v>EPA
United States
Environmental Protection
Agency
Atmospheric Sciences
Research Laboratory
Research Triangle Park NC 27711
Research and Development
December 1987
PROJECT REPORT
COMPLEX TERRAIN MODEL
DEVELOPMENT: FINAL REPORT
-------
EPA COMPLEX TERRAIN MODEL DEVELOPMENT:
FINAL REPORT
by
David G. Strimaitis1
Robert J. Paine2
Bruce A. Egan2
Robert J. Yamartino1
ISigma Research Corp.
Lexington, MA 02173
2ERT, Inc.
Concord, MA 01742
Contract No. 68-02-3421
Project Officer
Peter L. Finkelstein
Meteorology Division
Atmospheric Sciences Research Laboratory
Research Triangle Park, NC 27711
ATMOSPHERIC SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
NOTICE
The information in this document has been funded by the United
States Environmental Protection Agency Under Contract Mo. 68-02-3421
to ERT, Inc. It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication as an
EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
ii
-------
ABSTRACT
The Complex Terrain Model Development (CTMD) project has met its
original objectives of producing an atmospheric dispersion model
appropriate for regulatory agency application to elevated sources of
air pollutants located in mountainous terrain settings. The model
development effort has focused on predicting concentrations.during
stable atmospheric conditions.
The program, initiated in June 1980, has involved the performance
of 4 major field experiments which produced a wealth of data for model
development and verification purposes. The first experiment, held at
Cinder Cone Butte (CCB) in Idaho, involved the extensive use of a
mobile release system to provide a high capture rate of ground-level
concentrations resulting from elevated plumes flowing toward the
butte. The second experiment, at Hogback Ridge (HBR) near Farmington,
New Mexico, featured a very long ridge that provided a site for
testing the importance of terrain aspect ratio on the flow dynamics.
The final field experiments were held at the Tracy Power Plant (TTP)
near Reno, Nevada. This Full Scale Plume Study (FSPS) provided a
large-scale test of the modeling concepts developed. Data were also
obtained from a series of fluid modeling studies performed at EPA's
Fluid Modeling Facility. These tests provided confirmation of some of
the basic theoretical principles adopted in the modeling effort and
provided information on plume behavior as a function of systematic
changes in terrain shapes, release heights and distances to terrain
objects.
The Complex Terrain Dispersion Model (CTDM), fully described in
this report, is an advanced Gaussian model that uses a flow algorithm
to provide terrain-induced plume trajectory and deformation
information. CTDM is suitable for regulatory use, but it requires
substantially more information on terrain and local meteorology than
complex terrain screening models. With simpler data bases, it
demonstrates degraded performance.
The model evaluation effort concentrated first on the use of
field data collected within this program. Subsequent tests were made
with two other data sets obtained from S0£ monitoring networks near
a large paper mill and a large power plant, both located in complex
terrain. Statistical performance results of CTDM were compared with
those of other complex terrain models of current regulatory interest
or use. The model evaluation demonstrates that CTDM has superior
performance in the majority of tests and has consistently good
performance among all the sites. The statistical performance of CTDM
in complex terrain settings is shown to be comparable to the
performance of EPA's current refined flat terrain models in simple
terrain settings.
iii
-------
CONTENTS
Abstract iii
Figures vii
Tables x
Symbols and Abbreviations xii
Acknowledgments xix
1. Introduction 1
2. Overview of the CTDM Program 4
2,1 Background and Overall Program Plan ...... 4
2.2 Field Program 5
2.2.1 Goals and Design . 5
2.2.2 Field Program Results ' 7
2.3 Fluid Modeling Program 9
2.4 Model Design for Regulatory Use 9
2.5 Model Evaluation Program 10
3. Description of the Complex Terrain Dispersion
Model Design 12
3.1 Qualitative Overview 12
3.2 Use of Meteorological and Terrain Information . 20
3.2.1 Meteorological Data 20
3.2.2 Terrain Data 21
3.3 Derivation of Concentration Equations 22
3.3.1 Plume Rise Calculations 22
3.3.2 Dispersion Parameters 25
3.3.3 The LIFT Component 29
3.3.4 Terrain Factors for LIFT: Tz
and Ty 36
3.3.5 Internal Mixing Layer for LIFT 41
3.3.6 Flow Model for LIFT 44
3.3.7 The WRAP Component 64
3.3.8 Model for Streamlines 70
3.3.9 Receptors Not Influenced by Hills ... 74
4. CTDM Evaluation Analysis 76
4.1 Models Evaluated . 77
4.2 Data Sets Used for Evaluation 84
4.3 Statistical Tests and Case-Study Analyses ... 88
4.4 Results of Statistical Evaluation 93
4.5 Results of Case Study Evaluation 123
5. Sensitivity Tests 126
5.1 Workshop Recommendations for Sensitivity
Analyses 126
5.2 Test Design 128
5.3 Test Results 131
5.3.1 Sensitivity to Hc and Hill Shape . . . 131
5.3.2 Sensitivity to Source Position and
Wind Direction: Symmetric Hill and
2D Hill 135
3191F PB 876-260
-------
CONTENTS (Continued)
5.3.3 Sensitivity to Source Position and
Wind Direction: Asymmetric Hills .... 135
5,4 Operational Test on Hill Shape Sensitivity . . .
