P1.4
RESULTS OF PHOTOCHEMICAL SIMULATIONS
OF SUBGRID SCALE POINT SOURCE EMISSIONS
WITH THE MODELS-3 CMAQ MODELING SYSTEM
James M. Godowitch*
Atmospheric Sciences Modeling Division
Air Resources Laboratory
National Oceanic and Atmospheric Administration
Research Triangle Park, North Carolina
1. INTRODUCTION
Plume-in-grid approaches have been
incorporated into Euierian air quality grid models to
provide a more realistic treatment of the dynamic
and chemical processes governing pollutants
emitted from major point sources. Substantial
emissions of nitrogen oxides (NO,) and/or sulfur
oxides (SO*) are released from individual point
sources into plumes with horizontal widths
considerably smaller than the typical dimension of
regional photochemical model grid cells (e.g. 20-40
km). Additionally, the pollutant mixture of fresh
plumes from many point sources, particularly from
fossil-fuel power plants, can be characterized to be
in a high NO, / low VOC regime, while the ambient
background surrounding a plume is often in the
opposite chemical regime. The traditional Euierian
grid modeling approach has been to instantly mix
point source emissions into an entire grid cell
volume. This method, however, bypasses the
diffusion-limited, chemical evolution occurring in
subgrid scale plumes during transport downwind.
The effect of this overdiiution and the simultaneous
availability of high NO* emissions and volatile
organic compounds (VOCs) of anthropogenic
and/or biogenic emissions in a grid cell are to
promote the premature initiation of rapid
photochemical production of secondary species,
such as ozone (Os). Since a plume-in-grid
technique is designed to spatially resolve the
subgrid scale concentration gradients in a plume
and to simulate gradual plume growth in the
horizontal and vertical in response to
meteorological processes, photochemistry can be
captured in a more realistic manner.
Measurements of a variety of pollutants in plumes
downwind of major point sources during recent field
studies provide opportunities for plume-in-grid
evaluation efforts to assess the capabilities of
these subgrid scale plume treatments.
There have been a limited number of recent
plume-in-grid efforts with a reactive plume
algorithm being fully integrated into a 3-D
photochemical grid model. A method with a plume
composed of elliptical rings was embedded in the
* Author address: J. Godowitch, US EPA, NERL,
MD-80, RTF, NC 27711. On assignment to the
EPA National Exposure Research Laboratory;
e-mail: jug@hpcc.epa.gov
UAM-V mode! (Morris et al., 1992) and this same
plume approach was incorporated in the SAQM
model (Myers et al., 1996). Kumar and Russell
(1996) developed and applied a plume model in the
URM model with plume puffs treated for a limited
period followed by the plume material being
transferred into the grid. Karamchandani et al.
(2000) described the SCICHEM / SCIPUFF plume
model approach and they also presented
evaluation results of their model against recent
plume data.
A plume-in-grid (PinG) approach has been
incorporated into the Community Multiscale Air
Quality (CMAQ) modeling system (Byun et a!.,
1998). The CMAQ model system is composed of
state-of-science meteorological, emissions, and air
quality grid modeling components that reside in the
Models-3 system framework (Novak et al., 1998).
The latter also contains tools for model execution
and for various analyses of model outputs. The
PinG approach was specifically designed to
address the need to more realistically resolve the
spatial scale of plumes emanating from isolated,
high emission point sources within an Euierian
coarse grid framework. The key science codes of
the CMAQ plume-in-grid approach are a plume
dynamics model (POM) and a Lagrangian reactive
plume model, which is designated as the PinG
module since it is fully coupled with the CMAQ
Chemical Transport Model (CCTM). An overview of
these plume model components will be given.
The CCTM/PinG model was applied on a
domain encompassing the greater Nashville,
Tennessee region. Model simulations were
performed for selected days in July 1995 during the
Southern Oxidant Study (SOS) field study program,
which was conducted in the Nashville area. In
particular, five major point sources exhibiting a
range of NOX emission rates were selected for the
PinG treatment. Selected PinG model species
concentrations and representative samples of the
initial results of an ongoing evaluation of the PinG
model with the SOS/Nashville 1995 data are
presented to provide a preliminary demonstration of
the capability of the CMAQ/PinG approach. In
particular, modeled concentrations are compared
to plume data for pollutant species collected during
horizontal traverses by an instrumented helicopter
and research aircraft across different plumes.
