1
July
15,
2004
Analysis
of
8­
Hour
Ozone
Concentrations
in
Cass
County,
Michigan;
and
Muskegon
County,
Michigan
The
purpose
of
this
paper
is
to
summarize
ozone
modeling
and
data
analyses
for
Cass
County,
Michigan;
and
Muskegon
County,
Michigan,
to
support
reclassification
of
these
counties
from
moderate
to
marginal
nonattainment.

Introduction
On
April
15,
2004,
the
U.
S.
Environmental
Protection
Agency
(
USEPA)
designated
areas
as
attainment
(
or
nonattainment)
of
the
8­
hour
ozone
standard.
The
nonattainment
areas
were
sorted
under
either
Subpart
1
or
Subpart
2
of
the
Clean
Air
Act.
The
Subpart
2
areas
are
classified
as
marginal,
moderate,
serious,
or
severe
based
on
the
area's
8­
hour
design
value
calculated
using
the
most
recent
three
years
of
data.
Section
181(
a)(
4)
of
the
Clean
Air
Act
allows
Subpart
2
nonattainment
areas
to
be
reclassified
if
that
area
would
have
been
classified
in
another
category
if
the
design
value
in
the
area
were
five
percent
greater
or
less
than
the
level
on
which
the
classification
was
based.
For
example,
moderate
nonattainment
areas
may
be
"
bumped­
down"
to
marginal
nonattainment
areas,
if
their
design
value
is
96
ppb
or
less.
In
West
Michigan,
the
following
counties
are
eligible
to
be
bumped­
down
from
moderate
to
marginal:
Cass
County,
Michigan
(
2001­
2003
design
value
=
93
ppb);
and
Muskegon
County,
Michigan
(
95
ppb).

The
USEPA
identified
the
following
criteria
to
support
a
classification
adjustment
(
bumpdown
request
by
state,
discontinuity
(
i.
e.,
reclassification
must
not
result
in
an
illogical
or
excessive
discontinuity
relative
to
the
classifications
of
surrounding
areas),
attainment
(
i.
e.,
show
that
the
proposed
area
will
be
able
to
attain
by
the
earlier
date
specified
for
the
lower
classification);
emission
reductions
(
i.
e.,
show
that
the
area
will
achieve
the
appropriate
emission
reduction
necessary
to
attain
by
the
earlier
date);
and
trends.
This
paper
contains
information
on
ozone
air
quality,
including
trends
in
8­
hour
ozone
concentrations
for
the
two
counties
in
question,
and
modeling
results
for
2007,
the
attainment
year
for
marginal
nonattainment
areas.

Ozone
Air
Quality
A
few
simple
parameters
are
presented
here
to
characterize
the
change
in
ozone
air
quality
over
time:
number
of
exceedance
days,
number
of
site
exceedance
days,
and
design
value
(
i.
e.,
average
of
fourth
high
concentration
over
3­
year
period).

The
figure
below
shows:
(
a)
the
number
of
exceedance
and
site
exceedances
days
for
the
1­
hour
ozone
standard,
(
b)
the
number
of
exceedance
and
site
exceedances
days
for
the
8­
hour
ozone
standard,
and
(
c)
the
number
of
hot
days
and
cooling
degree
days
for
the
period
1987
 
2003
in
the
Lake
Michigan
region.
These
plots
show:

 
Ozone
is
strongly
influenced
by
meteorology.
The
number
of
exceedance
days
(
and
site
exceedance
days)
is
higher
during
the
hotter
summers.
 
There
appears
to
be
a
general
downward
trend
in
1­
hour
and
8­
hour
ozone
levels
from
the
late
1980s
through
the
early
1990'
s,
but
little
change
since
then.
2
July
15,
2004
 
