May
2005
NETAC
EAC
TSD.
wpd
Page
1
of
46
Attachment
1
MODELING
TECHNICAL
SUPPORT
DOCUMENT
for
Clean
Air
Act
Approval
and
Promulgation
of
The
Northeast
Texas
Area
Early
Action
Compact
for
Ozone
Prepared
by
U.
S.
EPA
Region
6
Air
Modeling
Group
May
2005
Modeling
Technical
Support
Document
for
May
2005
NETAC
EAC
TSD.
wpd
Page
2
of
46
Clean
Air
Act
Approval
and
Promulgation
of
The
Northeast
Texas
area
Early
Action
Compact
for
Ozone
TABLE
OF
CONTENTS
1.
BACKGROUND
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
3
of
48
2.
MODELING
PROTOCOL
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
6
of
48
3.
CONCEPTUAL
MODEL
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
6
of
48
4.
EPISODE
SELECTION
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
7
of
48
5.
MODEL
SELECTION
&
SETUP
OPTIONS
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
9
of
48
6.
EMISSION
INVENTORY
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
16
of
48
1999
BASE
CASE
AND
2002
CURRENT
EIS
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
17
of
48
2007
and
2012
FUTURE
YEAR
EIs
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
26
of
48
7.
BASECASE/
CURRENT
MODELING
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
33
of
48
8.
FUTURE
YEAR
MODELING
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
43
of
48
9.
ATTAINMENT
DEMONSTRATION
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
45
of
48
10.
POST
ATTAINMENT
DEMONSTRATION
­
Future
Growth
Analysis
.
.
.
.
.
.
.
Page
46
of
48
1.
BACKGROUND
The
US
Environmental
Protection
Agency
(
EPA)
1­
hour
ozone
National
Ambient
Air
Quality
May
2005
NETAC
EAC
TSD.
wpd
Page
3
of
46
Standard
(
NAAQS)
has
a
threshold
of
0.12
ppm
(
124
ppb)
with
an
expected
exceedance
rate
of
no
more
than
once
per
year
over
three
consecutive
years
(
i.
e.,
with
complete
data
capture
compliance
with
the
1­
hour
ozone
NAAQS
requires
that
the
fourth
highest
daily
maximum
1­
hour
ozone
concentration
in
three
years
at
every
ozone
monitor
to
be
less
than
or
equal
to
0.12
ppm).
Areas
that
violate
the
1­
hour
ozone
NAAQS
are
classified
as
ozone
nonattainment
areas.
Ozone
nonattainment
areas
must
develop
an
ozone
emissions
control
plan
and
demonstrate
that
they
will
attain
the
ozone
NAAQS
by
the
date
specified
in
the
Clean
Air
Act
Amendments
(
CAAA)
in
a
State
Implementation
Plan
(
SIP).
The
SIP
ozone
attainment
demonstration
is
usually
accomplished
using
air
quality
modeling.
The
Northeast
Texas
area
is
currently
classified
as
being
in
attainment
with
the
1­
hour
ozone
standard.

In
1997,
EPA
promulgated
a
new
ozone
NAAQS
that
is
potentially
more
stringent
than
the
1­
hour
standard.
The
new
form
of
the
ozone
NAAQS
is
based
on
ozone
measurements
averaged
over
eight
hours;
a
violation
of
the
8­
hour
ozone
standard
occurs
when
the
average
of
the
fourth
highest
8­
hour
ozone
concentration
over
three
consecutive
years
exceeds
0.08
ppm
(
84
ppb).
The
8­
hour
ozone
nonattainment
are
designations
were
based
on
ambient
measurements
taken
during
the
three
years
between
2001­
2003.
Regions
that
are
currently
designated
as
nonattainment
of
the
1­
hour
ozone
NAAQS
must
still
attain
this
standard
(
i.
e.,
have
three
consecutive
years
over
which
the
fourth
highest
hourly
ozone
concentrations
at
all
monitors
are
124
ppb
or
less).
Once
an
ozone
nonattainment
region
attains
the
1­
hour
ozone
NAAQS,
the
1­
hour
standard
can
be
revoked
by
EPA
and
the
area
would
be
required
to
meet
only
the
8­
hour
standard.

The
Northeast
Texas
area
has
exceeded
the
8­
hour
ozone
standard
in
the
past.
Currently,
the
area
has
an
8­
hour
ozone
Design
Value
that
is
close
to
the
standard.
In
April
2004,
the
Northeast
Texas
Area
was
designated
as
attainment
of
the
8­
hour
ozone
standard
based
on
2001
 
2003
observed
ozone
data.
However,
the
areas
elected
to
stay
in
the
EAC
program
to
protect
against
being
declared
nonattainment
before
2007.

Early
Action
Compact
(
EAC)
Protocol
The
Texas
Natural
Resources
Conservation
Commission
(
now
Texas
Commission
on
Environmental
Quality,
TCEQ)
has
developed,
in
cooperation
with
the
US
EPA,
a
Protocol
for
Early
Action
Compacts
(
EACs).
The
TCEQ
EAC
was
finalized
in
March
2002.
The
basic
principals
of
the
EAC
are
for
local
air
quality
planners
to
commit
to
early
implementation
of
emission
controls
as
needed
to
achieve
the
8­
hour
ozone
standard
by
2007
in
return
for
which
EPA
will
defer
declaring
the
area
nonattainment
of
the
8­
hour
ozone
standard
until
2007.
In
order
for
an
area
to
be
allowed
to
opt­
in
to
an
8­
hour
ozone
EAC
they
must
currently
attain
the
1­
hour
ozone
standard.
If
an
area
opts­
in
to
an
8­
hour
ozone
EAC
then
they
must
meet
specific
milestone
deliverables
that
are
listed
in
Table
1­
1;
if
an
area
fails
to
meet
an
EAC
milestone
deliverable
or
attain
the
8­
hour
ozone
standard
in
2007,
they
May
2005
NETAC
EAC
TSD.
wpd
Page
4
of
46
revert
back
to
standard
8­
hour
ozone
nonattainment
and
must
meet
all
traditional
nonattainment
requirements.

In
December
2004,
the
State
of
Texas
submitted
to
the
EPA
a
revised
State
Implementation
Plan
(
SIPs)
regarding
the
Northeast
Texas
Area
Early
Action
Compacts.
The
State's
submission
includes:

(
1)
a
letter
from
Kathleen
Hartnett
White,
Chairman
TCEQ,
to
Richard
Greene,
Regional
Administrator
of
EPA
Region
6,
transmitting
the
revision;
and
(
2)
the
SIP
revision
from
Texas
Commission
on
Environmental
Quality
summarizing
the
Area's
photochemical
grid
modeling
results
and
any
adopted
control
measures.
The
State
of
Texas
supplemented
its
revisions
by
submitting
to
the
EPA,
technical
documents
on
the
attainment
demonstration
modeling,
which
contained
the
following:
the
data
and
results
from
all
the
episodes
modeled
using
the
MM5
and
CAMx
models,
control
strategy
modeling
analyses,
future
design
value
calculations,
etc.
Any
emission
controls
necessary
will
be
implemented
by
2005.

The
Northeast
Texas
Area
has
elected
to
opt­
in
to
the
8­
hr
ozone
EAC.
Thus,
they
are
required
to
develop
emissions
and
photochemical
modeling
databases
needed
to
prepare
an
8­
hr
ozone
attainment
plan
to
be
included
in
a
State
implementation
Plan
(
SIP)
that
was
submitted
to
EPA
in
December
2004.
The
key
objectives
in
developing
an
all
new
photochemical
modeling
database
for
Northeast
Texas
Area
are:
°
To
develop
a
Modeling
Protocol
and
a
Conceptual
Model
for
ozone
exceedances
and
then
select
an
8­
hour
ozone
modeling
episode(
s)
for
the
Northeast
Texas
area;

°
To
create
a
modeling
domain
with
a
coarse
grid
domain
extent
sufficiently
large
to
treat
multi­
day
transport
of
ozone
and
precursors
from
significant
source
areas
outside
of
the
Northeast
Texas
area
and
select
model
setup
options;

°
Generate
Emission
Inventories
(
basecase,
current,
and
future
year)
to
be
used
in
the
modeling
May
2005
NETAC
EAC
TSD.
wpd
Page
5
of
46
exercise;

°
To
create
a
Photochemical
modeling
tool
by
simulating
an
episode
(
Base
Case
simulation),
a
Current
year
simulation
if
appropriate,
and
a
future
year
modeling
analysis
that
incorporates
adopted
federal
and
state
emission
reduction
strategies
that
show
attainment
in
2007.

°
Conduct
an
analysis
of
the
impact
of
Future
Growth
by
evaluating
emission
levels
in
2012
and
their
potential
impact
on
ozone
levels.

The
subsequent
sections
of
this
document
provide
the
EPA's
analysis
of
the
EAC
modeling.
This
document
should
be
used
in
conjunction
with
the
State's
SIP
submittal
as
certain
modeling
requirements
performed
by
the
State
(
i.
e.,
details
of
the
quality
assurance
performed,
detailed
analysis
of
data
suitability,
complete
listings
of
all
data
inputs
and
outputs,
etc.)
are
not
necessarily
reproduced
in
this
TSD.

EPA
MODELING
GUIDANCE
In
1999,
in
anticipation
of
the
implementation
of
a
new
8­
hour
ozone
standard,
EPA
released
"
Draft
Guidance
on
the
Use
of
Models
and
Other
Analyses
in
Attainment
Demonstrations
for
the
8­
Hour
Ozone
NAAQS."
This
guidance
provides
an
entirely
new
approach
to
using
photochemical
models
for
attainment
demonstration
purposes.
Specifically,
this
approach
to
demonstrating
attainment
relies
on
the
use
of
modeling
results
in
the
relative
sense,
rather
than
the
absolute
sense.
The
approach
also
incorporates
information
on
observed
ozone
concentrations
(
in
the
form
of
a
design
value).
A
relative
reduction
factor
(
defined
as
the
ratio
of
the
future­
year
to
base­
year
maximum
simulated
ozone
concentration
for
an
area
around
a
monitor)
is
calculated
for
each
monitoring
site
and
multiplied
by
the
observation­
based
design
value
to
obtain
an
estimated
future
design
value.
This
estimated
design
value
is
then
compared
with
the
federal
ozone
standard
to
assess
whether
attainment
is
expected
by
the
future
year
predicted
in
the
modeling.
The
relative­
reduction
approach
is
applied
for
each
monitoring
site
(
an
attainment
test)
and
also
for
other
areas
in
the
domain
that
are
consistently
characterized
by
high
ozone
in
the
modeling
results
(
a
screening
test).

