United States
        Environmental Protection
        Agency
Office of Air Quality
Planning and Standards
Research Triangle Park. NC 27711
EPA-454/R-92-014
October 1992
vxEPA
     COMPARISION OF A REVISED
      AREA SOURCE ALGORITHM
              FOR THE
    INDUSTRIAL SOURCE COMPLEX
         SHORT TERM MODEL
       AND WIND TUNNEL DATA

-------
                                  EPA-454/R-92-014
Vf
^
-4
O
        COMPARISION OF A REVISED
        AREA SOURCE ALGORITHM
                  FOR THE
       INDUSTRIAL SOURCE COMPLEX
            SHORT TERM MODEL
         AND WIND TUNNEL DATA
               U.S. Environments' F,v...-:;v->n Agency
               Region 5, Library :p[ -12J)
                                        rf'V
              Office Of Air Quality Planning And Standards
                  Office Of Air And Radiation
               U. S. Environmental Protection Agency
                Research Triangle Park, NC 27711

                    October 1992

-------
This report  has been reviewed by the  Office  Of Air Quality Planning And Standards,  U. S.
Environmental Protection Agency, and has been approved for publication.   Any mention of trade
names or commercial products is not intended to constitute endorsement or recommendation for use.
                                    EPA-454/R-92-014

-------
                             PREFACE

     The ability to accurately estimate pollutant concentration
due to atmospheric releases from area sources is important to the
modeling community, and is of special concern ^for Superfund where
emissions are typically characterized as area sources.
Limitations of the Industrial Source Complex  (ISC2)  model  (dated
92273) algorithms for modeling impacts from area sources,
especially for receptors located within and nearby the area, have
been documented in earlier studies.  An improved algorithm for
modeling dispersion from area sources has been developed based on
a numerical integration of the point source concentration
function.  Information on this algorithm is provided in three
interrelated reports.

     In the first report (EPA-454/R-92-014),  an evaluation of the
algorithm is presented using wind tunnel data collected in the
Fluid Modeling Facility of the U.S. Environmental Protection
Agency.  In the second report (EPA-454/R-92-015),  a sensitivity
analysis is presented of the algorithm as implemented in the
short-term version of ISC2.  In the third report
(EPA-454/R-92-016), a sensitivity analysis is presented of the
algorithm as implemented in the long-term version of ISC2.

     The Environmental Protection Agency must conduct a formal
and public review before the Agency can recommend for routine use
this new algorithm in regulatory analyses.  These reports are
being released to establish a basis for reviews of the
capabilities of this methodology and of the consequences
resulting from use of this methodology in routine dispersion
modeling of air pollutant impacts.  These reports are one part of
a larger set of information on the ISC2 models that must be
considered before any formal changes can be adopted.
                               111

-------
                        ACKNOWLEDGEMENTS

     This report has been prepared by Pacific Environmental
Services, Inc., Research Triangle Park,  North Carolina.  This
effort has been funded by the U.S. Environmental Protection
Agency under Contract No.  68D00124, with Jawad S. Touma as Work
Assignment Manager.  The wind tunnel data were collected at the
EPA Fluid Modeling Facility, with William Snyder as Project
Director.
                                IV

-------
                             CONTENTS


PREFACE	iii

ACKNOWLEDGEMENTS	iv

FIGURES	    vii

TABLES  	 X

1.0 INTRODUCTION  	   1

2 .0 SUMMARY OF THE WIND TUNNEL DATA BASE	   3
     2.1 Base Case Wind Tunnel Data Description	   3
     2.2 Rotated Rectangular Area Source Scenarios   	   4
     2.3 Circular Area Source Scenario  	   8
     2.4 Other Related Data	   8

3 .0 METHODOLOGY OF THE EVALUATION	14
     3.1 Setting Up The ISCST2 Simulations	14
     3.2 Physical Analysis   	   17
     3.3 Statistical Analysis 	   19

4.0 RESULTS FOR THE BASE CASE SCENARIO	22
     4.1 Results Of The Physical Analysis	22
     4.2 Results Of The Statistical Analysis	27

5.0 RESULTS FOR THE 45 DEGREE ROTATED SCENARIO	31
     5.1 Results Of The Physical Analysis	31
     5.2 Results Of The Statistical Analysis   	   36

6.0 RESULTS FOR THE 90 DEGREE ROTATED SCENARIO	40
     6.1 Results Of The Physical Analysis	40
     6.2 Results Of The Statistical Analysis   	   40

7.0 RESULTS FOR THE CIRCULAR SOURCE SCENARIO   	   45
     7.1 Results Of The Physical Analysis	45
     7.2 Results Of The Statistical Analysis   	   49

8.0 THE WIND TUNNEL DISPERSION PARAMETERS  	   52

9.0 CONCLUSIONS	60

10.0 REFERENCES	62

APPENDIX A. POINT-TO-POINT COMPARISON RESULTS FOR THE BASE CASE
     SCENARIO

APPENDIX B. POINT-TO-POINT COMPARISON RESULTS FOR THE 45-DEGREE
     ROTATED CASE SCENARIO
                                v

-------
APPENDIX C. POINT-TO-POINT COMPARISON RESULTS FOR THE 90-DEGREE
     ROTATED CASE SCENARIO

APPENDIX D. POINT-TO-POINT COMPARISON RESULTS FOR THE .CIRCULAR
     CASE SCENARIO

APPENDIX E. RESULTS OF STATISTICAL ANALYSES IN TABULAR FORM
                                VI

-------
                             FIGURES

Figure                                                       Page

2-1. Three-dimensional schematic diagram of wind tunnel
     experiment showing the three measurement planes for the
     Base Case scenario	   5

2-2. Comparison of wind tunnel lateral dispersion parameter
     with Pasquill-Gifford ay  values	12

2-3. Comparison of wind tunnel vertical dispersion parameter
     with Pasquill-Gif ford az  values	13

3-1. An Example Of The ISCST2 Input File For The Base Case.  .  14

3-2. An example of the graphical presentation of the
     physical analysis	18

3-3. An example of the graphical presentation of the
     statistical analysis	21

4-1. Comparison of lateral profile of normalized
     concentration for the Base Case at a downwind distance
     of 120 meters from the center of the area	24

4-2. Comparison of vertical profile of normalized
     concentration for the Base Case at a downwind distance
     of 120 meters from the center of the area	25

4-3. Comparison of lateral profile of normalized
     concentration for the Base Case at a downwind distance
     of 840 meters from the center of the area and a
     receptor elevation of 55.8 meters above ground	26

4-4. Fractional Bias of normalized concentration as a
     function of downwind distance for the Base Case using
     dispersion parameters fitted to wind tunnel data.   ...  29

4-5. Fractional Bias of normalized concentration as a
     function of receptor height for the Base Case using
     dispersion parameters fitted to wind tunnel data.   ...  30

5-1. Comparison of lateral profile of normalized
     concentration for the 45  Degree Rotated Case at a
     downwind distance of 340  meters from the center of the
     area	33

5-2. Comparison of vertical profile of normalized
     concentration for the 45  Degree Rotated Case at a
     downwind distance of 340  meters from the center of the
     area	34
                               VI1

-------
5-3. Comparison of lateral profile of normalized
     concentration for the 45 Degree Rotated Case at a
     downwind distance of 840 meters from the center of the
     area and a receptor elevation of 55.8 meters above
     ground	'.	35

5-4. Fractional Bias of normalized concentration as  a
     function of downwind distance for the 45-Degree Rotated
     Case using dispersion parameters fitted to wind tunnel
     data	38

5-5. Fractional Bias of normalized concentration as  a
     function of receptor height for the 45-Degree Rotated
     Case using dispersion parameters fitted to wind tunnel
     data	39

6-1. Comparison of lateral profile of normalized
     concentration for the 90 Degree Rotated Case at a
     downwind distance of 360 meters from the center of the
     area	42

6-2. Comparison of vertical profile of normalized
     concentration for the 90 Degree Rotated Case at a
     downwind distance of 360 meters from the center of the
     area	43

6-3. Fractional Bias of normalized concentration as  a
     function of downwind distance for the 90-Degree Rotated
     Case using dispersion parameters fitted to wind tunnel
     data	44

7-1. Comparison of lateral profile of normalized
     concentration for the Circular Case at a downwind
     distance of 204 meters from the center of the area.   .  .  46

7-2. Comparison of vertical profile of normalized
     concentration for the Circular Case at a downwind
     distance of 204 meters from the center of the area.   .  .  47

7-3. Comparison of lateral profile of normalized
     concentration for the Circular Case at a downwind
     distance of 840 meters from the center of the area and
     a receptor elevation of 55.8 meters above ground.  ...  48

7-4. Fractional Bias as a function of downwind distance for
     the Circular Case using dispersion parameters fitted to
     wind tunnel data	50

7-5. Fractional Bias as a function of receptor height for
     the Circular Case using dispersion parameters fitted to
     wind tunnel data	51
                              VI11

-------
8-1.  Comparison of the wind tunnel data and power-law fit
     for 0y	54

8-2.  Comparison of the wind tunnel data and power-law fit
     for az	55

8-3.  Comparison of the wind tunnel data and ISCST2 results
     at a downwind distance of 240 meters for the Base Case
     with various dispersion parameters	56

8-4.  The mean fractional bias (MFB)  of normalized
     concentration as a function of downwind distance for
     the Base Case with various dispersion parameters.  ...  57

8-5.  Comparison of the wind tunnel data and ISCST2 results
     at a downwind distance of 240 meters for the Circular
     Source Case with various dispersion parameters	58

8-6.  The mean fractional bias (MFB)  of normalized
     concentration as a function of downwind distance for
     the Circular Source Case with various dispersion
     parameters	59
                               IX

-------
                             TABLES

Table                                                        Page

2-1. Wind Tunnel Measurement Scenarios and Filenumbers for
     the Base Case Rectangle	   6
2-2. Wind Tunnel Measurement Scenarios and Filenumbers for
     the 45-Degree Rotated Rectangle  	 .
2-3. Wind Tunnel Measurement Scenarios and Filenumbers for
     the 90-Degree Rotated Rectangle  	 .
2-4. Wind Tunnel Measurement Scenarios and Filenumbers for
     the Circular Source	10

-------
                         1.0  INTRODUCTION

     Previous studies indicate that the currently implemented
ISCST2 area source algorithm, based on a finite line segment
approximation, produces unrealistic predictions of the
concentration distribution, especially for receptors located
within and nearby the area source  (EPA, 1989) .   Based on the same
report, the integrated line source algorithm for modeling impacts
from area sources used in the Fugitive Dust Model (FDM)  and the
Point, Area, and Line Source  (PAL) model appears to provide a
more adequate treatment of near-source geometry and reasonable
far-field behavior.  Based on these performance evaluations and
limited field data, the integrated line source algorithm has been
recommended as a candidate to substitute for the current ISCST2
area source algorithm.

     Under EPA Contract No. 68D00124,  Work Assignment No. 27, an
improved area source algorithm for ISCST2 was developed and
tested.  A sensitivity analysis of the algorithm was performed,
comparing it to the current regulatory algorithm in ISCST2 based
on a finite line segment approach  (EPA, 1992a).   The new
algorithm is based on a numerical integration of the point source
concentration function over the area,  and makes use of the
Romberg integration technique (Press,  et al, 1986)  to improve the
efficiency of the calculations.   The new numerical integration
algorithm has been shown to be essentially equivalent to the
convergent modes of the PAL and FDM integrated line source
algorithms  (EPA, 1992b).  The numerical integration algorithm,
which is based on modifications to the PAL model code,  was then
incorporated into the ISCST2 model.  The new algorithm has been
shown to perform very well in terms of efficiency and in terms of
the reasonableness of the results.

     In order to recommend the new algorithm for regulatory
modeling applications, it is desirable to compare the model
predictions to measured results.  Field studies of impacts within
and nearby area sources are quite few and limited in their scope.
The time and resources required to conduct a detailed field
experiment make a wind tunnel experiment an attractive
alternative means of collecting such comparison data.  In a wind
tunnel simulation, the meteorological conditions,  source-receptor
relationship and the time scales of the experiments can be easily
controlled. ' Furthermore, a series of modeling scenarios can be
set up intentionally to simulate different source-receptor
geometries, and a relatively large number of measurements can be
taken to provide a detailed picture of the concentration
distribution in three dimensions.

     In this report,  the improved ISCST2 area source algorithm is
evaluated using the wind tunnel data collected by the EPA Fluid
Modeling Facility and presented in an EPA data report (Snyder,
1991).  Both a qualitative physical analysis and a quantitative
statistical analysis are presented.  The ISCST2 model has been

-------
set up to predict the concentration distribution for each of the
equivalent wind tunnel experiment cases.  For the physical
analysis, the results of the model predictions are plotted
against their wind tunnel experiment counterparts in point-to-
point comparisons.  These point-to-point comparisons provide a
qualitative assessment of the ability of the new algorithm to
model the dispersion from area sources.  In the statistical
analysis, some of the statistical methods recommended by Cox
(1988) are used to provide a quantitative comparison of model
predicted to wind tunnel measured concentrations.  Specifically,
the mean fractional bias and the standard deviation of the
fractional bias across the plume are calculated as a function of
downwind distance and as a function of receptor elevation for
each of the wind tunnel scenarios.

     Section 2 of this report provides a summary of the wind
tunnel experiment.  The methodology of the model evaluation is
described in Section 3.  Sections 4 to 8 present the results of
the evaluation study.  Results for each of the wind tunnel
experiment scenarios are presented separately, beginning with the
rectangular base case.  Appendix A shows the complete results of
the physical analysis  (point-to-point comparisons)  for the "base
case" rectangle.  Appendix B shows the results for the "45-degree
rotated case,"  Appendix C shows the results for the "90-degree
rotated case,"  and Appendix D shows the results for the
"circular source case."  The results of the statistical analysis
are presented in tabular form in Appendix E for all four
scenarios.

-------
             2.0 SUMMARY OF THE WIND TUNNEL DATA BASE

     Wind tunnel data  simulating area source releases typical  of
a  superfund  site have  recently been  collected in the Fluid
Modeling Facility  (FMF) of EPA's Atmospheric Research and
Exposure Assessment Laboratory  (EPA-AREAL).  This wind  tunnel
experiment was  summarized by Snyder  (1991).  A typical  superfund
site is simulated by a ground-level  rectangular source  with  an
aspect ratio of 3 to 1.  The dimensions of the area source are
1200mm x 400mm, which  translate to full-scale dimensions of  720m
x  240m using a  scaling ratio of 600:1.  The source is oriented
with its long dimension perpendicular to the wind flow  with  a
free-stream wind speed of 4 m/s.  The surface of the wind tunnel
is covered by small irregularly shaped objects  (called  Sanspray)
providing a  surface roughness of about 0.2 meters in full scale
units.  This wind tunnel scenario is referred to as the "Base
Case."  Additional wind tunnel scenarios were conducted for
source orientation effects with the  same rectangular source
rotated by 45 degrees  and 90 degrees counterclockwise from the
base-case orientation.  In addition, a circular area source  of
equivalent size was used to examine  the effects of source shape.

     One of the purposes of this study is to examine the
performance of  the algorithm for predicting concentrations inside
and nearby the  area source.  Therefore,  lateral (crosswind)
profiles were measured at several distances relative to the
upwind edge of  the area for each scenario.  The measurement
locations correspond to 0.25 source width upwind from the center
of the area source, the center of the area, 0.25 source width
downwind from the center of the source,  the downwind edge of the
area source, and additional locations out to several source
widths downwind from the center of the area source.   The data
collected for each case is summarized in the following sections.

2.1 Base Case Wind Tunnel Data Description

     The first phase of the evaluation effort focused on the
"base case," which is  the measurement program of the rectangular
source oriented with its long dimension perpendicular to the wind
in a simulated atmospheric boundary layer with a 0.2 meter (full
scale)  roughness length and with a free-stream wind speed of 4
m/s.  All of the distances and measurements presented in this
report for the wind tunnel experiment reflect full scale units.
As noted above,  wind tunnel measurements were convertad to full-
scale units using a scaling ratio of 600:1.  The wind tunnel
experiment also included further analyses where che base case
wind speed was halved  and doubled,  and with a lower surface
roughness for the 4 m/s case.   The results of the high and low
wind speed cases showed that the normalized concentrations (\\i/Q)
were independent of wind speed.   The ground-level results for  the
case with lower surface roughness were slightly higher than the
base case results,  as  expected.