The
improvement
in
ozone
air
quality
since
the
late
1980s
is
consistent
with
the
reduction
in
regional
Volatile
Organic
Compound
(
VOC)
emissions
in
upwind
states
due
to
motor
vehicle
control
programs
(
e.
g.,
inspection
and
maintenance,
and
reformulated
gasoline);
area
source
control
programs;
and
stationary
source
controls.
While
these
control
programs
have
been
successful
in
"
shaving"
the
peak
1­
hour
ozone
levels,
additional
emission
reductions
are
needed
to
lower
8­
hour
ozone
levels.
Further
progress
in
reducing
8­
hour
ozone
levels
is
expected
to
come
from
reductions
in
regional
Oxides
of
nitrogen
(
NOx)
emissions
(
i.
e.,
NOx
SIP
Call,
Wisconsin's
NOx
rule,
federal
nonroad
standards,
and
possible
additional
federal
programs,
such
as
the
proposed
Interstate
Air
Quality
Rule).
(
Note,
additional
discussion
of
these
trends
and
the
expected
future
improvements
is
provided
in
a
Lake
Michigan
Air
Directors'
Consortium
(
LADCO)
document
titled
"
Mid­
Course
Review
for
1­
Hour
Ozone
in
the
Lake
Michigan
Region",
May
10,
2004.)

0
10
20
30
40
50
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
1­
hr
Exceedance
Days
0
50
100
150
200
250
1­
hr
Site
Exceedance
Days
Exceedance
Days
Site
Exceedance
Days
0
15
30
45
60
75
90
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
8­
hr
Exceedance
Days
0.00
150.00
300.00
450.00
600.00
750.00
900.00
8­
hr
Site
Exceedance
Days
Exceedance
Days
Site
Exceedance
Days
3
July
15,
2004
0
10
20
30
40
50
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
Days
>
90
F
0.00
300.00
600.00
900.00
1200.00
1500.00
Cooling
Degree
Days
Days
>
90F
Cooling
Degree
Days
Figure
1.
Trends
in
Ozone
and
Weather
in
Lake
Michigan
Region
Spatial
plots
of
8­
hour
design
values
below
also
indicate
a
decrease
since
the
late
1980s,
but
little
change
since
then.
4
July
15,
2004
Figure
2.
8­
Hour
Design
Values
for
1987­
1989,
1990­
1992,
1993­
1995,
1996­
1998,
and
1999­
2001
5
July
15,
2004
For
the
two
counties
in
question,
the
trends
in
8­
hour
ozone
(
e.
g.,
number
of
days
above
standard
and
design
value)
indicate
a
decrease
from
the
late
1980s
through
the
early
1990'
s,
but
little
change
since
then.

0
6
12
18
24
30
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
8­
Hour
Exceedance
Days
Cass
County
Muskegon
County
70
85
100
115
130
80­
82
81­
83
82­
84
83­
85
84­
86
85­
87
86­
88
87­
89
88­
90
89­
91
90­
92
91­
93
92­
94
93­
95
94­
96
95­
97
96­
98
97­
99
98­
00
99­
01
00­
02
01­
03
Design
Value
Cass
County
Muskegon
County
Figure
3.
Trends
in
8
 
Hour
Ozone
in
Two
Counties
Given
the
effect
of
meteorology
on
ozone,
it
is
necessary
to
adjust
the
ozone
trends
for
meteorological
influences.
A
simple
metric
was
considered
here
(
i.
e.,
number
of
exceedance
days
divided
by
the
number
of
hot
days).
The
plot
of
this
metric
below
shows
that
the
8­
hour
trends
are
relatively
flat
over
the
past
decade.
6
July
15,
2004
0
0.4
0.8
1.2
1.6
2
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
Cass
County
Muskegon
County
Figure
4.
Trends
in
Weather­
Adjusted
Ozone
Metric
(
No.
Exc.
Days/
No.
Hot
Days)
in
Two
Counties
More
rigorous
(
statistical)
adjustments
were
used
to
support
ozone
trends
analyses
for
sites
in
the
Lake
Michigan
region
(
see
"
Mid­
Course
Review
for
1­
Hour
Ozone
in
the
Lake
Michigan
Region",
May
10,
2004).
The
analyses
show
a
slight
downward
trend
in
8­
hour
ozone
levels
over
the
past
10
years
for
Muskegon,
although
the
trend
is
not
statistically
significant.
These
results
suggest
that
additional
emission
reductions
(
i.
e.,
regional
NOx
emissions)
are
needed
to
improve
ozone
air
quality
in
the
region.
The
effect
of
these
emission
reductions
is
discussed
in
the
following
section.

Modeling
Results
I
A
preliminary
8­
hour
assessment
was
conducted
by
LADCO
using
regional
modeling
data
and
the
USEPA's
recommended
8­
hour
attainment
test
(
see
"
8­
Hour
Ozone
Assessment",
May
2,
2001).
1
The
modeled
future
year
design
values
for
2007
are
shown
in
the
table
below.

