-------
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
-------