The
relative­
reduction
approach
was
designed
for
use
in
conjunction
with
a
new
8­
hour
ozone
standard.
It
can
also
be
applied
using
modeling
results
to
the
examination
of
the
1­
hour
ozone
standard
as
specified
in
the
draft
guidance
or
as
adapted
according
to
"
Guidance
for
Improving
Weight
of
Evidence
Through
Identification
of
Additional
Emission
Reductions,
Not
Modeled"
(
EPA,
1999b).
The
relative­
reduction
approach
also
accommodates
the
use
of
other
analysis
results
or
corroborative
evidence
to
support
a
finding
of
attainment.

In
addition,
unlike
the
previous
guidance
for
1­
hr
ozone
modeling
(
EPA,
1991),
EPA
now
recommends
that
models
be
selected
on
a
`
case­
by­
case'
basis
with
appropriate
consideration
being
given
to
the
candidate
model's:
(
a)
technical
formulation,
capabilities
and
features,
(
b)
pertinent
peer­
May
2005
NETAC
EAC
TSD.
wpd
Page
6
of
46
review
and
performance
evaluation
history,
(
c)
public
availability,
and
(
d)
demonstrated
success
in
similar
regulatory
applications.
All
of
these
considerations
should
be
examined
for
each
class
of
models
to
be
used
(
e.
g.,
emissions,
meteorological,
and
photochemical)
in
part
because
EPA
no
longer
recommends
a
specific
model
or
suite
of
photochemical
models
for
regulatory
application.

2.
MODELING
PROTOCOL
The
area
developed
a
modeling
protocol
in
January
2001
for
the
Flexible
Attainment
Region
(
FAR)
SIP
and
updated
this
modeling
protocol
in
October
2003.
The
original
and
updated
protocols
included
a
discussion
of
the
conceptual
model
and
how
the
episodes
fit
the
conceptual
model;
original
episode
selection;
models
selection
and
their
proposed
settings;
and
methodologies
on
conducting
the
meteorological,
emissions,
and
photochemical
modeling
analyses.
EPA
reviewed
the
proposed
protocol
and
provided
comments
to
the
areas/
State
and
the
contractor.

3.
CONCEPTUAL
MODEL
The
area
and
their
contractor
developed
an
analysis
of
1995­
1999
ozone
episodes
in
a
conceptual
model
description
in
September
2000.
This
original
analysis
indicated
that
stagnant
conditions
were
the
most
common
meteorological
pattern
during
ozone
exceedances.
The
conceptual
model
was
updated
and
finalized
in
January
2004.
The
updated
analysis
included
looking
at
ozone
episodes
from
2000
­
2003,
additional
species(
volatile
organic
species
and
SO2)
ground
monitoring
at
the
Longview
monitor,
several
aircraft
flights
in
East
Texas,
look
at
long­
term
ozone
levels
and
temperature.
This
work
was
done
in
accordance
with
EPA's
1­
hr
and
8­
hr
ozone
modeling
guidance
and
guidance
given
by
EPA
Region
6
personnel.
This
analysis
included
analysis
also
included
evaluation
of
the
surface
winds
and
conducting
HYSPLIT
runs.
The
updated
conceptual
model
analysis
further
confirmed
that
calm
wind
conditions
(
stagnant
air
masses)
with
high
pressure
aloft
were
the
most
common
meteorological
conditions
when
high
ozone
occurred.
The
updated
conceptual
model
also
indicated
that
elevated
levels
of
SO2
were
also
common
with
the
high
ozone
indicating
the
combustion
of
fossil
fuel
was
partially
responsible
for
the
elevated
ozone
levels.
The
area
is
dominated
by
VOCs
and
mostly
NOx
limited.
See
Figure
3­
1
for
the
annual
4th
high
8­
hour
ozone
levels
from
1995
to
2003.
May
2005
NETAC
EAC
TSD.
wpd
Page
7
of
46
Figure
3­
1.
Annual
fourth
highest
daily
maximum
8­
hour
ozone
at
Northeast
Texas
and
northeastern
Louisiana
monitoring
sites:
1995
 
2003.

4.
EPISODE
SELECTION
The
Northeast
Texas
area
followed
the
procedures
for
selecting
8­
hr
ozone
modeling
episodes,
outlined
in
the
EPA's
draft
guidance
(
EPA,
1999).
This
procedure,
in
part,
considers
selecting
episode
days
with
observed
8­
hour
ozone
concentrations
close
to
each
current
monitors'
Design
Value
and
consistent
with
the
form
of
the
NAAQS
(
i.
e.
the
ozone
levels
that
lead
to
nonattainment
designation);
representing
the
range
of
meteorological
conditions
that
accompany
exceedances
of
the
8­
hour
ozone
standard;
selecting
periods
for
which
adequate
emissions,
air
quality
and
meteorological
data
are
available
for
model
testing
and
application;
and
accounting
for
the
frequency
of
occurrence
of
the
relevant
aerometric
conditions,
appropriately
excluding
rare
or
extreme
events.
A
search
through
1997
to
1999
data
identified
four
candidate
episodes:

1.
August
26
to
September
4,
1998
May
2005
NETAC
EAC
TSD.
wpd
Page
8
of
46
2.
August
2
to
August
7,
1999
3.
August
15
to
August
22,
1999
4.
September
15
to
September
20,
1999
The
modeling
episode
period
for
Northeast
Texas
was
selected
for
the
August
15
­
August
22,
1999
ozone
episode.
The
modeling
period
was
expanded
to
August
13
­
August
22,
1999
to
include
two
spin­
up
days
before
the
start
of
the
episode
to
reduce
the
influence
in
the
modeling
of
initial
conditions.
This
period
included
combined
influences
from
a
high
regional
ozone
background
and
local
emissions,
and
included
a
complete
cycle
of
transport
winds
followed
by
local
stagnation
returning
to
transport
winds
at
the
end
of
the
episode.

High
ozone
levels
were
recorded
at
three
Continuous
Air
Monitoring
Stations
(
CAMS)
during
the
period
(
see
Table
4­
1).
There
was
a
build
up
between
August
15
and
August
17
with
the
onset
of
meteorological
stagnation
on
August
16
continuing
through
August
18.
On
August
18
and
August
19
the
ozone
levels
were
similarly
high
at
all
three
sites
consistent
with
a
high
regional
background
of
ozone.

TABLE
4­
1.
Maximum
observed
8
­
hour
ozone
concentration
(
ppb)
at
sites
in
the
Northeast
Texas
Region
on
the
modeling
episode
days.

Date
Longview
Maximum
Temperature
(
oF)
Maximum
8­
hour
Ozone
(
ppb)

Longview
CAMS
19
Tyler
CAMS
82
Cypress
River
CAMS
50
8/
15/
99
93
66
73
55
8/
16/
99
95
105
92
71
8/
17/
99
96
110
97
90
8/
18/
99
99
88
74
91
8/
19/
99
102
91
85
81
8/
20/
99
97
80
86
70
8/
21/
99
95
87
92
67
8/
22/
99
96
91
77
82
May
2005
NETAC
EAC
TSD.
wpd
Page
9
of
46
5.
MODEL
SELECTION
&
SETUP
OPTIONS
The
NETAC
area
selected
the
meteorology
and
photochemical
models
and
the
modeling
domains
and
grid
specifications
of
the
models
for
the
8­
hour
Early
Action
Compact
Study
in
accordance
to
the
EPA
draft
8­
hour
ozone
modeling
guidance.
The
NETAC
used
the
Comprehensive
Air
Quality
Model
with
Extensions
(
CAMx,
Version
4.0)
photochemical
grid
model,
the
Emissions
Processing
System
(
EPS2x)
along
with
meteorological
modeling
inputs
generated
from
PSU/
NCAR's
MM5
Prognostic
Meteorological
model
for
the
Northeast
Texas
EAC
study.

The
NETAC's
modeling
activities
were
performed
as
outlined
in
the
modeling
protocols,
according
to
EPA's
"
Guideline
for
Regulatory
Application
of
the
Urban
Airshed
Model"
(
Guideline).
The
protocol
document
provides
information
regarding
the
organizational
structure
of
the
modeling
study,
study
participants,
communication
structures,
and
the
resolution
of
technical
difficulties.
The
protocol
also
provides
detailed
information
on
each
element
of
the
modeling
analysis
including
selection
of
the
primary
modeling
tools,
methods
and
results
of
the
episode
selection
analysis,
modeling
domain,
model
input
preparation
procedures,
model
performance
evaluation,
use
of
diagnostic
and
sensitivity
analysis,
future­
year
modeling,
application
of
the
EPA
ozone
attainment
demonstration
procedures,
and
documentation
procedures.
The
protocol
was
reviewed
by
EPA.
Through
this
model
development
process,
the
EPA
became
was
involved
with
the
modeling
activities
of
the
area,
and
has
lent
considerable
technical
support
to
the
area's
effort.