-------
     For the Base Case, there are three groups of wind tunnel
data files, corresponding to a series of profiles taken from
three basic measurement planes.  Figure 2-1 provides a three-
dimensional schematic diagram illustrating the three measurement
planes for the Base Case scenario.  Table 2-1 also describes the
various measurements taken,  with a cross reference to the FMF
area source filenumbers.

     The first group of data files is for the crosswind
measurements taken in the ground-level horizontal (XY) plane.
They contain all the measurements at a total of 8 downwind
distances and include 30 to 50 crosswind points for each downwind
distance.  All of the receptors in the group have a uniform
height above ground (i.e., flagpole height) of 1.8 meters.  They
are located at downwind distances of -60, 0, 120, 240, 360, 840,
1800 and 2580 meters in full scale units, relative to the center
of the area source.  The second group of data files is for the
measurements taken in the centerline vertical (XZ) plane.  It
includes data for the same 8 downwind distances mentioned above
with 20 to 40 vertical points for each of the downwind distances.
The last group of data files is for the measurements in the
vertical crosswind (YZ) plane at a downwind distance of 840
meters.  The data points are distributed at 12 flagpole heights
with 17 crosswind points for each of these heights.

     These source-receptor relationships are consistent with the
main goal of the study, which is to demonstrate the concentration
distribution inside,  nearby and downwind from a typical ground-
level area source.

2.2 Rotated Rectangular Area Source Scenarios

     To evaluate the effects of source orientation,  the base case
rectangle was rotated 45 degrees and 90 degrees counterclockwise.
For the case with the rectangle rotated 45 degrees,  there are
also three groups of data files.  These measurements are
summarized in Table 2-2.   The first group is for the measurements
taken in the ground-level horizontal (XY) plane  (see Figure 2-1).
They contain all the measurements at a total of 8 downwind
distances and include 30 to 50 crosswind points for each downwind
distance.  They are located at downwind distances of -170, 0,
170, 340, 480, 600, 840,  and 1800 meters in full scale units,
relative to the center of the area.  The second group is for the
measurements in the csntsrline vertical  (XZ) plane.   It includes
data for the same 8 downwind distances mentioned before with 20
to 40 vertical points for each of the downwind distances.  Tlie
last group of data files is for the measurements in the vertical
crosswind  (YZ) plane at a downwind distance of 840 meters.  The
data points are distributed at 12 flagpole heights with 17
crosswind points for each of these heights.

-------
        NOT TO SCALE
    /-60 0  120 240 360
  -120
                                           YZ Plane
         840

    Downwind Distance C
                                         1800
                                                           XZ Plane
                                                  2580
Figure 2-1.
Three-dimensional schematic diagram of  wind tunnel
experiment  showing the  three measurement  planes
for the Base Case scenario.   Dashed lines indicate
approximate locations of measurement profiles.

-------
                   Table 2-1.
Wind Tunnel Measurement Scenarios and Filenumbers
           for the Base Case Rectangle
Type of Profile

Long. (y=0, z=3mm (1.8m))
Laterals (z=3mm (1.8m))

V
C


^
3



Verticals (y=0mm (Om) )









Laterals (x=1400mm (840m))
















X
mm (m)
-100 (-60)
0 (0)
200 (120)
400 (240)
600 (360)
1400 (840)
3000 (1800)
4300 (2580)
X
mm (m)
-100 (-60)
0 (0)
200 (120)
400 (240)
600 (360)
1400 (840)
3000 (1800)
4300 (2580)
z
mm (m)
3 (1.8)
18 (10.8)
33 (19.8)
48 (28.8)
63 (37.8)
78 (46.8)
93 (55.8)
108 (64.3)
123 (73.8)
138 (82.8)
153 (91.8)
168 (100.8)
AREAMC
Filenumber
19
AREAMC
Filenumber
5, 18
4,17
16
1,15
9, 10, 11
7,14
6, 12
13
AREAMC
Filenumber
27
26
25
24
23
22
21
20
AREAMC
Filenumber
800
806
801
807
802
808
803
309
804
810
805
811

-------
                    Table  2-2.
Wind Tunnel Measurement Scenarios and Filenumbers
       for the 45-Degree Rotated Rectangle
Type of Profile

Long. (y=0, z=3mm (1.8m))
Laterals (z=3mm (1.8m))



X

/^^^^^^^^^^^^^^P^ r en
^^^^^^^^^^^^^^




_PO
Verticals (y=0mm (Om) )

•







(y=-275mm (-165m) )
(y=-238mm (-143m) )
Laterals (x=1400mm (840m) )
















X
mm (m)
-283 (-170)
0 (0)
283 (170)
566 (340)
o /-\ /~i / A f\ r\ \
oUU (*"±o(j )
1000 (600)
1400 (840)

3000 (1800)

X
mm (m)
-283 (-170)
0 (0)
283 (170)
566 (340)
800 (480)
1000 (600)
1400 (840)
3000 (1800)
800 (480)
1000 (600)
z
mm (m)
3 (1.8)
33 (19.8)
63 (37.8)
93 (55.8)
123 (73.8)
153 (91.8)
183 (109.8)
213 (127.8)
243 (145.8)
273 (163.8)
303 (181.8)
333 (199.8)
AREAMC
Filenumber
69
AREAMC
Filenumber
58
57
56
55
r- A
b4
53
52

51

AREAMC
Filenumber
68
67
66
63
62
61
60
59
65
64
AREA
Filenumber
820
821
322
823
824
825
826
827
828
829
830
831

-------
     For the case with the rectangle rotated 90 degrees
counterclockwise (i.e., with the short dimension oriented
perpendicular to the wind flow),  there are only two groups of
data files, which are described in Table 2-3.  These correspond
to the ground-level horizontal (XY) plane and the centerline
vertical (XZ) plane.  The source-receptor relationship is similar
to the 45 degree case except that the 8 downwind distances are
changed to -180, 0, 180, 360, 480,  600, 840,  and 1800 meters,
relative to the center of the area.

2.3 Circular Area Source Scenario

     To examine the effects of source shape,  a circular source
of equivalent area was also used.  The circular source has a
diameter of 470 meters, and was examined in the same boundary
layer used in the base case.

     For the circular source, there are three groups of data
files, which are described in Table 2-4.  The first group is for
the measurements taken in ground-level horizontal (XY)  plane.  It
contains all the measurements at a total of 9 downwind distances
and includes 30 to 50 crosswind points for each downwind
distance.  All the receptors in this group have an uniform
flagpole height of 1.8 meters.   They are located at downwind
distances of -102,  0, 102, 204, 240, 360,  480,  840,  and 1800
meters, relative to the center of the area source.  The second
group is for the measurements in the centerline vertical (XZ)
plane.  It includes data for the same 9 downwind distances
mentioned above with 20 to 40 vertical points for each of the
downwind distances.  The last group of data files is for the
measurements in the vertical crosswind  (YZ) plane at the downwind
distance of 840 meters.  The data points are distributed at 12
flagpole heights with 17 crosswind points for each of these
heights.

2.4 Other Related Data

     As part of the wind tunnel experiment, the lateral and
vertical dispersion parameters were determined as a function of
downwind distance for a ground-level point source.  While the
wind tunnel boundary layer is characterized by neutral stability,
i.e., there is no surface heating present, the dispersion
parameters do not correspond precisely with the Pasquill-Gifford
 (PG) dispersion parameters used by the ISCST2 oicael for 3tabili;y
class D  (neutral).   This is because the surface roughness in the
wind, tunnel is somewhat: different: chan the surface rougnness en
which the PG dispersion curves are based.   The results of the
point source evaluation show that the dispersive behavior of the
simulated atmospheric boundary layer in the wind tunnel is near
the PG stability class C  (slightly unstable)  close to the source,
and tends toward the PG stability class D  (neutral)  farther
downwind.

-------
                   Table  2-3.
Wind Tunnel Measurement Scenarios and Filenumbers
       for the 90-Degree Rotated Rectangle
Type of Profile

Long. (y=0, z=3mm (1.8m))
Laterals (z=3mm (1.8m))

1

X •* - 3QQ 	 ;


1





•
3?^P™$§&gS r iprp
SjSSSSSKSSSK^ k id'jj
SsSeSssS 8»«ss f" D"l
*jggg$K$ggggggga v. u j
B
s SSSSSwwSSSsSS
g ^^^W^^OT C^SO!)
8 ^K^SS^a r -ici-o

Verticals (y=0mm (Om))












X
mm (m)
-300 (-180)
0 (0)
300 (180)
600 (360)
800 (480)
"t f^ f\ f\ I /• /^ r\ \
1000 (600)
1400 (840)
3000 (1800)


X
mm (m)
-300 (-180)
0 (0)
300 (180)
600 (360)
800 (480)
1000 (600)
1400 (840)
3000 (1800)
AREAMC
Filenumber
70
AREAMC
Filenumber
78
77
76
75
74
•"•? "^
73
72
71


AREAMC
Filenumber
86
85
84
86
82
81
80
79

-------
                   Table  2-4.
Wind Tunnel Measurement Scenarios and Filenumbers
             for the Circular Source
Type of Profile

Long. (y=0, z=3mm (1.8m))
Laterals (z=3mm (1.8m))

i
i


v -|-m r-10?->

^^^^^^^^^^^^^^^^^^^^^^^ rrn
S^^^^^^^^^i^^^^^^^^^^^^^^^

IT) ^^^^^^^^^^^^^^^^^^^^^^^^^ f "I1"10"1)


340 ^ocsjsjjsjgjajjis^ t j
i
Verticals (y=0mm (Om))










Laterals (x=1400mm (840m))
















X
mm (m)


-170 (-102)
0 (0)
170 (102)
340 (204)
400 (240)
600 (360)
800 (480)
1 yl /"» A t Q A f\ \
14UU ( o4U )
3000 (1800)

X
mm (m)
-170 (-102)
0 (0)
170 (102)
340 (204)
400 (240)
600 (360)
800 (480)
1400 (840)
3000 (1800)
z
mm (m)
3 (1.8)
33 (19.8)
63 (37.8)
93 (55.8)
123 (73.8)
153 (91.8)
183 (109.8)
213 (127.8)
243 (145.8)
273 (163.8)
303 (181.8)
333 (199.8)
AREAMC
Filenumber
69
AREAMC
Filenumber


96
95
94
88
93
92
91
Q A
yu
89

AREAMC
Filenumber
105
104
103
102
101
100
99
98
97
AREA
Filenumber
840
841
842
843
344
845
346
847
848
849
850
851
                        10

-------
     Figures 2-2 and 2-3 show the comparison between the
dispersion parameters based on wind tunnel measurements and the
PG class C and D curves for lateral (cry)  and vertical  (crz)
dispersion, respectively.  These comparisons are fairly typical
of the wind tunnel boundary layer.  The differences are
presumably due to the slightly rougher surface (higher
dispersiveness near the area source),  and the lack of large
scale, low frequency wind fluctuations in the wind tunnel
boundary layer.  The wind tunnel derived dispersion parameters
exhibit a tendency toward weaker dispersiveness relative to the
PG curves farther downwind.  In order to focus the evaluation on
the performance of the numerical integration algorithm and avoid
overshadowing the interpretation of the results with differences
due to the representativeness of the PG dispersion parameters to
the wind tunnel boundary layer,  the ISCST2 model results used in
the evaluation are based on a power-law fit to the wind-tunnel
derived dispersion parameters.  This approach is discussed in
more detail in Section 8.

     As noted in the wind tunnel data report (Snyder,  1991),  the
lateral dispersion parameter  (ay)  is,  for all practical  purposes,
independent of the elevation at which it is measured.   This is a
usual assumption in dispersion models, such as ISC2,  but data are
seldom available to verify the assumption.
                                11

-------
  Sigma Y £rrT)

  1000r	
   100
    10 -
     100
             1000
        Downwind Distance
                                                          PG Class  C
                                                          PG Class  D
                                                        Wind Tunnel  Data
10000
Figure  2-2.
Comparison of  wind tunnel lateral  dispersion
parameter with Pasquill-Gifford  ay values.
                                   12

-------
  Sigma Z

  100Q
   100
    10
                                                          PG Class C
                                                          PG Class D
                                                        Wind Tunnel Data
     100
             1000
        Downwind  Distance
                                                      10000
Figure  2-3.
Comparison of wind tunnel  vertical  dispersion
parameter with  Pasquill-Gifford crz values.
                                   13

-------
                     3.0  METHODOLOGY OF THE EVALUATION

3.1 Setting  Up The  ISCST2  Simulations

       Based on the scenarios of  the wind  tunnel  experiment,
described in Section 2  above, the  input  files for the  ISCST2
model  with the improved area source algorithm were  set  up
accordingly.   The detailed procedures  for setting up an input
file  for  ISCST2  can  be  found in Volume I  of  the ISC2 User's Guide
 (EPA,  1992c).   Figure 3-1  shows an example of an ISCST2 input
file  used for one of the simulations.
    CO STARTING
    CO TITLEONE Simulation of The Wind Tunnel Experiment:  Case I:F
    CO MODELOPT DFAULT CONC  RURAL
    CO AVERT IME  1
    CO FLAGPOLE  1.8
    CO POLLUTID OTH
    CO RUNORNOT RUN
    CO FINISHED

    SO STARTING
    **        Srcid Srctyp   Xs    Ys    Zs
    SO LOCATION    1  AREA -120.  -360.    .0000
    **        Srcid       Qs    Hs  Xinit   Yinit
    SO SRCPARAM    1      5.79   0.0  240.   720.
    SO EMISUNIT    1. (GRAMS/SEC/M**2)        grams/cubic-meter
    SO SRCGROUP ALL
    SO FINISHED

    RE STARTING
    RE GRIDCART C1 STA
    RE GRIDCART C1 XPNTS -60. 0. 120. 240. 360. 840. 1800. 2580.
    RE GRIDCART C1 YPNTS -600. -576. -552. -528. -504. -480. -456. -432. -408. -384.
    RE GRIDCART C1 YPNTS -360. -336. -312. -288. -264. -240. -216. -192. -144. -120.
    RE GRIDCART C1 YPNTS  -96. -72. -48.  -24.   0.   24.   48.  72.   96.
    RE GRIDCART C1 YPNTS  120. 144. 192.  216. 240.  264.  288. 312.  336. 360.
    RE GRIDCART C1 YPNTS  384. 408. 432.  456. 480.  504.  528. 552.  576. 600.
    RE GRIDCART C1 END
    RE FINISHED

    ME STARTING
    ME INPUTFIL  cardc.met
    ME ANEMHGHT    10.0 METERS
    ME SURFDATA  99999 1990          SURFNAME
    ME UAIRDATA  99999 1990          UAIRNAME
    ME WINDCATS   1.54   3.09   5.14   8.23  10.80
    ME FINISHED

    OU STARTING
    QU DAYTABLE ALLAVE
    OU POSTFILE 1 ALL PLOT WINDT1F.PLT
    OU FINISHED
Figure 3-1.     An Example  Of The  ISCST2  Input File  For  The  Base
                   Case.
                                         14

-------
3.1.1 Modeling Options

     For purposes of this evaluation, the regulatory default
option was specified, which forces the use of the recommended
modeling options.  The rural dispersion option was selected,
although this switch had no real effect since the dispersion
parameters were based on a fit to wind tunnel data as discussed
in Section 2.4.  The model was setup to calculate 1-hour
averages, which are comparable to the wind tunnel measurements
based on an assumption of steady state conditions.  According to
the wind tunnel data report, all of the ground-level receptors
should.be modeled with a height above ground  (flagpole height) of
1.8 meters in full scale units.  The pollutant type was set to
other.

3.1.2 Source-Receptor Relationship

     Section 2 provided a summary of the various scenarios of the
wind tunnel experiment. To simulate the area source setup in the
base case of the wind tunnel experiment, a corresponding
rectangular area with a location (southwest corner) of  (X = -120,
Y = -360) meters, an alongwind (X)  dimension of 240 meters, a
crosswind  (Y) dimension of 720 meters, and a release height of
0.0 meters, is used.  These inputs are defined on the SO LOCATION
and SO SRCPARAM cards.  To simulate the 45-degree rotated case
and the 90-degree rotated case, the same rectangular area source
is rotated 45 degrees and 90 degrees counterclockwise,
respectively, relative to the center of the area.  (The model
interface was later modified to rotate the area relative to the
vertex defined on the source location card.)   For the circular
area,  the ISCST2 model was modified to read in a series of
vertices, (x and y coordinates)  to define the area source shape.
For the purposes of this evaluation study,  the circular source
was approximated by a 20-sided polygon.

