
page
appendix,
Appendix
A.
41
The
following
examples
serve
to
illustrate
the
level
of
specificity
in
Appendix
A:

 
It
includes
two
tables
(
one
for
HCl
and
one
for
Mn),
each
of
which
lists
132
"
allowable
toxicity­
weighted
emission
rate[
s],"
corresponding
to
132
stack­

38
255
F.
3d
at
863­
65
(
quoting
42
U.
S.
C.
§
7412(
d0(
2)(
A)­(
E)).
39
42
U.
S.
C.
§
7607(
d)(
3)(
C).
40
Id.
§
7607(
d)(
7)(
B).
41
69
Fed.
Reg.
at
55,282.
14
height/
distance­
to­
property­
boundary
pairings,
for
use
in
determining
that
a
particular
facility's
HCl­
or
Mn­
equivalent
emissions
rates
qualify
for
the
riskbased
exemptions;
42
 
It
identifies
a
specific
EPA
technical
resource
document,
"
Air
toxics
Risk
Assessment
Reference
Library,
Volume
2,
Site­
Specific
Risk
Assessment,"
as
"[
a]
n
example
of
one
approach
for
performing
a
site­
specific
compliance
demonstration
for
air
toxics";
43
 
It
contains
detailed
instructions
on
how
monitoring
data
that
do
not
detect
HCl
or
Mn
(
so­
called
"
nondetect
data")
should
be
handled
in
determining
emissions
rates
from
a
source
facility;
44
and
 
It
lists
certain
elements
that
a
facility's
"
site­
specific
compliance
demonstration"
must
contain,
including
an
estimate
of
"
long­
term
inhalation
exposures"
from
the
facility,
and
an
estimate
of
"
inhalation
exposure
for
the
individual
most
exposed
to
the
facility's
emissions."
45
In
contrast
to
Appendix
A,
the
proposal's
discussion
of
risk­
based
exemptions
was
both
muddy
and
absurdly
scant.
For
example,
the
proposal
identified
"
three
mechanisms
that
 
could
be
used"
46
to
implement
the
exemptions,
the
first
and
third
of
which
("
an
applicability
cutoff
for
threshold
pollutants
 
under
 
section
112(
d)(
4)"
47
and
"
a
concentration­
based
applicability
threshold,"
48
respectively)
are
not
in
fact
two
mechanisms
but
one
 
a
fact
the
final
rule,
which
adopts
this
mechanism,
apparently
recognizes.
49
Even
more
concerning,
the
proposal
only
identified
HCl
as
a
possible
subject
of
a
risk­
based
exemption,
never
mentioning
an
exemption
for
Mn.
Finally,

although
the
proposed
rule
indicated
both
(
1)
that
EPA
had
"
chose[
n]
to
use
HCl
as
a
42
Id.
at
55,286,
Tables
2
and
3
to
Appendix
A
of
Subpart
DDDDD.
43
Id.
at
55,283
§
7.
44
Id.
at
55,283
§
4(
f).
45
Id.
at
55,283­
84
§
7(
c).
46
68
Fed.
Reg.
at
1688
(
emphasis
added).
47
Id.
48
Id.
49
Both
involve
exemptions
for
individual
facilities
that
"
can
demonstrate
that
their
emissions
of
threshold
pollutants
would
not
result
in
air
concentrations
above
the
threshold
levels."
Id.
at
1689.
See
also
id.
at
1692
(
defining
the
third
approach
as
"
an
applicability
cutoff
for
the
threshold
pollutant
hydrogen
chloride"
for
"
individual
facilities
that
can
demonstrate
that
their
emissions
of
hydrogen
chloride
will
not
result
in
air
concentrations
above
the
inhalation
reference
concentration
for
hydrogen
chloride").
15
surrogate
 
[
for]
inorganic
HAP,"
50
and
(
2)
that
the
Agency
was
considering
adopting
a
risk­
based
"
applicability
cutoff"
for
HCl,
51
the
proposal
never
acknowledged
the
deleterious
implications
of
the
latter
policy
choice
for
the
former.

With
respect
to
the
technical
details
of
the
risk­
based
exemption
mechanism,
the
proposed
rule
stated
only
that
EPA
was
considering
a
"[
t]
iered
analytical
approach."
52
The
entirety
of
the
Agency's
description
of
this
approach
was
as
follows:

Establishing
that
a
facility
meets
the
cutoffs
established
under
CAA
section
112(
d)(
4)
will
necessarily
involve
combining
estimates
of
pollutant
emissions
with
air
dispersion
modeling
to
predict
exposures.
The
EPA
envisions
that
we
would
promote
a
tiered
analytical
approach
for
these
determinations.
A
tiered
analysis
involves
making
successive
refinements
in
modeling
methodologies
and
input
data
to
derive
successively
less
conservative,
more
realistic
estimates
of
pollutant
concentrations
in
air
and
estimates
of
risk.

As
a
first
tier
of
analysis,
EPA
could
develop
a
series
of
simple
look­
up
tables
based
on
the
results
of
air
dispersion
modeling
conducted
using
conservative
input
assumptions.
By
specifying
a
limited
number
of
input
parameters,
such
as
stack
height,
distance
to
property
line,
and
emission
rate,
a
facility
could
use
these
look­
up
tables
to
determine
easily
whether
the
emissions
from
their
sources
might
cause
a
hazard
index
limit
to
be
exceeded.

A
facility
that
does
not
pass
this
initial
conservative
screening
analysis
could
implement
increasingly
more
site­
specific
but
more
resourceintensive
tiers
of
analysis
using
EPA­
approved
modeling
procedures,
in
an
attempt
to
demonstrate
that
exposure
to
emissions
from
the
facility
does
not
exceed
the
hazard
index
limit.
The
EPA's
guidance
could
provide
the
basis
for
conducting
such
a
tiered
analysis.
[
FN8]

FN8
"
A
Tiered
Modeling
Approach
for
Assessing
the
Risks
due
to
Sources
of
Hazardous
Air
Pollutants."
EPA­
450/
4­
92­
001.
David
E.
Guinnup,
Office
of
Air
Quality
Planning
and
Standards,
USEPA,
March
1992.

The
EPA
requests
comment
on
methods
for
constructing
and
implementing
a
tiered
analytical
approach
for
determining
applicability
of
50
Id.
at
1671.
51
Id.
at
1692.
52
Id.
at
1691.
16
the
CAA
section
112(
d)(
4)
criterion
to
specific
industrial
boiler
and
process
heater
sources.
It
is
also
possible
that
ambient
monitoring
data
could
be
used
to
supplement
or
supplant
the
tiered
modeling
approach
described
above.
It
is
envisioned
that
the
appropriate
monitoring
to
support
such
a
determination
could
be
extensive.
The
EPA
requests
comment
on
the
appropriate
use
of
monitoring
in
the
determinations
described
above.
53
Even
a
reader
new
to
"
tiered
modeling
approach[
es]"
can
easily
recognize
that
innumerable
details
are
missing
from
this
description.
For
example:
What
air
dispersion
models
will
be
acceptable?
Will
the
lookup
tables
be
published
as
part
of
the
final
rule
or
included
in
some
external
EPA
database?
How
will
EPA
arrive
at
the
emissions
levels
listed
in
those
tables?
What
are
"
conservative
input
assumptions"?
54
Will
facilities
that
resort
to
site­
specific
modeling
be
required
to
account
for
such
seemingly
relevant
sitespecific
factors
as
weather
(
e.
g.,
wind
speed,
average
precipitation
levels,
etc.)
and
altitude?
What
might
"
appropriate
monitoring"
55
include?

These
omissions
might
be
less
concerning
if
the
report
on
which
EPA
apparently
relied
in
crafting
the
exemption
proposal,
David
E.
Guinnup's
"
A
Tiered
Modeling
Approach
for
Assessing
the
Risks
due
to
Sources
of
Hazardous
Air
Pollutants,"
56
were
included
in
the
docket.
But
not
only
is
that
report
missing
from
the
docket,
it
appears
nowhere
in
the
final
rule.
If
EPA
in
fact
relied
on
the
report
in
crafting
either
the
proposal
or
the
final
rule,
the
CAA
requires
that
the
relevant
portions
of
the
report
be
included
in
the
docket.
57
If
the
final
version
of
the
risk­
based
exemptions
did
not
derive
53
68
Fed.
Reg.
at
1691.
54
Id.
55
Id.
56
Id.
57
See
42
U.
S.
C.
§
§
7607(
d)(
3)
("
All
data,
information,
and
documents
referred
to
in
this
paragraph
on
which
the
proposed
rule
relies
shall
be
included
in
the
docket
on
the
date
of
publication
of
the
proposed
rule."),
7607(
d)(
4)(
B)(
i)
("
All
documents
 
which
the
Administrator
determines
are
of
central
relevance
to
the
rulemaking
shall
be
placed
in
the
docket
as
soon
as
possible
after
their
availability.").
17
from
this
report,
that
fact
serves
only
to
underline
the
proposal's
paucity
of
details
about
the
possible
contours
of
those
exemptions.

