X-/EPA
United States
Environmental Protection
Agency
Environmental Sciences Research"
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-81-102 Aug. 1981
Project Summary
ENAMAP-1 Long-Term
SO2 and Sulfate Pollution
Model: Further Application of
Eastern North America
C. M. Bhumralkar, R. L Mancuso, D. E. Wolf, K. C. Nitz, W. B. Johnson, and
T. L Clark
Under contract to the U.S. Environ-
mental Protection Agency (EPA), SRI
International (SRI) developed and
evaluated an Eastern North American
Model of Air Pollution (ENAMAP)
(Bhumralkar et al., 1980). The
ENAMAP-1 * model, which is a modi-
fied version of the SRI-developed
European Model of Air Pollution
(EURMAP) t was specifically designed
to study long-term transboundary air
pollution processes over eastern North
America. The model can be used to
calculate monthly, seasonal, and an-
nual values of sulfur concentrations
and depositions and to quantify inter-
regional exchanges of airborne sulfur
between various selected Canadian
and EPA regions. ENAMAP-1 has
been shown to be highly suitable for
application to assess the long-term
transboundary sulfur pollution prob-
lem in eastern North America, because
of its realistic treatment of precipita-
tion scavenging and wet deposition,
and its consideration of both SOa and
SO! emissions over a very large region.
"SRI is currently developing an Improved version of
ENAMAP-1 which will be designated ENAMAP-2.
This will include, among other things, effects of
' complex terrain and emissions released at higher
elevations.
tEURMAP-1 was developed by SRI under the
sponsorship of the Environmental Agency
(Umweltbundesamt) of the Federal Republic of
Germany (FRG). For a detailed description of this
model, see Johnson et al. (1978).
This report describes the results of a
study funded by EPA with the objec-
tives of applying the ENAMAP-1
model to further test the model and to
study the variability of the model's
seasonal calculations of sulfur con-
centrations and depositions due to
year-to-year changes in the wind and
precipitation patterns. Section 2 of
this report reviews the basic structure
of the ENAMAP-1 model with respect
to model grid boundaries and other
variables. Section 3 presents a review
of the data bases, including the air-
quality, emission, and meteorological
data used with ENAMAP-1. The origi-
nal report discusses the monthly and
annual results obtained from ENAMAP-
1 using weather data for the four years
1975, 1976, 1977, and 1978 and
sulfur emissions data for 1977. Al-
though the original report discusses
seasonal and annual variations in SO2
and SO! concentration and deposition
patterns, as wall as seasonal variations
in concentration and deposition con-
tributions from large emission regions
to the same receptor regions, only the
annual and seasonal variations in the
SO! concentrations patterns are dis-
cussed in this report.
Thi$ Project Summary was devel-
oped by EPA's Environmental Sciences
Research Laboratory, Research Tri-
angle Park, NC, to announce key find-
-------
ings of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction and Review of
the ENAMAP-1 Model
ENAMAP-1 is a practical air pollution
model designed to have minimum com-
putational requirements for use in
making long-term calculations econom-
ically, while at the same time offering
acceptable realism in simulating the
most important processes involved in
the transboundary air pollution problem.
The ENAMAP-1 model can be used to
calculate monthly, seasonal, and annual
SC>2 and 864 air concentrations; S02
and SO< dry and wet deposition patterns;
and interregional exchanges resulting
from the S02 and SOI emissions over
eastern North America. The model uses
long sequences of historical meteoro-
logical data as input, retaining all the
original temporal and spatial detail
inherent in the data (Bhumralkar et al.
(1980)).
