-------
better results. For example, Miller and Holzworth have computed the
city-wide average concentration for selected early morning and after-
noon two-hour periods in Nashville on 31 selected days. The frequency
distributions of predicted and observed concentrations obtained from
this analysis are shown in Figures 37 and 38. The Miller and Holzworth
model is a rather extreme simplification of the Gaussian plume model
for urban diffusion analysis which completely ignores spatial varia-
tions in emission rates by use of a city-wide average. Therefore, no
resolution of the spatial distribution of concentrations is possible
with this model. This model has been recommended as a method of esti-
mating regional air quality where suitable monitoring observations are
not available and no single source is the principal cause of pollution
levels (Federal Register, 36, August 14, 1971, Part II).
Turner, who devised the scheme (based on the degree day con-
cept) for estimating diurnal variations in SCL emission rates in the
Nashville data used by Miller and Holzworth, used a more extensive set
of the same data to compute 24-hour average concentrations which included
consideration of the diurnal variation in emission rates (Turner 1964).
His results are not reported in enough detail to construct frequency
distributions; however, he reports that 43.7 percent of his predicted
two-hour concentrations at seven stations were within +0.01 ppm
(about 27 mg/m ) of the observed concentration. For 24-hour observa-
tions at the same seven stations he found that 58.1 percent of predicted
values were within +_ 0.01 ppm.
-------
bo
I
A)
1000
9
8
7
6
5
4
3
2
1001
9
8
7
6
5
4
3
2
10 1
g
7
5
4
3
2
11
Cumulative Percentage
2:-. 5 10 15 20 30 40 50 60 70 80 85 90 95 9
/
/
*/
/
'/'
.
/
X
/
\r
X
^
j/'/
^ /
/
^
*T
PRE
x>
DIG"
OBS
FED
/
ERVE
X"
, — -*
D
NASHVILLE
31 Selected Days
(1958-1959)
7 STATIONS
^-
SO Concentrations
(0400-0600 CST)
31 c.
ises
Figure 37. Observed and Predicted Frequency Distributions of Early Morning Concentrations
Reported by Miller and Holzworth (1967)
-------
a
o a
O ui
« S
O "
u 2
1000
9
8
7
6
5
4
3
2
lOOl
9
8
7
6
5
4
3
2
101
9
8
7
6
5
4
3
2
11
Cumulative Percentage
2'-. 5 10 15 20 30 40 50 60 70 80 85 90 95 98
'
X
/
/
p
X
/
/
/
REDIC'
^
FED/
/ ^
7
^
OBSE
-^x
RVED
^
*
m -*
'Z—
31
SC
.— —
-^
-^
NASHVH
Selected
(1958-19
7 STATIC
) Concen
^— '
.LE
Days
59)
DNS
trations
(1400-1600 CST)
31 cases
Figure 38. Observed and Predicted Frequency Distributions of Afternoon Concentrations
Reported by Miller and Holzworth (1967)
-------
More recently, Fortak (1969) reported results with a Gaussian
plume type of urban diffusion model. The frequency distribution of pre-
dicted and observed hourly SCL concentrations at Sites #1 and #4 in
Bremen, Germany, for the 1967-68 heating season are shown in Figures 39
and 40. Average daily emission rate estimates were determined by
Fortak for these calculations. The graphs included represented the
best and the worst agreement obtained by Fortak at four sites. He
points out that Site #4 was in the vicinity of a large plant, and he
attributes the observed high concentrations to uncontrollable, and
unaccounted for, low-level emissions from the nearby plant. At Site #1
the agreement between the distribution of predicted and observed
concentrations is almost as close as that obtained in this study for
the St. Louis data.
The results cited above, and those from this study, show that
the use of temporal variations in SOp in emission rates in concentration
calculations leads to a realistic determination of the frequency distribu-
tion of short-term concentrations over a seasonal period, as well as a
more accurate estimate of the seasonal mean concentration.
4.4 FINDINGS
A summary of the preceding results on the validity of the
Gaussian plume type of multiple source urban diffusion model is given
below. These results are based on the predicted and observed concentra-
tions of SCL at 8 locations in Chicago during January 1967 and 10 stations
in St. Louis during December 1964 to February 1965. The predictions
used hourly estimates of meteorological and emission parameters. The
atmospheric stability was estimated from hourly weather observations
-------
1000
9
8
7
6
5
4
3
2
100 1
9
8
7
6
5
4
3
2
101
9
8
7
6
5
4
3
2
ll
Cumulative Percentage
2-: 5 10 15 20 30 40 50 60 70 80 85 90 95 98%
'
/
/
S*
X
PREDICTED^
^ ^
x|X
•XOBSERV]
S
-"X
X
D
/
X -X
>/^
X
X
x^
Bremen, Gern
^
any
Heating Season 1967-1968
Site 1
9
8
7
6
5
4
3
2
1
9
8
7
6
5
4
3
2
1
9
8
7
6
5
4
3
2
Figure 39. Observed and Predicted Distributions of Hourly Concentrations
for Site 1 in Bremen, Germany, Reported by Fortak (1969)
-------
•fc
a.
y
a
1000
9
8
7
6
5
4
3
2
100 1
9
8
7
6
5
4
3
2
101
9
8
7
6
5
4
3
2
1 1
Cumulative Percentage
2:. 5 10 IS 20 30 40 50 60 70 80 85 90 95 98%
/
'
/
X
OBSEr^
/
/
/
f£DS
/
f
H(
/
.X"
/
/PI
/
Brem
>ating .
X
x
IEDI
en,
5ea«
Sit
X
X
x
CTEC
Gerrr
an 19
e4
X
^x
>. '
'
)
any
57-1968
^
9
8
7
6
5
4
3
2
1
9
8
X
6
S
4
3
2
1
9
8
7
6
5
4
3
2
Figure 40. Observed and Predicted Frequency Distribution of Hourly Concentrations
for Site 4 in Bremen, Germany, Reported by Fortak (1969)
-------
from an adjacent airport using the McElroy-Pooler diffusion parameters
based on Turner's definitions of stability categories. The mixing layer
ceiling was estimated from radiosonde observations taken twice daily
from remote locations (100 to 200 miles away). The wind speed and
direction were hourly averages of several 3 in St. Louis, 8 in Chicago)
continuous records. The emission rates of the largest sources were
identified and located individually. For other sources a mean emission
rate per unit area was estimated for a square gridwork of points with
a one mile spacing between adjacent points. Each emission rate was
related to hourly estimates of space heating and other operating
requirements.
The findings are summarized as follows:
1. Predicted long-term (month or season) concentrations
averaged over several locations are in good agreement with observed
concentrations.
2. Predicted long-term concentrations at individual locations
show a root-mean-square-error equal to about half the mean and indicate
a slight tendency to overestimate more often than underestimate observed
concentrations.
3. Predicted short-term (one or two hours) concentrations at
individual stations show larger deviations from observed concentrations
than'do the long-term predictions. However, over a period of a month,
or a season, the overall distribution of predicted short-term concentra-
tions closely approximates the distribution of observed concentrations.
4. Proportionate stratified sampling is an effective method of
selecting a limited set of short-term periods which adequately define a
representative distribution of short-term concentrations in a long-term
period. One hour out of 24 is a sufficient sample size for a three month
period if the selected hour is varied to include all hours of the day.
5. The calm, or light wind, case is not adequately treated by
the Gaussian plume type of urban diffusion model. Further study of
procedures for applying the model to this type of situation is needed.
-------
Section 5.0
-------
Section 5.0
SENSITIVITY ANALYSIS
Sensitivity is formally defined as the partial derivative of
a model's output with respect to its input. In the case of complex
models, however, a more practical definition, which is often employed
for analytical purposes, is the incremental change in output resulting
from an incremental change in input.
In a numerical simulation model as complex as the one utilized
in this report, it is not possible to study all aspects of sensitivity
analytically, nor is it safe to infer sensitivities from the form of the
Gaussian model, which is the kernel of the simulation, because of the
numerous interactions involved. Proper analyses require appropriate
numerical exercising of the complete simulation.
Sensitivity analyses of urban pollution models have been
reported by Hilst (1970) and Milford, et al. (1970a; 1970b). Hilst
applied sensitivity analysis concepts in an example case study of the
TRC Region Model, developed for the State of Connecticut, which involved
a combination of a trajectory-oriented Gaussian model with the puff
version for area sources and the plume version for major point sources.
Hilst's results are of interest, but as he attests, limited in scope by
the case study nature of the analysis.
The work reported by Milford, et al., deals with the variations
of the model developed for the New York/New Jersey/Connecticut Air Quality
Region. The studies which they have reported describe a model which
bears a general resemblance to the steady-state plume model implemented
-------
in the study covered in this report. Their work, however, is very
specifically oriented in form, input, and application to the greater
New York area, and their reported sensitivity results focus largely on
highly specific individual case examples, rendering useful generaliza-
tions difficult. The study in this report attempts to derive broad-
scale significant sensitivity findings from a generally applicable model.
This section describes the work performed to analyze the
sensitivity of the output concentrations of the multiple-source Gaussian
plume diffusion model to model input parameters. Important questions
to be answered by the analysis, which concentrates on the sensitivity
of the short-term version of the model, with reference to the longer
term climatological version where appropriate, are presented. The
parameters and their value ranges are discussed, the methodology is
described, and the analysis and results are presented. In the discussion
of this section, the broad-scale significant findings, where sensitivity
exists, are presented in summarized form. Appendix F contains descrip-
tions and samples of the computer printouts which give complete listings
of the sensitivity computations.
5.1 ELEMENTS INVESTIGATED
The principal points which the sensitivity analysis addresses
are presented in question format as follows. The analysis was approached
from this point of view in order to focus on questions which are
considered to be of the greatest practical significance, and the results
are intended to be definitive with regard to these questions. The
questions are identified in terms of type of model input, and each is
subsequently discussed in greater detail in the cited sections.
-------
1. Spatial Variability of Emission Rates. The question
here concerns the scale of variability in area source emission rates
which can impact significantly on model predictions. The two considera-
tions of primary interest in this question are the fineness of the grid
used to represent area sources, and the basis used for separating
significant point sources from area sources (Section 5.4.1).
2. Vertical Distribution of Area Source Emissions. This
question concerns the extent to which various vertical distributions
in the assumed emission height, and buoyancy rise, of area source
emissions may affect model predictions (Section 5.4.2).
3. Vertical Diffusion Parameters. This question addresses
the extent to which variations in the power law functions, which are
used to represent the vertical spread of pollutants with travel distance,
affect the model outputs (Section 5.4.3).
4. Decay Rate. The question here concerns definition of the
conditions under which this parameter significantly affects outputs
(Section 5.4.4).
5. Wind Speed and Hind Profile Power Law. The question of
model sensitivity to wind speed is normally straightforward, and becomes
complicated only when a decay rate exists. For zero decay, the model
sensitivity to wind speed is only slightly complicated by interaction
with the wind profile power law. Thus, the question of model sensitivity
-------
to both wind speed and the vertical wind speed profile power law is
examined (Section 5.4.5).
6. Mixing Ceiling. This question concerns the significance
of uncertainties in this parameter (Section 5.4.6).
7. Wind Direction. The question here concerns the degree
of resolution in wind direction to which the model output is sensitive
(Section 5.4.7).
8. Diurnal Variation in Emission Rate. The main question
here concerns the effect, on the predicted long-term average concentra-
tions, of any correlation of diurnal variations in emission rates with
diurnal variations in meteorological conditions (Section 5.4.8).
5.2 PARAMETER RANGES AND COMBINATIONS
Each of the sensitivity points raised in Section 5.1 focuses
on the sensitivity of the model to certain specific model inputs, and
all of the model inputs are incorporated in one or more of these questions.
In order to design model sensitivity experiments, it is necessary to
define a reasonable range of interest for each model input and to select
combinations of values of all input parameters to use in testing for
sensitivity. It will be seen that the inputs can be represented by a
small number of values scattered over the total range of values of
interest.
Sensitivity analysis in this program is focussed on changes
in calculated concentrations which are associated with changes in input
-------
for a given set of input values. In view of the large number of such
comparisons which are possible in the context of the preceding eight
questions, the first step is to define reasonable ranges of interest
for each parameter; subsequently, determination was made of which para-
meters have little influence on output over their defined range of
values, and the remaining parameters were analyzed in more detail.
The initial set of parameters and the specific input values
selected for use in the sensitivity analysis are shown in Table 25.
The values selected were based on the judgement and experience of the
in-house staff of diffusion meteorologists, as well as, to some extent,
on the numerical information developed in the course of the validation
study (Section 4.0). The following comments on the values selected for
certain of the parameters are in order at this point:
• Decay Half-Life: One obvious choice is for no decay
(infinite half-life); the other (30 minute half-life)
represents a moderately reactive material. (A third
extremely short half-life (5 minutes) was experimented
with to a small degree, with results reflected in
Section 5.4.3.)
• Hind Speed: The three values are intended to reflect
light, moderate and strong anemometer-level winds.
• Wind Prof i1e Power: Two values, arbitrarily chosen as
depicting the range.
• Wind Direction: (See Sections 5.4.1 and 5.4.7.)
• Mixing Ceiling: A very low value, an intermediate
(characteristic) value, and a high value.
• Piffusion Function and Stabi1ity C1 ass: The three
cases represent extreme atmospheric stability, neutral
stability, and extreme instability.
-------
Table 25. Sensitivity Parameters, Ranges and Selected Values
Parameters
METEOROLOGICAL AND
POLLUTANT:
Pollutant Half-Life*
Wind Speed
Wind Profile Power
Wind Direction
Mixing Ceiling
Diffusion Function and
Stability Class
EMISSION AND RECEPTOR:
Number of Point Sources
Area Source Grid Spacing
Distribution of Emission
Heights for Area Sources
Treatment of Diurnal
Variations in Emissions
Receptor Location with
Respect to Source Area
Units
min
m/sec
--
azimuth deg.
m
m
—
miles
—
—
--
Range
0-«>
1-20
0. 1-0. 5
0-360
100-00
--
—
—
—
—
Selected Values
30;oo
2, 6, 18
0.15, 0.3
—
100; 500; 2500
Pasquill Class E
McElroy- Pooler Class D
McElroy-Pooler Class 1
(1) All major sources (51 for
St. Louis)
(2) All with Annual Emissions
within 10% of Largest
Emitter (19 for St. Louis)
(3) None (all aggregated into
area sources)
(1) 0.25
(2) 1
(3) 4
(1) AU at Mean Height
(2) 50% at Mean Height, 25%
at 1/2 Mean Height and
25% at 3/2 Mean Height
(See Section 5.4.8)
(1) Upwind Zone
(2) Central High Emission
Zone
(3) Downwind Zone
*An "infinite" half-life represents a material that is essentially stable, or non-reactive, in the
atmosphere, and corresponds to a zero decay rate. A 30-minute half-life is equivalent to a decay
rate of 0.023J/min.
-------
Number of Point Sources: See discussion in Section
5.4.1.
Area Source Grid Spacing: See discussion in Section
5.4.1.
Distribution of Emission Heights for Area Sources:
See discussion in Section 5.4.2.
Treatment of Diurnal Variation in Emissions: See
discussion in Section 5.4.8.
Receptor Location with Respect to Source Area: See
discussion in Sections 5.4.1 and 5.4.7.
In addition to the parameters listed in Table 25, there is a
need to define a selection of basic geographic patterns of emission rates,
as well as the selection of receptor locations relative to this pattern
at which to measure sensitivity effects. It seems reasonable to define
three principal, general situations concerning the relationship of a
receptor relative to an emission pattern:
1. The receptor is in an area of relatively uniform emis-
sions with no significantly strong upwind sources (upwind receptor).
2. The receptor is in an area of high emission rates
surrounded by noticeably lower upwind emission rates (as in the center
of urban area), (center receptor).
3. The receptor is in area of light or moderate emissions
with significant upwind sources (downwind of urban center), downwind
receptor).
In the validation analysis the relative contribution to the total emis-
sions arising from point sources in contrast to area sources had already
been defined for the receptor locations which represent sampling stations.
From these results it was noted that, in by far the majority of the
-------
cases examined, the overall contribution from point sources was small.
There were notable exceptions in which the point sources were the dominant
contributors to particular receptor locations under particular conditions.
The impact of such special situations is examined further in Section 5.4.7.
Meanwhile it was determined that the first order of importance was to
examine realistic patterns of area source emissions, in order to assist
in the definition of emission rate patterns and the associated receptor
locations for use in the sensitivity analysis.
The area source emission rates for each square mile and each
hour in the 2036-hour St. Louis data sample had been previously stored
on magnetic tape for use in the validation analysis. These data were
retrieved and used to generate hourly contour maps of the area source
emission rates, which were then studied to determine whether consistent
patterns were present. Three sample maps are shown in Figures 41
through 43. These represent relatively extreme variations in emission
patterns over the 89-day period. These figures illustrate the general
observation that, although the magnitude of emissions at any point may
vary by a factor of as much as ten, the distribution of the pattern
remains relatively consistent. No outstanding variation in the general
shape of the pattern was noted with time of day or day of week. As a
result, it was decided that the three receptor location characteristics
described above (upwind, center and downwind) could be reasonably
represented by a single emission pattern with three such receptor loca-
tions. The selected pattern is that shown in Figure 42. The wind
-------
SYMBOL RATE, G/SEC
0 0.0 TO 0.01
1 0.01+ TO 0.03
2 0.03+ TO 0.10
3 0.10+ TO 0.30
4 0.30+ TO 1.00
SYMBOL RATE, G/SEC
5 1.00+ TO 3.00
6 3.00+ TO 10.00
7 10.00+ TO 30.00
8 30.00+ TO 100.00
9 OVER 100.00
40
39
38
37
36
35
34
33
32
31
30
29
N 28
0 27
R 26
T 25
H 24
23
C 22
0 21
0 20
R 19
D 18
I 17
N 16
A 15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
T
E
EAST COORDINATE
123
123456789012345678901234567890
01122222222222022466667(O66>5433
022221212222222223N77676^432333
2222222222000222224*6.167*53333
22222222220001223222222$6*4333
2222222222333222222202236D0333
222222222233333322220022^11333
022220203323333333000005502224
222203233323334444400021033344
25S033433333444444330023333444
5/70344443334444445430034333444
5^60224444445544444450444440444
2200233544!
2033445544X
2333555<
3233,555!
3233445,<
333344 5\
2S333344545«
4333443434!
33334443357
333345444!
323433335y667767"
323313356667667/J
i55443341444444344
.5544(0204444433334
1004444443334
:3443044
.4333034
>43033344
'544333435
04444^7
\554t7
5545?
'55555
544
.. . „ ^. ,76460454
44449^^5 555^67777645 ^6A6/5i554
444443*6j5555,666776r045445i t
455433554/SS6667/0344445433$.666
344334544V5^5.6€/254544433333>76
3333344444444>?4445444333333^
33334444444554200feK443333333l
33333344444443204444433333334!
333333344444431134444333333033
333333444344430444444434433333
333334334433330044444434533333
333333353334424445544333333333
333333333333300044444433333333
333333333343233444444333333333
333333333232332333344433330333
333333333323233333344443333333
333333333233303333344444334333
FIGURE 41.
ST. LOUIS AREA SOURCE EMISSION RATES,
1AM DECEMBER 2, 1964.
-------
f Receptor Location
SYMBOL RATE* G/SEC
0 0.0 TO 0.01
1 0.01+ TO 0.03
2 0.03+ TO 0.10
3 0.10+ TO 0.30
4 0.30+ TO 1.00
SYMBOL RATE, G/SEC
5 1.00+ TO 3.00
6 3.00+ TO 10.00
7 10.00+ TO 30.00
8 30.00+ TO 100.00
9 OVER 100.00
EAST COORDINATE
12 3
123456789012345678901234567890
40
39
38
37
36
35
34
33
32
31
30
29
N 28
0 27
R 26
T 25
H 24
23
C 22
0 21
0 20
R 19
D 18
I 17
N 16
A 15
T 14
E 13
12
11
10
9
8
7
6
5
4
3
2
1
FIGURE 42.
33
444
54333
4333
00333
1333
'03224
44445
02233322333233022'
03333232333333332:v
333333333300033332N
33333333330002333322;
22333223223333223333022]
33332233333\J33333222002;
033220203335333333000001
33220433443334444440002!
35£fl44444444V;44444440024444444
676/45554335444455430044444444
^vltf 2344445 5/^44454(5)045 5550444
320123365496^5543442455554444
20445 5 5 54>776$5 54/7)304444544444
'005544554444
4344b^66677666Yr7rOOZ66>4554044
434349i66777667Y7fcOO?78$>4444045
4444446^4»5666"
m4434555>
^3335 54545)6 ^B98§88(7 7/520555*
333355444«77^aa8«8V7S&«^55!
444455555>6776^8ere88&^8>67555!
43444443077i777888f6j69a77556^5
33444445Jb77(|)77e8a8$7p89'>5^5555
4444<6666.6666777|8/^p52<77^66^55
45A4 5T&&S2 6 7 7 7vB[$ 5 5 5 5\^S?>
4^^34,6^5^66767^0465555444X667"
454434(d555;^67>254554444444&77
44444445555597/5445^444444444476
4444444444 5 5&8200fe)4V44444443«>J
4443334444455420554V444444445T
433333444444433344444444444044
33333344444443055544V444434444
33334444444443 00 5 5 544445(6)44444
333444454444425555554444444444
334444444344300055555^44444444
344444444444244455555V44444444
344444444342 3 33 3 334 5 5«4440444
344444444323333343344IR4444444
L2fc7754044444
'80554445*
Wind Direction
ST. LOUIS AREA SOURCE EMISSION RATESt
-------
SYMBOL RATEt G/SEC
0 0.0 TO 0.01
1 0.01+ TO 0.03
2 0.03+ TO 0.10
3 0.10+ TO 0.30
4 0.30+ TO 1.00
SYMBOL RATE, G/SEC
5 1.00+ TO 3.00
6 3.00+ TO 10.00
7 10.00+ TO 30.00
8 30.00+ TO 100.00
9 OVER 100.00
40
39
38
37
36
35
34
33
32
31
30
29
N 28
0 27
R 26
T 25
H 24
23
C 22
0 21
0 20
R 19
D 18
I 17
N 16
A 15
14
13
12
10
9
8
7
6
5
4
3
2
1
EAST COORDINATE
123
123456789012345678901234567890
022222222222220224/f666($CrT6u&'5533
022222222222222223x77776\2452444
222222222200022222WxZ4>553333
222222222200022232222226.764333
222222222233322232220223]6p0333
222222222233333322220022/60333
022220203333333333000004^03224
222204233423344444400021033444
255033444434444444440023333444
445554334444445430044343444
23444455A554444450455550444
2201233^\5436'555543443455554444
2033553655^6&/6)554/4404444433444
23335596l§A776/a5JJ56)004444444434
33335566666666T77/602xf£ifo4444044
333345fe677766777aOOfl78/64444044
333445^6676755297644044344
A333454545y67
3333454447
88a77X5205544/67
988\76#55*554(77
T
E
5fe767777j888/77|8
fe67677"388>»555ii£7>254544444444)b76
3433445545555^4445444444444476
3333455444556^200(6)!f4444444435s7
33333344444553204544434444445!
333333444444443344444444444033
333333444444430555444444434444
333344334444430055444445(S>4444
333444454444425555554444444444
333444444333300055554444444444
344444444443233455555444444444
334444444342333333455434440444
334444444323333343344444444444
344444444233303333344444444444
FIGURE 43.
ST. LOUIS AREA SOURCE EMISSION RATES,
7AM DECEMBER 3, 1964.
-------
direction was defined to be as shown (changes in direction are examined
in Section 5.4.7), and the three solid circles were selected as receptor
locations representing upwind, center (high emission), and downwind
receptor zones.
The sensitivity analysis was therefore performed in the context
of this representative background pattern of emissions, and the para-
meters listed in Table 25 were varied against this background.
5.3 METHODOLOGY
The methodology followed is a straightforward manipulative
one, in which changes in input are used to define changes in output.
The output changes are then examined as relative or absolute changes
and ranked to determine those which are most significant. Short-term
concentrations are emphasized in the analysis, and the sensitivity of
long-term concentrations is addressed in the text as appropriate. The
first step in the analysis consisted of computer runs representing
a full factorial replication of all combinations of the inputs identified
in Table 25 which are relevant to short-term concentrations. This
includes all but wind direction and the magnitude of diurnal variations
in emissions. The total number of combinations is 5832 (i.e., 2 pollutant
half-lives x3 wind speeds x2 wind profile powers x3 mixing ceilings
x3 diffusion functions x3 sets of point sources x3 area source grid
spacings x2 sets of emission heights for area sources x3 receptor locations)
Within this total, for each model input, two to three thousand sets of
variations (depending on.the input parameter involved) in model output
were therefore generated for each specified variation in input.
-------
Inputs which show little output variation over all sets, or
almost all sets, were accordingly identified as insensitive inputs, and
variation in these insensitive parameters was not considered further.
Each of these is discussed in Section 5.4. The selection of a criterion
which would represent a significant change in output over a range of
input values was governed by the validation findings. As a result of
those findings, it was decided that input changes must generate at least
a 50 percent change in output for the range of input values considered
in order to qualify as a significantly sensitive input. Where signifi-
cant change was found, the analysis was pursued in greater depth, as is
described in Section 5.4.
5.4 SENSITIVITY ANALYSIS RESULTS
In the following discussion, the results of the analyses of
the computer runs described in Section 5.3 are presented in detail. The
analytical procedure for defining the impact of each input parameter
consists of identifying "sensitive" changes in calculated concentrations
resulting when an input parameter is varied. A "sensitive" change is
defined to be a change in the input parameter which results in a greater
than 50 percent change in the calculated concentration. Such cases are
then subjected to more detailed analysis.
5.4.1 Spatical Variability of Emission Rates
Urban diffusion modelers usually identify only the most
significant point sources, and obtain reasonably accurate estimates of
emission rates for these sources. Emissions from other sources are
-------
treated as uniformly distributed over a segment of area, and estimates
of the emission rate per unit area are made for such convenient squares
or blocks of the urban area. A square mile is a frequently used block
size.
The actual computational treatment employed in evaluating the
effects of these selected block sizes on urban air pollutant concentra-
tions may sometimes involve further assumptions regarding the distribu-
tion of source within each block (e.g., use of point sources, normal
line source, virtual point source, or uniform area source concepts to
represent each block mathematically.)
In this study the overall area source input data are represented
by a gridwork of point locations, each with its own emission rate per
unit area, and thus describing a smooth continuous surface (in the
mathematical sense) of area source emission rates. In the program
computations, linear interpolation between points is used to define the
emission rate as a continuous function of position in evaluating the
effects of area source emissions. For sensitivity analysis purposes
variation in the fineness of the area source representation is accomplished
by changing the spacing between grid points and the corresponding block
size in the area source emission inventory. The two questions of concern
here are, What is the real spatial variability in emission rates? and,
How accurately should the real spatial variability be reflected in the
model? Since the real spatial variability is not known, except as
estimated for square mile blocks, this parameter has been hypothesized
for testing sensitivity. The assumption has been made that when a unit
square mile area is divided into 16 quarter-mile squares, the emission
-------
rate per unit area for each subdivision will be approximately normally
distributed (in a statistical sense) about a mean value with a standard
deviation equal to one-half the mean. In the sensitivity analysis, mean
values for square mile areas were used with a random number generator
to define emission rates appropriate to the smaller quarter-mile squares.
Standard IBM computer routines for random number generation and inverse
normal function evaluation were used.
The basic computer model defined in Section 3.0 and the
selected set of S02 emission data drawn from the St. Louis data sample
were used to test whether changing the grid resolution (1 mile) to a
finer (0.25 mile) or a coarser (4 mile) mesh had a significant effect
on the calculations.
As a further sensitivity test, the effect of various levels
of aggregation of point source emission data into the general area
source emission rate was examined. This was done by comparing calcula-
tions when all (51), some (19), or none of the major sources were merged.
The point source emission rates associated with the selected area source
emission pattern are listed in Table 26. In addition to the two extreme
cases of merging none or all of the point sources, the effect of merg-
ing just those points, whose emission rates were less than 10 percent
of the largest emission rate, was examined (i.e., merging all but the
highest 19 emission rates). The location of the point sources relative
to the three sensitivity receptor locations is shown in Figure 44. In
order to examine the effect of large point sources on the center receptor
location, the wind direction was shifted as indicated in Figure 44 to
see what effect that would have on the sensitivity results.
-------
Table 26. Ranked List of St.
Rates for 1300 LST,
Louis Point Source Emission
December 5, 1964
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Identi-
fication
Number
44
48
49
47
30
51
43
50
45
36
37
46
42
8
4
31
41
22
2
6
SO2
Emission
Rate, g/sec
2680
1560
1310
1280
1050
752
706
604
601
591
586
568
529
488
433
384
353
305
277
191
Rank
21
22
23
24
25
26
24
28
29
30
31
32
33
34
35
36
37
38
39
40
No.
35
11
27
32
26
1
20
14
33
24
23
19
40
29
3
39
12
21
9
34
Rate
188
174
163
120
115
92
91
77
77
68
61
49
43
40
36
35
33
33
29
22
Rank
41
42
43
44
45
46
47
48
49
50
51
No.
10
17
16
25
5
18
13
28
7
15
38
Rate
19
19
18
18
17
15
11
11
9
4
0
-------
43
35
30
25 _
20
15
10
11
10 Point Source
Receptor
10
15
20
25
30
Figure 44. Location of Point Sources in St. Louis Relative to Wind Directions and
Receptor Locations in Sensitivity Analysis
-------
Parenthetically, it will be seen in Section 5.4.2 and 5.4.5
that, when the initial large-scale screening analysis was performed,
two parameters were immediately demonstrated to have only insignificant
effects on the output concentrations. These two, the vertical distribu-
tion of area source emissions (5.4.3) and the profile power (5.4.5), were
accordingly eliminated from the analysis in considering sensitivity to
other inputs; at the same time, it was found possible to reduce the
number of mixing ceilings considered (Section 5.4.6) from three (100,
500 and 2500m) to two (100 and 500m) for most of the detailed comparisons.
Thus, the number of combinations considered in this section was reduced
from 5832 to 972 (i.e., two pollutant half-lives x3 wind speeds xl wind
profile power x2 mixing ceilings x3 diffusion parameters x3 sets of
point sources x3 area grid spacings xl area source emission height
x3 receptor locations). Of these, 324 represent the effect of a changing
grid mesh size as a function of meteorological conditions in the presence
of the standard of 51 individually identified point sources. When the
324 are examined in detail, almost half of the cases (141 cases or 47%)
show significant changes in concentration (40% change) when the one-
and four-mile spacings are compared. Similarly, 78 cases (24%) show
show significant changes when the 1- and the 0.25-grid sizes were
compared. When a shift in wind direction was considered, as shown in
Figure 44, the number of sets yielding significant changes with a change
in grid mesh size was reduced slightly.
These results show that averaging area source emission rates
over areas larger than 1 square mile can lead to significant errors in
estimated pollutant concentrations. Furthermore, if the standard
-------
deviation in emission rates of quarter mile squares (about 6 city blocks)
is of the order of 50 percent of the mean over a square mile area, (as
postulated in the beginning of this section), then even the use of square
mile average emission rates can lead to significant errors in estimated
pollutant concentrations.
Now, considering the impact of merging point sources into the
area sources, we recall that this examination focuses primarily on the
overall impact of such changes, rather than on the fine scale details
of effects on specific receptors in the vicinity of significant point
sources. The latter aspect is dealt with by Mil ford, et al. (1970a),
and, because of its wind direction dependence, in Section 5.4.7 of this
report. In the broader context, then, the following findings apply.
For the smaller grid sizes (1 and 0.25 miles) only a negligible number
(1-4%) of the cases show significant changes when the number of individ-
ual point sources is reduced from 51 to 19, and then to zero. For the
coarse 4-mile grid, we find 4 percent of the cases showing significant
concentration changes when the 19-source case is compared to the
51-source case, increasing to 18 percent when the "no-point-sources-
considered" case is thus compared.
Therefore, we see that the broad-scale concentration picture
is little affected either by treating individually, or by merging,
various numbers of point sources when the area source grid scale is of
the order of 1 mile or smaller; however, with a larger grid scale (of
the order of 4 miles), failure to take into account at least the main
point sources individually can cause problems. The 4-mile grid
-------
dimension was already questionable, of course, from the previously
stated findings on grid size alone.
5.4.2 Vertical Distribution of Area Source Emissions
Area source emissions of SOp consist of a large variety of
individual sources including stores, small plants, apartment buildings
and small single and multi-family homes, to mention a few of the more
common types. Since the emissions from these sources are primarily
contained in burned fuel exhaust, the emissions are hot. As a result,
the emissions are released from a variety of heights with a variety of
plume rise effects. Although it is convenient to treat all emissions
from a particular area as emanating from the same height, it may be
unrealistic to do so. Various devices may be employed to simulate
the vertical distribution, such as the use of multiple area source
heights or the assumption of a vertical dimension in the initial plume.
An initial vertical distribution of pollutants has been
simulated in this study by using multiple emission heights, and allo-
cating the pollutant emission rate among those selected heights. In the
sensitivity analysis, the results obtained by allocating 25 percent of
the area emission rate to a height of one-half the mean emission height,
and 25 percent to one and one-half times the mean emission height
(leaving 50% emitted at emission height), are compared with those when
all emissions in an area are at the same height. In terms of the initial
set of 5832 input combinations there were 2916 pairs of such comparisons.
In none of these was the.concentration resulting from one distribution
50 percent greater than from the other. In fact, only in the case of a
-------
combination of high decay rate, low wind speed, and stable diffusion
parameters, did one exceed the other by more than 25 percent. In
general, the difference between the vertical distributions was negligible.
As a result, in additional calculations, all emissions in any given area
source were treated as emanating from a single height.
5.4.3 Vertical Diffusion Parameters
One way of expressing a basic hypothesis (the narrow plume
concept), which was found to be acceptable in the model implementation
described in Section 3.0, is that the scale of variability in emission
rates (i.e., crosswind distance between significant changes in emissions)
is large relative to the scale of variability in plume concentrations
(i.e., the diffusion parameter a ). Furthermore, as was seen in
Section 5.4.1, the contribution of point sources relative to that from
area sources is significant only a small percentage of the time in terms
of the broad concentration picture. As a result, the crosswind diffusion
parameter is of only minor interest in the model. The impact of atmos-
pheric turbulence and stability is manifested primarily in the vertical
diffusion description.
The critical effect of choice of diffusion parameters for a
"one-class change" in stability for the basic plume equation (single
point source) is illustrated in Figure 45. This figure shows normalized
concentrations (xu/Q) along the plume axis, as a function of downwind
distance, from a point source at a height of 20 meters. The concentration
is shown for the four combinations of two mixing-layer ceiling heights,
100 meters and 1000 meters, and two sets of diffusion parameters. One
set corresponds to stable conditions using the E class of the Pasquill
-------
A
frrn
10
-4
10
.-5
rt
a
a
"0
-------
parameters; the other set corresponds to the McElroy-Pooler (1968)
parameters based on the neutral Turner D stability category. The con-
centrations vary by a factor of 10 for the two sets of diffusion para-
meters when the mixing ceiling is only 100 meters and vertical diffusion
is thus severely restricted.
From an analysis conducted separately from the general sensi-
tivity study, the sensitivity of the model to the choice of a system of
diffusion parameters is further illustrated in Figure 46. Model predic-
tions for a three-week portion of the 89-day set of St. Louis data were
made using first the McElroy-Pooler system of diffusion parameters based
on the Turner stability classification system. The predictions were
then repeated using the Pasquill system and Turner's stability criteria.
The resulting frequency distributions of predicted 2-hour concentration
are plotted along with the observed distribution. The distribution using
the Pasquill-Turner system yields concentrations which are 40 to 70
percent higher than the McElroy-Pooler system for corresponding fre-
quencies. This sensitivity examination is somewhat unusual in that it
represents the effect of changing all stability inputs from one set to
another.
A more detailed examination of the effect of variations in
the vertical diffusion parameter (a ) is presented in Table 27. These
values were selected from the extensive set of combinations of model
inputs used in the general sensitivity analysis. They illustrate the
complexity of the interrelationships of diffusion parameters with decay
constant, wind speed and mixing ceiling. The most noticeable effect
-------
.8
00
a"
•S
I
o
a
1000
9
8
7
6
5
4
3
2
1001
9
8
7
6
5
4
3
2
101
9
8
7
6
S
4
3
2
11
Percentage
2:-. 5 10 15 20 30 40 50 60 70 80 SS 90 95 98%
/
/
.
/
/
t
/
'
/
/
C
H
/
f
/
/
//
/
PASQUILL
Mean 258.
BSER
lean :
/
//
y
VED;
73—
/
//
/
t
•/-*
X
•IcELF
1
/
f
CTrSr
Viean 1
X
^
300
70
X
X
[£R
^S"
^
9
8
7
6
S
4
3
2
1
9
8
6
S
4
3
2
1
9
8
7
6
S
4
3
2
Figure 46. Comparison of Distributions of Two-Hour Model Predictions with Observations
Using Two Different Systems for Assigning Diffusion Parameters and St. Louis Data Set
for Ten Stations Combined
-------
Table 27. Changes in Predicted Concentrations Resulting from
Changes in the Vertical Diffusion Parameter
Recpetor
Location
Center
Center
Center
Center
Center
Center
Upwind
Center
Downwind
Center
Center
Center
Downwind
Downwind
Downwind
Downwind
Downwind
Downwind
Half-Life
No Decay
No Decay
No Decay
30 Min.
30 Min.
30 Min.
No Decay
No Decay
No Decay
No Decay
No Decay
30 Min.
No Decay
No Decay
No Decay
*5 Min.
*5 Min.
*5 Min.
Wind
Speed, m/sec
2
6
18
2
6
18
2
2
2
6
18
6
2
6
18
2
6
18
Mixing
Ceiling, m
100
100
100
100
100
100
500
500
500
500
500
500
2500
2500
2500
2500
2500
. 2500
Concentrations (/ig/m3) as a Function
of Diffusion Parameters
Pasquill,
Class E
1259
547
182
1146
478
174
12
1635
1526
545
182
476
1629
543
181
3
4
14
McElroy-
Pooler,
Class D
1207
402
134
819
347
127
3
703
440
234
78
214
200
67
22
3
2
2
McElroy-
Pooler,
Class 1
1274
425
142
850
364
134
2
520
525
173
58
158
113
38
13
2
1
1
* A small sample of computer calculation runs was run using a very large decay (5-minute half-life).
-------
is the interaction between mixing ceiling and the diffusion parameter.
Under a low mixing ceiling (100m) differences resulting from changes
in a are minimal, while under a high mixing ceiling (2500m) the dif-
fusion parameter differences yield large differences in predicted
concentrations.
The results presented here indicate that, except in the case
of a very low mixing ceiling, the variations in diffusion parameter
categories can result in large variations in predicted concentrations.
Uncertainty exists, both in defining differences in stability categories
through suitable meteorological measurements, and in relating those
stability characterisitcs to specific values of diffusion parameters.
In view of the demonstrated sensitivity of the model to changes in the
diffusion parameter values, there is a need to develop a more definitive
system which relates diffusion parameters to objectively definable
meteorological characteristics.
5.4.4 Pollutant Half-Life
The effect of pollutant half-life due to atmospheric removal
processes on short-term model (one-hour) concentrations was examined
to determine whether it was significant, and, if so, under what conditions
it was most significant. A total of 486 pairs of model inputs repre-
senting no decay and a one-half hour half-life (486 of each) were
compared to examine this question. These consisted of the 972 selected
combinations discussed in Section 5.4.1.
From these pairs it is found that a 30-minute half-life (a
decay rate of 0.0231/min) causes a significant reduction in concentration
-------
in 45 percent of the cases. The most pronounced of these many cases are
associated with relatively light wind speeds (e.g., 2 m/sec), and a
receptor location which is located significantly downwind of the high
emission area. The 30-minute half-life resulted in concentration
reductions by a factor of 50 (98%) in the most extreme case. A selected
tabulation showing variations in the effect of the decay rate with
receptor location and wind speed is given in Table 28, for a 100-meter
mixing ceiling with the Pasquill Class E (stable) diffusion parameters
and for a 500-meter mixing ceiling with the McElroy-Pooler Class 1
(unstable) diffusion parameters.
The results show that the existence of a noticeable depletion
process will have a significant effect on concentrations in the low wind
situation. This effect will be especially pronounced downwind of high
emission rate areas, where the effects are noticeable even at high wind
speeds.
5.4.5 Wind Speed and Profile Power Law
The sensitivity of the model to wind speed measured in the city
was divided into two components: the wind speed at a reference height
of 20.8 meters (a convenient and representative height which corresponded
to some available data), and the power which defines the vertical profile
of wind speed according to the relationship:
u= U] ( £) P (41)
where
u = wind speed.at height h
u-, = reference wind speed at height z-,
p = power which is a function of atmospheric stability.
-------
Table 28. Model Concentrations With and Without a 30-Minute Decay Half-Life for Selected Combinations of Model Inputs
(1)
a. Comparisons When the Mixing Ceiling is 100 m and the Diffusion Parameters are Pasquill, Class E
With 30 Min.
Half-Life
Without
Decay
Model Concentrations (/Ag/m^) for Indicated Receptor Location, Wind Speed and Decay
Upwind
2 m/sec
7
11
6 m/sec
3
4
18 m/sec
1
1
Center
2 m/sec
1146
1259
6 m/sec
478
547
18 m/sec
174
182
Downwind
2 m/sec
41
2061
6 m/sec
145
687
18 m/sec
134
229
b. Comparisons When the Mixing Ceiling is 500 m and the Diffusion Parameters are McElroy-Pooler, Class 1
With 30 Min.
Half- Life
Without
Decay
Model Concentrations (ju,g/m3) for Indicated Receptor Location, Wind Speed and Decay
Upwind
2 m/sec
1
2
6 m/sec
1
1
18 m/sec
0
0
Center
2 m/sec
412
520
6 m/sec
158
173
18 m/sec
56
58
Downwind
2 m/sec
12
525
6 m/sec
41
175
18 m/sec
35
58
tn
01
i
(1) All Concentrations were Computed Using the St. Louis Data with 51 Point Sources, an Area Grid Mesh of 0. 25 Miles, One Area Source
-------
In the initial set of 5832 combinations of model inputs defined
for the sensitivity analysis there were 2916 pairs of comparisons in
which all model inputs were identical except that the power varied from
0.15 to 0.30. In none of these comparisons did the resulting concentra-
tions vary by as much as 10 percent. As a result, this parameter was
eliminated from further sensitivity consideration and a value of 0.15
was adopted as the standard value.