Due
to
the
lack
of
detail
in
the
proposal,
most
of
our
objections
to
the
risk­
based
exemptions
arose
upon
publication
of
the
final
rule,
after
the
period
for
public
comment
had
closed.
Yet
those
objections,
detailed
below,
are
centrally
relevant
to
the
final
rule,

as
even
EPA's
"
preliminary
`
rough'
assessment
of
the
large
solid
fuel
subcategory"

indicates
that
over
800
large,
solid­
fuel
boilers
could
take
advantage
of
the
exemptions
to
evade
otherwise
applicable,
CAA­
mandated
MACT
standards.
58
Therefore,
under
CAA
§
307(
d)(
7)(
B),
EPA
must
reconsider
both
the
concept
and
the
practicalities
of
the
exemptions.
In
so
doing,
the
Agency
should
take
into
account
the
following
legal
and
analytical
objections,
each
of
which
calls
into
question
the
validity
of
those
exemptions
and
is
thus
"
of
central
relevance
to
the
outcome"
of
the
final
ICI
boiler
and
process
heater
rule.
59
b.
CAA
§
112(
d)(
4)
does
not
support
the
adoption
of
individualized,
risk­
based
exemptions.

EPA
calls
its
§
112(
d)(
4)­
based
MACT
exemptions
"
health­
based
compliance
alternatives,"
60
but
this
euphemism
cannot
disguise
the
exemptions'
true
nature:
They
are
a
clear
attempt
to
exempt
individual
HAP
sources
in
a
listed
category
from
CAAmandated
HAP
standards.
That
attempt
contravenes
the
Act
and
is
arbitrary
and
capricious.

58
69
Fed.
Reg.
at
55,244.
59
42
U.
S.
C.
§
7607(
d)(
7)(
B).
60
E.
g.
69
Fed.
Reg.
at
55,227.
18
i.
The
plain
language
of
the
Act
specifies
that
EPA
can
neither
refuse
to
set
emissions
standards
for
listed
sources
nor
exempt
individual
sources
from
otherwise­
applicable
standards.

As
discussed
above,
the
CAA's
MACT
requirements
are
unequivocal:
When
a
source
category
is
listed,
the
Agency
must
establish
emission
standards
that
apply
to
the
entire
category.
Section
112(
c)(
2)
of
the
Act,
"
Requirement
for
emissions
standards,"

states
that
"[
f]
or
the
categories
and
subcategories
the
Administrator
lists,
the
Administrator
shall
establish
emission
standards
under
subsection
(
d)
of
this
section."
61
Likewise,
§
112(
d)(
1)
instructs
the
Agency
to
"
promulgate
regulations
establishing
emission
standards
for
each
category
or
subcategory
of
major
sources
and
area
sources
of
hazardous
air
pollutants
listed
for
regulation
pursuant
to
subsection
(
c)
of
this
section."
62
Applying
this
plain
language,
National
Lime
squarely
holds
that
EPA
is
not
allowed
to
make
a
"
no
control"
determination
for
a
pollutant
emitted
by
a
listed
source
category.
63
Against
this
background,
the
import
of
§
112(
d)(
4)
is
clear.
The
provision
gives
EPA
authority
to
"
consider"
a
HAP's
established
health
threshold
in
setting
emission
standards
under
the
Act.
64
In
other
words,
if
there
is
a
known
"
threshold"
level
below
which
the
HAP
causes
no
adverse
health
effects,
EPA
may
set
a
category­
or
subcategory­
wide
control
standard
at
a
level
more
or
less
stringent
than
the
MACT
provisions
would
otherwise
require.

Section
112(
d)(
4)
does
not,
however,
allow
EPA
to
make
facility­
by­
facility
exemptions
from
otherwise­
applicable
MACT
standards.
The
section
provides,
"[
w]
ith
respect
to
pollutants
for
which
a
health
threshold
has
been
established,
the
Administrator
61
42
U.
S.
C.
§
7412(
c)(
2).
62
Id.
§
7412(
d)(
1).
63
National
Lime,
233
F.
3d
at
633­
34.
64
42
U.
S.
C.
§
7412(
d)(
4).
19
may
consider
such
threshold
level,
with
an
ample
margin
of
safety,
when
establishing
emissions
standards
under
this
subsection."
65
The
words
"
under
this
subsection"
require
EPA
to
apply
any
§
112(
d)(
4)­
influenced
standard
to
entire
categories
or
subcategories
of
sources,
because
subsection
(
d)
includes
no
mention
of
individualized
exemptions
or
individualized
standard­
setting
and
therefore
must
be
read
to
require
the
regulation
of
HAP
emissions
on
a
category­
wide
basis.
Moreover,
§
112(
d)(
4)
is
only
available
when
the
Agency
is
"
establishing"
 
not
applying
 
emissions
standards.
66
This
interpretation
of
the
statute's
plain
meaning
is
confirmed
by
the
fact
that
the
only
time
§
112
mentions
individualized
standard­
setting
is
under
the
so­
called
"
MACT
hammer"
provisions
of
§
112(
j).
In
the
event
that
EPA
misses
a
statutory
deadline
to
establish
emissions
standards
for
a
category
or
subcategory
of
sources,
§
112(
j)
directs
the
States
to
set
MACT
standards
for
each
facility
via
Title
V
permitting
programs.
67
The
language
of
§
112(
j),
which
directs
"
the
owner
or
operator
of
any
major
source"
to
submit
a
Title
V
permit
application
in
this
circumstance,
68
indicates
that
Congress
knew
how
to
regulate
on
a
source­
by­
source
basis;
had
legislators
wished
to
create
a
similar,

source­
by­
source
regulatory
regime
for
threshold
HAP
emissions
from
listed
sources,

therefore,
one
can
only
presume
they
would
have
included
similar
language
in
§
112(
d)(
4).
Instead,
§
112(
d)(
4)
clearly
mandates
that
EPA
promulgate
category­
wide,

MACT­
based
emissions
standards
for
all
sources
in
a
listed
category
or
sub­
category.
69
65
Id.
66
Id.
(
emphasis
added).
67
Id.
§
7412(
j)(
2)­(
5).
68
Id.
§
7412(
j)(
2)
(
emphasis
added).
69
Id.
§
7412(
d)(
4);
see
also
id.
§
7412(
c)(
2)
("
For
the
categories
and
subcategories
the
Administrator
lists,
the
Administrator
shall
establish
emissions
standards
under
subsection
(
d),
according
to
the
schedule
in
this
subsection
and
subsection
(
e).").
20
The
only
alternatives
to
this
requirement
are
(
1)
to
delist
an
entire
category
of
facilities
via
section
112(
c)(
9),
70
or
(
2)
to
delist
a
specific
chemical
under
subsection
(
b)(
3).
71
ii.
Congress
did
not
intend
to
permit
EPA
to
exempt
facilities
simply
because
they
emit
threshold
HAPs.

In
creating
individualized,
risk­
based
exemptions
for
facilities
that
emit
HCl
and
Mn,
EPA
invokes
an
authority
that
Congress
expressly
declined
to
grant
the
Agency.

Indeed,
as
discussed
below,
both
Houses
of
Congress
considered
source­
by­
source
riskbased
exemptions
for
threshold
HAPs,
but
both
chose
not
to
adopt
such
exemptions.

Emissions
standards
under
§
112
apply
to
"
major
sources,"
and
the
Act
specifies
that
a
source
is
"
major"
if
it
emits
in
excess
of
10
tons
per
year
of
a
given
HAP.
72
A
prior
version
of
the
bill,
however,
would
have
allowed
EPA
to
decline
to
regulate
sources
that
emit
greater
quantities
of
threshold
pollutants.
73
As
the
Senate
report
describing
the
provision
explained:

[
U]
nder
section
112(
c)(
5)
the
Administrator
may
set
a
lower
boundary
for
the
category
of
major
sources
which
is
higher
than
10
tons
per
year
(
but
which
provides
an
ample
margin
of
safety
to
protect
public
health).
If
there
are
no
sources
emitting
the
pollutant
in
that
amount
in
a
particular
category,
then
no
regulation
under
section
112(
d)
need
be
promulgated.
74
70
Id.
§
7412(
c)(
9).
71
Id.
§
7412(
b)(
3).
72
Id.
§
7412(
a)(
1).
73
See
S.
Rep.
No.
101­
228,
at
519
(
1989),
reprinted
in
V
A
LEGISLATIVE
HISTORY
OF
THE
CLEAN
AIR
ACT
AMENDMENTS
OF
1990,
at
8338,
8859
(
Comm.
Print
1993)
(
hereinafter
"
LEGISLATIVE
HISTORY")
(
noting
that
the
Senate's
proposed
version
of
§
112
would
have
permitted
EPA
to
"
establish
a
minimum
emissions
rate
of
more
than
ten
tons
for
a
category
or
subcategory
and
a
pollutant
for
which
a
health
effects
threshold
can
be
established,
provided
that,
the
minimum
emissions
rate
assures,
with
an
ample
margin
of
safety,
such
threshold
will
not
be
exceeded
within
the
vicinity
of
the
sources
in
the
category
and
that
no
additional
adverse
environmental
effects
will
occur
as
the
result
of
emissions
from
the
sources
individually
or
in
combination
with
emissions
from
other
similar
sources").
74
S.
Rep.
No.
101­
228,
at
176,
reprinted
in
V
LEGISLATIVE
HISTORY,
at
8516.
21
The
fact
that
such
a
provision
did
not
become
law
indicates
that
the
current
CAA
does
not
permit
EPA
to
avoid
regulating
facilities
by
pointing
to
evidence
of
a
health
threshold.