-------
2, MODEL OVERVIEW
The PinG approach is based on the Lagrangian
reference frame with a continuous plume
represented by a series of moving plume sections
released at hourly intervals from the location of a
point source. The PinG module simulates one or
multiple point source plumes and is designed to
treat the physical and chemical processes during
the entire diurnal cycle. The PDM processor
generates plume dimensions, plume section
position on the grid, and other relevant parameters
needed by the CCTM/PinG. The Lagrangian PinG
module is fully coupled with the CCTM and is
executed simultaneously with the grid model
(Godowitch et at., 1999). Horizontal resolution
across a plume section is achieved with a
contiguous array of attached plume cells.
Currently, each plume section is composed of 10
plume cells in addition to a left and a right
boundary cell. The plume boundary conditions,
representing the ambient background, are provided
throughout the simulation by the appropriate CCTM
grid cell concentrations. In the current PinG
version, the plume cells represent a single vertical
layer. Plume cells in a plume section have the
same plume bottom and a common top height.
Relevant processes have been incorporated
into the plume equation for the mass balance of
individual pollutant species in each plume cell. The
key plume processes include dilution and
entrainment due to vertical expansion, dilution and
entrainment / detrainment from horizontal
expansion, crosswind plume diffusion between
plume cells, gas-phase chemistry, surface dry
deposition, and surface emissions. The detailed
mathematical formulation of these processes is
documented in Gillani and Godowitch (1999).
Since plume chemical evolution is strongly
influenced by plume expansion, plume growth must
be realistically simulated with time after release.
The PDM processor determines plume rise and
provides the plume dimensions and growth rate
parameters for use by PinG in the physical
processes mentioned above. Parameterizations
are employed in PDM to determine the turbulence
and wind shear contributions to plume growth
based on 2-D and 3-D meteorological fields
generated by the Penn State/NCAR mesoscale
model (MM5). For consistency with the grid model,
PinG applies the same gas-phase chemical
mechanisms and chemical solvers used by the
CCTM. Currently, the RADM2 or CB-4 chemical
reaction schemes can be selected for
photochemistry. The chemical solvers include a
quasi-steady state (QSSA) method and the sparse-
matrix vectorized Gear (SMVGEAR) technique
(Gipson and Young, 1999). PinG also applies the
same pollutant deposition velocities used by the
CCTM for dry deposition and surface emission
fluxes are injected into the surface-based plume
sections. Plume transport is determined from
mean wind components, which are computed by
averaging winds from model layers spanned by
each plume section.
3, MODEL SIMULATIONS AND RESULTS
The Eulerian modeling domain used in the
CCTM/PinG simulations consisted of 21 x 21
horizontal grid cells with a 36 km grid cell size and
21 vertical layers. This model domain was
centered on the greater Nashville, Tennessee area
and ft covered most of the southeastern US. It also
is a subdomain of a much larger 36 km gridded
modeling domain encompassing the entire eastern
half of the US. By using a subdomain targeted on
the experimental study region of interest, the
computational time and file sizes were greatly
reduced. Additionally, initial and boundary
conditions for this subdomain were provided from
CCTM modeled concentrations already generated
from simulations on the large domain. Likewise,
meteorological and emissions data sets for this
application were extracted from data files for the
larger domain using existing CMAQ processor
programs. The CCTM/PinG simulations were for
24-hour periods starting at 00 GMT. The model
results to be shown were obtained from simulations
using the CB-4 chemical mechanism and QSSA
solver, which required less computational time than
the RADM2 mechanism.
A set of five major NO« point sources was
selected for the plume-in-grid treatment. The point
source emissions treated in PinG were from the
Shawnee (SH), Paradise (PA), Cumberland (CU),
Johnsonville (JV), and Gallatin (GA) fossil-fuel
power plants. Continuous Emissions Monitoring
(GEM) data from the 1995 NET (National Emission
Trends) inventory provided the hourly emissions for
each point source. All other point sources in the
domain were modeled with the traditional Eulerian
grid treatment and were included the 3-D emissions
input data file. Based on the total daily NOx
emissions for a typical day (e.g., July 7, 1995), the
lowest to highest NO* emission sources were GA,
SH, JV, CU, and PA. Using GA as a reference
source (32 tons of NO«/day), the daily NOx
emissions from SH, JV, CU, PA were a~factor of
3.0, 3.7, 15.0, and 15.2 times higher, respectively.