     In order to compare the ISCST2 model results with the
normalized concentrations reported for the wind tunnel data,  the
emission rate and wind speed were set to provide a normalized
concentration in the model output.   To generate the normalized
results,  the total source strength is set to be 1.0 x 106  grams
per second, and is assumed to be uniformly distributed over the
area.   This is equivalent to an emission rate of 5.79 g/(s-m2)
for the ISCST2 model input.   The factor of 106 was  included in
the wind tunnel measured data.

     For each wind tunnel scenario,  three sets of recepcor
locations were used.  Receptor locations in full scale"units were
extracted from the wind tunnel  data sets,  and were used to setuo
receptor networks on the ISCST2 input files.   A separate model
run was constructed for each of the wind tunnel measurement
planes,  i.e.,  the ground-level  horizontal (XY)  plane,  the
centerline vertical (XZ)  plane, and the vertical crosswind (YZ)
                                15

-------
plane.  The vertical  crosswind plane was located at  a downwind
distance of 840 meters.

3.1.3 Dispersion Parameters

     The modules in the  ISCST2 model used to calculate  the
Pasquill-Gifford  (PG)  sigmas were replaced with modified  code  to
incorporate the lateral  dispersion parameter  (
-------
3.2 Physical Analysis

     The purpose of this kind of analysis is to ensure that the
results generated by the ISCST2 model with the numerical
integration area source algorithm are physically reasonable
across a range of source-receptor geometries.  This involves a
qualitative evaluation based on point-to-point comparisons of
modeled versus measured results.

     A computer program was developed to merge the raw data set
provided by the wind tunnel experiment and the post-processing
file generated by ISCST2 for each case.  The resulting merged
data file was then imported into a spreadsheet/graphics package
to generate plots of the point-to-point comparisons.

     Figure 3-2 provides an example of the point-to-point
comparison results.  From graphs such as these, one can
qualitatively evaluate the performance of the new ISCST2
numerical integration area source algorithm.  These figures
illustrate the shape of the plume from the area source as it
disperses downwind, and illustrate the relative magnitudes of the
modeled versus measured results as a function of receptor
location.  The y-axis of Figure 3-2 uses a logarithmic scale
instead of a linear scale to show the normalized concentration.
This is necessary due to the large range of concentration values
across the plot.  Normalized concentration values of  less than
0.01 were excluded from the point-to-point plots.
                               17

-------
  Normalized Concentration CCU/Q,  1/m**23
   100
    10
   0 1
   o  O-T
                                                                ISCST2
                                                              Wi nd  Tunnei
     -600 -500 -400  -300 -200 -100  0   100  200  300  400  500  500
                           Crosswind Distance Cm!)
Figure 3-2.     An  example  of the  graphical presentation of  the
                 physical analysis.
                                    18

-------
3 .3 Statistical Analysis

     The purpose of the statistical analysis is to provide a
quantitative comparison of the modeled versus measured
concentrations.  The fractional bias is used as the fundamental
measure of discrepancy between the measurement -based and
dispersion model -based results.  The fractional bias, FBj,  for
the i* data point in a data series can be calculated as follows:

where:    OB;=  the observed  (wind tunnel) data for the 1th data
               point
          PRt=  the predicted  (ISCST2 model) result for the  i*
               data point

The mean fractional bias  (MFB) across a subset of the data  can be
estimated as follows:
                                    FBi                      (3-2)
                                 i=l
The sample standard deviation  (SD) of the fractional bias is  then
given by:
                    SD*
                                           \°'5
                                                            (3-3)
     Figure 3-3 shows an example of the results for the
statistical ^analysis.  The mean fractional bias and the standard
deviation of the fractional bias are first calculated for a
subset of data points, such as the set of ground-level crosswind
receptor points at a given downwind distance.  Similar statistics
can be calculated at each downwind distance for a particular
scenario, and the results plotted, as in Figure 3-3, to show  the
fractional bias as a function of downwind distance.  For
regulatory modeling applications, the maximum concentrations  are
of greater importance than concentrations at the small end of the


                                19

-------
spectrum.  In order to avoid having the fractional bias statistic
overly influenced by points near the edge of the plume, which may
have a large difference for relatively small values,  measurement
points with either a normalized concentration of less than 0.01
or less than 0.01 times the peak wind tunnel value for that
distance where excluded from the calculations.   This  approach
provides a consistent way of selecting data points for the
statistical analysis, given the varying crosswind widths of the
area sources for the four scenarios.
                                20

-------
  Fract ionaI Bias

   2
                                 MFB+SD  MFB

                                 MFB-SD   *
  - 1
  -2
        -60      Q      120
               240     360     340
              Downwind Distance C
-------
              4.0  RESULTS  FOR  THE  BASE  CASE  SCENARIO

     The base case scenario consists of a 720m x 240m rectangular
source oriented with its long dimension perpendicular to the wind
flow, with a 0.2 meter roughness length and a free-stream wind
speed of 4 m/s.

     For the first group of measurement points from the wind
tunnel experiment, which are located in the ground level XY
plane, an 8x49 Cartesian receptor network was setup.  The
crosswind distances of the receptors range from -600 meters to
600 meters.   The grid coordinates match the wind tunnel sensor
locations in full scale units.  Downwind distances (measured from
the center of the area)  of -60,  0, 120, 240, 360,  840,  1800, and
2580 meters were used.  The first two distances include receptors
located within the area source,  while the third distance
corresponds to the downwind edge of the area.

     For the second group of measurement points,  which are
located in the XZ plane along the centerline, 8 discrete
Cartesian receptor points were used for each of the 28 flagpole
heights.  The 8 horizontal locations correspond to the 8 downwind
distances mentioned above, and the 28 flagpole heights of the
receptors ranged from 1.8 meters to about 344 meters in full
scale units.

     For the last group of the measurement points,  which are
located in the YZ plane at a full scale downwind distance of 840
meters, a 12x17 Cartesian network is set up to match the wind
tunnel measurement locations.

4.1 Results Of The Physical Analysis

     Figure 4-1 shows the comparison of the wind tunnel
experiment data and ISCST2 results for ground-level receptors
located at the downwind edge of the base case area source (X =
120 meters).  The downwind edge of the area source deserves
significant attention, since the highest concentration values for
area sources are expected to occur at this location.   The figure
shows very good agreement between the wind tunnel  and the ISCST2
results, both in terms of magnitude of the results and in terms
of the overall shape of the plume.  The high values near the
centerline match very well, and the low values near the edge of
the plume also show only insignificant differences.  These
results illustrate that the new algorithm is successfully
treating the edge effects for this scenario.  The  wind tunnel
results show some scatter above and below the model results,
which may be attributed to microscale effects due  to the surface
roughness elements close to the area source, and to the finite
averaging time.  Appendix A provides a collection of all the
point-to-point comparison results for the base case scenario.
The overall trends are that the predictions of the ISCST2 model
tend to match the wind tunnel experiment results inside and


                                22

-------
nearby the area source, but they tend to overpredict the area
source impact at ground-level locations farther downwind.

     Figure 4-2 shows the comparison of the wind tunnel and
ISCST2 centerline vertical concentration distributions at the
downwind edge of the area source.  The ISCST2 results match the
wind tunnel data very well in the lower levels (< 30 meters),  but
tend to underpredict the concentration values for the upper
levels.  The centerline vertical distributions at other downwind
distances are presented in Appendix A.  These figures show that
plume material in the wind tunnel simulation is being dispersed
more rapidly in the vertical direction than predicted by the
ISCST2 model, even though the dispersion parameters are based on
a fit to the wind tunnel data.  This suggests that the vertical
dispersion of the area source plume is non-Gaussian.

     Figure 4-3 shows the comparison of the wind tunnel and
ISCST2 lateral concentration distributions at a downwind distance
of 840 meters and a receptor height of 55.8 meters above ground.
This figure shows very good agreement in magnitude and shape for
the elevated plume at this downwind distance.  The lateral
distributions at other receptor heights for this downwind
distance are shown in Appendix A.  These figures show that while
the lateral plume shape is matched well by the model results at
all levels, the model tends to overpredict the magnitude of the
normalized concentrations for receptor elevations below about 60
meters and to underpredict for receptor elevations above 60
meters.  This is consistent with the vertical profiles of the
plume discussed above.

     The ability of the model to match the lateral shape of the
plume observed in the wind tunnel suggests that the Gaussian
assumption for lateral plume dispersion implicit in the model is
a reasonable one.  The tendency for the model to underpredict the
concentrations for receptors well above ground-level,  especially
at the farther downwind distances,  suggests that the Gaussian
assumption for vertical plume dispersion does not work as well.
However, for the receptors of most concern in typical regulatory
applications, i.e.,  ground-level receptors located near the
source, this deficiency is not a very serious one.
                                23

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	
   100 -
    10
    1 -
   0 1
   0 0
                                                            ISCST2
                                                          Wi nd Tunnei
     -600 -500 -400 -300 -200 -100  0   100  200  300  -400  500  600
                          Crosswind Distance CmD
Figure 4-1.
Comparison of lateral profile of normalized
concentration for the Base Case at a  downwind
distance  of 120 meters  from the center of the
area.   ISCST2 model results are based on
dispersion parameters fitted to wind  tunnel data
                                  24

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
   0
                                                           ISCST2
                                                          Wind Tunnel
    0 01
0,1         1          10         100
      Normalized Concentration CCU/Q, 1/ rn
                                                        1000
Figure 4-2.
 Comparison of vertical  profile of normalized
 concentration for the Base Case at  a  downwind
 distance of 120 meters  from the center of the
 area.   ISCST2 model results are based on
 dispersion parameters fitted to wind  tunnel data
                                  25

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000r:	
   100
    10
   0 1
   0 0
                                                           ISCST2
                                                          Wi nd TunneI
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance CmD
Figure 4-3.
Comparison of lateral profile of normalized
concentration for the Base  Case at a downwind
distance  of 840 meters  from the center  of  the area
and a.receptor elevation  of 55.8 meters  above
ground.   ISCST2 model results are based  on
dispersion parameters fitted to wind tunnel data.
                                  26

-------
4.2 Results Of The Statistical Analysis

     Figure 4-4 shows the mean fractional bias and the standard
deviation of the fractional bias as a function of downwind
distance for the ground-level horizontal  (XY) plane for the base
case scenario.  The scale of the vertical axis is for fractional
bias, which varies from -2 for extreme overprediction to +2 for
extreme underprediction.  A fractional bias of zero (0) indicates
perfect agreement between the observed (wind tunnel) and
predicted (ISCST2) values.  A fractional bias of +/- 0.67
indicates agreement within a factor of 2.

     For receptors located inside the area source (downwind
distances of less than 120 meters),  the standard deviation of the
fractional bias is larger, indicating that the concentration
values have relatively large fluctuations.  The standard
deviation of the fractional bias then decreases for receptors
located within a few source widths outside the area source.  The
physical analysis results discussed above and shown in Appendix A
also display some larger discrepancies between the wind tunnel
data and ISCST2 prediction inside the area source,  mainly because
of scatter in the wind tunnel data.   The plume is still
relatively shallow over the area, and microscale effects caused
by the surface roughness elements are considered to be the likely
cause for these fluctuations.

     The mean fractional bias (MFB)  for the ground-level
horizontal (XY) plane varies from about -0.4 inside the area
source to its minimum value of -0.02 at a downwind distance of
240 meters (120 meters downwind from the edge of the area), and
then increases in magnitude to about -0.3 for distances beyond
840 meters.   These comparison results are quite promising,  with
differences of less than 5 percent for the important receptor
locations at the downwind edge and within one source width of the
area.  The negative value of the fractional bias indicates that
the-ISCST2 model tends (slightly) to overpredict the results.

     Figure 4-5 shows the mean fractional bias and the standard
deviation of the fractional bias as  a function of receptor height
for the vertical crosswind (YZ)  plane at a downwind distance of
840 meters for the base case scenario.   Since this plane is
located well downwind of the area,  the standard deviation of the
fractional bias does not vary much with height.  This  indicates
that the wind tunnel data set is relatively smooth for these
receptor locations (i.e.,  not influenced by oiicroscale effects/ .

     The mean fractional bias varies with height from about -0.3
near- the surface to about 1.0 for receptors located 100.8 meters
above ground.   This means that the ISCST2 model overpredicts
slightly in the lower levels (less than about 60 meters),  and
underpredicts in the upper levels.   This result is also shown by
the physical analysis.   Again,  the results are very encouraging'


                               27

-------
for ground-level receptors, which are important for regulatory
applications.  The tendency for larger overprediction at ground-
level receptors at larger distances downwind,  together with
underprediction at elevated receptors, is attributed to the
apparent non-Gaussian character of the plume in the vertical.
The model predicts a slower rate of vertical dispersion than is
observed in the wind tunnel.

     The results of the statistical analysis are also presented
in tabular form in Appendix E.
                                28

-------
  Fract i onaI  Bias

  2
                               MFB+5D  MFB

                               MFB-SD   X
  -1
  -2
       -60
120     240     360
      Downwind Distance
                              840
                                                    1800
Figure 4-4
Fractional Bias of normalized concentration as  a
function of downwind  distance for the Base Case
using dispersion parameters fitted to wind tunnel
data.  Overbar is fractional bias plus  (+) and
underbar is fractional  bias minus (-) one standard
deviation  (SD); symbol  (*)  is mean fractional bias
(MFB).
                                29

-------
  Fract ionaI  Bias

  2
   1 -
_,,„_. i ir-r^ i—r^

iviro-^ou Mpg
i jr-n r- r^
ivir u OLJ TK
  -2
                                    I
           10 8  19 8  28 8  37 8 46 8 55,8 64 8  73,8  82 8  91 8  100 8
                             Receptor  Height
Figure 4-5.
Fractional Bias of normalized concentration as a
function of receptor height  for the Base Case
using dispersion parameters  fitted to wind tunnel
data.  Overbar is fractional bias plus (+)  and
underbar is fractional bias  minus (-)  one standard
deviation  (SD); symbol  (*)  is mean fractional bias
(MFB).
                                30

-------
          5.0 RESULTS FOR THE 45 DEGREE ROTATED SCENARIO

     The 45 degree rotated scenario is the measurement program of
the base case rectangle  rotated 45 degrees counterclockwise.
Because of the crosswind asymmetry for the 45 degree rotated
rectangle, it is perhaps the most interesting and challenging of
the tests of the numerical integration algorithm.  Results are
presented for the three  basic measurement planes, i.e., the
ground-level horizontal  (XY) plane, the centerline vertical  (XZ)
plane, and the crosswind vertical  (YZ) plane.

5. 1 Results Of The Physical Analysis

     Figure 5-1 shows the comparison of the wind tunnel
experiment data and  ISCST2 results for ground-level receptors
located at the downwind  edge  (corner)  of the 45 degree rotated
rectangle (X = 340 meters).  The downwind edge of the area source
deserves significant attention, since the highest concentration
values for area sources  are expected to occur at this location.
As with the base case scenario described in Section 4,  the figure
shows very good agreement between the wind tunnel and the ISCST2
results, both in terms of magnitude of the results and in terms
of the overall shape of  the plume.   The asymmetry of the
crosswind profile is very evident at this distance,  with the peak
concentration location corresponding to the furthest downwind
corner of the area.  Appendix B provides a collection of all the
point-to-point comparison results for the, 45 degree rotated
scenario.  The overall trends are similar to those observed for
the base case, with the  ISCST2 results matching the wind tunnel
results inside and nearby the area source,  but tending to
overpredict the area source impact at ground-level locations
farther downwind.   It is particularly interesting to note how
well the ISCST2 model results match the asymmetry of the
crosswind profile of the observed plume within and nearby the
area source.

     Figure 5-2 shows the comparison of the wind tunnel and
ISCST2 centerline vertical concentration distributions  at the
downwind edge of the area source.   The ISCST2 results match the
wind tunnel data very well in lower levels  (<40 meters),  but tend
to underpredict the concentration values for the upper  levels.
The centerline vertical distributions  at other downwind distances
are presented in Appendix B.   These figures show the same trend
as observed for the base case with the discrepancies for the
elevated receptors increasing as the downwind"distance  increases.