Moreover,
EPA's
current
interpretation
of
§
112
(
d)(
4)
directly
parallels
another
idea
advanced
and
rejected
during
the
debate
over
the1990
Amendments
to
the
CAA
(
CAAA).
The
House
bill
would
have
allowed
an
individual
facility
to
escape
MACT
if
a
risk
analysis
showed
that
the
source
posed
a
negligible
hazard.
Specifically,
the
bill
would
have
allowed
States
to
permit
"
a
major
source
to
comply
with
alternative
emission
limitations
in
lieu
of
standards
under
this
section,
if
the
owner
or
operator
presents
sufficient
evidence
to
demonstrate
that
emissions
from
the
source
in
compliance
with
such
limitations
present
a
negligible
risk
to
public
health
under
criteria
issued
by
the
Administrator."
75
This
proposal
was
not
in
the
Senate
bill,
however,
and
again
the
conference
rejected
the
idea
in
the
final
legislation.
76
Finally,
the
conference
report
on
the
1990
CAAA
states,
"[
i]
t
is
the
conferees'

intent
that
EPA
not
use
the
[
§
112(
j)]
permit
hammer
approach
(
case­
by­
case)
to
avoid
or
delay
meeting
MACT
requirements."
77
In
other
words,
Congress
expressed
its
desire
for
a
category­
based
approach
to
MACT
standard­
setting
rather
than
a
time­
consuming
and
delay­
inducing
source­
by­
source
approach.

c.
EPA's
authority
under
§
112(
d)(
4)
does
not
extend
to
HCl
and
Mn.

75
H.
R.
3030,
101st
Cong.
§
112(
g)(
1)(
A),
reprinted
in
II
LEGISLATIVE
HISTORY,
at
3939.
76
See
Senate
Debate
on
the
Clean
Air
Act
Amendments
of
1990
Conference
Report
(
hereinafter
"
Senate
Debate"),
reprinted
in
I
LEGISLATIVE
HISTORY,
at
866
(
Statement
of
Senator
Durenberger)
("
The
authority
for
such
exemptions
was
not
present
in
the
Senate
bill,
and
the
House
receded
to
the
Senate
on
this
point.
The
provision
was
deleted
in
conference.").
77
Clean
Air
Act
Amendments
of
1990
Conference
Report,
at
340
(
hereinafter
"
Conference
Report"),
reprinted
in
I
LEGISLATIVE
HISTORY,
at
1790.
22
i.
Section
112(
d)(
4)
applies
only
to
HAPs
for
which
a
health
threshold
has
been
established.

Whatever
the
legality
of
invoking
CAA
§
112(
d)(
4)
to
support
adoption
of
individualized
risk­
based
exemptions
from
otherwise
applicable
MACT
standards,
the
plain
language
of
that
section
limits
its
applicability
to
HAPs
for
which
a
"
health
threshold
has
been
established."
78
Three
aspects
of
this
phrase
are
noteworthy.
First,
the
word
"
threshold"
requires
that
EPA
have
direct
evidence
that
the
HAP
in
question
has
no
health
effects
below
an
identified,
non­
zero
level.
79
This
requirement,
in
turn,
precludes
the
extension
of
§
112(
d)(
4)
authority
to
carcinogens,
which
are
presumed
to
have
no
such
level.
Indeed,
the
Senate
CAAA
report
made
precisely
this
point,
noting
that
invoking
§
112(
d)(
4)
is
appropriate
only
when
"
the
pollutant
presents
no
risk
of
other
adverse
health
effects,
including
cancer,
for
which
no
threshold
can
be
established.
.
.
."
80
Recognizing
this
Congressional
intent,
EPA
traditionally
has
interpreted
§
112(
d)(
4)
to
exclude
consideration
of
carcinogens.
81
Second,
the
phrase
"
has
been"
demonstrates
that
Congress
intended
for
EPA
to
use
its
§
112(
d)(
4)
authority
only
when
a
pollutant's
health
threshold
is
already
accepted
in
the
scientific
community.
That
is,
Congress
did
not
intend
for
EPA
to
spend
time
and
resources
seeking
such
a
threshold.
This
reading
is
bolstered
by
the
fact
that
one
draft
of
78
42
U.
S.
C.
§
7412(
d)(
4).
79
The
Senate
report
on
the
CAAA
states
that
the
section
allows
EPA
to
use
the
"
no
observable
effects
level"
(
NOEL).
S.
Rep.
No.
101­
228,
at
171,
reprinted
in
V
LEGISLATIVE
HISTORY,
at
8511.
80
S.
Rep.
No.
101­
228,
at
171,
reprinted
in
V
LEGISLATIVE
HISTORY,
at
8511
(
emphasis
added);
see
also
Senate
Debate,
reprinted
in
I
LEGISLATIVE
HISTORY,
at
876
(
Statement
of
Senator
Durenberger)
("
With
respect
to
pollutants
for
which
a
safe
threshold
can
be
set,
the
authority
to
set
a
standard
less
stringent
than
the
maximum
achievable
control
technology
is
contained
in
subsection
(
d)(
4).
With
respect
to
carcinogens
and
other
non­
threshold
pollutants,
no
such
authority
exists
in
subsection
(
d)
or
any
other
provision
of
the
Act."
(
emphasis
added)).
81
See,
e.
g.,
National
Emission
Standard
for
Hazardous
Air
Pollutants,
Proposed
Standards
for
Hazardous
Air
Pollutants
From
Chemical
Recovery
Combustion
Sources
at
Kraft,
Soda,
Sulfite,
and
Stand­
alone
Semichemical
Pulp
Mills,
63
Fed.
Reg.
18,754,
18,
765
(
proposed
Apr.
15,
1998)
(
to
be
codified
at
40
C.
F.
R.
pt.
63)
(
noting
that
EPA
has
not
applied
section
112(
d)(
4)
to
carcinogens
because
Congress
"
clearly
intended
them
to
be
nonthreshold
pollutants").
23
the
CAAA
 
later
rejected
 
would
have
made
the
§
112(
d)(
4)
authority
contingent
only
on
a
finding
that
a
threshold
"
can
be
established."
82
Third,
Congress's
use
of
the
word
"
established"
 
rather
than,
for
example,

"
estimated"
or
"
approximated"
 
indicates
that
EPA
must
have
a
high
degree
of
scientific
certainty
before
using
its
§
112(
d)(
4)
authority.
83
Again,
the
CAAA
legislative
history
supports
this
interpretation;
even
the
Senate,
whose
bill
would
have
allowed
EPA
to
invoke
section
(
d)(
4)
when
a
threshold
"
can
be
established,"
did
not
intend
for
the
Agency
to
take
less­
than­
definitive
evidence
of
a
threshold
into
account
in
setting
emissions
standards:

[
T]
hat
would
jeopardize
the
standard­
setting
schedule
imposed
under
this
section
with
the
kind
of
lengthy
study
and
debate
that
has
crippled
the
current
program.
But
where
health
thresholds
are
well­
established,
 
and
the
pollutant
presents
no
risk
of
other
adverse
health
effects,
including
cancer,
for
which
no
threshold
can
be
established,
the
Administrator
may
use
the
threshold
with
an
ample
margin
of
safety
(
and
not
considering
cost)
to
set
emissions
limitations
for
sources
in
the
category
or
subcategory.
84
Overall,
then,
EPA
may
not
rely
on
§
112(
d)(
4)
to
support
risk­
based
exemptions
for
known
or
potential
carcinogens.
It
may
not
invoke
the
section
unless
evidence
of
a
no­
effects
threshold
for
a
pollutant
already
exists
in
the
scientific
literature.
And
it
may
not
speculate
from
inconclusive
evidence;
it
must
be
certain
that
exposure
to
the
pollutant
at
the
identified
"
threshold"
level
has
no
deleterious
health
effects.

ii.
No
health
threshold
has
been
established
for
HCl.