The PinG results from model simulations for July 7
and July 16, 1995 are emphasized in this paper
because the CU plume, in particular, was sampled
extensively at various downwind distances during
afternoon hours by the instrumented airborne
platforms.
3.1 Selected Results From PinG Simulations
The evolution of ozone (Oa) and nitrogen oxide
(NO) concentrations in a plume section released at
1400 GMT (i.e., 0900 local daylight time) from CU
and JV are displayed in Figures 1 and 2,
respectively. The set of both figures illustrates the
typical daytime chemical evolution occurring in
rather large, isolated power plant plumes, which
-------
nS MODS. RESULTS - Cumberland
JOT 7, IKS : nilllll THm - MOO OUT
-a -in -18 -w -a o 5 » » » ag
MIMCZ ACROSS
Figure 1 a. Ozone concentrations relative to
background values in the CU plume section
released at 1400 GMT for various times and
downwind distances on July 7,1995.
PWG MODEL RESULTS - Johrwonvms
.**)> 7, 1WB : »•*•••* Tton* - MOO awr
i|ifir|nii|iiii-[i m|iiii|ini[nr
-ao -w -w ~B e * w « ao zs
omtNCE Acnoae
Figure 2a. Ozone concentrations relative to
background levels in the JV plume section released
at 1400 GMT on July 7,1995.
PlnQ MODB. RESULTS : Cumberland
July 7, IMS : MMM Itm » MOO OMt
•KMW—
1 1.00^
i |
0.10 -s
0.01 —
nmu |QMI) /
OotwMlnd OUt*n» »»n)
"-*-*' KSOQ2S
*™-^— 1BOQW7
•-^-•ITSQeO
*-*-* «3»»ff? ^
«-^-»2100«7« f
<|j
A
p\
R\
1%.
H i 1 f . * i1! pi H | r i t • j i i i j i i i i j i i i I j || , 1 | M i r | r n j
e$ .ag -is -10 -ao a to is ao n
DtiTANCE ACROSS PUJME &onj
Rnd MODEL RESULTS : JohnsorMlla
<«iy 7. tgas : ngian T>T» * MOO OMT
^0,,OQ —
i ""]
0.01 —
TkwOMI)/
"• • * iraoao
-~M
J**»
A
^.v.-**"**-^ -*—
-29 -90 -» -« -» 0 B 10 « 9O 28
DISTANCE MWOBG FUME |r^
Figure 1b. Same as in Fig.la, except for NO
concentrations in the CU plume section.
consists of three distinct stages. In stage 1, the
relatively fresh plume is dominated by primary
emissions, and a deficit of Os relative to
surrounding ambient background concentrations
exists in the plumes due to substantial titration from
NO. As a plume expands and experiences mixing
with background concentrations, in this case
provided by the CCTM gridded concentrations, a
transition to stage 2 occurs with rapid Oa
production near the plume edges. During this
stage, a plume displays the characteristic 'wings'
near the plume edges as elevated Os values
exceed background values. Ozone concentrations
in the plume core also exhibit a strong recovery but
remain depressed relative to those found further
from the plume centertine. In stage 3, the plume
sections from both point sources have experienced
considerable dilution due to horizontal expansion
and are chemically mature. A noticeable 63 bulge
is found across the entire plume. The plume
sections from both sources exhibit NOx-limited
Figure 2b. Same as Fig. 2a, except for NO
concentrations in the JV plume section.
conditions in stage 3 as NO concentrations have
been reduced to near background values.
It is also apparent from these figures that the
O3 recovery and chemical evolution in the JV
plume section is more rapid than for the CU plume
section. The notable difference between these
plume sections was that the NO emission rate at
JV was considerably lower, by about a factor of 4,
than from CU at this release time and the NO
concentration differences between these plume
sections reflect this emission rate difference.