6. Model Applicability and Limitations ..........
7. Conclusions and Recommendations ............
References ........................... 156
APPENDIXES
A. Exact Solutions to the Linearized Equation for Stratified «.
Flow Over Terraom in Multidimensional Space ...... 161
B. Evaluation Results for Concentrations Paired in Tune,
unpaired in Space .......... ..... .... 180
C. Evaluation Results for Concentrations Paired in Space,
Unpaired in Tisne .......... .......... 192
D. Evaluation Results for Concentrations Paired in Time and
Space .............. ..... ...... 204
E. Evaluation Results by Meteorological Category for
Concentrations Paired in Tissef Unpaired in Space .... 216
F. Scatter Plots of Peak Hourly Model Predictions and
Observations ........ ....... ....... 266
G. Summary of Case Study Analyses of CTDM Predictions at
the Tracer Sites ............. ..... . . 355
H. Contributions of the Fluid Modeling Facility to EPA'S
Complex Terrain Model Development Program. ....... 402
VI
-------
FIGURES
Humber Page
1 Idealized picture of how the dividing-streamline plane
and the plane of stagnation streamlines "cut" into a
plume, allowing material nearer the center of a plume
to contact the surface of a hill . 15
2 Illustration of terrain effect on the vertical
distribution of plume material above He as
modeled in LIFT 16
3 Depiction of plume behavior in CTDH as it is
deflected around a hill as seen from above 17
4 Depiction of plume behavior in CTDM as it passes
over a hill as seen from above 19
5 Relationship among length scales used to specify
inverse polynomial, Gaussian, and elliptical
profiles 23
6 Illustration of the relationship between the crosswind
average concentration profiles at so and-s, and the
plume from one of many point-source elements
representing the flux of material across the plane at
s0 31
7 Illustration of the factor T^ in finite-difference
form 37
8 Illustration of the structure assumed for the
developing internal mixing layer over the hill
above Hc . 43
9 Derivation of the factor Tj, in finite-difference
form for the "backwards-looking" formulation 48
10 Side view of computed streamlines of flow passing
from left to right over a symmetric Gaussian hill
of aspect ratio 1 54
11 Side view of computed streamlines of flow passing
from left to right over a symmetric Gaussian hill
of aspect ratio 2 55
12 Side view of computed streamlines of flow passing
from left to right over an asymmetric Gaussian
hill of aspect ratio 2 along the flow, and 10
across the flow 56
vii
-------
FIGURES (Continued)
Number Pa&§
13 End-on view of streamlines at x=0 over a symmetric
Gaussian hill of aspect ratio 1, and at x=-« . . . . . 57
14 Comparison of lateral positions of streamlines in the
plane ae=0 over symmetric Gaussian hills of aspect
ratio 1 and 2. . 58
15 Illustration of streamline positions in the plane x=0
for an asymmetric hill of aspect ratio 1 and 2 ..... 60
16 Comparison of streamline positions in the plane x=0
observed in tow-tank simulations by Snyder et al. with
those obtained from CTDM 61
17 Typical streamline patterns in two-dimensional flow
around an elliptical cylinder. ............. 65
18 Top view of a plume in two-dimensional flow around
a hill . 67
19 Sketch of the flow around an ideal cylinder of
elliptical cross-section . 69
20 Definition of modeling variables, illustrating in
particular the coordinate system in which the xg-axis
is aligned with the tangent to the stagnation
streamline at the impingement point. .......... 72
21 Tracer gas sampler locations on Cinder Cone Butte ... 85
22 Field experiment layout in the vicinity of the
Hogback Ridge. ..................... 86
23 Terrain features surrounding the Tracy Power Plant as
modeled by CTDM 87
24 Terrain features and monitors in the vicinity of
the Westvaco Luke Mill sa
25 1980 Widows Creek monitoring network with outlines of
terrain features used in CTDM. ............. 90
26a Sample map showing the distribution of CTDM predictions
at tracer sample locations for owe of the Hogback Ridge
experiments. •••••••.........„..... 94
26b Sample map showing observed tracer concentrations
corresponding to the CTDM predictions in Figure 26a.
95
-------
FIGURES (Continued)
Number Page
27 Illustration of receptor radials, source locations,
and wind directions used in the sensitivity analysis
of hill 3-2 130
28 Scatter plots of sigma-w, sigma-v, and Brunt-Vaisala
frequency versus wind speed 132
29 Terrain-effect facjbor versus Hc for wind and source
aligned perpendicular to the. center of the hill 133
30 Terrain-effect factor versus He for symmetric
3-D hill and a long ridge 136
31 Terrain-effect factor versus He for hills of aspect
ratios 3-2 and 5-2 138
32 Digitized contours of Sand Mountain, located southeast
of the Widow's Creek Steam Station ..... 142
33 Elliptical fits to contours with hill "center"
positioned relatively close to the edge of the
plateau 143
34 Elliptical fits to contours with hill "center" located
far front the edge of the plateau 144
-------
TABLES
Number P^S
1 Comparison of the Lateral Position of Streamlines
Modeled- by Snyder et al. using FFT's and the CTDM
Algorithm for a Froude Number of 2 .......... 63
2 Features of Complex Terrain Models Used in the
Evaluation: CTDM. ....... ...... ..... 78
3 Features of Complex Terrain Models Used in the
Evaluation; HTDM. .................. 80
4 Features of Complex Terrain Models Used in the
Evaluation: COMPLEX I ..... ........... 82
5 Data Subsets and Evaluation Tests for Tracer and
Conventional Data Bases. . . ............ . 92
6a Summary Statistics for Data Unpaired in Time
and Space. ...................... 97
6b Summary Statistics for Data Unpaired in Time
and Space. ...................... 98
7a Summary Statistics for Data Unpaired in Time
and Space. . . . . « ................. 99
7b Summary Statistics for Data Unpaired in Time
and Space. .......... ..... ....... 100
8a Summary Statistics for Data Unpaired in Time
and Space. ...................... ioi
8b Summary Statistics for Data Unpaired in Time
and Space. ...................... 102
9a Summary Statistics for Data Paired in Time
Hot in Space . . ............... .... 103
9b Summary Statistics for Data Paired in Time,
Hot in Space .....................
lOa Summary Statistics for Data Paired in Time,
Hot in Space
lOb Summary Statistics for Data Paired in Time,
Hot in Space ..... ..... .......... . 107
-------
TABLES (Continued)
Number Page
lla Summary Statistics for Data Paired in Time,
Not in Space 108
lib Summary Statistics for Data Paired in Time,
Not in Space . 109
12a Summary Statistics for Data Paired in Time,
Not in Space 110
12b Summary Statistics for Data Paired in Time,
Not in Space Ill
13a Summary Statistics for Data Paired in Time,
Not in Space 112
13b Summary Statistics for Data Paired in Time,
Not in Space 113
14 Summary of Model Evaluation Results Data Subset:
Highest Value Unpaired in Time and Space 116
*
15 Summary of Model Evaluation Results Data Subset:
Average of Top N Values Unpaired in Time and Space . . 117
16 Summary'of Model Evaluation Results Data Subset:
Average Peak Hourly Value, Paired in Time, Unpaired
in Space 118
17 Summary of Model Evaluation Results Data Subset:
Average Peak Hourly Value, Paired in Time, Unpaired
in Space 119
18 Summary of Model Evaluation Results Data Subset:
Average Peak Hourly Value, Paired in Time, Unpaired
in Space 120
19 Summary of CTDM Case Studies: Prediction Bias as a
Function of Plume Location Relative to Hc 124
20 Peak 1-Hour S02 Concentrations Predicted by CTDM for
Sand Mountain Monitors Using Two Different Hill
Configurations 145
21 Summary Of M Values from the Complex Terrain
Evaluation Data Bases 152
22 Summary of M Values from Tracer Experiments for CTDM
and Other Refined Air Quality Models 154
-------
LIST OF SYMBOLS AND ABBREVIATIONS
SYMBOL
a
C
d
d6/dz
f
S
G
Major axis Length of an ellipse, m (also denoted as
La)
Term used in Fz formulation
Minor axis length of an ellipse, m (also denoted as
Lb)
Factors involved in B^, 83 terms in WRAP
calculation
Function of h/L used in computation of wind direction
adjustment with height
Parameter used in analytical expression foe
convolution integral I
Factors involved in vertical term in WRAP
concentration calculation
CTDM modeled concentration, yg/m^
Stack top inner diameter, m; also crosswind distance
from plume centerline, m
Source strength of a point source element at the
impingement point on the hill, g/sec
Vertical potential temperature gradient, 8K/m
Coriolis parameter = 1.458xlO~4«sin(latitude).