82
S.
1630,
101st
Cong.
§
112(
d)(
4)(
A),
reprinted
in
III
LEGISLATIVE
HISTORY,
at
4425
(
emphasis
added).
83
The
most
relevant
dictionary
definition
of
"
established"
is
"
to
put
beyond
doubt;
prove."
WEBSTER'S
NEW
COLLEGIATE
DICTIONARY
388
(
1980).
84
S.
Rep.
No.
101­
228,
at
171,
reprinted
in
V
LEGISLATIVE
HISTORY,
at
8511
(
emphasis
added).
24
These
three
considerations
squarely
preclude
establishment
of
a
§
112(
d)(
4)­
based
exemption
for
HCl.
To
begin
with,
it
is
possible
that
HCl
is
carcinogenic.
In
its
Integrated
Risk
Information
System
(
IRIS)
database,
EPA
states
that
HCl
"
has
not
undergone
a
complete
evaluation
and
determination
 
for
evidence
of
human
carcinogenic
potential."
85
Elsewhere,
the
Agency
states
that
"
no
information"
exists
on
HCl
carcinogenicity
in
humans.
86
Further,
the
International
Agency
for
Research
on
Cancer
(
IARC)
identifies
several
human
and
animal
studies
that
suggest
HCl
may
have
carcinogenic
effects:

One
US
study
of
steel­
pickling
workers
showed
an
excess
risk
for
cancer
of
the
lung
in
workers
exposed
primarily
to
hydrochloric
acid.
An
increased
risk
for
laryngeal
cancer
was
observed
in
the
same
cohort;
however,
no
analysis
was
performed
of
workers
exposed
to
hydrochloric
acid.
None
of
three
US
industry­
based
case­
control
studies
suggested
an
association
between
exposure
to
hydrogen
chloride
and
cancers
of
the
lung,
brain
or
kidney.
In
one
Canadian
population­
based
case­
control
study,
an
increased
risk
for
oat­
cell
carcinoma
was
suggested
in
workers
exposed
to
hydrochloric
acid;
however,
no
excess
risk
was
observed
for
other
histological
types
of
lung
cancer.
 

In
one
lifetime
study
in
male
rats
exposed
by
inhalation
at
one
dose
level,
hydrogen
chloride
did
not
produce
a
treatment­
related
increase
in
the
incidence
of
tumours.
Hydrogen
chloride
was
tested
at
one
dose
level
in
combination
with
formaldehyde
by
inhalation
exposure
in
the
same
longterm
experiment
in
male
rats.
Hydrogen
chloride
did
not
influence
the
nasal
carcinogenicity
of
formaldehyde
when
mixed
with
it
upon
entry
into
the
inhalation
chamber.
When
the
two
compounds
were
premixed
before
entry
into
the
inhalation
chamber,
an
increased
incidence
of
nasal
tumours
was
observed
over
that
seen
in
animals
treated
with
the
combination
mixed
on
entry
or
with
formaldehyde
alone.
87
85
U.
S.
EPA,
Integrated
Risk
Information
System:
Hydrogen
Chloride,
http://
www.
epa.
gov/
iris/
subst/
0396.
htm.
86
U.
S.
EPA,
Technology
Transfer
Network
Air
Toxics
Website:
Hydrochloric
Acid
(
Hydrogen
Chloride),
http://
www.
epa.
gov/
ttnatw01/
hlthef/
hydrochl.
html.
87
International
Agency
for
Research
on
Cancer,
Hydrochloric
Acid,
Summary
of
Data
Reported
and
Evaluation,
http://
www­
cie.
iarc.
fr/
htdocs/
monographs/
vol54/
03­
hydrochloric­
acid.
htm.
25
From
these
studies,
IARC
concludes
only
that
"[
t]
here
is
inadequate
evidence
for
the
carcinogenicity
in
humans
of
hydrochloric
acid,"
88
but
these
studies
and
the
absence
of
further
data
are
patently
insufficient
for
EPA
simply
to
assume
HCl
is
not
carcinogenic.

A
further
flaw
in
the
Agency's
decision
to
treat
HCl
as
a
threshold
pollutant
concerns
its
conclusion
that
the
HCl
reference
concentration
(
RfC,
or
level
at
which
no
adverse
health
effects
are
expected89)
is
0.02
mg/
m3
(
0.0134
ppm90).
91
EPA
selects
this
RfC
value
based
on
rat
studies
that
demonstrated
hyperplasia
of
the
nasal
mucosa
(
the
protective
cell
lining
of
the
nasal
tract
and
cavities),
larynx,
and
trachea
in
response
to
HCl
exposure.
92
But
these
rat
studies
only
investigated
effects
on
respiratory
tract
organs;
the
Agency
identifies
no
studies
that
indicate
whether
exposure
to
HCl
 
at
0.02
mg/
m3
or
any
other
concentration
 
harms
other
bodily
systems.
In
addition,
the
studies
failed
to
identify
a
no­
observed­
effects­
level
(
NOEL).
As
a
result,
the
RfC
in
the
final
ICI
boiler
and
process
heater
rule
is
based
on
the
lowest­
observed­
adverse­
effects­
level
(
LOAEL)
 
that
is,
the
lowest
dose
to
which
study
rats
were
exposed,
and
at
which
measurable
adverse
health
effects
occurred
in
treated
animals.
93
88
Id.
(
emphasis
in
original).
89
Specifically,
the
RfC
is
an
estimate
(
with
uncertainty
spanning
perhaps
an
order
of
magnitude)
of
a
continuous,
lifetime
inhalation
exposure
that
is
unlikely
to
pose
significant
risks
of
deleterious
noncancer
effects
in
human
subjects
(
including
sensitive
subgroups).
90
To
convert
concentrations
in
air
(
at
25
oC)
from
ppm
to
mg/
m3:
mg/
m3
=
(
ppm)
x
(
molecular
weight
of
the
compound)/
24.45.
For
hydrochloric
acid:
1ppm
=
1.49
mg/
m3.
91
69
Fed.
Reg.
at
55,283,
Appendix
A
§
6(
a)(
1)
(
referring
facilities
to
EPA's
Technology
Transfer
Network
Air
Toxic
Website
for
an
RfC
for
HCl).
92
R.
E.
Albert,
et
al.,
Gaseous
formaldehyde
and
hydrogen
chloride
induction
of
nasal
cancer
in
rats.
68
J.
Natl.
Cancer
Inst.
597­
603
(
1982).
93
The
Albert,
et
al.
study
(
discussed
in
detail
by
A.
R.
Sellakumar,
et
al.,
Carcinogenicity
of
formaldehyde
and
hydrogen
chloride
in
rats.
81
Toxicol.
Appl.
Pharmacol.
401­
406
(
1985))
reported
data
from
a
chronic
inhalation
exposure
study
in
rats.
One
hundred
male
Sprague­
Dawley
rats
were
exposed
to
10
ppm
hydrogen
chloride
for
6
hours/
day,
5
days/
week
(
duration­
adjusted
concentration
=
2.5
mg/
m3)
for
their
lifetimes.
Only
the
organs
of
the
respiratory
tract
were
subjected
to
detailed
histological
analysis.
Even
at
this
lowest
dose,
there
was
a
24%
incidence
of
hyperplasia
of
laryngeal­
tracheal
segments
in
HCl­
exposed
rats
(
larynx
2/
22,
trachea
6/
26)
versus
6%
in
the
controls.
26
It
is
highly
likely
that
at
doses
below
those
in
the
study,
there
are
alterations
in
normal
physiological
and
biological
functions
in
exposed
animals,
and
consequent
adverse
health
effects.
Further,
it
is
reasonable
to
presume
that
organs
other
than
those
associated
with
the
nasal
and
respiratory
tracts
are
vulnerable
to
damage
or
destruction
through
acute
or
long­
term
HCl
inhalation.
After
all,
clinical
symptoms
of
HCl
toxicity
in
humans
include
gastritis,
dermatitis,
and
adverse
reproductive
outcomes.

iii.
No
health
threshold
has
been
established
for
Mn.