The
NETAC
selected
the
modeling
domains
and
grid
specifications
of
the
models
for
the
Northeast
Texas
8­
hour
Early
Action
Compact
Study
in
accordance
to
the
EPA
draft
8­
hour
ozone
modeling
guidance.
The
modeling
domain
for
application
of
the
CAMx
was
designed
to
accommodate
both
regional
and
subregional
influences
as
well
as
to
provide
a
detailed
representation
of
the
emissions,
meteorological
fields,
and
ozone
(
and
precursor)
concentration
patterns
over
the
area
of
interest.
The
modeling
domains
were
defined
on
a
MM5
Lambert
Conformal
Projection
system
as
presented
in
Figure
5­
1.
The
ozone
modeling
uses
nested
36
km,
12
km
and
4
km
grids.
The
36
km
grid
extends
as
far
as
the
Midwest
to
account
for
2­
3
days
of
potential
regional
transport.
The
12
km
grid
includes
all
of
the
areas
in
eastern
Texas
that
are
conducting
ozone
modeling
so
that
a
consistent
12
km
grid
can
be
used
in
all
studies.
In
addition,
the
12
km
grid
includes
areas
that
would
be
upwind
of
Texas
during
an
ozone
episode
with
easterly
or
northeasterly
winds.
The
intention
is
to
accurately
model
potential
transport
of
ozone
from
areas
at
a
distance
upwind
of
about
one
State.
The
4
km
grid
covers
Northeast
Texas
and
immediately
adjacent
major
urban
areas
and
major
sources.
The
vertical
extent
of
the
CAMx
domain
used
in
the
Northeast
Texas
study
is
defined
by
fifteen
(
15)
vertical
layers
that
was
identical
to
the
MM5
layer
definitions
to
minimize
any
distortion
of
the
meteorological
variables
(
see
Table
5­
1).
May
2005
NETAC
EAC
TSD.
wpd
Page
10
of
46
Figure
5­
1.
Northeast
Texas
Early
Action
Compact
8­
Hour
Ozone
Study
Modeling
Domain
May
2005
NETAC
EAC
TSD.
wpd
Page
11
of
46
Table
5­
1.
Comparison
of
CAMx
and
MM5
Vertical
Grid
Structures.

The
CAMx
photochemical
model
requires
hourly,
gridded
input
fields
of
wind,
temperature,
watervapor
concentration,
pressure,
vertical
exchange
coefficients
(
Kv),
cloud
cover,
and
rainfall
rate.
These
meteorological
inputs
were
prepared
for
the
applications
using
the
Pennsylvania
State
University/
National
Center
for
Atmospheric
Research
(
PSU/
NCAR)
Mesoscale
Model
(
MM5
v5).
May
2005
NETAC
EAC
TSD.
wpd
Page
12
of
46
MM5
was
conducted
utilizing
its
Four
Dimensional
Data
Assimilation
(
FDDA)
capabilities
with
gridded
meteorological
analyses
derived
from
the
Eta
Data
Assimilation
System
(
EDAS),
which
is
archived
at
the
National
Center
for
Atmospheric
Research
(
NCAR).
The
EDAS
provides
3­
hourly
gridded
meteorological
fields
developed
from
the
initialization
cycle
runs
of
the
National
Weather
Service's
Eta
operational
forecast
model,
which
ingests
observations
from
a
combination
of
several
systems
(
routine
measurements
from
surface
and
upper
air
sites,
radar
networks,
and
satellite
profilers).
The
EDAS
domain
covers
most
of
the
North
American
continent
on
a
Lambert
Conformal
grid
with
40­
km
grid
spacing,
and
extends
vertically
from
the
surface
to
50
mb
(~
20­
km)
with
more
than
20
pressure
levels
of
data.
The
EDAS
data
were
supplemented
with
specialized
data
in
the
southcentral
U.
S.
to
maximize
performance
in
the
areas
of
interest.
These
supplemental
data
included
a
wind
profiler
network
(
operated
by
the
Forecast
Systems
Laboratory
of
NOAA),
the
EPA
AIRS,
and
observations
from
the
Big
Bend
Regional
Aerosol
and
Visibility
Observation
Study
(
BRAVO),
which
was
operated
in
Texas
between
July
and
October
1999.

The
locations
of
the
NWS
upper
air
meteorological
sounding
data
available
from
NCAR
archives
are
presented
in
Figure
5­
2.
Meanwhile,
measurement
data
include
hourly
surface
and
aloft
wind
speed,
wind
direction,
temperature,
moisture,
and
pressure
were
obtained
from
many
Class
I
airports,
i.
e.,
larger­
volume
civil
and
military
airports
operating
24­
hour
per
day.

The
NETAC
also
used
the
National
Center
for
Atmospheric
Research
(
NCAR)
terrain
databases
for
topographic
information.
For
instance,
for
the
36
km
and
12km
grids,
the
5
min
topographic
information
derived
from
the
Geophysical
Data
Center
global
data
set
was
used
while
the
30
second
resolution
data
set
was
used
for
the
4
km
grid.
Vegetation
type
and
land
use
information
was
based
on
the
NCAR/
PSU
10
min.
(~
18.5
km)
databases
for
the
36
km
grid
and
from
the
United
States
Geological
Survey
(
USGS)
data
for
the
12
km
and
4
km
grids.
May
2005
NETAC
EAC
TSD.
wpd
Page
13
of
46
Figure
5­
2.
Upper
Air
Sounding
Sites
Throughout
the
U.
S.
Used
in
MM5
Modeling
for
The
Northeast
Texas
EAC
8­
Hour
Ozone
Study.

MODEL
PHYSICS
CONFIGURATION
AND
OTHER
PARAMETERS
The
MM5
model
provides
a
wealth
of
options
to
configure
the
model
for
various
parameterizations
and
physics
packages.
A
number
of
combinations
of
MM5
model
physics
configurations
and
FDDA
May
2005
NETAC
EAC
TSD.
wpd
Page
14
of
46
nudging
were
tested
before
finalizing
on
the
following
parameters
for
the
NETAC
episode
application
which
are
summarized
as
follows:
°
Simple­
ice
microphysics
was
employed
for
all
domains
°
Kain­
Fritsch
cumulus
parameterization
scheme
was
invoked
for
108/
36/
12­
km
grids.
No
cumulus
parameterization
scheme
was
invoked
for
the
4­
km
grid,
as
convection
is
explicitly
fully
resolved
at
this
resolution
scale.
°
The
RRTM
radiation
scheme
was
used
for
all
the
grids.
°
Two­
way
interactive
108/
36/
12/
4­
km
grids
were
used.
°
The
Pleim­
Xiu
LSM
coupled
with
its
own
Planetary
Boundary
Layer
(
PBL)
scheme
was
employed
in
this
work.
The
soil
moisture
nudging
techniques
in
this
scheme
was
also
used
and
the
soil
moisture
was
initialized
by
the
EDAS
data
through
REGRID
program.
°
FDDA
analysis
nudging
on
the
108/
36/
12­
km
grids:
o
3D
analysis
nudging
above
the
boundary
layer
­­
MM5
was
nudged
toward
3­
hourly
EDAS
analysis
of
wind,
temperature,
and
humidity,
which
are
improved
by
the
surface
and
upper­
air
station
observation
data.
o
Surface
analysis
nudging
within
the
boundary
layer
­­
MM5
was
nudged
toward
3­
hourly
gridded
surface
analysis
data
generated
by
RAWINS
program.
In
addition,
the
soil
moisture
nudging
was
also
employed.
°
FDDA
observation
nudging
of
wind,
temperature
and
moisture
on
the
12­
km
and
4­
km
grids
from
routine
and
special
measurement
data
set
available
from
NCAR
that
includes
data
from
NOAA
profiler,
NWS
Surface
and
Upper
Air.

BOUNDARY
CONDITIONS
In
addition
to
the
meteorological
inputs,
the
CAMx
modeling
system
requires
a
number
of
additional
input
files
that
contain
information
on
pollutant
concentrations
at
the
initial
simulation
time
and
along
the
boundaries
of
the
modeling
domain,
land
use,
albedo,
ozone
column,
photolysis
rates,
and
chemical
reaction
rates.
There
are
three
CAMx
air
quality
input
files
that
define
pollutant
concentrations
for
each
of
the
CAMx
state
species
(
1)
throughout
the
three­
dimensional
grid
at
the
initial
simulation
time
(
coarse­
grid
only),
(
2)
along
the
lateral
boundaries
of
the
modeling
domain
for
each
hour
of
the
simulation
period,
and
(
3)
along
the
top
of
the
modeling
domain
for
the
entire
simulation
period.
The
initial
conditions
(
ICs)
(
Item
1
above)
are
the
pollutant
concentrations
specified
throughout
the
modeling
domain
at
the
start
of
the
simulation.
Boundary
conditions
(
BCs)
(
Items
2
and
3
above)
are
the
pollutant
concentrations
specified
at
the
perimeter
of
the
modeling
domain
throughout
the
simulation.
The
boundary
condition
assumptions
are
discussed
because
they
played
a
role
in
achieving
good
ozone
model
performance.
The
boundary
conditions
are
shown
in
Table
5­
2.
The
ozone
BC
was
set
to
40
ppb,
which
is
the
value
commonly
considered
to
be
the
continental
background
and
used
for
ozone
modeling
studies.
The
NOx
BC
was
set
to
1.1
ppb.
The
VOC
BCs
varied
by
boundary
segment
over
a
range
from
9
to
50
ppbC
according
to
broad
differences
in
land
cover.
The
higher
VOC
BCs
in
the
Northeast/
East
boundary
segment
are
for
areas
with
higher
biogenic
emissions.
The
lower
VOC
BCs
along
the
West
boundary
segment
are
for
dryer
areas
with
lower
biogenic
emissions.
The
lowest
VOC
BCs
are
over
the
Gulf
of
Mexico
and
these
low
values
were
also
used
for
all
May
2005
NETAC
EAC
TSD.
wpd
Page
15
of
46
boundaries
above
an
altitude
of
1700
m.
The
initial
conditions
throughout
the
modeling
domain
were
set
to
the
lowest
(
Gulf
of
Mexico)
BC
values.

Figure
5­
3.
CAMx
36
km
regional
modeling
domain
showing
boundary
segments
that
are
assigned
different
boundary
conditions
(
BCs).