Buoyancy flux, mVsec3
Vertical distribution function in Gaussian plume
equation
Acceleration due to gravity =9.8 m/sec2
Soil heat flux, watts/m2; also, a constant used in
determining vector wind speed from scalar wind speed
and OQ; also denotes Green's function
Xll
-------
h Mixed layer height, m; also final plume height, m;
also hill height function, m
hp Release height above reference base height, m
hjj Height of receptor above reference base height, m
h^ Modified height of receptor (relative to plume height)
due to terrain effects, ra
hg Stack height, m
H Height at the top of the hill, m; also the sensible
heat flux, watts/m^
Hc Critical dividing streamline height, m
iy Horizontal component of the turbulence intensity
iz Vertical component of the turbulence intensity
I Convolution integral of Green's function with the hill
height function
k von Karman constant
Ky Eddy diffusivity, horizontal component (m^/sec)
K£ Eddy diffusivity, vertical component (m^/sec)
I Mixing length, m; also crosswind distance from plume
centerline to receptor, m
(Ljj Mixing length in neutral conditions, m
13 Mixing length in stable conditions, m
L Monin-Obukhov length, m
La Major axis length of an ellipse, m (also denoted as
"a")
Lb Minor axis length of an ellipse, m (also denoted as
"b")
Ljp Lz Parameters used in analytical expression for
convolution integral I
LX Length scale of hill in along-wind direction, m
Ly Length scale of hill in cross-wind direction, tn
2 21/2
m Modified inverse length scale = n(l-HL /L )
x y
Xlll
-------
n ratio of N/u, m-1
N Brunt-Vaisala frequency, sec"1
Nm Average value of N in a Layer, sec"1
N0 Value of N at so (flat terrain value)
Q Emission rate, g/see
r Ratio of ellipse major axis length to minor axis
length, r^a/b
R Displacement from a reference point
(R2= x2+y2+z2),m
R^ Richardson number
RL Parameter used in analytical expression for
convolution integral I
s Square of the Brunt-Vaisala frequency, sec"2
so Downwind distance at which the plume impinges upon or
is deflected by terrain, m
ST Speed of the incident flow at the tower location, m/sec
S«B Speed of the incident flow approaching a hill, m/sec
t Travel time, sec
Ta Ambient air temperature, °K
Tn Factor for streamline distortion in the vertical
(height correction factor)
Tj, Distortion factor due to terrain effects in the
horizontal direction
TL Lagrangian time scale, sec.
TLO Value of TL at s0 (flat terrain value), sec.
Ts Stack gas temperature, °K
Tu Factor for wind speed distortion due to terrain
influences
Ty Terrain>effect factor equal to ratio of
Oy*/0y*'
Tz Terrain^effect factor equal to ratio of
az*/az*'
XIV
-------
Tov Terrain effect factor equal to ratio of ov/ovo
TOW Terrain effect factor equal to ratio of ow/owo
u Scalar wind speed, along-flow component (in/sec)
u* Friction velocity, m/sec.
Ug Geostrophic wind speed, m/sec
Um Average wind speed in a layer, m/sec
us Stack-top wind speed, m/sec
uv Vector wind speed, m/sec
u',v*,w* Perturbation wind velocities (Boussinesq flow) in
downwind, crosswind, and vertical directions,
respectively (m/sec)
v Cross-flow wind speed component, m/sec
w Vertical component of the wind speed, m/sec
w* Convective velocity scale, m/sec
W(j Mean downdraft velocity in unstable conditions, m/sec
wg Stack gas exit velocity, m/sec
x Distance in downwind direction, m
xo Downwind distance at so, the impingement point of
the approach flow on the hill, m
XQ Length parameter used in analytical expression for
convolution integral I
xr»yr>zr Position in space of the plume release
y Distance in crosswind direction, m
yg Crosswind distance of receptor from plume centerline, m
yg Modified crosswind distance of receptor from plume
centerline due to terrain effects, m
z Measurement height, m; also receptor height, m
z0 Surface roughness length, m
xv
-------
z^ Height of a streamline above the base of the hill
upwind, m
2' Height above terrain surface, m
a Wind speed shear (vertical component), sec"1
alta2 Parameters used in analytical expression for
conduction integral I
o^ Direction of the incident flow approaching a. hill,
degrees
& Strain function from theory of Hunt and Mulhearn
(1973) used in computation of Ty, Tz, also,
rotation angle between wind flow and stagnation
streamline
&z Vertical deflection experienced by a streamline due to
terrain influences, m
&a Wind direction change within the mixed layer, degrees
Ah,AH Plume rise, m
e Variable used to relate scalar and vector wind speeds,,
a function of OQ
A© Jump in potential temperature at the top of the mixed
layer, °K
n Height of a streamline above the surface of the hill
Y Constant of proportionality between the mixing length
in a stable atmosphere and «3w/ia°, a=O.S2
T Constant of proportionality between the mixing length
in a neutral atmosphere and z, f = 0.36
p,v Elliptical coordinates
l»o Value of w along the boundary of an ellipse
ws«us Elliptical coordinates of the point of the plume
release
VT^VT Elliptical coordinates of the point where wind speed
and direction are measured (tower)
$T Wind direction measured at the tower location
¥ Orientation angle of Gaussian hill
xvi
-------
¥m Stability correction for wind profile formulation in
the surface layer
?s Stream function through the source
Po Initial unperturbed fluid density in Boussinesq flow
a Generic representation of either oy or az due
to ambient turbulence, m
Oy, oz due to buoyancy-enhanced dispersion, m
Oy, oz due to source-induced turbulence, m
OQ Standard deviation of wind direction, degrees
ov Standard deviation of the crosswind component of the
wind speed, m/sec.