     Figure 5-3 shows the comparison of the wind tunnel and
ISCST2 lateral concentration distributions  at a downwind distance
of 840 meters and a receptor height of 55.8 meters above ground.
This figure shows  very good agreement  in magnitude and  shaoe for
the elevated plume at this downwind distance.   The model matches
the slight asymmetry of the observed plume  fairly well.   The
lateral distributions at other receptor heights for this downwind


                                31

-------
distance are shown in Appendix B.  These figures show that while
the lateral plume shape is matched well by the model results at
all levels, including the slight asymmetries,  the model tends to
overpredict the magnitude of the normalized concentrations for
receptor elevations below about 60 meters and to underpredict for
receptor elevations above 60 meters.  This is similar to the Base
Case, and is consistent with the vertical profiles of the plume
discussed above.
                                32

-------
   Normalized Concentration CCU/Q,  1/m**2}

   1000c	
   100
    10
   0 1
   0 01
                     I  I  I
                                 _J	1	L.
                                                            I SCST2
                                                           Wi nd Tunne I
     -500 -500 -400  -300 -200 -100  0   100  200   300  400  500   600
                          Crosswind Distance
Figure  5-1.
Comparison of lateral  profile of normalized
concentration for the  45 Degree Rotated Case  at  a
downwind distance of 340 meters from the center  of
the area.   ISCST2 model  results are  based on
dispersion parameters  fitted to wind tunnel data.
                                  33

-------
  Receptor Height

  300
  250 -
  200 -
  150
  100 -
   50 -
    0
    0 01
                                                           ISC5T2
                                                         Wi nd TunneI
0-1         1          10         100
      Normalized Concentration CCU/Q,
                                                       1000
Figure 5-2.
 Comparison of vertical profile of normalized
 concentration for the 45 Degree Rotated Case at a
 downwind distance of 340 meters from the  center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel  data.
                                 34

-------
  Normalized Concentration CCU/Q,  1/m**2;)

  1000c	
   100
    10
   0 1
   0.0
                               I
                                                           ISCST2
                                                         Wi nd Tunne I
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance Cm3
Figure 5-3.
Comparison of lateral profile of normalized
concentration for the 45 Degree Rotated Case at a
downwind  distance of 840 meters from the  center of
the area  and a receptor elevation of 55.8 meters
above ground.  ISCST2 model  results are based on
dispersion parameters fitted to wind tunnel data.
                                 35

-------
5.2 Results Of The Statistical Analysis

     Figure 5-4 shows the mean fractional bias and the standard
deviation of the fractional bias as a function of downwind
distance for the ground-level horizontal (XY)  plane for the 45
degree rotated rectangle.  The scale of the vertical axis is for
fractional bias, which varies from -2 for extreme overprediction
to +2 for extreme underprediction.  A fractional bias of zero  (0)
indicates agreement between the observed (wind tunnel) and
predicted (ISCST2) values.  A fractional bias of +/- 0.67
indicates agreement within a factor of 2.

     For receptors located inside the area source,  especially
upwind of the center, results show that the concentration bias
values have relatively large fluctuations.   The absolute mean
fractional bias is still quite small, less than about 0.15 for
receptors located over the area.  The standard deviation of the
fractional bias decreases for receptors located outside the area
source.  The physical analysis results discussed above and shown
in Appendix B also display some larger discrepancies between the
wind tunnel data and ISCST2 prediction inside the area source,
mainly because of scatter in the wind tunnel data.   Also, as
noted above, the concentration distributions for the rotated
rectangle are more complex, and there are somewhat larger
discrepancies for the plume edges over the area source than for
the base case.

     The mean fractional bias (MFB) for the ground-level
horizontal  (XY) plane varies from about -0.15 inside the area
source to its minimum value of -0.05 at the center of the area,
and then increases in magnitude to about -0.4 at a downwind
distance of 1800 meters.  These comparison results are quite
promising, with the ratio of the results showing differences of
less than about 10 percent for the important receptor locations
at the downwind edge of the area.  The negative value of the
fractional bias indicates that the ISCST2 model tends (slightly)
to overpredict the results at ground-level receptors.

     Figure 5-5 shows the mean fractional bias and the standard
deviation of the fractional bias as a function of receptor height
for the vertical crosswind (YZ)  plane at a downwind distance of
840 meters for the 45 degree rotated rectangle.  The standard
deviation of the fractional bias tends to increase with height
for this case, which appears to be related to some slight
differences in the shape of the profiles due co cha rccacicn
effect.

     The mean fractional bias varies with height from about -0.3
near the surface to about 1.3 or greater for receptors located
above  100 meters above ground.  This means that the ISCST2
overpredicts slightly in the lower levels (less than about 55
meters), and underpredicts in the upper levels.  This result is


                                36

-------
also shown by the physical analysis.   Again,  the results are very
encouraging for ground-level receptors near the source,  which are
important for regulatory applications.

     The results of the statistical analysis are also presented
in tabular form in Appendix E.
                               37

-------
  Fract i onaI  Bias

  2
.,_, i tr- r~\ t^r\


IVII'LJ— OU
MFB
*
  -1
  -2
       -170
       170     340     480     600
             Downwind Distance CmD
340
Figure 5-4
Fractional Bias of normalized  concentration as a
function of downwind distance  for  the 45-Degree
Rotated Case using dispersion  parameters fitted to
wind tunnel data.  Overbar  is  fractional bias plus
(+) and underbar is fractional bias  minus (-) one
standard deviation  (SD);  symbol (*)  shows mean
fractional bias  (MFB).
                                38

-------
  Fract i onaI  B i as

  2
i it-r~i

ivir u
i 
-------
          6.0  RESULTS  FOR THE  90  DEGREE  ROTATED  SCENARIO

     The 90 degree rotated scenario is the measurement program of
the base case rectangle rotated 90 degrees clockwise.  This
corresponds to having the short dimension of the area source
oriented perpendicular to the direction of wind flow.  Results
for this case are only presented for two measurement planes, the
ground-level horizontal  (XY)  plane, and the centerline vertical
(XZ) plane.  There were no measurements made for the crosswind
vertical  (YZ)  plane.

6.1 Results Of The Physical Analysis

     Figure 6-1 shows the comparison of the wind tunnel
experiment data and ISCST2 results for ground-level receptors
located at the downwind edge of the 90 degree rotated rectangle
(X = 360 meters).   The downwind edge of the area source deserves
significant attention, since the highest concentration values for
area sources are expected to occur at this location.  As with the
base case and the 45 degree rotated scenarios described in
previous sections, the figure shows very good agreement between
the wind tunnel and the ISCST2 results,  both in terms of
magnitude of the results and in terms of the overall shape of the
plume.  Appendix C provides a collection of all the point-to-
point comparison results for the 90 degree rotated scenario.  The
overall trends are similar to those observed for the base case,
with the ISCST2 results matching the wind tunnel results inside
and nearby the area source, but tending to overpredict the area
source impact at ground-level locations farther downwind.

     Figure 6-2 shows the comparison of the wind tunnel and
ISCST2 centerline vertical concentration distributions at the
downwind edge of the area source.  The ISCST2 results match the
wind tunnel data very well in lower levels (<50 meters), but tend
to underpredict the concentration-values for the upper levels.
The centerline vertical distributions at other downwind distances
are presented in Appendix C.   These figures show "the same trend
as observed for the base case with the discrepancies for the
elevated receptors increasing as the downwind distance increases.

6.2 Results Of The Statistical Analysis

     Figure 6-3 shows the mean fractional bias and the standard
deviation of the fractional bias as a function of downwind
distance for the ground-level horizontal  (XY) plane for the 90
iecr^ss rotstsci "**2ct3.no'Is.  """he seals of the vertical i£> zCJT
fractional bias, which varies from -2 for extreme overprediction
to +2 for extreme underprediction.  A fractional bias of zero  (0)
indicates agreement between the observed  (wind tunnel') and
predicted  (I3CST2) values.  A fractional bias of +/- 0.57
indicates agreement within a factor of 2.
                                40

-------
     The mean fractional bias (MFB)  for the ground-level
horizontal (XY)  plane varies from a maximum of about -0.2 at a
downwind distance of -180 meters to its minimum value of less
than about 0.01 over the center of the area,  and then increases
in magnitude to about -0.3 at a downwind distance of 1800 meters.
These comparison results are quite promising,  with the ratio of
the results showing differences of less than 5 percent for the
important receptor locations at the downwind edge of the area.

     There were no vertical crosswind measurements made for this
scenario, but the centerline vertical results  are consistent with
the results seen for the other rectangular cases.

     The results of the statistical analysis are also presented
in tabular form in Appendix E.
                               41

-------
  Normalized Concentration CCU/Q, 1/m**2)

  1QOQp	
   100
    10
   0 1
   0 011
                                                            ISCST2
                                                          Wind TunneI
     -500 -500 -400  -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure  6-1.
Comparison of lateral profile of normalized
concentration for the 90  Degree Rotated Case at  a
downwind  distance of 360  meters from the center  of
the area.   ISCST2 model results are  based on
dispersion parameters fitted to wind tunnel data.
                                  42

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
   0
    0 01
                                                           ISCST2
                                                         Wi nd Tunne!
01         1          10         100
      Normalized Concentration CCU/Q,
                                       1000
Figure 6-2.
Comparison  of vertical profile  of normalized
concentration for the 90 Degree Rotated Case  at  a
downwind distance of 360 meters from the center  of
the area.   ISCST2 model results are based on
dispersion  parameters fitted  to wind tunnel data.
                                 43

-------
  Fract i onaI  Bi as

  2
                               MFB+SD  MFB

                               MFB-SD   ¥
                                      J_
       -180
180     360     480
      Downwind Distance
                             600
843
Figure 6-3
Fractional. Bias of normalized concentration as a
function of downwind  distance for the 90-Degree
Rotated Case using dispersion parameters fitted to
wind tunnel data.  Overbar is fractional bias plus
(+) and underbar  is minus  (-)  one standard
deviation  (SD); symbol  (*)  shows mean fractional
bias (MFB).
                                 44

-------
           7.0 RESULTS FOR THE CIRCULAR SOURCE SCENARIO

      The circular area source scenario includes  measurements for
 a circular source with a diameter of 470  meters,  which
 corresponds  to the same total area as the rectangular source
 discussed in previous sections.   As described in Section 3,  the
 circular area was modeled as  a 20-sided polygon  by making a minor
 modification to the ISCST2 model  code.  Results  for the  circular
 source are presented for three measurement planes,  the ground-
 level horizontal (XY)  plane,  the  centerline vertical (XZ)  plane,
 and the crosswind vertical  (YZ) plane.

 7.1 Results  Of The Physical Analysis

      Figure  7-1  shows the comparison of the wind tunnel
 experiment data  and ISCST2 results for  ground-level receptors
 located at a downwind distance of 204 meters,  relative to the
 center of the area.   This distance is about 30 meters  upwind of
 the furthest downwind point on the circular area.   As  with the
 rectangular  scenarios described in previous sections,  the figure
 shows very good  agreement between the wind tunnel  and  the ISCST2
 results,  both in terms  of magnitude of  the results  and in terms
 of  the overall shape of  the plume.   The ISCST2 results match the
 wind tunnel  results  inside and nearby the  area source, but tend
 to  overpredict the area  source  impact at ground-level  locations
 farther downwind.  Appendix D provides a collection of all the
 point-to-point comparison results  for the  circular  source
 scenario.

      Figure  7-2  shows  the comparison of the  wind tunnel  and
 ISCST2  centerline  vertical concentration distributions at  the
 same distance downwind  (X = 204 meters).   The ISCST2 results
 match the  wind tunnel data very well  in the  lower levels  (<30
 meters), but  tend  to underpredict  the concentration values for
 the  upper  levels.  The centerline  vertical  distributions  at  other
 downwind distances are presented  in Appendix D.  These figures
 show the same trend  as observed for  the base case,  with  the
 discrepancies for  the elevated receptors increasing as the
 downwind distance  increases.

      Figure  7-3  shows the comparison  of the wind tunnel  and
 ISCST2  lateral concentration  distributions at a downwind distance
 of  840 meters  and  a  receptor  height of 55.8 meters  above ground.
 This  figure  shows  very good agreement in magnitude  and shape for
 the  elevated  plume at this downwind distance.  The  lateral
 distributions  at other receptor heights for this downwind
 distance are  shown in Appendix D.   These figures show that while
 the  lateral plume  shape is matched well by the model results at
 all  levels, the model tends to overpredict the magnitude of the
 normalized concentrations for receptor elevations below about 60
meters and to underpredict for receptor elevations above 60
meters.  This  is similar to the Base Case,  and is consistent with
 the vertical profiles of the  plume discussed above.


                                45

-------
  Normalized Concentration CCU/Q, 1/m**23
   100
    10
   0 1

                                                            SCST2
                                                          Wi nd Tunne I
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswmd Distance
Figure 7-1.
Comparison of lateral profile of normalized
concentration for the Circular Case at  a downwind
distance  of 204 meters  from the center  of the
area.   ISCST2 model results are based on
dispersion parameters fitted to wind tunnel data.
                                  46

-------
  Receptor Height

  300
  250
  200
  150
  100
    0
    0 01
                                                            SCST2
                                                          Wi nd TunneI
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**2;)
Figure 7-2.
 Comparison of vertical  profile of normalized
 concentration for the Circular Case at  a downwind
 distance of 204 meters  from the center  of the
 area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel data.
                                 47

-------
  Normalized Concentration CCU/Q,  1/m**2;i

  1000p	
   100
   0 1
   0 OT
                                                           5CST2
                                                         Wind Tunne
     -6QO -500 -400 -300 -200 -100  0   100 200  300  400  500  600
                         Crosswind Distance
Figure 7-3.
Comparison  of  lateral profile  of normalized
concentration  for the Circular Case at a downwind
distance  of 840 meters from  the center of  the  area
and a receptor elevation of  55.8 meters above
ground.   ISCST2 model results  are based on
dispersion  parameters fitted to wind tunnel  data.
                                 48

-------
7.2 Results Of The Statistical Analysis

     Figure 7-4 shows the mean fractional bias and the standard
deviation of the fractional bias as a function of downwind
distance for the ground-level horizontal  (XY) plane for the
circular source scenario.  The scale the vertical axis is for
fractional bias, which varies from -2 for extreme overprediction
to +2 for extreme underprediction.  A fractional bias of zero  (0)
indicates perfect agreement between the observed  (wind tunnel)
and predicted  (ISCST2) values.  A fractional bias of +/- 0.67
indicates agreement within a factor of 2.

     The mean fractional bias (MFB) for the ground-level
horizontal (XY) plane varies from about 0.2 inside the area
source to its minimum value of about 0.01 near the downwind edge
of the area (X = 240 meters),  and then increases in magnitude to
about -0.3 at a downwind distance of 1800 meters.  These
comparison results are quite promising,  with the ratio of the
results showing differences of less than 5 percent for the
important receptor locations near the downwind edge of the area.
Unlike the rectangular source scenarios described in previous
sections,  the ISCST2 model is slightly underpredicting
concentrations over the circular source area, while still
overpredicting at distances farther downwind.  The reason for
this slight difference is not clear from this analysis.

     Figure 7-5 shows the mean fractional  bias and the standard
deviation of the fractional bias as a function of receptor height
for the vertical crosswind  (YZ)  plane at a downwind distance of
840 meters for the circular source scenario.  This figure shows a
similar trend to that observed for the rectangular source for the
model to overpredict near the ground-level and to underpredict
for elevated receptors.

     The mean fractional bias varies with .height from about -0.3
near the surface to about 1.1 or greater for receptors located
above 100 meters above ground.   This means that the ISCST2
overpredicts slightly in the lower levels  (less than about 60
meters),  and underpredicts in the upper levels.   This result is
also shown by the physical analysis.   Again, the results are very
encouraging for ground-level receptors,  which are important for
regulatory applications.