In the absence of decay and plume rise the insensitivity of
concentrations to the power law parameter makes it clear that the model
concentration is inversely proportional to wind speed. This can also
be clearly seen in the model formulations. If the wind speed is constant
for all emission sources (no power law effect) it can be taken outside
the integral of Equation (9) and the summation in Equation (8) and
becomes a common factor in the summation of Equation (10). However, the
inclusion of a decay constant complicates the relationship. At the
upwind and center receptor locations the effect of decay is to slightly
reduce the inverse relationship. At the downwind location, the existence
of decay causes a reversal in the wind speed relationship over the light
to moderate wind speed range (2 to 6 m/sec). These effects are demon-
strated by the results presented in Table 29.
The modeling difficulties encountered when the wind speed is
of the order of 2 m/sec and less have been discussed in Section 4.2.3
of the validation analysis. To this must also be added the well known
measurement problems associated with obtaining a representative value
of the urban wind speed in light wind cases. While this is an infre-
quent occurrence (for the time periods reported in Section 4.0, winds
-------
Table 29. Model Concentrations as a Function of Wind Speed With a 30-Minute
Decay Half-Life for Selected Combinations of Model Inputs (*)
a. Comparisons When the Mixing Ceiling is 100 m
With Wind
Speed ofi
2 m/sec
6 m/sec
18 m/sec
Model Concentrations (jitg/m^
Pasquill, Class E
Upwind
7
3
1
Center
1146
478
174
Downwind
41
145
134
for Indicated Diffusion Parameters, Receptor Location and Wind Speed
McElroy-Pooler, Class D
Upwind
4
2
1
Center
819
347
127
Downwind
37
147
136
McElroy-Pooler, Class 1
Upwind
4
2
1
Center
850
364
134
Downwind
43
162
147
b. Comparisons When the Mixing Ceiling is 500 m
With Wind
Speed of:
2 m/sec
6 m/sec
18 m/sec
Model Concentrations (jug/m
Pasquill, Class E
Upwind
7
3
1
Center
1143
476
173
Downwind
36
109
100
for Indicated Diffusion Parameters, Receptor Locations and Wind Speed
McElroy-Pooler, Class D
Upwind
2
1
0
Center
554
214
76
Downwind
11
32
29
McElroy-Pooler, Class 1
Upwind
1
1
0
Center
412
158
56
Downwind
12
41
35
(1) All Concentrations were Computed Using the St. Louis Data with 51 Point Sources, an Area Grid Mesh of 0. 25 Miles, One Area Source
-------
were 2 m/sec and less in St. Louis 1.5% of the time, and in Chicago,
6% of the time), it remains a subject for further study from both the
modeling and the measurement points of view.
5.4.6 Mixing Ceiling
The sensitivity of model concentrations to changes in the
mixing ceiling was examined using mixing ceilings of 100, 500 and
2500 meters. Effects associated with 486 pairs of comparisons of 100-
and 500-meter ceiling heights were examined based on the 972 input
combinations obtained as described in Section 5.4.1. The most pronounced
influence was observed for diffusion parameters associated with unstable
meteorological conditions (McElroy, Class 1). The least pronounced
influence was observed for diffusion parameters associated with stable
meteorological conditions (McElroy-Pooler, Class 1). The least pronounced
locations where minimal travel from the principal effective source region
was involved. These effects may be clearly discerned in the selected
results listed in Table 30. In addition to the 100-and 500-meter mixing
ceilings, results from a 2500-meter ceiling have been added for the
downwind receptor. These results show that under stable conditions the
increased mixing ceiling has no effect, but under neutral and unstable
conditions a noticeable effect occurs. These results suggest that, at
downwind locations where significant travel from the primary emission
area is involved, the concentration is nearly inversely proportional to
the mixing ceiling. With increasingly stable diffusion parameters this
effect is reduced until with the stable type of diffusion parameter the
effect of the mixing ceiling is negligible.
-------
Table 30. Model Concentrations as a Function of Mixing Ceiling With a Wind
Speed of 6 m/sec for Selected Combinations of Model Inputs ' '
a. Comparisons When there is No Decay
With Mixing
Ceiling ofc
100m
500 m
2500 m
Model Concentrations (/tg/m^) for Indicated Diffusion Parameters, Receptor Location and Mixing Ceiling
Pasquill, Class E
Upwind
4
4
Center
587
545
Downwind
687
509
509
McElroy-Pooler, Class D
Upwind
3
2
---
Center
402
234
—
Downwind
698
147
67
McElroy-Pooler, Class 1
Upwind
3
1
Center
425
173
—
Downwind
748
175
38
b. Comparisons When the Decay Half- Life is 30 Minutes
With Mixing
Ceiling of:
100m
500 m
Model Concentrations (/ig/m^) for Indicated Diffusion Parameters, Receptor Location and Mixing Ceiling
Pasquill, Class E
Upwind
3
3
Center
478
476
Downwind
145
109
McElroy-Pooler, Class D
Upwind
2
1
Center
347
214
Downwind
147
32
McElroy-Pooler, Class 1
Upwind
2
1
Center
364
158
Downwind
162
41
01
(1) All Concentrations were Computed Using the St. Louis Data with 51 Point Sources, an Area Source Grid Mesh of 0. 25 Miles, One Area
-------
In general, the results obtained indicate that the pollutant
concentrations are approximately inversely proportional to the mixing
ceiling (as defined in these calculations) under conditions which reflect
greatest sensitivity. Under such conditions this result is in agree-
ment with the predictions of a box model, in which the pollutant is
uniformly dispersed in the vertical. Under stable conditions, or when
the predominant pollutant travel distances are small, the model is less
sensitive to the mixing ceiling. Under the defined sensitive conditions
the model is thus subject to prediction errors associated with inac-
curacies in mixing ceiling estimates. Inaccuracies will occur in the
presently used, rather indirect methods by which hourly variations in
ceiling heights must be estimated with currently available meteorological
data.
5.4.7 Wind Direction
The influence of wind direction is most critical with regard
to specific receptor locations which may, or may not be, influenced by
a strong upwind source, depending on small variations in wind direction.
Thus, receptor locations in the center of a strong emission area are
equally affected by all wind directions, while locations outside a
strong emission area are strongly influenced by whether they are directly
downwind of the high emission area.
The effect of various possible errors in wind direction on
short-term concentrations was tested using the emission data discussed
in Section 5.2. The effect of errors in the wind direction estimate
of 3, 10 and 45 degrees was examined for the two wind directions shown
previously in Figure 44. The error was allowed to vary to either side
-------
of the true direction, and the absolute values of the resulting errors
in concentration were averaged for summarizing purposes. Resultant
concentrations were evaluated at the three selected receptor locations
shown in Figure 44, using combinations of the following parameter values:
Parameter
wind speed
mixing ceiling
diffusion parameters
decay half-life
wind profile exponent
number of point sources
area source grid spacing
distribution of area
source emission heights
Value
2, 6 and 18 m/sec
100 and 500 meters
Pasquill Class E;
McElroy-Pooler Class D;
McElroy-Pooler Class 1
no decay
0.15
51
0.25 miles
all at 30 meters
Table 31 contains selected results for the case where all
sources (area and point) are considered, and the wind is first taken to
be from 349° (Table 31 a) (few large upwind point sources; see Figure 44),
and then from 020° (Table 31b) (many large upwind point sources). The
other fixed conditions are a neutral atmosphere stability and a 500 meter
mixing ceiling. The results for the 349-degree direction (Table 31a)
demonstrate the variations in model concentration which can occur with
various values of wind error, and show the change in effect of such an
error depending upon the specific receptor location considered. Errors
can go as high as about 25 percent at the central receptor, and up to
-------
Table 31. Model Concentrations with Various Degrees of Error in the Wind Direction Estimate for Selected Combinations of Model Inputs (1)
a. Comparisons when the Wind Direction is 349 (See Figure 44)
Model Concentration:
No Wind Error
Absolute Error in Mo del Concentration:
' With 3 Degree Wind Error
10 Degree Wind Error
45 Degree Wind Error
Model Concentrations (/tg/m ) and Absolute Errors for Indicated Receptor Location and Wind Speed
Upwind
2 m/sec
3
0
1
28
6 m/sec
1
0
0
9
18 m/sec
0
0
0
3
Center
2 m/sec
603
76
73
143
6 m/sec
201
25
24
48
18 m/sec
67
8
8
16
Downwind
2 m/sec
469
80
227
445
6 m/sec
156
27
76
148
18 m/sec
52
9
25
49
o
b. Comparisons when the Wind Direction is 020 (See Figure 44)
Model Concentration:
No Wind Error
Absolute Error in Model Concentration:
With 3 Degree Wind Error
10 Degree Wind Error
45 Degree Wind Error
Model Concentrations (/xg/m ) and Absolute Errors for Indicated Receptor Location and Wind Speed
Upwind
2 m/sec
7
12
18
75
6 m/sec
2
4
6
25
18 m/sec
1
1
2
8
Center
2 m/sec
1055
227
716
308
6 m/sec
352
76
239
103
18 m/sec
117
25
80
34
Downwind
2 m/sec
23
18
154
125
6 m/sec
8
6
51
42
18 m/sec
3
2
17
14
en
ro
i
(1) All concentrations were computed using the St. Louis data with 51 point sources, an area grid mesh of 0. 25 miles, one area source height,
-------
100 percent at the downwind location. Note that the pattern is consistent
at the downwind location (increasing concentration error with increasing
wind error), but not at the central site. Changes in the stability
condition (to unstable, or to stable) and/or the mixing ceiling
(to 100m) show different absolute values of concentration and error,
but quite similar patterns. The results in Table 31 b for the 020®
direction should be interpreted bearing in mind that this wind shift
consideration transfers the relative locations of the up- and downwind
sites to some extent in the acrosswind direction (see Figure 44).
Errors at the central site can now become significantly larger (up to
about 70% ) because of the significant upwind point sources, and the
other two sites show increased errors, with the "downwind" site losing
the consistent increase in error with increased direction error which
it had in Table 31a.
These results show the extreme variability of this effect,
and the importance of obtaining a representative wind direction to
enable adequate definition of the individual short-term concentrations
at certain specific types of site locations.
For long-term concentrations, errors associated with the mean
wind direction during any given period tend to be compensated for during
other periods, when a sufficiently large sample is used to construct the
long-term mean and frequency distribution of short-term concentrations.
The use of a statistical sampling plan to select wind directions in
constructing long-term averages obviates the need for considering the
problem of defining an appropriate class interval size for characteriz-
ing wind directions in a long-term concentration model. The results of
-------
statistical sampling plan evaluation discussed in Section 4.3.2 show
that long-term concentrations consturcted using an unbiased sample of
about 100 short-term periods will properly reflect the effects of wind
direction variations.
5.4.8 Diurnal Variation in Emission Rates
The sensitivity of the model to diurnal variations in S02
emission rates is evaluated in this section by analyzing the factors
which affect the emission rate estimations. These factors are repre-
sented in algorithms which are used to define the spatial distribution
of emissions for any given hour and the variations of the emissions
from hour to hour. The algorithms used in the validation analysis of
this study to represent St. Louis and Chicago emissions are given in
Appendixes B and C, respectively. The inputs to these algorithms which
vary diurnal ly, and thus give rise to the diurnal emission variations,
are temperature, electric power load at generating stations, and hour
of the day. Since the errors associated with measurement of these
inputs are small, it is clear that the errors which are more critical
to model sensitivity are those associated with the assumptions in the
emission algorithms which convert these inputs into the distributions
of SOp emissions. In the discussion which follows, these sources of
error and their impact on model calculations are identified and
characterized.
Sulfur dioxide emissions are characterized as arising from
one of three types of operations, namely, space heating, electric power
generation, and industrial processing. All three operations emit S02
-------
as a result of fuel consumption. However, industrial processing may
also include some direct emissions of SOp. The amount of annual emis-
sions associated with each type of operation for each point source and
each subdivision of an area source is determined from emission inventory
surveys. These surveys may provide more detailed breakdowns including
seasonal or even monthly emissions. These annual (or other more
frequent) values of emissions determine the spatial distribution of
emissions by type of operations. These emission estimates also act
as scaling factors for diurnal variations which are applied to each
type of operation. Thus, they determine the magnitude of the diurnal
variation at each location. An error in one of the estimates creates
a systematic error in the model predictions for locations in the vicinity
of the error estimate. However, the occurrence of this type of error
will be distributed randomly over all sources. If concentrations are
considered at a number of locations widely dispersed over the urban
area, the overpredictions and underpredictions due to this type of error
will tend to balance at any given time.
Diurnal variations in emissions from electric power generation
are estimated by means of linear relationships with hourly electric power
loads of specific generating units (e.g., see Section 4.1). The error
associated with these estimates is small. The uncertainty of emission
estimates for proposed power plants would be greater. However, the
emission rate estimates for this type of operation are judged to be
sufficiently well represented that they have less impact on model sensi-
tivity than other errors which need to be considered.
-------
Diurnal variations in emissions from industrial processing
are allocated on the basis of scaling factors which define the percent
of the peak operating capacity which is applicable to each day of the
week and to each eight-hour shift of the day. The algorithms used for
St. Louis and Chicago allow for three types of days of the week, namely,
weekday, Saturday, and Sunday or holiday. An input to the algorithm
designates the type of day which is assigned to each particular hour
for which a set of emission rates is requested. In the Chicago case,
sufficient data were collected on each point source to assign a specific
type of day to each day of the week. For example, for some sources
every day is treated as a weekday, and for some other sources, Monday
is designated as a holiday. For national holidays, all sources are
assigned holiday schedules. During any particular day and hour,
variations in actual emissions from these scheduled average emissions
may be considerable. This is because industrial operations must respond
to fluctuations in demand and to breakdowns in equipment. These influences
will result in errors in emission rate estimates in the immediate vicinity
of individual sources. However, it is probable that these errors will
be randomly distributed over an urban area at any given time. When
concentration predictors are considered over a number of widely dispersed
locations, the overprediction and underprediction errors will tend to
balance. Only detailed analysis of production records of individual
plant operations can be expected to yield more accurate emission
estimates.
-------
Emissions from space heating operations reflect diurnal cycles
in demand for heat and in temperature fluctuations. Nhile these two
factors tend to have opposite diurnal cycles, the resulting diurnal
pattern of heating operations is rarely uniform. The emission rates for
space heating operations are estimated by the following equation:
Q(t) = qT [TR - T(t) - A(t,d)], [T(t) + A(t,d)] £TR (42)
where
Q(t) = emission rate for time t
qT = emission rate per degree
TR = reference temperature (usually 65°F)
T(t) = temperature for time t
A(t,d) = temperature correction for time t and type of day d.
The temperature correction is an empirical factor to account for the
diurnal variation in the activities of a city which affect its demand
for fuel. Corrections were determined for St. Louis for each hour of
the day for weekdays, Saturdays and Sundays by Turner (1968). Two sets
of correction factors were derived. One is applicable to residential
space heating. The other is applicable to commerical and industrial
space heating. A further correction has been applied to emissions for
the Chicago area where residential heating emissions were borken down
between large apartment buildings (20 dwelling units or more) and small
apartment buildings and residences (low-rise). A stoking factor, or
"janitor" function, which' permits no emissions from 11 p.m. to 5 a.m.
-------
(3 a.m. if the temperature is below 5°F), and requires a 50 percent excess
for the first two hours of the day (recommended by Roberts, et al., 1970)
was applied to low-rise residential emissions.
Random errors associated with individual sources of these types
may be expected. However, these are probably small. A more serious type
of potential error is associated with assumptions regarding the response
of such emissions to sudden temperature changes and to unseasonal temper-
atures. This type of error will be systematic and city-wide. For example,
since buildings provide insulation between outside air and inside air,
there will usually be a significant time lapse between a sudden temperature
change and the time when its influence on inside air requires full com-
pensation by increased fuel consumption. However, the emission algorithm
assumes this adjustment takes place immediately. The effect of this lag
on fuel consumption was pointed out by Turner (1968). His findings for a
severe temperature change of 15°F in one hour suggest that a six to
eight hour period of adjustment is required. The error in emission
rates during the hour in which the change occurs is shown in Turner's
example to be factor of two too high and to return to no error in
six to eight hours. While this type of error is critical to short-term
model sensitivity, it is not important for long-term mean concentrations
because sharp temperature changes are rare over a long period.
Systematic errors due to unseasonal temperatures can also
result from the fact that many heating systems, especially in large
buildings, are only partially controlled by temperature thermostats.
Large amounts of excess heat may be generated when the temperature is
unseasonably warm, and the emission estimate will be, correspondingly,
-------
too low. Similarly, insufficient heat will be generated during
unseasonably cold periods, with an emission overestimate. This type
of error will persist for periods of several hours to several days
depending on the duration of the unseasonable temperatures. These
errors should tend to compensate each other over long-term periods.
However, systematic errors will exist in short predictions and, because
the contribution of these sources frequently represents most, or all, of
the affected concentration, the concentration error will be directly
proportional to the emission error. A careful study of the magnitude
and duration of heating emissions is required to determine the nature
of heating system response to these types of situations.
It is concluded that mean long-term concentrations are not
sensitive to errors in the diurnal variations in emission rates. However,
it is clear that individual short-term concentrations will be proportionally
sensitive to significant errors in the diurnal variation factor. During
winter seasons, when ground level concentrations are primarily due to
emissions from space heating operations, the short-term concentrations
may be in error by as much as a factor of two due to errors in temperature
dependent emission rates.
5.5 FINDINGS
A summary of the preceding results on the sensitivity of the
Gaussian plume type of multiple source urban diffusion model is given
below. These results are based on calculated changes in short-term
concentrations at a point associated with changes in model inputs. Three
types of receptor locations are considered within a selected representative
-------
pattern of urban area source emissions. Combinations of two or three
values for each of 10 input parameters were examined in drawing conclusions
regarding the sensitivity of the model. A change in concentrations at a
point of at least 50 percent due to a change in an input value was
considered a "sensitive" effect. Sensitivities in many cases are complex
and interrelated, and the individual section analyses should be consulted
for details.
The sensitivity findings for each input parameter are sum-
marized as follows:
• Spatial Variability of Emissions. The fineness of the grid
spacing used to represent area sources must be consistent
with the dimensions of real spatial variability. For
example, if the standard deviation of emission rates defined
by quarter-mile squares is as much as 50 percent of the
mean emission rate for a square-mile, a change from a grid
spacing of a mile to a quarter-mile produces "sensitive"
changes. Furthermore, when a small grid spacing is used,
large individual sources may be aggregated into the area
source without producing "sensitive" changes.
• Vertical Distribution of Area Source Emissions. On the
basis of comparisons between concentrations calculated
using a single height for area source emissions, with
concentrations calculated using a distribution of heights
(50% of emissions at mean height, 25% at one-half the
mean height and 25% at 1.5 times the mean height), it is
concluded that this is not a "sensitive" input. No
"sensitive" changes were observed in the comparisons.
• Vertical Diffusion Parameter. Under conditions which do
not involve low (e.g., 100m) mixing ceilings, calculated
short-term concentrations were shown to vary by a factor
of from 3 to 10 or more when the diffusion parameters
are varied from the Pasquill Class E to the McElroy-
Pooler Class 1 (stability classes defined in Section 2.4.4).
This large sensitivity indicates the need for measure-
ments of atmospheric conditions which are clearly related
to differences in diffusion conditions (i.e., values
of az).
-------
Pollutant Half-Life. In order to adequately estimate
concentrations at locations downwind of major emission
areas, information on the pollutant half-life due to
atmospheric removal processes must be available. As an
extreme example of the wide spectrum of comparisons
obtained, highly significant effects were computed for
a location 30 km downwind of the center of the urban area
when comparing no decay with a 30-minute half-life.
Concentrations with no decay were 50 times greater than
concentrations with decay.
Mind Speed and Profile Parameter Value. Calculated
concentrations are insensitive to changes in the wind
profile power law exponent. Maximum changes in calculated
concentrations were 10 percent when the exponent was varied
from 0.15 to 0.3. Calculated concentrations are inversely
proportional to wind speed when pollutant decay is negligible,
Decay acts to reduce the "sensitivity" effect due to wind
speed at locations downwind of the primary emission area.
Mixing Ceiling. Changes in the mixing ceiling produce
the greatest sensitivity at locations for which most of
the pollutant arrives after traveling large distances.
At locations 30 km downwind from the primary emission
area, the calculated concentration varied inversely with
the mixing ceiling in unstable conditions. This maximum
"sensitivity" effect decreases with increasing stability
and is negligible in stable conditions.
Wind Direction. A highly variable sensitivity exists for
short-term single station concentrations. Small changes
in wind direction (e.g., 3° azimuth) may result in extremely
"sensitive" effects at some locations. Short-term con-
centration estimates are thus dependent on accurate wind
direction estimates.
Diurnal Variation in Emission Rates. Errors in reported
annual or seasonal emissions from particular sources will
result in proportionally systematic errors in calculated
diurnal variations in concentrations in the vicinity of
the emission error. Because the emission algorithms are
temperature based, unseasonable temperatures or sudden
atmospheric temperature changes will result in correspond-
ing systematic errors in predicted short-term concentrations
at all locations for a short lag period, until an adjust-
ment in space heating operations occurs. Over a long-
term period errors in short-term concentrations due to
differences, between actual and estimated emission rates
will tend to be balanced in the combined frequency
distribution of short-term concentrations for several
locations.
-------
Long-Term Concentrations. Sensitivity of the long-term
model is defined as it relates to the model described
in Section 4.3. Errors in model inputs will include the
sensitivity effects summarized above for short-term
concentrations. However, over a long-term period the
random errors tend to compensate each other. It is
evident that this is reasonably true for model inputs used
for validation analysis in this study, since long-term con-
centration calculations averaged over several locations
were generally found to be in agreement with observed
concentrations. Furthermore, the combined frequency
distribution of observed short-term concentrations at all
locations considered were well represented by the model
calculations.
The long-term model can be expected to show comparable
sensitivity to some but not all, of the systematic (as
opposed to random) changes in the parametric inputs
described above for the short-term model. Those which can
impact are the spatial variability of emissions, the
vertical diffusion parameter selection, pollutant half-life,
wind speed and mixing ceiling.
-------
Section 6.0
-------
Section 6.0
CONCLUSIONS AND RECOMMENDATIONS
The results of this evaluation of the validity and the sensi-
tivity of the Gaussian plume type of urban diffusion model provide basic
definition of the capabilities and limitations of the model for simulating
urban air quality. The following represent the important conclusions
derived from the validation and sensitivity findings, respectively.
6.1. CONCLUSIONS FROM VALIDATION ANALYSIS
From the findings in the validation study in which comparisons
are made with data from two cities for one and three month periods, and
the model is implemented as described in Section 4.0, it is concluded
that:
1. For individual values of short-term (1- or 2-hour) concen-
trations at individual receptor locations, the predicted concentrations
show large deviations from the observed concentrations. However, a large
number of such comparisons over a month, or a season, produce frequency
distributions of predicted concentrations which compare quite well
with the observed distributions. The individual frequency deciles are
generally within a factor of two or less of each other. No single con-
trolling factor could be found which consistently accounted for a signif-
icant fraction of the deviations of the individual short-term values.
2. Predicted long-term (monthly or seasonal) concentrations
based on averaging of the calculated short-term concentrations, show
consistent good agreement with observations, with a root-mean-square
error equal to about half the mean, and a slight tendency to over-
estimate. This contrasts with results from other long-term models which
generally overestimate significantly; the improvement is concluded to
be largely due to the combination of two factors, one being the process
of accounting for the diurnal correlation of meteorological and emission
parameters, and the other, the use in this model of urban-derived
diffusion parameters (McElroy-Pooler parameters based on the Turner
stability classifications).
-------
3. A technique has been devised for calculating the long-term
estimates described above without the necessity of calculating every
short-term concentration involved. This is the statistical process of
proportionate stratified sampling, and is an effective method for select-
ing a limited set of short-term periods which adequately define the
distribution in the long-term period. As few as 5 to 10 percent of the
total number of short-term periods involved will describe the distribution
adequately, if the sampling is done as described in Section 4.3.2.
As noted in Section 4.2.3, the calm or light wind case repre-
sents a special problem; while suggestions for empirical treatment are
given in that section, the subject requires further study.
6.2 CONCLUSIONS FROM SENSITIVITY ANALYSIS
From the findings in the sensitivity analysis, it is concluded
that, generally, concentrations predicted by the short-term model were
found to be insensitive to changes in the vertical distribution of area
source emissions and the wind profile parameter value. The concentrations
are sensitive, under some circumstances, to changes in spacial variability
in emissions, vertical diffusion parameter, pollutant half-life, wind speed,
mixing ceiling, wind direction and diurnal variation in emission rates.
The long-term model (defined in Section 4.3) is sensitive in some cases to
the spatial variability of emissions, vertical diffusion parameters,
pollutant half-life, wind speed and mixing ceiling.
More specifically, selected principal sensitivity results are
as follows:
1. A fine grid spacing (0.25 mile on a side) is necessary for
area source definition if the emissions vary significantly over a square
mile (standard deviation of emission rates as much as 50 percent of the
mean). Conversely, if a .coarse grid spacing is used, then more care
must be taken in determining the significant point sources to be considered.
-------
2. The model shows sensitivity to selection of the stability
class assigned in any given calculation of short-term concentration, and
additionally, overall sensitivity to the choice of one set of diffusion
parameters (e.g., McElroy-Pooler) over another (e.g., Pasquill).
3. Mhen pollutant decay is negligible, the model shows an
inverse proportionality relationship with wind speed; in the presence of
a pollutant half-life of 30 minutes, the variations of concentration with
wind speed at downwind locations are markedly reduced.
4. Mixing ceiling shows its maximum effect in terms of sensi-
tivity at large downwind distances under stable conditions.
5. Sensitivity to wind direction is highly variable, and of
course shows the most impact at individual stations downwind of major
sources.
6. Diurnal variations in emission rates are governed in the
model algorithms by air temperatures, and in unseasonably warm or cold
periods will result in proportionate over- or underestimates of emissions,
and hence concentrations; these will exist for a short (several hours to
several days) period, gradually recovering to correct the emission values.
6.3 RECOMMENDATIONS
On the basis of this study, it is recommended that:
1. Consideration should be given to EPA's promulgation, for
general use, of the long-term version of the Gaussian plume type of
urban diffusion model described in this report, using the proportionate
stratified sampling concept, for the calculation of long-term means and
short-term distributions of pollutant concentrations. Appropriate
documentation should be prepared, together with processing procedures
for meteorolgical and emission inputs, and program manuals for an
optimized computer program (to be developed).
2. Study should be instituted on the various measurement and
classification problems described in this report, particularly (1) the
relationship of the appropriate meteorological measurements to the
corresponding stability categories, (2) mixing ceilings, and (3) wind
directions. In addition, the calm or light wind case should be studied to
to determine a satisfactory and objective method of predicting pollutant
concentrations under these circumstances.
-------
Section 7.0
-------
Section 7.0
REFERENCES
Booras, S.G. and C. E. Zimmer. 1968. "A Comparison of Conductivity
and West-Gaeke Analyses for Sulfur Dioxide." Journal of the Air
Pollution Control Association, 18(9), p. 612.
Briggs, G. A. 1969. Plume Rise. U.S. Atomic Energy Commission,
Division of Technical Information Extension, Oak Ridge, Tennessee.
Calder, K. L. 1969. A Narrow Plume Simplification for Multiple Source
Urban Pollution Models"! (Informal unpublished note.] December 31, 1969,
Calder, L. K. 1970. "Some Miscellaneous Aspects of Current Urban
Pollution Models." (Stern (ed.), Proceedings of Symposium on Multiple
Source Urb a n Pif f us i o n Mode1s. U.S. Environmental Protection Agency,
Air Pollution Control Office, Research Triangle Park, North Carolina.)
Calder, K. L. 1970. A Climatological Model for Multiple Source Urban
Air Pollution. (Presented at the First Meeting of NATO/CCMS Panel
on Modeling, October 9, 1970, Frankfurt, Germany.)
Chamot, C. et al. 1970. A Computerized Air Pollution Data Management
System. ANL/ES-CC-006, Argonne National Laboratory, Argonne, Illinois.
Clarke, J. F. 1964. "A Simple Diffusion Model for Calculating Point
Concentrations for Multiple Sources." Journal of the Air Pollution
Control Association, Vol. 14, No. 9, pp. 347-352.
Cramer, H.E. et al. 1964. Meteorological Prediction Techniques and
Data System. GCA Technical Report, No. 64-3-6. Bedford, Massachusetts.
Geophysics Corporation of America.
Croke, E. J. et al. 1968. City of Chicago Air Pollution System Model.
(First Quarterly Progress Report.) ANL/ES-CC-001.Argonne National
Laboratory, Argonne, Illinois.
Croke, E. J. et al. 1968. City of Chicago Air Pollution System Model.
(Second Quarterly Progress Report.)ANL/ES-CC-002.Argonne National
Laboratory, Argonne, Illinois.
Croke, E. J. et al. 1968. City of Chicago Air Pollution System Model.
(Third Quarterly Progress Report.)ANL/ES-CC-003.Argonne National
Laboratory, Argonne, Illinois.
Croke, E. J. and J. J. Roberts 1971. Chicago Air Pollution Systems
Analysis Program Final Report. ANL/ES-CC-009. Argonne National
Laboratory, Argonne, Illinois.
-------
Davidson, B. 1967. "A Summary of the New York Urban Air Pollution
Dynamics Research Program." Journal of Air Pollution Control
Association, 17, pp. 154-158.
DeMarrais, G. A. 1959. "Wind-Speed Profiles at Brookhaven National
Laboratory." Journal of Meteorology, 16 (4), pp. 181-190.
Fortak, H. G. 1969. "Numerical Simulation of the Temporal and
Spatial Distributions of Urban Air Pollution Concentrations."
Presented at Symposium on Multiple Source Urban Diffusion Models 27-30
October 1969, Chapel Hill, North Carolina.
Fuguay, James J., Charles L. Simpson, and W. Ted Hinds. 1964. "Predic-
tion of Environmental Exposures from Sources Near the Ground Based
on Hanford Experimental Data." Journal of Applied Meteorology, 3(6),
pp. 761-770.
Gifford, Frank A., Jr. 1961. The Problem of forecasting Dispersion
in the Lower Atmosphere. U.S. Atomic Energy Commission, Division of
Technical Information Extension, Oak Ridge, Tennessee.
Hansen, Morris H., William N. Hurwitz, and William G. Madow. 1953.
Sample Survey Methods and Theory. Vol. II, Theory. John Wiley &
Sons, New York.
Hay, J. S. and F. Pasquill. 1957. "Diffusion from a Fixed Source at
a Height of a Few Hundred Feet in the Atmosphere." Journal of Fluid
Dynamics, Vol. 2, pp. 299-310.
Hilst, G. R. and N. E. Bowne. 1966. A Study of Diffusion of Aerosols
Released from Aerial Line Sources Upwind of an Urban Complex. The
Travelers Research Center, Inc.
Hilst, G. R. 1970. "The Sensitivities of Air Qualtiy Predictions to
Input Errors and Uncertainties." (Stern (ed.), Proceedings of
Symposium on Multiple-Source Urban Diffusion Models. U.S. Environ-
mental Protection Agency, Air Pollution Control Office, Research
Triangle Park, North Carolina.)
Holland, J. Z. 1953. A Meteorological Survey of the Oak Ridge Area:
Final Report Covering the Period 1948-52. USAEC Report ORO-99.
Oak Ridge, Tennessee: U.S. Weather Bureau.
Holzworth, G. C. 1967. "Mixing Depths, Wind Speeds, and Air Pollution
for Selected Locations in the United States." Journal of Applied
Meteorology, Vol, 6, No. 6, pp. 1039-1044.
Holzworth, G. C. 1964. "Estimates of Mean Maximum Mixing Depths in the
Contiguous United States." Monthly Weather Review, Vol. 92,
pp. 235-242.
Islitzer, N. F. 1961. "Short Range Atmospheric - Dispersion Measure-
ments from an Elevated Source." Journal of Meteorology, 18(4).
-------
Koogler, J. E. et al. 1967. "A Multivariable Model for Atmospheric
Dispersion Predictions." Journal of the Air Pollution Control
Association, Vol. 17, No. 4, pp. 211-214.
Landsberg, H. H., L. L. Fischran, and J. L. Fisher. 1963. Resources in
America's Future. Johns Hopkins Press. Baltimore, Maryland.
Lettau, H. H. 1957. "Computation of Richardson Numbers, Classification
of Wind Profiles, and Determination of Roughness Parameters." Explor-
ing the Atmosphere's First Mile, Vol. 1, New York, Perganen Press,
pp. 328-331.
Lucas, D. H. 1958. "The Atmospheric Pollution of Cities." International
Journal of Air Pollution, Vol. 1, pp. 71-86.
Ludwig, F. L., W. B. Johnson, et al. 1970. A Practical Multipurpose
Urban Diffusion Model for Carbon Monoxide. Contracts CAPA-3-68 and
and CPA 22-69-64. Menlo Park, California: Standford Research
Institute.
Manowitz, B. et al. 1970. "The Isotope Ratio Tracer Method: Applica-
tions in Atmospheric Sulfur Pollution Studies." Paper presented at
the Second International Clean Air Congress of the International Union
of Air Pollution Prevention Association, December 6-11, 1970,
Washington, D. C.
Marsh, K. J. and V. R. Withers. 1969. "An Experimental Study of the
Dispersion of the Emissions from Chimneys in Reading - III: The
Investigation of Dispersion Calculations." Atmospheric Environment,
Vol. 3, pp 281-302.
Martin, Delance 0. and Joseph A. Tikvart. 1968. A General Atmospheric
Diffusion Model for Estimating the Effects of One or More Sources on
Air Quality. Presented at 61st Annual Meetings of the Air Pollution
Controls Association, St. Paul, June 1968. APCA-68-148.
Martin, D. 0. 1971. "An Urban Diffusion Model for Estimating Long
Term Average Values of Air Quality." Journal of the Air Pollution
Control Association. Vol. 21, No. 1, pp. 16-19.
McCaldin, R. 0. and R. S. Sholtes. 1970. Mixing Height Determinations
by Means of an Instrumented Aircraft. Contract No. CPA 22-69-76.
Gainesville, Florida: University of Florida.
McElroy, James L. 1969. "A Comparative Study of Urban and Rural
Dispersion." Journal of Applied Meteorology. (February) 8(1),
pp. 19-31.
-------
McElroy, James L. and F. Pooler, Jr. 1968. St. Louis Dispersion Study
Vol. II - Analysis. Publication No. AP-53. Raleigh, North Carolina,
National Air Pollution Control Administration.
Milford, S. N. et al. 1970. "Air Pollution Models of the New York/
New Jersey/Connecticut Air Quality Region." Presented at the 63rd
Annual Meeting of the Air Pollution Control Association, St. Louis,
June 14-18, 1970.
Milford, S. N. et al. 1970. "Comparisons of Air Pollution Models with
Aerometric Data for the Air Quality Region Centered on New York City."
Presented at the Second International Clean Air Congress of the
International Union of Air Pollution Prevention Association,
December 6-11, 1970, Washington, D. C.
Miller, Marvin E. and George C. Holzworth. 1967. "An Atmospheric
Diffusion Model for Metropolitan Areas." Journal of the Air Pollution
Control Association, Vol. 17, No. 1, pp. 46-50"! (January).
Milly, G. H. 1958. Atmospheric Diffusion and Generalized Munitions
Expenditures. ORG Study 17. Operations Research Group, U.S. Army
Edgewood Arsenal, Maryland.
Munn, R. E. 1966. Descriptive Micrometeorology. Advances in Geophysics,
Supplement 1. New York: Academic Press.
Ozolins, G. and R. Smith. 1966. A Rapid Survey Technique for Estimating
Community AirPollution Emissions. U.S. Department of Health, Education
and Welfare, Public Health Service Publication No. 999-AP-29. Raleigh,
North Carolina: National Air Pollution Control Administration.
Pasquill, F. 1962. Atmospheric Diffusion. London: D. VanNostrad Co.
Ltd.
Pooler, Francis, Jr. 1961. "A Prediction Model of Mean Urban Pollution
for Use with Standard Wind Roses." International Journal of Air and
Water Pollution. Vol. 4, No. 314, September 1961.
Roberts, J. J. et al. 1970. Chicago Air Pollution Systems Analysis
Program: A Multiple-Source Urban Atmospheric Dispersion Model.
ANL/ES-CC-007. Argonne, Illinois: Argonne National Laboratory.
Shikiya and MacPhee. 1969. "Multi-Instrument Performance Evaluation of
Conductivity-Type Sulfur Dioxide Analyzers." Journal of the Air
Pollution Control Association, 19 (12), pp. 943-945.
Singer, I. A. et al. 1970. Comparative Studies of Urban and Rural Wind
Climatology. Unpublished note carried out under auspices of U.S.
Atomic Energy Commission.
-------
Singer, I. G. and Maynad E. Smith.
Brookhaven National Laboratory."
Water Pollution. 10, 125-135.
1966. "Atmospheric Dispersion at
International Journal of Air and
Meteorology and Atomic Energy.
Slade, D. H. (Ed.). 1968.
Laboratories, Research Laboratories
Administration for the Division of Reactor
Published by U.S. Atomic Energy Commission,
Information, Oak Ridge, Tennessee.
Air Resources
Environmental Science Services
Development and Technology
Division of Technical
Smith, E. (Ed.). 1968. Recommended Guide for the Prediction of the
Dispersion of Airborne Effluents.
Society of Mechanical Engineers.
New York, N. Y.: The American
Smith, M. E. and I. A. Singer. 1966. "An Improved Method of Estimating
Concnetrations and Related Pheomena from a Point Source Emission."
Journal of Applied Meteorology, Vol. 5, No. 5, pp. 631-639,
October 1966.
Smith, M. E. and I. A. Singer. 1970. "Applications of a Multi-Source
Model to an Urban Study." Presented at the Second International
Clean Air Congress of the International Union of Air Pollution
Prevention Association, December 6-11, 1970, Washington, D. C.
Stern, A. C. (Ed.).
Diffusion Models.
Control Office,
from GPO.)
1970. Proceedings of Symposium on Multiple-Source
_ U.S. Environmental Protection Agency, Air Pollution
Research Triangle Park, North Carolina. (Available
Sutton, 0. G.
New York.
1953. Micrometeorology. McGraw-Hill Book Co., Inc.
Turner, D. Bruce. 1969 and 1970. Workbook of Atmospheric Dispersion
Estimates. (Revised). PHS. Pub. No. 999-AP-26. U.S. Department
of Health, Education and Welfare.
Turner, D. Bruce. 1968. "The Diurnal and Day-to-Day Variations of
Fuel Usage for Space Heating in St. Louis, Missouri." Atmospheric
Environment, Vol. 2, pp. 334-351.
Turner, D. Bruce. 1968. Unpublished notes. (Personal Communication.)
Turner, D. Bruce. 1964. "A Diffusion Model for an Urban Area."
Journal of Applied Meteorology, 3(1), 83.
Turner, D. Bruce and N. G. Edmisten. 1968. St. Louis SOg Dispersion
Model Study - Basic Data. (Unpublished). Raleigh, North Carolina:
National Air Pollution Control Administration.
-------
TRW Systems Group. 1969. "Air Quality Display Model." Contract
No. PH 22-68-60. Prepared for Department of Health, Education and
Welfare, Public Health Service, Consumer Protection and Environmental
Health Service, National Air Pollution Control Administration,
Washington, D. C. (Available from CFSTI as PB 189-194.)
U.S. Department of Commerce, Bureau of Census. 1962. Statistical
Abstracts o^f the U.S. Washington, D. C.
Urone, P. and W. H. Schroeder. 1969. "S02 in the Atmosphere: A Wealth
of Monitoring Data, but Few Reaction Rate Studies." Environmental
Science and Technology, Vol. 3, No. 5.
Weber, E. 1970. "Contribution to the Residence Time of Sulphur Dioxide
in a Polluted Atmosphere." Journal of Geophysical Research, Vol. 75,
No. 15, pp. 2909-2914.
Weber, E. 1970a. "Determination of the Life Time of S02 by Simultaneous
COg and S02 Monitoring." Paper presented at the Second International
Clean Air Congress of the International Union of Air Pollution
Prevention Associations, December 6-11, 1970, Washington, D. C.
-------
Appendix A
A SUMMARY OF PREVIOUS IMPLEMENTATIONS OF THE GAUSSIAN PLUME TYPE
-------
Appendix A
A SUMMARY OF PREVIOUS IMPLEMENTATIONS OF THE GAUSSIAN PLUME TYPE
OF URBAN DIFFUSION MODEL
This appendix contains summaries of implementations of the
Gaussian plume model to analyze pollutant concentrations from multiple
sources in an urban environment. The summaries include:
t Scope of application addressed by the investigator
• Treatment given to emission data
• Computational equations used
• Meteorological and other non-source data
• Validation results obtained.
The implementations summarized are identified by the name of
the investigator and indexed by exhibit number as follows:
Investigator Exhibit
Lucas (1958) A-2
Pooler (1961) A-3
Turner (1964) A-4
Clarke (1964) A-5
Miller & Holzworth (1967) A-6
Koogler, et al. (1967) A-7
Martin & Tikvart (1968) A-8
Fortak (1969) A-9
Mil ford, et al. (1970) A-10
Ludwig, et al. (1970) A-ll
Calder (1970) A-12
Singer & Smith (1970) A-13
A list of the symbols used in the summaries is presented as
Table A-l. Symbols used only with regard to one investigator's work
are introduced as they are encountered.
-------
The descriptions here are not comprehensive in that they do
not attempt to present or discuss all the investigator's findings. The
summaries present the basic method used to implement the model and the
results of validating this model. Further results obtained by fitting
the model to data or examining the effect of additional refinements
beyond the scope of the basic model are not discussed.
-------
Table A-l. Symbols
Symbol
Explanation
C or
Q, Qi
V
D
AH, AHi
h, h.
S
y v v
A > A-j > ' »
yv
' i
6i
6
U
Seasonal or long-term mean concentration over a
sequence of short-term (steady-state) periods
Concentration due to area sources
Concentration due to point sources
Short-term (steady-state) concentration
Emission rate per unit area (area source)
Emission rate (point source)
Stack effluent velocity (point source)
Stack diameter (point source)
Stack effluent temperature (point source)
Physical height of source
PI ume ri se
Effective source height
Alongwind length of area source
Alongwind distance between source and receptor
locations
Crosswind distance between source and receptor
locations
Azimuth from receptor to source
Mean wind direction
Mean wind speed
Representative wind speed for jth wind speed class
Empirical diffusion parameter
(Continued)
-------
Table A-l. Symbols (Concluded)
Symbol
Explanation
f(
L'Lk
T
'50
t, t
"V
Frequency distribution of specified variables
Mixing layer depth or ceiling
Ambient air temperature
Pollutant airborne half-life
Travel time from source to receptor with mean
wind speed
Horizontal diffusion parameter
Vertical diffusion parameter
-------
Exhibit A-2. Model Implementation by Lucas (1958)
I. Scope:
Short-term and seasonal mean concentrations from domestic fires
in an urban area.