EPA's
decision
to
treat
Mn
as
a
threshold
pollutant
is
equally
ill­
founded.
94
For
one
thing,
Mn,
like
HCl,
may
have
carcinogenic
effects.
On
this
subject,
EPA's
IRIS
database
indicates
that
"[
e]
xisting
studies
are
inadequate
to
assess
the
carcinogenicity
of
manganese,"
95
but
the
database
also
identifies
two
mouse
studies
in
which
exposure
to
Mn
compounds
induced
lung96
and
lymph97
tumors,
respectively.
Moreover,
EPA
concedes
that
the
results
of
the
former
study
"
are
suggestive
of
carcinogenicity."
98
Studies
by
the
National
Toxicology
Program
at
the
National
Institute
of
Environmental
Health
Sciences
(
NTP)
provide
further
evidence
of
Mn's
potential
carcinogenicity.
NTP
performed
toxicology
and
carcinogenesis
studies
on
rats
and
mice,
finding
that
adding
manganese
(
II)
sulfate
monohydrate
to
the
animals'
feed
marginally
increased
incidences
of
thyroid
gland
follicular
cell
adenoma
and
significantly
increased
incidences
of
94
The
flaws
in
the
Mn
exemption
are
particularly
worthy
of
reconsideration,
as
EPA's
proposed
ICI
boiler
and
process
heater
rule
provided
absolutely
no
notice
that
the
Agency
was
considering
Mn
as
a
possible
subject
of
a
risk­
based
exemption
under
§
112(
d)(
4).
95
U.
S.
EPA,
Integrated
Risk
Information
System:
Manganese,
http://
www.
epa.
gov/
iris/
subst/
0373.
htm.
96
G.
D.
Stoner,
et
al.,
Test
for
carcinogenicity
of
metallic
compounds
by
the
pulmonary
tumor
response
in
strain
A
mice.
36
Cancer
Res.
1744­
1747
(
1976).
97
J.
A.
DiPaolo,
The
potentiation
of
lymphosarcomas
in
mice
by
manganous
chloride.
23
Fed.
Proc.
393
(
1964)
(
abstract).
98
U.
S.
EPA,
Integrated
Risk
Information
System:
Manganese,
http://
www.
epa.
gov/
iris/
subst/
0373.
htm.
27
follicular
cell
hyperplasia.
99
Clearly,
then,
it
is
arbitrary
 
and
hence
unlawful
 
for
the
Agency
to
assume
that
Mn
is
noncarcinogenic
 
as
it
must
to
treat
Mn
as
a
threshold
pollutant
to
which
CAA
§
112(
d)(
4)
may
be
applicable.

In
addition,
as
for
HCl,
the
Agency's
studies
of
the
inhalation
effects
of
Mn
(
on
which
the
Mn
RfC
is
based)
identify
only
a
LOAEL,
not
a
NOEL.
100
Again,
this
strongly
suggests
that
physiological
and
biological
impairments
occur
at
Mn
doses
below
those
tested
in
the
studies.
Moreover,
according
to
EPA's
database,
the
two
studies
in
question
involved
men
exposed
to
Mn
compounds
in
their
workplace.
101
That
is,
the
Agency
only
considered
presumptively
healthy,
adult
men's
exposure
to
Mn
compounds
 
an
egregious
flaw
given
that
Mn
is
a
well­
established,
potent
developmental
neurotoxicant,

known
to
be
most
damaging
to
the
central
nervous
systems
of
fetuses,
infants,
and
young
children.
102
Finally,
the
rule
indicates
that
in
treating
Mn
as
a
threshold
pollutant,
the
Agency
considered
only
the
health
effects
of
chronic
exposure,
ignoring
"
potential
acute
effects"

because
"[
a]
screening
assessment
 
for
[
those]
effects"
showed
"
no
exceedances."
103
99
National
Toxicology
Program,
National
Institute
of
Environmental
Health
Sciences,
Study
Abstract
for
Manganese
Sulfate
Monohydrate,
http://
ntp­
apps.
niehs.
nih.
gov/
ntp_
tox/
index.
cfm?
fuseaction
=
abstracts.
abstract&
nextcircuit=
longtermbioassaydata&
study_
no=
C61143B&
abstract_
url=
http%
3A%
2F%
2Fntp%
2Dserver%
2Eniehs%
2Enih%
2Egov%
2Fhtdocs%
2FLT%
2DStudies%
2FTR428%
2Ehtml&
test_
type
=
Long­
Term&
study_
length=
2%
20Years&
cas_
no=
10034%
2D96%
2D5&
chemical_
name=
Manganese%
20Sulfate%
20Monohydrate.
100
U.
S.
EPA,
Integrated
Risk
Information
System:
Manganese,
http://
www.
epa.
gov/
iris/
subst/
0373.
htm.
101
Id.
(
describing
H.
Roels,
et
al.,
Epidemiological
survey
among
workers
exposed
to
manganese:
Effects
on
lung,
central
nervous
system,
and
some
biological
indices.
11
Am.
J.
Ind.
Med.
307­
327
(
1987),
and
H.
Roels,
et
al.,
Assessment
of
the
permissible
exposure
level
to
manganese
in
workers
exposed
to
manganese
dioxide
dust.
49
Br.
J.
Ind.
Med.
25­
34
(
1992).
102
See,
e.
g.,
L.
Takser
L.,
et
al.,
Manganese,
monoamine
metabolite
levels
at
birth,
and
child
psychomotor
development.
24
Neurotoxicology
667­
74
(
2003)
(
describing
a
prospective
epidemiological
study
in
247
healthy
pregnant
women
and
their
babies
to
determine
the
long­
term
effect
of
in
utero
exposure
to
environmental
low­
levels
of
Mn,
and
reporting
evidence
of
impaired
psychomotor
development
by
age
six).
See
also
L.
Normandin,
et
al.,
Manganese
distribution
in
the
brain
and
neurobehavioral
changes
following
inhalation
exposure
of
rats
to
three
chemical
forms
of
manganese.
25
Neurotoxicology
433­
41
(
2004).
103
69
Fed.
Reg.
at
55,220.
28
Amazingly,
however,
the
cited
screening
assessment
appears
nowhere
in
the
docket
for
the
final
rule.
Moreover,
a
memorandum
describing
that
screening
assessment,
which
does
appear
in
the
docket,
104
indicates
that
the
assessment
relied
on
"[
a]
cute
 
toxicity
dose­
response
values
 
taken
from
the
EPA
Air
Toxics
website
(
http://
www.
epa.
gov/
ttn/

atw/
toxsource/
summary.
html)"
105
 
even
though
that
website
identifies
no
acute
toxicity
dose­
response
value
for
Mn.
106
That
is,
EPA
appears
arbitrarily
to
have
concluded
that
Mn
emissions
from
regulated
facilities
never
exceed
the
level
at
which
acute
effects
may
occur
even
though
the
Agency
has
no
data
indicating
what
that
level
might
be.
107
d.
The
final
rule
makes
no
attempt
to
account
for
background
levels
and
alternative
sources
of
HCl
and
Mn.

A
further
problem
with
the
ICI
boiler
and
process
heater
rule's
risk­
based
exemptions
is
the
Agency's
failure
to
account
for
either
background
levels
or
alternative
sources
of
HCl
and
Mn.
The
final
rule
permits
an
individual
facility
to
demonstrate
that
it
qualifies
for
the
risk­
based
exemption
for
HCl
or
Mn
simply
by
submitting
data
showing
that
emissions
from
the
facility
do
not
exceed
the
RfC
for
that
HAP,
whether
or
not
there
are
other,
co­
located
sources
of
the
HAP
or
other
pathways
of
exposure.
108
As
a
consequence,
a
facility
with
both
an
ICI
boiler
or
process
heater
and
other
HCl
or
Mn­

104
Memorandum
from
Scott
Jenkins,
Risk
and
Exposure
Assessment
Group,
to
Dave
Guinnup,
Leader,
Risk
and
Exposure
Assessment
Group,
"
Screening
Assessment
of
Central
Nervous
System
Hazardous
Air
Pollutants
from
Wood­
Fired
Industrial
Boilers"
(
February
25,
2004),
OAR­
2002­
0058­
0608.
105
Id.
at
1.
106
U.
S.
EPA,
Acute
Dose­
Response
Values,
http://
www.
epa.
gov/
ttn/
atw/
toxsource/
table2.
pdf
(
listing
acute
exposure
guideline
levels
(
AEGLs)
for
various
air
toxics,
but
leaving
blank
the
AEGLs
corresponding
to
"
Manganese
compounds").
107
The
table
does
include
one
value
for
Mn
 
the
"
IDLH"
level,
or
level
at
which
the
compound
is
immediate
dangerous
to
life
or
health.
Relying
on
this
value
to
determine
when
and
whether
emissions
from
regulated
facilities
might
cause
acute
health
effects
is
completely
inappropriate,
as
the
IDLH
level
indicates
only
the
concentration
at
which
Mn
would
be
deadly
or
irreversibly
dangerous
to
health,
not
the
concentration
at
which
lesser
(
but
still
measurable
and
adverse)
acute
effects
would
be
observed.
108
69
Fed.
Reg.
at
55,283,
55,285,
Appendix
A
§
§
7(
a)(
b),
13.
29
emitting
units,
or
a
facility
located
in
an
industrial
park
near
other
similar
facilities,
or
a
facility
located
in
an
area
with
high
background
concentrations
of
either
HAP,
could
qualify
for
the
HCl
or
Mn
risk­
based
exemption
yet
still
pose
a
significant
risk
to
people
living
or
working
nearby.