In the ENAMAP-1 model, discrete-
puffs of S02 and SOI are assumed to be
emitted at equal time increments from
cells of an emission grid. This type of
treatment provides a realistic repre-
sentation of area sources. For a point
source, it assumes that the pollutant
expands initially to fill uniformly the
volume of the cell from the point within
the cell where the source is actually
located. [In this application, seasonal
emission data were conveniently avail-
able on an 80- by 80-km Universal
Transverse Mercator (UTM) grid.] For
each of the emission cells, the average
annual or seasonal emissionsaredivided
into discrete emission puffs released at
12-hour intervals and tracked at 3-hour
time steps, until either they move
outside the region of analysis or their
mass drops to an insignificant level (10
tons of SO2 and 1 ton of SOI). The
individual puffs are transported accord-
ing to a transport wind field that is
derived objectively from the available
upper-air wind observations.
Since diffusion on the regional scale
is not as significant as the transport and
removal processes, very simple treat-
ments of vertical and horizontal diffu-
sion have been used. Upon release,
each puff is assumed to undergo instan-
taneous vertical diffusion to give a
uniform concentration in the layer
between the surface and the top of the
mixing height. Horizontal diffusion is
treated by allowing the area of the puff
to increase linearly with time on the
basis of Fickian diffusion, assuming a
horizontal eddy diffusivity of 36 km2rf1.
During the transport of the puff, the
model assumes that the pollutant con-
centration within a puff is always
uniform.
The amount of pollutant mass that is
removed from a puff during each 3-hour
time step is dependent on the specified
dry and wet deposition rates that are
used; these amounts are deposited
within the appropriate 70- and 70-km
cells of the receptor grid. At each time
step, a fraction of the S02 is transformed
to SOI at the specified transformation
rate. Figure 1 shows the eastern sector
of the North American continent over
which the ENAMAP-1 model has been
applied. This sector covers the region
between 30°N and 50°N latitudes and
65°W and 105°W longitudes. Figure
1(a) shows the EPA regions and sub-
regions used in this study; southern
portions of Quebec and Ontario prov-
inces of Canada are also included.
Figure 1(b) shows the model receptor
grid; each receptor-cell is 70 by 70km.
The pollutant depositions are accumu-
lated and concentrations are averaged
in these receptor cells. The values for
the basic model elements that have
been used are listed in Table 1. These
values are based on reviews of recent
field, laboratory, and theoretical studies
and on evaluation studies (Mancuso el
al. [1978]).
In this study, the basic model was rur
for the months of January, April, July
and October of the years 1975, 1976
and 1978 and January, April, August
and October of 1977 using the meteoro-
logical data for each of the appropriate
months. The emission data of 1977
were assumed to apply for all four years,
mainly because no data base of similar
quality and resolution was available for
the other years. The particular months
were selected in order to examine the
seasonal variations in the results. [In
1977, August rather than July was
chosen as representative of the summer
because of the availability of a greater
amount of Sulfate Regional Experiment
(SURE) air quality data.] For each of the
four months of each of the years, fields
of S02 and SOI concentrations, dry
depositions, and wet depositions result-
ing from the S02 emissions were calcu-
lated, stored, and displayed graphically.
Interregional exchange tables were also
generated.
Annual depositions for each of the
years were estimated by assuming thai
the results for each of the four months
were representative of seasonal values,
totalling the four monthly deposition
values, and multiplying by three. Simi-
larly, estimates of annual average con-
centrations were obtained by averaging
values for the four months of that year.
Tablet.
Element Values Used in the ENAMAP-1 Application to Eastern
North America
Element
Values
Emission rate
Transport windspeed (V) fms'1) and
direction
Mixing height (km)
h = h0- f A»
SOz deposition rates (hr'^J
Dry
Wet
SOI deposition rates (hr'^J
Dry
Wet
SO2/SOI transformation rate fhr'1)
Data provided by SURE and NEDS
Derived by integrating winds over
boundary layer using upper-air wind
data
h0=1.3
£ = 0.15
0.037
0.28W
0.007
0.07/?f
0.01
*A = +1 in winter. -1 in summer, and 0 in spring and fall.
t/? is the precipitation rate in mm /v~1.