Table
5­
2.
Boundary
concentrations
for
different
boundary
segments
shown
in
Figure
5­
3.
May
2005
NETAC
EAC
TSD.
wpd
Page
16
of
46
CAMX
OPTIONS
There
are
four
model
options
that
must
be
chosen
for
each
project:
the
choice
of
advection
scheme,
the
plume­
in­
grid
scheme,
the
chemical
mechanism
and
the
chemistry
solver.
CAMx
includes
a
Plumein
Grid
module
that
tracks
emissions
from
large
point
sources
until
the
plume
size
is
commensurate
to
the
grid
cell
size
at
which
the
emissions
are
released
to
the
grid.
Plume­
in­
grid
treatment
was
used
for
some
NOx
point
sources.
Within
the
4­
km
modeling
domain
2
tons
NOx
on
any
episode
day
wass
the
criteria
for
selection.
For
the
extended
regional
emissions
grid,
the
NOx
criteria
was
25
tons
per
day
on
any
episode
day.
The
other
options
used
in
the
final
modeling
included
the
PPM
advection
scheme,
Carbon
Bond
4
(
CB4)
chemistry
mechanism,
and
the
CMC
chemical
solver.

6.
EMISSION
INVENTORY
The
NETAC
developed
three
major
versions
of
modeling
emission
inventories,
a
1999
version
representing
the
actual
emissions
that
occurred
during
the
specific
episodes,
a
2002
version
representing
the
best
available
emission
estimates
for
the
mid­
year
of
the
2001­
2003
DV
that
attainment
designation
was
based,
and
a
2007
version
representing
the
projected
emissions
expected
to
occur
in
the
year
2007.
The
episode
specific
modeling
emissions,
termed
the
"
base
case",
were
used
to
evaluate
the
MM5
and
CAMx
modeling
systems
reliability
in
replicating
the
high
ozone
concentrations
that
occurred
during
the
episode.
The
area
had
a
significant
drop
in
ozone
DV
for
the
2001­
2003
period
compared
to
the
1998­
2000
DV.
Several
power
plants
made
sizable
emission
reductions
that
could
have
impacted
the
2001­
2003
DV.
The
2001­
2003
period
was
also
a
period
less
atmospheric
stability
locally
and
more
tropical
storm
impacts
and
lower
high
temperatures
compared
to
the
1998­
2000
DV
period.
The
area
created
a
2002
EI,
termed
the
"
Current
Year"
EI,
to
take
into
consideration
the
lower
DV
and
the
impacts
of
NOx
reductions
at
many
local
power
plants.
The
2007
projected
modeling
emissions,
termed
the
"
future
year",
were
used
to
estimate
the
overall
level
of
reductions
in
VOC
and
NOx
needed
to
maintain
attainment
of
the
NAAQS.

The
NETAC
used
the
EPS2x
emissions
processing
system
for
developing
base
case,
current
case,
and
future
case
for
this
submittal.
This
system
consists
of
series
of
computer
programs
designed
to
perform
the
intensive
data
manipulation
necessary
to
adapt
a
county­
level
annual
or
seasonal
emission
inventory
for
modeling
use.
The
EPS2x
system
consists
of
a
series
of
modules
that
incorporate
spatial,
temporal,
and
chemical
resolution
into
an
emission
inventory
used
for
photochemical
modeling.
Point,
area,
non­
road
and
on­
road
mobile
source
emissions
data
were
processed
separately
through
the
EPS2x
system
to
facilitate
both
data
tracking
for
quality
control
and
the
use
of
data
in
evaluating
the
effects
of
alternative
proposed
control
strategies
on
predicted
future
air
pollutant
concentrations.
EPS2x
was
used
for
generating
emissions
input
files
for
the
CAMx
model.
Inventories
were
developed
for
1999,
2002,
and
2007
in
accordance
with
the
modeling
protocol.

1999
BASE
CASE
AND
2002
CURRENT
EIS
May
2005
NETAC
EAC
TSD.
wpd
Page
17
of
46
Point
Sources
The
modeling
inventories
prepared
for
the
modeling
episodes
were
based
on
the
following
information:


Version
2
of
the
1999
National
Emission
Inventory
(
NEI99
v2)


Point
source
and
county
level
emissions
data
for
the
State
of
Texas
provided
to
NETAC
from
TCEQ

Wildfires
emissions
obtained
from
WRAP
Fire
Forum

Un­
permitted
oil
and
gas
production
wells
emissions
provided
by
the
Texas
Oil
and
Gas
Association
Most
of
the
point
source
data
were
developed
for
a
typical
peak
ozone
(
PO)
season
weekday
and
weekend
day.
The
exception
was
for
Texas
EGUs,
which
are
hourly
episode
day
specific
data,
based
on
continuous
emissions
monitor
(
CEM)
data
that
were
reported
to
EPA's
"
Acid
Rain"
database.
For
all
states
in
the
12
km
grid,
other
than
Texas,
the
National
Emission
Inventory
(
NEI)
1999
Version
2
for
Criteria
Pollutants
data
was
used.
The
Texas
Commission
on
Environmental
Quality
(
TCEQ)
Point
Source
Data
Base
(
PSDB)
version
15a
for
1999
is
the
basis
of
the
non­
EGU
Texas
data
which
was
provided
by
TCEQ.
For
all
remaining
states
in
the
final
extended
regional
36
km
grid
the
1999
NEI
Final
Version
2
for
Criteria
Pollutants
data
was
used.

The
2002
base
year
emission
inventory
for
the
Northeast
Texas
EAC
study
was
based
on
the
EPA's
1999
National
Emission
Inventory
(
NEI99).
These
1999
emission
estimates
were
projected
to
the
2002
base
year
using
the
EPS2x
growth
and
projection
modules
with
growth
factors
developed
with
the
EGAS
model.
However,
emissions
from
stationary
point
sources
in
the
Northeast
Texas
region
were
provided
by
NETAC.
The
remaining
areas
within
the
State
of
Texas
were
based
on
stationary
point
source
emissions
data
from
the
NEI99.
Meanwhile,
the
Texas
Oil
and
Gas
Association
provided
emission
estimates
for
un­
permitted
oil
and
gas
production
wells
in
the
northeast
region
of
the
state.
All
the
point
source
data
were
processed
for
a
typical
peak
ozone
season
weekday
and
weekend
day.

On­
Road
Mobile
Meanwhile,
on­
road
mobile
emission
sources
were
processed
for
the
State
of
Texas.
These
link­
based
emissions
data
were
derived
from
the
U.
S
EPA's
MOBILE6,
which
was
used
to
developed
on­
road
emission
factors,
along
with
link­
based
activity
data
developed
for
the
Longview
and
Tyler
areas.
The
Texas
Transportation
Institute
(
TTI)
prepared
mobile
source
emissions
for
all
Texas
counties
under
contract
to
the
TCEQ.
Emission
factors
are
from
the
EPA's
MOBILE6
model.
Vehicle
miles
traveled
(
VMT)
for
1999
are
based
on
transportation
models
in
all
NNA
counties
that
have
a
complete
transportation
model
and
were
based
on
a
rural
HPMS
method
elsewhere.
All
other
on­
road
mobile
emission
estimates
were
based
on
the
NEI99
database.
These
data
were
May
2005
NETAC
EAC
TSD.
wpd
Page
18
of
46
again
projected
to
the
2002
base
year
using
EPS2x
with
growth
factor
calculated
using
the
EGAS
model.

Area
and
Off­
Road
For
Texas,
the
NEI99
is
the
basis
for
the
area
and
off­
road
mobile
regional
emissions
inventory.
The
NEI
1999
area
and
off­
road
emission
inventory
were
processed
to
extract
the
typical
peak
ozone
season
day
data
and
processed
with
EPS2x.
These
data
were
again
processed
using
EPS2x
and
spatially
allocated
to
grid
cells
using
gridding
surrogates
based
on
the
EPA's
gridding
surrogate
database.

BIOGENICS
Biogenic
emissions
were
calculated
for
the
36­
km,
12­
km,
and
4­
km
modeling
grids
using
GloBEIS
3.1
for
the
Northeast
Texas
EAC
modeling
study.
These
emissions
were
calculated
for
each
episode
day
for
each
of
the
grids.
GloBEIS3
requires
domain
definition,
land
use,
temperature,
photosynthetically
active
radiation
(
PAR),
wind
speed,
and
humidity
input
files.
Biogenic
emissions
can
vary
greatly
from
day
to
day
and
from
area
to
area
depending
on
ambient
conditions.
To
account
for
these
expected
variations
for
the
Northeast
Texas
application,
hourly
surface
temperature,
wind
speed
and
humidity
estimates
were
extracted
from
the
output
of
the
MM5
meteorological
modeling.
Meanwhile,
hourly
solar
radiation
based
on
GOES
satellite
data
as
analyzed
by
the
University
of
Maryland.
The
drought
index
input
file
was
generated
from
Palmer
Drought
Index
(
PDI)
data
obtained
from
the
Climate
Prediction
Center.
Biogenics
for
2002
were
kept
the
same
as
1999
estimates.

International
and
Offshore
Emissions
Offshore
emissions
for
the
Gulf
of
Mexico
were
obtained
from
the
TCEQ
and
are
the
same
emissions
1­
hour
ozone
for
Houston/
Galveston
and
Dallas­
Fort
Worth.
2002
emissions
were
generated
based
on
1999
estimates
and
growth
projections.

Summary
Tables
6­
1
and
6­
2
are
summaries
of
1999
Base
Case
NOx
and
VOC
emissions
respectively,
by
county
for
each
category
of
emissions
for
the
five
county
NETAC
area.
Table
6­
3
and
6­
4
are
summaries
of
2002
"
Current"
EI
NOx
and
VOC
emissions
respectively
for
the
five
county
NETAC
area
and
the
two
county
Shreveport
area..
May
2005
NETAC
EAC
TSD.
wpd
Page
19
of
46
Table
6­
1.
1999
episode
average
daily
NOx
emissions
(
tons/
day)
for
the
NETAC
area.

