PB92-152982
Sensitivity of RADM to Point Source Emissions Processing
(U.S.) Environmental Protection Agency, Research Triangle Park, NC
1992

-------
PB9 2-152982
EPA/600/A-92/035
SENSITIVITY OF RADM TO POINT SOURCE EMISSIONS PROCESSING
Daewon W. Byun
Computer Sciences Corporation.
P.O. Box 12767
Research Triangle Park, North Carolina 27709
Frauds S. Binkovvski1
Atmospheric Sciences Modeling Division
Air Resources Laboratory
National Oceanic and Atmospheric Administration
Research Triangle Park, North Carolina 27711
1. INTRODUCTION
The Regional Acid Deposition Model (RADM) and !
associated Engineering Model have been developed to.
study episodic source-receptor relationships on a regional
scale. The RADM includes transport, chemical
transformation, and deposition processes as well as input of
emissions into the vertical layers using a plume-rise
submodel. As wind speed and direction change with '
height, and as atmospheric turbulence varies significantly
with distance from the earth's surface, the model;
predictions can be affected considerably by the height at.
which tte emissions are injected. Atmospheric emissions
can be divided into area and point sources. The area
sources include mobile emissions, biogenic emissions, and
other emissions that are often related to the surface area of
the earth. Area sources are more difficult to control and
have inherent diversities in their physio-chemical
processes. Point sources include stacks at power'plants and
manufacturing facilities. Usually, point sources are the
target of emission reduction programs.
The comprehensive emissions system for RADM
includes processing of point sources, area sources, mobile
sources, and biogenic emissions. In order to reliably
evaluate effectiveness of future emission control strategies
in reducing acid deposition, quantification of the influence :
of emissions processing on the RADM predictions is
necessapr. One important part of point source missions
processing is the plume-fise computation. The effect of
using different plume-rise algorithms on the RADM's
prediction of the sulfur deposition needs to be investigated.
The objective of the present study is to compare two
different plume-rise calculation methods that are described
below.
2. FLUME-RISE SUBMODELS
Two types of plume rise estimation algorithms that
have been used for the RADM emissions processing are
compared here. Method-1 is based on the early formulas
by Briggs (1969) and uses wind and ambient temperature
data at stack height However, wind and temperature data
are available at a number of layers above the ground for the
RADM system. Thus, a plume-rise algorithm that uses
these additional meteorological data was developed.
Method-2 uses more recent plume rise formulas by
Briggs (1975, 1984) to replace earlier versions (Briggs,
1969). The procedure is similar to the one proposed by
Turner (1985), which is intended for use in Gaussian-type
plume dispersion, but has been somewhat modified for use
in a regional-scale model such as RADM. It requires point -
source stack parameters and meteorological input files
generated by a meteorological model, such as MM-4
(Anthes and Warner, 1978). The NAPAP point-source
stack parameters include: stack flow rate, stack height,
temperature of the plume, stack diameter and effluent
velocity, and stack identification code.
1. Os usigumtat to the Atmospheric Reicartb aad Exposure
Aisewaneal Laboratory, V.S. Environmenti! Protection Afcncy.
2.1 Meteorological Data Processing
The RADM's point-source plume-rise submodel utilizes
wind, temperature, and moisture profiles generated by the
MM-4 model. A meteorological preprocessor for RADM
estimates surface friction velocity (u») and sensible heat
flux based on the surface layer similarity theory. For the
unstable case, mixing height is estimated from the virtual
potential temperature profile. For neutral and stable cases,
dynamic boundary layer heights are estimated based on the
boundary layer theory. Estimates of wind speed and
ambient temperature at the sack top are interpolated based
on the surface layer similarity theory when stack top is
within the surface layer and simple linear interpolation
when it is located above the surface layer.
22 Plume-rise Calculation
Initial buoyancy flux (Ft) of the plume is computed
using the equation
hff,	if Tt>Ttt
Fb~0, '	UT,STa
(1)
where: Ft is the initial buoyancy flux
fis gravitational acceleration (m/s*)
\ is the temperature of the plume (K)
T, is the ambient temperature at stack height (K)
v, is the initial plume effluent velocity (mis)
d is the 'diameter of the stack at stack height (m).
Sine* buoyant plume rise is very sensitive to atmospheric
stability, a layer-by-layer plume penetration and plume rise
concept is utilized for calculating buoyant plume rise.