This
dubious
approach
differs
from
that
advanced
in
the
proposed
rule.
109
In
the
proposal,
EPA
made
clear
that
any
credible
assessment
of
the
risks
posed
by
a
regulated
facility
must
include
background
pollution
and
emissions
from
co­
located
sources.
The
proposal's
discussion
of
this
issue
proceeded
in
three
steps.
First,
the
Agency
explained
that
it
would
use
the
concept
of
a
hazard
index
(
HI)
or
hazard
quotient
(
HQ)
to
quantify
the
risks
from
regulated
plants.
110
(
As
defined
in
the
final
rule,
an
HI
is
the
sum
of
individual
HQs
"
for
multiple
substances
and/
or
multiple
exposure
pathways,"
while
an
HQ
is
"
the
ratio
of
the
predicted
media
concentration
of
a
pollutant
to
the
media
concentration
at
which
no
adverse
effects
are
expected."
111)
Second,
EPA
sought
comment
on
a
range
of
HI
(
or
HQ)
values
that,
if
exceeded
at
any
particular
plant,
would
indicate
unacceptably
high
emissions
of
one
or
more
HAPs.
Third,
the
Agency
observed
that
simply
ensuring
that
HAP
emissions
from
an
individual
facility
result
in
an
HI
less
than
or
equal
to
1.0
 
that
is,
ensuring
that
emissions
from
the
facility
do
not
exceed
the
threshold
concentration
for
the
relevant
pollutant
but
ignoring
other
sources
of
exposure
 
inevitably
underprotects.
Specifically,
the
Agency
stated:

109
Because
EPA's
final
approach
differs
from
that
championed
in
the
notice
of
proposed
rulemaking,
the
grounds
for
our
objection
arose
after
the
period
for
public
comment
had
ended.
The
objection
is
thus
appropriately
raised
in
this
petition.
See
42
U.
S.
C.
§
7607(
d)(
7)(
B).
Moreover,
the
objection
is
"
of
central
relevance
to
the
outcome
of
the
rule,"
id.,
because
it
demonstrates
that
the
exemptions
contravene
the
CAA
and
are
arbitrary
and
capricious.
110
68
Fed.
Reg.
at
1689­
1691.
111
69
Fed.
Reg.
at
55,285,
Appendix
A
§
13.
For
inhalation
exposures,
the
rule
goes
on
to
note,
"
the
HQ
is
calculated
as
the
air
concentration
divided
by
the
Rfc."
Id.
30
One
option
is
to
allow
the
hazard
index
posed
by
all
threshold
HAP
emitted
from
sources
at
the
facility
to
be
no
greater
than
one.
This
approach
is
protective
if
no
additional
threshold
HAP
exposures
would
be
anticipated
from
other
sources
in
the
vicinity
of
the
facility
or
through
other
routes
of
exposure
(
e.
g.,
through
ingestion).
 
Because
noncancer
risk
assessment
is
predicated
on
total
exposure
or
dose,
and
because
risk
assessments
focus
only
on
an
individual
source,
establishing
a
hazardous
index
limit
of
0.2
would
account
for
an
assumption
that
20
percent
of
an
individual's
total
exposure
is
from
that
individual
source.
 
If
the
facility
is
allowed
to
emit
HAP
such
that
its
own
impacts
could
result
in
HI
values
of
one,
total
exposures
to
threshold
HAP
in
the
vicinity
of
the
facility
could
be
substantially
greater
than
one
due
to
background
sources,
and
this
would
not
be
protective
of
public
health,
since
only
HI
values
below
one
are
considered
to
be
without
appreciable
risk
of
adverse
health
effects.
112
Thus,
at
the
proposal
stage,
EPA
apparently
concluded
that
an
HI
limit
of
1.0
could
only
protect
public
health
"
with
an
ample
margin
of
safety"
(
a
protective
cushion
required
by
§
112(
d)(
4)
113)
in
the
absence
of
other
sources
of
the
HAP
in
question.

This
conclusion
is
unsurprising.
Indeed,
any
other
conclusion
would
be
untenable.
The
lungs
do
not
distinguish
between
pollution
from
boilers
and
similar
pollution
from
other
plants
or
mobile
sources.
If
the
ambient
concentration
of
a
particular
pollutant
is
at
or
near
the
safe
level,
therefore,
an
additional
source
of
the
pollutant
can
push
the
exposure
over
the
threshold
even
if
the
additional
source
emits
the
pollutant
at
low
levels.
And
similarly,
if
the
source
is
emitting
the
pollutant
at
levels
close
to
the
RfC,
even
low
background
pollution
could
lead
to
unsafe
overall
conditions.

To
determine
whether
an
HI
(
or
HQ)
of
1.0
for
HCl
or
Mn
emissions
from
an
ICI
boiler
or
process
heater
is
insufficiently
protective,
therefore,
one
need
only
ask
whether
"
additional
[
HCl
or
Mn]
exposures
would
be
anticipated
from
other
sources
in
the
vicinity
of
the
facility
or
through
other
routes
of
exposure."
114
The
answer
to
both
questions
is
clearly
yes.
With
respect
to
HCl,
EPA
elsewhere
identified
a
study
 
112
68
Fed.
Reg.
at
1691
(
emphasis
added).
113
42
U.
S.
C.
§
7412(
d)(
4).
114
68
Fed.
Reg.
at
1691.
31
performed
by
a
different
regulated
industry,
lime
manufacturing
plants
 
that
demonstrated
that
"
the
mean
national
HCl
concentration
correspond[
s]
to
an
HQ
of
0.06
and
the
95th
percentile
national
HCl
concentration
correspond[
s]
to
an
HQ
of
0.2."
115
With
respect
to
Mn,
the
National
Library
of
Medicine's
Toxmap116
indicates
that
in
2002,

(
1)
nationwide,
over
1,000
facilities
reported
on­
site
Mn
releases,
and
(
2)
within
a
50­

mile
radius
of
most
major
urban
centers,
there
were
multiple
(
and
often
many)
such
releases.
117
Despite
these
alternative
sources
of
HCl
and
Mn,
however,
EPA's
final
ICI
boiler
and
process
heater
rule
settles
on
an
HI
(
or
HQ)
limit
of
1.0,118
thus
failing
to
ensure
 
on
the
Agency's
own
terms
 
that
sources
are
truly
"
without
appreciable
risk
of
adverse
health
effects."
119
The
rule
therefore
violates
the
CAA
and
is
arbitrary
and
capricious.

e.
EPA
may
not
simultaneously
adopt
a
risk­
based
exemption
for
HCl
emissions
and
use
HCl
as
a
proxy
for
other
unidentified
HAPs.

A
further
flaw
in
the
risk­
based
exemptions
concerns
the
Agency's
simultaneous
creation
of
an
exemption
for
HCl
emissions
and
adoption
of
HCl
as
a
surrogate
for
other
non­
metallic
inorganic
HAPs.
The
obvious
and
entirely
unlawful
consequence
of
this
maneuvering
is
that
large,
solid­
fuel
facilities
whose
HCl
emissions
fall
below
the
Agency's
identified
risk­
based
standard
are
permitted
to
evade
the
MACT
standards
for
115
Proposed
Rules,
National
Emission
Standard
for
Hazardous
Air
Pollutants
for
Lime
Manufacturing
Plants,
67
Fed.
Reg.
78,046,
78,056
(
proposed
Dec.
20,
2002)
(
to
be
codified
at
40
C.
F.
R.
pt.
63).
116
National
Library
of
Medicine,
National
Institutes
of
Health,
Toxmap
 
Environmental
Health
E­
Maps,
http://
toxmap.
nlm.
nih.
gov/
toxmap/
main/
index.
jsp.
117
For
example,
Toxmap
indicates
that
there
were
over
30
such
releases
within
a
50­
mile
radius
of
Chicago,
about
15
within
a
similar
distance
of
Dallas,
over
10
for
Los
Angeles,
about
40
for
Milwaukee,
and
almost
20
for
Philadelphia.
Id.,
Toxmap
 
Environmental
Health
E­
Maps,
MANGANESE
(
7439­
96­
5),
http://
toxmap.
nlm.
nih.
gov/
toxmap/
releases/
searchChemical.
do.
118
69
Fed.
Reg.
55,283,
Appendix
A
§
7(
a),
(
b).
119
68
Fed.
Reg.
at
1691.
32
HCl
and
in
turn
to
avoid
adopting
controls
for
other
non­
metallic
inorganic
HAP.