Source
Category
Gregg
County
Harrison
County
Rusk
County
Smith
County
Upshur
County
Area
12.0
7.8
12.9
5.0
8.3
Off­
road
4.4
6.0
2.5
6.5
2.4
On­
road
22.9
17.7
4.1
25.5
2.7
Point
Source
14.7
45.5
79.9
3.6
1.0
Subtotal
54.0
77.0
99.4
40.6
14.4
Biogenics
0.2
0.5
0.5
0.7
0.4
Total
54.2
77.5
99.9
41.3
14.9
Table
6­
2.
1999
episode
average
daily
VOC
emissions
(
tons/
day)
for
the
NETAC
area.

Source
Category
Gregg
County
Harrison
County
Rusk
County
Smith
County
Upshur
County
Area
13.6
12.7
11.3
13.0
13.1
Off­
road
2.5
1.5
1.1
4.2
0.5
On­
road
6.5
5.5
3.2
10.5
2.1
Point
Source
3.4
15.3
2.0
8.5
0.8
Subtotal
26.0
34.9
17.5
36.2
16.5
Biogenics
65.0
316.8
271.8
253.9
157.1
Total
91.0
351.7
289.4
290.1
173.5
May
2005
NETAC
EAC
TSD.
wpd
Page
20
of
46
Table
6­
3.
2002
NETAC
and
two
Shreveport
EAC
counties
NOx
emission
summaries
for
each
day
of
the
episode
for
each
source
category.
May
2005
NETAC
EAC
TSD.
wpd
Page
21
of
46
May
2005
NETAC
EAC
TSD.
wpd
Page
22
of
46
Table
6­
4.
2002
NETAC
and
two
Shreveport
EAC
counties
VOC
emission
summaries
for
each
day
of
the
episode
for
each
source
category.
May
2005
NETAC
EAC
TSD.
wpd
Page
23
of
46
May
2005
NETAC
EAC
TSD.
wpd
Page
24
of
46
2007
and
2012
FUTURE
YEAR
EIs
The
2007
future­
year
baseline
simulation
incorporates
the
effects
of
population
and
industry
growth
as
well
as
national
or
statewide
control
measures
or
programs
that
are
expected
to
be
in
place
by
2007.
The
future­
year
baseline
emissions
inventory
is
based
on
typical
summer
day
emissions,
with
adjustments
for
source­
specific
and
episode­
specific
information.
Future
year
emission
estimates
for
oil
and
gas
production
in
the
Northeast
Texas
Basin
were
provided
by
the
Texas
Oil
and
Gas
Association.

Point
Sources
Point
source
data
were
obtained
from
several
different
sources,
processed
separately
and
merged
prior
to
modeling.
The
data
include:
°
Texas
electric
generating
units
(
EGUs)
°
Texas
non­
EGU
point
sources
°
Facility
specific
data
°
Texas
minor
point
sources
°
Other
State
point
sources
The
2007
Texas
point
source
data
were
provided
to
NETAC
by
TCEQ
in
EPS2x
AFS
input
format.
The
hourly
EGU
data
are
developed
from
the
EPA's
Acid
Rain
Program
Database
and
are
based
on
30­
day
peaks
at
each
facility
in
the
summer
quarter
of
1997,
1998
and
1999.
These
data
include
`
new'
sources
within
100
miles
of
the
non­
attainment
areas.
Controls
weree
applied
to
the
EGU
data
to
represent
TCEQ's
NOx
rules.
The
TCEQ
Point
Source
Data
Base
(
PSDB)
was
the
basis
of
the
non­
EGU
Texas
data.
These
data
were
provided
as
2007
estimates
and
incorporated
growth
and
controls.

Many
facilities
in
the
Northeast
Texas
region
provided
future
year
emission
estimates
in
developing
the
Northeast
Texas
Region
Ozone
SIP
Revision
which
are
used
in
this
modeling
inventory.
These
sources
were
removed
from
the
Texas
files
listed
above
and
replaced
with
the
provided
SIP
data.
In
addition,
permits
for
new
EGU
units
in
the
Northeast
Texas
region
were
added.

For
all
states
other
than
Texas
the
U.
S.
EPA
2007
national
inventories
developed
to
assist
future
modeling
of
the
Heavy­
Duty
Engine
and
Vehicle
Standards
and
Highway
Diesel
Fuel,
henceforth
referred
to
as
2007
HDD
inventory,
were
downloaded
from
EPA's
ftp
site.

The
NOx
criterion
for
selecting
plume
in
grid
treatment
remained
the
same
as
in
the
base
case
and
"
Current"
runs.

Mobile
Sources
The
Texas
Transportation
Institute
(
TTI)
prepared
mobile
source
emissions
for
all
Texas
counties
under
contract
to
the
TCEQ.
Emission
factors
are
from
the
EPA's
MOBILE6
model.
Vehicle
miles
traveled
(
VMT)
for
2007
are
based
on
transportation
models
in
all
NNA
counties
May
2005
NETAC
EAC
TSD.
wpd
Page
25
of
46
that
have
a
complete
transportation
model
and
were
based
on
a
rural
HPMS
method
elsewhere.

The
2007
TTI
data
were
processed
using
the
same
methods
described
for
1999,
above.
The
result
of
this
processing
was
a
mobile
emissions
inventory
that
accurately
reflects
the
temperature
and
humidity
in
a
given
county
during
the
modeling
period.
This
included
benefits
from
TxLED
in
the
2007
estimates.
The
other
states
are
based
on
Mobile6.2
emission
factors
for
typical
summer
day
conditions
(
as
used
in
the
NEI99v2)
with
EPA
data
for
2007
VMT
and
fleet
turnover.

Area
Sources
Area
1999
emissions
estimates
for
the
counties
within
the
East
Texas
NNA
were
developed
specifically
for
the
NETAC
area.
The
1999
TCEQ
area
source
data
outside
the
East
Texas
NNA
is
the
basis
of
the
other
Texas
counties.
The
area
source
data
were
grown
to
2007
estimates
with
factors
by
source
classification
code
generated
using
EGAS
4.0.
The
exception
is
for
oil
and
gas
production
which
is
projected
using
the
ratio
of
2007
to
1999
production
values.
In
addition,
control
factors
were
applied
by
county
based
on
the
documented
SIP
rules.
This
included
benefits
due
to
the
impact
of
TCEQ's
water
heater
rule.
TCEQ
approximated
the
benefits
from
this
rule
for
modeling
purposes
as
a
half
a
ton
of
NOx
per
day
reductions
in
both
the
DFW
and
HGB
areas.
For
all
remaining
areas,
EPA's
2007
HDD
inventories
are
the
basis
for
the
area
regional
emissions
inventory.

Off­
Road
Sources
Off­
road
2007
emissions
estimates
for
the
counties
within
the
East
Texas
NNA
were
generated
using
NonRoad
v2002
with
local
data
for
mining
and
construction
equipment.
Aircraft
and
railroad
emissions
estimated
for
1999
were
grown
using
EGAS
growth
factors.
NonRoad
v2002
with
input
data
developed
by
TCEQ
was
run
to
estimate
off­
road
emissions
for
the
other
Texas
counties.
The
aircraft,
commercial
marine
and
railroad
emissions
are
taken
from
the
TCEQ
1999
offroad
inventory
and
projected
with
EGAS
growth
factors.
For
all
other
states,
NonRoadv2002
was
used
to
estimate
emissions.
The
aircraft,
commercial
marine
and
railroad
emissions
were
taken
from
EPA's
2007
HDD
off­
road
inventory.

Biogenic
Sources
Biogenic
emissions
were
maintained
the
same
as
those
developed
for
the
1999
base
case
modeling.

Summary
Tables
6­
5
and
6­
6
are
summaries
of
NOx
and
VOC
emissions
respectively,
by
county
for
each
category
of
emissions
for
the
NETAC
area.
May
2005
NETAC
EAC
TSD.
wpd
Page
26
of
46
Table
6­
5.
2007
NETAC
and
two
Shreveport
EAC
counties
VOC
emission
summaries
for
each
day
of
the
episode
for
each
source
category.
May
2005
NETAC
EAC
TSD.
wpd
Page
27
of
46
May
2005
NETAC
EAC
TSD.
wpd
Page
28
of
46
Table
6­
6.
2007
NETAC
and
two
Shreveport
EAC
counties
VOC
emission
summaries
for
each
day
of
the
episode
for
each
source
category.
May
2005
NETAC
EAC
TSD.
wpd
Page
29
of
46
May
2005
NETAC
EAC
TSD.
wpd
Page
30
of
46
DATA
SOURCES
FOR
2012
One
of
the
requirements
of
an
EAC
SIP
is
to
show
that
8­
hour
ozone
attainment
after
2007
would
not
be
jeopardized
by
future
growth
of
emissions
in
the
area.
This
was
done
by
projecting
local
area
emissions
(
NETAC
area)
to
2012
and
comparing
with
2007
emission
estimates.
For
the
NETAC
EACs
this
analysis
is
shown
in
Figure
10­
1.

The
point
source
inventory
was
assumed
constant
from
2007.
The
1999
NEI
version
2
was
the
basis
for
the
2012
area
estimates.
The
area
source
data
were
projected
to
2012
estimates
with
factors
by
source
classification
code
generated
using
EGAS
4.0.
The
nonroad
Model
was
used
to
generate
all
off­
road
sources
with
the
exception
of
aircraft,
railroad
and
commercial
marine.
The
2012
non­
nonroad
data
for
NETAC
was
based
on
the
1999
inventory
and
projected
to
2012
estimates
using
factors
generated
with
EGAS4.0.
NETAC
area
on­
road
mobile
estimates
are
based
on
link­
level
activity
and
MOBILE6.2
emission
factors.
May
2005
NETAC
EAC
TSD.
wpd
Page
31
of
46
7.
BASECASE/
CURRENT
MODELING
METEOROLOGICAL
MODELING
Output
from
MM5
was
compared
against
meteorological
observations
from
the
various
networks
operating
in
Texas
and
throughout
the
south­
central
U.
S.
This
was
done
both
graphically
and
statistically
to
evaluate
model
performance
for
winds,
temperatures,
boundary
layer
heights,
and
the
placement,
intensity,
and
evolution
of
key
weather
phenomena.
The
focus
of
the
evaluation
centered
on
performance
in
the
4­
km
grid,
although
a
regional
analysis
was
also
done
for
different
sections
of
the
12­
km
and
36­
km
MM5
domain.

