August
3,
2004
Mr.
Donald
S.
Welsh
(
3RA00)
Regional
Administrator
U.
S.
Environmental
Protection
Agency
Region
III
Office
1650
Arch
Street
Philadelphia,
PA,
19103­
2029
Dear
Mr.
Welsh:

We
recently
received
a
copy
of
the
Virginia
Department
of
Environmental
Quality's
(
VADEQ)
petition
to
EPA,
to
reclassify
the
Richmond
ozone
nonattainment
area
(
NAA)
from
a
"
moderate"
classification
to
a
"
marginal"
classification.
We
have
also
received
the
technical
support
document,
entitled
"
Richmond
Ozone
Nonattainment
Area
Petition
for
Reclassification
Technical
Support
Document,"
dated
July
12,
2004
(
hereafter
referred
to
as
TSD).
As
a
downwind
state
from
the
Richmond
NAA,
Delaware
is
providing
EPA
with
specific
comments
on
VADEQ's
TSD
relative
to
their
air
monitoring
data
analysis,
ozone
modeling
analysis,
and
precursor
emissions
analysis.

As
you
know,
over
the
past
30
years
we
have
learned
that
our
ozone
nonattainment
problem
is
pervasive,
and
that
both
the
extent
of
the
problem
and
the
emissions
that
cause
the
problem
are
very
large
in
scale.
Given
the
pervasive
nature
of
ozone,
Delaware
believes
that
the
only
solution
is
to
subject
upwind
areas
to
the
same
level
of
control
as
downwind
nonattainment
areas
(
see
Delaware's
July
14,
2003,
February
2,
2004,
and
February
27,
2004
letters
for
a
detailed
discussion
of
this
issue).
Delaware
believes
VADEQ's
petition
effectively
removes
the
Richmond
NAA
from
its
responsibility
as
an
upwind
area
to
Delaware,
and
therefore
requests
that
it
not
be
approved
by
the
EPA.

I.
Air
Monitoring
Data
Analysis
The
EPA
guidance
for
nonattainment
"
bump­
downs"
requires
that
both
near­
and
long­
term
trends
in
emissions
and
ambient
air
quality
support
any
downward
reclassification,
2
and
that
historical
air
quality
data
should
indicate
substantial
air
quality
improvement
(
Ref.
1).
It
is
Delaware's
position
that
the
VADEQ's
petition
does
not
include
historical
data
that
demonstrates
a
substantial
air
quality
improvement.
The
TSD
primarily
focuses
on
ambient
concentrations
which
indicate
a
near­
flat
trend
at
three
out
of
the
four
Richmond
NAA
monitors.
The
fourth
monitor
actually
shows
an
increasing
trend.
The
monitoring
data
therefore
does
not
meet
EPA's
guidance
for
a
bump­
down.
Additionally,
this
fourth
monitor
is
the
primary
downwind
monitor,
which
calls
into
question
the
effect
of
transport
from
the
Richmond
area
to
other
downwind
non­
attainment
areas
(
discussed
below).

The
TSD
states
that
"
three
of
the
four
monitors
in
the
Richmond
metro
area
have
and
continue
to
record
levels
consistent
with
a
marginal
nonattainment
classification."
It
should
be
pointed
out
that
ozone
nonattainment
classifications
focus
on
the
highest
representative
monitor
in
the
designation
area,
and
even
if
other
monitors
in
the
area
report
lower
concentrations,
they
should
not
have
any
impact,
unless
a
strong
argument
can
be
made
and
supported
that
the
highest
monitor
is
non­
representative
or
biased
(
Ref.
2).
The
VADEQ
failed
to
make
such
an
argument,
but
rather
stated
that
the
highest
monitor
is
in
an
area
of
lower
population
so
that
most
people
are
exposed
to
concentrations
in
the
marginal
range
rather
than
the
moderate
range.
This
specific
argument
is
not
consistent
with
EPA's
standard
for
classification.