Results of PinG model results in Godowitch and
Young (2000) revealed the NO, oxidation rate was
inversely related to the NO* emission rate. Thus,
slower NOX oxidation rates occurred in the modeled
plumes from the highest NOX point source
emissions. This feature has also been supported
by plume observational results from these same
point sources in Ryerson et al. (1998).
The temporal behavior of Oa in the plume core
for the 1400 GMT release is depicted in Figure 3
for three of the point sources to demonstrate the
impact from different NO« emission rates. It is
-------
evident that the time required for Oa in the plume
core to recover to the background value and the
time of occurrence of the peak Oa differ greatly
depending on the NOX emission rate. Ozone in the
GA plume with the lowest NO* emission recovered
quickest and reached maturity fastest, although it
also exhibited the lowest peak Os relative to
background compared to either JV or CU. In this
case, the CU plume section required almost 4
hours for Oa in the plume core to reach background
levels and about i hours of travel time to reach its
peak value of about 20 ppb above background.
Interestingly, the JV plume section also achieved a
peak O3 value of about 20 ppb above background,
however, it occurred much sooner after release
than the CU peak value. This indicates the CU
peak O3 was found much further downwind. These
model results are also comparable to observational
results in Ryerson et al. (1998) who reported plume
ozone exceeded background by as much as 23
ppb. They also noted the peak Cb was found
closer to the source for the JV plume based on
aircraft traverses downwind of JV and CU during
the afternoon of July 7.
A similar plot for N02, representing the sum of
all nitrogen species generated in the photochemical
process, relative to background levels is depicted
in Figure 4, Photochemical production leads to
higher NQZ levels as NOX source strength
Increased. The peak NO2 values also occurred
concurrently with peak Os values. Additionally, it is
evident that photochemical production of nitrogen
species occurs more rapidly after release since
NO2 exceeds background levels in each point
source plume section within an hour after release.
Nitric acid (HNOs) made up the largest fraction of
NOZ in these plumes.
3.2 Comparisons of Model and Observed
Species Concentrations in Plumes
Plume concentrations generated from
CCTM/PinG simulations are compared directly to
observed species concentrations obtained during
aircraft and helicopter flights from the SOS field
study in the Nashville region. The weather
conditions during July 7, 1995 were quite favorable
for experimental sampling of isolated major point
source plumes in the greater Nashville area. With
a steady northwesterly wind flow with speeds of 5-7
m/s in the afternoon convective boundary layer, the
plumes from the major point sources remained
separated from each other and from the Nashville
urban area (Godowitch and Young, 2000), which
would have complicated interpretation of results.
The PinG modeled concentrations from plume
sections closest in time and space to the plume
measurements made during a horizontal flight path
on July 7th (Nunnermacker et al., 2000) by the
Department of Energy (DOE) G1 research aircraft
are displayed in Figure 5. The modeled
concentrations of superimposed Os, SOz, and NOy
(sum of all nitrogen species) are on the plumes
PinG MODEL RESULTS
i i, wee ; muni nra - woo OMT
TOAVB. 1ME txxi^
Figure 3. Variation of ozone in the plume core
relative to background with time after release for
the plume section emitted at 1400 GMT from the
different point sources.
PinG MODEL RESULTS
MV 7, 1SBB : ngliMB Ttm - 1400 OMT
=
°~i
I *^
9 *•=
2 O ~
-1-r
«-»-» cxmb«rtirw ,»-*H1**^i
+~*-+Qman ff'
S1
//^
y^>i ^-^.^
jf »-*».
,C** ,_~»~-»»**H
•s.
^
V
«-,!
^•vt*^
f
-»-( i | i | i j r i r i r T~
0123460
TRAVB. TWC fwunQ
1 1 '
7 a
Figure 4. Same as in Fig. 3, except for plume NOZ
after removing the background value.
intercepted at about 46 km and 36 km downwind of
JV and CU, respectively, during the aircraft flight.
The pollutant signature of each plume reflects the
impact of primary emissions from each point
source. Relatively high SOz emissions compared to
lower NO, emissions exist at JV, whereas CU emits
relatively high NO, emissions compared to low SOz
emissions. The PinG model concentrations of SOz
and NOy are strongly supported by the data
obtained through each plume, which suggests the
emissions were accurately specified and the plume
dispersion and chemistry processes, in particular,
were treated realistically out to these downwind
distances. It is also encouraging that the PinG
model simulates the plume Oa rather closely at
these distances for both sources with Figure 5
showing a mature JV ozone plume while the CU
plume still displays an ozone deficit.