-------
South Quebec
VIII-
South
fa) EPA Regions Used in This Study
40 43
(t>) Emission Grid and Model Domain
Figure 1. Eastern North American modeling domain and emission/receptor
regions used in this study.
Review of the Data Bases
The ENAMAP-1 model uses three
types of data—meteorological (winds
and precipitation), emission (862 and
SOI), and air quality (S02 and SO«
concentration measurements). The
main purpose of this study was to
determine the effects of seasonal varia-
tions in meteorological patterns on the
ENAMAP-1 calculations. Therefore,
actual meteorological data for the years
1975 through 1978 have been used,
with the emission data for the year
1977, permitting a direct determination
of the effects of weather on the results.
However, the calculated results are
thus strictly correct only for the year
1977, and the comparisons with the air
quality data are most valid for that year.
A detailed description of the data bases
is given by Bhumralkar et al. (1980); a
brief review is given below.
Meteorological Data
Historical meteorological data for this
study (upper-air wind data for the
United States and precipitation data for
the United States and Canada) were
obtained from the National Climatic
Center (NCC) in Asheville, North Carolina.
The basic analyses were made with a
computer program that generated both
transport winds and precipitation
amounts at 3-hourly intervals for the
70- by 70-km ENAMAP-1 receptor grid.
The upper-air wind data set consisted of
the 12-hourly observations from about
50 sites in the U.S. The precipitation
data set included hourly data from about
2,000 U.S. stations and six-hourly data
from nearly 200 U.S. and Canadian
sites; the analyzed values were expressed
as rainfall rates in mm/nr and were
used directly in the wet deposition
calculations.
Emission Data
Emission data have been collected for
a number of years and maintained by
the National Emissions Data Systems
(NEDS) of the EPA. The NEDS data cover
the entire U.S. portion of the ENAMAP-
1 study area and provide relatively high
spatial resolutions. Specialized emission
data also have been compiled for the
SURE program. These data are complete
for sources existing in July 1977 and
effectively represent emissions for the
1977 time period. This SURE emission
data base is essentially a refinement of
the NEDS data base; that is, the NEDS
data were updated and screened for
errors and inconsistencies, especially
with respect to electric power plants.
-------
The 1977 seasonal emission data
base that was used in this study was
based, wherever possible, on the spe-
cialized emission data of the SURE
program. The SURE data, which are
arranged on an 80- by 80-km UTM grid,
cover almost the entire area considered
in this study except for Texas and the
Great Plains states. For this western-
most portion of the ENAMAP-1 domain,
gridded emission data were supplied by
NEDS.
Air Quality Data
Two sets of air quality data were used
for ENAMAP-1 model validation: Electric
Power Research Institute - Sulfate
Regional Experiment (EPRI-SURE) and
the EPA Storage and Retrieval of Aero-
metric Data (SAROAD). The former set
contained hourly averaged S02 and
daily averaged SOI concentrations for'9
sites in the northeastern quadrant of the
United States for the August 1977 to
October 1978 period. For August and
October 1977 and January, April, July,
and October 1978, the same type of data
were available for an additional 45 sites
in the same region. These 54 rural
monitoring sites were free from local
effects. The latter set of data contained
daily averaged SO2 concentrations from
many more sites. Data from sites thought
to be influenced by local sources were
ignored. This data set contained daily
averaged SOI concentrations, from
about 70 monitoring sites. However, the
typical frequency of monitoring was
three days per month, which meantthat
the monthly averaged SOI concentra-
tion at each site was rather poor.
These air quality data were used to
determine monthly averages for 140 by
140-km grid squares (2 by 2 receptor
grid cells of 70-km dimensions). These
averages were compared to those com-
puted by the model.
Annual and Seasonal
Variations in SO4
Concentration Patterns
The calculated and measured SOI
concentrations (pig/m3) patterns for
January 1975 - 1978; July 1975 and
1976; August 1977; and July 1978 are
depicted in Figures 2-9, respectively.