Observed
Future
Year
Design
Value
(
2007)
Site
County
Design
Value
CAA
Controls
Regional
Strategy2
Cassopolis
Cass
93
88
81
Muskegon
Muskegon
97
92
86
1
This
modeling
was
performed
to
support
the
1­
hour
attainment
demonstration
for
the
Lake
Michigan
area
and,
as
such,
there
are
limitations
with
using
it
to
assess
8­
hour
ozone.
For
example,
the
episodes
and
modeling
domain
were
selected
for
the
Lake
Michigan
area
and
may
not
accurately
represent
other
cities
in
the
modeling
domain,
such
as
St.
Louis
and
Detroit,
and
the
modeling
reflects
a
2007
future
year
scenario
(
note:
the
actual
8­
hour
attainment
date
is
expected
to
be
about
2012).
On
the
other
hand,
it
should
be
noted
that
three
of
the
four
modeled
episodes
are
representative
periods
for
high
8­
hour
ozone
in
the
Lake
Michigan
area
and
basecase
model
performance
for
8­
hour
ozone
was
found
to
be
as
good
as
(
or
better
than)
that
for
1­
hour
ozone.

2
The
1­
hour
regional
control
strategy
includes
CAA
controls,
Tier
II/
low
S,
NOx
SIP
call,
and
WI's
NOx
rule.
7
July
15,
2004
The
modeling
results
indicate
that
the
(
1­
hour)
regional
control
strategy
will
reduce
8­
hour
ozone
levels
but
may
not
be
enough
to
provide
for
attainment
of
the
8­
hour
standard
at
all
locations.
It
should
be
noted,
however,
that
the
observed
design
values
are
based
on
the
average
of
the
design
values
of
the
three
3­
year
periods
which
include
1996
(
i.
e.,
the
base
inventory
year
used
in
the
modeling).
The
three
3­
year
design
values
for
each
site
are
as
follows:

Site
1994­
1996
1995­
1997
1996­
1998
Average
Cassopolis
94
94
92
93
Muskegon
101
99
91
97
The
USEPA's
draft
modeling
guidance
("
Draft
Guidance
on
the
Use
of
Models
and
Other
Analyses
in
Attainment
Demonstrations
for
the
8­
Hour
Ozone
NAAQS,"
May
1999,
EPA­
454/
R­
99­
004)
recommends
using
the
higher
of
the
3­
year
period
"
straddling"
the
inventory
year
(
i.
e.,
1995­
1997
for
a
1996
inventory)
and
the
3­
year
period
used
to
designate
the
area
nonattainment.
The
design
values
associated
with
this
approach,
compared
to
the
alternative
approach
above
generally
shows
similar
results.

Higher
of .
Average
of
Three
Site
1995­
1997
2001­
2003
3­
year
Periods
Cassopolis
94
93
93
Muskegon
99
95
97
Modeling
Results
II
Additional
8­
hour
ozone
results
are
available
from
recent
modeling
conducted
by
LADCO
to
assess
the
impact
of
the
USEPA's
proposed
Clean
Air
Interstate
Rule
(
CAIR)
­
e.
g.,
"
Interstate
Air
Quality
Rule:
Modeling
Assessment,"
March
26,
2004.

The
resulting
modeled
design
values
are
shown
in
the
table
below.
Observed
Site
County
Design
Value
2010
base3
2010
IAQR4
Cassopolis
Cass
90
80.5
(
79.6)
79.8
(
79.1)

Muskegon
Muskegon
89
81.0
(
79.6)
80.1
(
79.0)
*
=
result
based
on
IPM
source­
specific
data
3
The
2010
base
inventory
for
all
sectors,
except
EGUs,
was
developed
using
economic
and
population
growth
projections,
along
with
emission
reductions
from
current
regulations,
including
the
NOx
SIP
Call;
Tier
II
vehicle
standards;
heavy
duty
diesel
vehicle
standards;
non­
road
diesel
proposed
standards;
NOx
and
VOC
reductions
from
recreational
vehicle/
large
spark
ignition
engine
rules;
SO2
and
PM2.5
reductions
from
the
industrial
boiler
MACT;
mostly
VOC
reductions
from
a
large
number
of
earlier
MACTs;
PM,
SO2,
and
NOx
reductions
from
a
small
set
of
MACTs;
and
VOC
reductions
from
national
rules
for
Marine
Vessel
Loading
of
Petrol
Liquids,
TSDFs,
and
Landfills.
It
does
not
include
reductions
from
two
other
recent
MACTs
(
Gas
Turbines
and
Reciprocating
Internal
Combustion
Engines),
but
these
are
relatively
small
reductions.