II. Treatment of Sources:
Represent emissions by a uniform area source (i.e., q, S and h)
for London
ft SO
q = 1.7 x 10"8 ? 2 (1954 estimate)
ft^sec
Under "normal" meteorological conditions
h = 30 + 5 (jj)3
Under stable meteorological conditions
u, ft/sec
h, ft
<1
80
2
45
3
35
>5
30
III. Computational Equations:
Approximate a Gaussian plume for a point source by a uniform cone
(allowing for impenetrability of the ground and mass continuity)
and integrate the resulting equation over the source area to get
concentration as a function of distance from the upwind edge of
the source area. For short-term with x < S,
/? Ku
With x > S,
C2(x) =
h'
4
K2x2 K4x4(2)2! K6x6(3)3!
- S)
-------
Estimate long-term concentrations using short-term concentrations
associated with wind speed classes as follows,
C(x) = E f(u..) Ck(x).
J
IV. Meteorological Parameters:
Use mean wind speed or frequency distribution of wind speed classes,
The diffusion parameter K is 0.041 for "normal" conditions; 0.014
for stable conditions (based on ft, sec unit system). Use normal
value for long-term estimates.
V. Validation Results:
Pollutant, Location, Time Observed Calculated
S02, London, 1954-55 0.114 ppm 0.11 ppm
winter season
S02, London, 5-8 Dec., 1952 1.5 ppm 9.0 ppnT3'
(a) A two-level diurnally varying emission rate (i.e., constant
for 15 hours, zero for 9 hours) was used in this prediction. The
calculation as presented graphically by Lucas appears to be related
to a puff, rather than a plume, type of calculation in which emis-
sions during one period are traced during the succeeding period in
which zero emissions occur. The details of the calculation are
not reported.
-------
Exhibit A-3. Model Implementation by Pooler (1961)
I. Scope:
Mean monthly pollutant concentrations resulting from multiple
pollutant sources at a gridwork of receptor locations.
II.. Treatment of Sources:
Represent emissions (Q) from each square mile of an urban area by
a point source in the center. Treat all emissions as having an
effective height of 30 m. Receptor concentrations are calculated
for the same locations used as point sources. Emissions (q ) from
the square centered on a receptor are treated as a uniform
circular area source.
III. Computational Equation:
Treat the plume from a point source as having a Gaussian distri-
bution of pollutant concentrations in the vertical and a uniform
distribution over a sector of angle -rr/8 in the horizontal (allow-
ing for impenetrability of the ground and mass continuity). Let
f (u-,e) denote the relative frequency of occurrence of a wind
J
speed class with an average value of u- and a wind direction
J
class with median value e. The concentration per unit emission
rate at a receptor, due to emissions from a point source a
distance x. and azimuth e. from the receptor is:
2
exp (- -^ )
where
(-} xiuic
\ol I J
., - 2KU/-V
a,e = empirical diffusion parameters
-------
o
Over a circular area source of 1 mi the concentration per unit
emission rate per unit area is:
Use linear interpolation between wind direction class medians to
obtain the concentration due to each source and sum,
o
o o . i i IT i i i i
where
4>.. = clockwise most adjacent wind direction class median
. are in radians.
IV. Meteorological Parameters:
Joint frequency distributions of wind speed classes having average
values of u. and wind direction classes having median values of e
J
are required to be known for the period of interest. Over periods
of a month the following values of the meteorologically oriented
diffusion parameters were used for Nashville, Tennessee:
a = 0.6, 6= 1.5, K = 0.06.
V. Validation Results:
An analysis of the regression of observed monthly lead peroxide
candle sulphation rate on concentrations derived by interpolating
isopleths drawn for the gridwork of calculated mean monthly con-
centrations was made for available data from 123 sampling locations
in Nashville, Tennessee, for the months of November 1958 through
March 1959. One-half of the observed concentrations were reported
to be between 80 and 125 percent of the regression relationship.
Less than 5 percent of the observed values deviated from the
regression relationship by more than a factor of two.
-------
Exhibit A-4. Model Implementation by Turner (1964)
I. Scope:
Estimate 24-hour gaseous pollutant concentrations (e.g., S0?) as
the mean of calculated 2-hour steady state concentrations at a
gridwork of receptor locations in an urban area.
II. Treatment of Sources:
Represent emissions from each square mile of a rectangular grid-
work covering the urban area source by a normal line source
centered on the square, oriented perpendicular to the wind, having
a standard deviation (a) of 402m and having an emission rate Q..
Treat all source emissions as having an effective height (h) of
20m. Use available data to estimate emission rates for specific
periods (e.g., 6 A.M. to 8 A.M., Nov. 12, 1958).
III. Computational Equations:
A. For short-term steady state,
Represent the emissions from each normal line source as a
Gaussian plume (allowing for impenetrability of the ground
and mass continuity) using a modification of the Pasquill
diffusion parameters a and a (see meteorological parameters).
Treat S0? as being subject to exponential decay with a half-
Sum concentrations from all sources.
life (t50) of 4 hours.
C = z
0
For long term,
Divide lonq-term period (e.g., 24 hours) into short-term
quasi-steady state periods (e.g., 2 hours) and compute mean
of steady state estimates, e.g.:
1 12
w 1/1 / vi
k=l
-------
IV. Meteorological Parameters:
Determine mean wind speed, wind direction and atmosoheric stability
class for each short-term period. Stability classes are defined
as described in oreceding text (see Tables 2 and 3). Using a
wind speed of 5 m/sec, the Pasquill diffusion parameters (a ' and
0 ) were converted to functions of time and stability class (see
Turner, 1964). Use t. and these realtionships to get (a ). and
(oy-).. Then, (ay). =a+ (oy')r
V. Validation Results:
Calculated concentrations were estimated for sampling locations
by interpolating from an isopleth analysis of initial calculated
values. In 1036 comparisons (available from 32 locations and 35
24-hour periods without precipation, in Nashville, Tennessee)
the following results were obtained for calculated and
observed concentrations rounded to nearest pphm.
Number of zero di fferences 263
Number of 1 pphm differences 602
Average absolute error 2.06 pphm
Root mean square error 3.28 pphm
Skill score 0.13
In 2707 comparisons of 2-hour concentrations (7 sampler locations)
selected from the same period it was noted that 43.7% of the
differences were within 1 ophm.
-------
Exhibit A-5. Model Implementation by Clarke (1964)
I. Scope:
Utilize accepted diffusion coefficients and readily available
meteorological data in a simple model not requiring an electronic
computer to estimate averaqe daily urban pollution concentrations
and diurnal variations.
II. Treatment of Sources:
Estimate short-term emission rates of significant point sources
(e.g., 2-hour rate for power plants; however, Clarke was only able
to obtain a constant emission rate for a single significant
Cincinnati power plant, i.e., 77 tons/day of SO^ and 4 tons/day
of NO ). Using emission inventory data grouped by homogeneous
/\
areas, estimate zero, average and maximum degree day emission
rates for each subdivision of a circular gridwork of model oriented
areas centered on a receptor location. Use daily degree day
observations to interpolate between these values. To define the
diffusion model oriented areas, determine the radial distances
ri > ro» r-3 ancl r „ which for Q. = 1 and u = 1 satisfy:
i i. o max i
r r r r
nf ] C dx = „ / 2 C dx = „ / 3 C dx = / max C dx
0 A rl A r2 A r3 A
For a given stability class these integrals are functions of dis-
tance only and the divisions may be determined graphically. Wind
direction class sector lines and the radial distances define the
area subdivisions. For the area source H = 30m for SOp and 20m
for NO . For point sources the plume rise is given by the Davidson-
/\
Bryant formula (e.g., see Briggs 1969, p.23) and h.. = H. + AH..
III. Computational Equations:
A. For short-term concentrations C, = C + C^
-------
N 2 Q.
CD = I _ - \ N exp <-
p i=l u/Mf) xi(oz).
where N = number of point sources whose azimuth from receptor
to source lies within the prevailing wind direction
class.
4 2 Q.
CA = L. ^ .,. V . exP {~
2(aJ
2
.
IZ-j V 2.
where Q. = emission rate of ith sector segment for sectors
lying within prevailing wind direction class
x. = /-pCr,-2 + r. -,2) = source to receptor alongwind distance
(r, - 0)
B. For long-term concentrations, compute the mean of the N com-
ponent short-term period concentrations, e.g.:
IV. Meteorological Parameters:
For each quasi steady-state period determine the wind direction
class, wind speed and Turner (1964) stability category (e.g., see
Tables 2 and 3 of text).
Use Turner's (1964) graphs to determine a values as functions of
the stability category and travel time t. .
V. Validation Results:
Comparing 19 observed and calculated 24-hour concentrations for
Cincinatti :
-------
CAMP Station Kettering
Laboratory
NOX S02 S02
Number of differences of 2 pphm 14 18
or less
Number of calculations within 18 15
a factor of 2 of observa-
tions
Regression of observed on calculated values (pphm)
Correlation coefficient 0.67 0.71 0.81
Slope 0.66 0.42 0.34
Intercept 3.4 1.1 0.45
-------
Exhibit A-6. Model Implementation by Miller and Holzworth (1967)
I. Scope:
Maximum and average short-term concentrations over a metropolitan
area without the use of an electronic computer.
II. Treatment of Sources:
Represent emissions as a continuous and uniform area source with
infinite crosswind dimension and a ground-level source height.
III. Computational Equations:
Treat area source as a continuum of infinite crosswind line
sources, each with emission rates q dx. Represent the plume from
each line by a vertical Gaussian distribution down to the distance
where 1.25 times the diffusion parameter a equals the mixing layer
ceiling (i.e., travel time t.); beyond this distance use a uniform
vertical distribution. Only emissions with a travel time of 50
seconds or more are considered at any receptor.
A. Maximum concentration at downwind edge of source area:
C = q
| p -I
L dt + . / s dt
,n - — .
50 L
where tg = - = time required to travel across the source area
with the mean wind speed, sec
B. Average of city area:
C = M / L / L 1— dt dt + . / s . / s 7- dt dt
O OU DU /^— t. t. L
IV. Meteorological Parameters:
Meteorological parameters required in the above model are mean
wind speed and mixing layer height. The diffusion parameters are
-------
taken from Turner's (1964) time-oriented interpretation of the
Pasquill diffusion parameters. Class C was used for afternoons
and Class D for mornings. The afternoon mixing layer ceiling was
defined to be the height above the surface at which a dry adiabatic
lapse rate from the maximum surface temperature intersected the
morning radiosonde temperature sounding. The morning mixing layer
ceiling was defined to be the height at which a dry adiabatic
lapse rate from a surface temperature of 5° C greater than the
morning minimum intersects the morning radiosonde temperature
sounding.
-------
V. Validation Results:
Pollutant
NO
• x
so2
NO
X
NO
X
SO
Location
Los Angeles
Nashville
Washington
(a)
Los Angeles
Nashville
Time
Of Day
1500-1700
1200-1400
1300-1700
0700-0900
0400-0600
Observing
Period
1963
Oct - Mar
1958-59
1962-64
1963
Oct - Mar
1958-59
No. of
Periods
36
31
58
35
31
No. of
Stations
7 to 9
7
1
7 to 9
7
Range of
Observations
pphm
4. 5 to 23. 5
0 to 4. 2
1 to 17
5 to 56
0 to 9. 8
Number of
Comparisons
Within
+2 pphm
30 of 36
90% within
+ 1 pphm
(b)
51% within
5 pphm of
regression line
All within
3 pphm of
regression line
Regression of Observed
On Calculated, pphm
Correlation
Coefficient
0.88
0.84
0.83
0.80
0.84
Intercept
0.06
0.25
0.77
5.23
1.77
Slope
1.02
1.16
0.92
0.58
0.41
(a) Calculated concentrations are for downwind edge of city.
-------
Exhibit A-7. Model Implementation by Koogler,
Sholtes, Davis and Harding (1967)
I. Scope:
Estimate continuous or 24-hour mean ground-level concentrations
from multiple sources in an urban area.
II. Treatment of Sources:
4
Sources with an emission rate greater than 10 g/hr are treated as
point sources. All other sources are treated as an area source.
Estimate plume rise (AH) using the Holland formula (i.e., use
stack diameter, gas velocity and gas heat content; see Slade 1968).
The effective source height is:
h = (H + AH)
2_
3
If an inversion base exists aloft, h equals the height of the
inversion base. If an inversion exists at ground level and an
adiabatic lapse rate exists above 1.25 times H, no ground-level
concentrations are calculated. Divide the area source into square
mile subdivisions. Each square is treated as a uniform crosswind
line source 1609m lone.
III. Computational Equations
Divide 24-hour interval into periods of constant meteorological
conditions (i.e., wind direction, wind speed and stability classi-
fication) and duration t. .
A) For point sources such that x < ut.
C = —^ exp
p _.,_ _ r
-------
B) For finite line sources such that x < ut.
Cfl = 9 esp(-0.693
M
.5 - y erf804.5
Contributions due to emissions during one preceding
period are also used as follows:
C) When the wind direction is unchanged from the preceding
period, apply the above equations to sources with an
alongwind distance from source to receptor such that
u?t. <_ x < (u, + u?)t, where u, is the wind speed of the
preceding period and u? is the wind speed of the current
period.
D) When the wind direction differs from the preceding
period, the plume from each source during the preceding
period is treated as a line source. The concentration
distribution along the new line source is the crosswind
integral of the ground-level concentration. The line
source is allowed to diffuse in the new wind direction
allowing only vertical dispersion and decay effects.
The total concentration at a point during a period is
the sum of all point, finite line, and preceding period
emissions. The concentration for a 24-hour period is the
average concentration over all subdivisions of the period.
IV. Meteorological and Other Non-Source Parameters:
For each period of constant meteorological conditions, determine
the wind direction, wind speed and stability classification. The
-------
wind speed is treated as constant to 65m. Above 65m the relation
of wind speed (u,) at height (z,) to the wind speed (iu) at height
(z2) is:
ul = U2
where n is a stability oriented parameter varying from 0.18 for
extremely unstable conditions to 0.51 for stable conditions.
Estimate stability classifications defined by Gifford (1961) from
vertical temperature measurements. Estimate diffusion parameter
values for a and a for each stability classification from
Gifford's presentation (1961). Use 4 hours for the half-life (t5Q)
of S02.
Validation Results:
Calculated and observed 24-hour average SOp concentrations were
compared for 11 sampling stations. Data were obtained for 12 days
during .December 1965 and January 1966 and resulted in 111 compar-
isons. The observed values ranged from 0 to 3 pphm. When values '
are categorized to the nearest 0.5 pphm, a chi square test of
significance on a two-way contingency table (observed versus
predicted) showed the calculated prediction was significant at the
0.1% level. Furthermore, 95% of the computed concentrations were
within ±1 pphm of the observed concentrations.
-------
Exhibit A-8. Model Implementation by Martin and
Tikvart (1968) (also Martin 1971)
I. Scope:
Average seasonal concentrations of pollutants at multiple receptors
from multiple sources.
II. Treatment of Sources :
Treat sources with emission rates greater than 100 tons/year as
elevated point sources. Estimate plume rise from the Holland
equation with a stability oriented correction factor (e.g., see
TRW Systems Group 1969). Treat all other sources as an area
source. Subdivide the area source into squares of 1 to 10 km on
a side and determine the rate of emission and average effective
height h. Each area source square is treated as a virtual point
which emits at a rate equal to the emission rate of the total
square at a distance (r ) upwind of the center of the square
dependent on the size of the area, i.e.:
r -
r "
o " 0.393
III. Computational Equations:
Both the area source and large point sources are represented by a
set of point sources. All wind directions in a 22.5° sector
centered on a 16-point compass azimuth are assumed to occur with
equal probability. The emission rate is assumed to be constant or
at least independent of meteorological conditions. Beyond a dis-
tance of 2 r the vertical distribution of pollutants is uniform
due to the influence of the mixing layer ceiling, where rm is the
travel distance such that a (r ) = 0.47 L. . In terms of the joint
i HI K
frequency distribution of meteorological parameter classes, the
concentration at a receptor point from all sources is given as
follows:
-------
For r .1 r
m
N 6 5
20,
sk)
'> K
exp
For r >_ 2r
N 6 5
1=1 j=l k=l
For r < r < 2r
L,u.
k j 16
r - r.
C (r) • C
-------
Use the Pasquill diffusion parameters for a, expressed as mathe-
K
matical equations of the form a = ax + c where the parameters
a, b and c are defined separately for each stability class and
various ranges of x. The mixing layer ceiling is taken as a
function of the stability class and the climatological afternoon
mixing layer height (L ), varying from 100m for class 5 to 1.5 L
c
for class 1.
V. Validation Results:
Mean winter concentrations were calculated for 40 St.Louis stations.
A regression analysis of observed values on calculated values was
performed for the 40 locations. The regression analysis was
repeated with 5 stations which were strongly influenced by point
sources omitted. The results were:
40 Stations 35 Stations
Correlation coefficient 0.60 0.84
Slope of regression line 0.266 0.56
Intercept of regression line 0.026 0.011
-------
Exhibit A-9. Model Implementation by Fortak (1970)
I. Scope:
Estimate long-term concentrations of ground-level S0? concentra-
tions from multiple sources and derive the associated frequency
distribution of short-term concentrations.
II. Treatment of Sources:
Treat sources with an emission rate greater than 1 kg/hr as point
sources. Estimate the plume rise using Stumke's empirical formula
which is similar to the CONCAWE formula (e.g., see Briggs 1969).
Treat small industrial sources and emissions during space heating
by dwellings as an area source. Estimate the seasonal emission
rate (QA) for each 500m by 500m square of the area source. Use
an emission height of 25m for squares in a downtown area and
15m for squares in a suburban area. A simple, but unspecified,
procedure is used to estimate the effective height of an area
source. Represent each square by a set of n by n indentical point
sources, equally spaced over the square. Each point has an emis-
2 2
si on rate of Q./n . Values of n from 81 to 144 are recommended.
III. Computational Equations:
Consider a fixed retangular horizontal coordinate with 5-axis
pointing east and n-axis pointing north. For wind direction e,
measured from north in radians, the alongwind distance x. and
crosswind distance y. between a point source at location (?., n-)
and a receptor at location UR> nR) are:
x.. = UR - £.) cos (|ir - e) + (nR - n.j) sin (|IT - e)
3 3
yi = ^nR " V COS (2* ' Q) ' (£R ' £-j) sin (jf* ' e)
The concentration at a receptor from all point sources (including
area source representations) during a specific combination of
meteorological conditions is:
-------
c =
where 0o(v, w) = )
3 ^
The concentrations associated with each possible combination of
the classes of the three meteorological parameters which have a
non-zero frequency of occurrence may be calculated and ordered
from low to high. Summing the frequency of occurrence associated
with ordered concentrations a frequency distribution of expected
concentrations at a point may be constructed. From the constructed
frequency distribution the concentration for any probability level
may be estimated. The frequencies of occurrence may also be used
to calculate the long-term mean or the mean for any particular
wind direction class.
IV. Meteorological and Other Parameters:
Determine the joint frequency distribution of meteorological
parameters including wind direction, wind speed and stability
class for the long-term period of interest (e.g., month, seasonal,
heating period). Use 36 wind direction classes, 7 wind speeds
and the 5 stability classes of Turner (see text, Tables 2 and 3)
to define the meteorological conditions. For low-level emissions
(i.e., area sources) use the Pasquill diffusion parameters for
a and a (e.g., see text, Figures 2 and 3). For high-level
emissions (i.e., large point sources) use a slightly modified
version of the Brookhaven diffusion parameters presented by
Fortak (see Stern 1970). Treat the power-law used to extrapolate
observed wind speed class values to values at various stack
-------
heights as a function of atmospheric stability using values
reported in the literature. Use 500m for the height of the mixing
layer ceiling (L).
Validation Results:
Calculated and observed S0? concentrations were compared at 4
locations in Bremmen, Germany, for the 1967-68 heating season (i.e.,
November through May). The cumulative frequency distributions of
calculated and observed concentrations for the season were com-
pared graphically at each location. In general, the observations
were overestimated at three stations and underestimated at one.
Monthly and seasonal means for the four locations were also com-
pared. The observed seasonal means were 0.06 mg/m at one station
3
and 0.08 mg/m at the other three. The calculated minus observed
seasonal concentrations were -0.04, 0.01, 0.02 and 0.04 mg/m .
The observed monthly means (for 4 stations times 7 months) ranged
3
from 0.03 to 0.13 mg/m . The calculated minus observed monthly
3
concentrations for the 28 values ranged from -0.05 to 0.06 mg/m .
-------
Exhibit A-10. Model Implementation by Mil ford,
McCoyd, Aronowitz and Scanlon (1970)
I. Scope:
Use short-term pollutant concentration estimates to give long-term
concentrations.
II. Treatment of Sources:
Treat electric power plants as point sources. Determine the
emission rate and effective height of each plant. Treat other
sources as an area source. Area source characteristics are
available from emission inventory surveys by square subdivisions.
Combine characteristics in the outer regions of the source area to
define characteristics in terms of coarser subdivisions. Represent
each square by treating it as a double virtual point, i.e., as a
two-dimensional plume with assumed vertical and horizontal diffu-
sion parameter values at the center of the square. Determine the
emission rate, initial vertical and horizontal diffusion parameter
value, and mean effective source height for each subdivision.
III. Computational Equations:
Since both the area source and point source are treated as a set
of point sources the standard Gaussian plume equation is applicable
to each calculation. Use a version (specific equation not speci-
fied) which includes wind speed, wind direction, stability class,
mixing ceiling, source location relative to a receptor, source
emission rate and source effective height as parameters. Parameter
values are fixed for a short-term period. W-ith an inversion aloft,
assume no effect due to the inversion up to a certain distance.
At some greater distance assume uniform vertical mixing. Use
linear interpolation at intermediate distances.
-------
IV. Meteorological Parameters :
Make estimates of the wind direction, the wind speed and the mixing
layer ceiling representative of the region for a selected period.
Estimate a stability class aporopriate for use with the McElroy-
Pooler (1968) diffusion parameters. For use with virtual point
sources it is necessary to determine a virtual distance associated
with the assumed initial diffusion parameter. Determine diffusion
parameters for each downwind travel distance from source to recep-
tor by determining the diffusion value corresponding to a distance
equal to the sum of the virtual distance olus the travel distance.
V. Validation Results:
Comparisons were made between calculated and observed S0? concen-
trations at 10 telemetry stations in New York City for the July
through August 1969 period. Calculations were made for each hour
of the period using wind speed and direction data from La Guardia
Airport, an infinite mixing layer ceiling, plume rise estimates
for large point sources based on a 10 mph wind speed using the
CONCAWE formula (e.g., see Briggs 1969), and the McElroy-Pooler
stability class 4 (see text, Table 5) diffusion parameters. The
means of all calculated concentrations for all hours and all
stations were 6.8 and 6.6 pphm for July and August, respectively.
The corresponding observed means were 7.0 and 7.5 pphm. Of the
20 sets of individual station monthly means, 15 calculated means
were within a factor of 2 of observed means.. The largest discre-
pancy was an overprediction by a factor of 4. A summary of statis-
tics reported for comparisons of hourly comparisons is as follows:
Comparisons of all sta-
tion means
Range of comparisons at
individual stations
July 1969
Mean Rel .
Error
-0.36
-3.4 to
0.42
Std. Dev.
of Rel.
Error
2.3
0.66 to
5.1
August 1969
Mean Rel .
Error
-0.58
-4.3 to
0.52
Std. Dev.
of Rel .
Error
2.8
0.69 to
6.0
-------
Exhibit A-ll. Model Implementation by Ludwig, Johnson,
Moon and Mancuso (1970)
I. Scope:
Calculate short-term (single steady state) carbon monoxide (CO)
concentrations in urban areas for producing (1) concentration
isopleths for a specific time, (2) concentration histories for a
specific location and (3) long-term climatological summaries of
concentrations for specific locations.
II. Treatment of Sources:
Treat all CO emissions as a ground-level area source. For each
receptor location considered, estimate the emission rate per unit
area applicable to angular segments centered on a line pointing
upwind from the receptor. The angular segments have a width of
45° out to 1 km. Beyond 1 km the angular segments have a width of
22.5 . The segments are bounded by arcs with radii of 0.125,
0.25, 0.5, 1, 2, 4, 8 and 16 km.
III. Computational Equations:
Use the Gaussian plume diffusion model for treating emissions over
travel distances which are less than the distance (r-r) at which
the Gaussian plume model concentration equals that obtained from
uniform mixing beneath the mixing layer ceiling.
rT = <*7T> 1J where rN < rT < Vl
I J
Beyond this distance uniform vertical mixing is assumed. The
resultant concentration at a receptor from eight sectors is:
-------
u 1
rN+l '
"N-I r ]-bu /'"id"
V n i+1 " i
L «1 a^ll - b. .)
^l i £
1 i ' N n I* . Y* \
i=N+l
/ 1 -bM •
O.sir J
- r ^
rN >
aNj(1 - V
IV. Meteorological Parameters:
Estimate wind direction, wind speed, mixing layer ceiling, and
stability class. Use wind direction and speed observations from
airport weather stations. Estimate mixing layer ceilings using
the nearest 1200Z radiosonde data and the maximum afternoon tempera
ture (T ) by the following values:
Time of Day L
Midnight to 1200 GMT
1200 GMT to 1400 1ST
m
- T
Ll "
m
'Ta - Tm,
. 29.3
,\j rn UN i yn u up
r A -T
Oorjol m s
. £11 Ol •
Pm + PS /Tm - Ts^
I 2 \Pm-Psy
-j - 0.0633
/T + T \
\ n °°7 m s
°-t87\ 2 /
m
where T = surface temperature in 1200Z radiosonde
p = surface .pressure in 1200Z radiosonde
T = temperature of first significant level in 1200Z radiosonde
-------
p = pressure of first significant level in 1200Z radiosonde
<)> = population of the urban area
T = maximum afternoon surface temperature
a
p = surface pressure corresponding to T
a a
p = pressure in the 1200Z radiosonde corresponding to a
/\
potential temperature given by Tg and pg
T = temperature in the 1200Z radiosonde corresponding to p
X X
H = hour of interest
L = L for 1200Z radiosonde of following day
Determine which of five Pasquill stability classes is appropriate
from the solar elevation angle («), the fraction of sky covered
by clouds (N), and the wind speed using the following table:
Wind Speed (Knots
I3
3-6
6-10
10-12
I13
0
-------
For the source segment closest to the receptor:
bu = °
a.. = a for class j at x = 125m
»J ^
For other parameter values, see author's text (Ludwig, et al. 1970)
V. Validation Results:
Comparisons were made between calculated and observed hourly con-
centrations of CO during weekdays of March to December 1964 for
the St. Louis CAMP station. The correlation coefficients ranged
from 0.16 to 0.45 with a mean of 0.31. The RMSE ranged from 4.7
to 8.8 ppm with a mean of 6.2 ppm. A good portion of the RMSE is
attributable to a background level which was estimated for each
month and found to vary between 3.5 and 7.0 ppm.
-------
Exhibit A-12. Model Imolementation by Calder (1970)
I. Scope:
Estimate lonq-term average concentrations of gaseous pollutants
using point and area source emission inventory data together with
climatological frequency data for wind speed, wind direction,
atmospheric stability and mixing depth.
II. Treatment of Sources:
Treat large sources with emission rates in excess of 100 tons/yr
as elevated point sources. Determine the seasonal emission rate,
physical stack height, and additional stack parameters required to
estimate plume rise (e.g., see Briggs 1969). Treat other emissions
as an area source. Determine seasonal emission rates for sub-areas
of the area source which are chosen in relation to the spatial
uniformity of the source distribution. Determine the uniform
effective height which is most applicable for the area source
(e.g., 20m was selected for St. Louis).
III. Computational Equations:
The total averaqe concentration at a receptor due to both point
and area sources is (T = C. + C where:
C = P. V V V Q1 f (ei' J> k) S (Xi' ^ V k)
y\ •
N
1=1 j k
where
j = wind speed class
k = Pasquill stability class as defined by Turner (see
Tables 2 and 3)
zr = height of receptor location
-------
The form of the factor S (x., z ; u , k) is dependent on the degree
' ' J
to which the mixinq layer ceiling (Lk) restrains vertical mixing
as follows:
L.
Foraz(x.; k) £ ^ ,
S(x., z: p., k) =
7r Uj oz(x.; k)
exp<-
21,
k
For 0 (x.; k) >_ -5--^ ,
£- \ C* • I O
exo<-
2o22(xf; k)
+ H. +AH.)2V
22(Xl;k) /.
S(xr zr; y k) =
l k
For 0 !;.- < a (x.; k) < 0 !;.- , use linear interpolation between the
d. I b z 1 c.. I b
two forms.
16 ,<
2TT 6
/16
fc
M=l
:*>!!
j k
S(x
, zr; u., k)>dx
where q (x) = / Q(x, e) de (0
ei
17
Apply the trapezoid rule to numerically integrate the equations.
The e. values are 22.5° apart. The q.(x) integral may be evaluated
using 2.25° increments for e. The C. integral may be evaluated
using 100m increments from the receptor to the edge of the source
area.
-------
IV. Meteorological and Diffusion Parameters:
The meteorological parameters in the above model are wind speed,
wind direction, stability classification, and climatological
mixing layer ceiling (e.g., as tabulated by Holzworth, 1964).
Generate the joint frequency distribution of wind speed, wind
direction, and stability classification using standard Weather
Bureau airport observations. On this basis, there are 16 wind
direction categories with dividing azimuths 6^, 6 wind speed
categories (each with a representative value u-)» and 5 stability
J
categories corresponding to the A to E Pasquill diffusion parameter
classes, but determined using Turner's criteria (see text, Figures
2 and 3). The mixing layer ceiling L is determined from the
seasonal average daily maximum mixing death (L in Meters) and the
stability classification as follows:
Stability Classification Lk (meters)
A 1.5 L
B, C, D (day) , L
D (night) 0.5 (100 + L)
E 100
The Pasquill diffusion parameters o (x; k) are determined from the
following relationships fitted to curves presented by Gifford (1961)
az(x; k) = ak. x ki + cki
For each value of k there are three possible sets of coefficients
(a^, b, . c, .). The proper set depends on which of 3 distance
ranges contains x (i.e., <100, 100-100C, >1000; see Calder (1970)
for the 45 coefficient values).
V. Validation Results:
Calculated and observed seasonal mean S0? concentrations were com-
pared for St. Louis winter season data (Dec. 1, 1964 to Feb. 28,
-------
1965). A regression analysis of observed on calculated values
o
(pg/m ) for 40 sampling stations resulted in a correlation coeffi-
cient of 0.775 an intercept of 19.98 and a slope of 0.26.
-------
Exhibit A-13. Model Implementation by Singer
and Smith (1970, 1966)
I. Scope:
Estimate short-term concentrations from multiple single sources
for stable pollutants.
II. Treatment of Sources :
Identify the pollutant emission rate, geographical location, stack
height and stack and emission characteristics related to plume
rise for each source. Treat each emission as a continuous point
source.
III. Computational Equations:
The concentration at each receptor point is the sum of contribu-
tions from each of N sources.
Qi / (Hi +AHi)2
c =
IV. Meteorological and Diffusion Parameters:
The wind speed is assumed to be horizontally uniform but varies
with height as a power function in which the power (q) is given
by the atmospheric stability. In order to account for the effect
of the wind profile on the growing vertical dimension of the plume
from a point source, a "mean equivalent wind speed" is defined.
The height in the wind profile at which the "mean equivalent wind
speed" occurs was found to be z" = 0.62a .
For Hi + AHi £ 0.62(az).
u = S(Hi + AH..)q
For H. + AH. > 0.62(a )
1 "• . z i
u = S|0.62(az)
-------
a
6 may be estimated from a wind speed u at height z . 6 = — -
a z
a
q varies from 0.15 for unstable conditions to 0.50 for stable
conditions.
The diffusion parameters are given in terms of distance x. and
Brookhaven gustiness classes (Singer and Smith 1966).
(oz). - bk x,
There is a set of ak, b. and p. values specified for each of the
4 Brookhaven gustiness classes as follows:
Brookhaven Gustiness Class
B2
Bl
C
D
ak
0.40
0.36
0.32
0.31
bk
0.41
0.33
0.22
0.06
Pk
0.91
0.86
0.78
0.71
(o ) is further limited by the mixing layer ceiling restriction
that (az) ±Y25 '
V. Validation Results:
Calculated and observed 6-hour mean SO,, concentrations were com-
pared. The data was selected from 20 sources, 5 S0? sampling
locations, Weather Bureau airport observations, supplementing wind
data and aircraft observations. The mean observed and calculated
concentrations for all 6-hour periods at each location were:
-------
Station
Ferry
Whitney
Count
Tempi e
Fairhaven
Mean
Observed
.049
.029
.035
.075
.048
Mean
Calculated
.043
.014
.008
.024
.033
Number of
Comparisons
31
47
31
8
8
-------
Appendix B
-------
Section 1.0
INTRODUCTION
This appendix summarizes the St. Louis data which have been
obtained and identifies their sources.
Punched cards were received from the government (Division of
Meteorology, National Environmental Research Center, Research Triangle
Park, EPA) which contain emission information, meteorological observa-
tions, air quality observations, and algorithms for estimating S02
emission rates for St. Louis. The data covered the period
1400 December 1, 1964 to 1400 February 28, 1965. The characteristics
and sources of these data are briefly reviewed below.
-------
Section 2.0
EMISSION DATA
Turner (1968b) has developed an algorithm for estimating S02
emissions as a function of location, temperature, time of day and day of
week. This algorithm represents emissions in terms of two types of sources,
point sources and area sources. The computation formulas are presented
below for each type of source both in terms of the input provided by the
punch cards and in terms of the original information which is the basis
for the punched card data. Much of the original information is not
available; however, formulations in terms of original data help to
indicate the nature of input errors and uncertainties associated with
the data and the derived emission estimates.
2.1 POINT SOURCES
Information regarding utility power plants and industrial
plants are available in different levels of detail. As a result, each
is discussed separately.
2.1.1 Industrial Sources
Industrial SOp emissions are the sum of emissions from process
and space heating fuel consumption. The emissions from process fuel
requirements are estimated by multiplying a peak process emission rate by
utilization factors related to the day of the week and the hour of the
day. The emissions from space heating requirements are determined in
terms of the outside air temperature deficit from 65°F. The algorithm
-------
used for St. Louis is:
Q(t) = Qp Fd(t) Fh(t) + qx Dc(t) (1)
where
Q(t) = SCL emission rate
Q = peak process SOg emission rate
F.(t) = fraction of peak for day of week
Fh(t) = fraction of peak for hour of day
q = heating fuel S0? emission rate per degree (i.e., per
J\ £
degree of ambient air temperature below 65°F)
Dc(t) = 65 - [T(t) + Ac(t)]
T(t) = ambient air temperature, °F
Ar(t)= commerical correction factor, °F (Turner 1968a).
c
The quantities Q , F.(T) and F. (t) were obtained by Turner from
survey questionnaires submitted by plant operators. The quantity qv may
A
be obtained from annual emission information by the following formulae:
<
-------
q = S09 emission rate per degree
A C-
(F ). = fraction of annual quantity of fuel j used in winter
J season
D = winter season degree days.
W
In terms of initial source data:
Q(t) = Qp Fd(t) Fh(t) + [65 - T(t) - Ac(t)] \- \ Wj Sj (4)
In addition to the emission rate parameters, i.e., Q , F^,
Fh, q and A , the furnished punched card data included location
11 A w
coordinates of the source, the physical height of the source stack
(effective height if no plume rise data is given) and the plume-rise,
wind-speed product. The plume-rise, wind-speed product was estimated
by means of Holland's (1953) formula using data obtained in the course
of an inventory survey. Only the resultant product data were available
for this study.
2.1.2 Power Plant Sources
Power plant SOp emissions were estimated by two different
methods, each developed to fit a certain type of available data. For
four plants, stack operating characteristics were obtained as a function
of plant output. For these plants hourly outputs in megawatts were
available for the entire data period. For one plant this information
was broken down by generating unit. The information included graphs
of fuel weight flow rate, stack temperature and stack exit gas volume
flow rate as functions of power output. The fuel flow rates were
-------
found to be appropriately represented by a linear relationship of the
form
F = A1 + A2 L (5)
where F = fuel weight flow rate, Ib/min
L = power output, megawatts
A,,A2 = empirical parameters.
The stack temperature and volume flow rate were also estimated by a
linear relationship with power output. However, it was found necessary
to divide the range of power outputs from zero to a peak value into
3 equal parts and to use a linear approximation over each part or class
of the range. Two values of stack temperature and of stack gas flow
rate were selected to define the end points of a linear approximation
for each class of power output values. However, a single end point
was selected for adjacent classes. As a result, four values of stack
temperature and flow rate and the peak power output allow one to
construct a linear approximation of these parameters as a function of
power output. Specifically,
Ts • T» + T^TT
-------
T = stack temperature for power load of L , °F
Jt X»
T = stack temperature for power load of L + •=• L
u (upper limit of class), °F. £ J p
Similarly, the stack volume flow rate for power output L lying in the
power output class with lower limit L is
J6
vs • \ + TJ/T 'vu - V
where V = stack gas volume flow rate, 10 ft /min
fi *?
V = stack gas volume flow rate for power load of L , 10 ft /min
J6 J6
V = stack gas volume flow rate for power load of L + •*• L ,
u c o a o p
10° fr/min.
Values of the class limits and the peak power load for seven emission
points in the St. Louis area are given in Table B-l . Values of the
empirical parameters for estimating fuel weight flow rate are also shown.
The fuel flow rate may be converted to an S0? emission rate by means of
a proper SOp emission factor and the sulfur content of the fuel as
follows:
where Q = S02 emission rate, Ib/min
F = fuel weight flow rate, Ib/min
E = S02 emission factor, Ib S02/lb S (~ 2)
S = sulfur content of fuel, percent.
-------
Table B-l. Stack Emission Parameters for St. Louis Power Plants (1964-65)
Plant
Meramec
Meramec
Meramec
Venice
Cahokia
Ashley
Unit
Number
1 and 2
3
4
Coefficient of
Fuel Flow Rate
(F, Ib/min)
F=A!+A2L
L=Power output, mw
Al
50
100
100
20
0
0
A2
10.89
10.43
10.75
13.0
16.67
1.536
Peak
Fuel Flow
Rate,
Ib/min
L
P
150
300
405
480
150
1500
Stack Temperature
Class Limits, F
Tl
281
242
236
252
350
296
T2
294
262
267
293
406
323
T3
312
290
323
312
420
342
T4
343
336
372
352
434
356
Stack Flow Rate
Class Limits,
106 ft3/min
Vl
40
210
260
30
20
5
V2
170
375
490
625
260
225
V3
320
625
840
1180
520
440
V4
500
1015
1460
1840
835
665
-------
For another power plant site, an average emission rate was
estimated for each two-hour period of the day for each of four emission
sites. A corresponding estimate of the wind speed, plume rise product
was also made for each site. These estimates were assumed to be repre-
sentative of all days in the data period. The estimates were derived
by government personnel based on discussions with the plant operator.
The estimates were part of the set of punched card data furnished by
the government.
In addition to the emission data described above, the govern-
ment furnished punched card data included location coordinates and
physical stack heights.
2.2 AREA SOURCES
Area source emissions for SOp were available for St. Louis in
5000 ft by 5000 ft grid squares. Emission data for several categories
of sources (including residential, commercial, river vessels, auto-
mobiles, railroads, backyard burning and industrial) were available on
punched cards for each grid square. The information amounted to a
parameter estimate for use in the following algorithm developed by
Turner (1968b):
Q(t) = Qr + qr Dr(t) + Qc Fc(t) + qc Dc(t) + QV + Q
a
Dr(t) = 65 - T(t) - Ar(t)
Dc(t) = 65 - T(t) - Ac(t)
-------
where Q(t) = SCL emission rate
Q = base residential S02 emission rate
q = residential heating SCL emission rate per degree
T(t) = ambient air temperature
Ar(t) = residential correction factor (Turner 1968a)
A (t) = commercial correction factor (Turner 1968a)
Q = base commercial S02 emission rate
FJt) = commercial diurnal variation factor
c
qc = commercial heating SCL emission rate per degree
Q = river vessel SCL emission rate
CL = base automotive S00 emission rate
a c.
FJt) = automotive diurnal variation factor
d
Q., = railroad S00 emission rate
W c.
Q. = backyard burning S02 emission rate
qx = industrial heating SCL emission rate per degree
Q = base industrial process emission rate
= industrial day of week variation factor
= industrial diurnal variation factor
The parameters and variables in the above algorithm reflect
a combination of basic data and assumptions. As a result, an attempt
has been made to divide the above parameters into more fundamental com-
ponents so that assumed and reported or observed values can be more
easily identified. In the notation below subscripts i and j are intro-
duced on parameters which vary from one area source to another. Subscript
-------
k denotes fuel types such as gas, oil and coal. Subscript a is used to
designate an annual or national average.
J DaRa k-1 EkHk
where H = average annual U.S. household space heating energy
requirement
D= = average annual U.S. degree days
a
Ra = average number of rooms per U.S. household
a
R. . = average number of rooms per dwelling unit in grid
J square (i ,j)
$k = sulfur dioxide emitted per unit of fuel k
E. = heating efficiency of fuel k
H. = heat content of fuel k
(N.).. = number of residential dwelling units using fuel k in
J grid square (i ,j)
K = number of fuels.
(n ).. = s / j p-
wr'ij 1 - FS VHr'ij w
where F = summer day fuel consumption as fraction of average
winter day
F" = average winter degree day.
W
where At = duration of season
C. . = number of commercial sources in grid square (i,j)
' J
W^ = annual quantity of fuel k used by £th source
(F ). = fraction of annual quantity of fuel k used by £th
source used in summer season
Sk = sulfur dioxide emitted per unit of fuel k.
B-10
-------
where D,, = winter season degree days
W
(F ). = fraction of annual quantity of fuel k used by £th source
in winter season.