For neutral stability, Briggs' equation for buoyant
plume rise at break-up is:
th -1.2 [Ft /(u-uJ) ]se [hs * thp^
(2)
where: th is the plume rise above the stack (m)
v is the avenge wind speed at sock height (m/i)
kt is stack height (m)
An approximate solution of above, suggested by Briggs
(personal communication, April 23,1983), is used in actual
computation:
M « UfFrfvuJ) p* [ht+13FrfwuJ) P*
(3)
For stable cases, the new plume-rise algorithm defines
plume rise as the minimum of (a) Ah calculated by the
. above approximate equation for neutral stability and (b) M
! estimated by Briggs' equation for the rise of bent-over
1 plumes:
£Ji-Z6fFi,/(u's)]"3
e 06
where: s ¦ f- j- is the stability parameter.
For both neutral and stable cases, if the projected plume
; rise is found to be located at the layer above the top of the
current layer, the rise is limited to the top of the current
(4)

-------
layer and residual buoyancy flux is computed. With the
residual buoyancy flux, the above procedures for the
buoyancy plume rise are repeated until all of the buoyancy
of the plume is exhausted Finally, plume top and bottom
heights are estimated using (a) the assumption of a top-hat-
shaped concentration distribution, and (b) the assumption
that plume thickness is the same as plume rise (Turner et
al., 1986).
For unstable cases, it is assumed that the plume is
dispersed throughout the mixing layer. However, if the
distance between the stack top and the top of the mixing
layer (zi") is less than 200m, fractional plume penetration
ip) is computed using Briggs" equation (1984):
p = 0, if	> 3.9[Fail's)}"3
p«1, if	V < 1.3 [Ft/(ws))m
p=	t	Otherwise	(5)
When the stack top is located above the mixing height, the
minimum of: (a) the approximation of plume rise for the
neutral case fEq. (2)} and (b) £Ji estimated by the rise of
bent-over plumes fEq. (4)] is used to estimate the plume
center height Again, the thickness of the plume is assumed
to be the same as plume rise for this condition.
The fractional portion of the plume for each layer is
determined using hourly plume top and bottom heights and
RADM's layer interface heights computed by the
hypsometric equation. Figure 1 provides a schematic of the
plume partitioning procedure. Note that the RADM layere
are defined in o-level surfaces, so the actual height varies
depending on the surface pressure values. Table 1 provides
Layer definitions for Mayer RADM.
*Top-hat shaped"
concentration distribution
Plume Top
plume top - plume bottoi
* Plume Center. /V"'
Plume Bottom
(k-1,)-th layer
'k-l
Fig. 1. Schematic of plume partitioning.
Table 1. Laver definitions of 6-laver RADM
Layer (k) I 2 3 4 5 6
g-level LP \99 .96 .90 CT.7 0.4 0.0
where o = -£-^- , Pup=100mb in RADM.
* top'
3. POINT SOURCE EMISSION PROCESSING
Using stack parameters as well as SO2, NOx, and total
hydrocarbon (THC) data from 1985 NAPAP point-source
emissions inventory, estimates of plume rise and emission
distributions for the RADM layers were determined by two
versions of plume-rise algorithms. The NAPAP point-
source data contains general identifying information (such
as Source Classification Code, latitude and longitude of
stack), stack parameters, major pollutant annual emissions,
temporal allocation factors, and speciation factors. Since
the total number of stacks in the RADM modeling domain
exceeds 50,000, it is very expensive to process plume rise
for all stacks for each meteorological episode. Therefore,
we classified point tources as major and minor stacks,
based on annual emission tonnage. We classified stacks as
major if:
SO2 emissions exceed500tonsfyr, or
NOj emissions exceed 210 wns/yr, or
Total hydrocarbon (THC) emissions exceed 400 tonsfyr.
Hie remaining stacks were classified as minor. Out of the
53,386 point sources in the 1985 NAPAP data, 5,924 were
classified as major and 47,462 as minor. Among the major
sources selected, stacks with physically consistent stack
parameters were used in the plume-rise calculations and the
remaining sources are combined with the minor stacks.
The emissions from major stacks were allocated to six
model layers using the plume-rise algorithms described
above, rot minor stacks, we computed seasonal average
diurnal plume fractions for each layer and applied them to
corresponding episodic case. Ate speciation and temporal
allocation, the minor source emissions were summed by
grid cell and by layer together with the major sources.
Figure 2 shows the flow of point-source emissions
processing in the RADM system.