4
The
2010
CAIR
strategy
includes
the
2010
base,
plus
the
SOx
and
NOx
reductions
from
the
proposed
Interstate
Air
Quality
Rule.
8
July
15,
2004
These
results,
which
are
consistent
with
the
USEPA's
modeling
for
the
proposed
rule
(
see
"
Technical
Support
Document
for
the
CAIR,
Air
Quality
Modeling
Analyses"
January
2004),
show
that
the
two
counties
in
question
are
expected
to
be
in
compliance
with
the
8­
hour
ozone
standard
by
2010.

The
relevance
of
this
modeling
for
2007
needs
to
consider
the
difference
in
emissions
(
especially,
NOx
emissions)
between
2007
and
2010.
An
approximation
of
this
difference
can
be
made
comparing
the
2007
inventory
developed
by
the
USEPA
for
its
heavy­
duty
diesel
(
HDD)
rulemaking
and
the
2010
(
base)
inventory
developed
by
the
USEPA
for
the
proposed
CAIR.
(
Note,
the
derivation
of
the
2010
CAIR
inventory
relied
on
the
2007
HDD
inventory5,
as
well
as
the
2010
inventory
developed
by
the
USEPA
for
its
proposed
land­
based
nonroad
diesel
engine
(
LNDE)
rulemaking6.)
A
simple
comparison
of
the
2007
HDD
(
with
an
adjustment
for
nonroad
emissions)
and
the
2010
CAIR
inventories
shows
about
a
nine
percent
difference
in
NOx
emissions.
(
Note,
the
VOC
emissions
for
these
two
inventory
years
are
within
a
few
percent
of
each
other).
Furthermore
discussion
of
the
difference
in
NOx
emissions
is
provided
below
by
source
sector.

Point
Sources:
NOx
emissions
from
EGUs
are
about
eight
percent
less
in
the
2010
CAIR
(
base)
inventory
compared
to
the
2007
HDD
inventory.
The
major
control
program
affecting
EGUs
is
the
NOx
SIP
call.
The
NOx
SIP
Call
requires
22
States
and
the
District
of
Columbia
to
submit
State
Implementation
Plans
that
address
the
regional
transport
of
ground­
level
ozone
through
reductions
in
NOx
emissions.
The
rule
affects
EGUs,
as
well
as
large
non­
utility
point
sources
(
i.
e.,
large
industrial
boilers
and
turbines,
large
internal
combustion
engines,
and
cement
manufacturing).
The
compliance
date
for
the
NOx
emission
budgets
is
May
2004.

On­
Road
Highway
Vehicles:
NOx
emissions
from
on­
road
sources
are
about
17
percent
less
in
the
2010
CAIR
(
base)
inventory
compared
to
the
2007
HDD
inventory.
The
major
control
programs
affecting
on­
road
sources
are
Tier
2/
low
sulfur
gasoline
and
the
HDD
rule
(
2007
Heavy­
Duty
Highway
Rule).
The
Tier
2/
low
sulfur
gasoline
program
establishes
new
tailpipe
standards
for
all
classes
of
passenger
vehicles,
light­
duty
trucks,
and
SUVs
that
will
be
phased­
in
between
2004
and
2007,
and
requires
that
the
level
of
sulfur
in
gasoline
be
reduced
by
up
to
90
percent
in
phases
between
2004
and
2007.
The
HDD
rule
establishes
new
emission
standards
for
heavy­
duty
highway
engines
and
vehicles
that
will
be
phased­
in
between
2007
and
2010,
and
requires
that
the
level
of
sulfur
in
highway
diesel
fuel
be
reduced
by
97
percent
by
mid­
2006.