-------
PlnQ RESULTS AND AWCOAFT DATA
juv T. igga : TVT» - 1*15 OMT
Figure 5. Comparison of PinG concentrations for
ozone (solid lines), NOy (short dashed lines), and
S02 (long dash) versus measured Os (o), SO2 (A)
and N0y (•) from the DOE aircraft flight across the
JV (extreme left plume) and CD plume on the
afternoon of July 7,1995. Interceptions of the JV
and CD plumes were made at downwind distances
of 46 km and 36 km, respectively.
Another model/aircraft data comparison is
depicted in Figure 6, which contains the data
collected during traverse 3 of the NOAA WP-3
aircraft flight on July 7th (Ryerson et a!., 1998).
The flight path was along a SW to NE line across
the region and intercepted the JV, CU, and PA
plumes from left to right in Figure 6. PinG model
concentrations of Os and NO are very comparable
for the JV and CU plumes at 80 km downwind.
However, for the modeled PA plume, a notable
underestimate in Os exists, although the plume Os
structure is similar to the observed plume. This
shift in the magnitude of Os for the modeled PA
plume is largely attributable to underpredicted
boundary concentrations provided by the grid
model. It became apparent from an examination
of the observed SO2 concentrations (not shown
here) and this O3 data series that another mature
plume from a major point source further upwind of
PA exists on the right of the PA plume in Figure 6.
This example demonstrates the impact on the PinG
model from grid model boundary conditions.
Measurements made during several horizontal
traverses of the CU plume at different downwind
distances from the TVA helicopter flights on the
afternoon of July 7th are compared to PinG
concentrations. Figure 7a displays in-plume Os
concentrations after subtracting model background
values. Likewise, Figure 8a contains the observed
in-plume Os structure after subtracting observed
background values. In addition, the plume width
from the centeriine to each edge was employed to
normalize the distance across each side of the
observed and model plumes at these different
downwind positions. The modeled evolution and
magnitudes for ozone and NOy shown in Figure 7a
PlnQ RESULTS AND AIRCRAFT DATA
Figure 6. Comparison of PinG ozone (thick solid
lines) and NO (dashed lines) concentrations for the
JV, CU and PA plumes (from left to right) vs. the
NOAA WP-3 aircraft data sampled at 1-s intervals
at about 80 km downwind of JV and CU, and about
100 km downwind of PA for the horizontal traverse
starting at 2006 GMT on July 7, 1995.
and 7b are quite similar with the observed patterns
depicted in Figures 8a and 8b.
Statistical results computed from paired model
and observed plume concentrations from helicopter
sampling on July 7 and 16, 1995 are presented in
Table 1 and 2, respectively. Owing to differences
between model and observed background values
of Os in these cases, Os background values were
subtracted in order to focus on the relative O3
concentrations in the modeled and observed
plumes for this species. This procedure was not
performed for other pollutants. Concentrations of
NO, in particular, are noticeably lower for the
slower wind case on July 16, which is indicative of
a somewhat 'older' plume at the same downwind
distance. More quantitative results at other
downwind distances for various species are
expected to provide a better overall picture of
model performance.
4. SUMMARY AND ONGOING WORK
Model simulations of the CCTM/PinG have been
performed to provide modeled plume species
concentrations from selected major point sources
for quantitative comparisons with plume
measurements from the SOS summer 1995
Nashville field study. These initial results and
comparisons of modeled and observed plume
concentrations are encouraging with the PinG
model exhibiting the capability of realistically
simulating the observed photochemical behavior
for Os and other species for these case studies.
Similar analyses will be performed for other case
study days. Additional PinG simulations are also
anticipated with a CCTM fine grid (e.g. 12 km)
domain and with the RADM2 chemistry
mechanism.
-------
RnO MODEL RESULTS
JULY;, m
ID-
S' "
I °;
8 ~*°T
X**' *^>^_ !
