United States
Environmental Protection
Agency
Atmospheric Sciences ^,-
Research Laboratory
Research Triangle Park NC 27711 ",
Research and Development
EPA/600/S3-87/027 Dec. 1987
Project Summary
A Mesoscale Acid Deposition
Model: Preliminary
Applications and a Guide for
User Interface
Gregory R. Carmichael and Leonard K. Peters
In 1984 the U.S. EPA initiated a
Mesoscale Acid Deposition Study. This
study was organized to assess and
understand the interactions and impor-
tance of local emissions in mesoscale
precipitation scavenging. One impor-
tant component of the project is the
simulation of the field events using a
detailed meteorological and chemical
modeling framework (called Meso-
STEM). A dynamic meteorological
model (MASS—meteorological atmos-
pheric simulation system) is used to
predict the meteorological fields,
including precipitation rates. These
fields are used as inputs to a compre-
hensive chemical model (STEM-II—
Sulfur Transport Eulerian Model). The
3-dimensional models are used to
quantify the relationships between
emissions, chemical production and
wet deposition on the mesoscale. The
linkage of the models, and the appli-
cation of MesoSTEM to the May 2nd
and 3rd, 1985 Philadelphia data set is
the subject of this progress report.
This Project Summary was devel-
oped by EPA's Atmospheric Sciences
Research Laboratory, Research Trian-
gle Park, NC, to announce key findings
of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
The STEM (Sulfur Transport Eulerian
Model) models were developed to provide
a theoretical basis to investigate the
relationships between the emissions, the
atmospheric transport, chemistry and
removal processes, and the resultant
distribution of air pollutants and depo-
sitions. The development of an Eulerian
model, the STEM-I model, to describe the
transport/transformation and removal of
SOz and sulfate began in 1975. This
STEM-I model simulates the transport
and chemical processes of S02 and
sulfate in three-dimensions and makes
use of a set of 27 reactions to describe
SOz gas phase photo-oxidation.
Work began in 1980, with NASA
funding, to extend the STEM model to
include a more detailed treatment of
NxOy, hydrocarbon, and HxOy species and
wet removal processes. This activity, and
subsequent work with Pacific Northwest
Laboratories related to wet removal, led
to the development of the STEM-II model,
which currently treats some 60 chemical
species. The important atmospheric
processes are incorporated into the
model using chemical, dynamical and
thermodynamical parameterizations
having sufficient detail to accommodate
boundary layer-free troposphere
exchange in cloudy and cloud-free
environments, an in-cloud and below-
cloud wet removal and chemical pro-
cesses. The STEM-II model and its
components have already been used for
a variety of applications. Some of these
are: (a) transport of SO2/NOx/hydrocar-
bon urban plumes to the background
troposphere; (b) sulfate and nitrate
formation in the presence of land/sea
breeze; (c) effects of in-cloud and below-
cloud scavenging on homogeneous gas-
phase chemistry, (d) detailed analysis of
-------
mixing-limited chemical reaction of NO
emissions; and (e) the investigation of
sulfate production in orographic storms.
Many other applications are currently
ongoing and some will be discussed
laterm this report. Also, the STEM-II
model has been applied to problems on
various spatial scales utilizing horizontal
grid sizes from 80 km down to 5 km.
In 1984 the U.S. EPA initiated a
Mesoscale Acid Deposition Study. This
study, involving scientists from EPA,
Brookhaven National Laboratory, NASA,
Meso, Inc., and the Universities of Iowa
and Kentucky, was organized to assess
and understand the interactions and
importance of local emissions in meso-
scale precipitation scavenging. The study
consists of a field component, diagnostic
calculations, and the application of a
combined mesoscale meteorological and
chemical modeling framework (called
MesoSTEM). The STEM-II model is being
used in this project. This report summar-
izes our activities during the two year
period May 1985-May 1987. These
activities have focused on the linkage of
the dynamic mesoscale meteorological
model (i.e., the MASS (Mesoscale
Atmospheric Simulation System)) and
the STEM-II model, the testing of the
MesoSTEM model, and the diagnostic
analysis of field data collected on May
2nd and May 3rd, 1985 around Phila-
delphia, PA.