Daily
statistical
results
for
the
several
MM5
simulations,
including
the
final
MM5
simulation,
on
the
4­
km
grid
were
conducted
for
winds,
temperature,
and
humidity
and
compared
to
the
NCAR
observation
database.
This
comparison
of
daily
statistics
from
NCAR
observation
database
with
the
statistical
performance
benchmarks
indicated
that
the
MM5
statistical
performance
mostly
meets
all
the
MM5
benchmark
goals
(
See
Table
7­
1).
Table
7­
2
includes
the
same
statistics
for
the
12­
km
domain.
The
MM5
benchmark
goals
have
been
developed
by
ENVIRON
and
TCEQ
based
on
analysis
of
previous
SIP
modeling
efforts
and
how
well
they
performed.

The
performance
evaluated
was
based
on
the
surface
domain­
average
analysis
and
the
performance
measures
can
potentially
average
out
over­
and
under­
predictions
and
mask
local
performance
problems.
There
are
only
7
observation
sites
in
the
4­
km
domain,
so
this
may
not
be
as
much
of
an
issue,
but
the
lack
of
monitors
in
the
4­
km
does
make
it
a
less
robust
statistical
data
set.
The
wind
observations
showed
that
calm
and
light
wind
conditions
were
dominant
over
the
much
of
the
episode.
Most
mesoscale
meteorological
models
usually
cannot
handle
such
conditions
well.
During
most
of
the
episode,
the
wind
speed
was
over­
predicted
in
both
Runs5b
and
6.
The
wind
speed
from
Run
6
performed
slightly
better
(
less
over
prediction)
during
the
first
few
days
and
near
the
end
of
the
episode.
The
diurnal
pattern
of
temperature
was
fairly
well
replicated.
Like
the
temperature
in
the
12­
km
grid,
the
daily
maximum
temperatures
for
Run
6
were
over­
estimated
and
the
nighttime
temperatures
were
mostly
modeled
well.

The
hourly
humidity
time
series
indicate
show
similar
over­
prediction
and
underprediction
patterns
to
those
on
the
12­
km
grid
for
both
Run
6
and
5b.
The
gross
error
and
bias
for
all
parameters
were
relative
higher
given
the
finer
resolution
and
few
observation
sites
(
statistics
tend
to
be
more
influenced
by
stochastic
noise
as
the
number
of
observation­
prediction
pairings
decrease).
The
bias
and
gross
error
for
Run
6
wind
speed
were
smaller
than
those
from
Run
5b
during
the
first
three
and
last
three
episode
days.
Run
5b
performed
better
during
the
light
or
calm
wind
conditions
in
the
middle
of
the
episode.
The
Run
6
IOA
for
wind
speed
was
slightly
better
during
most
of
the
episode
days.
The
wind
directions
from
both
runs
were
replicated
well
with
the
bias
in
the
range
of
±
20
degrees.
May
2005
NETAC
EAC
TSD.
wpd
Page
32
of
46
The
temperature
bias
from
Runs
6
and
5b
were
in
the
opposite
direction
during
most
of
the
episode
days.
Run
6
temperatures
were
generally
over
predicted,
while
Run
5b
temperatures
were
mostly
under
estimated.
However,
the
IOA
from
both
runs
were
both
quite
high.
The
humidity
bias
and
gross
error
from
Run
6
were
relatively
worse
compared
to
those
of
Run
5b,
just
like
results
for
the
12­
km
grid.
However,
the
humidity
IOA
from
both
runs
were
comparable,
ranging
from
0.5
to
0.83.
Comparisons
of
4­
km
daily
statistics
from
Runs
3b,
5b,
and
6
against
statistical
benchmarks
are
summarized
in
the
tables
below.
The
daily
minimum
to
maximum
range
over
the
episode
and
daily
average
values
are
shown;
values
in
red
denote
statistics
outside
the
benchmarks.
In
general,
wind
performance
did
not
change
significantly
from
Run
3b
through
6.
Temperature
performance
was
also
similar
among
all
three
runs.
The
humidity
over
prediction
in
Run
3b
was
improved
with
Run
5b,
but
turned
to
an
under
prediction
in
Run
6;
however,
the
IOA
was
much
improved
in
Runs
5b
and
6.
Overall,
the
modifications
made
in
Runs
5b
and
6
did
not
lead
to
any
different
performance
in
terms
of
these
daily
statistics,
except
possibly
the
intradiurnal
variability
in
humidity.

A
number
of
graphical
wind
field
plots
and
boundary
layer
plots
were
also
provided
and
evaluated.
The
graphical
plots
indicate
that
the
MM5
simulation
run
6
(
final
run)
is
doing
fairly
well
at
replicating
atmospheric
conditions.
While
the
surface
statistics
did
not
show
as
much
difference
in
model
performance
the
Run
6
had
a
much
improved
PBL
prediction
compared
to
Run
5b.

Table
7­
1.
Comparison
of
the
MM5
model
performance
statistics
in
the
NETAC
4­
km
grid
domain
for
August
1999
episode
with
ENVIRON's
statistical
performance
benchmark
values.
May
2005
NETAC
EAC
TSD.
wpd
Page
33
of
46
Table
7­
2.
Comparison
of
the
MM5
model
performance
statistics
in
the
NETAC
12­
km
grid
domain
for
August
1999
episode
with
ENVIRON's
statistical
performance
benchmark
values.

In
conclusion,
for
most
days
the
MM5
simulation
for
the
August
1999
episode
is
exhibiting
fairly
good
model
performance
that
mostly
meet
the
statistical
performance
benchmarks
established
based
on
analysis
of
past
MM5
and
RAMS
meteorological
modeling
to
support
air
quality
modeling.
The
MM5
meteorological
model
simulation
for
the
August
1999
NETAC
ozone
episode
exhibited
reasonably
acceptable
performance
for
surface
winds,
temperature
and
humidity.
The
MM5
model
was
also
able
to
reasonably
replicate
the
observed
upper­
level
meteorological
observations
in
the
NETAC
area.
Taking
both
statistical
and
graphical
analyses
into
consideration
the
MM5
modeling
Run
6
is
working
the
best
overall
of
the
simulations
tested
and
is
performance
is
acceptable
for
conducting
attainment
May
2005
NETAC
EAC
TSD.
wpd
Page
34
of
46
demonstration
photochemical
modeling.

CAMx
In
the
Northeast
Texas
EAC
modeling
study,
NETAC
followed
EPA
1­
hour
and
8­
hour
ozone
modeling
evaluation
guidance
to
assess
model
performance
for
all
four
episodes.
These
evaluations
consisted
of
diagnostic
analyses,
graphical
and
statistical
performance
measures.
These
performance
measures
are
to
be
used
in
conjunction
with
one
another
in
evaluating
model
performance.

The
NETAC
also
conducted
diagnostic
tests
as
part
of
the
model
performance
to
identify
flawed
model
simulations
and
to
implement
improvements
to
the
model
input
files
in
a
logical,
defensible
manner.
For
instance,
a
series
of
diagnostic
model
tests
were
performed
examining
the
role
of
initial
and
boundary
conditions
and
anthropogenic
emissions.
In
addition,
following
EPA
8­
hour
and
1­
hour
ozone
modeling
evaluation
guidance,
NETAC,
as
a
part
of
modeling
evaluation
process,
produced
various
graphical
products
including
spatial
maps
of
predictions
and
observations,
scatter
and
Q­
Q
plots,
and
time
series
plots
at
each
monitoring
station
across
the
36
km,
12
km,
and
4
km
domains
for
all
pollutant
species
for
which
measurements
were
available.
Overall,
the
8­
hour
spatial
plots
show
fairly
good
agreement
with
observed
concentrations
compared
with
the
1­
hr
plots.
Scatter
plots
and
Q­
Q
plots
of
8
hour
and
1­
hour
concentration
distributions
did
not
show
spurious,
obviously
flawed
trends.
Figure
7­
1
shows
the
1­
hour
ozone
values
for
the
August
15
­
22,
1999
episode.
Figure
7­
2
shows
the
8­
hour
ozone
values
for
the
same
period.

In
addition
to
spatial
and
other
graphical
analyses,
a
number
of
statistical
metrics
were
also
evaluated
in
accordance
with
EPA
guidance.
EPA's
1­
hour
ozone
SIP
modeling
guidance
lists
three
performance
goals:
°
Unpaired
1­
Hour
Ozone
Peak

20%
°
Normalized
Bias
for
Hourly
Ozone

15%
°
Normalized
Gross
Error
for
Hourly
Ozone
<
35%
EPA's
draft
8­
hour
ozone
modeling
guidance
also
suggests
using
the
bias
and
gross
error
performance
goals
for
daily
maximum
8­
hour
ozone
performance
evaluation.
These
ozone
statistics
were
evaluated
for
both
the
1­
hour
statistics
and
8­
hr
statistics
and
included
in
the
SIP
modeling.
Only
some
of
the
statistics
for
the
1­
hour
ozone
are
being
documented
in
this
write­
up,
but
the
full
set
of
statistics
(
1­
hour
and
8­
hour)
that
were
evaluated
are
available
in
NETAC's
and
TCEQ's
submissions.
The
spatially
paired
procedure
is
the
current
recommended
method
for
conducting
these
statistics
and
the
closest
value
is
an
alternative
approach
to
the
EPA
recommended
method.
Table
7­
3
presents
the
1­
hour
ozone
statistics
and
the
maximum
predicted
and
monitored
values
for
each
day
of
the
modeling
episode.
The
statistics
for
both
of
these
group
of
monitors
is
within
EPA
recommendations
with
the
exception
of
the
Unpaired
peak
on
the
18th
and
19th,
which
is
slightly
above
the
20%
recommended
level
and
the
Normalized
Bias
is
out
of
EPA's
recommended
range
on
the
16th,
20th
and
21st.
In
an
area
where
monitor
network
is
not
very
dense,
such
as
the
NETAC
area
during
the
1999
episode,
an
over­
prediction
bias
in
the
unpaired
domain
wide
peak
is
not
of
major
concern
in
this
case
since
this
is
probably
due
to
the
lack
of
monitor
coverage
in
the
area.
The
Normalized
Bias
calculations
are
also
more
sensitive
to
individual
monitor
performance
issues
with
a
small
set
of
monitors.
The
Normalized
May
2005
NETAC
EAC
TSD.
wpd
Page
35
of
46
Gross
Error
statistics
were
within
EPA's
recommended
levels
for
all
episode
days
excluding
ramp­
up
days.