K !ii
I i\ I j
(a ) = -— T. S £ (F ). W. M4}
»Ty'ii n \f * w'ko 1
-------
2.3 SUMMARY OF ST. LOUIS EMISSION PARAMETERS
Table B-2 lists the source parameters for estimating SOg emission
rates for St. Louis point sources, the basis of the parameter estimate and
the source of the information. Table B-3 shows the derived average
diurnal variations in power plant SOp emission rates and plume rise factors
for one plant with four emission sites. Table B-4 shows the diurnal
variation in the commerical and residential temperature corrections for
space heating requirements. Table B-5 lists the area source parameters
available for estimating SC^ emission rates for St. Louis area sources.
As in Table B-2, the basis for the estimate and the source of information
are also included. Table B-6 shows the diurnal variation in the com-
merical base emission factor and in the automotive emission factor.
Miscellaneous additional emission parameters are listed in Tables B-7,
B-8, and B-9.
-------
Table B-2. Point Source Emission Data
Source Parameter
QP
*d(*)
Fh(t)
W.
SJ
Da
(Fw).
Dw
Stack height
Effective stack
height
Plume rise-wind
speed product
Q(t) for special
power plant site
AcW
T(t)
L.Lg.Lp
A1'A2
Te>Tu
VC'Vu
Basis of Estimate
Annual SC>2 emissions
Usual days worked per week
Usual shifts worked per day
Annual fuel consumption
Chemical analyses
Annual degree days
Percent Wj used in winter
Winter season degree days
Stack height
Stack height plus judgment
Stack diameter, exit gas temp.
and flow rate
Fuel use, output loads
Temperature correlation with
steam heat loads
Airport hourly temp. obs.
Plant output records
Plant operating characteristics
Stack operating characteristics
Stack operating characteristics
Source of
Information
Questionnaire
Questionnaire
Questionnaire
Questionnaire
NOAA
Questionnaire
NOAA
Questionnaire
Questionnaire
Plant oper.
Turner (1968a)
NOAA
St. Louis
Union Elec. Co.
St. Louis
Union Elec. Co.
St. Louis
Union Elec. Co.
St. Louis
Union Elec. Co.
Parameter Values
Stored on punched cards
Stored (Saturday and Sunday listed
as fraction of week day)
Stored (midnight and swing shifts
listed as fraction of day shift)
Contained in q (stored on punched
cards)
See Table B-8
Contained in qx (stored on punched
cards)
Contained in qx (stored on punched
cards)
Contained in q (stored on punched
cards)
Stored on punched cards
Stored on punched cards
Stored on punched cards ( also see
Table B-3)
See Table B-3
See Table B-4
Stored on punched cards
Stored on punched cards
(L , see Table B-l)
See Table B-l
See Table B-l
See Table B-l
-------
Table B-3. Diurnal SO, Emission Rates and Wind Speed- Plume Rise
Products for Four Emission Sites of a St. Louis Power Plant
Two
Period
Ending
0200
0400
0600
0800
1000
1200
1400
1600
1800
2000
2200
2400
Stack A
Emission
Rate,
g/sec
106
106
106
106
529
423
353
212
176
318
176
106
Wind Speed
Plume Rise
Product,
m /sec
40
40
40
40
201
161
134
80
67
120
67
40
Stack B
Emission
Rate,
g/sec
106
106
106
106
529
529
529
529
529
529
529
106
Wind Speed
Plume Rise
Product,
2
m /sec
40
40
40
40
201
201
201
201
201
201
201
40
Stack C
Emission
Rate,
g/sec
282
282
282
600
706
706
706
706
706
706
706
282
Wind Speed
Plume Rise
Product,
m2/sec
102
102
102
217
256
256
256
256
256
256
256
102
Stack D
Emission
Rate,
g/sec
1341
1552
2364
2681
2681
2681
2681
2681
2681
2681
2681
1799
Wind Speed
Plume Rise
Product.
2
m /sec
504
583
888
1007
1007
1007
1007
1007
1007
1007
1007
676
-------
Table B-4. Diurnal Residential and Commercial Temperature Corrections (Turner, 1968a)
Hour
Ending
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
2100
2200
2300
2400
AC<'>
Commercial Temperature Corrections, F
Weekday
13.32
13.23
12.54
10.43
5.64
-1.75
-8.04
-11.69
-13.91
-12.94
-12.43
-12.53
-12.39
-11.19
-9.62
-7.88
-4.37
0.56
4.55
6.62
9.08
10.41
11.53
13.19
Saturday
17.03
18.19
16.30
14.55
10.17
3.19
-0. 13
-4.13
-7.14
-6.74
-7.00
-6.98
-7.11
-7.01
-3.84
-1.84
-0.66
2.60
5.97
7.92
8.90
11.47
13.48
15.86
Sunday
16.87
17.82
18.43
16.90
15.15
12.86
9.47
8.63
6.01
4.82
2.64
1.38
0.30
0.55
2.98
4.49
6.25
8.70
9.92
10.10
10.47
12.01
12.23
12.03
*r(t>
Residential Temperature Corrections, F
Weekday
11.11
10.61
9.69
8.54
7.08
3.13
-2.15
-7.32
-7.61
-8.85
-8.44
-7.46
-6.73
-6.25
-5.11
-4.08
-3.17
-2.41
-0.77
-0.01
2.56
3.22
5.33
9.11
Saturday
10.08
11.97
9.69
8.43
6.65
4.24
1.85
-0.73
-6.30
-8.01
-7.26
-9.34
-8.28
-8.07
-7.78
-6.14
-5.28
-3.82
-1.73
-0.86
2.31
3.85
5.71
8.74
Sunday
8.11
9.07
9.12
8.15
6.64
4.76
1.83
0.15
-5.60
-7.61
-8.72
-7.84
-5.55
-5.87
-4.09
-2.86
-1.50
-1.43
-0.41
-0.61
-1.49
-0.60
1.23
4.78
-------
Table B-5. Area Source Emission Parameters
Source Parameter
F
s
H
a
. Da
Na
i3
kl
(Fw)kl
(Qv)ij
(Qa)..
Fa(t)
(9w)ij
(Qb)ij
[Fd(t)]j.
[FhWhj
(QP)e
Dw
T(t)
Ar(t)
Basis of Estimate
Fuel company information
National fuel data
National meteorological data
Census information
1960 Census (St. Louis SMS A1 s)
1960 Census (St. Louis SMSA's)
Assumed fuel heat content
Assumed combustion efficiency
Average fuel sulfur content
Emission survey data
Reported annual fuel use
Reported % of W, , used in
summer
Reported % of Wkl used in
winter
Not reported
Not reported
Not reported
Not reported
Not reported
Usual industry working days
Usual industry working shifts
Reported annual SC>2 emissions
1934 to 1964 St. Louis temp, data
Hourly airport temperature
Temperature correlations with
gas sendouts
Source of
Information
La Clede Gas Co.
Landsberg et al,
1963
Landsberg et al,
1963
Stat. Abstracts
1962
U. S. Census
Bureau
U. S. Census
Bureau
Turner, 1968b
Turner, 1968b
Turner, 1968b
Not reported
Questionnaire
Questionnaire
Questionnaire
Not reported
Not reported
Not reported
Not reported
Not reported
Questionnaire
Questionnaire
Questionnaire
NOAA
NOAA
Turner, 1968 a
Parameter Values
0.2
70 x 106 Btu/yr
4600 °F days/yr
4. 9 rooms/household
Contained in Q and q
Contained in Q and q
See Table B-7
See Table B-7
See Table B-8
See Table B-6
Contained in Qc, q^, and qx
(stored on 1200 punched cards
Contained in Qc and q^ (stored
on 1200 punched cards)
Contained in q and q (stored
on 1200 punched cards)
Stored on 1200 punched cards
Stored on 1200 punched cards
See Table B-6
Stored on 1200 punched cards
Stored on 1200 punched cards
Stored on 1200 punched cards
Stored on 1200 punched cards
Contained in Q (stored on
1200 punched cards)
31.6°F
Stored on punched cards
See Table B-3
(Continued)
-------
Table B-5. Area Source Emission Parameters (Concluded)
Source Parameter
Ac(t)
Dw
Source height
Basis of Estimate
Temperature correlations with
steam heat loads
Winter season airport tempera-
ture
Estimate guided by building
heights
Source of
Information
Turner, 1968a
NOAA
Visual reports
Parameter Values
See Table B-3
Contained in q and q (stored
on 1200 punched cards)
Stored on 1200 punched cards
-------
Table B-6. Base Commercial and Automotive Emission Factors for St. Louis Area Sources
Two Hour
Period
Ending
0200
0400
0600
0800
1000
1200
1400
1600
1800
2000
2200
2400
Fc(t)
Fraction of Mean Commercial
Base Emission Rate
0.20
0.20
0.20
1.96
1.82
1.75
1.69
1.62
1.48
0.68
0.20
0.20
Fa(t)
Fraction of Mean Automotive
Emission Rate
0.176
0.037
0.111
1.575
2.261
1.001
0.788
1.325
2.632
1.084
0.573
0.417
-------
Table B-7. Miscellaneous Emission Parameters
Area Source Emission Data
1. Average energy use for space heating per household (USA wide) =
70 x 106 Btu/yr
Ref: Resources in America's Future; H. H. Landberf, L.L. Fischran,
J. T. Fisher; Johns Hopkins Press, 1963 (Baltimore, Md.).
2. Average degree days for the country = 4600 °F day/yr
Ref: ditto (1)
3. Average size of household =4.9 rooms/dwelling unit
Ref: Statistical Abstracts of U.S.; 1962; Department of Commerce,
Bureau of Census, Washington, D.C.
4. Fossil fuel characteristics:
Fuel Heat Content (H,J Combustion Efficiency (E,J
" " " " " ' K •~1--| -L -..».- ** - |l^
Coal 26 x 106 Btu/ton 0.5
Oil 145,000 Btu/gal 0.6
Gas 1,000 Btu/ft3 0.75
5. Summer fuel use = 20% of winter day
6. Average St. Louis winter degree day = 31.6 °F day
Ref: Dec, Jan, Feb 1934 to 1964 Weather Bureau records.
7. Information from U.S. Census of Housing (estimated from SMSA
data) obtained for source area:
(a) Number of dwelling units
(b) Rooms per dwelling unit
(c) Number of coal burning units
(d) Number of gas burning units
(e) Number of oil burning units
(f) Population.
8. Survey questionnaire information requested in St. Louis Intra-
state State Study (Turner 1968b):
(a) Source Location
(b) Type of fuel and annual use required
(c) Percent used by season
(d) Stack height
(e) Process emissions in tons per year.
-------
Table B-8. St. Louis SO- Emission Parameters
Fuel
Coal
Oil
Gas
Fuel Requirements
For average
U. S. household
(4-9 rooms/
household) ,
units per house-
hold per
degree day
0. 0012 tons
0. 18 gal
22. 5 ft3
For average
U. S. room,
units per room
per degree day
2.45xlO"4tons
3. 67X10" 2 gal
4. 59 ft3
SO2 Emission
SO2 emission
lbSO2
per unit
(Sk)
38p per ton
157p per
103 gal
2. 86s per
102 ft3
SO2 heating
emission rate,
grams per
second per room
per degree ^a'
4. 88pxlO~5
3.025pxl0~5
6. 89s x ID' 8
St. Louis Values
Sulfur
content
3.3%
1.5%
0.3 grains
per 102 ft3
SO2 heating
emission rate,
grams per
second per room
per degree *a'
1.61X10"4
4. 54X10" 5
2. 08X10" 8
p = % sulfur content by weight
s = grains sulfur content per 100 ft
. o
(a) degree = number degrees average outside temperature is below 65 F
Table B-9. Average Sulfur Content of Coal Used by St. Louis Power Plants
Plant
Meramec
Venice
Cahokia
Ashley
Sulfur Content, %
2.86
2.52
1.96
3.39
-------
Section 3.0
METEOROLOGICAL DATA
The following meteorological data were obtained for the St. Louis
area:
1. Hourly Lambert Field aviation weather observations
2. Hourly Scott Field aviation weather observations
3. Lindbergh High School wind, temperature and relative
humidity hourly averages from strip charts
4. State Police Station C hourly averages of wind,
temperature and relative humidity
5. Hazelwood High School hourly averages of wind, temperature
and relative humidity
6. TV tower hourly averages of temperature and wind at
3 heights and a bivane measured standard deviation in
wind direction.
The parameters and units which are available on punch cards are
listed in Tables B-10 and B-ll.
-------
Table B-10. Summary of St. Louis Hourly Airport Weather Observations
(Available for Lambert Field and Scott Field)
Parameter
Units
Ceiling
Sky Cover
Visibility
Weather Elements
Temperature
Dew Point
Wind Direction
Wind Speed
Peak Gust
Altimeter Setting
Precipitation
(Lambert Field Only)
Hundreds of Feet or Code
Code
Tenths of Mile
Code
Tens of Degrees
Knots
Knots
Hundredths of Inches of Mercury
Hundredths of Inches
-------
Table 11. Hourly Average Weather Observations
Observation Point
Lindbergh HS
Lindbergh HS
Lindbergh HS
Lindbergh HS
State Patrol Station C
State Patrol Station C
State Patrol Station C
State Patrol Station C
Hazelwood HS
Hazelwood HS
Hazelwood HS
Hazelwood HS
TV Tower, 127 ft
TV Tower, 127 ft
TV Tower, 255 ft
TV Tower, 255 ft
TV Tower, 459 ft
TV Tower, 459 ft
TV Tower 124 ft
TV Tower
TV Tower
TV Tower
TV Tower
Parameter
Wind speed
Wind direction
Temperature
Relative humidity
Wind speed
Wind direction
Temperature
Relative humidity
Wind speed
Wind direction
Temperature
Relative humidity
Wind speed
Wind direction
Wind speed
Wind direction
Wind speed
Wind direction
Temperature
124 ft to 249 ft
temperature gradient
124 ft to 452 ft
temperature gradient
249 ft to 452 ft
temperature gradient
Bivane standard
deviation
Units
Miles per hour
Tens of degrees
°F
%
Miles per hour
Tens of degrees
°F
%
Miles per hour
Tens of degrees
°F
%
Miles per hour
Tens of degrees
Miles per hour
Tens of degrees
Miles per hour
Tens of degrees
°F
°F
°F
°F
degrees
-------
Section 4.0
AIR QUALITY DATA
Two types of SCL measurements are available for St. Louis.
Ten stations with average two-hour observations are available, and
40 stations with average 24-hour observations are available. The
sampling period for 24-hour observations began and ended at 2 p.m.
daily. This time of day permits bubbler collectors to be switched at
a time when ambient concentrations are expected to be a minimum. Both
2-hour and 24-hour samplers used a bubbler collection assembly to
measure SOp concentrations. The basic National Air Sampling Network (NASN)
gas sampling bubbler, with slight modifications, was used. Each bubbler
consisted of a polypropylene centrifuge tube (4 inches long by 1 inch
diameter) fitted with a two-hole rubber stopper. A glass tube extended
to approximately 5/16-inch from the bottom of the bubbler tube. The
airflow was regulated with a standard Gel man orifice assembly. Flow
rate determinations were made weekly with a calibrated flow meter. In
the two-hour samplers approximately 120 liters of air was bubbled through
the sampler each two hours. An external vacuum pump draws air through
the bubbler where sulfur dioxide is stripped from the air stream by the
complexing action of sodium tetrachloro-mercurate absorbing reagent. The
collected samples were transported to a laboratory where they were
analyzed for S02 concentrations. Twelve two-hour samples were picked
up for analysis each day and analyzed in a specially set up local
laboratory. The analytical method used was that of West and Gaeke as
modified by Welch and Terry. Duplicate sets of 24-hour samples were
-------
available at the same locations as the two-hour samples. A comparison
of the 24-hour concentrations measured by two-hour samplers and by the
24-hour samplers revealed that two-hour samplers averaged about 15 per-
cent higher than the 24-hour samplers.
-------
Section 5.0
ST. LOUIS PREPROCESSING PROGRAM LISTINGS
C ST LOUIS DATA
DIMENSION TFR(24,3), TFC I ( 24 , 3 ) , DFC ( 24 ),DFA(24),SULFR(7),PLOAD(8)
DIMENSION HA(1200),Z(3),TFR2(24),TFCI2(24),HA2(380)
DIMt.MSION XR<50) ,YR(50),ZR(50),UBS02(50) ,XP(51),YP(5 1 )
DIMENSION ZP(51),QP(51),CAS02(40),QB(1200),QRH(1200),OCB{1200 ) ,
1QCIH< 1200),QAU(1200),QSA(1200),SATA( 1200) ,SUNA(1200),SMI DAI 1200),
2 SWIGA(1200) ,QIJ(3600)
EQUIVALENCE (TFR2(1) ,TFR(1,3)),(TFCI 2(1),TFCI(1,3 ) ),(HA2(1),HA(821
1) )
DATA DLAST/22814./
DATA NH/3/
DATA Z/20.,30.,45./
DATA DLTA/1524./
DATA NRECP/50/
DATA NR1/40/
DATA CAS02/40*0./
DATA GX/30./
DATA GY/40./
DATA NIPS/44/
DATA NUS/7/
DATA NRPNT/51/
DATA XP / 3.94, 9.32,14.04,14.88,14.60,19.20,18.76,20.26,19.92,
1 20.84,22.00,24.88,24.38,26.52,31.00,16.52,16.96,13.84,
2 16.08,15.46,18.00,18.14,18.38,19.14,20.88,20.86,21.14,
3 20.96,21.90,21.82,21.88,21.62,22.72,24.46,23.54,24.62,
4 26.14,29.86,18.16,18.68,22.34,22.36,22.38,22.40,10.84,
5 10.80,10.76,10.74,19.88,19.41,19.67/
DATA YP /13.20,28.80,26.88,12.34,13.50,18.20,22.66,17.06,16.52,
1 38.14,19.56,34.76,34.70,34.66,30.20,21.56,15.84, 8.66,
2 18.90,23.00,17.00,17.00,21.58,21.84,16.88,24.30,25.00,
3 24.80,17.64,24.14,37.66,37.98,19.40,20.82,38.30,35.84,
4 34.72,11.08,18.34,19.08,36.52,36.52,36.52,36.52, 2.84,
5 2.85, 2.86, 2.87,21.78,17.50,19.87/
DATA HA/343*20.,30.,29*20.,2*30.,29*20.,30.,22*20.,3*30.,3*20.,3*
1 30.,20*20.,4*30.,2*20.,5*30.,21*20.,10*30.,21*20.,10*30.,20.,3*
2 30.,16*20.,8*30.,2*40.,20.,4*30.,16*20.,2*30.,2*20.,3*30.,40.,50.
3,3*20.,3*30.,15*20.,7*30.,2*40.,20.,30.,20.,3*30.,15*20.,9*30.,21*
4 20.,8*30.,2*20.,30.,19*20.,30.,2*20.,3*30.,23*20.,2*30.,3*20.,2*
5 30.,4*20.,3*30.,16*20.,30.,4*20.,30.,20.,30.,23*20.,3*30.,27*20./
DATA HA2 /3*30.,29*20.,30.,221*20.,30.,28*20.,2*30.,26*30.,30.,
1 20.,30.,27*20.,30.,38*20./
DATA XR/19.42,20.16,18.66,20.24,17.72,21.18,18.12,20.76,16.22,20.4
10, 16.52,22.48,18.04,22.32,16.32,20.42,14.14,21.02,14.16,22.44,14.9
26,24.14,16.86,20.24,18.50,18.88,15.42,22.30,13.78,23.26,11.14,24.7
36,10.34,27.10,10.68,26.04,13.96,23.54,14.74,16.64,18.66,16.32,14.1
44,16.88,10.34,20.24,20.48,22.48,22.30,26.04/
DATA YR/ 20.86,20.06,19.16,22.36,20.14,21.20,22.50,19.20,21.16,
117.88,18.92,18.64,17.76,16.64,24.12,25.20,21.46,23.84,20.26,20.64,
218.04,19.78,15.88,15.80,28.42,14.94,27.88,26.86,25.06,25.04,23.64,
323.22,20.34,19.98,17.18,17.06,16.06,14.52,14.32,13.04,19.16,24.12,
421.46,15.88,20.34,22.36,17.88,18.64,26.86,17.06/
DATA ZR/50*0./
DATA TFR/ 8.11, 9.07, 9.12, 8.15, 6.64, 4.76, 1.83, 0.15,
1 -5.60, -7.61, -8.72, -7.84, -5.55, -5.87, -4.09, -2.86,
2 1.50, -1.43, -0.41, -0.61, -1.49, -0.60, 1.23, 4.78,
3 11.11, 10.61, 9.69, 8.54, 7.08, 3.13, -2.15, -7.32,
4 -7.61, -8.85, -8.44, -7.46, -6.73, -6.25, -5.11, -4.08,
5 -3.17, -2.41, -0.77, -0.01, 2.56, 3.22, 5.33, 9.11/
DATA TFR2/
6 10.08, 11.97, 9.69, 8.43, 6.65, 4.24, 1.85, -0.73,
7 -6.30, -8.01, -7.26, -9.34, -8.28, -8.07, -7.78, -6.14,
8 -5.28, -3.82, -1.73, -0.86, 2.31, 3.85, 5.71, 8.74/
DATA TFCI/16.87, 17.82, 18.43, 16.90, 15.15, 12.86, 9.47, 8.63,
1 6.01, 4.82, 2.64, 1.38, 0.30, 0.55, 2.98, 4.49,
2 6.25, 8.70, 9.92, 10.10, 10.47, 12.01, 12.23, 12.03,
3 13.32, 13.23, 12.54, 10.43, 5.64, -1.75, -8.04,-11.69,
4 -13.91,-12.94,-12.43,-12.53,-12.39,-11.19, -9.62, -7.88,
-------
DATA TFCI2/
6 17.03, 18.19, 16.30, 14.55, 10.17, 3.19, -0.13, -4.13,
7 -7.14, -6.74, -7.00, -6.78, -7.11, -7.01, -3.84, -1.84,
8 -0.66, 2.60, 5.97, 7.92, 8.90, 11.47, 13.48, 15.86/
DATA DFC/0.200,0.200,0.200,0.200,0.200,0.200,1.960,1.960,
1 1.820,1.820,1.750,1.750,1.690,1.690,1.620,1.620,
2 1.480,1.480,0.680,0.680,0.200,0.200,0.200,0.200/
DATA DFA/0.176,0.176,0.037,0.037,0.111,0.111,1.575,1.575,
1 2.261,2.261,1.001,1.001,0.788,0.788,1.325,1.325,
2 2.632,2.632,1.084,1.084,0.573,0.573,0.417,0.4177
DATA SULFR/4*2.863,2.517,1.963,3.3947
DATA IfU/57
DATA IW1/67
C XR,YR VALUES IN LOCATIONS 41-50 ARE THE LOCATIONS FOR THE 2HR SAMPLERS
C IN THE SEQUENCE THE 2HR CONC DBS ARE LISTED ON THE DATA CARDS
C SULFR DATA IS THE MEAN VALUE UF 12 WEEK PERIOD
C INPUT OUTPUT DEVICE NUMBERS USED AS FOLLOWS
C IR1 CARD READER
C IW1 ON LINE PRINTER
REWIND 10
REWIND 11
REWIND 12
REWIND 13
REWIND 14
REWIND 15
REWIND 16
1R1 = 5
READ(IRlt9000) IH1,IH2
9000 FORMAT(10I5)
DO 1 I=1,NRPNT
XP(I) = 1524. * XP( I )
YP( I) = 1524. * YP(I )
1 CONTINUE
CALL INA(QB,QRH,QCB,QCIH,QAU,QSA,SATA,SUNA,SMIDA,SWIGA,IR6,GX,GY)
I I = IH1 - 1
IF (II) 8,8,2
2 CONTINUE
DO 3 1=1,11
READ (10)
READ (11)
READ (15)
READ (16)
3 CONTINUE
II = 11/24
IF (II) 8, 8,12
12 CONTINUE
DO 13 1=1,11
READ (12)
13 CONTINUE
II = (IH1 - l)/2
IF (II) 8,8,14
14 CONTINUE
DO 15 1=1,11
READ (13)
15 CONTINUE
8 CONTINUE
DO 2000 IRR = IH1,IH2
JN = IRR
C READ MET DATA
CALL METIN( JN,YMDH,YEAR,AMNTH,DAY,HOUR,I DOW,WS,WH,P,WD,
1 INDEX,CIGMX.SIGA,RIB,PCPN,WGLD,STAPR,TG, TEMP)
IF (I - 614) 11,10,11
10 CONTINUE
YMDH = 2704.
11 CONTINUE
C LOAD EMISSION DATA
CALL SHIFT(ID,IS,IDOW,HOUR)
-------
DDR = 65. - (TEMP + TFRUH.IDM
IF (DDR) A,5,5
4 DDR = 0.
5 CONTINUE
DDC = 65. - (TEMP + TFCI(IH,ID)»
IF (DDC) 6,7,7
6 DDC =0.
7 CONTINUE
CALL IEMIT(NIPS,WS,WH,P,DDC,IS,ID,IH,CIGMX,TG,ZP,QP)
C CALL INPUT ROUTINE TO READ PLOAD DATA -INPLD
CALL INPLDl JN,PLOAD,PMDH)
IF( PMDH - YMDH) 100,200,100
100 CONTINUE
WRITElIW1,9200)YMDH,YEAR,PMDH
9200 FORMAT(«0 DATE TIME GROUP FROM PLOAD DATA AND MET DATA DO NOT MATC
IH*** MET DATA YMDH= ',F7.0,« YEAR= • ,F3.0,/,58X,' PLOAD DATE GROU
2P= «,F7.0,///)
200 CONTINUE
PLOAD(7) = PLOAD17) * 14.07 + 20. + PLOAD(8)
IU = NIPS + 1
CALL UEMIT(NIPS,NUS,TEMP,STAPR,TG,WS,WH,P,PLOAD,SULFR,CIGMX,ZP(IU)
1,QP(IU))
CALL AEMIT(HA,GX.GY,ID.IS,DDR,DDC,DFC(IH),DFA(IH),QB,QRH,QCB,OCIH,
1QAU,QSA,SATA,SUNA,SMIDA,SWIGA,QIJ)
C DETERMINE IF HOUR OF DATA IS ODD OR EVEN—IF ODD FILL THE S02 ARRAY
C WITH ZEROS IF EVEN READ THE 2 HR S02 OBS
DO 1300 NEO = 2,24,2
FNEO = NEO
IF (HOUR -FNEO) 1300,1400,1300
1300 CONTINUE
DO 1310 I=1,NRECP
OBS02(I) = 0
1310 CONTINUE
GO TO 1200
1400 CONTINUE
C SUBROUTINE NS02 READS IN THE 2HRLY S02 OBSERVATIONS
CALL NS02( JN,OBS02,SMDH3,NR1,NRECP)
IF (SMDH3 - 120000.) 1420,1410,1410
1410 CONTINUE
SMDH3 = SMDH3 - 120000.
1420 CONTINUE
C CHECK 2HRLY DTG WITH MET DATA YMDH FOR CORRECT S02 OBS.
IF ( YMDH - SMDH3) 1500,1700,1500
1500 CONTINUE
C WRITE ERROR MESSAGE FOR NONMATCHING DTG FROM 2 HR S02 OBS
WRITE (IW1.9600) SMDH3,YMDH
9600 FORMAT ('0 DATE-TIME FROM 2HR S02 OBS DOES NOT MATCH DATE-TIME OF
1 MET DATA - DTG FROM S02 IS ',F7.0,» DTG OF MET. IS ',F7.0///J
1700 CONTINUE
IF(HOUR - 14.) 1100,300,1100
300 CONTINUE
C CALL IN 24 HOUR S02 OBSERVATIONS WITH DIG READ FROM EACH CARD
CALL NS024 ( JN,OBS02,SMDH1,NR1)
DTGC = SMDH1 - 120000.
IF (DTGC ) 1000,1000,800
800 CONTINUE
IF ( DTGC - YMDH) 900,1200,900
900 CONTINUE
WRITE (IW1.9500) SMDH1.YMDH
9500 FORMAT («0 DATE-TIME OF 24 HR S02 DATA READ BY NS024 DOES NOT MATC
IH DATE-TIME OF MET. DATA S02 DTG= ',F7.0,' MET DTG= «,F7.0,///)
GO TO 1200
1000 CONTINUE
IF ( YMDH - SMDH1) 900,1200,900
1100 CONTINUE
DO 1110 1=1,NR1
OBS02U) = 0
-------
1200 CONTINUE
CALL HROUT ( IW2,YMDH,YEAR,AMNTH,DAY,I DOW,HOUR,WS,WH,P,WD,
1 INDEX,CIGMX,SIGA,RIB,PCPN,WGLD,STAPR,TG, TEMP,NRECP,NR1,XR,YR,
2 ZR,OBS02,CAS02,GX,GY,DLTA,NH,Z.NRPNT,XP,YP,ZP,QP)
CALL OUTA(YMDH,GX,GY,NH,QIJ,IW3)
WRITE (IW1,9700) YMDH
9700 FORMAT (' RECORD WITH INDEX =•, FIO.O,1, WRITTEN ON UNIT 15 AND 16
I1 )
IF {YMDH - DLAST) 1800,2100,2100
1800 CONTINUE
2000 CONTINUE
2100 CONTINUE
END FILE 15
END FILE 16
REWIND 10
REWIND 11
REWIND 12
REWIND 13
REWIND 14
REWIND 15
REWIND 16
CALL EXIT
END
SUBROUTINE AEMIT(HA,GX,GY,ID,IS,DDR,DDC,DFC,DFA,QB,QRH,QCB,QCIH,
1QAU,QSA,SATA,SUNA,SMIDA,SWIGA,QIJ)
C THIS ROUTINE COMPUTES A THREE DIMENSIONAL ARRAY OF EMISSION RATES FOR
C AREA SOURCE
DIMENSION QIJ(l),HA(1),QB(1),QRH(1),QCB(1),QCIH(1),QAU(1),QSA(1),
1SATA(1),SUNA(1),SMIDA(1),SWIGA(1)
NI = GX * GY
DO 20 I=1,NI
C GET DAY-OF-WEEK EMISSION FACTOR
GO TO (I,2t3) tlD
1 DOWF = SUNA(I)
GO TO A
2 DOWF = 1.
GO TO 4
3 DOWF = SATAU )
4 CONTINUE
C GET HOUR-OF-DAY EMISSION FACTOR
GO TO (5.6,7),IS
5 SHFTF = SMIDA(I)
GO TO 8
6 SHFTF = 1.
GO TO 8
7 SHFTF = SWlGAd )
8 CONTINUE
QA = QB(I) + QRH(I) * DDR + QCB(I) * DFC + QCIH(I) * DDC + QAU(I)
1 * DFA + QSA(I) * DOWF * SHFTF
QIJ(I ) = 0.
K = I + NI
QIJ(K) = 0.
L = K + NI
QIJ(L) = 0.
IF (HA(I) - 30.) 9,10,11
9 QIJU ) = QA
GO TO 12
10 QIJ(K) = QA
GO TO 12
11 QU(L) = QA
12 CONTINUE
20 CONTINUE
RETURN
-------
SUBROUTINE UEMI T ( NIPS,NUS , TBAR , PRESS , TG,U1 , Zl ,P , PLOAD , SULFR ,C IGMX ,
1ZP,QP)
C THIS ROUTINE COMPUTES EMISSION RATE (
C AND EFFECTIVE HEIGHT (ZP) FOR UTILITY POINT SOURCES
C USES INPUTS
C TBAR - AIR TEMPERATURE, DEG K
C PRESS - AIR PRESSURE, IN HG
C NUS - NUMBER OF STACKS
C PLOAD - POWER LOAD, MEGAWATTS
C SULFR - SULFUR CONTENT OF FUEL, PERCENT
C A - PARAMETERS OF FUEL - LOAD RELAT I ONSHI P , LB/MIN -
C B - PARAMETERS OF FLUE GAS TEMP - LOAD RELATIONSHIP,
C C - PARAMETERS OF FLUE GAS FLOW RATE - LOAD RELATION
DIMENSION A(3,7),B(4,7),C(4,7) ,PLOAD(1) ,SULFR(1) ,HPT(7) , ZP(l) ,
IQP(l)
DATA EMIS/2./,FCON/0.075598/
DATA A/ 50. ,10. 89 , 150., 50. ,10. 89 , 150. , 100. , 10. 43 , 300.,
1 100. ,10. 75 , 405., 20. ,13. , 480., O.,16.67 , 150.,
2 0., 1.536, 1500. /
DATA B/281., 294., 3 12., 343., 281., 294., 312., 343., 242., 262., 290., 336.
1, 236., 267., 323., 372., 252., 293., 3 12. ,352. , 350. ,406. ,420., 434.
2, 296. ,323. ,342. ,356. /
DATA C/ 40., 170. ,320. ,500., 40. , 170. ,320. ,500. ,210. ,375. ,625. ,
11015. ,260., 490., 840., 1460. , 30. , 625. ,1180. ,1840. ,20. , 260. , 520. , 835 .
2, 5. ,225. ,440., 665. /
DATA HPT/ 2*76. 5, 2* 106. 7, 72. 3, 100.3,56.27
C CONVERT TBAR FROM DEC F TO DEG K
TEMP = (TBAR - 32) / 1.8 + 273.
C EMIS = GM S02 EMITTED PER GM SULFUR IN FUEL ANALYSIS
C FCON = FUEL UNIT CONVERSION FACTOR = (GM/LB) * (MIN/SEC) * (1./100PER
QCON = EMIS * FCON
DO 20 1=1, NUS
C COMPUTE S02 EMISSION
IF (PLOAD(I)) 2,2,3
2 OP( I) = 0.
GO TO 4
3 CONTINUE
FUEL = A(lil) + A(2,I) * PLOAD(I)
QP(I) = QCON * SULFR(I) * FUEL
4 CONTINUE
C COMPUTE FLUE GAS TEMPERATURE (DEG K) AND FLOW RATE (CU. CM/SEC)
CHKLD = PLOADl I ) / A(3,I )
PLFR = CHKLD - 0.66667
C
C
C
IF (PLFR) 10,13,13
10 PLFR = CHKLD - 0.33333
IF (PLFR) 11,12,12
11 K2 = 2
FR = 3. * CHKLD
GO TO 15
12 K2 = 3
GO TO 14
13 K2 = 4
14 FR = 3. * PLFR
15 Kl = K2 - 1
STKTP = (B(K1,I) + FR * (B(K2,I) - B(K1,IM - 32.) / 1.8 +
STKFL = (C(K1,I) + FR * (C(K2,I) - C(K1,I))) * 471.95E3
COMPUTE HEAT EMISSION
CONSTANT . 00206495=2. 553 ( SP. HEAT (CONST. VOL ) /GAS CONST ) *33863.9
IN.HG132F) )*2.38848E-8(CAL/DYNE CM)
273.
(DYNE/SQ
QHEAT = 0.00206495 * PRESS * STKFL * (STKTP - TEMP) / STKTP
WS = Ul * (HPT(I) / Z1)**P
CALL PLUMZITG, QHEAT, TEMP, HPT ( I),WS,ZP(I ) ,QP ( I ) ,C IGMX)
20 CONTINUE
RETURN
-------
c
c
SUBROUTINE
IEMIT(NIPS,Ul,Zl,P,DDC,IS,ID,IH,CIGMX,TG,ZPj,QP)
THIS ROUTINE COMPUTES EMISSION RATE
HEIGHT (ZP) FOR INDVSTRIAL POINT SOURCES
ZP( 1) ,QP(1)
QSI(44),QSI2(4),OHI(44),HPT(44),UDH(44)
SAT (44),SUN (44),SWIG (44),SMID (44)
Q(12,4),UDHS(12,4)
AND EFFECTIVE
DIMENSION
DIMENSION
DIMENSION
DIMENSION ...
EQUIVALENCE (QSI(37),QSI2(1 ) )
DATA QHI/190.86,575.63, 75.45, 74.60, 1.29,200.31, 17.82,203.19,
1 60.00, 0.00,361.00, 0.00, 22.70, 7.24, 7.66, 15.25,
2 31.19, 09.01,101.23,189.38, 69.44, 0.00, 0.11,141.83,
3 20.73,238.94, 0.00, 22.32, 27.86,279.60,798.18,176.01,
4 6.78, 3.62,390.81,533.55, 0.00, 0.00, 24.95, 62. 37/
DATA QSI/ .0383, 35.0118, .0009,397.5426, 16.7393, 94.1626,
1 24.8305,390.6217,18.9670, 19.3968, 0.0 , 33.3890,
2 7.5876, 73.3152, 3.2161, 10.7800, 4.1580, 10.3950,
3 57.6870, 20.2691, 0.0 ,305.3670, 60.9396, 63.6274,
4 7.8490, 31.5156,162.7519, 4.9417, 26.3866,911.3780,
5 0.2635, 35.2936, 73.1896, 20.2018, 18. 7807 .334.0416/
DATA QSI2/585. 5721, 25. 6327, 23.1000, 12.9362/
DATA HPT /40., 20. ,7*40. ,15. ,20. ,4*40. ,49. ,19. ,47. ,4*75. ,2*55.,
1 3*75., 2*55., 2*75. ,2*55. , 75. , 3*55. ,75. , 59. ,64. ,3*76.2,
2 107. 5/
DATA UDH /O., 806., 0., 287., 0., 246. ,0. ,428. ,2*0., 487. ,10*0. ,218.,0.,
I 77. ,0., 97. ,106.
DATA SAT/3*0.,1., .99, 1.
1 ,0. ,1.,0. , l.,0.,
DATA SUN/3*0.,1.,.99, 1.
1 ,0. , l.,0. , 1. ,0. ,
DATA SWIG /4*1.,.99,13*1
DATA SMID /2*0.,2*1.,.99
1 2*1. /
DATA Q/4*106. ,529. ,423.,
1 106. ,3*282. ,600. ,7*706.
DATA UDHS/4*40.,201., 161
1 3*102. ,217. ,7*256. ,102.
DATA ISTAR/0/
IF (ISTAR) 31,29,31
29 CONTINUE
ISTAR = 1
DO 30 1=1,40
QHKI ) = 0.01 * QHK I )
30 CONTINUE
31 CONTINUE
DO 10 1=1, NIPS
J = I - 40
IF (J) 20,20,15
15 CONTINUE
QHI( I = 0
SATU = 1.
SUNU = 1.
SW1G( ) = 1.
SMID( ) = 1.
11 = IH + 1) / 2
QSK I = Q( 11, J)
UDH( I = UDHS(IltJ)
20 CONTINUE
,2*0.
,0.,1
6*1.,
,0.,1
6*1.,
. ,0. ,
,5*1.
353.,
,282.
.,134
,504.
GET DAY-OF-WEEK EMISSION FACTOR,
GO TO (1,2, 3), ID
1 DOWF = SUN( I )
GO TO 4
2 DOWF = 1.
GO TO 4
3 DOWF = SAT( I )
4 CONTINUE
,202., 402. ,3*0. ,101. ,650. , 1030. , 3*0. /
.,0. ,1.,0. ,1. ,0. ,1. ,0. ,3*1. ,3*0. ,2*1.
0. ,2*1. ,0. ,2*1. /
. ,0.,1.,0.,1.,0.,1.,0.,3*1.,3*0.,2*1.
0. ,2*1. tO. ,2*1.7
8*1. ,.48,12*1.7
,0.,1.,0.,5*1.,3*0.,6*1.,.48,9*1.,0.,
212., 176., 318., 176. ,5*106. ,7*529. ,
,1341., 1552., 2364., 8*2681., 1799. /
. ,80., 67. ,120. ,67. ,5*40. ,7*201. ,40.,
,583., 888. ,8*1007., 676. 7
1=SUNDAY, 2=WEEKDAY, 3=SATURDAY
-------
C GET HOUR-OF-DAY EMISSION FACTOR, 1=01-08, 2=09-16, 3=17-24
GO TO (5,6,7),IS
5 SHFTF = SMID(I)
GO TO 8
6 SHFTF = 1.
GO TO 8
7 SHFTF = SWIG(I)
8 CONTINUE
QP(I) = QHI(I) * DDC + QSKI) * DOWF * SHFTF
WS = Ul * (HPT(I) /Z1)**P
ZP(I) = UDH(I)/ WS + HPT(I )
IF (TG - 999.) 9,13,9
9 IF (TG) 13,13,11
11 CONTINUE
ANUM = UDH(I) / (2.9 * WS)
DZCRI = 2. * ANUM * SQRTUNUM * (TG + 0.0098) / (TG * CIGMX))
IF (CIGMX - HPT(I) - DZCRI) 12,12,13
12 QP(1 ) = 0
13 CONTINUE
10 CONTINUE
RETURN
END
SUBROUTINE SHIFTlID,IS,KDOW,HOUR)
DETERMINE DAY OF WEEK
IF (KDOW - 2) 5,6,7
SUNDAY AND HOLIDAY
5 ID = 1
GO TO 9
WEEKDAY
6 ID = 2
GO TO 9
7 IF (KDOW - 7) 6,8,5
SATURDAY
8 ID = 3
9 CONTINUE
DETERMINE SHIFT
IF (HOUR - 9.) 10,11,11
MIDNIGHT SHIFT (01 - 08)
10 IS = 1
GO TO 14
11 IF (HOUR - 17.) 12,13,13
DAY SHIFT (09 - 16)
12 IS = 2
GO TO 14
SWING SHIFT (17 - 24)
13 IS = 3
14 CONTINUE
RETURN
-------
SUBROUTINE HROUT (IW2,YMDH,YEAR,AMNTHfDAY,I DOW,HOUR,WSfWHfPtWD,
1 INDEX,CIGMX,SIGA,RIB,PCPN,WGLD,STAPR,TG,AVTMP,NRECP,NR1,XR,YR,
2 ZR,OBS02,CAS02.GX,GY.DLTA,NH,Z.NRPNTfXP,YP,ZP,QP)
DIMENSION XRCI),YR(1),ZRC1),OBS02(1),Z(I),XP(I),YPI 1),ZP(1),QPC1)
1,CAS02(1)
HRI TEC 15) YMDH,YEAR,AMNTH»DAY,I DOW,HOUR.NRECP,NR1,(XR(N),N=1.