The
TSD
includes
a
graph
of
the
average
design
value
(
page
4)
among
the
four
monitors
in
the
Richmond
nonattainment
area,
and
tries
to
conclude
that
the
area's
design
values
"
are
consistent
with
a
marginal
nonattainment
classification."
As
we
understand
the
process,
spatial
averaging
is
not
an
acceptable
procedure
for
ozone
classifications.
Further,
this
same
graph
shows
an
increasing
trend,
rather
than
a
decreasing
or
stable
trend.
As
mentioned
above,
this
increasing
trend
fails
to
meet
EPA's
bump­
down
guidance
requiring
a
trend
of
substantial
air
quality
improvement.
On
the
same
page
of
the
TSD
is
another
graph
showing
the
number
of
exceedance
days.
This
graph
appears
to
show
a
flat
trend
for
the
most
recent
years
of
data,
which
again
does
not
demonstrate
a
substantial
improvement
in
air
quality.

Based
on
the
above
analysis,
we
believe
that
the
first
three
conclusions
in
the
Air
Monitoring
Data
Analysis
section
of
the
TSD
do
not
have
an
appropriate
or
acceptable
technical
foundation.
Also,
the
fourth
conclusion
("
no
exceedances
of
the
8­
hour
standard
have
been
recorded
in
the
Richmond
area
during
the
2004
ozone
season")
is
no
longer
valid,
since
on
July
21,
2004
the
Charles
City
County
monitor
recorded
an
exceedance
of
91
ppb.
It
also
should
be
pointed
out
that
in
addition
to
this
one
exceedance
of
the
8­
hour
standard
in
the
Richmond
areas
so
far
this
year,
(
1)
the
2004
ozone
season
is
not
yet
over,
and
(
2)
the
2004
ozone
season
may
be
an
unusually
wet
and
cool
one,
as
the
weather
has
demonstrated
thus
far,
and
thus
may
not
be
a
good
indicator
of
attainment
status.

II.
Regional
Ozone
Modeling
Analysis
The
VADEQ
fails
to
discuss
issues
regarding
ozone/
precursor
transport
and
contribution
from
Richmond
NAA
to
its
downwind
states.
Long­
range
transport
and
3
contribution
of
ozone
and
its
precursors
from
an
upwind
area
to
the
downwind
area
has
been
well
documented
and
recognized.
For
example,
the
NOx
SIP
Call
modeling
results
(
using
UAM­
V
and
CAMx)
have
indicated
that
Virginia
is
an
upwind
area
that
contributes
significantly
to
the
8­
hour
ozone
nonattainment
problem
in
Delaware.
In
particular,
the
UAM­
V
and
CAMx
models
have
estimated
that
Virginia's
contribution
to
Delaware
to
be
between
10
and
60
ppb
(
See
Attachment
1
to
this
letter,
and
Reference
3).

Similar
modeling
work
was
performed
by
EPA
for
its
Interstate
Air
Quality
Rule
(
IAQR).
In
this
more
recent
work,
the
CAMx
source
apportionment
modeling
results
have
demonstrated
that
Virginia
contributes
to
Delaware's
New
Castle
County
from
8
ppb
to
a
maximum
of
17
ppb
in
2010
(
Reference
4).
Delaware
believes
that
(
1)
this
contribution
is
definitely
significant,
and
(
2)
if
this
contribution,
and
like
contributions
from
other
upwind
states
are
not
removed,
it
will
be
very
difficult,
if
not
impossible,
for
Delaware
to
reach
attainment
by
2010.

Also,
recent
CALGRID
modeling
performed
by
Delaware
using
2010
OTC
NOx
Resolution
emissions,
which
accounts
for
22%
NOX
reduction
from
CSA2003
elevated
point
sources,
shows
as
much
as
24
ppb
contribution
from
Virginia
to
Delaware's
1­
hour
ozone
concentrations.
A
discussion
of
these
modeling
results
is
attached
to
this
letter
(
See
Attachment
3
to
this
letter).