Source
Sue*
Spceiti
Emiuena V I* SCC
Corwun Ematicre
toRADMSptCMi
to cntic Oau
Chmm Ejmosic
mcwi a*mu>ioft siu
Uyvr port-tour e* vmitsions
Came** •nwicril by gntf
Fig. 2. Flowchart of the point-source emissions processing
in the RADM system.

-------
4. SENSITIVITY OF RADM PREDICTIONS TO POINT-
SOURCE EMISSIONS PROCESSING
The objective of this study is to compare the effect of
two different plume-rise processing methods, method-1 and
method-2, on the sulfate deposition prediction by RADM
for two five-day periods - in summer (A9), and winter
(Al). Case A1 (30 Jan. 1982 - 4 Feb. 1982) was selected as
a representative winter episode. This case began with a
large l,031mb high pressure system over the eastern U.S.
A surface low pressure center tracked eastward across
southern Canada and a cold front that extended southward
from the centra- of surface low moved off the east coast of
the U.S. Very little precipitation was observed or simulated
in conjunction with the frontal system. Case A9 (2 Aug.
1983 - 7 Aug. 1983) was selected to represent a typical
summer episode. This case began with a high pressure cell
over the central U.S. A weakening cold front, located just
east of the high pressure cell, was moving slowly toward
the east coast During the last half of the case study period,
a weak quasi-stationary front that extended from Illinois to
New England induced light amounts of precipitation. Out
of five days in each episode, the first two days were used
for the initialization of RADM and remaining three days
were used for the simulation of acidic deposition.
<1 Differences in Emission Input
Figures 3-a,b show RADM-domain total hourly major
source plume fractions estimated with method-1 and
method-2 for winter Al case. Method-1 does not show a
distinct day-night difference. Method-2 estimates higher
plume rise especially during daytime because the plumes
are assumed to be well mixed within the mixed layer. For
summer case (Figures 4-a,b), method-2 produces very
pronounced diurnal variation of the plume rise pattern.
During mid-day, some portions of plume reaches the fourth
layer of the 6-layer RADM. Again, method-1 does not
show a distinct day-night difference: only a few sources
reach the fourth-layer even during a very convective period.
Since extremely small amount of plume reached above the
fourth layer, we limited our analysis up to fourth-layer of
the 6-layer RADM.
Method-2 plume-rise algorithm is very sensitive to the
variation of the atmospheric stability while method-1 does
not show a pronounced day-nighttime difference. The
difference between the two methods is more pronounced
for the summer case (A9) than the winter case (Al).
Gridded and layered major point-source emissions were
computed using the procedure described in the previous
section. Figures 5-a,b and 6-a,b present RADM domain-
total layered SO2 emissions allocated with both methods for
case Al and A9. Since taller stacks emit larger volumes of
pollutant than smaller ones, the total emission fractions
show heavier weighting in the upper layers than the layered
plume fractions. A basic difference between the two
methods is vertical allocation of the gridded emissions.
Depending on which level the pollutant is injected,
subsequent transport, diffusion, gas-phase chemistry
reaction, transformation inside clouds, and removal
processing will be affected. The SO2 emissions from the
major sources accounts for about 85 percent or higher of
the total emissions (from major and minor point sources
and area sources) in the RADM domain.
. 4.2 Sensitivity of RADM
Atmospheric deposition processes are significantly
influenced by where we inject point-source emissions.
1 Method-2, which distributes point-source emissions
throughout the depth of the mixed layer during daytime
when deposition velocities are large, results in more dry
deposition of primary pollutants such as SO2 than the
method-1. On the other hand, method-2 sometimes injects
point-source emissions into one upper layer. Then the
(a) Method-1, Case A1
HOUR
(b) Method-2, Case A1
HOUR
P Layer 1 gg Layer 2 gg Layer 3 g Layer 4
Fig. 3. Domain-averaged percentage of plume
by RADM layers (A 1; Winter)
(a) Method-1, Case A9
0%
°S!S85gES88S3
-------
(a) Method-1, Case A1
6 4x10
„ 3*10
®	c
* 2x10
8 2
Hour
(b> Method-2, Case A1
8 8 R
7x10°
6x105
g 5x10®
X 4x105
o 3x10®
* 2X105
1x10®
0x10°.