\
\
\
« . . to nn
-1» -1.0
^
\
-05 OJO Q
/
/
/
a 1J9 ^
•CALB9 DBWNGB W3M MM CBBB1**
Figure 7a. PinG model ozone concentrations
relative to background in the CU plume at
downwind distances corresponding to the
observed plumes on the afternoon of July 7,
1995 shown in Fig. 8a. Position within each
piume has been normalized by the distance
from plume centerline to each edge for each
downwind distance.
HEUCOPTER PLUME DATA
10—
S~ n ~
°-
IM
8 -»€
.JO—
^i^«""^
\
\
Dovwiwttal
*"""* * 40 IBTI
"*""•*"*• 80 fem
-w -1.0
•CMLB3
JULV7, W»5
*' -*. --*••••-_ '* — \
/ *
\_>
-ct$ (to as t« ts
DBPWCe WOM POME OBNTfEHUME
Figure 8a. Observed ozone relative to
background in the CU plume at three
downwind distances from horizontal plume
interceptions by the TVA helicopter during the
afternoon of July 7,1995.
10043-
wa-
i*-
0.1-
FVC MODEL RESULTS
JULY 7, M89
_,.•''' xv
••;>-' *\\
.A -1.0 -Of CLO 03 1.0 1-6
BCALBJ DBT/MCE mow PLUME OBTBUNE
Figure 7b. Same as in Fig. 7a, except for total
nitrogen species (NOy) concentrations.
ACKNOWLEDGEMENT
Thanks are due to the NOAA Aeronomy Laboratory
for submitting the WP-3 aircraft data and to the
Dept. of Energy for providing the G1 aircraft data to
the EPA SOS data base. Appreciation is due to
the TVA Atmospheric Sciences and Environmental
Assessments Dept. for providing the helicopter
plume model evaluation data set.
DISCLAIMER
This paper has been reviewed in accordance with
the U.S. Environmental Protection Agency's peer
HELICOPTER PLUME DATA
WO.G-
10*-
•
f
1,0-
ai-
^-—•^
*"" \
^x^l^^ "••->
/ / X
F ^ i S | 1 1 1 1 j ! 1 1 1 | t 1 1 1 [ I 1 I 1 j 1 I I i |
A -10 -o& ouo as to t»
KSM2) OBTANCE FROM FUME caJTSUJTE
Figure 8b. Same as in Fig. 8a, except for observed
N0y concentrations.
review and administrative review policies and
approved for publication. Mention of trade names
or commercial products does not constitute
endorsement or recommendation for use.
REFERENCES
Byun, D., et al., 1998: Description of the Models-3
Community Multiscale Air Quality (CMAQ) Modeling
System, 10* Joint Conf. on Applications of Air Poll.
Meteorol. with A&WMA., 11-16 Jan. 1998, Phoenix,
AZ, p 264-268.
-------
Giliani, N.V, and J.M. Godowitch, 1999: Plume-in-
grid treatment of major point source emissions.
Chap. 9, EPA/600/R-99/030, Research Triangle
Park,NC, URL://http://www.epa.gov/asmdnerl/
models3/doc/science
Gipson, G. and J.O. Young, 1999: Gas-phase
chemistry, Chap. 8, EPA/600/R-99/030, Research
TrianglePark.NC, URL:http://www.epa.gov/
asmdnerl/modelsS/doc/science/
Godowiteh, J.M., et al., 1999: Photochemical
plume-in-grid simulations of major point sources in
the CMAQ modeling system, Symp. on
Interdisciplinary Issues in Atmos. Chem., 10-15
Jan. 1999, Dallas, TX, p 121-124.
Godowitch, J.M. and J.O. Young, 2000:
Photochemical simulations of point source
emissions with the Models3 CMAQ plume-in-grid
approach., A&WMA 91S| Annual Conf., 18-22 June
2000, Salt Lake City, UT., 14 pp.
Karamchandani, et al., 2000: Development and
evaluation of a state-of-science reactive plume
model, Environ. Sci. Technol., 34, 870-880.
Kumar, N. and A.G, Russell, 1996: Development of
a computationally efficient reactive subgrid-scale
plume model, J. Geophys. Res.,101,16737-16744.
Morris, R.E., et al., 1992: Overview of the variable-
grid UAM-V. A&WMA 85'
1992, Kansas City, MO.