5
See
"
Procedures
for
Developing
Base
Year
and
Future
Year
Mass
and
Modeling
Inventories
for
the
Heavy­
Duty
Engine
and
Vehicle
Standards
and
the
Highway
Diesel
Fuel
(
HDD)
Rulemaking"
(
EPA/
420­
R­
00­
020,
October
2000)
and
"
Data
Summaries
of
Base
Year
and
Future
Year
Mass
and
Modeling
Inventories
for
the
Heavy­
Duty
Engine
and
Vehicle
Standards
and
the
Highway
Diesel
Fuel
(
HDD)
Rulemaking
 
Detailed
Report"
(
EPA/
420­
R­
00­
019,
October
2000).

6
See
USEPA's
"
Regulatory
Impact
Analysis"
EPA420­
R­
04­
007,
May
2004
9
July
15,
2004
Nonroad
Sources:
NOx
emissions
from
nonroad
sources
are
about
six
percent
less
in
the
2010
CAIR
(
base)
inventory
compared
to
the
2007
HDD
inventory.
The
major
control
program
affecting
nonroad
sources
is
the
USEPA's
May
2004
Clean
Air
Nonroad
Diesel
Rule.
This
rule
requires
pollution
controls
on
diesel
engines
used
in
industries
such
as
construction,
agriculture
and
mining,
and
it
will
reduce
sulfur
content
of
diesel
fuel.
Standards
for
new
engines
will
be
phased­
in
starting
with
the
smallest
engines
in
2008,
until
all
but
the
very
largest
diesel
engines
meet
both
NOx
and
PM
standards
in
2014.
Some
of
the
largest
engines,
750+
horsepower,
will
have
one
additional
year
to
meet
the
emissions
standards.
Diesel
fuel
currently
contains
about
3,000
parts
per
million
(
ppm)
sulfur.
The
new
rule
will
cut
that
to
500
ppm
in
2007
and
15
ppm
by
2010.

This
information
suggests
that
a
significant
difference
in
NOx
emissions
is
expected
between
2007
and
2010.
As
such,
it
may
be
worthwhile
to
conduct
a
2007
sensitivity
analysis
by
adjusting
(
increasing)
the
NOx
emissions
as
follows:
EGUs
(
increase
by
ten
percent);
On­
Road
(
increase
by
fifteen
percent);
and
Non­
Road
(
increase
by
five
percent).

ADDITIONAL
PHOTOCHEMICAL
MODELING
Photochemical
Model
Selection
Several
one­
atmosphere
photochemical
models
treat
the
physical
processes
and
chemistry
that
form
ozone.
These
models
include
the
Community
Multiscale
Air
Quality
modeling
system
(
CMAQ)
and
the
Comprehensive
Air
Quality
Model
with
Extensions
(
CAMx4)
by
ENVIRON.
Fast
simulation
times
and
full
science
make
the
CAMx4
model
the
ideal
choice
for
modeling
grid
simulations
over
regional
domains
and
multiple
month
episodes.
The
summers
(
i.
e.,
June
through
August)
of
2001
and
2002
were
used
for
this
analysis
to
capture
the
variety
of
high
ozone
episodes
that
occurred
across
the
Upper
Midwest.

ENVIRON
developed
an
ozone
source
attribution
approach
that
has
become
known
as
the
"
Ozone
Source
Apportionment
Technology,"
or
OSAT
(
Yarwood
et
al.,
1996).
This
method
was
originally
implemented
in
the
urban
airshed
model
and
was
then
built
into
CAMx.
The
OSAT
provides
a
method
for
estimating
the
contributions
of
multiple
source
areas,
categories,
and
pollutant
types
to
ozone
formation
in
a
single
model
run.
The
OSAT
also
includes
a
methodology
for
diagnosing
the
temporal
relationships
between
ozone
and
emissions
from
groups
of
sources.

The
OSAT
allows
CAMx
to
track
source
region
and/
or
source
emissions
category
contributions
to
predicted
grid
cell
ozone
concentration;
thus,
for
any
selected
receptor
point
and
time,
the
model
gives
a
clear
picture
of
the
likely
distribution
of
ozone
and
ozone
precursors
by
source
emissions
category
and/
or
source
region,
as
well
as
an
indication
as
to
whether
the
ozone
at
the
selected
time
and
location
would
more
likely
respond
to
upwind
NOx
or
VOC
controls.