ovo Value of ov at s0 (flat terrain value)
ow Standard deviation of vertical wind speed, m/sec.
owo Value of 0^ at so (flat terrain value)
Oy Crosswind distance from plume centerline in a Gaussian
plume at which the concentration falls to e~1/2 of
the value at the plume centerline, m
OyO Value of Oy at s0, m
Oy* Growth in Oy between so and the receptor
position of interest for flat terrain, m
Oy*' Growth in Oy between so and the receptor
position of interest with terrain influences
considered, m
Oya Oy due to ambient turbulence, accounting for
terrain-influenced changes in turbulence parameters, m
Oye Effective Oy, accounting for effect of strain in
the flow over terrain on horizontal diffusion
oz Vertical distance from plume centerline in a Guassian
plume at which the concentration falls to e~^/2 Of
the value at the plume centerline, m
ozo Value of oz at so, m
oz* Growth in oz between s0 and the receptor
position of interest for flat terrain
oz*' Growth in oz between so and the receptor
position of interest with terrain influences
considered, m
xvn
-------
ABBREVIATIONS
AGA
ARLFRD
CCB
CTDM
ERF
EPA
EPRI
FMF
FSPS
GLC
HER
LHS
MSB
NMSE
HHS
RMSE
BTDM
SHIS
TPP
VPTG
oz due to ambient turbulence, accounting for
terrain-influenced changes in turbulence parameters,
Effective e»z, accounting for effect of strain in
the flow over terrain on vertical diffusion
Surface potential temperature (at z=o). °K
Mean potential temperature of the mixed layer, eK
American Gas Association
Air Resources Laboratory Field Research Division
Cinder Cone Butte
Complex Terrain Dispersion Model
Error Function
Environmental Protection Agency
Electric Power Research Institute
Fluid Modeling Facility
Full Scale Plume Study, conducted at the Tracy Power
Plant near Reno, NV
Ground-Level Concentration
Hogback Ridge
Left Hand Side of Equation
Mean Square Error
Normalized Mean Square Error
Right Hand Side of Equation
Root Mean Square Error
Rough Terrain Diffusion Model
Small Hill Impact Study (#1 was at Cincer Cone Butte,
#2 was at the Hogback Ridge.
Tracy Power Plant
Vertical Potential Temperature Gradient
XVlll
-------
ACKNOWLEDGEMENTS
This project has been over seven years in duration. The number
of people who contributed significantly to the success of the program
and to whom we are indebted is correspondingly large.
First we wish to thank those at EPA who have provided consistent
and devoted support and encouragement.
• George Holzworth is a clear "father figure" for the project. His
initial clear definition of EPA's needs was essential and helped
to guide us throughout the project. We also appreciated George's
experience at the first field experiments and his staying with
the project even after his initial "retirement."
• Bill Snyder's insight and interest regarding the role that fluid
modeling should play in this effort was very influential and
benefited the program design. The experiments performed under
his direction at the EPA Fluid Modeling Facility provided data on
phenomena difficult to analyze in the field data as well as
verification of basic flow concepts incorporated in the final
dispersion models. His counsel on the photographic program and
his goat-like ability to find ways to get himself to remote but
significant locations for observations were major contributions
to the field experiments.
• Frank Schiermeier enthusiastically supported our efforts and kept
the project going and on target through the ups and downs of
fiscal policy. We appreciate his confidence in us and his
negotiative skills in assuring cooperation among all the
different organizations who contributed to the effort.
• Peter Finkelstein, our EPA project manager during the final
phases of our effort, has been another source of encouragement
and also a great resource to us. His willingness to test and
exercise early versions of the model provided us with otherwise
unobtainable feedback. We appreciate his more general
encouragement regarding the needs for good science and his
patience with our (sometimes meandering) resolution of problems.
• Steve Perry, having worked through our computerized products has
provided a number of suggestions for improvements. His diplomacy
in identifying 'glitches' along the way didn't go unnoticed.
Our five progress reports acknowledge many of those external to
ERT and EPA who contributed to our field programs. The groups form a
substantial list.
xix
-------
• North American Weather Consultants was responsible for the field
operations at Cinder Cone Butte. Tim Spangler and George Taylor
showed ingenuity and patience in developing and executing the
logistics of this complicated effort. Their efforts were
essential to the success of these experiments.
• Wynn Eberhard and his colleagues from the National Oceanographic
and Atmospheric Administration's Wave Propagation Laboratory
(NOAA WPL) generated the ground- based lidar data for all the
field experiments. Bill Neff provided the Doppler acoustic
sounders and, later, sonic anemometers and other instruments to
supplement our more conventional meteorological measurements.
The unique information gained by these operations has contributed
significantly to our understanding of the phenomena observed. We
appreciate the dedication these groups showed in their efforts.
• Ray Dickson, Gene Start, and their associates at the NOAA ARLFRD
were responsible for the tracer release, sampling, and analysis
and many of the meteorological measurements at the Hogback Ridge
and Tracy Power Plant experiments. These operations, which
involved a large number of people and many items of equipment,
required significant logistics, and we credit their leadership
and experience for contributions to this program.
• Norm Ricks, first from ERT, then from Morrison/Knudsen, should
receive the "most-devoted contributor" award through his
dedication to obtaining good photographic and video tape data.
Those cold hours atop our 150-m tower at HBR exemplified Norm's
extraordinary efforts. His wizardry in keeping equipment going
was also outstanding.
• The field crews from NOAA's ATOL who contributed and operated
their tethersonde throughout the long nights deserve our
appreciation.
• Julian Hunt and Rex Britter of Cambridge University consulted
with us at various stages. They put fluid modeling results in
perspective and explored theoretical aspects of the effort. We
benefited from our several discussions with them.
• The Electric Power Research Institute, under the inspiring
guidance of Glenn Hilst, provided direct support of the field
experiments at Tracy. They co-sponsored the entire first effort
and contributed the airborne lidar capability to the final
experiment. EPRI's participation, via their subcontractors
Rockwell International, SRI International, TRC Environmental
Consultants, and the Research Triangle Institute, added greatly
to the value of this program.