1 NRECP).(YR(N),N=1.NRECP),(ZRIN),N=l,NRECP),(OBS02CN),N=1,NRECP),
2 (CAS02CN),N=1,NR1),
3 GX,GY,DLTA,NH,(Z(N),N=1,NH),NRPNT,(XP(N),N=1,NRPNT),(YP(N),N=1,
4 NRPNT),CZP(N),N=1,NRPNT),(QP(N),N=l,NRPNT),WS,WH,P,WD,INDEX,
5 AVTMP,CIGMX,SIGA,RIB,PCPN,WGLD,STAPR
RETURN
END
SUBROUTINE INA (QB,RH,CB,QCIH ,AUM,QBA,SATA,SUNA,SMIDA, SWGA.I01,
IGXtGY)
DIMENSION QB(1),RH(1),CB(1),QCIH(1), AUM(1),QBA(1),
1 SATAIL)tSUNA(l),SWGA(l),SMIDA(l)
NG = GX * GY
DO 100 1= 1,NG
READC 14) RBtRHtI)VCB(I)tCHtVMfAUMd),RRM,BBM,
1 QBA(I),QHA ,SATA(I),SUNA(I),SWGA{I).SMIDACI)
QB(I) = RB + VM + RRM + BBM
QCIHCI) = CH + QHA
100 CONTINUE
RETURN
END
SUBROUTINE INPLD (IR2,PLOAD,PMDH)
DIMENSION PLOAD(l)
READC11) PMDH,CPLOADCN),N=1,8)
RETURN
END
L A l^ L/ (_ /\
RETURN
END
YEAR,AMNTH,DAY,HOUR,IDOW,WS,WH,P,WD,
I,WGLD,STAPR,TG,AVTMP )
ITH,DAY,HOUR,I DOW,WS,WH,P,WD,
ItHGLDtSTAPRiTGtAVTMP
SUBROUTINE NS02C IRA,OBS02,SMDH3,NR1,NRECP)
DIMENSION OBS02C1)
Nl = NR1 + 1
READ (13) SMDH3,( OBS02CN),N=N1,NRECP)
RETURN
-------
SUBROUTINE NS024(IR3,OBS02, SMDH1,NR1)
DIMENSION OBS02( 1)
READ (12) SMDH1,(OBS02(N),N=1,NR1)
RETURN
END
SUBROUTINE OUTA(YMDH,GX,GY,NH,QIJ,IW3)
DIMENSION QIJ(l)
N = GX * GY * NH
WRITE (16) YMDH,GX,GY,NH,(QIJ(I),I=1,N)
RETURN
END
SUBROUTINE PLUMZ(TG,QH,TBAR,STHGT,WS,EFFHT,QP,CIGMX)
C ROUTINE TO COMPUTE EFFECTIVE SOURCE HEIGHT USING BRIGGS PLUME RISE EQS
C INPUTS ARE
C TG - VERTICAL TEMERATURE, GRADIENT, DEG K / METER
C QH - HEAT EMISSION, CAL/SEC
C TBAR - TEMPERATURE, DEG K
C STHGT - STACK HEIGHT, METERS
C WS - WIND SPEED AT STACK HEIGHT, METERS/SEC
C DIST - TRAVEL DISTANCE, METERS
C PLUME RISE IS BASED ON TRAVEL DISTANCE OF FIVE TIMES DISTANCE AT WHICH
C TURBULENCE DOMINATES ENTRAINMENT, I.E. X/X* = 5
F = 0.000037 * QH
C IF TG IS MISSING OR NON-POSITIVE, USE UNSTABLE OR NEUTRAL FORMULAS
IF (TG - 999.) 1,4.4
1 IF (TG + 0.008) 4,4,2
C COMPUTE STABILITY PARAMETER S
2 THG = TG +0.0098
S = 9.8 * THG / TBAR
C IF PLUME PENETRATES INVERSION, SET QP = 0
IF (STHGT - CIGMX) 8,8,3
3 QP = 0.
8 CONTINUE
C COMPUTE STABLE PLUME RISE
DH = 2.9 * (F / (WS * S))**0.33
GO TO 10
C FIND RAPID RISE DISTANCE XI FOR NON-STABLE CONDITIONS
4 IF (STHGT - 305.) 5,5,6
5 XI = 2.16 * F**0.4 * STHGT**0.6
GO TO 7
6 XI = 67.3 * F**0.4
C COMPUTE NON-STABLE PLUME RISE
7 CONTINUE
XOX1 = 5.
DH = 1.6 * F**0.33 * Xl**0.67 * (0.4 + 0.6*XOX1 + 2.2*XOX1*XOX1)
I/ (WS * (1. + 0.8*XOX1)**2)
C ADD PLUME RISE TO STACK HEIGHT
10 EFFHT = STHGT + DH
RETURN
-------
Appendix C
-------
Section 1.0
INTRODUCTION
This appendix presents descriptions, sources and summaries of
SOp emission, meteorological, and observed SOp concentration data for
Chicago which were used in the validation analysis of the study pre-
sented in the main body of this report. The principal data were obtained
on magnetic tape from Argonne National Laboratory and were generated by
their APICS data management system (Chamot, et al., 1970). Additional
data on source locations and emission inventory results were obtained
on punched cards. The data used in the study cover the period 0000
January 1, 1967 to 2300 January 31, 1967. Additional data covering
the 13-month period of December 1966 to December 1967 were obtained but
not used. A review of these data revealed irregularities in the data
and large blocks of missing data. It was assumed that representative
judgments could be made from the one-month sample.
-------
Section 2.0
EMISSION DATA
Emission information is available for point sources (the larger
emission sources) and a grid work of square areas one mile on a side.
Location coordinates of the points and the areas are referenced to fixed
\
coordinates. Information regarding each of these two types of sources
is discussed separately in subsequent paragraphs. The development of
these data is described in reports from Argonne National Laboratory
CCroke, et a.!., 1968a,b,c; Roberts, et al., 1970; Chamot, et al., 1970).
2.1 AREA SOURCE DATA
For each square mile area an estimate of the annual SOp emis-
sion rate was obtained for each of three classes of emitters. These
include:
Class I: Low rise residential structures consisting
of 19 or less dwelling units,
Class II: High rise residential structures, consisting
of 20 or more dwelling units, and commercial
and institutional buildings, and
Class III: Industrial plants not large enough to be
treated as individual point emitters.
In addition to the annual emission rates, estimates regarding effective
stack heights and diurnal variations in emission rates were generated
in Argonne's extensive study of Chicago emission data. Algorithms for
estimating diurnal variations in emission rates are given by Equations
(6), (7), and (8) in the Table C-l. In Equations (6) and (7), 20%
of the annual emissions are attributed to hot water requirements and
distributed evenly over the year. The remaining emissions for Classes
I and II are attributed to space heating requirements and are allocated
-------
Table C-l. Chicago Emission Rate Algorithms
(Roberts, et al., 1970; Chamot, et al., 1970)
Utility Source Emission Rate Algortihms **
Ni
Q. = 0.1533 S .2 (w.). (H.l. + B,)
J I --I J ' I I »
Industrial Source Emission Rate Algorithm**
(0.1533 FCSC+.0.1097 F0SQ) (Wp
Additional Source Emission Rate Algorithms
(1) Uniform proration
QA
Q = 8760"*
(2) Degree day plus hot water
Q =
8760
(65-T)
24 H
D
(3) Pumping station pattern
Q =
8760
Area Source Emission Rate Algorithms
(1) Residential or commercial low rise
Q =
Fw
8760"
(65-T-AR) Uh
DAHD
(D
(2)
(3)
(4)
(5)
(6)
*8760 hours is one year [continued
**0.1533 = 0.01 1368U)/240, where 3680 = Ib S0? emitted per ton sulfur
in coal, 240 = heat content of coal (therms/ton); 0.1097 = 0.01
(15790)/1440, where 15790 = Ib S0? emitted per 1000 gal sulfur in oil,
1440 = heat content of oil (therms/1000 gal).
-------
Table C-l. Chicago Emission Rate Algorithms (continued)
(2) Residential or commercial high rise
Q =
-FW) (65-T-Ac)
(7)
(3) Industrial
Q =
8760
(8)
Definition of Terms
Q = emission rate, Ib/hr
Q. = emission rate of jth stack in multiple stack
source, Ib/hr
S = percentage of sulfur in fuel
(w.). = fraction of emissions from ith generating unit
J going to jth stack
L. = generating unit output, megawatts
A.. = regression coefficient, therms/megawatt
B. = regression coefficient, therms
wi = fraction of emissions going to jth stack
•J
F = fraction of fuel requirement filled by coal
\*
F = fraction of fuel requirement filled by oil
S = percentage of sulfur in coal fuel
c
S = percentage of sulfur in oil fuel
U = fraction of average monthly process fuel used
during a particular shift (midnight, day or
swing on a weekday, Saturday or Sunday (also
holiday))
II = fraction of maximum process fuel requirement
m
used during mth month
(continued)
-------
Table C-l. Chicago Emission Rate Algorithms (concluded)
L = maximum process fuel usage rate, therms/hr
L = maximum space heating fuel usage rate, therms/hr
T = temperature, °F
Q. = annual S02 emission, Ib/yr
U. = fraction average hourly fuel usage associated
with hth hour
FW = 0.2 = fraction of annual residential/commercial fuel
usage attributed to hot water requirements
DA = 6155 = annual degree days, °F day/yr
HD = hours of fuel usage per day, hr/day
AD = Turner's residential heating temperature correction,
K Op
Ar = Turner's commercial heating temperature correction,
C op
Uh Values
(1) Pumping station pattern
Uh = 0.429 (hours 0 to 6)
Uh = 1.29 (hours 7 to 23)
(2) Area source
Uh = 1.5 (T £ 5°F, hours 4 and 5)
(5°F < T < 65°F, hours 6 and 7)
Uh = 1.0 (T <. 5°F, hours 6 to 22)
(5°F < T < 65°F, hours 8 to 22)
Uh = 0 (T <_ 5°F, hours 0 to 3 and 23)
(5°F < T < 65°F, hours 0 to 5 and 23)
(T > 65°F, all hours)
Hp Values
HD = 19 (T <5°F)
H =17 (T > 5°F)
-------
o
on the basis of outside air temperature deficit below 65 F. Only the
first term is applicable in these equations when the outside air
temperature is over 65°F. Equation (6) includes a "janitor" factor
U. to account for "hold fire" periods after 10 PM and for a 50% increase
in the burn rate during the first two early morning start up hours
(starting at 4 AM when temperature is <5°F and 6 AM otherwise).
The following effective source heights were used for each
class of emitters:
Class Effective Source Height
I 50 ft
II 200 ft
III 150 ft
The original sources of the area source data were as follows:
Class Data Source
I 1968 survey by Markets and Rates Dept. of
Peoples Gas, Light and Coke (PGLC) Co.
II 1968 (Residential) and 1961 (Commercial)
surveys by PGLC Co.
Ill 1963 survey of annual fuel use by Chicago
Dept. of Air Pollution Control
2.2 POINT SOURCES
Point sources are of two types: industrial and power plants.
Data relating to power plant emissions include generator operating
characteristics, stack heights, location coordinates of the plant site,
and the sulfur content of the fuel used. Observed thermal input
-------
requirements (fulfilled by burning coal) for observed power output
were estimated by Argonne (Roberts, et al.,1970) in the form :
T = AL+B
where T = thermal input requirement, BTU/hr
L = power output, megawatts
A,B = empirical parameters
The fitted coefficents and the hourly log of power output for the data
period were obtained from Argonne. The percentage of the burned fuel
exhaust gases diverted to each plant stack (w.) were obtained for each
J
generator unit. This information is utilized in Equation (1) of the
Table C-l to estimate the emission rate from each stack. The heat
content of coal is taken to be 240 therms/ton (a therm is 10 BTU's).
The empirical parameters A and B, the fuel sulfur content, the
stack height, the plant location coordinates, and the generator-to-stack
exhaust coefficients (w.) are available on punched cards. The hourly
J
power outputs for each generator are available on magnetic tape.
Data relating to the larger industrial emissions treated as
point sources include the process operating characteristics and fuel
requirements, the plant space heating requirements, the relative amounts
of each type fuel used to meet fuel requirements, the sulfur content of
each fuel, the percentage of exhaust gas allocated to each plant stack,
and the outside air temperature. Fuel requirements for spacing heating
were estimated by Argonne to vary from zero at 55°F to a maximum peak
-------
value at -10°F. A linear relationship with outside air temperature is
assumed to be valid as follows:
H = 55"T L
where H = fuel requirements for space heat, therms/hr
T = outside air temperature, °F
L = peak space heating requirement, therms/hr
Fuel requirements for industrial processes are related to seasonal and
diurnal operating characteristics by means of utilization factors
determined from survey questionnaires and interviews with plant
operators processed by Argonne (Roberts, et al., 1970). The fuel
requirement may be stated as follows:
Hp • Us Um Lp
where H = fuel requirement for industrial process, therm/hr
U = shift utilization for shift of day (midnight, day or
swing) and day of week, fraction of monthly utilization
U = month utilization, fraction of peak rate
L = peak fuel requirement for industrial processes,
" therm/hr
All of the above characteristics are reflected in the emission
rate algorithm of Equation (2) in Table C-l. Outside air temperatures
for the data period are available on magnetic tape. All other data and:
contained on punched cards. The heat content of coal is taken to be
240 therms/ton. The heat content of oil is taken to be 1440 therms/
1000 gal.
-------
Data were also collected regarding certain additional sources
which were considered appropriate for treatment as point sources but
for which the available data were not compatible with the power plant or
industrial emission algorithms. The annual emissions were reported for
each source. In addition each source was judged to conform to one of
three types of diurnal emission patterns. These patterns consist of
uniform emission, hot water plus temperature dependence, and pumping
station pattern in which the nighttime emission rate is about one-third
the daytime emission rate. The algorithms for estimating emission rates
are listed as Equations (3), (4) and (5) in Table C-l. The annual
emission rate for each source and its classification by diurnal emis-
sion pattern are available on punched cards along with the source location
coordinates and stack heights. Emissions are allocated equally among
stacks where more than one is present.
Effective stack height was computed using Briggs' (1969)
equations for plume rise and the reported stack height for each stack.
The heat content of exhaust gases was taken to be 15% of the thermal
fuel requirement computed for each source. The amount allocated to each
stack is analogous to the emission rate allocated to each stack.
In the case of the additional point sources discussed above, an available
estimate of the sulfur content of the coal used by each source was used to
convert S02 emission rates to heat emission rates by the following equation:
240 Q
He = 0.15
W)
3680N
where H = heat emission rate, therms/hr
-------
Q = S02 emission rate, Ib/hr
S = fuel sulfur content, percent
N = number of stacks
The numbers 240 and 3680 represent the therms per ton of coal and the
pounds of SOp emitted per ton of sulfur in burned fuel, respectively.
The number 0.15 is the fraction of coal heat content contained in the
exhaust gases.
-------
Section 3.0
METEOROLOGICAL DATA
The meteorological data for Chicago which are required for the
validation analysis include wind speed and direction, temperature,
clo.ud cover and types, cloud heights and ceiling, and the height of the
top of the mixing layer. Three types of data were utilized. These
include TAM (Telemetered Air Monitoring) Station wind data, Midway
Airport hourly airway observations and hourly estimates of the top of
the mixing layer.
An average wind speed and direction based on continuous
measurements over a 75-minute period centered on each hour was available
for each TAM site. The anemometer and wind vane height at these stations
was generally above building heights and varied from 40 feet to 180 feet;
the average height was 80 feet. A vector average of the observations
for each hour was used to determine the mean wind speed and direction
for the validation study.
The Midway Airport observations were used to get the outside
air temperature for making temperature dependent emission estimates and
for determining stability classifications by means of the Turner (1964)
classification scheme.
Argonne had obtained hourly estimates of the top of the mixing
layer using Midway surface temperatures and rural vertical temperature
profiles. The rural vertical temperature profile was constructed using
Green Bay and Peoria radiosonde data for 0600 and 1800 CST and the
-------
Argonne surface temperature. An interpolation between the two soundings
was made to fit the Argonne surface temperature if appropriate. Other-
wise, one sounding or the other was shifted to fit to the Argonne tem-
perature. Linear interpolations were made for hourly intervals between
the 12-hour observation periods. The mixing layer ceiling was < .
estimated by the intersection of the Midway potential temperature with
rural temperature profile.
-------
Section 4.0
S02 OBSERVATIONS
Mean hourly concentrations of S02 at each of eight TAM stations
were obtained on magnetic tape for the data period.
Concentrations measured by TAM station analyzers and averaged
over 24-hour periods were compared with 24-hour concentrations measured
by the West-Gaeke method (Booras and Zimmer 1968). The West-Gaeke
method averaged about 20% lower than the conductivity method. However.
as shown in the 2-hour versus 24-hour comparisons for St. Louis
(Appendix B), this could be due to the use of 24-hour samples in the
West-Gaeke method. The methods showed large deviations in both direc-
tions (either one high relative to the other). Interference from other
pollutants was clearly evident at certain locations (e.g., TAM
Station #3 in the Chicago Loop area) where the mean concentration
measured by conductivity was twice that measured by the West-Gaeke
method. As a result of this analysis and those reported by other
investigators (e.g., Shikiya and MacPhee 1969), it is seen that observed
concentrations for single steady-state periods may be in error by a
factor of two.
-------
Section 5.0
CHICAGO PREPROCESSING PROGRAM LISTINGS
C CHICAGO DATA PREPROCESSOR
DIMENSION XP(200),YP(200),HP{200),NS(30),SPCT( 30),QPTOT( 30),
1NPATI 30),HGTI(4,100),COALP(100),01LP(100),SPCIC(100),SPC10(100 )
2,XR(10),YR(10),ZR( 10),HA(5),ZP(200) ,WDAY(7,100),P(5),QH(200)
DIMENSION STKPI(4,100),PMUF(12,100)tSFWP(9,100),SPHTG(100),QP(200)
DIMENSION HGTU(6,6),NSU(6),NUNIT(6),SPCTU(6),STKPU(24,6),A(4,6)
1,8(4,6),ULOAD(4,6),QHU(6),QPU(6),QPS(6)
DIMENSION PROCL(IOO),OBCON(10),QPI(4),QHI(4),EFHGT(6),UDH(200)
DATA NADPT/27/
DATA NINPT/52/
DATA NUTPT/6/
DATA NRECP/8/
DATA XR/5.3,10.5,13.3,14.4,11.1,12.7,6.7,7.9/
DATA YR/25.6,10.7,17.9,11.6,11.7,5.3,18.8,9.3/
DATA ZR/8*0./
DATA GX/20./
DATA GY/30./
DATA DLTA/1609.3/
DATA NH/3/
C HA IS HEIGHT OF AREA SOURCES IN METERS
DATA HA/30.5,45.7,61./
DATA Zl/20./
DATA P/0.1,0.I,0.15,0.2,0.3/
DATA ICARD/5/
DATA IPRTR/6/
READ (ICARD,1003) IREC1,NREC
1003 FORMAT (215)
C GET POINT SOURCE EMISSION PARAMETERS
CALL INC1(NADPT,XP,YP,HP,NS ,SPCT,QPTOT,NPAT,IPRTR)
Nl = NADPT
CALL INC2(N1,NINPT,XP,YP,WDAYtHGTItCOALPtOILPtSPCICtSPCIOfSTKPI
ItPMUF.SFWPfSPHTGtPROCLtIPRTR)
Nl = Nl * NINPT
CALL INC3(N1,NUTPT,XP,YP,HGTU,NSU,NUNIT,SPCTU,STKPU,A,B,IPRTR)
II = IREC1 - 1
IF (II) 2,2,1
1 CONTINUE
DO 3 1 = 1, II
READ (16)
READ (17)
3 CONTINUE
2 CONTINUE
DO 400 11=1,NREC
IR = II
CALL INC4(CY,CM,CD,CH,DOW,THTA,U1,TEMP,CEIL,PRES,CSUM4,TUNC,CIGMX
1,OBCON,IR)
TBAR = (TEMP - 32.) * 1.8 + 273.
IF (TUNC - 5.) 6,5,4
4 CONTINUE
WRITE(IPRTR,1000) TUNC
1000 FORMATt' STABILITY PARAMETER IS OUT OF RANGE, TUNC =',F6.1)
5 CONTINUE
TG = -0.0065
GO TO 7
6 CONTINUE
TG = 999.
7 CONTINUE
INDEX = TUNC
PWIND = P(INDEX)
CALL INC5(CYY,CMM,CDD,CHH,DOWW,NUTPT,NUNIT,ULOAD,IR)
; CHECK DATE/TIME DATA
IF (CY - CYY) 10,20,10
10 CONTINUE
WRITE (IPRTR.lOOl) CY,CM,CD,CH,DOW,CYY,CMM,CDD,CHH,DOWW
1001 FORMAT!• DATE/TIME DATA FROM MET AND LOADtFILES DISAGREE YEAR
1 MONTH DAY HOUR WEEKDAY«/40X'MET FILE•,5F8.0/39X'LOAD FILE1
2,5F8.0)
-------
20 CONTINUE
IF (CM - CMM) 10,30,10
30 CONTINUE
IF (CD - CDD) 10,40,10
40 CONTINUE
IF (CH - CHH) 10,50,10
50 CONTINUE
IF (DOW - DOWW) 10,60,10
60 CONTINUE
M = CM
C COMPUTE EMISSION RATE AND EFFECTIVE HEIGHT FOR ADDITIONAL POINTS
DO 100 I2=1,NADPT
CALL APSHE(TEMP,SPCT(I2),QPTOT(I2),NPAT(12),NS(I 2),CH,QP(12)
1,UH(12))
WS = Ul * (HPU2) / Z1)**PWIND
IF (QH(I2)) 70,70,80
70 CONTINUE
ZPU2) = HP (12)
GO TO 90
80 CONTINUE
CALL PLUMZ(TG,QH(I2),TBAR,HP(I 2),WS,ZP(I 2),QP(12),CIGMX)
90 CONTINUE
UDHU2) = WS * (ZP(I2) - HPU2J)
100 CONTINUE
C COMPUTE EMISSION RATE AND EFFECTIVE HEIGHT FOR INDUSTRIAL POINTS
DO 200 I3=1,NINPT
C GET SHIFT AND WEEKDAY INDEXES
IF (DOW - 7.) 110,110,105
105 CONTINUE
ID = 3
GO TO 115
110 CONTINUE
IDW = DOW
ID = WDAY(IDW,I3)
115 CONTINUE
CALL SHFTC(HOUR,IS)
CALL 1PSHE(SPHTG(I 3),PROCL(I 3),TEMP,COALP(I 3),01LP(13),SPCIC(I 3)
1,SPCIO(I3),STKPI(1,I3),PMUF(1,I3),SFWP(1,I3),M,IS,ID,QPI,QHI)
NI = NADPT +13
QP(NI) = 0
ZPINI) = 0
QH(NI) = 0
HPdMI ) = 0
UDH(NI) = 0
DO 140 1=1,4
IF (HGTI(I,I3)) 140,140,120
120 CONTINUE
IF (QHKI)) 140,140,130
130 CONTINUE
WS = Ul * (HGTI(I,I3) / Z1)**PWIND
CALL PLUMZ(TG,QHI(I),TBAR,HGTI(I,13),WS,EFHGT(I),QPI(I),CIGMX)
OP(NI ) = QP(NI ) + QPK I )
QH(NI) = QH(NI) + STKPI(I,I3) * QHKI)
HP(NI) = HP(NI) + STKPI(I,I3) * HGTI(I,I3)
UDH(NI) = UDH(NI) + STKPI(I,I3) * WS * (EFHGT(I) - HGTI(I,I3))
ZP(NI) = ZP(NI) + STKPKI.I3) * EFHGT(I)
140 CONTINUE
200 CONTINUE
C COMPUTE EMISSION RATE AND EFFECTIVE HEIGHT FOR UTILITY POINT SOURCES
Nl = NI
DO 300 I4=1,NUTPT
CALL UPSHE(NSU(14),NUNIT(14),SPCTU(14), STKPU(1,I4)
1,A(1,I4),B(1,I4),ULOAD(1,I4),QHU,QPU,QPS)
NI = Nl + 14
QP(NI ) = 0
ZP(NI) = 0
QH(NI) = 0
-------
UDH(NI) = 0
II = NSUU4)
DO 240 1=1,II
IF (QHUUM 240,240,230
230 CONTINUE
WS = Ul * (HGTU(I,I4) / Z1)**PWIND
CALL PLUMZ(TG,QHU(I),TBAR,HGTU(1,14),WS,EFHGT(I),QPS(I),CIGMX)
QP(NI ) = QP(NI) + QPS(I)
QHINI) = QH(NI) + QHU(I)
240 CONTINUE
OH(NI) = QH(NI) / II
IF (QP(NI)J 300,300,244
244 CONTINUE
DO 250 1=1,11
EHW = QPS(I) / QP(NI)
ZP(NI) = ZP(NI) -«• EHW * EFHGT(I)
HP(NI) = HP(NI) + EHW * HGTUU.I4)
UDH(NI) = UDH(NI) + EHW * (EFHGT(I) - HGTU(I,I4))
250 CONTINUE
300 CONTINUE
NRPNT =NI
WRITE OUTPUT RECORD
CALL OUTC(CY,CM,CD,CH,DOW,NRECP,XR,YR,ZR,GX,GY,DLTA,NH,HA,NRPNT,XP
1,YP,ZP,QP,U1,Z1,PWIND,THTA,INDEX,CIGMX,TEMP,OBCON,IR,XNDX,PRES)
IF (CH - 23.) 320,310,320
310 CONTINUE
WRITE (IPRTR.1002) XNDX
1002 FORMAT(' RECORD, WITH INDEX =I,F10.0,1, WRITTEN ON UNIT 19')
320 CONTINUE
400 CONTINUE
END FILE 19
REWIND 12
REWIND 13
REWIND 14
REWIND 16
REWIND 17
REWIND 19
CALL EXIT
END
SUBROUTINE I NCI(NADPT,XP,YP,ZP,NS,SPCT,QPTOT,NPAT,IPRTR)
DIMENSION XP( 1) ,YP(1),ZP(l),NS(1)tSPCT(l),QPTOT(1).NPATll)
DO 100 1=1,NADPT
READ( 12 ) J,XP(I),YP(I) ,ZP(I ),NS(I),SPCT(I) ,QPTOT(I),NPAT(I)
IF (J - I) 10,20,10
10 CONTINUE
WRITE!IPRTR,1000) I,J
1000 FORMAT!' ORDER ERROR DETECTED READING RECORD NO.',16,', RECORD NO.
ION FILE WAS',16,' {INC1)')
CALL EXIT
20 CONTINUE
IF(NS(I)) 425,425,450
425 CONTINUE
NSl1) = 1
450 CONTINUE
C CONVERT STACK HEIGHTS FROM FT TO M
ZP( I ) = 0.3048 * ZP(I)
100 CONTINUE
RETURN
-------
SUBROUTINE INC2(Nl ,NINPT.XP,YP,A, H,COALP,01LP,SPCTC*SPCTO,STAKP
1,PMUF,SFWP,SPHTG,PROCL,IPRTR)
DIMENSION XPd)iYP(l), H(4,1),CCALP(1),01LP(1),SPCTC(1),A(7,1)
1,SPCTO(1),STAKP(4,1),PMUF(12,1),SFWP(9,1),SPHTG(1),PROCL(1)
I = Nl
DO 100 J=1,NINPT
1 = 1 + 1
READ( 13 )K,XP(I),YP(I),(H(L,J) , STAKP(L,J) , L = 1,4),(A(L,J),L = 1,7)
l.SPHTG(J),PRCCL(J), (PMUF(L.J),L=1,12),(SFWP(L,J),L=l,9),SPCTC(J)
2.COALPU) ,SPCTO(J) ,OILP(J)
IF {J - K) 10,20,10
10 CONTINUE
WRITE (IPRTR,1000) J,K
1000 FORMATM ORDER ERROR DETECTED READING RECORD NO.Sib,', RECORD NO.
ION FILE WAS1,16,' {INC2)«)
CALL EXIT
20 CONTINUE
DO 30 L = l,4
C CONVERT STACK HEIGHTS FROM FT TO M
H(L,J ) = 0.3048 * H(L,J)
C CONVERT STAKP TO FRACTION
STAKP(L,J) = STAKP(L,J) / 100.
30 CONTINUE
C CONVERT COALP AND OILP TO FRACTION
COALP(J) = COALP(J) * .01
OILPlJ) = .01 * OILP(J)
C CONVERT PUMF TO FRACTION
DO 200 L=l,12
PMUF(L.J) = PMUF(L,J) * .1
200 CONTINUE
C CONVERT SFWP TO FRACTION
DO 250 L=l,9
SFWP(L,J)=SFWP(L,J)*.01
250 CONTINUE
100 CONTINUE
RETURN
END
SUBROUTINE INC3(Nl,NUTPT,XP,YP,HGTU,NSU,NUNIT,SPCTU,STKPU,A,B
1, IPRTR)
DIMENSION XP(1),YP(1), NSU(1),NUNIT(1),SPCTU(1),STKPU(24,1)
1,A(4,1),B(4,1),HGTU(6,1)
I = Nl
DO 100 J=l,NUTPT
I = I + 1
READ( 14 ) K,XP(I),YP(I).NSU{JJ,(HGTUIL,J),L=1,6),NUNIT(J)
1,SPCTU(J),(STKPU(L,J),L=1,24),(A(L,J),L=1,4),(B(L,J),L=1,4)
IF (J - K) 10,20,10
10 CONTINUE
WRITE (IPRTR,1000) J,K
1000 FORMATP ORDER ERROR DETECTED READING RECORD NO.«,I6,', RECORD NO.
ION FILE WAS1,16,' ( INC3) • )
CALL EXIT
20 CONTINUE
C CONVERT STACK HEIGHTS FROM FT TO M
DO 30 L=l,6
HGTU(L,J) = 0.3048 * HGTU(L,J)
30 CONTINUE
C CONVERT STAKP TO A FRACTION
DO 50 L=l,24
STKPU(L,J)=STKPU(L,J)*.01
50 CONTINUE
100 CONTINUE
RETURN
END
-------
SUBROUTINE INC4(CY,CM,CD,CH,DOW, WD,U1,TEMP,C£IL,PRES,CSUM4,TUNC
1,CIGMX,C8CON,IR)
DIMENSION OBCON(1)
READ( 16 ) I,U1,WD,TEMP,CEIL,PRES,CSUM4,TUNC,CIGMX,(OBCON(K),
1K=1,8),CH,DOW,CD,CM,CY
RETURN
END
SUBROUTINE INC5(CYY,CMM,CDD,CHH,DOWW,NUTPT,NUNIT,ULOAD,IR)
DIMENSION NUNIT(l),ULOAD(4,1),ALOAD(20)
READ( 17 )I,CYY,CMM,CDD,DOWW,CHH,(ALOAD(J),J=l,15)
K = 0
DO 20 I=lfNUTPT
JJ = NUNIT(I)
DO 10 J=1,JJ
K = K + 1
ULOADU, I ) = ALOAD(K)
10 CONTINUE
20 CONTINUE
RETURN
END
SUBROUTINE PLUMZ(TG,QH,TBAR,STHGT,WS,EFFHT,QP,CIGMX)
C ROUTINE TO COMPUTE EFFECTIVE SOURCE HEIGHT USING BRIGGS PLUME RISE EQS
C INPUTS ARE
C TG - VERTICAL TEMERATURE, GRADIENT, DEG K / METER
C QH - HEAT EMISSION, CAL/SEC
C TBAR - TEMPERATURE, DEG K
C STHGT - STACK HEIGHT, METERS
C WS - WIND SPEED AT STACK HEIGHT, METERS/SEC
C DIST - TRAVEL DISTANCE, METERS
C PLUME RISE IS BASED ON TRAVEL DISTANCE OF FIVE TIMES DISTANCE AT WHICH
C TURBULENCE DOMINATES ENTRAINMENT, I.E. X/X* = 5
F = 0.000037 * QH
C IF TG IS MISSING OR NON-POSITIVE, USE UNSTABLE OR NEUTRAL FORMULAS
IF (TG - 999.) 1,4,4
1 IF (TG + 0.008)4,4,2
C COMPUTE STABILITY PARAMETER S
2 THG = TG +0.0098
S = 9.8 * THG / TBAR
C IF STACK HEIGHT EXCEEDS MIXING CEILING, SET QP = 0
IF (STHGT - CIGMX) 8,8,3
3 QP = 0.
8 CONTINUE
C COMPUTE STABLE PLUME RISE
DH = 2.9 * (F / (WS * S))**0.33
GO TO 10
C FIND RAPID RISE DISTANCE XI FOR NON-STABLE CONDITIONS
4 IF (STHGT - 305.) 5,5,6
5 XI = 2.16 * F**0.4 * STHGT**0.6
GO TO 7
6 XI = 67.3 * F**0.4
C COMPUTE NON-STABLE PLUME RISE
7 CONTINUE
XOX1 = 5.
DH = 1.6 * F**0.33 * Xl**0.67 * (0.4 + 0.6*XOX1 + 2.2*XOX1*XOX1)
I/ (WS * (1. + 0.8*XOX1)**2)
C ADD PLUME RISE TO STACK HEIGHT
10 EFFHT = STHGT + DH
RETURN
-------
SUBROUTINE APSHE(TEMP,SULPC,QPTOT,NPAT,NS,HR,QAPHR,QHEAT)
C INPUTS
C TEMP TEMPERATURE {DEG F)
C SULPC SULFUR CONTENT OF FUEL (PERCENT)
C QPTOT ANNUAL EMISSION OF S02 00 LB/YR)
C NPAT EMISSION PATTERN - 1=UNIFORM,2=TEMP DEPENDENT,3=PUMP
C NS NUMBER OF STACKS
C HR HOUR OF DAY
C OUTPUTS
C QAPHR S02 EMISSION RATE (G/SEC)
C QHEAT HEAT EMISSION RATE (CAL/SEC)
C CHEAT CONVERTS ( THERMS/HR) TO (CAL/SEC)
DATA CHEAT/7.OOOE3/
C THTON IS COAL HEAT CONTENT ( THERMS/TON)
DATA THTON/240./
C S02SU IS S02 EMISSION FACTOR FOR SULFUR IN COAL (LB S02/TON SULFUR)
DATA S02SU/3680./
C CRATE CONVERTS LB/HR TO G/SEC
DATA CRATE/0.1260/
C DEGDA IS ANNUAL DEGREE DAYS
DATA DEGDA/6155./
C DETERMINE METHOD OF CALCULATION FOR EMISSION PATTERN
GO TO (100,200,300),NPAT
100 CONTINUE
C UNIFORM EMISSION
QAPHR=QPTOT/B760.
GO TO 400
200 CONTINUE
C TEMP DEPENDENT EMISSION
IFITEMP-65) 250,225,225
225 CONTINUE
TE = 0.
GO TO 275
250 CONTINUE
TE=1.
275 CONTINUE
QAPHR=QPTOT*(.2/18760.) + .8*TE* ( 65.-TEMP )/(24. * DEGDA))
GO TO 400
C PUMP STATION PATTERN
300 CONTINUE
TPUMP=.429
IF(6.9-HR) 350,375,375
350 CONTINUE
IF(HR-23.1) 370,375,375
370 CONTINUE
C HOUR IS LT 23.1 AND GT 6.9
TPUMP=1.29
375 CONTINUE
QAPHR= QPTOT*TPUMP/8760.
400 CONTINUE
C CALCULATE HEAT EMISSION FOR SINGLE STACK SIMULATION OF MULTIPLE STACKS
C ASSUMPTIONS-
C - STACK HEIGHTS ARE EQUAL
C - S02 AND HEAT DISTRIBUTION AMONG STACKS IS UNIFORM
C - COAL IS USE AS A FUEL
QHEAT = (QAPHR * THTON) / (0.01 * SULPC * S02SU * NS)
QHEAT = QHEAT * CHEAT
C QA IS IN UNITS OF LBS/HR CONVERT TO GRAMS/SEC
QAPHR = CRATE * QAPHR
RETURN
-------
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
SUBROUTINE IPSHE(SPHTG,PROCL,TEMP,COALP,01LP,SULPCtSULPO,STAKP,
CPMUF,SFWP,M,IS,ID, QS02S,QTSTK)
IPSHE CALCULATES THE HOURLY EMISSION FROM INDUSTRIAL POINT SOURCES
INPUTS
MAX. SPACE HEATING REQD. (THERMS/HR)
MAX. PROCESS LOAD (THERMS/HR)
TEMPERATURE (DEG F)
COAL LOAD (FRACTION)
OIL LOAD (FRACTION)
SULFUR CONTENT OF COAL (PERCENT)
OF OIL (PERCENT)
ALLOCATION (FRACTION)
ALLOCATION (FRACTION)
ALLOCATION (FRACTION)
INDEX
OUTPUTS
SPHTG
PROCL
TEMP
COALP
OILP
SULPC
SULPO
STAKP
PMUF
SFWP
M
IS
ID
SULFUR CONTENT
STACK EMISSION
MONTH EMISSION
EMISSION
OF YEAR
SHIFT
MONTH
SHIFT
INDEX
DAY OF WEEK
1=1-8, 2=9-16, 3=17-24)
INDEX (1=WEEKDAY,2=SAT,3=SUN
OR HOLIDAY)
SULFUR IN COAL (LB S02/TON SULFUR)
SULFUR IN OIL (LB S02/1000 GAL SULFR)
(FRACTION)
QS02S S02 EMISSION RATE (G/SEC)
QTSTK HEAT EMISSION RATE (CAL/SEC)
REAL LS,LP,L,LC,LO
DIMENSION PMUF(12),SFWP(3,3),STAKP(4),QS02S(4),QTSTK(A)
C HEATC IS COAL HEAT CONTENT (THERMS/TON)
DATA HEATC/240./
C HEATO IS OIL HEAT CONTENT (THERMS/1000 GAL)
DATA HEATO/1440./
C S02CO IS S02 EMISSION FACTOR FOR
DATA S02CO/3680./
C S020I IS S02 EMISSION FACTOR FOR
DATA S020I/15790./
C HEATE IS HEAT LOSS TO FLUE GASES
DATA HEATE/0.15/
C CHEAT CONVERTS THERMS/HR TO CAL/SEC
DATA CHEAT/7.OOOE3/
C CRATE CONVERTS LB/HR TO G/SEC
DATA CRATE/0.1260/
IF (TEMP - 55.) 10,10,20
10 CONTINUE
LS = 0.
GO TO 30
20 CONTINUE
LS = SPHTG*(55.-TEMP)/65.
30 CONTINUE
LP = PROCL*PMUF(M)*SFWP(IS,ID)
C LS AND LP ARE THE ADJUSTED HEATING AND PROCESS LOADS
L = LP+LS
C L IS TOTAL LOAD
LC = COALP*L
LO = OILP*L
C LC AND LO ARE COAL AND OIL LOADS
C = LC / HEATC
C C IS THE HOURLY RATE OF COAL USE
0 = LO / HEATO
C 0 IS THE HOURLY RATE OF OIL USE
QC = 0.01 * SULPC * C * S02CO
QS = 0.01 * SULPO * 0 * S020I
C QC AND QS ARE S02 EMISSION RATES FOR COAL AND OIL
QS02 = (QC + QS) * CRATE
C ALLOCATE S02 EMISSION AND THERMAL OUTPUT TO STACKS
DO 300 1=1,4
QS02S(I)= STAKPlI)*QS02
QTSTK(I) = STAKP(I) * HEATE * L * CHEAT
300
CONTINUE
RETURN
-------
UPSHE(NSTAK,NUNIT,SULPCt
C
c
C
c
c
c
c
c
c
c
c
c
c
SUBROUTINE
CS02SK)
CALCULATE UTILITY POINT SOURCE HOURLY EMISSIONS
INPUTS
NSTAK
NUNIT
SULPC
STAKP
A
STAKP,A,B,XL,THOSK,S02UN,
OUTPUTS
B
XL
NO. OF STACKS
NO. OF GENERATOR UNITS
SULFUR CONTENT OF COAL (PERCENT)
STACK EMISSION ALLOCATION BY GEN.
REGRESSION COEF (THERMAL LOAD ON
REGRESSION COEF ,INTERCEPT
POWER OUTPUT (MEGAWATTS)
UNIT (FRACTION)
POWER OUTPUT),SLOPE
THOSK HEAT EMISSION
S02UN S02 EMISSION
S02SK S02 EMISSION
DIMENSION THIUN(4)
CS02SM6) ,A(4),B(4),XL(4)
C HEATE IS HEAT LOSS TO FLUE GASES
DATA HEATE/0.15/
C HEATC IS COAL HEAT CONTENT (THERMS/TON)
DATA HEATC/240./
C S02CO IS S02 EMISSION FACTOR FOR SULFUR
DATA S02CO/3680./
C CHEAT CONVERTS THERMS/HR TO CAL/SEC
DATA CHEAT/7000./
C CRATE CONVERTS LB/HR TO G/SEC
DATA CRATE/0.1260/
DO 30 K=l,NUNIT
C CALCULATE THERMAL LOADS FOR EACH UNIT
THI UN(K) = A(K)*XL(K)+B(K)
IF (THIUN(K)) 20,25,25
20 CONTINUE
THIUN(K) = 0.
25 CONTINUE
THOUN(K) = HEATE * THIUN(K)
C CALCULATE S02 EMISSION FROM K TH UNIT
C = THIUN(K) / HEATC
S02UN(K) = 0.01 * SULPC * C * S02CO
30 CONTINUE
DO 100 I=1,NSTAK
S02SK(I)=0.
THOSK(I)=0.
100 CONTINUE
C START STACK ALOCATION
DO 200 J = l,NSTAK
DO 200 K=l,NUNIT
S02SK( J) = STAKP( J , K )*S02UN ( K) + S02SMJ)
200 CONTINUE
C S02SK IS IN UNITS OF
DO 400 J=l,NSTAK
S02SKU) a CRATE
DO 300 K=l,NUNIT
THOSK(J)= STAKP(J,K)*THOUN(K)+THOSK(J)
300 CONTINUE
- THOSK(J) * CHEAT
RATE
RATE BY GEN. UNIT
RATE BY STACK
, STAKP(6,4),THOUN(4),THOSK(6)
(FRACTION)
S02UN(4)
IN COAL (LB S02/TON SULFUR)
LOOP
LBS/HR - CONVERT TO GRAMS / SEC
* S02SKU)
400
THOSM J)
CONTINUE
RETURN
-------
SUBROUTINE SHFTC(HOUR,IS)
DETERMINE SHIFT
IF (HOUR) 13,13,9
9 CONTINUE
IF (HOUR - 17.) 10,13,13
10 CONTINUE
IF (HOUR - 9.) 11,12,12
MIDNIGHT SHIFT (01 - 08)
11 CONTINUE
IS = 1
GO TO 14
DAY SHIFT (09 - 16)
12 CONTINUE
IS = 2
GO TO 14
SWING SHIFT (17 -
13 CONTINUE
IS = 3
14 CONTINUE
RETURN
END
SUBROUTINE OUTC(CY,CM,CD,CH,DOW,NRECP,XR,YR,ZR,GX,GY,DLTA,NH,HA
1,NRPNT,XP,YP,ZP,QP,WS,WH,P,WD,INDEX,CIGMX,TEMP,OBS,IR,XNDX, STAPR)
DIMENSION XR(1),YR(1),ZR(1),HA(1) ,XP(1),YP(1),ZP(1),QP(1)
1,OBS(1)
C COMPUTE RECORD INDEX NUMBER
XNDX = CH -t- 100. * CD + 1.E4 * CM
C SET OTHER PARAMETER VALUES
IDOW = DOW
SIGA = 0.