The
CAMx
(
version
4.03),
with
the
OSAT
technology,
was
used
in
this
analysis
to
determine
the
geographic
source
(
i.
e.,
Chicago)
contribution
of
ozone
precursors
and
source
emissions
type
(
i.
e.,
on­
road
mobile)
contribution
at
specific
locations
such
as
the
Cassopolis
and
Muskegon
monitors.
10
July
15,
2004
CAMx
OSAT
Results
for
Cassopolis
Monitor
(
Cass
County)
To
provide
high
credibility
the
OSAT
results,
only
modeled
results
from
days
where
the
CAMx
predicted
concentrations
within
20
percent
of
actual
and
the
Cassopolis
monitor
recorded
an
actual
concentration
of
85
ppb
or
greater
were
used.
This
screening
criteria
yielded
18
modeled
days
for
analysis
during
the
simulated
summers
of
2001
and
2002.
Results
from
these
18
days
were
weighted
by
actual
ozone
concentration
to
give
greater
value
to
the
higher
ozone
days.

Geographic
regions
were
broken
down
by
states
surrounding
Michigan
and
by
counties
within
Michigan.
Emission
types
were
broken
down
by:
biogenics,
on­
road
mobile,
non­
road
mobile,
low
level
point
sources,
elevated
point
sources,
and
area
sources.
Results
of
local
contribution
compared
to
transport
concentrations
from
surrounding
areas
are
as
follows:

Contribution
Percentage
to
Cassopolis
Monitor
by
Geographic
Area
and
by
Emissions
Type
Biogenics
On­
Road
Non­
Road
Low
Point
Elevated
Point
Area
TOTAL
Local
Contribution
Cass
&
St.
Joseph
Counties
0.05%
0.27%
0.15%
0.02%
0.02%
0.07%
0.58%
Out­
of­
State
Transport
Chicago
Area
0.05%
3.06%
2.22%
0.50%
2.51%
1.32%
9.66%
Illinois
(
excluding
Chicago)
0.51%
2.01%
1.75%
0.39%
3.42%
0.55%
8.63%
Indiana
0.80%
8.65%
5.52%
0.41%
6.52%
2.83%
24.73%
Ohio
0.03%
0.59%
0.42%
0.01%
0.52%
0.29%
1.86%
Wisconsin
0.11%
0.88%
0.57%
0.02%
0.43%
0.01%
0.38%

Leftover
contributions
are
from
the
following
areas;
4.04
percent
from
remaining
Michigan
counties;
17.86
percent
from
other
non­
listed
states;
and
32.26
percent
from
background.

As
shown,
less
than
one
percent
of
the
ozone
recorded
at
the
Cassopolis
monitor
can
be
attributed
to
local
(
i.
e.,
Cass
Co.
plus
St.
Joseph
Co.)
emissions.
With
a
design
value
of
93
ppb,
ALL
Michigan
emissions
contributions
to
the
Cassopolis
monitor
(
e.
g.,
4.62
percent
of
total)
could
be
eliminated
and
the
monitor
would
still
be
in
violation
of
the
8­
hour
standard.
The
modeling
evidence,
in
conjunction
with
common
sense
analysis,
demonstrates
overwhelming
out­
of­
state
transport.

CAMx
OSAT
Results
for
Muskegon
Monitor
(
Muskegon
County)
Similar
to
the
previous
analysis,
only
modeled
results
from
days
where
the
CAMx
predicted
concentrations
within
20
percent
of
actual
and
the
Muskegon
monitor
recorded
an
actual
concentration
of
85
ppb
or
greater
were
used.
This
screening
criteria
yielded
12
modeled
days
for
analysis
during
the
simulated
summers
of
2001
and
11
July
15,
2004
2002.
Results
from
these
12
days
were
weighted
by
actual
ozone
concentration
to
give
greater
value
to
the
higher
ozone
days.