8
C\J
N
2 8 8 8
8
Hou
~ Layer 1 Layer 2 gj Layer 3 £ Layer 4
Fig. 5. Layered SOj emission rate for 6-layer RADM
(Al: Winter)
(a) Method-1, Case A9
Case A9
S
Hoir
(b) Method-2, Case A9
Case A9
4x10:
~ Layer 1 ^ Layer 2 gg Layer 3 g Layer 4
Fig. 6. Layered SO2 emission rate for 6-layer RADM
(A9: Summer)
Based on the figures above and other analyses, the
effects of using different plume-rise methods are
summarized as: (1) Method-2 produces 2-3 percent higher
domain total dry deposition and 2-3 percent lower domain
total wet deposition than method-1. Total (wet and dry)
deposition is affected by the episodic precipitation amount
and pattern. (2) Method-2 predicts higher dry deposition
closer to the sources than the method-1. (3) Maximum
difference in the dry/total sulfur deposition ratio is about IS
% near the high emission region with persistent airflow.
DISCLAIMER: The information in this document has been
funded wholly or in part by the United Slates
Environmental Protection Agency under contract 68-01-
7365 to Computer Sciences Corporation. It has been
subjected to Agency review and approved for publication.
5. REFERENCES
Anthes, R.A. and T.T. Warner, 1978: Development of
hydrodynamic models suitable for air pollution and
other mesometeorological studies. Mon. Wea. Review,
106,1045-1078.
Briggs, G.A., 1969: U.S. AEC Critical Review Series TID-
25075, USAEC Technical Information Center, Oak
Ridge, TN.
Briggs, G.A., 1975: Plume rise predictions. In: Lectures on
Air Pollution and Environmental Impact Analyses,
Workshop Proceedings, Boston, MA, Sept. 29 - Oct. 3,
1975. pp 59-111.
Briggs, G.A., 1984: Plume rise and buoyancy effects. In:
Atmospheric Science and Power Production, D.
Randerson, ed., DOE/TIC-27601 (DE84005177),
Technical Information Center, U.S. DOE, Oak Ridge,
TN, 850 pp.
Turner, D.B., 1985: Proposed pragmatic methods for
estimating plume rise and plume penetration through
atmospheric layers. Atmos. Environ., 19,1215-1218.
Turner, D.B., T. Chico, and J.A. Catalano, 1986: TUPOS—
A multi-source Gaussian dispersion algorithm using on-
site turbulence data. EPA/600/8-86/010, U.S. EPA,
Research Triangle Park, NC, 155pp.
»1I4
It I
<•.0
ft
I •
• I
9 •
.) •
mJv.c

Fig. 7. Percentage difference between plume-rise method-1
and method-2 in the ratio of dry to total (wet and dry)
sulfur deposition, (a) winter case A1, and (b) summer case
A9.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before comp
1. REPORT NO. 2.
EPA/600/A-92/035

4. TITLE and subtitle
Sensitivity of RADM to point source emissions processing
S. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHORCS! -I n
D. W. Byun and F. S. Binkowski
i. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING organization name and address
Computer Sciences Corporation, RTP, NC 27709
2AREAL, ASMD, RTP, NC 27711
10. PROGRAM ELEMENT NO.
N104/C/1001/22/J (FY-92)
11. CONTRACT/GRANf NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Atmospheric Research & Exposure Assessment Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Trianqle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Conference PaDer (3/88-9/901
14. SPONSORING AGENCY CODE
EPA/600/09
15, SUPPLEMENTARY NOTES
16. ABSTRACT
The Regional Acid Deposition Model (RADM) and associated Engineering Model have
been developed to study episodic source-receptor relationships on a regional scale. The
RADM includes transport, chemical transformation, and deposition processes as well as
input of emissions into the vertical layers using a plume-rise submodel. As wind speed
and direction change with height, and as atmospheric turbulence varies significantly with
distance from the earth's surface, the model predictions can be affected considerably
by the height at which the emissions are injected. Atmospheric emissions can be divided
into area and point sources. The area sources include mobile emissions, biogenic
emissions, and other emissions that are often related to the surface area of the earth.
Area sources are more difficult to control and have inherent diversities in their physio-
chemical processes. Point sources include stacks at power plants and manufacturing
facilities. Usually, point sources are the target of emission reduction programs.
17.	KEY WORDS AND DOCUMENT ANALYSIS
1. DESCRIPTORS
b-IDENTIFIERS/OPEN ENDED TERMS
c. cosati Fifid/C»oup
'

•
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report i
UNCLASSIFIED
21. NO..OF PAGES
b
20. SECURITY CLASS (Thu page)
UNCLASSIFIED
22. PRICE
f PA 2220-1 (R.», 4_7?) pkcviou* koitio* i» BMOttrt
-------