Overview of STEM-II Model
The diagnostic analysis in this project
is being carried out by use of the STEM-
II model. The STEM-II model is an
Eulerian combined transport/chemis-
try/removal model which treats chem-
ical species in the gas, cloud, rain and
snow phases. Thirty-nine species are
advected, while 21 species are short-
lived and are modeled using pseudo-
steady state methods.
The STEM-II model is structured to
treat wet removal processes in detail. A
schematic of the processes that are
included in the model analysis is pre-
sented in Figure 1. In this model, the in-
cloud scavenging of sulfate aerosol is
presumed to occur totally by nucleation.
Both cloud and aerosol size distributions
are assumed to be monodispersed. Once
the cloud forms, the number of sulfate
aerosols activated and dissolved to form
aqueous sulfate is assumed to be the
same as the number concentration of
cloud droplets. Rain drops are also
assumed to be monodispersed. Secon-
dary sulfate is treated in the same way
as primary sulfate and is allowed to be
reinjected into the gas phase as hydro-
meteors evaporate.
In this project the cloud parameters are
calculated by the Advanced Scavenging
Module (ASM) cloud scavenging model.
The ASM model treats the cloud micro-
physics, and calculates the cloud and
precipitation distributions, and intercon-
version rates. ASM calculates a self
consistent and realistic meteorological
data set of the simulated storm. The
precipitation formation processes are
treated in a parameterized fashion. The
scheme lumpswater into fourcategories:
water vapor, cloud water, rain and snow.
The cloud related parameters calculated
in ASM are transferred to the STEM-II
model where the transport/chemistry/
removal of pollutants is calculated.
For the gas phase, the reaction mech-
anism includes 85 reactions and 60
chemical species. Of these species, 39
long-lived species are advected while the
remaining 21 short-lived species such as
free radicals are modeled using the
Pseudo-Steady State Approximation.
For the liquid phase, 17 equilibrium
dissociation reactions, 3 chemical kinetic
expressions, and 22 ionic species are
included in the STEM-II code. Detailed
solution equilibria and chemical kinetics
are solved using a modified semi-implicit
method while satisfying the electrical-
neutrality constraint. The mass transfer
associated with absorption into and
desorption from droplets is also included.
Application of STEM-II to
May 2nd and 3rd Field Data
The STEM-II model described above is
being used to analyze field data collected
in the U.S. EPA Mesoscale Acid Depo-
sition Study. The model was imple-
mented on the VPS 32 machine at the
NASA Langley Research Center. The
current application work is focusing on
the May 2nd and 3rd, 1985 study
conducted in the Philadelphia area.
The modeling domain used for this
analysis is shown in Figure 2. Two
different grid systems are being used: a
coarse grid with horizontal grid spacings
of ~ 56 km; and a fine grid with horizontal
spacings of 20 km (see Figure 3). The
vertical domain is the same in each
system with 14 vertical grids non-
uniformly spaced between the surface
and a height of 6 km. The wind, temper-
ature, and surface precipitation fields
predicted by the dynamic meteorological
model MASS. The predicted precipitation
fields were used as inputs to the ASM
cloud model to generate the microphys-
ical parameters for the wet removal
calculations.
The emission fields for the simulations
are based on Version 5.2 of the U.S. E PA
National Acid Precipitation Assessment
Program (NAPAP) which includes SOX,
NOX and speciated hydrocarbons. The
point and area sources gridded to 20 km
are shown in Figures 4-8. Elevated point
sources are assigned to vertical grids by
distributing the point emissions over a
control volume centered around the grid
point. Presented are point source and
area emissions for NO, SO2, NHs, sulfate
and toluene. The spatial distributions of
the NO, SOZ, sulfate and toluene emis-
sions are similar with the major point
sources in the northeast section (around
New York) of the model domain and
around Philadelphia. The area sources
show a similar pattern but with appre-
ciable values at each grid point. The NO
area emissions are very large exceeding
1.1 x 1013 molecules/m2-s at several
locations. The area emissions of ammo-
nia are quite different from the other
species. As shown, there are very low
ammonia emissions in New Jersey and
relatively large emissions on the Penn-
sylvania side of the model region.