     The results,of the  statistical analysis are also presented
in tabular form in Appendix' E.
                               49

-------
  Fractional Bias

  2
i ir-r-> r- r-,

IVll O^OU
k 4i~ n T\

MFB
* -
  -1
  -2
        I
       -102
102     204    240    360
        Downwind Distance
                                480
840
Figure 7-4
Fractional Bias as a  function of downwind distance
for the Circular Case using dispersion parameters
fitted to wind tunnel data.   Overbar is fractional
bias plus  (+) and underbar is minus (-) one
standard deviation  (SD) ,-  symbol (*) is mean
fractional bias  (MFB).
                                50

-------
  Fractional Bias

   2
                               MFB+SD  MFB

                               MFB-SD   *
  -2
                                  I
                19 8      37 8     55 8     73 8
                             Receptor Height
                                   91 8
109  3
Figure 7-5
Fractional Bias  as  a  function of receptor height
for the Circular Case using dispersion parameters
fitted to wind tunnel data.   Overbar is fractional
bias plus  (+) and underbar is minus (-) one
standard deviation  (SD)/  symbol (*) shows mean
fractional bias  (MFB).
                                51

-------
            8.0 THE WIND TUNNEL DISPERSION PARAMETERS

     The differences between the Pasquill-Gifford  (PG) sigmas and
the lateral dispersion parameter (cry)  and vertical  dispersion
parameter  (crz)  observed  in  the  wind tunnel were  shown in  Figures
2-4 and 2-5.  It was shown that the dispersive behavior of the
simulated atmospheric boundary layer in the wind tunnel is near
the PG-C stability class (slightly unstable) close to the source
and tends toward the PG-D stability class (neutral)  farther
downwind.

     In order to focus the evaluation study on the performance  of
the numerical integration algorithm, rather than on  the
representativeness of PG dispersion parameters to the wind tunnel
boundary layer, the ISCST2 model results were generated using
dispersion parameters based on the wind tunnel point source data.
The ISCST2 model was modified to use a power-law fit of the wind
tunnel data, given by the following formulas:


                         o  = 0.73547 x°-64931                   (8-1)
                         O  =0.28565x°'71285                   (8-2)
                          z
where x is the downwind distance in meters.  Figures 8-1 and  8-2
provide comparisons of the wind-tunnel derived values and the
power-law fit for ay and  az, respectively.  Linear axes are used
in these figures in order to make the slight differences more
distinguishable .

     In order to evaluate the effects of using the wind-tunnel
derived dispersion parameters, two cases were examined using  PG
class C and PG class D dispersion parameters for comparison to
the results using wind tunnel data.  These comparisons are
presented for ground-level receptors for the base case scenario
and for the circular source case.  Point-to-point comparisons of
normalized concentration are presented for a downwind distance of
240 meters for both cases, using a linear concentration axis  to
make the differences more visible.  This distance corresponds to
the first distance downwind of the source, and is where the
modeled versus wind tunnel result? showed the best agreement .

          es 3-1 and 3-4 shew che ccrnpaTiscn. results fcr  zhe  -;a
case.  Figure 8-3 shows that the ISCST2 results based on  PG-C
stability class slightly underpredict the concentration values
near the rectangular source, while the PG-D stability class
resulcs overpredict the centerline concentration values at che
same distance.  The C stability results match the wind tunnel
concentrations better than  the D stability results, which is
consistent with the comparison of sigma values shown in Section

                                52

-------
2.  Figure 8-4 compares the mean fractional bias for ground-level
crosswind profiles as a function of downwind distance for the PG-
C, PG-D and wind tunnel dispersion parameters.

     Figures 8-5 and 8-6 show the comparison results for the
circular source case.  These figures show a similar pattern to
that seen for the base case scenario,  with PG-C stability
matching the wind tunnel data better close to the source,  but
tending to underpredict for receptors  located farther downwind.
The PG-D stability results match the wind tunnel sigmas better
farther downwind, but tend toward a larger overprediction of
concentrations.
                               53

-------
Sigma Y

100
500
                              1000

                          Downwind Distance
                                         1500
                                                         Raw Data
                                                        Power- I aw Fit
                                       2000
Figure  8-1.
              Comparison  of the wind tunnel  data and  power-law
              fit  for
                        y'
                                 54

-------
   S i gma Z £ rrf)
   100
   50
500
     1000
Downwind  Distance
                                            1500
                                                             Raw Data
                                                            Power-I aw F11
2000
Figure 8-2.     Comparison of  the wind tunnel  data and power-law
                 fit for crz.
                                   55

-------
  Normalized Concentration CCU/Q., 1/m**2}

  140
  120 -
  100 -
                                                        Fitted Parameters
                                                        Wind Tunnel  Data
                                                          PG  Class C
                                                          PG  Class D
    -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                        Crosswind Distance Cm!)
Figure 8-3.
Comparison of  the wind tunnel  data and ISCST2
results at a downwind  distance of 240  meters  for
the Base Case  with various dispersion  parameters
                                    56

-------
  Fractional Bias

  2
  -2
-60
                                                            F i t ted
                                                           PG Class C
                                                           PG Class D
                    120     240    360    840
                         Downwind Distance
                                        1800
2580
Figure  8-4
         The mean fractional bias (MFB) of normalized
         concentration as a function of downwind distance
         for the Base Case with various dispersion
         parameters.
                                  57

-------
  Normalized  Concentration CCD/Q^ 1/m**23

  300
                                                        Fitted Parameters
                                                        Wi nd TunneI  Data
                                                         PG Class C
                                                         PG Class D
    -600 -500 -400 -300 -200 -100  0   100  200 300 400  500  600
                        Crosswind Distance
Figure  8-5.
Comparison of  the wind  tunnel data and  ISCST2
results at a downwind distance of  240 meters for
the  Circular Source Case  with various dispersion
parameters.
                                   58

-------
  Fractional Bias

  2
  - 1
  -2
      -102
                                                            Fitted
                                                           PG Class C
                                                           PG Class D
  102    204    240    360
         Downwind Distance
                                          480
840
1800
Figure.8-6.
The mean  fractional bias  (MFB)  of normalized
concentration as a function of downwind  distance
for the Circular Source Case with various
dispersion parameters.
                                  59

-------
                         9.0 CONCLUSIONS

     This report documents an evaluation of the new ISCST2
numerical integration algorithm for modeling impacts from area
sources using data collected in a wind tunnel experiment.  Both a
qualitative physical analysis and a quantitative statistical
analysis were performed to evaluate the performance of the
algorithm.  In the physical analysis,  point-to-point comparisons
were made between the ISCST2 model results and the wind tunnel
measured results for a range of source scenarios and receptor
locations.  The overall magnitude of the normalized
concentrations and the shape of the plume were qualitatively
compared to determine the reasonableness of the model results.
In the statistical analysis, the mean fractional bias and the
standard deviation of the fractional bias were used as
quantitative measures of the performance of the algorithm.  The
fractional bias was determined for a series of crosswind receptor
locations, and the mean fractional bias and standard deviation
were plotted for ground-level receptors as a function of downwind
distance and for elevated receptors as a function of receptor
height at a downwind distance of 840 meters.

     A total of four wind tunnel experiment scenarios were
modeled by ISCST2 with the numerical integration area source
algorithm.  The base case scenario provides a detailed picture of
the impacts from a rectangular area source with the long
dimension oriented perpendicular to the direction of flow.  The
rectangular source has an aspect ratio of 3 to 1.   The
rectangular area source was then rotated 45 degrees and 90
degrees counterclockwise to show the effects of source
orientation relative to the wind flow.  Finally, a circular
source scenario was modeled by ISCST2 by approximating the circle
as a 20-sided polygon.  The circular source has a diameter of 470
meters, giving it the same area as the rectangular source.

     The results of the analyses show that the new ISCST2 area
source algorithm predicts the concentration distribution with
relatively good accuracy, especially for the ground-level
receptors of maximum impact located near the downwind edge of the
area source.  The normalized ISCST2 modeled concentrations
generally matched the wind tunnel measured concentrations to
within 10 percent, and the lateral shape of the plume was also
matched very well.  The overall trend is for the ISCST2 model to
overtjrsdict concentrations "for the ground-level recactors Iccacsd
farther downwind of the area and to underpredict concentrations
fcr =1evaded recaptcrs.  These rasults suggest thac 3iie Gaussian
assumption for lateral dispersion compares well with wind tunnel
observations for a neutral boundary layer, but that the plume
does not show a Gaussian distribution in the vertical.  The model
tends co underestimate che amount of vertical spread of the
plume.  The integration procedure used to model the area source
appears to provide a very accurate estimate of the concentrations
                                60

-------
for ground-level receptors located near the source,  which are of
most concern to regulatory modeling and Superfund applications.

     The comparison results for the various source shapes and
orientations demonstrate the ability of the model to accurately
depict the plume shape for complex source-receptor geometries.
The performance of the model is particularly noteworthy for the
45-degree rotated rectangle because of its ability to match the
lateral asymmetry of the plume over and near the source.   In
addition, for the circular source,  the model was modified to
allow inputting of arbitrary vertices for a multi-sided polygon,
indicating the flexibility of the model to handle irregularly-
shaped sources.  While the model algorithm is capable of handling
these irregular polygons,  the user interface for defining such
sources has not been fully developed.
                               61

-------
                         10.0 REFERENCES

Cox, W.M., 1988:  Protocol for Determining the Best Performing
     Model.  U.S. Environmental Protection Agency, Research
     Triangle Park,  NC.

Press, W., B. Flannery, S. Teukolsky, and W. Vetterling, 1986:
     Numerical Recipes. Cambridge University Press, New York, 797
     pp.

Snyder, W. H., 1991:  Wind-Tunnel Simulation of Dispersion from
     Superfund Area Sources,  Fluid Modeling Facility Internal
     Report,  U.S. Environmental Protection Agency, Research
     Triangle Park,  NC.

U.S. Environmental Protection Agency, 1989:  Review and
     Evaluation of Area Source Dispersion Algorithms for Emission
     Sources at Superfund Sites.  EPA-450/4-89-020.  U.S.
     Environmental Protection Agency, Research Triangle Park, NC.

U.S. Environmental Protection Agency, 1992a:  Sensitivity
     Analysis of a Revised Area Source Algorithm for the
     Industrial Source Complex Short Term Model.  EPA-454/R-92-
     015, U.S. Environmental Protection Agency, Research Triangle
     Park, NC.

U.S. Environmental Protection Agency, 1992b:  Summary of Quality
     Assurance and Equivalence Tests Performed on the Modified
     Area Source Algorithm for the ISCST2 Model.  Project report
     for WA No. 1-27,  68D00124, U.S. Environmental Protection
     Agency,  Research Triangle Park, NC.

U.S. Environmental Protection Agency, 1992c-.  User's Guide for
     the Industrial Source Complex  (ISC2) Dispersion Models,
     Volume I - User Instructions.  EPA-450/4-92-008a,  U.S.
     Environmental Protection Agency, Research Triangle Park, NC.
                                62

-------
           APPENDIX A









POINT-TO-POINT COMPARISON RESULTS




        FOR THE BASE CASE

-------
     This Appendix presents the point-to-point comparison results
for the Base Case scenario.  The Base Case consists of a 720m x
240m rectangular area with the long dimension oriented
perpendicular to the wind flow in a simulated boundary layer with
a 0.2 meter surface roughness length and a free-stream wind speed
of 4 m/s.  The ground-level lateral concentration profiles (XY
plane)  are presented first, followed by the centerline vertical
profiles  (XZ plane),  and then the vertical crosswind profiles (YZ
plane)  at a downwind distance of 840 meters.   The ground-level
profiles are based on a 1.8 meter receptor height above ground.
The vertical crosswind profiles for receptor heights above about
100 meters have not been included in this report.  Normalized
concentration values of less than 0.01 have also been excluded
from the graphs.  As noted in the captions, the ISCST2 model
results are based on the use of dispersion parameters fitted to
wind tunnel data.  The use of wind tunnel sigmas is discussed in
Section 8.
                               A-l

-------
  Normalized Concentration CCU/Q, 1/m**23
   100
   10
   0. 1
   0 0
                                                           ISCST2
                                                          Wind TunneI
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance Cm)
Figure A-l.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Base Case at a
downwind  distance of -60  meters from  the center  of
the area.   ISCST2 model results are based on
dispersion parameters fitted to wind  tunnel daca.
                                 A-2

-------
  Normalized Concentration CCU/Q,  1/m**2D
   100
   10
   0 1
   0 01
                                                           ISCST2
                                                          Wi nd Tunne I
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance Cm}
Figure A-2.
Comparison of ground-level lateral normalized
concentration profiles  for the Base  Case at a
downwind  distance of  0  meters from the center of
the area.   ISCST2 model results are  based on
dispersion parameters fitted to wind tunnel lata.
                                 A-3

-------
  Normalized Concentration CCU/Q,  Vm**2}

  1000F	
   100 -
   10 -
    1  -
   a 1-
   0.0'
                                       j	,
                                                           ISCST2
                                                          W i nd TunneI
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure A-3.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Base Case at a
downwind  distance of 120  meters from  the center  of
the area  (downwind edge).   ISCST2 model  results
are based on dispersion parameter
tunnel data.
                                 A-4

-------
  Normalized Concentration CCU/Q, 1/m**23

  1000^—	
   100
    10
   0 1
   0 0-
                                       _L
                                                            ISCST2
                                                          Wind Tunnel
     -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure A-4.
Comparison of ground-level lateral  normalized
concentration profiles  for the Base Case at a
downwind  distance of  240  meters from the center  of
the area.   ISCST2 model results are based on
dispersion parameters fitted to wind tunnel data.
                                 A-5

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	
   100
   10
   0 1
   0 0
                                                            ISCST2
                                                          Wi nd Tunne I
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance C"0
Figure A-5.
Comparison of ground-level lateral normalized
concentration profiles  for the Base  Case at a
downwind  distance of 360  meters from the center  of
the area.   ISCST2 model results are  based on
dispersion parameters fitted to wind tunnel data.
                                 A-6

-------
  Normalized Concentration CCU/Q, 1/m**2;)

  1000p	
   100
   0. 1
  0 01
                      _L
                                            ISCST2
                                                          Wi nd Tunne I
    -600 -500 -400 -300 -200 -100  0   100  200   300  400  500   600
                          Crosswind Distance
Figure A-6.
Comparison of ground-level lateral  normalized
concentration profiles  for the Base Case at a
downwind  distance of  840 meters from the center of
the area.   ISCST2 model results are based on
dispersion parameters  fitted to wind tunnel dara.
                                 A-7

-------
  Normalized Concentration CCU/Q., 1/m**2}
   100
    10
   0 01
                     I  I  I  !  I  I  I  I   I
                                                           ISCST2
                                                          W i nd Tunne I
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  BOO
                          Crosswind Distance
Figure A-7.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Base Case  at a
downwind  distance of 1800 meters from  the center
of the area.   ISCST2 model  results are based on
dispersion oarameters fitted to wind tunnel iaca
                                 A-8

-------
  Normalized Concentration CCU/Q,  1/m**23

  -lOOOr	———
   100
    10
   0 1
   0 0
                                                           ISC5T2
                                                          Wind Tunnel
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure A-8.
Comparison of ground-level lateral normalized
concentration profiles  for the Base  Case at a
downwind  distance of 2580  meters from the center
of the  area.   ISCST2 model results are based on
dispersion parameters fitted to wind tunnel data
                                 A-9

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
                                                           ISCST2
                                                          Wind Tunnel
    0 01
01         1          10         100        1000
      Normalized Concentration CCU/OJ 1/m**22)
Figure A-9.
 Comparison of vertical  normalized concentration
 profiles for the Base Case at a downwind distance
 of  -60  meters from the  center of the  area.  ISCST2
 model  results are based on dispersion parameters
 fitted to wind tunnel data.
                                A-10

-------
  Receptor Height

  300
  250
  200
  150
  100
   0
                                                           ISCST2
                                                          Wi nd Tunne I
    0.01
0.1         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**2}
Figure A-10.
 Comparison of vertical  normalized  concentration
 profiles  for the Base Case at a downwind distance
 of  0 meters from the center of the area.   ISCST2
 model  results are based on dispersion parameters
 fitted to wind tunnel data.
                                A-ll

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
                                                           ISCST2
                                                          Wind Tunnel
    0,01
0.1         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure A-11.
 Comparison of vertical  normalized concentration
 profiles for the Base Case at a downwind distance
 of  120  meters from the  center of the  area
  (downwind edge).  ISCST2  model results  are based
 on  dispersion parameters  fitted to wind tunnel
 data.
                                A-12

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    0 01
                                                           ISCST2
                                                          Wi nd TunneI
01         1          10         100        1000
      Normalized Concentration CCU/G, 1/m**23
Figure A-12.
 Comparison of vertical normalized concentration
 profiles  for the Base Case at a downwind distance
 of 240 meters from the center of the area.   ISCST2
 model results are based  on dispersion parameters
 fitted to wind tunnel data.
                                A-13

