Importance
of
Using
Plume­
in­
Grid
Modeling
to
Estimate
Power
Plant
PM
Impacts
Overview
°
Limitations
of
3­
D
grid
modeling
°
Plume
chemistry
and
PM
modeling
°
Errors
in
using
purely
gridded
approach
°
Plume­
in­
grid
(
PiG)
modeling
°
Results
from
PiG
modeling
°
Conclusions
Limitations
of
3­
D
Grid
Modeling
°
Artificial
dilution
of
stack
emissions
°
Unrealistic
near­
stack
plume
concentrations
°
Incorrect
representation
of
plume
chemistry
°
Incorrect
representation
of
plume
transport
Plume
Size
vs
Grid
Size
(
from
Godowitch,
2004)

Horizontal
resolution:
36
km
Plume
sections
are
less
than
grid
dimensions
for
several
grid
cells
downwind
of
source
locations
Plume
Chemistry
and
PM
Modeling
Early
Plume
Dispersion
No
PM2.5
Formation
1
2
Mid­
range
Plume
Dispersion
Sulfate
and
nitrate
formation
rates
in
plume
less
than
ambient
background
rates
Long­
range
Plume
Dispersion
3
Ambient
conditions
Potential
Errors
in
PM
Modeling
Using
a
Gridded
Approach
°
Near­
source
acid
formation
rates
downwind
of
large
NOx
point
sources
over­
estimated
 
Point
source
PM
sulfate
concentrations
overestimated
(
if
the
source
emits
SO2
in
addition
to
NOx)

 
Point
source
total
nitrate
concentrations
overestimated
 
Point
source
PM
nitrate
concentrations
may
be
overestimated
depending
on
ambient
conditions
Plume­
in­
Grid
Modeling
°
Plume
model
embedded
within
grid
model
("
host")

°
Plume
model
handles
transport
and
chemistry
of
point
source
emissions
in
initial
and
intermediate
plume
stages
°
When
certain
criteria
(
e.
g.,
plume
size,

chemical
maturity
of
plume)
are
satisfied,
the
plume
model
"
hands­
over"
calculations
to
"
host"
PiG
Models
with
CMAQ
as
Host
°
CMAQ­
PinG
(
EPA;
Godowitch,
2004)

 
October
2004
release
of
CMAQ
(
Version
4.4)

includes
a
plume­
in­
grid
treatment
for
PM
°
CMAQ­
APT
(
EPRI)

 
State­
of­
the­
science
treatment
of
stack
plumes
at
the
sub­
grid
scale
using
the
reactive
plume
model,

SCICHEM
 
Tested
for
ozone
and
gases
(
Karamchandani
et
al.,

2002)

 
PM
version
available
in
early
2005
(
Karamchandani
et
al.,
2004)
PiG
Application
with
CMAQ­
APT
°
Northeastern
U.
S.

 
July
11­
15,
1995
 
12
km
grid
resolution
 
30
largest
NOx
point
sources
selected
for
PiG
treatment
 
Total
NOx
emissions
from
selected
sources:

5765
Mg/
day
~
14%
of
domain­
wide
NOx
emissions
°
PiG
impacts
for
O3
and
HNO3
evaluated
Northeastern
U.
S.
Application
PiG
Sources
Effect
of
PiG
Treatment
on
HNO
3
Concentrations
(
July
13,
1995)

CMAQ­

APT
­

CMAQ
Point
Source
HNO
3
Contributions
from
Grid
and
PiG
Modeling
Base
CMAQ
CMAQ­
APT
Effect
of
PiG
Treatment
on
HNO
3
Concentrations
(
July
14,
1995)

CMAQ­

APT
­

CMAQ
Point
Source
HNO
3
Contributions
from
Grid
and
PiG
Modeling
Base
CMAQ
CMAQ­
APT
Effect
of
PiG
Treatment
on
HNO
3
Concentrations
(
July
15,
1995)

CMAQ­

APT
­

CMAQ
Point
Source
HNO
3
Contributions
from
Grid
and
PiG
Modeling
Base
CMAQ
CMAQ­
APT
HNO
3
Mass
Budgets
°
The
total
HNO3
mass
integrated
across
the
entire
domain
in
the
surface
layer
is
about
6%

lower
in
the
PiG
simulation
than
in
the
non­

PiG
simulation
°
The
total
HNO3
mass
integrated
across
the
entire
domain
over
all
layers
is
about
3%
lower
in
the
PiG
simulation
than
in
the
non­
PiG
simulation
Conclusions
°
Traditional
grid
modeling
over­
estimates
sulfate
and
total
nitrate
formation
in
plumes
from
large
NOx
point
sources
potentially
incorrect
estimates
of
their
contributions
to
PM2.5
°
Plume­
in­
grid
(
PiG)
PM
modeling
provides
a
better
representation
of
the
fate
of
point
source
emissions
°
Using
PiG
treatment
is
essential
to
calculate
PM
impacts
of
large
NOx
point
sources,
particularly
when
coarse
grid
resolutions
are
employed
in
the
grid
model