August 1977 was chosen as a modeling
period since intensive SURE air quality
data were available for that month.
Monthly-averaged wind fields accom-
pany the concentration analyses. The
• concentrations depicted in the figures
represent a spatial average of monthly
averaged concentrations across 140 by
140 km grid squares.
January 1975 - The calculated SOI
concentration pattern for January 1975
[Figure 2(a)] is a very interesting one. A
strip of higher SOI concentrations (>8
/ug/m3) extends across the Northeast
and upper Midwest, strongly reflecting
the anticyclonic pattern of the wind
field. This SOI field shows relatively low
values in the central Illinois-Indiana-
Ohio area, as did the S02 field; however,
the low SOI concentrations also extend
across Illinois into the western Kentucky-
Tennessee area.
Comparison of the calculated and
measured SOI concentrations for Janu-
ary 1975 [Figure 2(b)] is difficult because
of the lack of an extensive monitoring
network. However, where there are
data, the comparison is generally very
good. For example, the strip of high
calculated values mentioned above,
which extends across the southern tips
of the Great Lakes and across Pennsyl-
vania, appears to be depicted by the
measured data. The measured data
show relatively low values in the south-
eastern and northeastern United States,
in agreement with the calculated re-
sults. A few measured values in the
western part of the domain are relatively
high and inconsistent with the calculated
values; however, these measurements
are probably associated with local
sources.
January 1976 - The calculated SOI
concentration pattern for this period
[Figure 3(a)] shows a center of high
values (>8 (ig/m3) over the northeastern
United States, while west of the Missis-
sippi, the concentrations become quite
low «2 fjg/m3). The high values in the
northeast of the United States indicate
that there was a transport of the pol-
lutant into this area, although this is not
distinctly shown by the mean wind
pattern of Figure 3(c). The calculated
SOI concentrations for January 1976
[Figure 3(a)] and 1975 [Figure 2(a}] show
definite differences. The SOI concen-
trations in New England were relatively
low in January 1975, apparently because
of stronger transport winds over this
area. Also, the January 1976 pattern
does not show as much variability in
SOI concentrations as were produced in
the Southeast during January 1975 in
association with the anticyclonic circu-
lation.
The SOI monitoring sites in January
1976 were again sparse. The available
data [Figure 3(b)] appear to be generally
consistent with the calculated concen-
trations. A large value (15 /ug/m3) in
southern Michigan and a large value (11
//g/m3) in North Carolina are incon-
sistent with the calculated results.
January 1977 - The calculated S0«
concentrations for January 1977 [Figure
4(a)] show that the higher SOI concen-
trations (>8 //g/m3) are centered over
the northeastern United States as in
1976; however the "8" isoline extends
farther west, apparently because of
differences in the transporting winds
that are not evident from the monthly
mean fields. The high SOI concentra-
tions off the East Coast in 1975,1976,
and 1977 are a reflection of a prevailing
wind blowing from northwest to south-
east.
The number of available measured
SOI concentrations for January 1977
[Figure 4(b)] provide a more desirable
coverage. The measured data show
high values in the northeastern United
States that are consistent with the
calculated values (based on 1977 emis-
sion data). However, some high mea-
sured values (up to 10/ug/m3) that have
been recorded west of the Mississippi
are probably unrepresentative data.
January 1978 - The calculated SOI
concentration pattern for January 1978
[Figure 5(a)] is very similar to that for
January 1977 [Figure 4(a)]. However,
the 8 fjg/m3 isoline does not extend off
the coast into the Atlantic apparently
because the wind did not blow sulfur
pollution off the coast as frequently
during this January period as it did in
1977. The available measured data for
this period [Figure 48
pg/m3) is displaced to the southeast,
lying principally over the Virginia area
with a strip of relatively high values
extending south westward into Georgia.