RIB = 0.
PCPN = 0.
WGLD = 0.
WRITE ( 19 ) XNDX,CY,CM,CD,IDOW,CH,NRECP,(XR(I),I=1,NRECP)
1,(YR(I ),I = 1,NRECP),(ZR(I),I = 1,NRECP),(OBS(I),I=1,NRECP),GX,GY,DLTA
2,NH,(HA(I),I=1,NH),NRPNT,(XP(I), I=1.NRPNT), (YPU ),I=1,NRPNT)
3, (ZP( I ), I = 1,NRPNT),(QP(I), I = 1,NRPNT),WS,WH,P,WD,INDEX,TEMP,CIGMX
4,SIGA,RIB,PCPN,WGLD,STAPR
RETURN
END
-------
Appendix D
-------
Exhibit D-l
LISTING OF FORTRAN CODE COMPUTER PROGRAM
AND SUBROUTINES USED FOR VALIDATION CALCULATIONS*
*Additional subroutines are listed in Exhibit D-3.
-------
c
c
c
c
c
c
c
c
DIFFUS2
MAIN PROGRAM FOR VALIDATION CALCULATIONS
DIMENSION YEARR(2) ,AMONN(2) ,DAYY12) ,HOURR(2)
DIMENSION CNCT150), CAREA(50J,CPCIN(50),XRR(50),YRR(50)
1,ZRR(50),OBS02(50)
DIMENSION DISTX(1000), XDCAY(2000),SSIGZ(1000),EXPOZ(2000)
1,HA(5) ,UHA(5) ,QXY(5) ,Q(
-------
GO TO 160
22 CONTINUE
IF(NPS-NNPS) 23,23,21
23 CONTINUE
NGV = GX * GY * NH
IF (NGV - NA) 24,24,21
24 CONTINUE
IF(NH-NNH) 28,28,27
27 CONTINUE
WRITE (IPRTR,1005) NNH,NH
1005 FORMAT('NH EXCEEDS DIMENSION LIMIT OF',I 4,',NH=',14
1,'NH SET TO LIMIT')
NH=NNH
28 CONTINUE
AREA = DLTA * DLTA
DO 29 IG=1,NGV
Q(IG) = Q(IG) / AREA
29 CONTINUE
NPSS = NPS
XYMIN = 0.5 * DLTA
XSMAX = (GX + 0.5) * DLTA
YSMAX = (GY + 0.5) * DLTA
C CHECK VALIDITY OF WIND DATA, SKIP TO NEXT DATA SET IF WIND
C SPEED LIES OUTSIDE 0 TO 30 OR WIND DIRECTION LIES OUTSIDE
C -PI/2 TO 2.5*PI
IF (WSPD - 1.) 150,150,30
30 CONTINUE
IF (WSPD - 30.) 31,150,150
31 CONTINUE
IF (THTA - 7.9) 40, 40,150
40 CONTINUE
IF (THTA + 1.6) 150,150,41
41 CONTINUE
C SET COS AND SIN FACTORS FOR WIND/SOURCE COORDINATE CONVERSIONS
ISTAR = 0
CALL SCORD
CALL WCORD
CALL SIGZZ
IF (IERR) 50,45,50
45 CONTINUE
ISTAR = 1
CALL DISTB
IF (IERR) 60,60,50
50 CONTINUE
WRITE (IPRTRtlOOl) YEAR,AMON,DAY,HOUR
1001 FORMAT (• INPUT ERROR, GO TO NEXT DATA SET«,4F8.0)
GO TO 160
60 CONTINUE
C GET CONCENTRATION FOR EACH RECEPTOR LOCATION
ZR = -1
C NRECP = 10
DO 140 1 = 1,NRECP
XR=XRR(I) * DLTA
C XR = XRR(40 + I) * DLTA
YR=YRR(I) * DLTA
C YR = YRR(40+I) * DLTA
ZRL = ZR
ZR = ZRR(I)
IF (ZR - ZRL) 70,80,70
70 CONTINUE
CALL EXPZB
80 CONTINUE
CALL CONTR
CALL POINT
IF (IERR) 90,100,90
90 CONTINUE
WRITE (IPRTR,1001) YEAR,AMON,DAY,HOUR
GO TO 160
C
-------
100 CONTINUE
CAREA(I) = CONG
CPOIN(I) = CONPS
CNCT(I)= CONC+CONPS
C CONVERT CONCENTRATIONS FROM GRAM/SEC TO MICROGRAM/SEC
CNCT(I) = 1.E6 * CNCT(I)
140 CONTINUE
CALL OUTPT
GO TO 160
150 CONTINUE
WRITE( IPRTR,1003) WSPD,THTA
1003 FORMAT (' WIND INPUT IS UNACCEPTABLE,WSPD =',F6.1
1,•, THTA =',F7.3)
160 CONTINUE
CALL INCRT
IF(l.-STOP) 20tl70,20
170 CONTINUE
REWIND 18
CALL EXIT
END
SUBROUTINE OUTPT
DIMENSION CNCT(50),CAREA(50),CPCIN(50),XRR(50),YRR(50)
1,ZRR(50),OBS02(50)
DIMENSION DISTX(1000),XDCAY(2000),SSIGZ(1000),EXPOZ(2000)
ltHA(5),UHA(5),QXY(5),Q(4000)
DIMENSION XP (100),YP (100),ZP (100),QP(100)
COMMON/BASIC/IPRTR,I STAR,I ERR,ISIGD,NXL1M,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,I NDEX,THTA,CIGMX ,GX,GY , DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYM1N,XSMAX,YSMAX
COMMON/OUTPUT/YEAR,AMON,DAY,HOUR ,NRECP,CNCT,CAREA,CPGIN
1,XRR,YRR,ZRR,OBS02,IDOW,NR1,TEMP,SIGA,RIB,PCPN,WGLD,STAPR
COMMON/AREAS/NXI,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ,NH,HA
i,UHA,QXY,Q,CONC,NHl
COMMON/WNDSP/WSPD,WHGT,PWIND
COMMON/POINTS/NPS ,XP ,YP ,ZP ,OP,CONPS,NPSS
DATA IWOF/18/
CCOMPUTE RECORD INDEX NUMBER
Y = HOUR + 100 * DAY + 1.E4 * AMON
C WRITE DATA INTO FILE
WRITEUWOF )Y,YEAR.AMON , DAY , I DOW , HOUR , NRECP
1,(OBS02(N)fN=ltNRECP),(CNCT(N),N=1,NRECP)
2,WSPD,WHGT,PWIND,THTA,INDEX,TEMP,CIGMX,SIGA,RIB,PCPN,WGLD,STAPR
C 1, (OBS02(N),N=41,50),(CNCT(N),N=1,NRECP)
KRITE( IPRTR,107) Y,IWOF
107 FORMATS OUTPUT RECORD INDEX =',F10.0,', WRITTEN ON UNIT1,16)
RETURN
END
C
u
-------
SUBROUTINE DAFIL
C ROUTINE TO TRANSFER MODEL INPUTS FROM DISK AND TAPE TO CORE,
C ST.LOUIS DATA
DIMENSION CNCT150),CAREA(50),CPOIN(50),XRR(50),YRR(50)
1,ZRR(50),OBS02(50)
DIMENSION DISTX(1000),XDCAY(2000), SSIGZ(1000),EXPOZ(2000)
1,HA(5),UHA(5),QXY(5),Q(4000)
DIMENSION XP (100),YP (100),ZP (100),QP(100)
DIMENSION DUf'U 100)
COMMON/BASIC/IPRTR,I STAR,I ERR , ISIGD ,NXLIM,NXZLM,NLI.XONE
IfXMAXtCONX,DECAY,I NDEX,THTA,CIGMX ,GX ,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YH,ICARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/OUTPUT/YEAR,AMON,DAY,HOUR ,NRECP,CNCT,CAREA,CPCIN
1,XRR,YRR,ZRR,OBS02,IDOW,NR1,TEMP,SIGA,RIB,PCPN,WGLD,STAPR
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ,NH,HA
1,UHA,QXY,Q,CONC,NH1
COMMON/WNDSP/WSPD,WHGT,PWIND
COMMON/POINTS/NPS ,XP ,YP ,ZP ,QP,CONPS,NPSS
DATA NRECS/0/
CCOMPUTE RECORD INDEX NUMBER
YMDH = HOUR + 100 * DAY
IF (AMON - 12) 1,3,2
1 CONTINUE
YMDH = YMDH + AMON * 10000
3 CONTINUE
C COMPUTE RECORD NUMBER, EXIT IF REQUESTED DATA IS NOT IN FILE
IF (YEAR - 64) 2,4,18
2 CONTINUE
YMDH = HOUR + 100 * DAY + 10000 * AMON + 1000000.* YEAR
WRITE (IPRTR,100) YMDH
100 FORMAT (' YMDH =',F9.0,«, DATA PERIOD REQUESTED IS NOT'
1,' IN FILE1)
c CALL EXIT
t ' C CHECK REQUEST AGAINST DECEMBER 1964 DATES 01/1500 TO 31/2400
4 CONTINUE
NREC = -14
IF (AMON - 12) 2,6,2
v 6 CONTINUE
IF (DAY - 1) 2,8,10
8 CONTINUE
IF (HOUR - 15) 2,10,10
10 CONTINUE
IF (DAY - 31) 12,12,2
12 CONTINUE
IF (HOUR) 2,2,14
14 CONTINUE
IF (HOUR - 24) 16,16,2
16 CONTINUE
NREC = NREC + 24 * (DAY - 1) + HOUR
GO TO 30
18 CONTINUE
IF (YEAR - 65) 2,20,2
20 CONTINUE
IF (AMON - 1) 2,22,24
C CHECK REQUEST AGAINST JANUARY 1965 DATES 01/0100 TO 31/2400
22 CONTINUE
NREC = 730
IF (DAY) 2,2,23
23 CONTINUE
GO TO 10
f 24 CONTINUE
IF (AMON - 2) 2,26,2
C CHECK REQUEST AGAINST FEBRUARY 1965 DATES 01/0100 TO 28/1400
26 CONTINUE
r NREC = 1474
V- IF (DAY) 2,2,27
27 CONTINUE
IF (DAY - 28) 12,28,2
c
c
-------
28 CONTINUE
IF (HOUR - 14) 12,12,2
30 CONTINUE
READ DATA FROM DISK
IF (NRECS) 2,31,32
31 CONTINUE
IF (NREC - 1556) 32,32,131
131 CONTINUE
WORKING ON SECOND TAPE REEL,SPACE UNIT 15 DOWN 1556 RECORDS
NRECS = 1556
DO 132 1=1,1556
READ (15)
132 CONTINUE
32 CONTINUE
IF (NREC - NRECS - 1) 36,33,38
33 CONTINUE
READ ( 15 ) XNDX,YRF1,AMFI,DAFI,I DOW,HRFI,NR1,NRECP
CHECK DISK INDEX NUMBER AGAINST COMPUTED RECORD INDEX NUMBER
IF (XNDX - YMDH) 35,40,35
35 CONTINUE
WRITE( IPRTR,101 ) XNDX,YMDH
101 FORMAT (' XNDX =',F7.0,', YMDH =',F7.0,', DISK AND'
I,1 COMPUTED RECORD INDEX NUMBERS DO NOT AGREE')
CALL EXIT
36 CONTINUE
NRECC = NRECS - NREC + 1
DO 37 1=1,NRECC
BACKSPACE 15
BACKSPACE 16
37 CONTINUE
GO TO 33
38 CONTINUE
NRECC = NREC - NRECS - 1
DO 39 1=1,NRECC
READ ( 15 )
READ ( 16 )
39 CONTINUE
GO TO 33
40 CONTINUE
BACKSPACE 15
READ ( 15 ) XNDX,YRFI,AMFI,DAFI,IDCW,HRFI,NR1,NRECP,(XRR(I )
It I = lt NR1) t (YRR( I ) , I = i,NRl) , ( ZRR ( I ) ,I = 1,NR1) , (CJBS02( I )
2,I=1,NR1),(DUMKI),1=1,NRECP), GX ,GY,DLTA,NH,(HA(I),I=1,NH)
3,NPS,(XP(I),I=1,NPS),(YP(I),I=1,NPS),(ZP(I),I=1,NPS)
4,(QP(I),I=1,NPS),WSPDtWHGT.PWINDtTHTA,INDEX,TEMP,CIGMX
5,SIGA,RIB,PCPN,WGLD,STAPR
READ DATA FROM TAPE
NO = GX * GY * NH
READ ( 16 ) XNDH,GX,GY,NH,(Q (I),1=1,NO)
CHECK TAPE INDEX NUMBER AGAINST COMPUTED RECORD INDEX NUMBER
IF (XNDH - YMDH) 55,60,55
55 CONTINUE
WRITE( IPRTR,102) XNDH,YMDH
102 FORMAT (' XNDH =',F7.0,', YMDH =',F7.0,', TAPE AND'
1,' COMPUTED RECORD INDEX NUMBERS DO NOT AGREE1)
CALL EXIT
60 CONTINUE
NRECS = NREC
RETURN
END
o
C
o
-------
ASSUMING DATA FILE STARTS CC
YRR(50)
SUBROUTINE CHIDA
C ROUTINE TO GET CHICAGO DATA MODEL iNPUTSt
C ON JAN 1 67 AND ENDS 2300 JAN 31 67.
DIMENSION CNCT(50 ) ,CAREA(50),CPOIN(50),XRR(50)
1,ZRR(50)tOBSU2(50)
DIMENSION DISTXt1000),XDCAY(2000),SSIGZ<1000),EXPOZ(2000)
1,HA(5),UHA(5),QXY(5), QSIJ(4000)
DIMENSION XP (100),YP (100),ZP ( 100),QP(100)
COMMON/BASIC/IPRTR,I STAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGKX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/OUTPUT/YEAR,AMON,DAY,HOUR ,NRECP,CNCT,CAREA,CPGIN
1,XRR,YRR,ZRR,OBS02,I DOW,NR1,TEMP,SIGA,RIB,PCPN,WGLD,STAPR
COMMON/AREAS/NXI,NX,KUT6X,DISTX,XDCAY,SSIGZ,EXPOZ,NH,HA
1,UHA,QXY,Q,CONC,NH1
COMMON/WNDSP/WSPD,WHGT,PWIND
COMMON/POINTS/NPS ,XP ,YP ,ZP ,QP,CONPS,NPSS
DIMENSION RES(600),COM(600),XIND(600)
DIMENSION RESK 150),COM1(250)
EQUIVALENCE (RES(468),RES1(1)),
1*0. ,5.70E5,6.
48E6,4.80E5,9.
90E6,8.70E5,1
17E6,1.50E6,1
60E5,4.00E5,3
01E6,5.00E4,8
40E5,3.30E5,1
,50E5,3
tl
,1
10E5, 1
DATA RES /
,1.00E4,12*0,
, 1
,1
,7,
10*0
, 1
,2
,3
,5
,5
,7
,6
10E5,12*0. ,1,
20E5.3.28E6,1,
40E5,1.50E6f6,
,9.00E5,1,
,OOE5f11*0,
70E5,
10*0,
8.00E4,
90E5,6.70E5,2.40E5,12*0
40E5,6.20E5,4.00E5,13*0
70E5.2.30E5,11*0. ,
60E5,14*0. ,9.00E4,2*0. ,4
40E5, 1.10E5,7.00E4,1.40E5.4
,2*5.E4,12*0. , 1.00E5,1.21E6,2
OOE4,11*0. , 1.80E5, 1.85E6,!
70E5,3.00E4,O.OOEO,1.00E4,9*0,
40E5,7.60E5,3.00E4,1.
04E6,5.40E5,4.20E5,7,
29E6, 1.20E5,1.06E6,6.
80E5, 1. 10E5,3.50E5,7.
(COM(392),COM1(1))
80E5,18*0. ,1.22E6,9
OOE5,1.60E5,4.00E4,0
20E5,2*8.E4,3
40E5,1.90E5,7
40E5,5
12E6,3
40E5,7
OOE4
,3,
,1
,08E6,9
,05E6,4
,80E5
, 10E5
, 10E5,6
,70E5,3
,30E5,1
,60E5
,OOE4
,OOE4
,OOE4
,5,
,8,
,60E5,4.
1
80Eb,
OOE4,1
10E5,1
,9
,2
, 1
, 1
, 1
,2
07E6,4.
22E6,1.
14E6, 1.
80E5,2.
DATA RES1/
1.QOE5.9.00E4, 13*0
,5.00E4,3.00E4,14*0
14*0. ,4.00E4,2
14*0. ,5,
19E6,1.80E5,1,
46E6,7.50E5,7,
1.01E6,2,
50E5,1
50E5.5
40E5,3
20E5,1
90E5,5
40E5,1
40E5,4
10E5,12*0.
60E5,2.64E6,
10E5.1
,2
,1
,60E5,
,OOEO,
,OOE4,
,OOE4,
,OOE5,
,70E5,
,10E5,
, 15E6,
,40E5,
,92E6,
,90E5,
90E5
68E6
OOE5
OOE4,10*0
OOE4,2.00E4,11*0
20E5,12*0. ,5.00E4,3.
OOE4,12*0.,3.E4,5.E4,
OOE5
30E5,
02E6,
14E6,
90E5,
60E5,
2*0.,
4.E5,4*0.
8.E4
12*0.
12*0.
l.E4,5.E*
5E5
6E5
3E5
5E5
17E6
16E6
9E5
E4,2*0.
E4,l.Ef
4E5
88E6
3E5
9E5
E4
6.E4/
DATA COM
,2.46E6,1.51E6,5
,9
,2
, 1
,1
,4
,1
, 1
, 30E5.6
,88E6, 1
IOOE4,
,OOE5,
,70E5,3
.13E6,6
, 10E6,6
,20E6,2
11*0,
10E5,4
C2E6,9
9*0. ,6
2*0. ,6
30E5,2
30E5, 1
20E5,9
22E6,1
40E6,2
21E6,11*0,
13E6,9,
,1
, 1
,6. 10E5,13*0. ,1
,12*0. ,1.72E6,6
12*0 - --- •
i
,1
20E5,11*0
19E6,1
60E5,1
80E6.6
40E5,2
. 70E5,1
,4.40E5,3.40E5,3
,7*0. ,5.00E4,-1
DATA COM1/
9.00E4,8.00E4,1.40E5,9,
, 1.70E5,6.00E5,7.70E5,2,
, 1.60E5,8.50E5,6.20E5,5,
1.79E6f2
,20E5,10*0. ,9
70E5,2.00E5,1
, 1.80E5, 1
,3.30E5,1
OOE4,56*0./
/48*0.,9.70E5,1
,50E5,3*0. ,5
10E5,3.20E5,!
, 10E5,8.10E5,2
,70E5,3.90E6,1
1OOE4,9*0. ,2
70E5,1.10E5,5
,02E6,7.20E5,3
,60E5,7.20E5
,16E6,7.40E5,5
,86E6,1.81E6,9
84E6,3.15E6,2
" 03E6,1
,47E6,5
92E6,1
30E5,1
,39E6,2
,20E5,3
80E5,1.30E5,1
60E6,1
OOE4,2*0,
1.80E2f2.00E5,
30E5,1.00E4,2*3.E4,16*0,
E5,2.Ef
2E5
,41E6,1
,OOE4,7
.10E5, 1
, 10E5.1
,66E6,7
,06E6
,OOE4
160E5,2,
,9.50E5,4,
,80E5
,80E5
,94E6,2
02G6,1
50E5,5
,6
,1
,2
,5,
,1
,50E5,
,OOE4
,OOE5
,90E5
,40E5,1
,80E6,6
,86E6,6,
50E5.4.
60E5,2,
84E6
27E5,9
50E6,4
OOE4,8*0. ,2
30E5,2.20E5,2
90E5,8.10E5,5
OOE4,9*0
30E5.2
OOE5,4
OOE5,8
40E6,9
24E6,1
45E6,5
30E5,9
60E5.6
30E5,2
10E5,1
5*0.
8.00E4
OOE5
30E5
13E6.7
40E5,4
»
,2
,1.30E5,
,11*0
,1
,7
,70E5,
,OOE4,
,70E5,
,70E5,
,33E6,
,2
90E5,11*0
20E5,5.20E5,
70E5,4.50E5,
70E5,1
08E6, 1
60E5,9
40E5,3
60E5.4
50E5,5
69E6,6
40E5,2
10E5,5
10E5,2
,27E6,
,26E6,
,70E5,
,60E5,
,20E5,
,60E5,
,OOE5,
,48E6,
,OOE4,
,80E5,
11*0.
2.29E6
11*0.
4E5
15E6
5E5
15E6
12E6
10*0.
2.4E5
44E6
27E6
6E5
8E5
9E5
1E5
6E5
93E6
1E5
OOE4,7*0. ,1,
10E5,2.00E5,5.
60E5,4.10E5,3.70E5,5,
72E6,1.39E6,
50E5,9*0. ,
80E5,6.00E4,
3.1E5/
7.5E5
1.E4
-------
10
20
fjV*wWh.vsr«'*v/^SL~ i j * *^ • ^s w 7
9,' 3 . OOE4 I 0 I OOE01 2 I 24E6 ,'
DATA NRFf.S/f)/
J -«• w V
U W I r\ i^iv^^^^f
IF (NRECS)
CONTINUE
REWIND 19
CONTINUE
YMDH = HOUR + l.u
C NREC SET BY COMPUTING
C 0000 JAN 1 1967 PLUS 1
NREC = (DAY - 1.) *
C ORIENT DATA FILE READER
C NRECS = RECORD NUMBER OF
NRECC = NREC - NRECS - 1
IF (NRECC) 80,300,100
80 CONTINUE
DO 90 1 = 1,NRECC
BACKSPACE 19
90 CONTINUE
GO TO 300
100 CONTINUE
DO 110 1=1,NRECC
READ ( 19 )
110 CONTINUE
300 CONTINUE
C READ DATA FROM DISK
READ ( 19 i
1,GX,GY,DLTA,NH,(HA(
E2 * DAY + 1.E4 * AMON
IG NUMBER OF HOURS BY WHICH REQUESTED PERIOD EXCEEC
1
24. + HOUR + 1. +
AT PROPER RECORD
LAST RECORD READ
)
XNDX,YMDH
" " ', YMDH
C
CHECK DATA INDEX .-.U.-.^L.^ ~^~i,-,^,
IF (XNDX - YMDH) 340,350,340
340 CONTINUE
WRITE (IPRTRflOl) XNDX,
101 FORMAT (' XNDX =',F7.0, . ...„
1 INDEX NUMBERS DO NOT AGREE1)
REWIND 19
CALL EXIT
350 CONTINUE
NR1 = NRECP
NRECS = NREC
CONVERT FT TO M IN HA ARRAY
DO 355 I=1,NH
XNDX,CY,CM,CD,1DOW,CH,NRECP,(OBS(I),1=1,NRECP)
(I),1=1,NH),NRPNT
, (QP( I ) ,I=1,NRPNT),WS,WH,P,WD,INDEX,TEMP,CIGMX
D.STAPR
R AnATN<;T rnMPiiTpn TNinpy MIIMRPR
,,,,,,,,
2, (ZP(I ) , I = 1,NRPNT),(QP(I),I=1,NRPNT),WS,WH,P,WD,I
3,SIGA,RIB,PCPN,WGLD,STAPR
CK DATA INDEX NUMBER AGAINST COMPUTED INDEX NUMBER
IF (XNDX - YMDH) 340,350,340
= ' ,F7.0
-------
HA{ I ) = 0.3048 * HA( I )
355 CONTINUE
I HOUR = HUUR
IF (IHOUR) 360t360,370
360 CONTINUE
IHOUR = 24
370 CONTINUE
CALL DOWCHl IDOW, ID)
NG = GX * GY
DO 380 1=1, NG
II = NG + 1 - I
12 = 11+ NG
13 = 12 + NG
CALL ASHE(TEMP, ID , I HOUR, RES (I ) ,COM( I),XIND(I),QSIJ(I1),QSIJ(I3)
1,QSIJ( 12) )
380 CONTINUE
RETURN
END
SUBROUTINE DOWCH( I DOW , ID 1
C ROUTINE TO CLASSIFY DAYS OF WEEK
C DAY INPUT(IDOW) OUTPUT(ID)
C MON 1 2
C TUE 22
C WED 3 2
C THU 4 2
C FRI 52
C SAT 6 3
C SUN OR HOL 7 TO 17 1
IF ( IOOW - 6) 20,30,10
10 CONTINUE
ID = 1
GO TO 40
20 CONTINUE
ID = 2
GO TO 40
30 CONTINUE
ID = 3
40 CONTINUE
RETURN
END
SUBROUTINE PRAMB
COMMON/ B AS I C/ I PR TR, I STAR, I ERR , I S I GD ,NXLI M , NXZLM , NLI , XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
C NLI = MAXIMUM NUMBER OF LOG INCREMENTS IN INTEGRATION
C XONE = CLOSEST SOURCE DISTANCE USED (METERS)
C CONX= LOGARITHMIC INCREMENT FOR INTEGRATION VARIABLES
C ISIGD = INDICATES OPTION FOR DEFINING DIFFUSION PARAMETERS
C 1 = MCELROY-POOLER PARAMETERS USING TURNER STAB.
C 2 = MCELROY-POOLER PARAMETERS USING RICHARDSON NC.
C 3 = MCELROY-POOLER PARAMETERS USING BROOKHAVEN STA6
C 4 = PASQUILL PARAMETERS USING TURNER STABILITY CAT.
C DECAY = DECAY CONSTANT (PER SEC)
ICARD = 5
IPRTR = 6
NLI = 20
XONE = 50.
XMAX = 5.5E4
RN = 1. / (NLI - 1)
CONX = (XMAX / XONE)**RN
ISIGD = 4
DECAY=0.
RETURN
END
O
0
-------
SUBROUTINE ASHEdEMP, ID , I H, RES ,COM, XI NO ,OR , QC ,QI )
ASHE CALCULATES HOURLY EMISSION RATE FOR AREA SOURCES
C INPUTS TEMP - AVERAGE TEMPERATURE
C ID - DAY OF WEEK INDEX ( 1 -
C IH - HOUR OF DAY ( 1 - 24 )
C RES - RESIDENTIAL EMISSION RATE
C COM - COMMERCIAL EMISSION RATE
C XIND - INDUSTRAIL EMISSION RATE
DIMENSION TFR(24,3),TFCI (24,3)
DIMENSION TFR2(24) ,TFCI2(24)
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
C CON1
EQUIVALENCE
DATA TFR/ 8
-5
1
11
-7
-3
DATA TFR2/
10
-6
-5
DATA TFCI/16
6
6
13
-13
-4
DATA TFCI2/
17
-7
-0
(TFR2( 1),TFR( 1,3) )
• lli
.60,
.50,
.11,
.61,
.17,
.08,
.30,
.28,
.87,
.01,
.25,
.32,
.91,
.37,
.03,
.14,
.66,
= ANNUAL DEGREE
9.07,
-7.61,
-1.43,
10.61,
-8.85,
-2.41,
11.97,
-8.01,
-3.82,
17.82,
4.82,
8.70,
13.23,
-12.94,
0.56,
18.19,
-6.74,
2.60,
DAYS *
9. 12,
-8.72,
-0.41,
9.69,
-8.44,
-0.77,
9.69,
-7.26,
-1.73,
18.43,
2.64,
9.92,
12.54,
-12.43,
4.55,
16.30,
-7.00,
5.97,
24. =
, (TFCI2
8.15,
-7.84,
-0.61,
8.54,
-7.46,
-0.01,
8.43,
-9.34,
-0.86,
16.90,
1.38,
10.10,
10.43,
-12.53,
6.62,
14.55,
-6.78,
7.92,
6155. *
3 ,1=H,2=W
(YEARLY)
(YEARLY)
(YEARLY)
(1) ,TFC
6.64,
-5.55,
-1.49,
7.08,
-6.73,
2.56,
6.65,
-8.28,
2.31,
15.15,
0.30,
10.47,
5.64,
-12.39,
9.08,
10.17,
-7.11,
8.90,
24.
I (1
4
-5
-0
3
-6
3
4
-8
3
12
0
12
-1
-11
10
3
-7
11
,3 = S
,3))
.76,
.87,
.60,
.13,
.25,
.22,
.24,
.07,
.85,
.86,
.55,
.01,
.75,
.19,
.41,
.19,
.01,
.47,
1.83,
-4.09,
1.23,
-2. 15,
-5.11,
5.33,
1.85,
-7.78,
5.71,
9.47,
2.98,
12.23,
-8.04,
-9.62,
11.53,
-0.13,
-3.84,
13.48,
0
-2
4
-7
-4
9
-0
-6
8
8
4
12
-11
-7
13
-4
-1
15
.15,
.86,
.78,
.32,
.08,
.117
.73,
.14,
.747
.63,
.49,
.03,
.69,
.88,
.197
.13,
.84,
.867
C
C
DATA CON1/147720.7
MULTIPLY BY CUNCE TO CONVERT LBS/HR TO GRAMS/SEC
DATA CONCE/0.1260/
QRA= .9*RES
QRC= .1*RES/8760.
QRA = TEMPERATURE DEPENDENT PORTION OF RESIDENTIAL AREA SOURCE EMISSI
QRC = HOT WATER REQUIREMENT
DDR = 65. - (TEMP + TFRUH.IDM
IF (DDR) 4,5,5
4 DDR = 0.
5 CONTINUE
QR= QRC +QRA*DDR/CON1
C QR = ADJUSTED RESIDENTIAL SOURCE EMISSION RATE
QCA = .9 *COM
QCC= .1*COM/8760.
DDC = 65. - (TEMP + TFCI(IH,ID))
IF (DDC) 6,7,7
6 DDC = 0.
7 CONTINUE
QC= QCC-
QCA*DDC/CON1
QI=XIND/8760.
C QR,QI,QC ARE IN UNITS
QR=UR*CONCE
QC=QC*CONCE
QI=QI*CONCE
RETURN
END
OF LBS/HR- CONVERT TO GRAMS/SEC
C
c
-------
SUBROUTINE DISTB
C ROUTINE TO SET DISTANCE DEPENDENT ARRAYS.
DIMENSION DISTX( 1000) ,XDCAY(2000) ,SSIGZ(1000) ,EXPOZ(2000)
1.HA15) ,UHA( 5) , QXY(5) ,0(4000)
COMMON/BASIC/ I PRTR, I STAR, I ERR , I SI GD , NXL I M , NXZLM , NL I , XONE
1,XM AX, CUNX, DECAY, INDEX, THTA,CIGMX,GX,GY, DLTA, IND,SIGY
2,S1GZ,XR,YR,ZR,XS,YS,XW, YW, I CARD, 1C X , XYMI N , XSMAX , YSMAX
COMMON/ ARE AS /NX I ,NX,KUTCX,DISTX,XDCAY,SSIGZ,EXPOZ,NH,HA
1,UHA,QXY,Q,CONC,NH1
COMMON/ WNDSP/WSPDtWHGTfPWI NO
DATA ISTRT/0/
IP (ISTRT) 30,10»30
10 CONTINUE
ISTRT = 1
C SET DISTX ARRAY
DISTX(l) = XONE
DO 20 I=2,NXLIM
J = I - 1
DISTX( I ) = CONX* DISTX( J)
IF (DISTX(I) - DISTX(J) - DLTA) 16,16,14
14 CONTINUE
DISTX( I ) = DISTX( J) + DLTA
16 CONTINUE
IF (DISTX(I) - XMAX) 18,24,24
18 CONTINUE
20 CONTINUE
NMISS = (XMAX - DISTX(NXLIM) )/DLTA + 1.
WRITE (IPRTRtlOOl) NM I SS , XMAX , D I STX ( NXL I M )
1001 FORMAT( 1X18, ' MORE LOCATIONS REQUESTED FOR DISTX ARRAY, XMAX ='
1 E10.3,1, DISTX(LAST)=« ,E10.3)
CALL EXIT
24 CONTINUE
NLI = I
30 CONTINUE
C SET ARRAYS WHICH DEPEND ON DISTANCE AND METEOROLOGICAL CONDITIONS
C GET AREA SOURCE WIND SPEEDS FOR EACH EMISSION HEIGHT
DO 40 IH=1,NH
UHA(IH) =WSPD* (HA( I H ) /WHGT ) **PWI NO
40 CONTINUE
C CHECK THAT DIMENSION LIMIT NXZLM IS NOT EXCEEDED
NHLI=NH*NLI
IF (NHLI-NXZLM) 26,26,25
25 CONTINUE
WRITE ( IPRTR, 1004)NXZLM,NHLI
1004 FORMAT!1 NH * NLI EXCEEDS DIMENSION LIMIT OF',16,', NH,='
16,', NLI=',I6)
CALL EXIT
26 CONTINUE
C GET TRAVEL DISTANCE DECAY FACTORS
IK = 0
DO 60 1 = 1, NLI
IDCAY = 0
XI = DISTX( I )
DO 50 IH=lfNH
IK - IK + 1
IF (DECAY) 42,42,44
42 CONTINUE
XDCAY(IK) = 1.
GO TO 46
44 CONTINUE
XDCAY(IK) = EXP(-DECAY * XI / UHA(IH))
IF (XDCAY(IK) - l.E-6 ) 45,46,46
45 CONTINUE
IDCAY = IDCAY +• 1
IF (IDCAY - NH) 60,55,55
55 CONTINUE
NXI = I
GO TO 62
-------
46 CONTINUE
50 CONTINUE
60 CONTINUE
NXI = NLI
62 CONTINUE
GET SIGMAZ PARAMETERS
DO 130 1=1,NXI
XW = DISTXlI)
CALL SIGZZ
IF (IERR) 67,68,67
67 CONTINUE
RETURN
68 CONTINUE
SSIGZ(I) = SIGZ
IF(SIGZ-CIGMX)70,140,140
70 CONTINUE
130 CONTINUE
KUTEX = NXI
GO TO 160
140 CONTINUE
KUTEX = I
DO 150 J=I,NXI
SSIGZ(J) = C1GMX
150 CONTINUE
160 CONTINUE
RETURN
END
SUBROUTINE CONTR
C ROUTINE TO COMPUTE CONCENTRATION AT RECEPTOR XR,YR,ZR FROM AREA SOURCE
C GIVEN BY ARRAY Q WITH DIMENSIONS GX,GY,NH.
C THE COMPUTATION IS MADE BY INTEGRATING THE EFFECTS FROM EACH HGT. AND
C SUMMING.
DIMENSION DISTX{1000),XDCAY(2000),SSIGZ(1000),EXPOZ(2000)
1,HA(5) ,UHA(5),OXY(5),QC4000)
DIMENSION SXII5),
TERMA(5)
CUMMON/BASIC/IPRTRtISTARtI ERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ,NH,HA
liUHAfQXYtQtCONCtNHl
C CONST = 1/(2*SQRT(2*PI))
DATA CONST/0.199471/
C NHLIM IS DIMENSION LIMIT FOR SXI AND ' TERMA ARRAYS.
DATA NHLIM/5/
Yk = 0.
IF (NH - NHLIM) 3,3,2
2 CONTINUE
WRITE(IPRTR,1000) NHLIM,NH
1000 FORMAT (' NH EXCEEDS DIMENSION LIMIT OF',16,', NH =',16,', NH SET
1TO LIMIT' )
NH1= NHLIM
GO TO 4
3 CONTINUE
NH1 = NH
4 CONTINUE
DO 10 IH=1,NH1
SXK IH) = 0
TERMA(IH) = 0
10 CONTINUE
f XWL = DISTX(l) '
^ DO 110 1=1,NX
XK = DISTX{I)
CALL SCORD
-------
IF (IND) 120,50,120
50 CONTINUE
CALL RATE
IF (I - 1) 60,60,80
60 CONTINUE
DO 70 IH=1,NH1
TERMA(IH) = QXY(IH) * EXPOZ(IH)
70 CONTINUE
GO TO 100
80 CONTINUE
DO 90 IH=l,NHl
K = { I -1) * NH + IH
TERMB = QXY(IH) * EXPOZ(K)
SXKIH) = SXK1H) + (TERMA(IH) + TERMB) * (XW - XWL )
TERMA(IH) = TERMB
90 CONTINUE
100 CONTINUE
XWL = XW
110 CONTINUE
II = NX + 1
IF (II - NLI) 115,115,140
115 CONTINUE
XW = DISTX(II)
120 CONTINUE
DO 130 IH=1,NH1
SXKIH) = SXI(IH) + TERMA(IH) * ( XW - XWL)
130 CONTINUE
140 CONTINUE
CONC = 0
DO 150 IH=1,NH1
CONC = CONC + SXI(IH) / UHA(IH)
150 CONTINUE
CONC = CONST * CONC
RETURN
END
SUBROUTINE RATE
C A SUBROUTINE TO INTERPOLATE EMISSION RATE OF A POINT INTERMEDIATE TO
C POINTS ON A STANDARD GRID SYSTEM
DIMENSION DISTXl1000),XDCAY(2000),SSIGZ(1000),EXPOZ(2000)
1,HA(5),UHA(5),OXY(5),Q(4000)
COMMON/BASIC/IPRTR,I STAR,I ERR,I SIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/AREAS/NXI,NX,KUTEX,DISTXfXDCAYtSSIGZ,EXPOZ»NHtHA
1,UHA,OXY,Q,CONC,NH1
C INITIALIZE INTEGER CONSTANTS
IGRD = GX
JGRD = GY
NG = IGRD * JGRD
X = XS/ DLTA
IX = X
Y = YS/ DLTA
IY = Y
C CHECK IF POINT IS ON OUTSIDE FRINGE OF GRID, I.E. WITHIN 0.5 GRIDS CF
C EDGE. IF POINT IS IN FRINGE CORNER, USE CORNER GRID VALUES. CTHER
C POINTS, LINEARLY INTERPOLATE BETWEEN EDGE GRID POINTS.
IF (X - IGRD) 10,1,1
1 CONTINUE
IF (Y - JGRD) 5,2,2
2 CONTINUE
K = 0
3 CONTINUE
C USE CORNER VALUE.
-------
K = K + NG
OXY(I) = Q(K)
4 CONTINUE
RETURN
5 CONTINUE
IF (IY) 6,6,7
6 CONTINUE
K = IGRD - NG
GO TO 3
7 CONTINUE
Kl = IY * IGRD - NG
K2 = Kl + IGRD
DK = Y - IY
8 CONTINUE
C USE LINEAR INTERPOLATION ON EDGE
DO 9 1=1,NH1
Kl = Kl + NG
K2 = K2 + NG
QXY(I) = Q(K1) + DK * (Q(K2) - Q(KD)
, 9 CONTINUE
RETURN
10 CONTINUE
IF (IX) 11, 11,16
11 CONTINUE
1 IF (Y - JGRD) 13,12,12
12 CONTINUE
K = 1 - IGRD
f GO TO 3
C 13 CONTINUE
IF (IY) 14,14,15
14 CONTINUE
,- K = 1 - NG
V> GO TO 3
15 CONTINUE
K2 = IY * IGRD + 1 - NG
Kl = K2 - IGRD
v DK = Y - IY
GO TO 8
16 CONTINUE
IF (Y -JGRD) 18,17,17
17 CONTINUE
Kl = IX - IGRD
K2 = Kl + 1
DK = X - IX
GO TO 8
18 CONTINUE
IF (IY) 19,19,20
19 CONTINUE
Kl = IX - NG
K2 = Kl + 1
DK = X - IX
GO TO 8
20 CONTINUE
C DETERMINE WHICH TRIANGLE OF GRID POINTS WILL BE USED FOR INTERPOLATE
BX = X - IX
BY = Y - IY
IF (BX - BY) 200,100,100
100 CONTINUE
Kl = IX + (IY - 1) * IGRD - NG
r- DO 150 1 = 1,NH1
•-> Kl = Kl + NG
K2 = K1+ 1
K4 = K2 + IGRD
r QXY(I)=Q(K1) > BX * ( Q(K2 )-C(Kl ))+BY*(Q(K4)- Q(K2))
v- 150 CONTINUE
RETURN
200 CONTINUE
r Kl = IX + (IY - 1) * IGRD - NG
\^
-------
DO 300 1=1,NH1
Kl = Kl + NG
K3 = Kl + IGRD
KA = K3 + 1
QXY(I)=Q(K1) + BY * (Q(K3 )-Q(Kl ))+BX*(Q(K4) - Q(K3))
300 CONTINUE
RETURN
END
SUBROUTINE EXPZB
C ROUTINE TO COMPUTE VERTICAL DIFFUSION FACTOR INCLUDING EFFECTS OF
C DECAY AND GROUND REFLECTIONS FOR EACH OF NH SOURCE HEIGHTS.
C BASIC EQUATION IS
C EXPUZ = ( XDCAY / SIGZ ) *(EXP(-0.5*((HA-ZR)/SIGZ)**2) +
C EXP(-0.5*( (HA+ZR)/SIGZ)**2) )
DIMENSION DISTXl1000),XDCAY(2000),SSIGZ(1000),EXPOZ(2000)
1,HA(5),UHA{5),QXY(5),0(4000)
COMMON/BASIC/IPRTR,I STAR,I ERR,I SI GO,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX , XYMIN,XSKAX,YSMAX
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ,NH,HA
1,UHA,QXY,Q,CONC,NH1
C INPUTS
C ZR = RECEPTOR HEIGHT
C NH = NUMBER OF SOURCE HEIGHTS
C HA = ARRAY OF SOURCE HEIGHTS
C CIGMX = MIXING CEILING
C SSIGZ = VERTICAL DIFFUSION PARAMETER
C OUTPUT
C EXPOZ = VERTICAL DIFFUSION FACTOR
K = 0
DO 60 J=1,NXI
G = 2. *CIGMX/SSIGZ(J)
DO 50 1=1,NH
F = (HA(I) - ZR) /SSIGZ(J)
IF (F*F - 50.) 12,12,10
10 CONTINUE
El = 0
11 CONTINUE
E2 = 0
GO TO 15
12 CONTINUE
El = EXP(-0.5 * F * F)
F = (HA(I) + ZR) /SSIGZtJ)
IF (F- 7.) 13,13,11
13 CONTINUE
E2 = EXP(-0.5 * F * F)
15 CONTINUE
K = K + 1
EXPOZ(K) ={(E1 + E2) /SSIGZ(J)) * XDCAY(K)
r- 50 CONTINUE
lv 60 CONTINUE
NX = NXI
RETURN
C END
C
-------
C
SUBROUTINE INCRT
C THIS SUbROUTINE INCREMENTS THE TIME AT CONCENTRATIONS ARE CALC. BY THE
C DIFFUSION PROGRAM. IN ADDITION TO THIS IT GENERATES THE TIME INDEX
C WHICH IS USED TO GENERATE THE SOURCE MATRIX AND METEOROLOGICAL INPUTS
C CORRESPONDING TO THE TIME AT WHICH THE CONCENTRATION IS REQUIRED.