Geographic
regions
were
broken
down
by
states
surrounding
Michigan
and
by
counties
within
Michigan.
Emission
types
were
broken
down
by:
biogenics,
on­
road
mobile,
nonroad
mobile,
low
level
point
sources,
elevated
point
sources,
and
area
sources.
Results
of
local
contribution
compared
to
transport
concentrations
from
surrounding
areas
are
as
follows:

Contribution
Percentage
to
Muskegon
Monitor
by
Geographic
Area
and
by
Emissions
Type
Biogenics
On­
Road
Non­
Road
Low
Point
Elevated
Point
Area
TOTAL
Local
Contribution
Muskegon
County
0.01%
0.30%
0.10%
0.01%
0.17%
0.06%
0.65%
Out­
of­
State
Transport
Chicago
Area
0.05%
5.40%
4.32%
0.82%
3.16%
2.73%
16.48%
Illinois
(
excluding
Chicago)
0.25%
1.80%
1.52%
0.33%
2.65%
0.58%
7.13%
Indiana
0.13%
3.84%
2.67%
0.19%
4.23%
1.78%
12.84%
Ohio
0.01%
0.89%
0.64%
0.02%
0.77%
0.46%
2.79%
Wisconsin
0.05%
1.18%
0.69%
0.02%
0.57%
0.60%
3.11%

Leftover
contributions
are
from
the
following
areas;
11.07
percent
from
remaining
Michigan
counties;
18.11
percent
from
non­
listed
states;
and
27.82
percent
from
background.

As
shown,
less
than
one
percent
of
the
ozone
recorded
at
the
Muskegon
monitor
can
be
attributed
to
local
(
i.
e.,
Muskegon
Co.)
emissions.
With
a
design
value
of
95
ppb,
ALL
Michigan
emissions
contributions
to
the
Muskegon
monitor
(
e.
g.,
11.72
percent
of
total)
could
be
eliminated
and
the
monitor
would
still
have
a
design
value
of
84
ppb,
barely
below
non­
attainment
threshold.
The
modeling
evidence,
in
conjunction
with
common
sense
analysis,
demonstrates
overwhelming
out­
of­
state
transport.

Grid
Projection
and
Domains
The
CAMx
model
was
applied
with
a
Lambert
projection
centered
at
(­
97,
40)
and
true
latitudes
at
33
and
45.
The
photochemical
modeling
domain
consists
of
97
cells
in
the
X
direction
and
90
cells
in
the
Y
direction
covering
the
Central
and
Eastern
United
States
with
36
km
grid
cells.
This
is
shown
in
the
figure
to
the
right
as
the
dark
yellow
box.
The
lighter
yellow
box
shows
the
MM5
modeling
domain.
The
emissions
modeling
domain
is
similar
to
the
MM5
domain,
but
with
9
fewer
cells
in
each
direction.
12
July
15,
2004
CAMx4
is
applied
with
the
vertical
atmosphere
resolved
with
16
layers
up
to
approximately
15
kilometers
above
ground
level.

Grid
Cell
Size
X,
Y
Origin
(
km)
X,
Y
Offset
(
from
MM5
grid)
NX,
NY
Meteorological
36
km
(­
2952.,
­
2304.)
N/
A
165,
129
Emissions
36
km
(­
2628.,­
1980.)
9,
9
147,
111
Photochemical
36
km
(­
900.,­
1620.)
57,
19
97,
90
Meteorological
Inputs
Meteorological
input
data
for
the
CAMx
photochemical
modeling
runs
were
processed
using
the
NCAR's
5th
generation
Mesoscale
Model
(
MM5)
version
3.5
(
Dudhia,
1993).
Key
model
parameters
and
options
used
in
MM5
are
shown
in
the
table
below.
A
more
detailed
discussion
of
MM5
applications
to
support
photochemical
modeling
may
be
found
in
the
MM5
Modeling
Protocol
at
www.
ladco.
org
(
Baker,
2004).
The
parameterizations
and
modules
selected
were
determined
to
be
an
optimal
model
configuration
for
the
Upper
Midwest
based
on
extensive
sensitivity
simulations
(
Johnson,
2003).

The
meteorological
fields
output
by
MM5
are
prepared
for
use
by
the
photochemical
model
with
processing
utilities.
These
programs
translate
certain
meteorological
parameters
from
the
MM5
grid
to
the
photochemical
grid.
Additionally,
these
processors
must
estimate
parameters
that
are
not
explicitly
output
by
MM5.
Since
the
meteorological
processing
programs
for
each
model
not
only
translate
data,
but
also
diagnose
certain
key
parameters,
this
step
must
be
scrutinized
to
achieve
optimal
model
results.