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
                                                           ISCST2
                                                          W i nd TunneI
    0 01
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure A-13.
 Comparison of vertical  normalized concentration
 profiles for the Base Case at a downwind distance
 of 360  meters from the  center of the  area.   ISCST2
 model results are based on dispersion parameters
 fitted  to wind tunnel data.
                                A-14

-------
  Receptor Height

  300
  250 -
  200
  150
  100
   50
   0
    0 01
                                                           ISCST2
                                                          Wi nd Tunne I
01         1          10         100        1000
      Normalized Concentration CCU/Q., 1/fn**2}
Figure A-14.
 Comparison of vertical  normalized concentration
 profiles  for the Base Case at a downwind distance
 of 840  meters from the  center of the  area.   ISCST2
 model results are based on dispersion parameters
 fitted  to wind tunnel data.
                                A-15

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
                                                           ISCST2
                                                          W i nd TunneI
    0 01
01         1          10         100        10QO
      Normalized Concentrati an CCU/Q, 1/m**2}
Figure A-15.
 Comparison of vertical  normalized concentration
 profiles for the Base Case at a downwind distance
 of 1800  meters from the center of the  area.
 ISCST2 model results are based on dispersion
 parameters fitted to wind tunnel data.
                                A-16

-------
  Receptor Height Cm3

  300
  250
  200
  150
  100
   50
   0
                                                  i I  i
                                                           ISCST2
                                                          Wi nd TunneI
    0.01
0.1         1          10         100        1000
      Normalized Concentration (CU/Q, l/m**2;)
Figure A-16.
 Comparison of vertical  normalized concentration
 profiles  for the Base Case at a downwind distance
 of 2580 meters from  the center of the  area.
 ISCST2 model results are based on dispersion
 parameters fitted to wind tunnel data.
                                A-17

-------
  Normalized Concentration CCU/Q, 1/m**23

  1000c	
   100
   10
   0 1
  0 0
                i  I  i
                                            I  I  I
                                                           ISCST2
                                                          Wi nd TunneI
    -600 -500 -400 -300  -200 -100  0    100  200  300  400  500  BOO
                          Crosswind Distance CmD
Figure A-17.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind distance
of 840 meters from the  center of the area and a
receptor  elevation of 1.8 meters above  ground.
ISCST2 model results are based on dispersion
parameters fitted to wind tunnel data.
                                A-18

-------
  Normalized Concentration CCU/Q, 1/m**2;)

  1000e	
   100
    10
   0 1
   0 011
                                                           ISCST2
                                                          Wi nd Tunne I
     -600 -500 -400 -300  -200 -100  0    100  200  300  400  500  600
                          Crosswmd Distance
Figure A-18.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind  distance
of 840 meters from the  center of the area and a
receptor  elevation of 10.8  meters above ground.
ISCST2 model results are  based on dispersion
parameters fitted to wind tunnel data.
                                A-19

-------
  Normalized Concentration CCU/0J 1/m**2}

  1000|=	
   100
   10
   0 1
   0 0
                    i   i	i  i   i  i
                                                            5CST2
                                                          Wi nd Tunnei
    -600 -500 -400 -300  -200 -100  0    100  200  300  400   500  600
                          Crosswind Distance
Figure A-19.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind  distance
of 840 meters from the center of the area and a
receptor  elevation of 19.8  meters above ground.
ISCST2 model results are based on dispersion
parameters fitted to wind tunnel data.
                                A-20

-------
  Normalized Concentration CCU/Q, 1/m**2}

  lOOQc	
   100
    10
   0 1
   0.01
                                                           I SCST2
                                                          W i nd Tunne I
    -600 -500 -400 -300  -200 -100  0    100  200  300  400  500  600
                          Crosswind Distance
Figure A-20.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind  distance
of 840 meters from the center of the area and a
receptor  elevation of 28.8  meters above ground.
ISCST2 model results are based on dispersion
parameters fitted to wind tunnel data.
                                A-21

-------
  Normalized Concentration CCU/Q, 1/m**2}

  10QOp	
   100 -
   10
    1  -
   0 1-
  0 0
                                                           I SCST2
                                                          W i nd TunneI
    -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure A-21
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind  distance
of 840 meters from the center of the area and a
receptor  elevation of 37.8  meters above ground.
ISCST2 model results are based on dispersion
parameters fitted to wind tunnel data.
                                A-22

-------
  Normalized Concentration CCU/Q,  1/m**2;i

  1000c	
   100
    10
   0 1
   o oi
                                                           ISCST2
                                                          Wi nd Tunne I
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance CmD
Figure A-22.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind distance
of 840 meters from the  center of the area and a
receptor  elevation of 46.8  meters above  ground.
ISCST2 model results are  based on dispersion
parameters fitted to wind tunnel data.
                                A-23

-------
  Normalized Concentration CCU/Q, 1/m**2}
   100
   10
   Q 1
   0 0
                                                           SCST2
                                                         Wi nd TunneI
    -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance Cm3
Figure A-23.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind  distance
of 840 meters from the center of the area and a
receptor  elevation of 55.8  meters above ground.
ISCST2 model results are based on dispersion
parameters fitted to wind tunnel data.
                                A-24

-------
  Normalized Concentration CCU/Q, 1/m**2D

  -lOOOp—	
   100
   10
   0. 1
   0 0
                                          i  i  i  i  i   i  i
                                           ISCST2
                                                          Wi rid Tunne I
    -600  -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure A-24.
Comparison of lateral normalized concentration
profiles  for the Base Case at a downwind distance
of 840 meters from the  center of the  area and a
receptor  elevation of 64.8 meters above  ground.
ISCST2 model results are  based on dispersion
parameters fitted to wind tunnel data.
                                A-25

-------
  Normalized Concentration CCU/Q,  1/m**2}
   100
   10
   0 1
  0 0
                                                          ^ ISCST2
                                                          	*	

                                                          Wi nd TunneI
                                                          	1	
    -600  -500 -400 -300  -200 -100  0   100  200  300   400  500  600
                          Crosswind Distance Cm3
Figure A-25.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind distance
of 840 meters from the  center of the area and a
receptor  elevation of 73.8  meters above  ground.
ISCST2 model results are  based on dispersion
parameters fitted to wind tunnel data.
                                A-26

-------
  Normalized Concentration CCU/Q,
   100
    10
   0. 1
   o en
                                                           I5CST2
                                                          Wind TunneI
    -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance CmD
Figure A-26.
Comparison of lateral normalized concentration
profiles  for the Base Case  at  a downwind distance
of 840 meters from the center  of the area  and a
receptor  elevation of 82.8  meters above ground.
ISCST2 model results are based on dispersion
parameters fitted to wind tunnel data.
                                A-27

-------
  Normalized Concentration CCU/Q, 1/m**2}

  10QOc	
   100
   10
   0 1
  0 0
                                                           ISCST2
                                                          Wi nd TunneI
    -600 -500 -400 -300  -200 -100  0    100  200  300  400  500  600
                          Crosswind Distance
Figure A-27.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind  distance
of 840 meters from the  center of the area and a
receptor  elevation of 91.8  meters above ground.
ISCST2 model results are  based on dispersion
parameters fitted to wind tunnel data.
                                A-28

-------
  Normalized Concentration C.CU/Q, 1/m**2;)
   100 -
   0. 1
   o o-
                                                           I5CST2
                                                          Wi nd Tunne I
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure A-28.
Comparison of lateral normalized concentration
profiles  for the Base Case  at a downwind distance
of 840 meters from the  center of the area and a
receptor  elevation of 100.8 meters above ground.
ISCST2 model results are  based on dispersion
parameters fitted to wind tunnel data.
                                A-29

-------
           APPENDIX B









POINT-TO-POINT COMPARISON RESULTS




 FOR THE  45-DEGREE ROTATED  CASE

-------
     This Appendix presents the point-to-point comparison results
for the 45-Degree Rotated Case scenario.  The 45-Degree Rotated
Case consists of the 720m x 240m base case rectangle rotated 45
degrees counterclockwise in a simulated boundary layer with a 0.2
meter surface roughness length and a free-stream wind speed of 4
m/s.  The ground-level lateral concentration profiles (XY plane)
are presented first, followed by the centerline vertical profiles
(XZ plane),  and then the vertical crosswind profiles (YZ plane)
at a downwind distance of 840 meters.  The ground-level profiles
are based on a 1.8 meter receptor height above ground.   The
vertical crosswind profiles for receptor heights above about 100
meters have not been included in this report.  Normalized
concentration values of less than 0.01 have also been excluded
from the graphs.  As noted in the captions, the ISCST2 model
results are based on the use of dispersion parameters fitted to
wind tunnel data.  The use of wind tunnel sigmas is discussed in
Section 8.
                               B-l

-------
  Normalized Concentration CCU/Q, 1/m**2D
   100
   10
   0 1
  0 01
                I  I  I   I  I
                                 I	I
                                                            SCST2
                                                          Wind Tunnel
    -600 -500 -400 -300  -200 -100  0    100  200  300  400   500  600
                          Crosswmd  Distance
Figure B-l.
Comparison of ground-level  lateral normalized
concentration profiles  for  the 45-Degree  Rotated
Case at  a  downwind distance of -170 meters from
the center of the area.   ISCST2 model  results ar;
based on dispersion parameters fitted  to  wind
tunnel data.
                                 B-2

-------
  Normalized Concentration CCU/Q, 1/m**23

  1000F	
   100
    10
   0 1
   0 01
                                                           ISC5T2  '
                                                          	X	

                                                          Wind TunneI
                                                          	1	
     -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure B-2.
Comparison of ground-level lateral normalized
concentration profiles  for the 45-Degree Rotated
Case at  a downwind distance of 0, meters from the
center of the area.   ISCST2 model results are
based on dispersion parameters fitted to wind
tunnel data.
                                 B-3

-------
  Normalized Concentration CCU/Q, 1/m**23

  1000=	
   100
    10
   0 1
   0 0'
                         I  I  I  I  I
                                                            ISC3T2
                                                          	*	

                                                          Wi nd Tunne
                                                          	1	
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance CmD
Figure B-3.
Comparison of ground-level lateral normalized
concentration profiles  for the 45-Degree Rotated
Case at  a downwind distance of 170 meters from  the
center of the area.   ISCST2 model results are
based on dispersion parameters fitted to wind
tunnel data.
                                 B-4

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000p	
   100
    10
   0 1
                                                            SCST2
                                                          Wi nd Tunne
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure B-4.
Comparison of ground-level  lateral normalized
concentration profiles  for  the 45-Degree Rotated
Case at  a downwind distance of 340 meters from the
center of the area.   ISCST2 model results are
based on dispersion parameters fitted  to wind
tunnel data.
                                 3-5

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	
   100
   10
   0 1
   Q 01"
                                                           ISCST2
                                                          Wi nd TunneI
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure B-5.
Comparison of ground-level  lateral normalized
concentration profiles  for  the 45-Degree Rotated
Case at  a  downwind distance of 480 meters from the
center of  the area.   ISCST2 model results are
based on dispersion parameters fitted  to wind
tunnel data.
                                 E-6

-------
  Normalized Concentration CCU/Q, 1/m**2}
   100 -
   10
   0 1
   0 0
                                                           ISCST2
                                                          Wind Tunne
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure B-6.
Comparison of ground-level  lateral normalized
concentration profiles  for  the 45-Degree  Rotated
Case at  a  downwind distance of 600 meters from the
center of  the area.  ISCST2 model results are
based on dispersion parameters fitted  to  wind
tunnel data.
                                 B-7

-------
  Normalized Concentration CCU/Q, 1/m**2}

  -lODOc	:	
   100
    10
   0 1
   0 0
                                                           ISCST2
                                                          Wi nd TunneI
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure B-7.
Comparison of ground-level  lateral normalized
concentration profiles  for  the 45-Degree Rotated
Case at  a  downwind distance of 840 meters from the
center of'the area.  ISCST2 model results are
based on dispersion parameters fitted  to wind
tunnel data.
                                 B-8

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000^—	
   100 -
   10
   0 1
   Q 0
                                            ISCST2
                                                          Wi nd TunneI
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure B-8.
 Comparison  of  ground-level  lateral normalized
 concentration  profiles for  the 45-Degree  Rotated
'Case at a downwind distance of 1800.meters  from
 the center  of  the area.   ISCST2 model results are
 based on dispersion parameters fitted to  wind
 tunnel data.
                                 B-9

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
   0
                                                           ISCST2
                                                         Wi nd Tunne I
    0 01
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure B-9.
 Comparison of vertical  normalized concentration
 profiles for the 45-Degree Rotated Case  at a
 downwind distance of  0  meters from the center of
 the  area.   ISCST2 model results are based on
 dispersion parameters fitted to wind  tunnel data
                                B-10

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    0 01
                                                           ISCST2
                                                         Wi nd Tunne I
0.1         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure B-10.
 Comparison of vertical  normalized concentration
 profiles  for the 45-Degree Rotated Case  at a
 downwind  distance of  170  meters from  the center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind  tunnel data.
                                B-ll

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
                                                          ISCST2
                                                         Wi nd TunneI
    0 01
01         1          10        100
      Normalized Concentration CCU/Q,
                                       1000
Figure B-ll.
Comparison  of  vertical normalized concentration
profiles  for the 45-Degree Rotated Case at a
downwind  distance of 340 meters  from the center of
the area.   ISCST2 model results  are based on
dispersion  parameters fitted  to  wind tunnel data.
                                B-12

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    Q 01
                                                           ISCST2
                                                         Wi nd Tunne I
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/'m**23
Figure B-12.
 Comparison of vertical normalized concentration
 profiles  for the 45-Degree Rotated Case  at a
 downwind  distance of 480  meters from the center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel daca.
                                B-13

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
    0 01
                                                          ISCST2
                                                         	*	

                                                         Wind Tunnel
                                                         	1	
01         1          10         100
      Normalized Concentration CCU/Q.,
                                       1000
Figure B-13.
Comparison  of vertical normalized concentration
profiles  for the 45-Degree Rotated Case at  a
downwind  distance of 600 meters  from the center of
the area.   ISCST2 model results  are based on
dispersion  parameters fitted  to  wind tunnei daca.
                                B-14

-------
  Receptor Height

  3001
  250
  200
  150
  100
   50
    0
    0 01
                                                           ISCST2
                                                         Wind Tunnel
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure B-14.
 Comparison of vertical normalized concentration
 profiles  for the 45-Degree Rotated Case  at  a
 downwind  distance of 840  meters from the center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel data.
                                B-15

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
    0 01
                                                           ISCST2
                                                         Wi nd TunneI
01         1          10         100        1000
      Normalized Concentration (CU/Q, 1/m**2}
Figure B-15.
 Comparison of vertical  normalized concentration
 profiles for the 45-Degree Rotated Case  at a
 downwind distance of  1800  meters from  the center
 of  the  area.  ISCST2  model results are based on
 dispersion parameters fitted to wind tunnel data.
                                B-16

-------
Normalized Concentration

1000c
                           ,  1/m**2}
   100 -
   10
   0 1
   0 01	L
                                                           ISCST2
                                                         Wi nd Tunne i
     -600 -500 -400 -300 -200  -100  0   100  200  300  400  500  600
                        uCrosswind Distance CmD
Figure B-16.
              Comparison of lateral normalized  concentration
              profiles for the  45-Degree Rotated  Case at a
              downwind distance of  840 meters from the center  of
              the area and a  receptor elevation of 1.3 meters
              above ground.   ISCST2 model results are based en
              dispersion parameters fitted to wind tunnel data.
                                B-17

-------
  Normalized Concentration CCU/Q,  1/m**2}

  lOOOc	
   100
    10
   Q 1
   0.0
                                                           ISCST2
                                                         Wi nd Tunne
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance CrrQ
Figure B-17.
Comparison of lateral normalized concentration
profiles  for the 45-Degree  Rotated Case at  a
downwind  distance of 840 meters from the  center of
the area  and a receptor elevation of 19.3 meters
above ground.  ISCST2 model results are based on
dispersion paramecers ficted ~o wind cunnel aaca.
                                B-18

-------
  Normalized Concentration CCU/Q,  1/m**23
   100
   10
   0 1
   0.0
                                     I  I  I
                                                           ISCST2
                                                         W i nd Tunne
     -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance C"\)
Figure B-18.
Comparison of lateral normalized concentration
profiles  for the 45-Degree  Rotated Case at  a •
downwind  distance of 840 meters from the  center of
the area  and a receptor elevation of 37.8 meters
above ground.  ISCST2 model results are based on
dispersion parameters fitted to wind tunnel data.
                                B-19