Annual Conf., June
Myers, T., et al., 1996: The implementation of a
plume-in-grid module in SAQM. Report SYSAPP-
96-/06, Systems Applications Intl., San Rafael, CA.
Novak, J.N., et al., 1998: Models-3: A unifying
framework for environmental modeling and
assessment. 10* Joint Conf. on Appl. Of Air Poll.
Meteorol., 11-16 Jan. 1998, Phoenix, AZ.
Nunnermacker, L J., et al., 2000, NOy lifetimes
and O3 production efficiencies in urban and power
plant plumes: Analysis of field data, J. Geophys.
Res., 105, 9165-9176.
Ryerson, T. B., et al., 1998: Emissions lifetimes
and ozone formation in power plant plumes. J.
Geophys. Res., 103, 22569-22583
Table 1. Statistics of PinG model results and observed helicopter concentrations in the Cumberland
plume at 40 km downwind on July 7,1995 (1517 - 1630 GMT)
Species
Model
Ave SD
Observed
Ave SD
Peak
Model Obs
Bias
O3 - O3bg
NO
NOy
SO2
-24.2 ± 56.1
22.2 ± 89.1
49.1+143.8
5.2 ± 14.7
-18.3 ± 41.7
17.6 ± 57.8
42.5 ±112.0
3.7 ± 12.2
-5.7 1.3
59.7 39.6
103.2 79.6
10.7 7.4
-5.9
4.7
6.6
1.5
0.86
0.96
0.95
0.72
Table 2. Statistics of PinG model results and observed helicopter concentrations in the Cumberland
plume at 36 km downwind on July 16,1995 (1748 - 1951 GMT)
Species
Model
Ave SD
Observed
Ave SD
Peak
Model Obs
Bias
O3 - O3bg
NO
NOy
SO2
-4.8 ± 81.4
5.9+ 33.2
30.2 ±115.3
4.8 ± 18.2
0.8 ±79.8
7.4 + 40.4
25.5+111.0
3.4 ±13.2
11.5 22.9
17.1 21.3
65.5 55.9
10.4 7.0
-5.6
-1.5
4.7
1.4
0.76
0.74
0.79
0.70
Note: Bias = Model - Observed , SD = standard deviation , R = correlation coefficient
Concentration units = ppb
-------
NERL-RTP-AMD-00-208
TECHNICAL REPORT DATA
1. REPORT NO.
EPA/600/A-00/114
2.
3.RECIPIENT'S ACCESSION NO.
4. TITLE AND SOBTITLE
RESULTS OF PHOTOCHEMICAL SIMULATIONS OF SUBGRID
SCALE POINT SOURCE EMISSIONS WITH THE MODELS-3 CMAQ
MODELING SYSTEM
5,REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
James M. Godowitch
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Same as Block 12
10.PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Research and Development
National Exposure Research Laboratory
Research Triangle Park, NC 27711
13.TYPE OF REPORT AND PERIOD COVERED
Proceedings. FY-01
14. SPONSORING AGENCY CODE
EPA/600/9
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The Community Multiscale Air Quality (CMAQ) / PIume-in-Grid (PinG) model was applied on a domain
encompassing the greater Nashville, Tennessee region. Model simulations were performed for selected days in July
1995 during the Southern Oxldant Study (SOS) field study program which was conducted in the Nashville area. In
particular, five major point sources exhibiting a range of NOx emission rates were selected for the PinG treatment.
Selected PinG model concentrations and representative examples of the initial results of an ongoing evaluation of
the PinG model with the SOS/Nashville data are presented to provide a preliminary demonstration of the capability of
the CMAQ/PinG approach. In particular, modeled concentrations of ozone, S02, and nitrogen oxides are compared
to plume data collected during horizontal traverses by an instrumented helicopter and research aircraft across
different plumes. Statistical results are also provided at 40 km downwind of the largest point source. The
comparisons and quantitative results are encouraging as PinG exhibited the capability to realistically simulate the
observed photochemical evolution for ozone and other species at various downwind distances for these cases.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/ OPEN ENDED
TERMS
c.COSATI
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This
Report)
UNCLASSIFIED
21.NO. OF PAGES
7
20. SECURITY CLASS (This
Page)
UNCLASSIFIED
22. PRICE
------- |