Table
7­
4
presents
the
8­
hour
ozone
statistics
for
each
of
the
days
of
the
modeling
episode.
Some
of
the
negative
bias
values
may
result
from
spatial
miss­
matches
between
the
locations
of
modeled
and
observed
high
ozone.
Figure
7­
3
shows
a
scatter
plot
of
nearest
observed
and
predicted
8­
hour
ozone
(
ppb)
near
monitor
locations
in
Northeast
Texas.
This
figure
shows
values
centered
on
the
1:
1
line
with
no
clear
tendency
toward
over
or
under
prediction.
The
values
all
lay
within
20%
of
the
1:
1
line
which
shows
good
agreement.
The
correlation
coefficient
(
r2)
for
the
scatter
plot
of
predicted/
observed
pairs
is
0.76
which
meets
EPA
guidance
for
a
moderate
to
large
positive
correlation.
May
2005
NETAC
EAC
TSD.
wpd
Page
36
of
46
Figure
7­
1.
Spatial
Mean
1­
hr
Ozone
Concentrations
for
NETAC
base
case
monitors.
May
2005
NETAC
EAC
TSD.
wpd
Page
37
of
46
Figure
7­
2.
Spatial
Mean
8­
hr
Ozone
Concentrations
for
Episode.
May
2005
NETAC
EAC
TSD.
wpd
Page
38
of
46
Table
7­
3.
1­
hour
Ozone
Statistical
Performance
Evaluation
­
final
base
case
(
Base
5).

Episode
day
Peak
observed
(
ppb)
Peak
modeled
(
ppb)
Unpaired
peak
accuracy
(
±
20%)
Normalized
Bias
(
±
15%)
Gross
error
(
35%)

August
15
95
84
­
11
­
9
14
August
16
124
105
­
15
25
August
17
134
136
2
­
13
20
August
18
91
136
7
12
August
19
101
133
­
6
12
August
20
99
109
10
19
August
21
107
98
­
9
24
August
22
107
106
­
1
­
8
15
May
2005
NETAC
EAC
TSD.
wpd
Page
39
of
46
Table
7­
3.
CAMx
8­
Hour
Ozone
Model
Performance
Evaluation
Results.
Figure
7­
3.
Scatter
plot
of
nearest
observed
and
predicted
8­
hour
ozone
(
ppb)
near
monitor
locations
in
Northeast
Texas.
Quantiles
are
also
shown
as
circles
and
the
dashed
lines
show
+/­
20%
bias.
The
r2
value
is
the
correlation
coefficient.
May
2005
NETAC
EAC
TSD.
wpd
Page
40
of
46
SUMMARY
On
most
episode
days
the
Run
5
base
case
simulation
achieves
most
of
EPA's
statistical
model
performance
goals
for
8­
hour
and
1­
hour
ozone
concentrations.
The
1­
hour
and
8­
hour
ozone
hourly
performance
statistics
do
suggest
an
underestimation
tendency
that
is
likely
partly
due
to
understated
ozone
and
precursor
transport
and/
or
missing
or
understated
emissions.
However,
in
general
the
NETAC
Run5
base
case
model
simulation
is
exhibiting
sufficient
skill
and
meeting
most
performance
goals
so
that
it
may
be
used
to
project
future­
year
ozone
air
quality
and
8­
hour
ozone
attainment
recognizing
the
inherent
uncertainties
in
atmospheric
modeling
process.
EPA's
guidance
is
to
look
at
the
suite
of
1­
hour
and
8­
hour
metrics
and
the
different
graphical
techniques
as
an
overall
assessment
in
determining
if
acceptable
model
performance
was
achieved.
Therefore,
it
is
important
to
look
at
the
model
performance
evaluation
as
a
whole
and
seek
to
determine
whether
the
model
is
suitable
for
use
in
the
intended
purpose
of
an
8­
hour
ozone
attainment
demonstration.
Overall
conclusions
from
the
performance
evaluation
are:

°
Modeled
ozone
formation
is
consistent
with
conceptual
model
in
showing
that
high
ozone
levels
in
Northeast
Texas
resulted
from
a
combination
of
production
from
local
emissions
sources
combined
with
a
regional
background
and
transport
of
ozone.

°
Model
performance
is
fairly
good
overall
with
some
tendancy
towards
site­
specific
underestimation
bias.

°
Overall,
statistical
measures
for
the
fine
grid
are
within
the
EPA
recommended
ranges
much
of
the
time.
The
lack
of
local
monitors
(
sparse
network)
to
use
in
the
statistical
tests
make
the
statistical
tests
more
sensitive
to
paired
performance
issues
at
one
or
two
monitors.
May
2005
NETAC
EAC
TSD.
wpd
Page
41
of
46
8.
FUTURE
YEAR
MODELING
In
this
section
we
discuss
the
procedures
used
for
the
future­
year
modeling
and
the
projections
of
future­
year
8­
hour
ozone
Design
Values
for
the
2007
emission
scenarios.

FUTURE­
YEAR
MODELING
INPUTS
Development
of
the
2002
and
2007
future­
year
base
case
emission
inputs
was
described
previously.
The
same
August
1999
episode
meteorological
conditions
were
used
in
the
current
year
2002
and
future
year
2007
modeling
as
used
in
the
1999
base
case
modeling.
Thus,
effects
of
meteorological
cycles,
land
use
variations
and
other
phenomena
that
may
affect
meteorological
conditions
in
the
future
are
not
taken
into
account.
The
same
initial
and
boundary
conditions
that
were
used
for
the
1999
and
2002
modeling
were
also
used
for
the
2007
future­
year
modeling.

PROJECTION
OF
2007
8­
HOUR
OZONE
DESIGN
VALUES
The
EPA
draft
guidance
for
8­
hour
ozone
modeling
contains
specific
procedures
for
using
the
modeling
results
in
a
relative
fashion
to
scale
the
observed
8­
hour
ozone
Design
Values
to
project
future­
year
8­
hour
ozone
Design
Values
for
comparisons
with
the
standard.
These
procedures
were
used
to
estimate
2007
8­
hour
ozone
Design
Values
under
the
various
2007
emission
scenarios.

The
procedures
for
projecting
future­
year
8­
hour
ozone
Design
Values
starts
with
a
current
observed
8­
hour
ozone
Design
Value
for
each
monitor.
The
modeling
results
are
used
in
a
relative
fashion
to
scale
the
observed
8­
hour
ozone
Design
Values.
This
is
done
through
a
model
estimated
Relative
Reduction
Factor
(
RRF)
that
is
the
ratio
of
the
estimated
8­
hour
ozone
concentrations
from
the
futureyear
to
current­
year
emission
scenarios.
The
RRF
is
used
to
scale
the
current
year
observed
Design
Value
(
DVC)
to
estimate
the
projected
future­
year
8­
hour
ozone
Design
Value
(
DVF):

DVF
=
DVC
x
RRF
The
RRF
is
defined
as
the
ratio
of
the
average
of
the
maximum
8­
hour
ozone
concentrations
near/
at
each
monitor
for
the
future­
year
emissions
scenario
to
the
average
for
the
current
year
base
case
emissions
scenario.
Near
the
monitor
is
defined
by
an
array
of
7
x
7
grid
cells
centered
on
the
monitor
for
the
4
km
grid
cell
resolution
used
in
the
NETAC
modeling.
EPA's
1999
DRAFT
8­
hour
modeling
guidance
was
followed
to
estimate
the
future­
year
8­
hour
ozone
Design
Values
for
the
2007
emission
scenarios.

EPA's
1999
draft
8­
hour
ozone
modeling
guidance
includes
the
following
language
for
selecting
the
current­
year
observed
8­
hour
ozone
Design
Values
that
are
used
in
the
modeled
attainment
demonstration
test:

"
States
should
review
monitored
data
from
(
a)
the
3­
year
period
`
straddling'
the
year
May
2005
NETAC
EAC
TSD.
wpd
Page
42
of
46
represented
by
the
most
recent
available
emissions
inventory
(
e.
g.,
1995­
1997,
for
a
1996
inventory),
and
(
b)
the
3­
year
period
used
to
designate
an
area
`
nonattainment'.
The
current
design
value
used
in
the
modeled
attainment
and
screening
tests
is
the
higher
estimate
from
(
a)
and
(
b)."
(
EPA,
1999)."

For
the
first
criteria
and
the
NETAC
EAC
photochemical
modeling,
two
current­
year
base
case
emissions
inventories
were
estimated,
a
2002
EI
based
on
1999
NEI
v2
and
estimated
as
described
in
Section
6.
The
2002
EI
is
more
recent
than
1999
EI
but
some
concern
exists
that
the
2002
EI
is
not
a
true
2002
NEI
level
and
has
many
assumptions
and
projections
that
may
lead
to
uncertainties
in
actual
modeled
values.
For
the
second
criteria,
8­
hour
ozone
attainment
designations
are
being
based
on
2001­
2003
air
quality
data.
A
point
was
made
in
the
submittal
that
if
the
2002
EI
is
accurate
enough
that
both
criteria
(
a)
and
(
b)
in
EPA's
guidance
indicate
that
2001­
2003
observed
Design
Values
should
be
used
in
the
NETAC's
future­
year
Design
Value
projections.