Finally, we want to acknowledge the efforts of the many ERT staff,
past and present, who contributed beyond the call of duty in this*
program.
xx
-------
• Tom Lavery, project manager through the field experiment phases,
provided leadership and follow-through to our staff and our
subcontractors. His willingness to let others "do their thing"
helped provide the research environment needed.
• Ben Greene, our quality assurance conscience, was also our most
articulate phenomena reporter in the field experiments. His
devotion to understanding what the data collected really meant
was crucial to our subsequent interpretations of the field
studies. His persistence in addressing problems needing
resolution was needed and is appreciated.
• Akula Venkatram contributed many of the early theoretical
concepts which have carried through to the end product of the
CTDM. His intuition and suggestions for exploration provided the
team with avenues to pursue when the path to progress became
mired.
• Dan Godden manned our field headquarters and, with a sense of
humor in the middle of the night, kept our spirits up and eyes
open.
• Don DiCristofaro did much of the coding and testing of CTDM
during the development stage of the program, and was instrumental
in preparing The Modelers' Data Archives, which should be of use
to other for years.
• Liz Insley prepared many of the meteorological and terrain data
sets used in the final evaluation of CTDM, providing important
feedback on decisions faced by users of the model.
• Michael Mills joined the team in-time to develop the terrain
preprocessor programs necessary to run the model. The computer
graphics developed by Mike that are associated with the terrain
preprocessor, receptor generator, and CTDM postprocessor will go
a long way toward increasing user understanding and acceptance of
the CTDM package. His quick understanding of all that had been
done previously was a tremendous boost to the project.
• Michael Dennis became responsible for documenting and testing
much of the CTDM and its pre- and post-processors. His ability
to prevent and/or chase down problems is especially appreciated.
• Art Bass contributed significantly to our original proposal
effort and to the model development work plan. We appreciate the
long-lasting forethought he provided in these efforts.
• Steve Andersen, Chris Johnson, and others from ERT's Fort Collins
office were responsible for the 150-meter tower, other tower and
ground-based meteorological instrumentation, site preparation and
restoration, and the care and feeding of 40-year old arc lamps.
Their long hours, ingenuity, experience with instruments and data
acquisition systems, and lack of vertigo were vital to the
success of the field data collection.
There are others, too numerous to specifically call out, who
helped us accomplish this program. We thank them all!
xxi
-------
SECTION 1
INTRODUCTION
The Complex Terrain Model Development (CTMD) project, 'sponsored
by the U.S. Environmental Protection Agency (EPA), was initiated, in
1980. Its purpose is to develop, evaluate, and refine practical plume
models for calculating ground-level air pollutant concentrations that
result from elevated emission sources located in complex terrain
(i.e., terrain which rises to heights well above expected plume
levels). The primary objective of the project is to develop models to
simulate 1-hour average concentrations during stable atmospheric
conditions.
These models are to be used in a wide variety of applications,
such as the siting of new energy development facilities and other
sources of air pollution, regulatory decision making, and
environmental planning. Therefore, the models should be easy to
understand and use, with known accuracy and limitations.
The objectives of the program were described by Holzworth (1980)
and generally follow the recommendations of the participants of the
EPA-sponsored workshop to consider the issues and problems of-
simulating air pollutant dispersion in complex terrain (Hovind et al.
1979). The program was subsequently designed to include model
development efforts based on physical modeling, field experiments, and
theoretical work.
Four major CTMD field experiments have been completed during the
last seven years to collect data for development and evaluation of
various modeling approaches. The first field experiment, the. Small
Hill Impaction Study No. 1 (SHIS #1), was conducted during the fall of
1980 at Cinder Cone Butte (CCB). Idaho. CCB is a roughly
axisymmetric, isolated and approximately 100-m tall hill located in
the broad Snake River Basin near Boise, Idaho. The second field
experiment, SHIS #2, was performed during October 1982 at the Hogback
Ridge (HER) near Farmington, New Mexico. HER is a long, 90-m tall
ridge located on the Colorado Plateau near the western slopes of the
San Juan Mountains. Both small hill studies consisted of flow
visualization and tracer experiments conducted during stable flow
conditions with supporting meteorological, lidar, and photographic
measurements. At these sites, the tracer gases were released from
mobile cranes or a tower.
The third and fourth field experiments were conducted at the
Tracy Power Plant (TPP) located in the Truckee River Valley east of
the Reno,. Nevada. The third experiment, performed in November 1983,
was undertaken as a feasibility and design study for the Full Scale
-------
Plume Study (FSPS). It was co-sponsored by the Electric Power
Research Institute (EPRI). The November experiment not only
demonstrated the feasibility of conducting the FSPS at the TPP, but
with the expanded scope made possible by EPRI's participation, it also
produced a data base that itself is useful for modeling purposes. The
FSPS was subsequently performed at Tracy in August 1984.
The data bases compiled from the CCB, HER, and the FSPS
experiments are available from the EPA Project Officer. The data
bases compiled from each of the experiments include the following
components:
• Source information: emission rates, locations, and heights
of SF6, CF3Br, and oil-fog releases.
• Meteorological data: measurements of the approach flow as
well as information on flow and dispersion near the
terrain.
• Tracer gas concentrations: data from more than 50,000
individual samples collected during the experiments from as
many as 100 sampler locations in each experiment.
• Lidar data (archived at the Wave Propagation Laboratory):
sections across the plume characterizing the trajectory and
growth of the plume upwind of, interacting with, and
sometimes in the lee of, key terrain features.
• Photographic data: still photographs taken from fixed
locations, aerial photographs taken at CCB and Tracy from an
aircraft flying overhead, and (occasional) 16-mm and 8-mm
movies and videotapes.
During the course of the CTMD project, five Milestone Reports
(Lavery et al. 1982, Strimaitis et al, 1983, Lavery et al. 1983,
Strimaitis et al. 1985, and DiCristofaro, et al. 1986) have been
published. These reports, which are available from EPA, describe the
progress in developing and evaluating complex terrain models using the
CCB, HBR, and TPP data bases. They also describe in detail the two
small hill studies, the November 1983 Tracy study, the FSPS, and a
series of towing tank and wind tunnel studies performed at the EPA
Fluid Modeling Facility (FMF) in support of the modeling.
A preliminary partially validated model was delivered to the
Project Officer in October 1985. A workshop was held in February 1986
to present to the scientific community results from the field
experiments, model development activities, and related work done both
within and outside the CTMD project and to obtain feedback from users
working with the preliminary model.
This final report provides a brief overview of the program and
discusses the major accomplishments with respect to the model
development and validation efforts since the Fifth Milestone Report.