The May 2nd event was an example
of a slow moving frontal system which
produced very heavy precipitation in the
study area. The flow field in the modeling
region at 14:00 LT (local time) was
generally from the east and gradually
shifted to the northeast after 22:00 LT.
The level one wind speeds were high
throughout the period with horizontal
winds of 15 m/s. The precipitation came
from the south and reached Philadelphia
at 15:50 LT. Output from the MASS
model was used to provide the meteor-
ological inputs for the STEM-II model.
The measured wet deposition loadings
of nitrate and sulfate are shown in Figure
9. The general features show that for
both species the loadings were higher
on the Pennsylvania side. Furthermore,
the nitrate loadings were slightly higher
than those for sulfate on the New Jersey
side.
Coarse Grid Simulations
Coarse grid model predictions have
been obtained for an 18 hr period
corresponding to 14:00 May 2nd to 8:00
May 3rd, 1985. The model was initialized
by starting with rural conditions and then
seasoning under no-flow conditions with
real emissions for a period of 2-3 days.
-------
Autoconversion
Accretion
Absorption
Desorption
Evaporation
Rain Phase
3 Reactions
Transport (Convection)
Cloud Phase
3 reactions
Transport (Convection
and Diffusion)
Riming
o
I
o
I
Uj
Gas Phase + Aerosol
Homogeneous Chemistry
85 Reactions and 60
Chemical Species
Dry Deposition
transport (Convection
and Difusion)
Adsorption
Melting
Freezing
Snow Phase
Transport (Convection)
Meteorological Data Set
* Liquid Content
* Temperature
* Wind Velocity
* Eddy Diffusivity
* Interconversion Rates
* Air Density etc....
Figure 1. Interaction diagram of chemical/physical processes treated in the STEM+ASM model.
Initialization of the model is difficult
because of the small domain size and
large horizontal velocities. The model
was initialized under no-flow conditions
so that the initial conditions would reflect
the source characteristics of the model
domain. Since no measurements were
available to set the inflow boundary
conditions, the chemical composition of
the air masses transported into the model
region (through the eastern and northern
boundaries) was estimated by perform-
ing 1-dimensional model calculations
reflecting the air mass histories. For
example, the air entering through the
eastern boundary has relatively low
concentrations reflecting that it origi-
nated from the south and traveled along
a wet trajectory, whereas the air mass
entering the northern boundary has high
concentrations of pollutants reflecting
the high source regions upwind in New
York. The results of the coarse grid
simulation are discussed in a paper
presented at the 16th International
Technical Meeting on Air Pollution
Modeling and Its Applications.
Fine Grid Simulation
Results using a fine grid of 20 km
horizontal mesh spacing for the May 2nd
and 3rd, 1985 case have also been
obtained. The fine grid domain is shown
in Figure 3 and covers a region 180 km
x 180 km x 6 km with 1400 grid points
(i.e., 10x10x14). The emissions inventory
-------
NO Emissions
Figure 2. Philadelphia mesoscale acid
deposition study domain. The
shaded regions denote the
areas where the monitoring
stations were deployed. The
large slanted square repre-
sents the modeling region.