Emissions
Inputs
Emissions
data
was
processed
using
EMS­
2003.
The
EMS­
2003
model
is
selected
for
its
ability
to
efficiently
process
the
large
requirements
of
regional
and
daily
emissions
processing.
In
addition
to
extensive
quality
assurance
and
control
capabilities,
EMS­
2003
also
performs
basic
emissions
processes
such
as
chemical
speciation,
spatial
allocation,
temporal
allocation,
and
control
of
area,
point,
and
motor
vehicle
emissions
(
Janssen,
1998).
Outputs
from
EMS­
2003
include
a
coordinate­
based
elevated
point
source
file
and
gridded
emissions
estimates
for
low­
point,
area,
mobile,
and
biogenics
sources.
Anthropogenic
emission
estimates
are
made
for
a
weekday,
Saturday,
and
Sunday
for
each
month.
The
biogenic
emissions
are
day­
specific.
Volatile
organic
compounds
are
speciated
to
the
CB4
chemical
speciation
profile.
Configuration
36km
and
12km
Domains
Explicit
Moisture
Simple
ice
Cumulus
Kain­
Fritsch
PBL
Pleim­
Chang
(
ACM)
Radiation
RRTM
Multi­
Layer
Soil
Model
Pleim­
Xu
Shallow
convection
No
4­
D
Data
Assimilation
Analysis
nudging
on
above
PBL
Moist
Physics
Table
No
13
July
15,
2004
SPECIE
DESCRIPTION
ALD2
Aldehydes
ETH
Ethylene
FORM
Formaldehyde
ISOP
Isoprene
OLE
Olefins
­
Anthropogenic
OLE2
Olefins
­
Biogenic
(
OVOC)

PAR
Paraffins
TOL
Toluene
XYL
Xylene
NH3
Ammonia
CO
Carbon
monoxide
NO2
Nitrogen
dioxide
NO
Nitrogen
oxide
SULF
Sulfur
SO2
Sulfur
dioxide
PEC
Primary
PM­
fine
elemental
carbon
PNO3
Primary
PM­
fine
nitrate
POA
Primary
PM­
fine
organic
aerosol
PSO4
Primary
PM­
fine
sulfate
CCRS
Primary
PM­
coarse
crustal
FCRS
Primary
PM­
fine
crustal
CPRM
Primary
PM­
coarse
"
other"

FPRM
Primary
PM­
fine
"
other"
The
point
source
inventory
is
based
on
the
1999
National
Emission
Inventory
version
2.0.
Temporal
profiles
were
applied
to
all
CEM
units
located
in
the
Upper
Midwest
(
IL,
IN,
WI,
MI,
OH,
MN,
IA,
MO,
KY,
TN,
WV,
PA)
by
hour
of
the
day,
day
of
the
week,
and
month
of
the
year.
The
1995
Canadian
point
sources
are
included
in
the
elevated
point
source
inventory.
No
Mexican
point
source
emissions
are
included.

Area
sources
include
all
categories
that
are
not
included
in
the
point,
on­
road,
off­
road,
biogenic,
or
ammonia
inventory.
Categories
such
as
solvent
utilization
and
non­
point
fuel
combustion
are
included
in
the
area
inventory.
The
area
source
inventory
is
based
on
the
1999
National
Emission
Inventory.
A
90
percent
reduction
factor
to
all
dust
categories
was
applied
to
the
inventory
to
remove
the
non­
transportable
fraction
of
these
emissions.
This
area
inventory
also
includes
all
non­
point
emissions
from
the
Canadian
inventory,
which
includes
non­
road,
on­
road,
and
ammonia
estimates.

On­
road
emissions
are
estimated
with
the
MOBILE6
using
MM5
output
surface
temperature
and
15
m
relative
humidity.
The
default
temporal
tables
were
modified
to
represent
a
more
complex
distribution
of
vehicle
miles
traveled
for
weekend.
Off­
road
emissions
are
estimated
with
the
latest
release
of
USEPA's
NONROAD
2002
model.

The
biogenic
emissions
were
estimated
with
EMS­
2003
using
BIOME3/
BEIS3
and
the
BELD3
land
use
dataset.
Other
inputs
to
the
biogenic
emissions
model
include
hourly
satellite
photosynthetically
activated
radiation
(
PAR)
and
15
m
temperature
data
output
from
MM5.
The
15
m
temperature
data
was
selected
for
its
spatial
representation
of
the
tree
canopy
layer.