A detailed description of the final version of CTDM is provided in
-------
Section 3. CTDM now contains a flow model which computes the
deflection experienced by a streamline passing through a given point
over the hill. The model uses a "backwards-looking" solution of
linearized equations of motion of a steady-state Boussinesq flow. The
deflection calculation provides a more realistic and theoretically
satisfying method for estimating the distance between plume
center-lines and terrain surfaces and for estimating the distortion of
plume shape associated with streamline spacing changes. Appendix A
provides additional details of the algorithm.
The evaluation of CTDM with field data is described in
Section 4. The development of CTDM was guided by the findings and
observations from the field experiments conducted as part of this
project at CCB, HER and TPP (FSPS). These programs were designed
specifically to gather the meteorological information, observations
and ground level concentration data needed for the model development
effort. CTOM and several other models of regulatory interest were
tested with the same data sets. In addition, CTDM was tested against
two other data sets from studies at the Westvaco paper mill at Luke,
Maryland and at the Tennessee Valley Authority's Widow's Creek Steam
Plant. The data sets from these monitoring programs provide an
independent means of evaluating the models. Although the
meteorological and ground-level concentration data were not as
extensive as that collected at the field sites, full years of data are
available. The results of these tests are also presented in Section 4,
CTDM was also subjected to a series of sensitivity tests which
are reported on in Section 5. These tests largely emanated from
recommendations received at the CTDM workshop held in February 1986.
The analyses focus on the effect that variations in measured or
predicted quantities would have on the terrain effect in CTDM. The
terrain effect is defined as the ratio of the predicted peak
ground-level concentration to the concentrations at the plume
centerline for the same travel time. This measure of sensitivity is
appropriate for the intended use of the model as it focuses attention
to the magnitude of predicted concentrations rather than to changes in
the location of the peak values.
Limitations to the use and applicability of CTDM are addressed in
Section 6 of this report. The specific guidance emerges in part from
the results of the evaluation and sensitivity test efforts and also
from recognizing that testing of the model has not subjected it to all
of the important terrain obstacle shapes and configurations which may
be encountered. The limitations must be placed in the context of
implicit or explicit limitations applicable to EPA's prior complex
terrain modeling techniques.
Finally, a series of conclusions and recommendations which emerge
from this model development effort are presented in Section 7.
Appendices to the report include a paper on the flow model,
statistical results from the model evaluation efforts and a document
describing the contributions made to the CTMD program by EPA's Fluid
Modeling Facility.
-------
SECTION 2
OVERVIEW OF THE GTMD PROGRAM
2.1 Background and Overall Program Plan
The CTMD program was initiated by EPA in response to
long-standing controversies in the technical and regulatory
communities over the lack of reliable methods for predicting air
quality concentrations in regions of mountainous or complex terrain.
Of particular concern was the absence of a verified dispersion model
for predicting ambient air concentrations during stable atmospheric
conditions, the conditions expected to give rise to the highest
short-term concentrations. EPA had employed "screening" models (most
notably, the Valley model (Burt, 1977)) which were known to be
strongly biased toward overprediction in high terrain areas, but EPA
guidance did not recommend a "refined" model for use in such
settings. It was recognized that a major data collection and model
development effort would be required to develop such a capability.
EPA convened a workshop (Hovind et al., 1979) of specialists in
field measurement programs, fluid modeling and mathematical modeling
for purposes of helping to define the specific elements of such a
program. The workshop report recommended that a two-phased field
program be initiated with a first phase being performed at an isolated
and relatively small terrain feature of simple geometric shape. This
hill would be heavily instrumented and studied at relatively small
cost and the tracer release and sampling program would be capable of
being altered in response to different meteorological conditions. A
second phase field program would involve a "full-scale" site and
increased topographic complexity. An underlying concept was that the
dynamics of an elevated plume's interaction with a terrain feature
could be studied at smaller hills, whereas the effects of larger scale
meteorological flows on transporting plumes toward high terrain would
be better studied at large terrain shapes. In accordance with the
above concept, the workshop report also suggested the complementary
use of physical or fluid-modeling facilities to investigate the flow
dynamics involved in a systematic way with scaled-down models of
terrain features.
Mathematical model development was an essential, organizing
component of the proposed program because the ultimate product was to
be a model for regulatory use. The workshop recommended reliance upon
Gaussian-based or "K-Theory" models because of their conceptual
simplicity and ease of use. Holzworth (1980) stated the overall need
succintly as "the production of a useful model (or models) with
demonstrated reliability and prescribed applicability."
-------
The program was to focus on understanding the impact during
stable atmospheric conditions and for one-hour averaging periods. EPA
developed a request for proposals for the effort and awarded the
contract for the Complex Terrain Model Development program to ERT in
June, 1980. A parallel and complementary fluid modeling effort was
also expanded upon at the EPA Fluid Modeling Facility (FMF). Scale
model experiments were to be performed on a variety of terrain shapes
including that of the first field experiment site at Cinder Cone Butte
in Idaho.
The components of the CTMD program and the progress made are
well-detailed "Snd documented in a number of reports. In particular,
five Milestone Reports* were written during the course of the program
which describe in detail the model development efforts, the field
experiments, the data gathering and interpretation efforts, and the
model evaluation efforts performed during the course of this program.
These reports also contain, as appendices, relevant contributions
from EPA's Fluid Modeling Facility efforts. In addition, separate
documents (Greene and Heisler, 1982; Greene, 1985; and Greene, 1986),
describe the quality assurance aspects of various components of the
program. Another document (Lavery et al., 1986) summarizes the
results of a workshop held with users of an early version of the
Complex Terrain Dispersion Model (CTDM). In addition, a number of
other reports relating to this effort have been produced by members of
EPA's FMF and will be referenced as appropriate.
The remaining portions of this section describe the program
components and also identify specific documents containing more
detailed information about these components.
2.2 Field Program
2.2.1 Goals and Design
A program to gather high quality and relevant data from field
experiments was an essential element of the overall study. The
measurements desired were driven by the needs of the model development
effort. In accordance with EPA's conceptual plan, the first field
experiment was.held in the fall of 1980 at Cinder Cone Butte (CCB)
near Boise, Idaho. This volcanic hill stands about 100 m high and has
a nearly circular base about 1 kilometer in diameter. Set in a broad
section of the Snake River valley, CCB is an isolated feature
surrounded by relatively flat terrain for tens of kilometers.