Figure 3. Model region for MesoSTEM
application. Coarse grid and
fine mesh grid systems are
shown Solid dots show grid
points (8,2) and(4.8) in the New
Jersey and Pennsylvania
regions, respectively.
was discussed earlier and plotted for
selected species in Figures 4 to 8. The
initial conditions of the primary pollu-
tants (e.g., NO, NO2, and S02) showed
large concentrations around New York
City and Philadelphia areas. The initial
conditions of ammonia reflected the
large area sources on the Pennsylvania
side of the model domain. Ground level
NHa concentrations exceeded 5 ppb in
much of this region. The initial concen-
molecls. m sec
.1E+13<0
.5E+12<0<.1E+13
1E+12<0<5E+12
.JE+1KO<1E+12
NO Emissions(P)
molecls. m sec'1
.2E+09
-------
50s Emissions
molecls. m sec
.5E+12<0
,2E+12<0<.5E+12
.5E+1KCK.2E+12
.5E+10<(X.5E+11
SOi Emissions(P)
.SE+12<0<.1£+13
.1E+12<(X.5E+12
.1E+1KO<.JE+12
Figure 5. Elevated point sources (at level 2 corresponding to a height 200 mi and area
emissions for SOi.
The uniformity in the wet deposition is
reflected by the situation that the gas
phase concentrations of nitric acid and
sulfate in the high source regions are
only 20-30% higher than those in the
rural areas in spite of the fact that the
concentrations of the primary pollutants
are 5-10 times higher.
Further details of the deposition at the
New Jersey measurement sites and the
Pennsylvania sites can be seen by
examining the deposition as a function
of time at grid points (8,2) and (4,8) (see
Figure 3). The time profiles of surface gas
phase and rain phase concentrations of
selected species are presented for
comparison in Figure 14. The New Jersey
site is an area of low NOx and SOX sources
and for the meteorological conditions on
May 2nd the wind direction is from the
east; thus, relatively clean air is passing
to this site through the eastern boundary.
Precipitation occurs at this grid point for
the first 16 hours of the event. The rain
phase concentrations at this grid point
show that the soluble pollutants are
removed rapidly by the initial rain and
that the bulk of the deposition loading
occurs in the first hour of the rain event.
The liquid phase concentrations of
nitrate and H+ build up beginning at 5:00
LT on May 3rd. This corresponds to the
time that the wind shifts to a northeast-
erly flow transporting the polluted air
mass from New York.
The Pennsylvania site (grid point (4,8))
is located in a region with high SOX and
NOx emissions. The site is also downwind
of major sources throughout the mete-
orological conditions of May 2nd, 1985.
Precipitation began at this site at 17:00
LT May 2nd and continued until 4:00 LT
on May 3rd. The time profiles of liquid
phase concentrations at this site are also
presented in Figure 14. The time profiles
are quite different than those at the New
Jersey site. The liquid phase
concentrations of most species reach
maximum values at 19:00 LT. The sulfate
and nitrate concentration maximums are
about 150//M. The liquid phase concen-
tration profiles of H2O2 and bisulfite show
opposite trends. The H20z is high initially
and decreases as time progresses. This
indicates that at early times H2O2 is
present in excess of bisulfite, and the
bisulfite is consumed by the reaction with
H202 to produce sulfate. At later times
the H202 is depleted from the system and
the bisulfite concentration increases.
The gas phase profiles (not shown)
indicate that gaseous H2O2 is depleted
from the area by 22:00, whereas gaseous
S02 is maintained by local emissions.
The total deposition at the two sites
is also plotted. The total H+ deposition
is about 1500jumole/m2atthe Pennsyl-
vania site and ~ 1000 /i/mole/m2 at the
New Jersey site. The nitrate deposition
exceeds the sulfate deposition at each
point.
Budget calculations for various pro-
cesses over the modeling domain for the
18 hour simulation are presented for
HNOs, sulfate and S02 in Figure 15.
These results provide useful insights into
the relative importance of various pro-
cesses. The SO2 inventory shows that
local emissions during the 18 hour period
is equivalent to about 1/3 of the mass
of SOa initially present. (Furthermore, the
net advection (i.e., inflow-outflow) into
the region is equivalent to about 1 /3 of
the amount emitted.) The wet deposition
and dry deposition are of equal impor-
tance and together are equivalent in
magnitude to the emissions. The wet
removal processes consist of wet depo-
sition of S(IV) at the surface and the
conversion of S(IV) to sulfate in cloud and
rain water. As indicated, the importance
of liquid phase transformation of S(IV) is
much more important that the gas phase
reactions of S02. The predominant liquid
-------
Ammonia Emissions
molecls. m 2 sec"1
.1E+12<0
.5f+/7<0<./£+/2
.1E+1KCX.5E+11
Ammonia Emissions(P)
molecls. m 3 sec
Figure 6.