Note
that
all
of
the
above
mentioned
modeling
excursions,
which
include
the
reductions
from
the
federal
measures
VADEQ
mentions
in
their
TSD,
indicate
that
a
significant
contribution
from
Virginia
to
Delaware
remains
in
2010
(
i.
e.
Delaware's
attainment
date).
This
means
reductions
in
Virginia
beyond
these
federal
measures
are
needed
to
mitigate
the
contribution
from
Virginia
to
Delaware.
Delaware
believes
the
Richmond
NAA
moderate
classification
is
consistent
with
their
obligation
to
mitigate
this
negative
contribution.
For
example,
Rate­
of­
progress
(
ROP)
emission
reduction
requirements
are
mandated
in
moderate
NAA's
by
the
CAAA,
are
subject
to
EPA
and
public
review,
and
will
help
Virginia
mitigate
its
negative
impact
on
Delaware.
No
ROP
reductions
are
required
under
the
CAAA
for
marginal
areas,
so
if
the
bump­
down
were
approved
the
Richmond
NAA
would
not
be
required
to
reduce
emission
beyond
the
federal
measures
mentioned
in
the
TSD
(
i.
e.,
beyond
the
measures
that
modeling
clearly
shows
are
insufficient).
Given
the
proven
negative
impact
of
up­
wind
emissions
on
downwind
nonattainment,
it
is
not
appropriate
to
exempt
areas
with
significant
emissions
from
these
ROP
requirements.

III.
Precursor
Emission
Analysis
In
this
section
of
the
TSD,
VADEQ
presents
its
estimates
of
precursor
emission
reductions
between
2002
and
2007,
and
claims
that
"
significant
reductions
in
both
VOC
and
NOx
emissions
are
expected
in
the
Richmond
area."
Attachment
2
to
this
letter
presents
a
simple
comparison
of
emissions
between
Richmond
NAA
and
Delaware
NAA.
This
comparison
demonstrates
that
both
the
total
emissions
and
the
spatial
emission
intensity
(
SEI,
in
term
of
TPD
per
square
mile)
of
the
Richmond
NAA
are
much
higher
than
that
of
Delaware.
We
believe
that
total
emissions
and
the
SEI
are
useful
tools
to
help
evaluate
4
nonattainment
area
classifications
relative
to
large­
scale,
regionally
driven
air
quality
problems,
like
ozone
pollution.

EPA's
bump­
down
guidance
requires
that
growth
projection
and
emission
trends
support
a
bump­
down
request
(
Ref.
1).
The
VADEQ
does
not
appear
to
present
any
discussion
on
growth
projection
for
Richmond
NAA.
VADEQ
only
lists
three
major
categories
of
existing
and
future
federal
controls
on
which
the
Richmond
NAA
depends
for
significant
emission
reductions
in
2007,
some
of
which
appear
to
produce
reductions
only
after
2007.
These
controls
are
not
sufficient
to
mitigate
Virginia's
negative
impact
on
Delaware
(
see
modeling
discussion
above).

Based
on
total
emissions
and
the
SEI,
it
is
unreasonable
and
unacceptable
to
classify
Richmond
area
as
a
"
marginal
NAA"
that
would
be
subject
to
less
stringent
requirements
in
its
emission
reductions,
while
maintaining
Delaware,
which
has
smaller
total
emissions
and
SEIs,
as
a
"
moderate
NAA"
requiring
more
emission
reductions.
This
does
not
indicate
that
Delaware
supports
a
bump­
down
for
its
own
non­
attainment
counties,
rather
it
supports
that
all
areas
with
significant
emissions
that
impact
downwind
nonattainment
areas
be
subject
to
consistent
and
specific
ROP
emission
reduction
requirements.
Such
areas
include,
relative
to
Delaware,
those
in
Michigan,
Ohio,
Pennsylvania,
West
Virginia,
Virginia,
North
Carolina,
New
Jersey,
Delaware,
Maryland,
and
New
York.