EPA
Region
VI
has
indicated
that
interpretation
of
criteria
(
a)
"
most
recent
available
inventory"
refers
to
the
most
recently
available
NEI,
which
the
1999
NEI
at
the
time
was
the
most
recent
full
version
of
the
NEI.
EPA's
expressed
some
concern
that
the
EI
should
be
a
ground
up
estimate
for
2002
for
the
whole
modeling
domain
that
doesn't
include
projections
or
assumptions
of
no
growth
(
point
sources
were
assumed
to
be
equal
to
1999
for
the
2002
EI
used)
because
an
inaccurate
EI
could
give
misleading
conclusions.
Since
this
episode
had
a
strong
stagnation
component
for
much
of
the
episode,
the
updating
of
local
EI
and
other
areas
as
much
as
possible
to
input
actual
2002
estimates
was
deemed
acceptable.
The
SIP
included
modeling
that
was
conducted
with
a
future­
year
attainment
of
the
8­
hour
ozone
standard
calculated
using
the
2001­
2003
observed
8­
hour
ozone
Design
Values.
The
modeling
for
the
SIP
also
included
further
analysis
in
weight
of
evidence
discussions
that
also
projected
2007
8­
hour
ozone
Design
Values
in
NETAC
area.
The
projection
of
the
8­
hour
ozone
Design
Values
for
the
2007
Base
Case
and
monitoring
sites
in
NETAC
area
the
2001­
2003
observed
8­
hour
ozone
Design
Values
(
DVs)
are
shown
in
Tables
8­
1.

Table
8­
1.
Projected
2007
8­
hour
ozone
Design
Values
(
DVs)
in
NETAC
for
the
2007
Base
Case
and
with
previous
controls
put
in
place
under
the
area's
1­
hour
SIP
using
the
2001­
2003
observed
DVs.

Monitor
2001­
2003
DV
Modeled
RRF
Projected
2007
DV
Longview
82
0.981
80
Tyler
81
0.954
77
Karnack
84
0.958
80
Waskom
84
0.966
81
Notes:
The
Longview
and
Tyler
monitors
have
2001­
2003
DVs.
The
Karnack
and
Waskom
monitors
have
two
years
of
data
(
2002­
2003)
that
can
be
used
in
the
attainment
test.

Using
the
2001­
2003
observed
DVs
attainment
is
projected
at
all
sites
in
the
NETAC
area.
The
May
2005
NETAC
EAC
TSD.
wpd
Page
43
of
46
maximum
projected
2007
DV
using
the
observed
2001­
2003
DVs
at
the
Waskom
monitor
is
81
ppb.
To
demonstrate
attainment
the
maximum
projected
8­
hour
ozone
DV
must
be
84.9
ppb
or
lower.

EPA
Evaluation
Summary
The
EPA
believes
that
NETAC's
EAC
8­
hour
ozone
photochemical
modeling
study
developed
for
the
Northeast
Texas
area
meets
EPA
requirement.
The
NETAC
has
adequately
followed
the
EPA's
guidance
in
demonstrating
Northeast
Texas
Area
will
be
in
attainment
of
ozone
NAAQS
in
2007
and
2012.

The
modeling
results
show
the
maximum
predicted
2007
8­
hour
ozone
Design
Value
in
the
Northeast
Texas
region
in
the
future
will
be
well
below
the
8­
hour
ozone
NAAQS.
The
projected
2007
8­
hour
ozone
levels
at
Longview
and
Tyler
are
4
and
7
ppb,
respectively,
below
the
highest
level
that
would
demonstrate
attainment
(
84
ppb)
providing
a
margin
of
safety
for
the
study.
The
EPA
proposes
to
approve
NETAC's
EAC
8­
hour
ozone
attainment
demonstration
for
Northeast
Texas
region.

9.
ATTAINMENT
DEMONSTRATION
2007
CONTROL
SCENARIOS
In
the
NETAC
modeling
submitted
by
Texas
to
EPA
did
not
have
any
additional
local
control
strategies
modeled.
In
the
2007
base
case
modeling,
the
modeling
did
include
NOx
reductions
due
to
the
agreed
orders
on
power
plants
in
the
NETAC
area
and
reductions
at
Eastman's
plant
that
were
all
reductions
put
in
place
in
the
NETAC
1­
hour
ozone
SIP.
In
response
to
TCEQ
and
EPA's
comments
the
area
did
conduct
further
modeling
and
reported
results
in
a
September
2004
memo
from
ENVIRON.
This
modeling
included
a
sensitivity
run
that
include
the
impacts
due
to
the
state's
gas
can
rule
and
VOC
reductions
at
Eastman's
plant
that
are
being
implemented
by
through
voluntary
permit
modifications.
These
additional
measures
lowered
ozone
slightly,
but
with
the
truncation
method
used
in
calculating
the
Future
year
DV,
no
change
in
the
Future
year
DV
was
demonstrated.
No
other
modeling
was
conducted
that
attempted
to
include
reduction
in
emissions
due
to
voluntary
measures
that
the
area
has
implemented
and
included
in
their
EAC.

Ozone
attainment
is
projected
at
all
monitors
in
the
NETAC
area
for
2007
when
the
2001­
2003
observed
DVs
are
used
in
the
projections
(
See
Table
8­
1).

Application
of
EPA's
Screening
Test
EPA's
1999
Draft
guidance
has
a
screening
test
that
applies
the
modeled
attainment
test
for
grid
cells
without
a
monitor
that
consistently
exhibit
high
estimated
ozone
concentrations.
Screening
cells
are
those
that
exceed
the
estimated
ozone
at
a
near
by
monitor
by
5%
or
more
for
at
least
50%
of
the
days
modeled.
No
screening
test
was
required
since
the
threshold
was
not
reached.
May
2005
NETAC
EAC
TSD.
wpd
Page
44
of
46
Weight­
of­
Evidence
(
WOE)

Ozone
Air
Quality
and
Emission
Trends
and
Modeling
Trends
were
analyzed
of
annual
4th
highest
8­
hour
average
ozone
concentrations
at
monitoring
sites,
as
well
as
trends
in
projected
emissions
(
see
Section
10
for
details).
The
area
also
conducted
source
apportionment
modeling
that
further
corroborated
that
attainment
should
continue
to
be
achieved.

The
general
downward
trends
in
ozone
at
the
NETAC
monitors
over
recent
years
combined
with
continued
projected
downward
trends
in
VOC
and
NOx
emissions
in
the
NETAC
area
support
that
ozone
levels
should
continue
to
drop
in
the
NETAC
area
and
the
area
should
continue
to
attain
the
8­
hour
ozone
standard
in
2007.

Conclusions
of
NETAC
WOE
Current
air
quality,
trends
in
air
quality,
emission
trends,
and
corroborative
modeling
analysis
seem
to
indicate
that
the
NETAC
area
will
continue
to
attain
the
8­
hour
ozone
standard
in
2007.
It
should
be
pointed
out
that
based
on
the
DRAFT
1999
Modeling
guidance
the
NETAC
area
passed
the
attainment
test
and
did
not
need
to
conduct
a
weight­
of­
evidence,
but
the
additional
analysis
further
corroborates
the
passing
of
the
attainment
test
and
makes
a
stronger
technical
package.
The
NETAC
area
is
currently
monitoring
attainment.

10.
POST
ATTAINMENT
DEMONSTRATION
­
Future
Growth
Analysis
Comparison
of
2007
and
2012
Emissions
One
of
the
requirements
of
an
EAC
SIP
is
to
show
that
8­
hour
ozone
attainment
after
2007
would
not
be
jeopardized.
This
is
done
by
projecting
local
area
emissions
to
2012
and
demonstrating
that
ozone
precursor
emissions
between
2007
and
2012
are
projected
to
be
reduced.
For
the
NETAC
EAC
area
this
analysis
is
shown
in
Figure
10­
1.

Total
anthropogenic
NOX,
VOC
and
CO
emissions
in
the
NETAC
5
county
area
are
projected
to
be
further
reduced
between
2007
and
2012.
As
all
ozone
precursors
in
the
NETAC
5
county
area
are
being
projected
to
be
further
reduced
between
2007
and
2012,
then
attainment
of
8­
hour
ozone
standard
should
be
maintained
in
the
NETAC
area
after
2007.
May
2005
NETAC
EAC
TSD.
wpd
Page
45
of
46
Figure
10­
1.
Trends
in
Northeast
Texas
episode
average
anthropogenic
emissions
(
tons/
day)
from
1999
to
2012.

This
projected
2012
inventory
consisted
of
point
source
emissions,
on­
road
mobile
source
emissions,
non­
road
mobile
emissions,
oil
and
gas
production
emissions
forecasted
out
2012.
Specifically,
the
future
mobile
inventory
was
based
on
MOBILE6
was
run
using
default
input
parameters
for
the
year
2012.
The
NONROAD
model
was
likewise
run
for
2012
for
those
counties
to
provide
emissions
for
off­
road
sources.
Meanwhile,
area
sources
were
grown
from
2007
to
2012
using
the
EGAS
growth
and
projection
model.
Emission
estimates
for
2012
for
oil
and
gas
production
in
the
Northeast
Texas
area
was
provided
by
the
Texas
Oil
and
Gas
Association.
For
the
projected
2012
point
source
inventory,
NETAC
included
emission
estimates
for
two
proposed
plants.
The
summary
of
the
2012
May
2005
NETAC
EAC
TSD.
wpd
Page
46
of
46
modeling
inventory
for
the
Northeast
Texas
area
is
predicted
to
be
significantly
lower
than
the
2007
NOx
emission
inventory
showing
that
emissions
controls
outweigh
the
effects
of
emissions
growth
from
2007
to
2012.
As
indicated,
the
impact
of
year
2012
emissions
in
the
Northeast
Texas
region
is
projected
to
have
minimal
effect
on
8­
hour
ozone
concentrations.

NETAC
staff
have
indicated
that
they
will
continue
analyze
air
quality
and
related
data
and
analyses.
The
area
has
indicated
that
they
will
continue
to
analyze
and
update
emission
and
monitoring
trends
(
including
new
point
sources)
and
include
these
in
reports
that
will
track
and
document
progress
towards
attainment.