-------
"\ vr\;./£.$'2?8...9..8,
-f >:--••<; \ ?°0..?
i i \ ;.,,"-.'V.............
i \ 1 5."-' 5
'"- ] /"i r"\"--«../
1 J ! 5; \ /
V .' I S >''
-! t... \.....A ?
....
Local Maximum Values Shown apply at Points Marked by Plus
Signs
10 m/s
Figure 2. SOZ concentrations i/jg/m3) for January 1975. (a) calculated, (b) measured, (c) mean monthly transport winds.
-------
Local Maximum Values Shown apply at Points Marked by Plus
Signs
Figure 3.
10 m/s
SO* concentrations ifjg/m3) for January 1976. (a) calculated, (b) measured, (c] mean monthly transport winds.
-------
(a) .
Local Maximum Values Shown apply at Points Marked by Plus
Signs
.88
4. SO« concentrations (jjg/m*) for January 1977. (a) calculated, (b) measured, (c) mean monthly transport winds.
-------
/.oca/ Maximum Values Shown apply at Points Marked by Plus
Signs
(c)
-------
Z.oca/ Maximum Values Shown apply at Points Marked by Plus
Signs
10 m/s
Figure 6. SOt concentrations (fjg/m3) for July 1975. (a) calculated, (b) measured, (c) mean monthly transport winds.
-------
(b)
'",•""**.,„
I >.
J.w»"*«-»w.-t »•... . ...Jw ...
* ...-' 7>*
*_»•*». ••*•
\
• •
! •
13
Y
?;
5 :
Aoca/ Maximum Values Shown apply at Points Marked by Plus
Signs
Figure 7.
10 m/s
concentrations (jjg/m3) for July 1976. (a) calculated, (b) measured, (c) mean monthly transport winds.
10
-------
Local Maximum Values Shown apply at Points Marked by Plus
Signs
10 m/s
Figure 8. SOI concentrations (fig/m3) for August 1977. (a) calculated, (b) measured, (c) mean monthly transport winds.
11
-------
(b)
.••»••••••*••• I _
5 %--4,>/-\<
1 I * -* dV
• r •-.. -«~—a « \
•-*/ •• •••./^?J:-»x ?e ; \
'4 '4..6 -y r' ) rt sfci ,
-1—A. i\ *vw«M.:^
\ T \7/?^'8«~(''0*N V7./V-
V«»*«»»*^ -*o \//182.1 \11 13*^0
\ *y~"~ 's*
—y-n^Y
Local Maximum Values Shown apply at Points Marked by Plus
Signs
10m/s
Figure 9. SO4 concentrations dig/m3) for July 1978. (a) calculated, (b) measured, (c) mean monthly transport winds.
12
-------
This can be attributed to the stronger
winds [Figure 7(c)], which would have
moved the emissions in this direction.
The measured SOJ concentrations for
January 1976 [Figure 7(b)] are also
incomplete, but do not appear to show
the same southeastdisplacementasthe
calculated values. In the measured data,
the major SOJ pollution appears to be
located over the highly SOz-emitting
Pittsburgh area. Also, as was noted for
July 1975, the relatively high measure-
ment of $64 values suggests either that
the calculated values are slightly low or
that the SO* measurements are some-
what biased by local 804 sources.
August 1977 - The calculated SOI
concentrations for August 1977 [Figure
8(a)] show a distinct southwest-to-
northeast elongation of the concentra-
tion isolines. ENAMAP-1 appears to
have performed quite well in calculating
these S0< concentrations for this sum-
mer period, since they compare very
favorably with the measured values
[Figure 8(b)]. The higher SOS concen-
trations (>8 /ug/rn3) occur over the
northeastern United States in both the
calculated and measured fields, with
peak values (17 fjg/m3) occurring near
the Pittsburgh area. A region of low
concentrations is indicated in both the
calculated and measured fields over
eastern Kentucky, although the mea-
sured concentrations are somewhat
lower than the calculated.