Meteorological data showed that stable drainage flows occurred on a
relatively routine basis at night, but there was moderate variability
in wind direction, probably associated with the broadness of the
valley. The field program design had at its core several concepts:
*In the following discussion, these milestone reports will be referred
to as the First, Second, etc. Milestone Reports; they are formally
referenced in Section 1.
-------
(1) two different tracer gases would be released from a mobile crane
at a variety of heights relative to terrain and the critical dividing
streamline heights. The crane would be capable of being moved to
different positions upwind of the butte in accordance with predicted
wind directions, (2) tracer gas concentrations would be obtained at
approximately 90 locations on or around the butte (with sampling time
periods of one hour or less), (3) dense smoke releases would allow
photographic and lidar documentation of plume behavior, and
(4) extensive measurement of meteorological variables of importance to
the model development effort would be made with both fixed and remote
sensing instrumentation.
Because of the emphasis on understanding the flow dynamics under
stable atmospheric conditions, the field experiments were generally
begun in the evening and continued until after daybreak. Weather
forecasting and on-site monitoring of meteorological conditions were
relied upon to set the initial locations of the cranes releasing the
smoke and tracer gases. Participants at the experiments at CCB were:
ERT (overall project management); Western Scientific Services, Inc.
(fixed meteorological data collection); North American Weather
Consultants (smoke and tracer releases, tracer data collection,
photography and mobile meteorology); NOAA Wave Propagation Laboratory
(lidar systems); and TEC, Inc. (independent data audits).
The First Milestone Report provides a detailed description of the
first field experiment configurations at CCB including a discussion of
the quality assurance program.
The second field experiment took place at Hogback Ridge (HBR)
near Farmington, New Mexico in October 1982. This ridge was chosen
because it represents a more-or-less two-dimensional terrain feature,
having a height of about 85 meters, but extending several kilometers
to either side of the experiment site. An important modeling issue
which needed resolution was the importance of aspect ratio (of length
to height) to plume trajectory behavior. Also needed was an
understanding of how plumes would be transported under stable,
"blocking" conditions upwind of nearly two-dimensional terrain
features. The experimental setup at HBR was similar to that
established at CCB, but included an additional capability for
releasing tracer gases from different levels of a 150-ra tower.
Approximately 125 tracer gas samplers were set out for each
experiment, concentrating the coverage in the upwind and downwind
sides of the ridge nearest the tracer releases, but also providing
samplers at considerable distances along the ridge axis.
The NOAA ARLFRD (Air Resources Laboratory Field Research
Division) was responsible for the tracer gas releases, sampling and
analyses, the smoke visualization and the telemetry and meteorological
data archive and display systems for this experiment.
Morrison-Knudsen was responsible for the photographic program. The
NOAA WPL performed lidar measurements and contributed additional
meteorological instrumentation to the program. The HBR field
experiment design is described in the Third Milestone Report.
-------
The final field experiments were conducted at the Tracy Power
Plant (TPP) near Reno, Nevada. A feasibility study was conducted at
the site in November 1983 and the Full Scale Plume Study (FSPS) took
place in August 1984. The TPP is located in the Truckee River Valley
and has a 91.4-meter stack. Nocturnal drainage winds were shown to
routinely transport plumes toward the east. Mountain peaks rise to
several times stack and expected plume heights in the down-valley wind
direction. The site was chosen to test the concepts and preliminary
models developed on the basis of findings from the two prior field
experiment sites. The feasibility study conducted in 1983 was
co-sponsored by EPA and the Electric Power Research Institute. ERT
erected and instrumented the 150-ra tower on site. Almost all of the
planned tests for the FSPS were performed. The data collected showed
that the TPP plume would interact with high terrain, especially near a
bend in the valley, down-valley of the plant. The tracer gas release
and measurement systems were shown to work well. An airborne lidar
was added to the program and tested. Intermountain Film Productions
conducted the photographic program under the direction of ERT. EMSI
and TRC, Inc. provided the tracer gas sampling and on-site management,
respectively, for the EPRI-sponsored efforts.
The FSPS conducted in August 1984 continued with the use of two
tracer gases and extensive meteorological measurements. One-hour
samples of ground-level tracer gas concentrations were collected at
over 100 locations. Principal participants were ERT, NOAA/WPL, and
HOAA ARLFRD. Meteorological Standards Institute provided external
audits of instrumentation and Morrison-Knudsen collected the
photographic data. SRI International, under sponsorship of the EPRI,
made airborne lidar measurements. The design and results of the
preliminary field study at TPP are described in ERT's Fourth Milestone
Report. A full description of the final FSPS is provided in ERT's
Fifth Milestone Report. A separate document (Eberhard, 1986)
describes the NOAA Wave Propagation Laboratory's contributions to all
of the CTMD field programs.
2.2.2 Field Program Results
The Modelers' Data Archives
Data collected at each of the field experiments were archived for
use in the model development and evaluation phases of this program and
for future use. Recognizing the difficulty others have had trying to
utilize raw data from large field experiments, ERT also developed
Modelers' Data Archives (MDA's) for each experiment. The MDA's are
subsets of the complete data sets which are thought to be of most use
to those involved in dispersion model development. The MDA's also
contain the data in an organized format providing greatly increased
ease-of-use. The MDA's for the CCB and HER experiments are described
in the Third and Fourth Milestone Reports, respectively. The MDA for
the TPP experiment is described in a separate document (DiCristofaro,
1986). The MDA's have been made available to interested parties
through EPA's Terrain Effects Branch.
The raw data have been collected in a series of computer files
which are available on magnetic tape from the Terrain Effects Branch
of EPA. Three reports documenting these files (Truppi and Holzworth,
7
-------
1983; Truppi, 1985; and Truppi, 1986) have been prepared and are
available. In addition, these data files have been transferred to a
series of SAS™ data sets. They and a report describing them
(Truppi, 1987) are also available from the Terrain Effects Branch.
Field Program Findings
The field programs provided basic data on ground-level
concentrations as & function of source-hill geometries and
meteorological conditions. In addition, the flow visualization
through smoke releases and the photography program together with the
lidar measurements and observer notes provided other information on
the air flow dynamics encountered. Detailed descriptions of the
findings are a main component of the Milestone Reports. The First and
Second Milestone Reports discuss the findings of the CCB experiment.
The Third and Fourth Milestone Reports discuss the results of the HBR
experiment. The Fourth and Fifth Milestone Reports provide results
from the FSPS.
A progression of understanding emerged from these experiments.