DIMENSION YEARR(2) ,AMONN(2),DAYY(2) ,HOURR{2)
COMMON/DAT I ME /DH,STOP,YEARR,AMONN,DAYY,HOURR,HOURL
DIMENSION DIM(12)
C DIM IS AN ARRAY REPRESENTING THE NUMBER OF DAYS IN EACH MONTH
DATA DIM/31.,28.,31.,30.,31.,30.,31.,31.,30.,31.,30.,31./
C THE ARRAYS YEARR, AMONN, DAYY, HOURR CONTAIN THE YEAR (TWO DIGITS),MO.
C DAY, AND HOUR OF THE FIRST AND LAST CALCULATION (INDEX=1 AND INDEX=2
C RESPECTIVELY)
C CHECK FOR LEAP YEAR
TEST= YEARR(l) / 4.
ITEST=TEST
IF(TEST-ITEST) 200,100,200
100 CONTINUE
C YEAR( 1) IS A LEAP YEAR
DIM(2)=29.
GO TO 220
200 CONTINUE
DIM(2) = 28.
220 CONTINUE
C INCREMENT HOUR
HOURR(l) = HOURR(l) + DH
C CHECK TO SEE IF HOUR(1) IS IN THE SAME DAY
IF(HOURR(D-HOURL) 600,600,300
300 CONTINUE
C HOUR(l) IS NOT IN THE SAME DAY
HOURR(l) = HOURR(l) - 24.
DAYY( 1) = DAYY( 1 ) + 1.
C CHECK TO SEE IF DAY(l) IS IN THE SAME MONTH
MONTH=AMONN(1)
IF (DAYY(l) - DIM(MONTH)) 600,600,400
400 CONTINUE
C DAY(l) IS NOT IN THE SAME MONTH
AMONNt1) = AMONN(1) + 1.
DAYY(l) = DAYY(l) - DIM(MONTH)
C CHECK TO SEE IF MONTH(l) IS IN SAME YEAR
IF(AMONN(1 )-12.) 600,600,500
500 CONTINUE
C MONTH(l) IS NOT IN THE SAME YEAR
AMONN(l) = AMONN(l) - 12.
YEARRt1) = YEARR( 1) + 1.
600 CONTINUE
C CHECK TO SEE IF THIS IS LAST INCREMENT
IF (YEARR(l) - YEARR12)) 1100,700,1000
700 CONTINUE
C YEARS ARE THE SAME
IF(AMONN(1)-AMONN(2)) 1100,800,1000
800 CONTINUE
C MONTHS ARE THE SAME
IF (DAYY(l) - DAYY12)) 1100,900,1000
900 CONTINUE
C DAYS ARE THE SAME
IF (HOURR(i) - HOURR(2M 1100,1100,1000
1000 CONTINUE
C STOP INCREMENTING
STOP=1.
RETURN
1100 CONTINUE
STOP=0.
RETURN
END
-------
Exhibit D-2
LISTING OF FORTRAN CODE COMPUTER PROGRAM
AND SUBROUTINES USED FOR SENSITIVITY CALCULATIONS*
-------
C DIFFUS3 - SENSITIVITY ANALYSIS
DIMENSION DTHTA(IO)
DATA DTHTA/-45.,-10.,-3.,0.,3.,10.,45./
DIMENSION YEARR(2)|AMONN(2),DAYY(?),HOURR(2)
DIMENSION CNCT(50),CARE A(50),CPOIN(50),XRR(50),YRR(50),ZRR(50)
l,OBS02(50),OSET(12,54)
DIMENSION DISTX(200,7),XDCAY(600,7),SSIGZ(200,7),EXPOZ(600,7)
1,HA{5),UHA(5),QXY(5),Q(3600),QX(200,3,1,7),NQX(7,3)
DIMENSION XP (100),YP (100),ZP ( 100),QP(100)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGO,NXLIM,NXZLM,NLI,XONE,XMAX,CCN>
It DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA,IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/DATIME /DH,STOP,YEARR,AMONN,DAYY,HOURR
COMMON/OUTPUT/YEAR.AMON,DAY,HOUR ,NRECP,CNCT,CAREA,CPOIN,XRR,YRR
1,ZRR,OBS02,IDOW,NR1,TEMP,SIGA,RIB,PCPN,WGLD,STAPR,OSET,INDC
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ, NH,HA,UHA
It OXY,0 ,CONC,NH1,QX,NQX
COMMON/WNDSP/WSPD,WHGT,PWIND
COMMON/POINTS/NPS ,XP ,YP ,ZP , QP,CONPS,NPSS
C SET CONTROL PARAMETERS
IERR = 0
INDO = 0
ICARD = 5
, IPRTR = 6
NLI = 20
XMAX = 5.SEA
XONE = 1.
RN = 1. / (NLI - 1 )
CONX = (XMAX / XONE)**RN
C GET BASE EMISSION FIELD, RECEPTOR LOCATIONS AND WIND DIRECTION
CALL SENDA
WRIFElIPRTR,901) ICARD,IPRTR,XONE,XMAX,CONX
901 FORMAT(/' CONTROL PARAMETERS'/4X'1CARD',3X'IPRTR',8X'
1' ,8X'CONX'/1X2I8,3E12.3/)
XYMIN = 0.5 * DLTA
XSMAX = (GX + 0.5) * DLTA
YSMAX = (GY + 0.5) * DLTA
C GET OX ARRAY
20 CONTINUE
THTA1 = THTA
CALL GENQX
IF (IERR) 150,30,150
30 CONTINUE
10 = 0
C CYCLE ON DIFFUSION PARAMETERS
DO 148 18=1,3
GO TO (45,46,47),18
45 CONTINUE
ISIGD = 4
INDEX = 5
GO TO 48
46 CONTINUE
ISIGD = 1
INDEX = 4
GO TO 48
47 CONTINUE
ISIGD = 2
INDEX = 1
48 CONTINUE
C CYCLE ON MIXING CEILING
CIGMX = 20.
DO 147 17=1,2
CIGMX = 5. * CIGMX
,. ISTAR = 0
CALL SIGZZ
ISTAR = 1
IF (IERR) 150,50,150
r 50 CONTINUE
-------
C CYCLE ON WIND PROFILE POWER
PWIND = 0.075
DO 146 16=1,1
PWIND = 2. * PWIND
C CYCLE ON WIND SPEED
WSPD = 2. / 3.
DO 145 15=1,3
WSPD = 3. * WSPD
WRITE (IPRTR.900)
900 FORMAT(//'l INDO',2X'AREA CONC.•,3X'PT. CONC.1
A ,2X'TOT. CONC.',3X'ISIGD1,3X'INDEX',3X'CIGMX',3X
1'PWIND',4X«WSPD',3X'DECAY',3X'NH',4X'THTA',3X'NPS',2X'RECEPTOR'/)
C CYCLE ON DECAY CONSTANT
DO 144 14=1,1
GO TO (51,52,53) ,14
51 CONTINUE
DECAY = 0.
GO TO 54
52 CONTINUE
DECAY = 0.0003851
GO TO 54
53 CONTINUE
DECAY = 0.05
54 CONTINUE
C CYCLE ON DISTRIBUTION OF EMISSION HEIGHTS
DO 143 13=1,1
GO TO (55,56),13
55 CONTINUE
v NH = 1
HA(1) = 30.
GO TO 57
f 56 CONTINUE
<• > NH = 3
HA{1) = 15.
HA(2) = 30.
HA(3) = 45.
5 57 CONTINUE
CALL DISTC
IF (IERR) 150,58,150
58 CONTINUE
CALL EXPZC
C CYCLE ON GRID SPACING
DO 142 12=1,7
THTA = THTA1+ DTHTAU2) * 3.14159 / 180.
ISTAR = 0
CALL SCORD
CALL WCORD
ISTAR = 1
C CYCLE UN NUMBER OF POINT SOURCES
DO 141 11=1,1
GO TO (65,66,67),II
65 CONTINUE
NPSS = 51
GO TO 68
66 CONTINUE
NPSS = 19
GO TO 68
67 CONTINUE
NPSS = 0
68 CONTINUE
C CYCLE ON RECEPTOR LOCATION
DO 140 1=1,3
INDO = INDO + 1
r 10 = 10 + 1
( XR = XRR(I) .
YR = YRR(I)
CALL CONTRS(1,11,12)
£ CONPS =0.
-------
IF (NPS) 100,100,85
85 CONTINUE
CALL POINT
IF (IERR) 90,100,90
90 CONTINUE
WRITE(IPRTR,1002) INDO
1002 FORMATt/1 ERROR RETURN
GO TO 150
100 CONTINUE
10)
10
10
10
10
10
10
10
10
10
10
10
FROM POISN, INDO =',I8/)
1 ,
2,
3,
OSET(
OSET(
OSET(
OSETl
OSET(
OSET(
OSET(
OSET(
OSET(
OSET(10,
OSET( 11,
OSET(12,
8,
9,
= CONC * 1.E6
= CONPS * 1.E6
CONPS)
* 1.E6
=(CONC
= NPS
= THTA
= NH
= DECAY
= WSPD
= PWIND
= CIGMX
= ISIGD
= XRR( I )
WRITEl IPRTR, 902) INDO, ( OSET ( N , I 0 ) ,N= 1
1,PWIND,WSPD,DECAY,NH,THTA,NPS,XRR(I )
902 FORMAT! 1XI5,3E12.3,2I8,F8.0,F8.3,F8.1,F8.3,I5,F8
140 CONTINUE
141 CONTINUE
142 CONTINUE
143 CONTINUE
144 CONTINUE
CALL OUTSE
10 = 0
145 CONTINUE
146 CONTINUE
147 CONTINUE
148 CONTINUE
149 CONTINUE
THTA = 0
GO TO 20
150
3) , ISIGD,INDEX,CIGMX
3, I6,F9.0)
35
CONTINUE
END FILE
CALL
END
EXIT
12
c
C
SUBROUTINE GENQX
DIMENSION QA(3),QB(3),QC(3)
DIMENSION QR(4,4)
DIMENSION DTHTA(IO)
DATA DTHTA/-45.,-10.,-3.,0.,3.,10.,45./
DIMENSION CNCT(SO),CAREA(50),CPOIN(50),XRR(50),YRR(50),ZRR(50)
1,OBS02(50),OSET(12,54)
DIMENSION DISTX(200,7),XDCAY{600,7),SSIGZ(200,7),EXPOZ(600,7)
1,HA(5),UHA(5),QXY(5),0(3600),QX(200,3,1,7),NOX(7,3)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE,XMAX,CON)
1,DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA,IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YK,ICARD,ICX,XYMIN,XSMAX,YSN'.AX
COMMON/OUTPUT/YEAR,AMON,DAY,HOUR ,NRECP,CNCT,CAREA,CPCIN,XRR,YRR
1,ZRR,OBS02,IDOW,NR1,TEMP,SIGA,RIB,PCPN,WGLD,STAPR,OSET,INDO
COMMON/AREAS/NX I,NX,KUTEX,D1STX,XDCAY,SSIGZ,EXPOZ, NH,HA,UHA
1, QXY,Q ,CONC,NH1,QX,NQX
NXLIM = 200
YW = 0.
THTA1 = THTA
DO 10 J=l,3
DO 10 1=1,7
NQX(I,J) = 0
-------
10 CONTINUE
READ (ICARD,1000) IR1
1000 FORMAT (I 10)
WRITE(IPRTR,1001) IR1
1001 FORMAT(/' INITIAL RANDOM NO. INTEGER =',I10)
IRS = IR1
GX = 120.
GY = 160.
DLTA = 381.
DO 200 IB=1,7
IR1 = IRS
THTA = THTAU DTHTA(IB) * 3.14159 / 180.
ISTAR = 0
CALL SCORD
CALL WCORD
ISTAR = 1
C SET DISTX ARRAY
OISTX(lflB) = XONE
DO 20 I=2,NXLIM
J = I - 1
DISTX(I.IB) = CONX * DISTX(JflB)
IF (DISTXUfIB) - DISTX(J.IB) - DLTA) 16,16,14
14 CONTINUE
, DISTXdtIB) = DISTXU,IB) + DLTA
( 16 CONTINUE
IF ( DISTX(I,IB) - XMAX) 18,24,24
18 CONTINUE
f. 20 CONTINUE
' NMISS = (XMAX - DISTX(NXLIM,IB)) / DLTA + 1.
WRITE (IPRTR,1002) NMISS,XMAX,DISTX(NXLIM , IB)
1002 FORMAT! 1X18,' MORE LOCATIONS REQUESTED FOR DISTX ARRAY, XMAX ='
, 1 E10.3,', DISTX(LAST)=«,E10.3)
(> I6RR = 1
RETURN
24 CONTINUE
NLI = I
V DO 300 IC=lt3
XR = XRR( 1C)
YR = YRR(1C)
KLAST = 0
v DO 100 J=1,NLI
XW = DISTX(J,IB)
CALL SCORD
CALL GCHEKS
IF (IND) 110,40,110
40 CONTINUE
50 CONTINUE
1X1 = XS / 1524. + 0.5
IY1 = YS / 1524. + 0.5
K = ( IY1 - 1) * 30 + 1X1
IF (K - KLAST) 51,52,51
51 CONTINUE
C GENERATE SUB-GRID FOR BLOCK K EMISSIONS
KLAST = K
QM = Q(K)
CALL QAREAt IR1.QM.QR)
IF ( IERR) 360,52,360
52 CONTINUE
IX = XS / DLTA - 1.
f IY = YS / DLTA - 1.
1X2=IX-4*(IX1-1)
IY2 = IY - 4 * (IY1 - 1)
QX(J,1C,1,18) = OR(IX2,IY2)
r 100 CONTINUE
<-; NQXUB.IC) = NLI
GO TO 300
110 CONTINUE
p NQX(IB,1C) = J - 1
-------
1,1) -
1,1) -
300 CONTINUE
200 CUNTINUE
IF (NQX{
310 CONTINUE
IF (NQX(
320 CONTINUE
ICX = 1
GO TO 360
330 CONTINUE
ICX = 3
GO TO 360
340 CONTINUE
IF (NQX(1,2)
350 CONTINUE
ICX = 2
360 CONTINUE
RETURN
END
NQX(1,2)) 340,310,310
NQX(1,3)) 330,320,320
- NQX( 1,3)) 330,350,350
SUBROUTINE SENDA
DIMENSION ICP( 100) ,XPP(100),YPP(100),ZPP(100)
DIMENSION CNCT(50 ) ,CAREA{50),CPCIN(50),XRR(50), YRR(50),ZRR(50)
1,OBS02(50),OSET(12,54)
DIMENSION DISTX(200,7),XDCAY(600,7),SSIGZ(200,7),EXPOZ(600,7)
1,HA(5),UHA(5),QXY(5),Q(3600),QX(200,3,1,7),NQX(7,3)
DIMENSION XP (100),YP (100),ZP ( 100),QP(100)
COMMON/BASIC/IPRTR,I STAR,IERR,ISIGD,NXLIM,NXZLM,NLI, XONE,XMAX,CON>
1,DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA,IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW, I CARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/OUTPUT/YEAR,AMON,DAY,HOUR ,NRECP,CNCT,CAREA,CPCIN,XRR,YRR
l,ZRR,OBS02,IUOW,NR1,TEMP,SIGA,RIB,PCPN,WGLD,STAPR,OSET,INDO
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ, NH,HA,UHA
1, OXY,Q ,CONC,NH1,QX,NQX
COMMON/WNDSP/WSPD,WHGT,PWIND
50
COMMON/POINTS/NPS
XP ,YP ,ZP ,QP,CONPS,NPSS
DLTA,GX,GY,WHGT,THTA,NH,NPS,(XPP(N),YPP(N),ZPP(N)
55
60
70
* DLTA * TAN(THTA)
* DLTA * TAN(THTA)
O
READ (10) YMDH,
1,QP(N),N=1,NPS)
NG = GX * GY
DO 50 1=1,NH
Nl = ( I -1) * NG
READ (10) (Q(N+N1),N=1,NG)
CONTINUE
XRR(2) = 16.5 * DLTA
YRR(2) = 20.5 * DLTA
XRR( 1 ) = XRR( 2) •»• 9,
YRR(1) = 29.5 * DLTA
XRR(3) = XRR(2) - 18,
YRR(3) = 2.5 * DLTA
ZR = 0.
DO 55 1=1,NPS
ICP( I ) = I
CONTINUE
CALL SORTK ICP)
DU 60 1=1,NPS
J = ICP(I )
XP(I) = XPP(J)
YP( I ) = YPP(J)
ZP(I) = ZPP(J)
CONTINUE
AREA = DLTA * DLTA
DO 70 1 = 1,NG
0(1) = Q(I) + Q/U+NG) + Q(I+2*NG)
0(1) = Q(I) / AREA
CONTINUE
RETURN
END
-------
SUBROUTINE SORT1 (ICP)
DIMENSION ICP(l)
DIMENSION CP(100)
DIMENSION XP (100),YP (100),ZP (100) , QP(100)
COMMON/POINTS/NPS ,XP ,YP ,ZP ,OP,CONPS,NPSS
EUUIVALENCE (CP(1),QP(1))
ITEMS = NPS
NP = ITEMS
1 NP = NP/2
IF(NP) 7,7,2
2 K = ITEMS - NP
J = 1
3 I = J
M = I + NP
4 IF (CP( I ) - CP(M)) 6,6,5
5 SAVE = CPU )
CP(I) = CP(M)
CP(M) = SAVE
ISAVE = ICP( I )
ICP(I)= ICP(M)
ICP(M)= ISAVE
M = I
I = I - NP
IF (I - 1) 6,4,A
6 J = J + 1
IF (J - K) 3,3,1
7 RETURN
END
SUBROUTINE QAREA ( I R 1 , QM , QR )
COMMON/ BASIC/ I PR TR, I STAR, I ERR , I SI GD , NXL I M ,NXZLM , NL I , XCNE , XMAX , CCN>
1, DEC AY, INDEX,THTA,CIGMX,GX,GY, DLTA , I ND , S IGY , S IGZ , XR , YR , ZR , XS , YS
2,XW,YW,ICARD, I CX , XYM I N , XSMAX , YSMAX
DIMENSION QR(4,4)
DO 40 J = l,4
DO 30 1=1,4
CALL RANDU( IR1, IR2,RFL)
CALL NDTRI (RFL,QN,QD, IER)
IF (IER) 10,20,10
10 CONTINUE
WRITE( IPRTR, 1000) RFL,IR1,IR2
1000 FURMATt/' RANDOM NO. ERROR, RFL =',F10.5,« ,IR1 =',110,', IR2 ='
1, IIO/)
IERR = IER
RETURN
20 CONTINUE
IR1 = IR2
IF (QN + 2.) 25,25,26
25 CONTINUE
OR( I, J) = 0.
GO TO 30
26 CONTINUE
USE STD. DEV. = 0.5 * MEAN
QR( I, J) = (ON * 0.5 + 1. ) * QM
30 CONTINUE
40 CONTINUE
RETURN
END
C
-------
c
X)
C SUBROUTINE NDTRI
C COMPUTES X = P**(-1)(Y), THE ARGUMENT X SUCH THAT Y=P(X) =
C THE PROBABILITY THAT THE RANDOM VARIABLE U, DISTRIBUTED
C NORMALLYJO,1 ) , IS LESS THAN OR EQUAL TO X. FIX), THE
C ORDINATE OF THE NORMAL DENSITY, AT X, IS ALSO COMPUTED.
C DESCRIPTION OF PARAMETERS
C P - INPUT PROBABILITY.
C X - OUTPUT ARGUMENT SUCH THAT P = Y = THE PROBABILITY THAT
C U, THE RANDOM VARIABLE, IS LESS THAN OR EQUAL TO X.
C D - OUTPUT DENSITY, F(X).
C IER - OUTPUT ERROR CODE
C =-1 IF P IS NOT IN THE INTERVAL (0,1), INCLUSIVE.
C =0 IF THERE IS NO ERROR. SEE REMARKS, BELCW.
C REMARKS
C MAXIMUM ERROR IS 0.00045.
C IF P = 0, X IS SET TO -(10)**74. D IS SET TO 0.
C IF P = 1, X IS SET TO (10)**74. D IS SET TO 0.
C METHOD
C BASED ON APPROXIMATIONS IN C. HASTINGS, APPROXIMATIONS FCR
C DIGITAL COMPUTERS, PRINCETON UNIV. PRESS, PRINCETON, N.J.,
C 1955. SEE EQUATION 26.2.23, HANDBOOK OF MATHEMATICAL
C FUNCTIONS, ABRAMOWITZ AND STEGUN,DOVER PUBLICATIONS, INC.,
C NEW YORK.
SUBROUTINE NDTRI(P,X,D,IE)
IE = 0
I F ( P ) 11 4 , 2
1 IE=-1
GO TO 12
2IF(P-1.0)7,6,1
4 X=-.999999E+38
5 0=0.0
GO TO 12
6 X=.999999E+38
GO TO 5
7 D=P
IF(D-0.5)9,9,8
8 D-l.O-D
9 T2=ALOG(1.0/(D*D))
T=SQRT(T2)
X=T-(2.515517+0.802853*1+0.010328*12)/(I.0+1.432788*1+0.189269*12
1 +0.001308*T*T2)
IF(P-0.5)10,10,11
10 X=-X
11 D=0.3989423*EXP(-X*X/2.0)
12 RETURN
END
SUBROUTINE RANDU(I X,1Y,YFL)
IY=IX*899
IF(IY)5,6,6
5 IY=IY+2147483647+1
6 YFL = IY
YFL=YFL/2147483647.
RETURN
-------
SUBROUTINE RATEA(J,IB,1C,X,Y)
C A SUBROUTINE TO INTERPOLATE EMISSION RATE OF A POINT INTERMEDIATE TO
C POINTS ON A STANDARD GRID SYSTEM
DIMENSION QA(3),QB(3),QC(3)
DIMENSION D1STX(200,3).XDCAY(600,3),SSIGZ(200,3),EXPOZ(600,3)
1,HA(5),UHA(5),QXY(5),Q(3600),QX(200,3,3,3),NQX(3,3)
COMMON/BASIC/IPRTR,I STAR,I ERR,I SI GO,NXLIM,NXZLM.NLI,XONE,XMAX,CONX
1,DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA,IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW,ICARD,ICX,XYKIN,XSMAX,YSKAX
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ, NH,HA,UHA
1, QXY,Q ,CONC,NH1,QX,NQX
C INITIALIZE INTEGER CONSTANTS
1GRD = GX + 0.5
JGRD = GY
NG = IGRD * JGRD
IX = X
IY = Y
C CHECK IF POINT IS ON OUTSIDE FRINGE OF GRID, I.E. WITHIN 0.5 GRIDS OF
C EDGE. IF POINT IS IN FRINGE CORNER, USE CORNER GRID VALUES. OTHER
C POINTS, LINEARLY INTERPOLATE BETWEEN EDGE GRID POINTS.
IF (X - IGRD) 10,1,1
1 CONTINUE
IF (Y - JGRD) 5,2,2
2 CONTINUE
K = NG
3 CONTINUE
, C USE CORNER VALUE.
GO TO (60,31,32),IB
31 CONTINUE
QA(1) = Q(K)
,-. 1X1= X + 0.5
v> XI = 1X1* DLTA
IY1= Y + 0.5
Yl = IY1* DLTA
GO TO 33
v- 32 CONTINUE
1X1= X + 0.5
IY1= Y + 0.5
CALL QCOMB(1X1,IY1.QD)
QA(1) = QD
XI = ( 1X1- 0.375) * DLTA
Yl = ( IY1- 0.375) * DLTA
33 CONTINUE
QA{2) = QA(1)
QA(3) = QA(l)
IQI = 1
CALL ADDPO(IQI,X1,Y1,QA,QB,QC)
DO 35 1=1,3
QX(J, 1C, I,IB) = QA(I)
35 CONTINUE
GO TO 57
5 CONTINUE
IF (IY) 6,6,7
6 CONTINUE
^ K = IGRD
GO TO 3
7 CONTINUE
Kl = IY * IGRD
r K2 = Kl «• IGRD
Dl = Y - IY
IQI = 3
8 CONTINUE
r C USE LINEAR INTERPOLATION ON EDGE
^ GO TO (60,41,42),IB
41 CONTINUE
QA(1) = Q(K1)
-------
IF ( IQI - 3) 47,48,48
47 CONTINUE
XI = IX * DLTA
IY1 = Y * 0.5
Yl = IY1 * DLTA
GO TO 45
48 CONTINUE
1X1 = X + 0.5
XI = 1X1 * DLTA
Yl = IY * DLTA
GO TO 45
42 CONTINUE
IF ( IQI - 3) 43,44,44
43 CONTINUE
IY1 = Y + 0.5
Yl = IY1 * DLTA
XI = IX * DLTA
CALL QCOMB( IX , IY1,QD)
QA(l) = QD
1X1 = IX + 1
CALL QCCIMB(IXl.IYltQD)
QB( I) = QD
GO TO 45
44 CONTINUE
1X1 = X + 0.5
XI = 1X1 * DLTA
Yl = IY * DLTA
f CALL QCOMBl1X1,IY,QD)
<• QA( 1) = QD
IY1 = IY + I
CALL QCOMBt1X1,IY1.QD)
,- QB(1) = QD
V> 45 CONTINUE
QA(2) = QA(1)
QA{3) = QA(1)
, . QB(2) = QB(1)
v QB13) = QB(1)
CALL ADDPOlIQI,X1,Yl,QA,QB,QC)
DO 46 1=1,3
QX{J,1C,It IB) = QA(I) + D1*(QB(I) * QA(I))
v 46 CONTINUE
GO TO 57
10 CONTINUE
IF (IX) 11,11,16
11 CONTINUE
IF (Y - JGRD) 13,12,12
12 CONTINUE
K = 1 - IGRD + NG
GO TO 3
13 CONTINUE
IF (IY) 14,14,15
14 CONTINUE
K = 1
GO TO 3
15 CONTINUE
K2 = IY * IGRD + 1
Kl = K2 - IGRD
01 = Y - IY
IQI = 3
r- GO TO 8
16 CONTINUE
IF (Y -JGRD) 18,17,17
17 CONTINUE
r, Kl = IX - IGRD > NG
f^ K2 = Kl + 1
01 = X - IX
IQI = 2
GO TO 8
-------
18 CONTINUE
IF (IY) 19,19»20
19 CONTINUE
Kl = IX
K2 = Kl + 1
Dl = X - IX
IQI = 2
GO TO 8
20 CONTINUE
DETERMINE WHICH TRIANGLE OF GRID POINTS WILL BE USED FOR INTERPOLATION
Dl = X - IX
D2 = Y - IY
Kl = IX + ( IY - 1) * IGRD
IF (Dl - D2) 200,100,100
100 CONTINUE
K2 = Kl + 1
K3 = K2 + IGRD
IQI = 4
50 CONTINUE
GO TO (60,52, 53) , IB
52 CONTINUE
QA( 1) = Q(K1)
QB(1) = Q(K2)
QC( 1) = Q(K3)
XI = IX * DLTA
Yl = IY * DLTA
GO TO 56
53 CONTINUE
CALL QCOMB( IX, IY,QD)
QA( 1) = QD
1X1 = IX + 1
IY1 = IY + 1
CALL QCOMB( 1X1, IY1,QD)
QC( 1) = QD
XI = ( IX - 0.375) * DLTA
Yl = ( IY - 0.375) * DLTA
IF (IQI - 4) 54,54,55
54 CONTINUE
CALL QCOMB( 1X1, IY,QD)
QB( 1) = QD
GO TO 56
55 CONTINUE
CALL QCOMB( IX, IY1,QD)
QB( 1) = QD
56 CONTINUE
QA(2) = QA( 1
QA(3) = QA( 1
QB(2) = QB( 1
QB(3) = QB( 1
QC(2) = QC( 1
QC(3) = QC( 1
CALL ADDPOt I Q I , XI , Yl , QA, QB , QC )
DO 57 1=1,3
QX( J, 1C, I , IB) = QA(I) + D1*(QB(I) - QA(I)) + D2*(QC(I) - QB(I))
57 CONTINUE
60 CONTINUE
RETURN
200 CONTINUE
K2 = Kl + IGRD
K3 = K2 + 1
IQI = 5
AAA = Dl
Dl = 02
D2 = AAA
GO TO 50
END
C
-------
SUBROUTINE QCOMB(I X,IY,CD)
DIMENSION DISTX(200,3),XDCAY(600,3),SSIGZ(200,3),EXPOZ(600,3 )
i,HA(5),UHA(5),QXY(5),0(3600),QX(200,3,3,3),NQX(3,3)
COMMON/HAS I C/I PR TR, I STAR, IERR,ISIGD,NXLIM,NXZLM,NLI,XONE,XMAX,CCNX
1,DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA , IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ, NH,HA,UHA
1, QXY,Q ,CONC,NH1,QX,NQX
IF ( IX - 8) 20,10, 10
10 CONTINUE
NX = 2
GO TO 30
20 CONTINUE
NX = 4
30 CONTINUE
IF (IY) 31,31,32
31 CONTINUE
IY1 = 1
GO TO 33
32 CONTINUE
IY1 = IY
33 CONTINUE
IF (IX) 34,34,35
34 CONTINUE
1X1 = 1
GO TO 36
35 CONTINUE
1X1 = IX
36 CONTINUE
Kl = 120 * (IY1- 1) + 4 *(IX1- 1) - 30
QD = 0.
DO 50 J=l,4
Kl = Kl + 30
DO 40 1=1,NX
K = Kl + I
QD = QD + Q(K)
40 CONTINUE
50 CONTINUE
QD = QD / (4. * NX)
RETURN
END
SUBROUTINE DISTC
C ROUTINE TO SET DISTANCE DEPENDENT ARRAYS.
DIMENSION DISTX(200,7),XDCAY(600,7),SSIGZ(200,7),EXPOZ(600,7)
1,HA(5),UHA(5),QXY(5),0(3600),QX(200,3,1,7),NQX(7,3)
COMMON/BASIC/IPRTR,I STAR,I ERR,I SIGD,NXLIM,NXZLM,NLI,XONE,XMAX,CCN)
I,DECAY,INDEX,THTA,CIGMX,GX.GY, DLJA,IND»SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/AREAS/NXI,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ, NH,HA,UHA
1, QXY,Q ,CONC,NH1,QX,NQX
COMMON/WNDSP/WSPD,WHGT,PWIND
NXZLM = 600
C SET ARRAYS WHICH DE'PEND ON DISTANCE AND METEOROLOGICAL CONDITIONS
C GET AREA SOURCE WIND SPEEDS FOR EACH EMISSION HEIGHT
DO 40 IH=1,NH
UHA(IH) =WSPD* (HA{IH)/WHGT)**PWIND
40 CONTINUE
-------
DO 200 IB=1,7
NLI = NQX(IB,ICX)
C CHECK THAT DIMENSION LIMIT NXZLM IS NOT EXCEEDED
NHLI=NH*NLI
IF (NHLI-NXZLM) 26,26,25
25 CONTINUE
WRITE (IPRTR,1004)NXZLM,NHLI
1004 FORMAT!' NH * NLI EXCEEDS DIMENSION LIMIT OF1,16,', NH,=',I
16,', NLI=',I6)
IERR = 1
RETURN
26 CONTINUE
C GET TRAVEL DISTANCE DECAY FACTORS
IK = 0
DO 60 1=1,NLI
XI = DISTX(I,IB)
DO 50 IH=1,NH
IK = IK + 1
IF (DECAY) 42,42,44
42 CONTINUE
XDCAY(IK,IB) = 1.
GO TO 46
44 CONTINUE
XARG = DECAY * XI / UHAUH)
IF (XARG - 25.) 45,45,62
^ 45 CONTINUE
XDCAY(IK,IB) = EXP(-XARG)
46 CONTINUE
f- 50 CONTINUE
t- 60 CONTINUE
GO TO 65
62 CONTINUE
IL = NH * NLI
V DO 64 I=IK,IL
XDCAY(I,IB) = 0.
64 CONTINUE
65 CONTINUE
NX I = NLI
C GET SIGMAZ PARAMETERS
DO 130 1=1,NXI
XW = DISTX(I,IB)
CALL SIGZZ
IF (IERR) 67,68,67
67 CONTINUE
RETURN
68 CONTINUE
SSIGZ(I,IB) = SIGZ
IF(SIGZ-CIGMX)70,140,140
70 CONTINUE
130 CONTINUE
KUTEX = NXI
GO TO 160
140 CONTINUE
^ KUTEX = I
DO 150 J=I,NXI
SSIGZ(J,IB) = CIGMX
r 150 CONTINUE
^ 160 CONTINUE
200 CONTINUE
RETURN
C 6ND
C
-------
SUBROUTINE EXPZC
C ROUTINE TO COMPUTE VERTICAL DIFFUSION FACTOR INCLUDING EFFECTS OF
C DECAY AND GROUND REFLECTIONS FOR EACH OF NH SOURCE HEIGHTS.
C BASIC EQUATION IS
C EXPCZ = ( XDCAY / SIGZ )*{EXP(-0.5*((HA-ZR)/SIGZ)**2) +
C EXP(-0.5*l(HA+ZR)/SIGZ)**2))
DIMENSION DISTX(200,7),XDCAY(600,7),SSIGZ(200,7),EXPOZ(600,7)
1,HA(5),UHA(5),QXY(5),0(3600),QX(200,3,1,7),NQX(7,3)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE,XMAX,CCNX
1,DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA,IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
COMMON/AREAS/NX I,NX,KUTEX,DISTX,XDCAY,SSIGZ,EXPOZ, NH,HA,UHA
It OXY,Q ,CONC,NH1,QX,NOX
C INPUTS
C ZR » RECEPTOR HEIGHT
C NH = NUMBER OF SOURCE HEIGHTS
C HA = ARRAY OF SOURCE HEIGHTS
C CIGMX = MIXING CEILING
C SSIGZ = VERTICAL DIFFUSION PARAMETER
C OUTPUT
C EXPOZ = VERTICAL DIFFUSION FACTOR
DO 100 IB=1,7
NXI = NQX(IB,ICX)
, K = 0
( DO 60 J=1,NXI
G = 2. *CIGMX/SSIGZ(J,IB)
DO 50 1=1,NH
f F = (HA( I ) - ZR) /SSIGZ(J,IB)
( IF (F*F - 50.) 12,12,10
10 CONTINUE
El = 0
r- 11 CONTINUE
(> E2 = 0
GO TO 15
12 CONTINUE
El = EXP(-0.5 * F * F)
1 F = (HA(I) + ZR) /SSIGZ(J,IB)
IF (F- 7.) 13,13,11
13 CONTINUE
E2 = EXP(-0.5 * F * F)
15 CONTINUE
EXPOZ(K,IB) = ((El + E2) / SSIGZ(J,IB)) * XDCAY(K,IB)
50 CONTINUE
60 CONTINUE
NX = NX I
100 CONTINUE
RETURN
END
SUBROUTINE OUTSE
DIMENSION CNCT(50),CAREA(50),CPOIN(50),XRR(50),YRR(50),ZRR(50)
ltOBS02(50),OSET(12,5A)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE,XMAX,CCN)
1,DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA,IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
2,XW,YW,ICARD,ICX
c COMMON/OUTPUT/YEAR,AMON,DAY,HOUR ,NRECP,CNCT,CAREA,CPOIN,XRR,YRR
1.ZRR.OBS02,IDOWiNRl,TEMP,SIGA,RIBfPCPNtWGLDiSTAPRtOSET,INDO
DATA Jl/1/
WRITE(12) (J,(OSET(I,J-J1+1),I=1,12),J=J1,INDO)
r WRITE!IPRTR,1000) J1,INDO
^ 1000 FORMATC SETS',16,' TO',16,' ENTERED IN SENSITIVITY OUTPUT FILE')
Jl = INDO + 1
RETURN
c ENO
-------
SUBROUTINE CONTRS{J, I 1, 12 )
DIMENSION DISTX(200,7),XDCAY(600,7),SSIGZ(200,7),EXPOZ(600,7 )
1,HA(5),UHA(5) ,QXY(5),Q(3600),QX(200,3,1,7),NQX(7,3)
DIMENSION SXK5), TERMA(5)
COMMON/BASIC/IPRTR,I STAR,1 ERR,ISIGD,NXLIM,NXZLM,NLI,XONE,XMAX,CON>
1,DECAY,INDEX,THTA,CIGMX,GX,GY, DLTA,IND,SIGY,SIGZ,XR,YR,ZR,XS,YS
2,XW,YW,ICARD,ICX,XYMIN1XSMAX,YSMAX
COMMON/AREAS/NXI,NX,KUTtX,DISTX,XDCAY,SSIGZ,EXPOZ, NH,HA,UNA
1, QXY,Q ,CONC,NH1,QX,NQX
C CONST = 1/(2*SQRT(2*PI))
DATA CONST/0.199471/
YW = 0.
NX = NQXl12,J)
DO 10 IH=1,NH
SXH IH) = 0
TERMA(IH) = 0
10 CONTINUE
XWL = DISTXl1,12)
DO 110 1=1,NX
f XW = DISTXt1,12)
( IF (NH - 1) 40,40,50
40 CONTINUE
QXY(1) = QX(I,J,II,12)
f. GO TO 60
v 50 CONTINUE
QXY(2) = 0.5 * QX(I,J,I1,12)
OXY(1) = 0.5 * QXY(2)
OXY13) = QXY(1)
<•> 60 CONTINUE
DO 90 IH=1,NH
K = ( I -1) * NH + IH
IF (QXY(IH) - l.E-50) 70,70,80
70 CONTINUE
TERMB = 0.
GO TO 85
80 CONTINUE
TERMB = QXY(IH) * EXPOZ(K,I2)
85 CONTINUE
SXKIH) = SXI(IH) + (TERMA(IH) + TERMB) * (XW - XWL)
TERMAtIH) = TERMB
90 CONTINUE
100 CONTINUE
XWL = XW
110 CONTINUE
IX = NX + 1
IF (IX - NLI) 115,115,140
115 CONTINUE
XW = DISTX(IX,12)
120 CONTINUE
DO 130 IH=1,NH
SXKIH) = SXKIH) + TERMA(IH) * ( XW - XWL)
,- 130 CONTINUE
140 CONTINUE
CONC = 0
DO 150 IH=1,NH
f- CONC = CONC + SXKIH) / UHA(IH)
c 150 CONTINUE
CONC = CONST * CONC
RETURN
,- END
-------
Exhibit D-3
LISTING OF FORTRAN CODE SUBROUTINES USED WITH
BOTH VALIDATION AND SENSITIVITY PROGRAMS
-------
SUBROUTINE POINT
C DUE TO EMISSIONS FROM SPECIFIED POINT SOURCES
C BASIC EQUATION IS
C CONPI =(Q/(2*PI*U*SIGY*SIGZ))*EXP(-0.5*(Y/SIGY)**2-DECAY*X/U )
C *(EXP(-0.5*((Z-H)/SIGZ)**2)+EXP(-0.5((Z+H)/SIGZ)**2)))
C U = WSPD*(ZP/WHGT)**PWIND
C SIGY = HORIZONTAL DIFFUSION PARAMETER
C SIGZ = VERTICAL DIFFUSION PARAMETER
C Y = CROSSWIND DISTANCE BETWEEN SOURCE AND RECEPTOR
C Z = ZR
C H = ZP
DIMENSION XP (100),YP (100),ZP (100),QP(100)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
COMMQN/WNDSP/WSPD,WHGT,PWIND
COMMON/POINTS/NPS ,XP ,YP ,ZP ,QP,CONPS,NPSS
C INPUTS
C XR,YR,ZR = RECEPTOR LOCATION IN SOURCE GRID COORDINATES
C NPS = NUMBER OF POINT SOURCES
C XSSXP,YP,ZP = ARRAYS OF POINT SOURCE LOCATIONS IN SOURCE GRIDS
C QP = POINT SOURCE EMISSION RATE
C XONE = CLOSEST DISTANCE TO RECEPTOR
C ISIGD = DIFFUSION PARAMETER OPTION
C DECAY = DECAY CONSTANT
C THTA = WIND DIRECTION
C WSPD = WIND SPEED AT HEIGHT WHGT
C PWIND = WIND PROFILE PARAMETER
1 C INDEX = DIFFUSION STABILITY PARAMETER
C CIGMX = MIXING CEILING
C OUTPUTS
C IERR = ERROR INDICATOR FROM SIGYZ ROUTINE
( •• C CONPS = CONCENTRATION AT RECEPTOR FROM NPS POINT SCURCES
DATA PI/3.14159/
CONPS = 0
Nl = NPS - NPSS
v DO 50 I =N1,NPS
HH = ZP (I)
XS = XP (I)
YS = YP (I)
CALL WCORD
IFUW -XONE) 40,10,10
10 CONTINUE
CALL SIGYZ
IF (IERR) 20,30,20
20 CONTINUE
RETURN
30 CONTINUE
IF (HH - 1.) 28,29,29
28 CONTINUE
HH = 1.
29 CONTINUE
U = WSPD * (HH / WHGT)**PWIND
CONPI =QPlI)/(2*PI*U*SIGY*SIGZ)
FY = YW/SIGY
FY2 = FY * FY
IF (FY2 -50. ) 32,40,40
32 CONTINUE
CONP2 = EXP(-0.5*FY2)
/ 33 CONTINUE
r- FZ1 = (ZR-HH)/SIGZ
FZ2 = (ZR+HHJ/SIGZ
IF(FZ1 * FZ1 -50.) 35,34,34
r 34 CONTINUE
^ CONP3 = 0
GO TO 36
35 CONTINUE
/- CONP3 = EXP(-0.5*FZ1*FZ1)
-------
36 CONTINUE
IP (FZ2 - 7. ) 38,37,37
37 CONTINUE
CONP4 = 0
GO TO 39
38 CONTINUE
CONP4 = EXP(-0.5*FZ2*FZ2)
39 CONTINUE
IF (DECAY) 44,44,45
44 CONTINUE
CONP6 = 1.