IV.
Concluding
Remarks
Based
on
our
review
and
analysis
of
VADEQ's
TSD
and
relevant
data,
Delaware's
conclusions
and
position
relative
to
VADEQ's
petition
are
as
follows:

1.
The
VADEQ
does
not
provide
adequate
air
quality
analysis
and
arguments
to
support
its
petition.
2.
The
VADEQ
fails
to
adequately
address
issues
regarding
ozone/
precursor
transport.
Delaware
is
extremely
concerned
about
how
VADEQ
will
deal
with
Virginia's
significant
contributions
to
nonattainment
problems
in
its
downwind
states,
including
Delaware.
3.
Delaware
is
against
approval
of
such
a
petition,
unless
the
transport
and
contribution
issues
are
addressed
to
our
satisfaction.
4.
It
would
be
unreasonable
and
unfair
to
Delaware
(
and
to
other
downwind
states
as
well)
if
VADEQ's
petition
is
approved
to
bump­
down
the
Richmond
NAA,
since
the
Richmond
NAA
has
significant
emissions
that,
based
on
modeling
and
lacking
any
additional
local
controls,
will
continue
to
negatively
impact
Delaware
after
both
the
2007
marginal
and
2010
moderate
attainment
dates.

I
also
understand
that
requests
for
8­
hour
ozone
classification
"
bump­
downs"
have
been
requested
by
Maryland
and
Pennsylvania.
I
have
requested
copies
of
these
requests
from
EPA,
but
have
not
yet
received
them.
I
request
that
EPA
not
take
any
final
action
on
these
other
requests
until
Delaware
has
been
afforded
ample
time
to
review
them,
and
until
EPA
has
considered
any
comments
Delaware
may
have
in
making
their
final
determination.
5
Thank
you
for
considering
our
comments.
If
you
have
any
question,
or
would
like
to
discuss
this
matter
further,
please
contact
me
at
(
302)
739­
4791.

Sincerely,

Ali
Mirzakhalili
Administrator
cc:
John
A.
Hughes
John
B.
Blevins
Judith
Katz
David
Arnold
6
References
1.
"
Air
Quality
Designations
and
Classifications
for
the
8­
Hour
Ozone
National
Ambient
Air
Quality
Standards;
Early
Action
Compact
Areas
With
Deferred
Effective
Dates",
USEPA,
40
CFR
Part
81,
Federal
Register:
April
30,
2004
(
Volume
69,
Number
84),
Rules
and
Regulations,
Page
23857­
23951.

2.
"
Final
Rule
To
Implement
the
8­
Hour
Ozone
National
Ambient
Air
Quality
Standard­­
Phase
1",
USEPA
40
CFR
Parts
50,
51
and
81,
Federal
Register:
April
30,
2004
(
Volume
69,
Number
84),
Rules
and
Regulations,
Page
23951­
24000.

3.
Appendix
D:
8­
Hour
Upwind/
Downwind
Linkages,
Air
Quality
Modeling
Technical
Support
Document
for
the
NOx
SIP
Call,
US
EPA,
Office
of
Air
and
Radiation,
September
23,
1998.

4.
Appendix
G:
Metrics
for
8­
Hour
Ozone
Contributions
to
Downwind
Nonattainment
Counties
in
2010,
Technical
Support
Document
for
the
Interstate
Air
Quality
Rule
Air
Quality
Modeling
Analyses,
USEPA,
January
2004.
7
Attachment
1
Contributions
from
Upwind
States
to
Delaware's
8­
Hour
Ozone
Nonattainment.
UAM­
V
Zero­
Out
CAMx
Modeling
Contributing
State
Min,
ppb
Max,
ppb
Min,
ppb
Max,
ppb
Illinois
2
3
2
7
Indiana
2
3
2
6
Kentucky
2
4
5
9
Michigan
2
7
2
9
North
Carolina
5
22
5
28
Ohio
2
6
5
14
Pennsylvania
­­­
­­­
10
32
Tennessee
2
3
2
7
Virginia
10
34
10
60
West
Virginia
5
15
5
17
Note:
Data
in
this
table
are
compiled
from
NOx
SIP
Call
modeling
results
(
Ref.
3).