Adding
insult
to
injury,
the
Agency
fails
even
to
identify
which
non­
metallic
inorganic
HAPs
may
elude
regulation
in
this
way,
let
alone
to
explain
its
apparent
certainty
that
all
such
HAPs
are
"
pollutants
for
which
a
health
threshold
has
been
established,"
120
such
that
they
may
be
eligible
for
consideration
under
§
112(
d)(
4).
Instead,
EPA
says
only
that
the
"
available
 
emissions
test
information
 
indicate[]
that
the
primary
inorganic
HAP
emitted
from
boilers
and
process
heaters
is
HCl,"
121
that
"[
m]
uch
smaller
amounts
of
hydrogen
fluoride
and
chlorine
are
emitted,"
122
and
that
the
Agency
"
do[
es]
not
expect
hydrogen
cyanide
emissions
from
boilers
covered
under
the
final
rule."
123
Even
if
one
accepts
these
unsubstantiated
and
self­
serving
assertions
as
true,
their
direct
and
necessary
implication
is
that
facilities
that
qualify
for
the
HCl
exemption
need
not
control
their
hydrogen
fluoride
(
HF)
emissions
either.
As
discussed
above,
however,

whatever
the
general
lawfulness
of
§
112(
d)(
4)­
based
exemptions,
EPA
may
not
adopt
such
exemptions
for
pollutants
whose
health
thresholds
have
not
been
established.
124
Yet
HF
falls
squarely
in
this
category.
Although
EPA
does
not
consider
the
chemical
in
its
IRIS
database,
125
the
database
does
include
information
about
fluorine,
to
which
HF
rapidly
breaks
down.
According
to
that
information,
no
data
are
available
to
determine
an
RfC
for
chronic
inhalation
exposure
to
fluorine,
nor
has
fluorine
"
undergone
a
complete
evaluation
and
determination
 
for
evidence
of
human
carcinogenic
potential."
126
Moreover,
EPA
elsewhere
acknowledges
data
suggesting
that
those
with
120
42
U.
S.
C.
§
7412(
d)(
4).
121
69
Fed.
Reg.
at
55,230.
122
Id.
123
Id.
at
55,244.
124
See
supra
at
pp.
20­
21.
125
See
U.
S.
EPA,
Integrated
Risk
Information
System,
http://
www.
epa.
gov/
iris.
126
U.
S.
EPA,
Integrated
Risk
Information
System:
Fluorine,
http://
www.
epa.
gov/
iris/
subst/
0053.
htm.
33
occupational
exposure
to
HF
have
greater
than
normal
occurrences
of
cancer.
127
Because
EPA
has
not
established
that
HF
is
noncarcinogenic,
it
may
not
consider
HF
a
threshold
pollutant.
Further,
even
if
an
RfC
for
HF
were
established,
EPA
offers
no
justification
for
its
apparent
assumption
that
plants
whose
emissions
meet
the
RfC
for
HCl
will
necessarily
also
meet
the
RfC
for
HF.
For
this
reason
alone,
EPA's
rule
contravenes
the
CAA
and
is
arbitrary
and
capricious.

Far
worse
than
the
fact
that
some
facilities'
HF
emissions
may
go
uncontrolled,

however,
is
the
fact
that
EPA
also
allows
facilities
that
qualify
for
the
HCl
exemption
to
ignore
emissions
of
all
other
inorganic
non­
metallic
HAPs,
some
of
which
may
not
even
be
identified.
The
Agency
itself
concedes
that
it
has
"
limited
emissions
information
for
 
inorganic
HAP[
s]"
other
than
HCl
and
metals,
128
yet
it
identifies
only
HCl,
HF,
and
chlorine
as
inorganic,
non­
metallics
likely
to
be
emitted
from
large,
solid­
fuel
boilers
 
an
oversimplification
at
best
naïve
and
at
worst
disingenuous
given
the
exhaustive
list
of
fuel
types
burned
by
such
facilities
(
including
"
coal,
wood,
biomass,
tires,
plastics,
and
other
nonfossil
solid
materials"
129).
Moreover,
EPA
expressly
discounts
hydrogen
cyanide
emissions
without
offering
any
support
for
the
claim
that
regulated
boilers
are
unlikely
to
emit
this
toxin.
Without
a
thorough
study
of
emissions
from
large,
solid­
fuel
boilers,
EPA
cannot
reasonably
adopt
a
risk­
based
HCl
exemption
that
effectively
exempts
all
inorganic
non­
metallic
HAPs
 
whatever
their
individual
toxicities
and
carcinogenicities
 
from
MACT
compliance.
For
these
additional
reasons,
the
HCl
exemption
violates
the
Act
and
is
arbitrary
and
capricious.

127
U.
S.
EPA,
Technology
Transfer
Network
Air
Toxics
Website:
Hydrogen
Flouride,
http://
www.
epa.
gov/
ttn/
atw/
hlthef/
hydrogen.
html.
128
69
Fed.
Reg.
at
55,230.
129
Id.
at
55,269.
34
f.
The
Mn
exemption
effectively
permits
sources
with
relatively
low
Mn
emissions
to
avoid
controlling
emissions
of
any
other
non­
mercury
metals.

The
Mn
exemption
is
just
as
flawed
as
the
HCl
exemption,
although
EPA
has
gone
out
of
its
way
to
conceal
the
flaw.
Because
EPA
does
not
expressly
use
Mn
as
a
surrogate
for
any
other
HAPs,
the
Mn
exemption's
only
direct
effect
is
to
permit
facilities
to
avoid
their
control
obligations
for
Mn.
In
practice,
however,
the
exemption
also
enables
facilities
to
avoid
controlling
emissions
of
other,
non­
mercury
metals.

To
see
this,
one
must
first
understand
both
how
the
ICI
boiler
and
process
heater
rule
regulates
emissions
of
non­
mercury
metals
and
how
the
Mn
exemption
operates.

The
rule
first
identifies
eight
non­
mercury
metallic
HAPs:
arsenic,
beryllium,
cadmium,

chromium,
lead,
manganese,
nickel,
and
selenium.
The
rule
then
adopts
particulate
matter
(
PM)
as
a
surrogate
for
these
HAPs,
because
"[
m]
ost,
if
not
all,
non­
mercury
metallic
HAP
emitted
from
combustion
sources
will
appear
on
the
flue
gas
fly­
ash,"
so
"
the
same
control
techniques
that
would
be
used
to
control
the
fly­
ash
PM
will
control
non­
mercury
metallic
HAP."
130
Because
"
some
sources
burn
fuels
containing
very
little
metals,
but
would
have
sufficient
PM
emissions
to
require
control
under
the
PM
provisions
of
the
proposed
rule,"
however,
the
final
rule
also
includes
"
an
alternative
[
total
selected
metals
(
TSM)]
limit.
 
A
source
may
choose
to
comply
with
the
alternative
[
TSM]
emissions
limit
instead
of
the
PM
limit
to
meet
the
final
rule."
131
The
Mn
exemption
operates
against
this
background.
Sources
must
first
submit
data
establishing
to
EPA's
satisfaction
that
(
1)
their
Mn
emissions
fall
below
certain
listed
levels,
or
(
2)
the
maximum
predicted
concentration
of
Mn
at
their
site
falls
below
130
Id.
at
55,223.
131
Id.
35
the
RfC
for
Mn.
132
Sources
able
to
meet
one
of
these
requirements
may
then
exclude
their
Mn
emissions
from
their
calculations
when
they
seek
to
establish
compliance
with
the
otherwise­
applicable
MACT
standard
for
non­
mercury
metallic
HAPs.
For
sources
that
choose
to
comply
with
the
PM
standard,
this
deduction
is
irrelevant.
Sources
that
choose
to
comply
with
the
TSM
standard,
however,
receive
a
windfall:
They
are
permitted
to
show
that
their
emissions
of
seven
non­
mercury
metals
(
arsenic,
beryllium,

cadmium,
chromium,
lead,
nickel,
and
selenium)
fall
below
the
standard
EPA
adopted
for
eight
such
metals.
This
is
particularly
problematic
for
facilities
whose
Mn
emissions
are
high
relative
to
their
emissions
of
other
non­
mercury
metals
 
to
instruct
such
facilities
to
total
only
their
emissions
of
the
seven
non­
mercury
non­
Mn
metals
in
determining
their
non­
mercury
metallic
HAP
emissions
is
like
instructing
banks
to
ignore
all
$
20
bills
in
determining
the
contents
of
their
ATM
machines.
Provided
such
facilities
can
comply
with
the
Mn
exemption,
their
emissions
of
the
remaining
seven
metals
will
almost
inevitably
fall
below
the
TSM
standard,
and
they
will
therefore
be
able
to
avoid
taking
any
steps
to
control
those
emissions
(
even
though
the
Agency
nowhere
asserts,
let
alone
establishes,
that
the
remaining
seven
metals
are
"
threshold
pollutants"
to
which
§
112(
d)(
4)
could
even
arguably
apply).
Indeed,
EPA
effectively
concedes
this
point
at
various
places
in
the
final
rule,
when
it
labels
the
Mn
exemption
a
"
TSM
compliance
alternative,"
133
and
observes
that
almost
400
biomass­
fired
boilers
"
could
be
potentially
eligible"
for
this
alternative.
134
This
flaw
in
the
Mn
exemption
is
not
only
illogical
but
unlawful.
As
noted
elsewhere,
the
CAA
requires
EPA
to
adopt
emissions
controls
comparable
to
those
132
See
id.
at
55,283,
Appendix
A
§
5(
d).
133
E.
g.,
id.
at
55,227,
55,244
(
emphasis
added);
id.
at
55,283,
Appendix
A
§
6(
b)
(
emphasis
added).
134
Id.
at
55,244.
36
achieved
by
the
best
performing
similar
sources.
To
comply
with
this
requirement
(
even
assuming
the
lawfulness
of
the
Mn
exemption
itself,
which
we
do
not
concede)
EPA
should
have
adopted
two
TSM
standards:
one
for
facilities
summing
emissions
of
all
eight
non­
mercury
metallic
HAPs,
and
one
for
facilities
excluding
their
Mn
emissions.