GO TO 46
45 CONTINUE
AAA = DECAY*XW/U
IF (AAA - 25.) 48,40,40
48 CONTINUE
CONP6 = EXP(-AAA)
46 CONTINUE
CONP7 = CONP3 + CONP4
IF (CONP7 - l.E-10) 40,40,47
47 CONTINUE
CONPI = CONP1*CONP2*CCNP7*CONP6
CONPS = CONPS + CONPI
40 CONTINUE
50 CONTINUE
RETURN
END
C-
SUBROUTINE SIGYZ
C ROUTINE TO CALL SUBPROGRAM DESIGNATED BY ISIGD
COMMON/BASIC/I PRTR,I STAR,I ERR , I SI GD ,K'XL I M ,NXZLM , NL I , XONE
1tXMAXtCONXtDECAYtINDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
C ISIGD = INDICATES OPTION FOR DEFINING DIFFUSION PARAMETERS
C 1 = MCELROY-POOLER PARAMETERS USING TURNER STAB.
C 2 = MCELROY-POOLER PARAMETERS USING RICHARDSON NO.
C 3 = MCELROY-POOLER PARAMETERS USING BRCOKHAVEN STAE
C 4 = PASQUILL PARAMETERS USING TURNER STABILITY CAT.
C DECAY = DECAY CONSTANT (PER SEC)
IERR = 0
GO TO ( 400, 500, 600, 700),ISIGD
400 CONTINUE
CALL SIGY1
CALL SIGZ1
GO TO 800
500 CONTINUE
CALL SIGY2
CALL SIGZ2
GO TO 800
r 600 CONTINUE
CALL SIGY3
CALL SIGZ3
GO TO 800
r 700 CONTINUE
(- CALL SIGY4
CALL SIGZ4
800 CONTINUE
r. RETURN
^- END
C
-------
c
c
o
SUBROUTINE SIGY1
C MCELROY-POOLER PARAMETERS BASED ON TURNER-PASQUILL STABILITY CATEGORY
C THE CALCUALTION USES A POWER LAW - A(INDEX)*X**P(INDEX)
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C RESTRICTION XW MUST BE POSITIVE
C IF RESTRICTION IS VIOLATED THE CALCULATION IS NOT MADE AND
C MESSAGE IS RETURNED
DIMENSION A(5)fP(5)
COMMON/BASIC/IPRTR,I STAR,I ERR,ISI GO,NXLIM,NXZLM,NLI,XCNE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGNX,GX,GY,DLTA,IND,SIGY
2fSIGZfXRtYRtZRfXSfYSfXNfYWfI CARD,ICX,XYMIN,XSMAX,YSMAX
C VALUES OF A AND P ARE GIVEN IN THE DATA STATEMENTS FOR EACH ARRAY
DATA A/ 0., 1.42,1.26,1.13,0.992/
DATA P/0.f.7A5fO.73tO.71tO.65/
C TEST FOK RESTRICTION
IF (XW) 10,10,20
10 CONTINUE
WRITE(IPRTR,15) XW
15 FORMATdH ,'SUBROUTINE SIGY1 - BAD INPUT X =',F15.2)
I ERR = 1
RETURN
20 CONTINUE
SIGY=A(INDEX)*XW**P(INDEX)
RETURN
END
SUBROUTINE SIGY2
C MCELROY-POOLER PARAMETERS BASED ON RICHARDSON NO.
C THE CALCUALTION USES A POWER LAW - A(INDEX)*X**P(INDEX)
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C IERR = ERROR RETURN
C RESTRICTION XW MUST BE POSITIVE
C IF RESTRICTION IS VIOLATED THE CALCULATION IS NOT MADE AND
C MESSAGE IS RETURNED
DIMENSION A(5) t P{5)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
DATA A/1.49,1.4,1.26,1.14,0.945/
DATA P/.76 1,0.719,0.712,0.698,0.648/
C TEST FOR FIRST RESTRICTION
IF(XW )5,5,10
5 CONTINUE
WRITE( IPRTR, 15) XW •
15 FORMAT!1H ,'SUBROUTINE SIGY2 - BAD INPUT X =',F15.2)
IERR = 1
RETURN
S1GY=A(INDEX)*XW**P(INDEX)
RETURN
-------
SUBROUTINE SIGY3
C MCELROY-POOLER PARAMETERS BASED ON MODIFIED BROOKHAVEN STABILITY CAT.
C THE CALCUALTION USES A POWER LAW - A(INDEX)*X**P(INDEX)
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C RESTRICTION XW MUST BE POSITIVE
C IF RESTRICTION IS VIOLATED THE CALCULATION IS NOT MADE AND
C MESSAGE IS RETURNED
DIMENSION A(4),P(4)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,1CARD,ICX,XYMIN,XSMAX,YSMAX
DATA A/1.48425,1.74942,1.50373,1.39026/
DATA P /.73727,.68741,.66b52,.61474/
C TEST FOR FIRST RESTRICTION
IF (XW) 10,20,20
WRITE( IPRTR, 15) XW
15 FORMAT!1H ,'SUBROUTINE SIGY3 - BAD INPUT X =',F15.2)
IERR = 1
RETURN
20 CONTINUE
SIGY=A(INDEX)«XW**P(INDEX)
RETURN
END
SUBROUTINE SIGY4
C PASQUILL-GIFFCRD CALCULATION FOR SIGY
C THE CALCUALTION USES A POWER LAW - A(INDEX)*X**P
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C RESTRICTION XW MUST BE POSITIVE
DIMENSION A(6)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
DATA A/.3658,.2751,.2089,.1474,.1046,.0722/
DATA P/ .9031/
C TEST FOR FIRST RESTRICTION
IF (XW) 10,10,20
10 CONTINUE
WRITE(IPRTR,15) XW
15 FORMATUH , 'SUBROUTINE SIGY4 - BAD INPUT X =',F15.2)
IERR = 1
RETURN
20 CONTINUE
SIGY=A(INDEX)*XW**P
RETURN
END
SUBROUTINE GCHEK
C ROUTINE TO DETERMINE IF POINT X,Y IS OUTSIDE RECT. GRID AREA DEFINED
C BY DIAGONAL FROM 0.5*DELTA,0.5*DELTA TO (GX+0.5)*DELTA,(GY+0.5)*DELTA
C IND=0 FOR ON GRID, IND=-i FOR OUTSIDE GRID
COMMON/BASIC/IPRTR,I STAR,I ERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
f IF (XS- XSMAX) 10,50,50
10 IF (YS- YSMAX) 20,50,50
20 CONTINUE
IF (XS- XYMIN) 50,50,30
r 30 CONTINUE
'- IF (YS- XYMIN) 50,50,40
40 CONTINUE
IND = 0
r RETURN
^ 50 CONTINUE
IND = -1
RETURN
C ENO
-------
SUBROUTINE SIGZZ
COMMON/BASIC/IPRTR,I STAR,I ERR,ISIGD,NXL1M,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
IERR = 0
GO TO ( 400, 500, 600, 700),ISIGD
400 CONTINUE
CALL SIGZ1
GO TO 800
500 CONTINUE
CALL SIGZ2
GO TO 800
600 CONTINUE
CALL SIGZ3
GO TO 800
700 CONTINUE
CALL SIGZA
800 CONTINUE
RETURN
END
c.
SUBROUTINE SIGZ1
C MCELROY-POOLER PARAMETERS BASED ON TURNER-PASQUILL STABILITY CATEGORY
C THE CALCUALTION USES A POWER LAW - A(INDEX)*X**P(INDEX)
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C IERR = ERROR RETURN
C RESTRICTION - INDEX MUST BE BETWEEN 2 AND 5
C IF RESTRICTION IS VIOLATED AN ERROR MESSAGE IS RETURNED
DIMENSION A(5,2)»P(5.2)
COMMON/BASIC/IPRTR,I STAR,I ERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,OLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
DATA A/ 0.,.0926,0.0891,0.0835,0.0777,0.0,0.072,0.169,1.07,1.0 I/
DATA P/.0,1.18,1.11,1.08,0.955,0.0,1.22,1.01,0.682,0.554/
IF (ISTAR) 90,5,90
5 CONTINUE
r- C TEST FOR RESTRICTION
20 CONTINUE
IF (5-INDEX) 30,50,50
30 CONTINUE
r WRITEtIPRTR,35) INDEX
(-* 35 FORMATUH ,'SUBROUTINE SIGZ1 - BAD INPUT INDEX =',IA)
IERR = 1
RETURN
x-s 50 CONTINUE
C IFUNDEX-2) 30,60,60
o
o
-------
60 CONTINUE
IF(CIGMX)70,70,80
70 CONTINUE
WRITE (IPRTR.1000) CIGMX
1000 FORMAT (' PARAMETER OUT OF RANGE ,C I GMX= ' , F8 . 1 )
IERR = 1
RETURN
80 CONTINUE
XI DEFINES DISTANCE FOR WHICH SIGMAZ = 0.5*MIXING CEILING
DO 82 J=l,2
B = A( INDEX, J)
Q = P( INDEXi J)
XI = (CIGMX / (2. * B))**(l. / Q)
IF (XI - 600. ) 83,83,82
82 CONTINUE
83 CONTINUE
X2 DEFINES DISTANCE FOR WHICH SIGMAZ = MIXING CEILING
DO 85 J=l,2
B = A(INDEX,J)
Q = P( INDEX, J)
X2 = (CIGMX/B)**( l./Q)
IF (X2 - 600. ) 85,85,84
84 CONTINUE
85 CONTINUE
RETURN
90 CONTINUE
IF(XW-Xl) 100,200,200
X LESS THAN XI, NO MODIFICATION
100 CONTINUE
DETERMINE WHICH RANGE XW IS IN
IF (XW - 600. ) 110,110,120
110 CONTINUE
FIRST RANGE
J = 1
GO TO 130
120 CONTINUE
SECOND RANGE
J = 2
130 CONTINUE
B = At INDEX, J)
0 = P( INDEX, J)
SIGZ = B *XW**Q
RETURN
200 CONTINUE
IF (XW - X2) 300,400,400
X BETWEEN XI AND X2
300 CONTINUE
SIGZ = 0.5 * CIGMX * (XW + X2 - 2.*X1)/(X2 - XI)
RETURN
X GREATER THAN X2
400 CONTINUE
SIGZ= CIGMX
RETURN
END
C
C
-------
SUBROUTINE SIGZ2
C MCELROY-POOLER PARAMETERS BASED ON RICHARDSON NO.
C THE CALCUALTION USES A POWER LAW - A( I NDEX)*X**P(INDEX)
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C IERR = ERROR RETURN
C RESTRICTION INDEX MUST BE BETWEEN 1 AND 5
DIMENSION A(5,2)fP(5,2)fC(5)
COMMON/BASIC/IPRTR,I STAR,I ERR,I SIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
DATA A/2*.118,0.115,0.11,0.0954,2*0.00724,0.0581tO.llfO.478/
DATA P/2*1.02,1.,0.934,0.907,2*1.51,1.12,0.934,0.655/
DATA C/3*300.,10000.,600./
IF (ISTAR) 90,5,90
5 CONTINUE
C TEST FOR RESTRICTION
IF (5-INDEX) 30,50,50
30 CONTINUE
WRITE(IPRTR.35) INDEX
35 FORMATtlH ,'SUBROUTINE SIGZ2 - BAD INPUT INDEX =',I4)
IERR = 1
RETURN
50 CONTINUE
IFUNDEX-1) 30,60,60
60 CONTINUE
IF(CIGMX)70,70,80
70 CONTINUE
WRITE UPRTR,1000) CIGMX
1000 FORMAT (' PARAMETER OUT OF RANGE,CIGMX=',F8.1)
IERR = 1
RETURN
80 CONTINUE
C XI DEFINES DISTANCE FOR WHICH SIGMAZ = 0.5*MIXING CEILING
DO 82 J=l,2
B = A( INDEX,J)
Q = P( INDEX,J)
XI = (CIGMX / (2. * B))**(!. / Q)
IF (XI - C( INDEX) ) 83,83,82
82 CONTINUE
83 CONTINUE
C X2 DEFINES DISTANCE FOR WHICH SIGMAZ = MIXING CEILING
DO 85 J=l,2
B = A(INDEX,J)
Q = P(INDEX,J)
X2 = (CIGMX/B)**(l./Q)
IF (X2 - 600.) 85,85,84
84 CONTINUE
85 CONTINUE
RETURN
90 CONTINUE
IF(XW-Xl)100,200,200
C X LESS THAN XI, NO MODIFICATION
100 CONTINUE
C DETERMINE WHICH RANGE XW IS IN
IF (XW - C(INDEX)) 110,110,120
110 CONTINUE
C FIRST RANGE
J = 1
GO TO 130
120 CONTINUE
C SECOND RANGE
J = 2
130 CONTINUE
B = A(INDEX,J) '
0 = P(INDEX,J)
SIGZ = B *XW**Q
RETURN
200 CONTINUE
-------
IF (XW - X2) 300,400,400
C X BETWEEN XI AND X2
300 CONTINUE
SIGZ = 0.5 * CIGMX * (XW + X2 - 2.*X1)/(X2 - XI)
RETURN
C X GREATER THAN X2
400 CONTINUE
S1GZ = CIGMX
RETURN
END
SUBROUTINE SIG23
C MCELROY-POOLER PARAMETERS BASED ON MODIFIED BROOKHAVEN STABILITY CAT.
C THE CALCUALTION USES A POWER LAW - A(INDEX)*X**P(INDEX)
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C RESTRICTION INDEX MUST BE BETWEEN 1 AND 4
C IF EITHER RESTRICTION IS VIOLATED THE CALCULATION IS NOT MADE AND
C MESSAGE IS RETURNED
DIMENSION A(4),P(4)
COMMON/BASIC/IPRTR,ISTAR,IERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,C1GMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSMAX
DATA A/.04952,.00837,.11285,.52039/
DATA P/l. 14047,1.48884,.9332,.646727
IF (ISTAR) 90,5,90
5 CONTINUE
C TEST FOR RESTRICTION
IF (4-INDEX) 30,50,50
1 30 CONTINUE
WRITE(IPRTR,35) INDEX
35 FORMATUH ,'SUBROUTINE SIGZ3 - BAD INPUT INDEX =',I4)
IERR = 1
RETURN
50 CONTINUE
IF (INDEX-1) 30,60,60
60 CONTINUE
IF(CIGMX)70,70,80
70 CONTINUE
WRITE (IPRTR.IOOO) CIGMX
1000 FORMAT (' PARAMETER OUT OF RANGE,CIGMX=',F8.1)
IERR = 1
RETURN
80 CONTINUE
B=A(INDEX)
Q=P(INDEX)
C XI DEFINES DISTANCE FOR WHICH SIGMAZ = 0.5*MIXING CEILING
RETURN
90 CONTINUE
IF(XW-Xl)100,200,200
C X LESS THAN XI, NO MODIFICATION
100 CONTINUE
f SIGZ = B *XW**Q
RETURN
200 CONTINUE
IF (XW - X2) 300,400,400
f C X BETWEEN XI AND X2
( 300 CONTINUE
SIGZ = 0.5 * CIGMX * (XW + X2 - 2.*X1)/(X2 - XI)
RETURN
C C X GREATER THAN X2 •
*-- 400 CONTINUE
SIGZ= CIGMX
RETURN
C END
-------
SUBROUTINE SIGZ4
C PASQUILL-GIFFORD CALCULATION FOR SIGZ
C THIS CALCULATION USES A POWER LAW - A*X**P
C INPUTS-XW IS DISTANCE FROM SOURCE, INDEX IS STABILITY CLASS NUMBER
C RESTRICTION INDEX MUST BE BETWEEN 1 AND 5
DIMENSION A(5,3),P(5,3),C(5,2)
COMMON/BASIC/IPRTR,I STAR,I ERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,I CARD,ICX,XYMIN,XSMAX,YSMAX
DATA A/.125,0.119,0.111,0.105,0.1,0.00883,0.0579,0.111,0.392,0.372
1,0.000226,0.0579,0.111,.948,2.867
DATA P/1.03,0.986,0.911,0.827,0.778,1.51,1.09,0.911,0.636,0.587
1,2.1, 1.09,0.911,0.540,0.3667
DATA C/250.,4*1000.,500.,4*10000.7
IF( ISTAR ) 84,5,84
5 CONTINUE
C TEST FOR RESTRICTION
IF(5-INDEX) 30,50,50
30 CONTINUE
WRITE( IPRTR.35) INDEX
35 FORMAT ( 1H , 'SUBROUTINE SIGZ4 - BAD INPUT ------ INDEX =«,I4)
IERR = 1
40 RETURN
50 CONTINUE
IF(INDEX-l) 30,60,60
60 CONTINUE
IF(CIGMX)70,70,80
70 CONTINUE
WRITE (IPRTR.IOOO) CIGMX
1000 FORMAT (' PARAMETER OUT OF RANGE , CIGMX= ', F8 . 1 )
IERR = 1
f- RETURN
(; 80 CONTINUE
C XI DEFINES DISTANCE FOR WHICH SIGMAZ = 0.5*MIXING CEILING
DO 82 J=l,3
B=A(INDEX,J)
11 Q=P(INDEX,J)
XI = (CIGMX/(2.*B) )**(!. /Q)
IF ( J - 3) 81,81,83
81 CONTINUE
v. IF (XI - C(INOEXfJ)) 83,83,82
82 CONTINUE
83 CONTINUE
C X2 DEFINES DISTANCE FOR WHICH SIGMAZ = MIXING CEILING
DO 92 J=l,3
B=A( INDEX, J)
0=P( INDEX, J)
X2 = (CIGMX/B)**( l./Q)
IF (J - 3) 91,91,93
91 CONTINUE
IF (X2 - CUNDEXfJM 93,93,92
92 CONTINUE
93 CONTINUE
RETURN
84 CONTINUE
IF (XW - XI) 86,200,200
^ C XW LESS THAN XI, NO MODIFICATION
C DETERMINE WHICH RANGE X IS IN
86 CONTINUE
r- IF (XW - C(INDEXtJ)) 85,85,90
^ 85 CONTINUE
C J=l INDICATES THAT X IS IN THE FIRST RANGE
J = l
n GO TO 120
O 90 IF (XW - C(INDEX,J)J 100,110,110
100 CONTINUE
C J=2 INDICATES THAT X IS IN THE SECOND RANGE
c J=2
-------
GO TO 120
110 CONTINUE
J=3 INDICATES THAT X IS IN THE THIRD RANGE
J = 3
120 CONTINUE
SIGZ= A( INDEX,J)*XW**P(INDEX,J)
RETURN
200 CONTINUE
IF (XW - X2) 300,400,AGO
X BETWEEN XI AND X2
300 CONTINUE
SIGZ = 0.5 * CIGMX * (XW + X2 - 2**X1)/(X2 - XI)
RETURN
X GREATER THAN X2
400 CONTINUE
SIGZ= CIGMX
RETURN
END
SUBROUTINE SCORD
COMMON/BASIC/IPRTR,I STAR,I ERR,ISIGD,NXLIM,NXZLM,NLI,XONE
1,XMAX,CONX,DECAY,INDEX,THTA,CIGMX,GX,GY,DLTA,IND,SIGY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,1CARD,ICX,XYMIN,XSMAX,YSMAX
IF (ISTAR) 20,10,20
10 CONTINUE
ALPH=3.14159/2.- THTA
COSA = COS(ALPH)
SINA = SIN(ALPH)
RETURN
20 CONTINUE
XS = XR + XW * COSA - YW * SINA
YS = YR + XW * SINA + YW * COSA
RETURN
END
SUBROUTINE WCORD
C THIS ROUTINE CONVERTS GRID COORDINATES ON STANDARD GRID REF. SYSTEM TC
C COORDINATES IN A SYSTEM WITH THE X AXIS ALIGNED WITH THE WIND.
C THTA IS WIND DIRECTION ALPHA IS THE ANGLE OF ROTATION FROM STANDARD
C TO WIND ORIENTED SYSTEM
COMMON/ BASIC/ I PRTR, ISTAR, I ERR, I S I GD , NXL I M , NXZLM , NL I , XONE
It XM AX, CONX, DECAY, INDEX, THTA , C IGMX ,GX ,GY , DLTA , I ND , S I GY
2,SIGZ,XR,YR,ZR,XS,YS,XW,YW,ICARD,ICX,XYMIN,XSMAX,YSKAX
IF (ISTAR) 20,10,20
10 CONTINUE
ALPHA=3. 14159/2.- THTA
COSA = COS (ALPHA)
SINA = SIN (ALPHA)
RETURN
20 CONTINUE
XP1 = (XS-XR)*COSA
XP2 = (YS-YR)*SINA
XW = XP1 + XP2
YP1 =(YS-YR)*CO'SA
YP2 =(XS-XR)*SINA
YW = YP1 - YP2
RETURN
END
-------
Appendix E
-------
Appendix E
SAMPLES OF VALIDATION DATA LISTINGS
*
This appendix describes and presents samples of the punched
cards and computer printouts which were generated in the course of this
study and which were reviewed to determine results and findings. Since
the total set of all pages of computer printouts consists of several
thousand pages, only samples of each type of printout are reproduced
here.
Two input data records were formed for each hour of validation
data. These data records are stored on magnetic tape and may be loaded
into disk files for convenient retrieval. An example of the information
contained in each record is shown by Figures E-l and E-2. Figure E-l
lists meteorological parameters, time and date information, sampler
observations, and point source emission information which are available
in one hourly input record. A description of the computer printout iden-
tifications shown in the figure is given in Table E-l. Figure E-2 is
a map of hourly area source emission rates obtained from the second type
of hourly input record. These figures are samples of the computer print-
outs generated to review input data for each hour of validation data.
Figures E-3 and E-4 are samples of the map printed for each
short-term (two-hour for St. Louis, one-hour for Chicago) validation
period. The X's show the relation of sampler locations relative to one
another. The upper printed number by each X is the observed value, and
the lower printed number is the predicted value. Stars designate missing
values. The principal meteorological inputs are also shown.
-------
I
ro
YFAP
64.
ORS
OBS
ORS
DBS
OP
OP
ZP
QP
ZP
OP
ZP
QP
ZP
OP
ZP
:. in OHM TF
214. 4 34
AMCN nAY HOUR
12. 2. 14.
381.^00 ?22
129.0^0 30
84.000 R]
150.000 23
2 °3 .000 oqqo
I
0.806F+02 0.2
'+0.00
11
0 . 1 5 2 F + 03 0.3
1^2.23
21
0.203F+02 0.3
31
1^7.20
0.3c3F+~3 0.5
106.87
o.^;
MP SIGA
NPFCP GV
«?0 30.
.000 126
.000 100
.000 22
.000 97
.000 '4 1 0
2
78E+03 0.
238.03
12
34P+02 0.
22
05F4-03 0.
125.00
32
10F+03 0.
55.00
42
29E+03 0.
122.21
R I n
GY HPLTA
40. 1^24.
.000 IS?.
.000 110.
.000 6.
.000 29.
.000 10.
3
31RE+0? 0.
40.00
13
172F+02 0.
40.00
23
61 OF. +02 0.
55.00
33
5^.00
706F+03 0.
134. 7Q
prpM WGUST
2.00 0.
MM MRPMT
3 51
000 174
000 173
000 34
000 236
4
42QF+03
111.34
14
764E+02
40.00
24
123F+03
73.38
34
217E+02
75.QO
44
268F+04
328.06
USPn
^.70
.000
.000
,000
,000
5
0. 173F
40
15
0.645F
40
PRFSS
27.74
20.8
114,000
174.000
lo.OOO
999.000
°°99.000
+02 0.1
.00
+01 0.1
.00
25
0.166F+02 0.1
75.00
35
0.184C
79
45
0.682F
257
+03 0.5
.10
+03 0.6
.16
40
WIND
0.128
362
290
2^9
9990
6
79E+03
101.15
16
72E+02
49.00
26
32F+03
97.25
36
59F+03
210.12
46
82F+03
?57.16
THTA 1NOFX
1.500 4
.000
.0^0
.000
.000
.000
0.
0.
0.
0.
0.
116.
101.
40.
00.
7
323E+02
40.00
17
173E+02
19.00
27
163E+C3
90.31
37
586F+03
300.80
128F+04
399.66
CIGMX
441.
000
000
000
COO
000
0
0
0
0
0
. .104.000.
' 090.000
194.000
72.000
_999Q.OOO
. . 9
.476F+03
. 146.30
is
.142E+02
47.00.
28
,144=+02
55.00
38
.25&-+02
75.00.
48
.167E+04
463.70
-------
ST. LOUIS AREA SOURCE EMISSION RATES
0 -
1 -
2 -
3 -
4 -
5 -
0.
0.
0.
0.
0.
1.
0
01 +
03 +
10 +
30+
00 +
TO
TO
TO
TO
TO
TO
0.
0.
0.
0.
1.
3.
01
03
10
30
00
00
6 -
7 -
8 -
9 -
115.
3.
10.
30.
OVER
00 +
00 +
00 +
100
TO
TO
TO
.00
10.
30.
100.
00
00
00
1
123456789012345678901234567890
1 344444444233303333344444444444
2 334444444323333343344444444444
3 334444444342333333455434440444
4 34444444^444244455555444444444
5 333444444344300055554444444444
6 333444454444425555554444444444
7 333344334444430055444445644444
8 333333444444430555444444434444
9 333333444444433344444444444044
10 333333444445532045444344444455
11 333344444455652006444444444357
12 444444454555575445444444444476
13 454434655556677254544444444677
14 466434665666767704555554445667
15 444455656656767860555465448555
16 444456656666677875452666565555
17 4444666576767778B8447865777555
18 334444356777778888878887555555
19 434444336777777888768987755555
20 4444 5555567765889885 6~8 57755555
21 333345444677888988764556655477
22 433345454567888888776520555467
23 834334555666688888677644444435
24 333344566566677865267644044344
25 333345667776677760067864444044
26 3333~55666666667770026664444044
27 233355665777556566004544444434
28 203355565576665547304444544444
29 320123365456655543443455554444
30 570234444556554444450455550444
31 676445554335444455430044444444
32 255044444434444444440024444444
33 332204233433344444400021044444
34 033220203333333333000006603224
35 333322333233333332220022611333
36 223332232233332233330223600333
37 333333333300023333222326754333
38 33333-3333300033332577767553333
39 033332323333333323787862462444
40 022333223332330224766680665533
Figure E-2. Sample Map of Hourly Area Source Emission Rates
-------
Table E-l. Description of Computer Printout Identifications
Identification
Description
REC. ID
DOW
TEMP
SIGA
RIB
PCPN
WCUST
PRESS
NRI
AMON
DAY
NRECP
GX
GY
DELTA
NH
NRPNT
USPD
UHGT
PWIND
THTA
INDEX
CIGMX
OBS
QP
ZP
Record identification number
Day of the week
Ambient air temperature (°F)
Standard deviation in horizontal wind direction
Bulk Richardson number
Precipitation in 0.01 in/hr
Peak wind gust, knots
Atmospheric pressure, in Hg
= 40, number of 24-hour samplers
Month of the year
Day of the month
= 50, number of 24-hour and 2-hour samplers
Number of East-West coordinates in area source grid
Number of North-South coordinates in area source grid
Spacing between area source grid points, m
= 3, number of area source heights
Number of point sources
Wind speed at height UHGT, m/sec
Reference height for wind speed, m
Wind profile exponent
Wind direction, radians from North
Stability class for diffusion parameters
Mixing ceiling, m
o
Observed SOp concentration, yg/m (first 40 are 24-
hour average, last 10 are 2-hour average)
Point source emission rate, g/sec
Effective point source height, m
-------
ST LOUIS i
$ * * * * ft * * * :
*
*
*
*
'I*
*
*
*»"
*
*
*
*
*
*
*
*
*
*
*
*
*
*
;£ ftfc ^:*** **=;;* **=!:« ^t * * ^: * * **** ****** **** ****** ****** $X: ********
X*******
in.644
X 103.000
203.038
X 130.000
151.624
X 736.000
58^.737
X 60.000
53.715
X 653.000
923.716
X 349.000
640.000 X
S56.345
784.873
X
424.?04
X 58.000
644.03]
*
*
*
*
*
*
*
*
J,
1*
*
*
*
*»-
^»
*
*
*
*
*
*
*
J,
1*
*
*
*
*
*
r * # * =Jc #$$ * $ **** t- *** * *********************
HOI)!? 5,
'-iP'JD nmrCTICN (OFGPEES)
WIND SPEFT (M/SEC )
S T A P I L ! T Y C L \ S S
Tf^PFRATURE (HEG.F )
MlX.rFU.lNG (METERS)
PKrflPITATION
-34. A
2.8
5
13.0
352.0
C.O
HOUR 6 .
-45.8
5
13.0
352.0
O.C
Figure E-3. Sample Computer Generated Map Used to
Review St. Louis Validation Results
-------
*
*
*
*
*
*
« * ft A -|: * -;t .-^ ft * *•
x" "oV^i" •>
C . I H 1
* * * * * * * * *
X *******
0.7X0
0 . S 4 7
n. 6 M
1,1 T ' I ri p y 'J -r r- T J p N' J • ) f. f; n p r (
WT'-jH C.TCP-, f 1/C)cr )
S"r/\n 11 i TY <~\ ACC
T F •' P r i? ^ T ij ^ P ( T P r,, F-)
M| Y .r n] , y»|- ( MCTfiot;
P'?-f j o f TATf PN
"!. r
Figure E-4. Sample Computer Generated Map Used to
Review Chicago Validation Results
-------
Figure E-5 is an example of the statistical summary of valida-
tion results for a single Chicago station (90 and 100 percent values of
the frequency distribution are not shown in this example although they
are included in complete computer printout). Similar summaries were
generated for each observing location and for all stations combined.
Summaries were generated for the entire validation period and for
selected subsets of the period, e.g., all hours for which the wind
speed had some specified range.
The validation results are also available on punched cards.
The format for the punched card data is listed in Table E-2.
-------
I
00
STATIST
BEGINNI
ENDING
ICAL SUMMARY FOR 1 STATIONS
NG DATE 67. 1. 1. 0.
DATE 67. 1. 31. 23.
NUMBER
OF CASES
OBSERVED 673
PREDICTED 673
DRSRVD-PRtDICTED 673
STATION
SUM
0.22119E 02
0.48279E 02
-0.26162E 02
INDEX NUMBERS USED IN THIS RUN ARE
1
SUM OF STANDARD MEAN ABSOLUTE
SQUARES MEAN DEVIATION DIFFERENCE
0.29168E 01 0.32866E-01 0.57085E-01
0.14854E 03 0.71737E-01 0.46465E 00
0.14405E 03 -0.38874E-01 0.46135E 00 0.62249E-01
SLOPE
INTERCEPT
REGRESSI
B(
B(
ON COEFFICIENTS
1)= 0.0146
0)= 0.0318
DECILE 0
OBSERVED 0.0
PREDICTED 0.000
10 20
0.0 0.
0.005 0.
FREQUENCY DISTRIBUTION BY DECILES
30 40 50 60 70 80
0 0.0 0.010 0.010 0.010 0.030 0.050
007 0.009 0.012 0.014 0.017 0.027 0.042
OBSERVED
MINUS
PREDICTED -10.913
-0.049 -0.
027 -0.015 -0.011 -0.007 -0.004 0.000 0.011
-------
Table E-2. Punch Card Format for Validation Data
I. Format for St. Louis Validation Data at 10 Stations
Card
1
2
Columns
1-8
9-10
11-12
13-52
53-76
77-79
80
1-8
9-24
25-28
29-32
33-36
37-40
41
42-45
46-50
51-55
56-62
63-66
67-70
72-75
77-79
80
Format*
18
12
12
1014
614
3X
11
18
414
F4.1
F4.1
F4.3
F4.2
11
F4.1
15
F5.1
F7.3
F4.2
F4.1
F5.2
4X
11
Units
None
None
None
pg/m
ug/m
None
None
None
yg/m3
m/sec
m
None
Radians
None
°F
m
"Azimuth
None
0.01
in/hr
knots
in Hg
None
None
Description
Output record index number
Day of the week
= 10
Observed concentrations for stations
3, 15, 17, 23, 33, 4, 10, 12, 28
and 36, respectively
Calculated concentrations for sta-
tions 3, 15, 17, 23, 33 and 4,
respectively
Blank
= 1
Output record index number
Calculated concentrations for sta-
tions 10, 12, 28 and 36, respec-
tively
Wind speed for reference height
Wind speed for reference height
Wind profile exponent
Wind direction
Stability class index
Air temperature
Mixing layer ceiling
Wind turbulence statistics (a,)
Bulk Richardson number
Precipitation rate
Peak wind gust
Atmospheric pressure
Blank
= 2
(Continued)
-------
Table E-2. Punch Card Format for Validation Data (Concluded)
II. Format for Chicago Validation Data at 8 Stations
Card
1
2
Columns
1-10
11
12
13-44
45-76
77-80
1-8
9-24
25-80
Format*
IX
11
814
814
16X
Units
None
None
mg/m
mg/m
None
Description
See Part I above
Blank
= 8
Observed concentrations for stations
1 to 8, respectively
Calculated concentrations for sta-
tions 1 to 8, respectively
See Part I above
See Part I above
Blank
See Part I above
*Standard FORTRAN code notation.
-------
Appendix F
-------
Appendix F
SAMPLES OF SENSITIVITY DATA LISTINGS
This appendix describes and presents samples of the computer
printouts which were generated to permit review of the sensitivity
results. The total set of all printed results consists of several
hundred pages. Only samples of each type of printout are reproduced
here.
An output record was generated for each set of input values
considered in the sensitivity analysis. A listing of the concentrations
and input values for each of 5832 records was made in the format shown
in Figure F-l. A description of the computer printout identification
in Figure F-l is given in Table F-l. The sensitivity output records
were stored in disk files for subsequent statistical analysis and tabu-
lation of results.
One type of summary which was generated is illustrated in
Figure F-2. The mean concentration of all records with a given input
value was determined. The results in Figure F-2 are for the entire set
of sensitivity results. Similar summaries were generated from selected
subsets of the total set of results in order to evaluate special sensi-
tivity considerations.
Another type of listing which provides a convenient review of
model sensitivity results is illustrated in Figure F-3. This figure is
a partial listing of all comparisons of calculated concentrations in
which the only input changed is the mixing ceiling. Similar listings
were generated for all inputs considered in the sensitivity analysis.
-------
A quick review of this type of printout easily identifies the combina-
tions of input values (if any) for which "sensitive" results are generated
with input parameter.
A specialized type of summary was generated to examine the
effects of changes in wind direction input. An example of this special-
ized printout is given in Figure F-4.
-------
INDO TOT. CONC.
ISIGD
INDEX
CIGMX
PWIND
WSPD DECAY
NH
DLTA NPS RECEPTOR
-n
co
1
2 '"
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
' 20.
21
22
23
24
25
26
27
0.291E+02
0.304E+04
0.109E+03
0.291E+02
0.122E+05
0.106E+03
0.291E+02
0.462E+05
0.106E+03
0.360E+02
0.321E+04
0.127E+03
0.360E+02
0.485E + 0-4
0.124E+03
0.360E+02
0.868E+04
0.124E+03
0.420E+Q2
0.302E+04
0.191E+03
0.433E+02
0.358E+04
0.190E+03
0.433E+02
0.244E+04
0.190E+03
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
100.
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
0.150
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
381.
381.
381.
381.
381.
381.
381.
381.
381.
1524.
1524.
1524.
1524.
1524.
1524.
1524.
1524.
1524.
6096.
6096.
6096.
6096.
6096.
6096.
6096.
6096.
6096.
51
51
51
19
19
19
0
0
0
51
51
51
19
19
19
0
0
0
51
51
51
19
19
19
22366.
25146.
30707.
22366.
25146.
30707.
22366.
25146.
30707.
22366.
25146.
30707.
22366.
25146.
30707.
22366.
25146.
30707.
22366.
_^5_146.__
30707.
22366.
25146.
30707.
0 22366.
0 25146.
0
30707.
-------
Table F-l. Description of Computer Printout Identifications
Identification
Description
INDO
TOT. CONC.
ISIGD
INDEX
CIGMX
PWIND
WSPD
DECAY
NH
DLTA
NPS
RECEPTOR
Output record index number
3
Calculated concentration at receptor location, yg/m
Index to designate system of diffusion parameters used
Stability index for diffusion parameters
Mixing ceiling, m
Wind profile exponent
Wind speed at reference height, m/sec
Decay constant (inverse of mean decay time), sec
Number of area source heights
Spacing between area source grid points, m
Number of point sources
East-West coordinate of receptor location, m
-------
PARAMETER
VALUE MEAN CONCENTRATION
RECEPTOR LOCATION (M.,EAST)
RECEPTOR LOCATION (M.,EAST)
RECEPTOR LOCATION (M.,EAST)
NO. OF PUINT SOURCES
MO. OF POINT SOURCES
NO. OF POINT SOURCES
AREA SOURCE GRID SPACING (M.)
AREA SOURCE GRID SPACING (M.)
AREA SOURCE GRID SPACING (M.)
NO. AREA SOURCE EMISSION HGTS
DECAY CONSTANT
DECAY CONSTANT
WIND SPEED (M/SEC)
WIND SPEED (M/SEC)
WIND SPEED (M/SEC)
WIND PROFILE PARAMETER
MIXING CEILING (M.)
MIXING CEILING (M. )
DIFFUSION FUNCTIONS (TYPE)
DIFFUSION FUNCTIONS (TYPE)
DIFFUSION FUNCTIONS (TYPE)
0.2237E+05
0.2515E+05
0.3071E+05
0.5100E+02
0.1900E+02
0.0
0.3810E+03
0.1524E+04
0.6096E+04
0.1000E+01
0.0
0.3851E-03
0.2000E+01
0.6000E+01
0.1800E+02
0.1500E+00
0.1000E+03
0.5000E+03
0.4000E+01
0.1000E+01
0.2000E+01
0.6002E+01
0.3856E+03
0.3962E+03
0.2371E+03
0.2616E+03
0.2891E+03
0.2769E+03
0.2735E+03
0.2373E+03
0.2626E+03
0.3921E+03
0.1331E+03
0.5112E+03
0.1983E+03
0.7831E+02
0.2626E+03
0.3455E+03
0.1797E+03
0.3520E+03
0.2188E+03
0.2169E+03
Figure F-2. Summary of Sensitivity Results by Input Parameter
-------
CO N C E.NJ.RA LION S_EOR
M.
I
CT>
DIFFUS.
FN.
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
MIXING
100
11.
1259.
2061.
11.
1259.
1849.
11.
1259.
1750.
15.
1211.
1987.
15.
1238.
2811.
15.
1238."
2711.
25.
829.
1402.
48.
1205.
1738..
48.
1929.
2621.
7.
1146.
41.
7.
1146.
39.
7.
1146.
35.
9.
785.
36.
CEILH
500
12.
1635.
1526.
12.
JLfclSj.
1376.
12.
1635.
132 1 .
15.
_12iZ*
1469.
.15..
1239.
_2124.
15.
_1Z19_*
2069.
25.
830.
J.^2^
48.
1207.
1269.
48.
1960.
1899.
7.
1143.
36.
7.
1143.
34.
7.
1143.
32.
9.
784.
31.
RECEPTOR
LOCATION
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
UPWIND
CENTER
DOWNWIND
NO. OF
POINT
SOURCES
51
51
51
19
19
19
0
0
0
51
51
51
19
19
19
0
0
0
51
51
51
19
19
19
0
0
0
51
51
51
19
19
19
0
0
0
51
51
51
AREA
GRID
MI
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
1.00
1.00
1.00
1^00
1.00
1..QQ
1.00
1.00
1.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
0.25
0.25
0.25
(L. 23
0.25
0.25
0.25
0_.2.5
0.25
1.00
1.00
1.00
NO.
AREA
HGTS
1
1
1
1
1
1
1
1
1
1
1
OF DECAY
CONST
/SEC
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
1 0.000
1 0.000
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000385
0.000385
0.000385
0.000385
0.000385
0.000385
0.000385
0.000385
0.000385
0.000385
0.000385
0.000385
WIND
SPEED
M/S
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
WIND
POWER
LAW
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
-------
AREA SOURCE, WIND AZIMUTH IS 349
LOCA-
TION
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
UPWIND
CENTER
DNWIND
STA-
BILITY
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
STABLE
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
NEUTRAL
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
UNSTAB
MIXING
CEIL.
(M)
100.
100.
100.
100.
100.
100.
100.
100.
100.
500.
500.
500.
500.
500.
500.
500.
500.
500.
100.
100.
100.
100.
100.
100.
100.
100.
100.
500.
500.
500.
500.
500.
500.
500.
500.
500.
100.
100.
100.
100.
100.
100.
100.
100.
100.
500.
500.
500.
500.
500.
500.
500.
500.
500.
WIND
SPEED
(M/S)
2.
2.
2.
6.
6.
6.
18.
18.
18.
2.
2.
2.
6.
6.
6.
18.
18.
18.
2.
2.
2.
6.
6.
6.
18.
18.
18.
2.
2.
2.
6.
6.
6.
18.
18.
18.
2.
2.
2.
6.
6.
6.
18.
18.
18.
2.
2.
2.
6.
6.
6.
18.
18.
18.
UNITS ARE MICROGRAMS/CU.M.
MODEL
CONCEN- ABS. ERROR BY WIND ERROR
TRATION 3 DEG 10 DEC 45 DEC
MEAN
12,
1576,
1890.
4,
525,
630,
1,
175,
210,
12,
1566,
1422,
4,
522,
474.
1,
174,
158,
9,
1063,
1879,
3,
354,
626,
1,
118,
209,
3,
571,
382,
1,
190,
127,
0,
63,
42,
9,
1065,
1879,
3,
355,
626,
1,
118,
209,
2,
382,
379,
1,
127,
126,
0,
42,
42,
377,
1,
32,
489,
0,
11,
163,
0,
4,
54,
1,
29,
360.
0,
10.
120,
0.
3.
40,
1.
33.
487,
0,
11,
162,
0.
4.
54,
0.
68,
99.
0,
23.
33,
0.
8.
11,
1,
31.
487.
0,
10.
162.
0.
3,
54,
0,
82,
98.
0.
27.
33,
0,
9,
11,
61,
2,
106,
1247,
1,
35.
416,
0,
12,
139,
2,
107,
935.
1,
36,
312,
0,
12.
104,
I,
60.
1243.
0.
20.
414.
0,
7.
138.
1.
30.
252,
0,
10.
84,
0.
3.
28,
1.
58.
1243.
0.
19.
414.
0,
6.
138.
0.
26,
250,
0,
9.
83,
0,
3,
28,
149,
66,
526,
1823,
22,
144,
607,
7,
48,
202,
59,
453.
1363,
20.
151,
454.
7.
50.
151,
65,
245.
1815.
22.
82.
605,
7,
27,
202.
14.
173,
367,
5.
58.
122.
2,
19.
41,
65.
246,
1815.
22.
82.
605,
7,
27.
202,
13.
83,
365.
4.
28,
122,
1.
9,
41,
255,
Figure F-4. Sample of Sensitivity Results for
Changes in Wind Direction Input
-------