-------
  Normalized Concentration CCU/Q, 1/m**23

  1000c	
   100
   10
   0 1
  0.0
                                                           ISCST2
                                                         Wi nd Tunne
     •600 -500 -400 -300 -200  -100  0   100  200  300  400  500  600
                         Crosswind Distance Cm3
Figure B-19.
Comparison of lateral normalized concentration
profiles  for the 45-Degree  Rotated Case at  a
downwind  distance of 840 meters from the  center of
the area  and a receptor elevation of 55.8 meters
above ground.  ISCST2 model results are based on
dispersion parameters fitted to wind tunnel data.
                                B-20

-------
  Normalized Concentration CCU/Q,  1/m**2}

  1000c	
   100
    10
   0 1
   0.0
                                                           ISCST2
                                                         Wi nd Tunne I
     -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  500
                         Crosswind Distance (rrO
Figure B-20.
Comparison of lateral normalized concentration
profiles  for the 45-Degree  Rotated Case at  a
downwind  distance of 840 meters from the  center of
the area  and a receptor elevation of 73.8 meters
above ground.  ISCST2 model results are based on
dispersion parameters fitted to wind tunnel daca.
                                B-21

-------
  Normalized Concentration CCU/Q,  1/m**2}

  1000r	
   100
    10
   0 1
   0 0
                         ,  I  i
                                                           ISCST2
                                                         Wi nd Tunne I
     -600 -500 -400 -300 -200  -100  0   100  200  300  400  500  600
                         Crosswind Distance
Figure B-21.
Comparison  of lateral normalized concentration
profiles  for the 45-Degree Rotated Case at  a
downwind  distance of 840 meters from the center of
the area  and a receptor elevation of 91.8 mecers
above ground.  ISCST2 model  results are based en
dispersion  paramecers fitted to wind tunnel daca.
                                B-22

-------
  Normalized Concentration CCU/Q,  1/m**2}

  1000c	
   100
    10
   0 1
   0 0
                                                           ISC5T2
                                                         Wi nd Tunne I
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance C^O
Figure B-22.
Comparison of lateral normalized concentration
profiles  for the 45-Degree  Rotated Case at  a
downwind  distance of 840 meters from the center of
the area  and a receptor elevation of 109.8  meters
above ground.  ISCST2 model results are based on
dispersion parameters fitted to wind cunnel data.
                               • B-23

-------
           APPENDIX  C









POINT-TO-POINT COMPARISON RESULTS




  FOR THE  90-DEGREE ROTATED CASE

-------
     This Appendix presents the point-to-point comparison results
for the 90-Degree Rotated Case scenario.  The 90-Degree Rotated
Case consists of the 720m x 240m base case rectangle rotated 90
degrees counterclockwise (short dimension perpendicular to the
wind flow) in a simulated boundary layer with a 0.2 meter surface
roughness length and a free-stream wind speed of 4  m/s.   The
ground-level lateral concentration profiles (XY plane) are
presented first, followed by the centerline vertical profiles  (XZ
plane).  There were no vertical crosswind profile (YZ plane)
measurements taken for this scenario.  The ground-level profiles
are based on a 1.8 meter receptor height above ground.
Normalized concentration values of less than 0.01 have been
excluded from the graphs.  As noted in the captions, the ISCST2
model results are based on the use of dispersion parameters
fitted to wind tunnel data.  The use of wind tunnel sigmas is
discussed in Section 8.
                               C-l

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	
   100
    10
   0 1
                                                           ISCST2
                                                          Wi nd Tunne I
     -600 -500 -400 -300 -200 -100  0   100  200  300   400  500  600
                          Crosswind Distance CnO
Figure C-l.
Comparison of ground-level lateral normalized
concentration profiles  for the 90-Degree Rotated
Case at  a downwind distance of -180 meters from
the center of the area.   ISCST2 model  results are
based on dispersion parameters fitted  to wind
tunnel data.
                                 C-2

-------
  Normalized Concentration CCU/Q,  1/m**2}
  ioooF	:	
   100
    10
   0 1
   0 0
                                                     I
                                                            ISCST2

                                                          	*	

                                                          Wi nd TunneI

                                                          	1	
     -600 -500 -400 -300 -200 -100  0   100  200  300  -400  500  60Q
                          Crosswind Distance (rrO
Figure C-2.
Comparison of ground-level lateral  normalized
concentration profiles  for the 90-Degree Rotated
Case at  a downwind distance of 0 meters from the
center of the area.  ISCST2 model results are
based on dispersion parameters fitted to wind
tunnel data.
                                 C-3

-------
  Normalized Concentration CCU/Q, 1/rn**2}

  1000c	
   100
    10
   0 1
   0 0
                                                           I SCST2
                                                          Wi nd TunneI
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure C-3.
Comparison of ground-level lateral normalized
concentration profiles  for the 90-Degree Rotated
Case at  a downwind distance of 180 meters from  the
center of the area.  ISCST2 model results are
based on dispersion parameters fitted to wind
tunnel data.
                                 C-4

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000p	
   100
    10
   0 1
   0 0
          I
                                                           ISCST2
                                                          Wi nd Tunne I
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure C-4.
Comparison of ground-level  lateral normalized
concentration profiles  for  the 90-Degree Rotated
Case at,a downwind distance of 360 meters from the
center  of the area.   ISCST2 model results are
based on dispersion parameters fitted  to wind
tunnel  data.
                                 C-5

-------
  Normalized Concentration CCU/Q, 1/m**2}

  -lOOOp	
   100 -
    10
    1  -
   0 1-
   0 0
        I  I  I
                                                           ISCST2
                                                           i nd Tunne
     -600 -500 -400 -300  -200 -100  0   100  200  300   400  500  600
                          Crosswind Distance
Figure C-5.
 Comparison  of  ground-level  lateral normalized
 concentration  profiles for  the 90-Degree  Rotated
 Case at a downwind distance of 480 meters from the
-center of the  area.  ISCST2 model results are
 based on dispersion parameters fitted  to  wind
 tunnel data.
                                 C-6

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000e	
   100 -
   10
   0 1
   0 0
                                                           I SCST2
                                                          Wi nd Tunne !
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  500
                          Crosswind Distance C^O
Figure C-6.
Comparison of ground-level lateral normalized
concentration profiles  for the 90-Degree Rotated
Case at  a downwind distance of 600 meters from  the
center of the area.   ISCST2 model results are
based on dispersion parameters fitted  to wind
tunnel data.
                                 C-7

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	
   100
   10
   0 1
   Q 0
                                                            ISCST2
                                                          Wi nd TunneI
    -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  500
                          Crosswind Distance Cm}
Figure C-7.
Comparison of ground-level lateral normalized
concentration profiles  for the 90-Degree Rotated
Case at  a downwind distance of 840 meters from  the
center of the area.  ISCST2 model results are
based on dispersion parameters fitted to wind
tunnel data.
                                 C-8

-------
  Normalized Concentration CCU/Q, 1/m**2;)
   100
    10
   0 1
   0 0
                                                           ISCST2
                                                          Wi nd TunneI
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance CnO
Figure C-8.
Comparison of ground-level  lateral normalized
concentration profiles  for  the 90-Degree Rotated
Case at  a downwind distance of 1800 meters from
the center of the area.   ISCST2 model  results are
based on dispersion parameters fitted  to wind
tunnel data.
                                 C-9

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
    0 01
                                                           ISCST2
                                                         Wi nd Tunne I
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure C-9.
 Comparison of vertical normalized concentration
 profiles  for the 90-Degree Rotated Case  at a
 downwind  distance of  -180  meters from  the  center
 of  the  area.  ISCST2  model results are based on
 dispersion parameters fitted to wind tunnel data.
                                C-10

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    0 01
                                                           ISCST2
                                                         Wi nd TunneI
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**2;>
Figure C-10.
 Comparison of vertical  normalized concentration
 profiles  for the 90-Degree Rotated Case  at  a
 downwind  distance of 0  meters from the center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel data
                                C-ll

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    0 01
                                                           ISCST2
                                                         Wi nd TunneI
01         1          10
      Normalized Concentration
                                             100        1000
                                             ,  1/m**2}
Figure C-ll.
 Comparison of vertical normalized concentration
 profiles  for the 90-Degree  Rotated Case at  a
 downwind  distance of 180 meters from the center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel daca.
                                C-12

-------
  Receptor  Height

  300
  250
  2DO
  150
  100
   50
   0
    0 01
                                                           1SCST2
                                                         Wi nd TunneI
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure C-12.
 Comparison of vertical normalized concentration
 profiles  for the 90-Degree  Rotated Case  at  a
 downwind  distance of 360 meters from the center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel data.
                                C-13

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
                                                           ISCST2
                                                         Wi nd Tunne I
               0 1
          1         10        100       1000
     Normalized Concentration CCU/Q,  1/m**23
Figure C-13.
Comparison of vertical normalized concentration
profiles  for the 90-Degree Rotated Case at  a
downwind  distance of 480 meters from the  center of
the area.   ISCST2 model results are -based on
dispersion parameters fitted to wind tunnel data.
                                C-14

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    0 01
                                                           ISCST2
                                                          Wi nd TunneI
0.1         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**2;)
Figure C-14.
 Comparison of vertical  normalized concentration
 profiles  for the 90-Degree Rotated Case  at a
 downwind  distance of  600  meters from  the center of
 the area.   ISCST2 model results are based on
 dispersion parameters fitted to wind  tunnel data.
                                C-15

-------
  Receptor Height

  300
                                                           I SCST2
                                                         Wi nd Tunne I
  250 -
  200 -
  150 -
  100 -
    0 01
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure C-15.
 Comparison of vertical normalized concentration
 profiles  for the 90-Degree Rotated Case  at  a
 downwind  distance of 840  meters from the center of
 the  area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel data.
                                C-16

-------
  Receptor Height

  300
  250 -
  200
  150
  100
   50
   0
                                                           ISCST2

                                                         	*	

                                                         Wi nd TunneI

                                                         	1	
    0 01
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**2;)
Figure C-16.
 Comparison of vertical normalized concentration
 profiles  for the 90-Degree  Rotated Case  at  a
 downwind  distance of 1800 meters from  the  center
 of the  area.  ISCST2 model  results are based on
 dispersion parameters fitted to wind tunnel data.
                                C-17

-------
           APPENDIX D









POINT-TO-POINT COMPARISON RESULTS




      FOR THE CIRCULAR CASE

-------
     This Appendix presents the point-to-point comparison results
for the Circular Case scenario.  The Circular Case consists of a
circular source with a diameter of 470 meters (equivalent area to
the base case rectangle) in a simulated boundary layer with a 0.2
meter surface roughness length and a free-stream wind speed of 4
m/s.  The ground-level lateral concentration profiles (XY plane)
are presented first, followed by the centerline vertical profiles
(XZ plane),  and then the vertical crosswind profiles (YZ plane)
at a downwind distance of 840 meters.  The ground-level profiles
are based on a 1.8 meter receptor height above ground.   The
vertical crosswind profiles for receptor heights above about 100
meters have not been included in this report.  Normalized
concentration values of less than 0.01 have also been excluded
from the graphs.  As noted in the captions,  the ISCST2  model
results are based on the use of dispersion parameters fitted to
wind tunnel data.  The use of wind tunnel sigmas is discussed in
Section 8.
                               D-l

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	
   100
   10
   0 '1
   0 0
                    I   I  I
                                                           ISCST2
                                                          Wi nd TunneI
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-l.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Circular Case at a
downwind  distance of  -102 meters from  the center
of the  area.   ISCST2  model  results are based on
dispersion parameters fitted to wind tunnel data.
                                 D-2

-------
  Normalized Concentration CCU/Q, 1/m**2}
   100 -
   10
   0 1
   0 0
                                                           I5CST2
                                                          Wi nd Tunne I
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-2.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Circular Case at a
downwind  distance of 0  meters from the  center of
the area.   ISCST2 model results are based on
dispersion parameters fitted to wind  tunnel data.
                                 D-3

-------
  Normalized Concentration CCU/Q, 1/m**2}

  -lOOOp	
   100
   0 1
  o o-
                    i   i  i
                                                           ISCST2
                                                          W i nd TunneI
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-3.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Circular Case at  a
downwind  distance of 102  meters from  the center  of
the area.   ISCST2 model results are based on
dispersion parameters fitted to wind  tunnel data.
                                 D-4

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	
   100
    10
   0 1
   0 0
                                                           ISCST2
                                                          Wi nd Tunne I
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-4.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Circular Case at  a
downwind  distance of 204  meters from  the center  of
the area.   ISCST2 model results are based on
dispersion parameters fitted to wind  tunnel data.
                                 D-5

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000n	
   100 -
    10 -
   0 1-
   0 0
                                                            ISCST2
                                                          Wind Tunne
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-5.
Comparison of ground-level lateral normalized
concentration profiles  for the Circular Case at  a
downwind distance of 240  meters from the center  of
the area.   ISCST2 model results are  based on
dispersion tjarameters fitted to wind tunnel data.
                                 D-6

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000c	—	
   100
    10
   0 1
   0 0
                                          ',  I  '.  'l   I
                                                           ISCST2
                                                          Wi nd Tunne
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-6.
Comparison of ground-level lateral normalized
concentration profiles  for the Circular Case at  a
downwind  distance of 360  meters from  the center  of
the area.   ISCST2 model results are based on
dispersion parameters fitted to wind  tunnel data.
                                 D-7

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000F	
   100
    10
   0 1
   0 Oi
        i  i  i
                     i  i  i
                                                           ISCST2

                                                             )K	

                                                          Wi nd TunneI
    -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-7.
Comparison of ground-level lateral normalized
concentration profiles  for the Circular Case at  a
downwind  distance of 480  meters from  the center  of
the area.   ISCST2 model results are based on
dispersion parameters fitted to wind  tunnel data.
                                 D-8

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000F	
   100
   10
   0 1
   0 CH
          I
                                           ISCST2

                                          	*	

                                          Wi nd Tunne

                                          	1	
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-8.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Circular Case at a
downwind  distance of 840  meters from  the center of
the area.   ISCST2 model  results are based on
dispersion parameters fitted to wind  tunnel data.
                                 D-9

-------
  Normalized Concentration CCU/Q, 1/m**2}

  1000p	
   100
    10
   0 1
   0 0
                                                           ISCST2
                                                          Wi nd Tunne I
    -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                          Crosswind Distance
Figure D-9.
Comparison of ground-level  lateral normalized
concentration profiles  for  the Circular Case at a
downwind  distance of 1800 meters from  the center
of the area.   ISCST2 model  results are based on
dispersion narameters fitted to wind tunnel data.
                                D-10

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
                                                           ISCST2
                                                         Wi nd Tunne I
    0 01
0.1         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure D-10.
 Comparison of vertical  normalized concentration
 profiles  for the Circular Case at a downwind
 distance  of -102 meters from the center of the
 area.   ISCST2 model results are based  on
 dispersion parameters  fitted to wind tunnel data
                                D-ll

-------
  Receptor Height

  300
  250
  200
  150
  100
   50
   0
    0 01
                                                           SCST2
                                                         Wi nd TunneI
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure D-ll.
 Comparison of vertical  normalized concentration
 profiles for the Circular Case at a downwind
 distance of 0 meters  from the center of  the area
 ISCST2  model results  are  based on dispersion
 parameters fitted to  wind tunnel data.
                                D-12

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    0 01
                                                           SCST2
                                                          Wind Tunnel
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure D-12.
 Comparison of vertical  normalized concentration
 profiles for the Circular Case at a downwind
 distance of 102 meters  from the center  of the
 area.   ISCST2 model  results are based on
 dispersion parameters  fitted to wind tunnel data.
                                D-13

-------
  Receptor Height

  300
  250 -
  200
  150
  100
   50
   0
    0 01
                                                           ISCST2
                                                         Wi nd TunneI
0.1         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**2;)
Figure D-13.
 Comparison of vertical  normalized concentration
 profiles for the Circular Case at a downwind
 distance of 204 meters  from the center of the
 area.   ISCST2 model  results are based  on
 dispersion parameters  fitted to wind tunnel dat;
                                D-14

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
    0 01
                                                           ISCST2
                                                         Wi nd Tunne I
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**23
Figure D-14.
 Comparison of vertical  normalized concentration
 profiles  for the Circular Case at a downwind
 distance  of 240 meters  from the center  of  the
 area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel data
                                D-15