Attachment
2
Comparison
of
2002
Ozone
Season
Daily
Emissions
Richmond
Delaware
Richmond
Delaware
NAA
NAA
NAA
NAA
VOC
VOC
N0x
N0x
Total
emission
in
2002,
TPD
139.5
117.1
238.5
207.0
Land
area,
sq.
miles
1686
1983
1686
1983
Emission
TPD
per
sq.
mile
0.08
0.06
0.14
0.10
Notes:
Richmond
emission
data
from
VADEQ's
Petition
Technical
Supporting
Document.
Delaware
emission
data
from
its
June
1,
2004
submission
to
EPA's
NEI
Database.
Land
area
data
from
US
Census
Bureau
website
http://
www.
census.
gov/
geo/
www/
ezstate/
poverty.
html
Attachment
3
Delaware's
Modeling
Results
for
Transport
and
Contribution
from
Virginia
to
Delaware
Delaware's
own
modeling
analyses
has
indicated
that
transport
from
Virginia
into
Delaware
is
significant,
and
cannot
be
ignored.
Figures
1,
2,
and
3
depict
48­
hour
back
trajectories
that
correspond
to
the
days
with
the
highest
ozone
values
in
Delaware.
Figure
1
clearly
indicates
that
the
trajectory
came
from
the
Richmond
ozone
non­
attainment
area
on
11th
June
2001,
a
day
for
which
Sussex
County
was
in
non­
attainment
for
ozone.
Therefore,
based
on
this
evidence
we
can
conclude
Richmond
ozone
non­
attainment
area
contributes
to
Delaware's
ozone
non­
attainment
on
11th
June
2001.
Figures
2
and
3
also
indicate
that
Virginia
causes
transport
of
ozone
and
its
precursors
thereby
aggravating
Delaware's
nonattainment
of
ozone
NAAQS.

We
performed
CALGRID
modeling
for
June
1995
episode
with
emissions
for
2010
OTC
NOx
Resolution,
which
accounts
for
22%
NOx
reduction
from
CSA2003
elevated
point
sources.
The
June
1995
episode
is
a
12­
day
episode
with
June
12th
as
the
first
day.
Virginia's
8
impact
is
assessed
by
zeroing
out
emissions
in
the
modeling
domain
from
all
states
but
for
Virginia's.
Edge
effects
due
to
boundary
conditions
are
modeled
by
zeroing
out
emissions
in
the
entire
modeling
domain.
Virginia's
impacts
are
assessed
by
subtracting
the
latter
concentrations
from
the
former.
Hourly
variations
of
such
impacts
are
captured
for
the
entire
modeling
episode.
Figures
4
and
5
clearly
show
that
Virginia
impacts
Delaware
on
both
days
­
June
21st
and
22nd.
These
figures
also
demonstrate
that
Virginia's
contribution
to
Delaware's
hourly
ozone
contribution
could
be
as
much
as
24
ppb
on
both
days.
Such
a
contribution
is
significant,
which
might
prevent
Delaware
from
attaining
the
8­
Hr
ozone
NAAQS
in
spite
of
its
local
control
measures.
In
other
words,
Delaware's
local
control
measures
alone
will
not
be
enough
to
bring
the
area
into
attainment
without
addressing
the
transport
from
Delaware's
upwind
states
including
Virginia.

The
evidence
presented
above
clearly
points
to
the
fact
that
Virginia
is
an
upwind
state
to
Delaware
and
contributes
significantly
to
Delaware's
ozone
non­
attainment.
Therefore,
EPA
should
account
for
Virginia's
ozone
transport
into
Delaware
as
one
of
the
major
factors
for
not
bumping
down
Richmond
metro
area
to
a
marginal
ozone
non­
attainment
area.
9
Figure
1:
The
48­
Hour
Back
Trajectories
for
June
11,
2002.
10
Figure
2:
The
48­
Hour
Back
Trajectories
for
June
27,
2001.
11
Figure
3:
The
48­
Hr
Back
Trajectories
for
July
17,
2001.
12
Figure
4:
Virginia's
contribution
to
its
neighboring
states
for
June
20,
1995
episode
day
for
2010
OTC
NOx
Resolution.
13
Figure
5:
Virginia's
contribution
to
its
neighboring
states
for
June
21,
1995
episode
day
for
2010
OTC
NOx
Resolution.