.1E+08<£<.2E+08
./£+07<£<./£+OS
.1E+06
-------
molecls. m sec
.1E+12<0
.5E+1KO<.1E+12
.1E+1KO<.5E+11
molecls. m'3 sec 1
.2E+07
-------
Toluene Emissions
Toluene EmissionsfP)
molecls. m 2 sec"1
.1E+)2<0
• 5E+1KCX.1E+J2
.JE+JKCX.SE+II
.1E+JO<0<.1E+J)
molecls. m 3 sec"1
.2E+07<£
.1E+07<£<. 25+07
.1E+06
-------
402
857
. 1180
(a)
2726
Storm 0503
H* Deposition
fjmol/m2
1392
' 1226
1503 '089
1547 '600
I™2- 1528 1217 *§L
'
0503
NOl Deposition
1462i ^ 0503
SOJ Deposition
g52 502 \ fJmol/m2
20km
652
Philadelphia
^^^ ^ 725 655
cC es7 71>2' ™
Cv" "
599
Figure 9. Measured total deposition of (a) H*. (b> HN03, and (c) SO4°. The underlined values
are for manually deployed bottle-funnel collectors and the others are for automatic
wetfall collectors.
-------
(a)
ppb
W.
-------
(c)
ppb
10.
-------
(a)
ppb
8.
-------
ppb
A
-------
umole m~
1000.
-------
(jmole m'2
300.
-------
D = Precip. Kate #5 mm hr 1
0 = Total Precip. mm
O = S04
A = HNO3
+ = Bisulfite
X =
D = Precip. Rate ttb mm hr
0 = Total Precip. mm
(d)
Legend
n = hT
o = so*=
A = HNOs
= Bisulfite #5.
X =W2O2 #5.
O = S04=
A = HNOa
+ = Bisulfite
x =
Figure 14. (a) Precipitation field at grid (8,2). (b) precipitation field at grid (4,3), (c) rain water concentrations at grid (8.2), (d) rain watt
concentrations at grid (4.8), (e) cumulative deposition at grid (8,2), and (f) cumulative deposition at grid (4,8).
1 6
-------
LZJ Total Inventory at t = 0
ESI Emission
17Z2 Advection
U\S Dry Deposition
$M Wet Deposition
LU Gas Phase Chemical Reaction
CvJ Liquid Phase Chemical Reaction
fa)
c *-:
Q •»
^
,, 0
O m
The amount of HN
0.0 J 0 2.0
"A
f
O
The amount o
ro 1.0 2.
I
f
f
1
g
Figure 15. Budget calculations for (a)
HNO3, (b) sulfate and(c) SO*
Gregory R. Carmichael is with Department of Chemical and Materials
Engineering, University of Iowa, Iowa City, I A 52242; and Leonard K. Peters
is with Department of Chemical Engineering, University of Kentucky,
Lexington. KY 4O506.
Francis S. Binkowski is the EPA Project Officer (see below).
The complete report, entitled "A Mesoscale Acid Deposition Model: Preliminary
Applications and a Guide for User Interface," (Order No. PB 87-227 658/
AS; Cost: $18.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Officer can be contacted at:
Atmospheric Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
1 7
-------
m
T O
o> =*
Q) O
5? 3 $-
8 *
00
J- CD
S8
CD
CD
CO
O
O
o
00 <
wsw
rm
m
*"*
O
o».
o
> m c
§' a
T)
O
CD
O
"
S. O =?
O 3 m
i 3
•^ 5'
01 o
fO 3
£ 3
00 CD
3
*•+
51
3D
CD
(A
O
c>
CD
3»'« 2
"" *»' "
------- |