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
   0
    0 01
                                                           ISCST2
                                                         Wi nd Tunne I
01         1          10         100        1000
      Normalized Concentration CCU/Q, 1/m**2)
Figure D-15.
 Comparison of vertical  normalized concentration
 profiles for the Circular Case at a downwind
 distance of 360 meters  from the center  of the
 area.   ISCST2 model results are based on
 dispersion parameters fitted to wind tunnel data
                                D-16

-------
  Receptor  Height

  300
  250
  200
  150
  100
   50
    0
                                                           ISCST2
                                                          Wind Tunnel
    0 01
01         1          10         100        1QOQ
      Normalized Concentration CCU/Q, 1/m**2}
Figure D-16.
 Comparison of vertical  normalized concentration
 profiles  for the Circular Case at a downwind
 distance  of 480 meters  from the center  of the
 area.   ISCST2 model' results are based on
 dispersion parameters  fitted to wind tunnel data
                                D-17

-------
  Receptor  Height

  300
                                                           ISCST2
                                                         Wi nd Tunne I
  250 -
  200 -
  150 -
  100 -
    0 01
01         1          10         100
      Normaliz.ed Concentration CCU/Q,
                                       1000
Figure D-17.
Comparison  of  vertical normalized concentration
profiles  for the Circular Case  at a downwind
distance  of 840 meters from  the center of the
area.   ISCST2  model results  are based on
dispersion  parameters fitted to wind tunnel data
                                D-18

-------
  Receptor  Height

  300
  250 -
  200 -
  150 -
                                                           ISCST2

                                                         	*	

                                                         Wi nd TunneI
                                                         	1	
  100 -
    0 01
01         1          10         100
      Normalized Concentration CCU/Q, I/ ' m
                                                       1000
Figure D-18.
 Comparison of vertical normalized concentration
 profiles  for the Circular  Case at a downwind
 distance  of 1800 meters  from the center  of  the
 area.   ISCST2 model results  are based on
 dispersion parameters fitted to wind tunnel data
                                D-19

-------
  Normalized Concentration CCU/Q,  Vm**2;)

  1000c	
   100
    10
   0 1
   0.0
                                              i	i
                                                           ISCST2
                                                         W i nd Tunne I
     -600 -500 -400 -300 -200 -100  0   100  200  300  400  500  BOO
                         Crosswind Distance Cm3
Figure D-19.
Comparison of lateral normalized concentration
profiles  for the Circular  Case at a downwind
distance  of 840 meters from the center of  the area
and a receptor elevation of 1.8 meters above
ground.   ISCST2 model results  are based on
dispersion parameters fitted co wind cunnel  daca.
                                D-20

-------
  Normalized Concentration CCU/Q,  1/m**23

  1000r	
   100
    10
   0 1
   o.oi
          !
                                                           ISCST2
                                                         Wi nd TunneI
     -600 -500 -400 -300  -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance
Figure D-2Q.
Comparison of lateral normalized concentration
profiles  for the Circular  Case at a downwind
distance "of 840 meters  from the center of  the area
and a receptor elevation of 19.8 meters  above
ground.   ISCST2 model results are based  en
dispersion parameters fitted to wind cunnel daca.
                                D-21

-------
  Normalized Concentration CCU/Q,  Vm**2}

  1000=	
   100
    10
   a 1
   o.o
                                                           ISCST2
                                                         Wi nd Tunne
     -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance
Figure D-21.
Comparison of lateral normalized concentration
profiles  for the Circular  Case at a downwind
distance  of 840 meters from the center of  the area
and a receptor elevation of 37.3 meters above
ground.   ISCST2 model results  are based on
dispersion parameters fitted to wind cunnel  daca.
                                D-22

-------
  Normalized Concentration CCU/Q,  1/m**23
   100
    10
   0 1
   0.0
                                 I  I  I
                                                           ISC5T2
                                                         Wi nd TunneI
     -600 -500 -400 -300 -200 -100  0   100 200  300  400  500  600
                         Crosswind Distance C^D
Figure D-22.
Comparison of lateral normalized concentration
profiles  for the Circular Case at a downwind
distance  of 840 meters from the center of  the area
and a receptor elevation of 55.8 meters above
ground.   ISCST2 model results  are based on
dispersion parameters fitted to wind tunnel  data.
                                D-23

-------
  Normalized Concentration CCU/Q,  1/m**2}

  1000F	
   100
   10
   0 1
   0.01
                                                  I  I  I
                                                           ISCST2
                                                         Wind Tunnel
     -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance
Figure D-23.
Comparison of lateral normalized concentration
profiles  for the Circular  Case at a downwind
distance  of 840 meters from the center of  the area
and a receptor elevation of 73.8 meters  above
ground.   ISCST2 model results are based  on
dispersion parameters fitted to wind tunnel data.
                                D-24

-------
  Normalized Concentration CCU/Q,  1/m**2;)

  -\ooo<=-
   100 -
                                                           ISCST2
                                                         Wi nd Tunne
     -BOO -500 -400 -300 -200 -100  0   100  200  300  400  500  600
                         Crosswind Distance
Figure D-24.
Comparison of lateral normalized concentration
profiles  for the Circular  Case at a downwind
distance  of 840 meters from the center of  the area
and a receptor elevation of 91.8 meters  above
ground.   ISCST2 model results are based  on
dispersion parameters fitted to wind tunnel  data.
                                D-25

-------
  Normalized Concentration CCU/Q,  1/m**2}

  1000c	
   100 -
   10
   0 1
  0 0
                I  I  I
                                                  I  I  I
                                                           ISCST2
                                                         	*	

                                                         Wind Tunnel
                                                         	1	
    -500 -500 -400 -300 -200  -100  0   100  200  300  400  500  BOO
                         Crosswind Distance
Figure D-25.
Comparison of lateral normalized concentration
profiles  for the Circular  Case at a downwind
distance  of 840 meters from the center of  the area
and a receptor elevation of 109.8 meters  above
ground.   ISCST2 model results are based on
disoersion oarametars fittad to wind tunnel  aa;a.
                                D-26

-------
           APPENDIX E









RESULTS OF STATISTICAL ANALYSES




        IN TABULAR FORM

-------
     This Appendix presents results of the statistical analyses
in tabular form for all four scenarios.   The tables provide the
values of mean fractional bias (MFB) and standard deviation of
the fractional bias (SD) used to generate the graphical plots in
the main body of the report.  Also included on the tables is the
mean ration of observed (wind tunnel)  to predicted (ISCST2)
values (OB/PR) and the number of data points used from each
profile.   For each scenario, the ground-level lateral
concentration profiles  (XY plane) are presented first, followed
by the centerline vertical profiles (XZ  plane),  and then the
vertical  crosswind profiles (YZ plane)  at a downwind distance of
840 meters.  The ground-level profiles are based on a 1.8 meter
receptor height above ground.   The vertical crosswind profiles
for receptor heights above about 100 meters were not included in
the graphical plots presented in the main body of the report,  for
consistency with the range of measurements used for the Base Case
and because the number of data points drops off.   Normalized
concentration values of less than 0.01 were also excluded from
the graphs.  The ISCST2 model results used to generate these
statistical results are based on the use of dispersion parameters
fitted to wind tunnel data.  The use of  wind tunnel sigmas is
discussed in Section 8.
                              E-l

-------
Table E-1.   Statistical Results for the XY Plane for the  Base Case
X (m)
-60.00
.00
120.00
240.00
360.00
840.00
1800.00
2580.00
Z 
-------
Table E-4.   Statistical Results for the XY Plane for the 45-Degree  Rotated Case
X (m)
-170.00
.00
170.00
340.00
480.00
600.00
840.00
1800.00
Z (m)
1.80
1.80
1.80
1.80
1.80
1.80
1.80
1.80
MFB-SD
-.88
-.62
-.41
-.40
-.43
-.37
-.45
-.56
MFB
-.14
-.05
-.13
-.13
-.20
-.22
-.28
-.40
MFB+SD
.60
.52
.16
.13
.03
-.07
-.12
-.24
SD
.74
.57
.29
.26
.23
.15
.17
.16
OB/PR
10.09
5.11
.91
.90
.83
.81
.76
.67
N
14
22
28
30
32
32
38
24
Table E-5.   Statistical Results for the XZ Plane for the 45-Degree  Rotated Case
X (m)
-170.00
.00
170.00
340.00
480.00
600.00
840.00
1800.00
Y (m)
.00
.00
.00
.00
.00
.00
.00
.00
MFB-SD
1.94
.06
-,07
-.22
-.24
-.26
-.24
-.44
MFB
1.97
.54
.40
.40
.47
.42
.56
.30
MFB+SD
2.00
1.02
.88
1.01
1.17
1.11
1.37
1.04
SD
.03
.48
.48
.61
.70
.68
.81
.74
OB /PR
284.36
2.17
1.86
2.17
3.06
2.43
4.51
2.52
N
3
13
16
17
22
25
17
26
Table E-6.   Statistical Results for the YZ Plane for the 45-Degree Rotated Case
X (m)
840.00
840.00
840.00
840.00
840.00
840.00
840.00
840.00
840.00
840.00
Z (m)
1.80
19.80
37.80
55.80
73.80
91.80
109.80
127.80
145.80
163.80
MFB-SD
-.34
-.38
-.38
-.04
.16
.54
.99
1.32
1.63
1.83
MFB
-.27
-.32
-.25
.10
.39
.88
1.26
1.55
1.73
1.87
MFB+SD
-.20
-.26
-.13
.24
.63
1.21
1.52
1.77
1.83
1.90
SD
.07
.06
.13
.14
.23
.33
.27
.23
.10
.03
OB/PR
.76
.72
.78
1.12
1.54
2.91
5.14
11.44
17.11
29.83
N
15
15
15
15
15
15
14
13
8
4
                                    E-3

-------
Table E-7.   Statistical Results for the XY Plane  for the 90-Degree Rotated Case
X (m) Z
-180.00
.00
180.00
360.00
480.00
600.00
840.00
1800.00
(m)
1.80
.80
.80
.80
.80
.80
.80
.80
MFB-SD
-.37
-.19
-.36
-.20
-.30
-.31
-.35
-.56
MFB
-.19
.01
-.13
-.04
-.09
-.18
-.25
-.32
MFB+SD
-.01
.21
.10
.13
.12
-.05
-.15
-.09
SD
.18
.20
.23
.17
.21
.13
.10
.24
OB/PR
.84
1.03
.90
.98
.94
.84
.78
.74
N
23
25
28
34
30
25
29
40
Table E-8.   Statistical Results for the XZ Plane  for the 90-Degree Rotated Case
X (m)
-180.00
.00
180.00
360.00
480.00
600.00
840.00
1800.00
Y (m)
.00
.00
.00
.00
.00
.00
.00
.00
MFB-SD
-.31
-.03
-.27
-.15
-.30
-.30
-.49
-.61
MFB
-.09
.22
.11
.15
.11
.23
.07
-.13
MFB+SD
.12
.48
.48
.44
.52
.76
.63
.36
SD
.21
.26
.37
.30
.41
.53
.56
.49
OB/PR
.93
1.30
1.21
1.22
1.24
1.56
1.41
1.03
N
11
17
15
16
16
19
14
18
                                    E-4

-------
 Table E-9.   Statistical Results for the XY Plane for  the  Circular Case
X (m)
-102.00
.00
102.00
204.00
240.00
360.00
480.00
840.00
1800.00
Z (m)
1.80
1.80
1.80
1.80
1.80
1.80
1.80
1.80
1.80
MFB-SD
-.17
-.07
-.12
-.05
-.09
-.14
-.19
-.28
-.43
MFB
.23
.09
.05
.10
.01
-.04
-.11
-.18
-.33
MFB+SD
.63
.25
.22
.24
.10
.07
-.02
-.09
-.23
SD
.40
.16
.17
.15
.09
.11
.08
.09
.10
OB/PR
1.63
1.11
1.07
1.12
1.01
.97
.90
.83
.72
N
28
27
32
29
31
33
35
38
49
Table E-10.  Statistical  Results for the XZ Plane for the Circular  Case
X (m)
-102.00
.00
102.00
204.00
240.00
360.00
480.00
840.00
1800.00
Y (m)
.00
.00
.00
.00
.00
.00
.00
.00
.00
MFB-SD
-.19
-.12
-.15
-.11
-.15
-.34
-.33
-.35
-.59
MFB
.22
.17
.22
.13
.13
.11
.19
.10
.00
MFB+SD
.62
.45
.59
.37
.41
.56
.72
.55
.60
SD
.40
.28
.37
.24
.28
.45
.53
.45
.60
OB/PR
1.37
1.23
1.36
1.17
1.20
1.28
1.50
1.24
1.37
N
12
10
15
15
15
13
16
13
19
Table E-11.   Statistical Results for the YZ Plane for the Circular Case
X (m)
840.00
840.00
840.00
840.00
840.00
840.00
840.00
840.00
840.00
840.00
Z (m)
1.80
19.80
37.80
55.80
73.80
91.80
109.80
127.80
145.80
163.80
MFB-SD
-.30
-.37
-.40
-.05
.12
.54
.96
1.33
1.65
1.89
MFB
-.23
-.29
-.21
.13
.30
.75
1.13
1.45
1.73
1.91
MFB+SD
-.17
-.20
-.01
.31
.48
.96
1.30
1.57
1.80
1.92
SD
.07
.09
.20
.18
.18
.21
.17
.12
.07
.01
OB/PR
.79
.75
.83
1.16
1.38
2.29
3.74
6.62
14.88
41.54
N
11
11
12
12
11
11
11
9
7
2
                                E-5

-------
                                    TECHNICAL REPORT DATA
                  (Please read Instructions on reverse before completing)
1. REPORT NO.
   EPA-454/R-92-014
                                                          3. RECIPIENT'S ACCESSION NO.
4.  TITLE AND SUBTITLE

   Comparison of a Revised Area Source Algorithm for
   the  Industrial Source Complex  Short Term Model and
   Wind Tunnel Data
                                                          5. REPORT DATE
                                                            October 1992
                                                          6. PERFORMING ORGANIZATION CODE
7.  AUTHOR(S)
                                                          8. PERFORMING ORGANIZATION REPORT NO.
9.  PERFORMING ORGANIZATION NAME AND ADDRESS

   Pacific Environmental Services
   5001  South Miami Boulevard
   Post  Office Box 12077
   Research Triangle  Park, NC  27709-2077
                                                          10. PROGRAM ELEMENT NO.
                                                          11. CONTRACT/GRANT NO.   WA No. 1-131
                                                             EPA Contract No. 68 D00124
12.
SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of  Air Quality  Planning and Standards
Technical  Support Division
Research Triangle Park,  NC  27711	
13. TYPE OF REPORT AND PERIOD COVERED
   Final Report
                                                             14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
   EPA  Work Assignment  Manager: • Jawad S.  Touma
16. ABSTRACT

     This  report summarizes the results  of comparison  between a new  numerical
integration algorithm  for modeling area  source dispersion,  as implemented in the
Industrial Source Complex Short Term  (ISCST2) model using wind tunnel  data collected  in
the U.S. Environmental Protection Agency's Fluid Modeling Facility.  Area sources
referred to can be characterized as low  level releases with little buoyance due to
either  momentum or temperature such as landfills or lagoons that are commonly found at
Superfund  sites.  The  results of the  analyses show that the new ISCST2 area source
algoroithm predicts the concentration distribution with relatively good accuracy,
especially for ground-level area sources.   This conclusion seems also  valid for the
various source shapes  and orientations thus suggesting the ability of  the model to
accurately depict the  plume shape for complex source-receptor geometries.
17.
                                   KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                             b. IDENTIFIERS/OPEN ENDED TERMS
                                                                              c. COSATI
                                                                              Field/Group
   Air Pollution
   Toxic Air  Pollutants
   Air Quality Dispersion  Models
                                         Dispersion Modeling
                                         Meteorology
                                         Air  Pollution Control
18. DISTRIBUTION STATEMENT

   Release  Unlimited
                                         19. SECURITY CLASS (Report)
                                             Unclassified
                                             20. SECURITY CLASS (Page)
                                                 Unclassified
                 21. NO. OF PAGES
                 180
                                                                           22. PRICE
EPA Form 2220-1 (Rev. 4-77)
                       PREVIOUS EDITION IS OBSOLETE

-------