July 1978 - The calculated SOI con-
centration pattern for July 1978 [Figure
9(a)] is similar to that for 1977 [Figure
8(a)j. The available measured data for
this period [Figure 9(b)] show a pattern
that is reasonably consistent with the
calculated one. However, the measured
SO* values appear to be significantly
higher than the calculated, possibly
because of an increase in the sulfur
emission in 1978 that was not accounted
for in the model calculations.
Conclusions
In this study, the SRI-developed
ENAMAP-1 model was applied to deter-
mine the significance of seasonal and
yearly variations in meteorological
patterns on the ENAMAP-1 model cal-
culations. To achieve this, the sulfur
emission data for 1977 were used with
meteorological data for four different
years: 1975, 1976, 1977, and 1978.
Model calculations were made for the
monthly and annual SO2 and SO* con-
centrations, depositions, and regional
exchanges. The calculated results were
generally consistent with the air quality
measurements, although discrepancies
may have been caused by the use of
1977 emissions for all four years and by
use of certain measurements that were
unrepresentative of average values
within receptor cells. The yearly varia-
tions in the meteorological data were
found:
• To produce changes in the S02S04
concentrations and depositions
that were consistent with the
changes in transport winds and
precipitation ambunts.
• To be most noticeable in the monthly
SOS concentrations and SOI wet
depositions; the latter being sensi-
tive to yearly variations in both the
boundary layer wind and precipita-
tion amounts.
• To have little effect on the monthly
S02 concentrations, which princi-
pally depict the high emission
areas.
• To have little effect on the annual
fields, since the results for a given
year were derived by averaging the
results for January, April, July, and
October of that year.
As noted in this study and its prede-
cessor (Bhumralkar et al., [1980]), there
are some differences between the
calculated and measured results, par-
ticularly in regard to seasonal and
latitudinal variation. These differences
appear to be partially caused by the
imperfect simulation of mixing height
and the vertical growth of puffs. Also,
the neglect of terrain influences is
noticeable in the Appalachian region'.
An improved version of the model
(called ENAMAP-2) is being developed
with a view to mitigate these limitations.
References
Bhumralkar, C.M., R.L Mancuso, D.E.
Wolf, R.A. Thuillier, K.C. Nitz, and
W.B. Johnson, 1980: "Adaptation
and Application of a Long-Term Air
Pollution Model ENAMAP-1 to Eastern
North America," Final Report, Con-
tract 68-02-2959, SRI International,
Menlo Park, California.
Johnson, W.B., D.E. Wolf, and R.L.
Mancuso, 1978: "LongTerm Regional
Patterns and Transfrontier Exchanges
of Airborne Sulfur Pollution in Europe,"
Atmos Environ., Vol. 12, pp. 511 -527.
Mancuso, R.L, C.M. Bhumralkar, D.E.
Wolf, and W.B. Johnson, 1978: "Eval-
uation and Sensitivity Analyses of the
European Regional Model of Air Pol-
lution (EURMAP-19," Progress Report,
Contract LUP-411 515/IIIA315, /SRI
International, Menlo Park, California.
Perhac, R.M., 1978: "Sulfate Regional
Experiment in the Northeastern United
States: The SURE Program," Atmos.
Environ., Vol. 12, pp. 641 -647.
J3
-------
C. M. Bhumralkar, R. L Mancuso, D. E. Wolf, K. C. Nitz, and W. B. Johnson are
with SRI International, Menlo Park. CA 94025; T. L. Clark is the EPA author
and also the EPA Project Officer /see below).
The complete report, entitled"ENAMAP-1 Long-Term SO sand Sulfate Pollution
Model: Further Application of Eastern North A merica," (Order No. PB 81-213 217;
Cost: $12.50, subject to change) will be available only from:
National Technical Information Service
5235 Port Royal Road
Springfield, V'A 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
it \JS GOVERNMENT PRINTING OFFICE, 1981 — 757-012/7342
14
-------
II
Is
o
5'
I.
ro
m > •
dk,
o
I
------- |