Instead,
EPA
permits
facilities
that
comply
with
the
Mn
exemption
to
compare
apples
and
oranges;
they
are
permitted
to
show
not
that
their
emissions
of
the
seven
remaining
non­
mercury
metallic
HAPs
are
comparable
to
the
same
set
of
emissions
from
the
best
performing
plants,
but
instead
that
their
emissions
of
those
seven
HAPs
are
comparable
to
the
best
performing
plants'
emissions
of
those
HAPs
plus
Mn.
This
result
violates
the
express
requirements
of
the
CAA
and
is
arbitrary
and
capricious.

g.
The
described
procedures
for
demonstrating
compliance
with
the
§
112(
d)(
4)
risk­
based
exemptions
are
significantly
flawed.

Appendix
A
to
the
final
rule
outlines
two
methods
by
which
sources
can
demonstrate
compliance
with
the
HCl
and
Mn
exemptions,
both
of
which
are
flawed.

First,
a
facility
may
conduct
emissions
tests
at
"
appropriate
emission
point[
s]"
135
under
"
worst­
case
operating
conditions,"
136
calculate
the
HCL­
equivalent
or
Mn
emissions
rate
for
each
emission
point,
and
establish
that
the
calculated
rate
falls
below
the
appropriate
value
in
the
Agency's
look­
up
tables.
Second,
a
source
that
fails
to
comply
with
this
"
look­
up
table
approach"
137
may
"
perform
a
site­
specific
compliance
demonstration"
and
submit
data
demonstrating
to
EPA's
satisfaction
that
the
source's
maximum
HI
for
HCl
and
chlorine,
or
HQ
for
Mn,
does
not
exceed
1.0.

135
Id.
at
55,283,
Appendix
A
§
5(
a).
136
Id.
at
55,282,
Appendix
A
§
4(
b)(
2).
137
Id.
at
55,283,
Appendix
A
§
7.
37
The
most
significant
flaw
in
the
look­
up
table
approach
is
that
EPA
instructs
sources
to
use
the
average
stack
height
of
their
multiple
emission
points
in
comparing
emissions
rates
to
the
values
in
the
tables138
 
an
oversimplification
that
may
understate
the
risks
posed
by
the
source.
For
example,
if
a
source
is
configured
such
that
its
shortest,
most
highly­
polluting
stack
is
located
closest
to
the
source's
neighbors,
a
calculation
that
estimates
risk
by
averaging
with
another
emission
point
that
is
taller,

cleaner,
and
farther
away,
may
not
accurately
reflect
the
risk
to
those
neighbors.
In
addition,
the
values
in
the
look­
up
tables
 
that
is,
the
maximum
permissible
emissions
rates
 
vary
depending
on
the
emission
point's
distance
from
the
property
boundary,
but
the
Agency
makes
no
attempt
to
account
for
other
variables
(
for
example,
topography
or
climatic
conditions,
or
proximity
of
local
population
centers)
that
could
also
significantly
affect
the
risks
a
plant's
emissions
pose.

Whatever
the
flaws
in
the
look­
up
table
approach,
however,
they
pale
in
comparison
to
the
inadequacies
of
the
site­
specific
compliance
demonstration,
at
least
as
the
latter
approach
is
described
in
the
final
rule.
The
only
guidance
EPA
gives
sources
with
respect
to
this
second
approach
is
that
they
must
use
a
"
scientifically­
accepted
peerreviewed
assessment
methodology,"
139
and
that
their
compliance
demonstration
must,
at
a
minimum,
(
1)
"[
e]
stimate
long­
term
inhalation
exposures
through
the
estimation
of
annual
or
multi­
year
average
ambient
concentrations;"
140
(
2)
"[
e]
stimate
the
inhalation
exposure
for
the
individual
most
exposed
to
the
facility's
emissions;"
141
(
3)
use
"
site­

138
Id.
at
55,282,
Appendix
A
§
6(
a)(
2).
139
Id.
at
55,282,
Appendix
A
§
7.
140
Id.
at
55,283­
84,
Appendix
A
§
7(
c)(
1).
141
Id.
at
55,283­
84,
Appendix
A
§
7(
c)(
2).
38
specific,
quality­
assured
data
wherever
possible;"
142
(
4)
use
"
health­
protective
default
assumptions
wherever
site­
specific
data
are
not
available;"
143
and
(
5)
"[
c]
ontain
adequate
documentation
of
the
data
and
methods
used."
144
This
cursory
outline
does
not
even
begin
to
place
real,
enforceable
limits
on
facilities'
assessment
methodology.
The
Agency
does
not,
for
example,
explain
how
facilities
are
to
estimate
inhalation
exposures.

Are
they
required
to
estimate
exposures
for
individuals
who
work
or
attend
school
in
the
vicinity
of
the
plant,
or
only
individuals
who
live
nearby?
Must
they
account
for
geographic
and
climatic
variations
that
could
affect
dispersion
of
the
pollutants,
such
as
variations
in
topography,
wind
speeds,
precipitation
levels,
or
humidity
levels?

Moreover,
EPA
uses
terms
like
"
quality­
assured
data"
and
"
health­
protective
default
assumptions"
that
are
entirely
meaningless
without
further
clarification.
Such
an
obscure
and
permissive
compliance
approach
cannot
possibly
satisfy
§
112(
d)(
4)'
s
"
ample
margin
of
safety"
requirement.
Indeed,
the
methodology
is
so
vague
and
unenforceable
that
one
wonders
whether
any
source
will
fail
to
qualify
for
the
exemptions,
no
matter
how
high
its
HCl
and
Mn
emissions.

As
discussed
above,
the
proposed
ICI
boiler
and
process
heater
rule
failed
to
discuss
the
Agency's
legal
justification
for
adopting
47
"
no
control"
standards
and
to
outline
the
possible
contours
of
the
risk­
based
exemptions
(
notably
omitting
any
mention
of
the
possibility
of
an
Mn
exemption).
Thus,
most
of
the
objections
raised
above
could
not
have
been
raised
at
the
comment
stage.
Yet
these
objections,
which
make
clear
that
neither
the
"
no
control"
standards
nor
the
exemptions
can
withstand
legal
challenge,
"
are
142
Id.
at
55,283­
84,
Appendix
A
§
7(
c)(
3).
143
Id.
at
55,283­
84,
Appendix
A
§
7(
c)(
4).
144
Id.
at
55,283­
84,
Appendix
A
§
7(
c)(
5).
39
of
central
relevance
to
the
outcome
of
the
rule."
145
Accordingly,
EPA
must
"
convene
a
proceeding
for
reconsideration
of
the
rule
and
provide
the
same
procedural
rights
as
would
have
been
afforded
had
the
information
been
available
at
the
time
the
rule
was
proposed."
146
Furthermore,
some
of
the
issues
we
raise
(
notably
the
creation
of
47
"
no
control"
standards
and
the
establishment
of
an
HCl
exemption
that
also
enables
facilities
to
avoid
controlling
emissions
of
other,
unidentified,
non­
metallic,
inorganic
HAPs)
are
so
crucial
to
the
rulemaking,
and
so
legally
deficient,
that
EPA
has
violated
its
"
duty
to
examine
key
assumptions
as
part
of
its
affirmative
`
burden
of
promulgating
and
explaining
a
non­
arbitrary,
non­
capricious
rule.'"
147
EPA
must
therefore
grant
reconsideration
of
those
issues
or
risk
the
rule's
invalidation
in
court.

Dated:
November
11,
2004
145
42
U.
S.
C.
§
7607(
d)(
7)(
B).
146
Id.
147
Small
Refiner
Lead
Phase­
Down
Task
Force
v.
EPA,
705
F.
2d
506,
534­
35
(
D.
C.
Cir.
1983)
(
quoting
National
Lime,
627
F.
2d
at
433).
See
also
Appalachian
Pwr.
Co.
v.
EPA,
135
F.
3d
791,
818
(
D.
C.
Cir.
1998).