Observations from CCB demonstrated the validity of the dividing
streamline concept or critical height, Hc, in simulating the flow
fields during stable atmospheric conditions. The critical height can
be defined in terms of the meteorological measurements of the wind
speed and temperature as a function of height. Tracers and smoke
released directly upwind of the hill and above Hc were generally
observed to flow up and over the hill, in accordance with theory.
Plumes released below this height were generally observed to pass
around to the side of the hill, also in accordance with theory.
Maximum ground-level concentrations at CCB were most commonly
associated with tracer gas releases near the calculated
dividing-streamline height.
At HBR, the dividing streamline concept also was shown to be
applicable, although the flow behavior below this height was different
from that observed at CCB. In particular, at HBR, the lower portion
of the flow was "blocked" behaving as a relatively stagnant flow with
correspondingly low wind speeds. This occurs also in accordance with
theoretical considerations a§ the two-dimensionality of th© ridge
shape offers only a very long path for the air to flow around to the
side. The largest concentrations observed at HBR occurred for
releases below Hc and the magnitude of the concentrations
(normalized by release rate) were much larger than those observed at
CCB,. or subsequently at the FSPS. Smoke and tracers released above
HC were observed to flow over the ridge and resulted in peak
concentrations near the top or on the lee side.
Observations from the FSPS at the Tracy site showed both kinds of
behavior as well. There were portions of the flow which, when they
encountered high terrain away from the river valley walls, became
relatively stagnant as observed at HBR. Plumes embedded in
down-valley flows and encountering terrain obstacles protruding from
the valley side walls, on the other hand, exhibited an ability to lift
-------
up and over the terrain or readily pass around the sides as seen at
CCB. In each case, Hc for each hill proved to be a reliable
parameter for differentiating between these two regimes. Peak
concentrations at the FSPS were most often associated with releases
near the calculated dividing streamline heights.
2.3 Fluid Modeling Program
An integral part of the CTMD program from the beginning was the
efforts undertaken at EPA's Fluid Modeling Facility (FMF) at Research
Triangle Park, NC. Theoretical aspects of the phenomena associated
with interactions of stably-stratified atmospheric flow with terrain
obstacles suggested that scaled-down, fluid modeling experiments could
be used to investigate many of the fluid mechanical issues. The
implications of the dividing streamline concept, in particular, were
especially well-suited to the types of systematic investigations which
could be undertaken with wind tunnels and with stratified towing
tanks. Although the field experiments provide "real world," "ground
truth" data on the relationship of ground-level concentrations to
emission rates, the results are specific to the field study
configuration and to the meteorological conditions encountered.
Through fluid modeling, one can investigate in a systematic and
reproducible manner, the effects of many geometric configuration or
flow parameter changes which would be virtually impossible and
prohibitively expensive through field experiments. A key
justification for reliance upon the fluid modeling results is a
demonstration that there is a correspondence between field
measurements and a physical model simulation of the field measurements,
Experiments at the FMF included simulations with models of CCB
and HBR and for conditions corresponding to some of the interesting
field experiments. Verification of the dividing streamline concept
was a central focus for experiments that included testing of the
effects of changes in release height and terrain shape. Another
series of experiments addressed the effects on maximum surface
concentrations of sources of different heights being placed upwind and
downwind of simply-shaped terrain obstacles. W. Snyder produced an
independent report summarizing the above and other contributions of
the FMF to the CTMD program. This report is self-standing and is
included in its entirety as Appendix H to this document.
2.4 Model Design for Regulatory Use
The CTMD program has an ultimate goal of providing a dispersion
model for routine use by the air pollution modeling community. CTDM
is to fill the need for a refined model for complex terrain settings
where terrain heights exceed the heights of the sources under review.
This goal provided specific guidance to the form of the model
development effort. First of all, the model had to be based upon
experimental data relevant to the kinds of meteorological and
topographical conditions of historic and expected future concern to
EPA. It was anticipated that many of the future regulatory permitting
requests would emerge from the Western USA. Past efforts had clearly
-------
shown that conditions under nighttime stable meteorological conditions
were the most constraining for compliance with the short-term (1-hour,
3-hour, or 24-hour) National Ambient Air Quality Standards (NAAQS) or
Prevention of Significant Deterioration (PSD) increments.
Demonstrations of compliance with NAAQS and PSD increments require
that the very highest values expected from hour-by-hour sequential
modeling up to five years of meteorological data be used.
Practicality calls for the need for relatively cost-effective computer
code for such demonstrations. In effect, this constrained the program
from pursuing advanced numerical simulation techniques as a modeling
method and supported the concept of advancing Gaussian plume modeling
types of methods to meet the program's needs. There was also a
concern that more advanced techniques might not display improved
performance given the realities constraining meteorological
measurement programs and inherent uncertainties associated with
atmospheric turbulent flows. Another concept influencing the design
of the dispersion model was that the model should be readily
understood and explainable to others who are affected by the results
of the model, but may have little background in theoretical fluid
mechanics. From this perspective, an approach based more on
analytical equations rather than on turbulence simulation methods was
preferred.
More specifics on the model development plan are presented in the
First and Second Milestone Reports. Subsequent Milestone Reports
describe improvements made to the model in the course of its
development. Section 3 describes the technical basis of the final
version of the CTDM in detail. Separate user's manuals have been
provided for those concerned with operating the model or the model
preprocessors.
2.5 Model Evaluation Program
The model development program called for use of findings from the
field experiments and the fluid modeling efforts in assisting with the
theoretical development of equations for the CTDM. In essence, this
took place in several stages. Initially, the confirmation of the
dividing streamline concept from the results from CCB contributed to
the development of separate models for addressing flows above and
below Hc. A series of tests were made with these models and with
models being used at the time by EPA in its regulatory practice.
Results of these early comparisons are presented in the First
Milestone Report. Subsequent model evaluation efforts focused also on
case-study analyses of model predictions with observations for a large
number of the hours at each of the field sites. Results of these
analyses are presented in each of the subsequent Milestone Reports. A
preliminary version of the model was completed in October 1985 and was
provided to a number of interested groups for independent evaluation.
Representatives of these groups convened at a workshop in February
1986 to share the results of experiences with the model. The results,
especially those from some further fluid modeling tests, suggested
that additional modifications be made to CTDM which would incorporate
the growth of an internal boundary layer as the approach flow above
Hc was influenced by the terrain.
10
-------
Upon completion of a "final" model, the model evaluation effort
focused on several existing data bases from S02 measurement programs
near large point sources in complex terrain settings. This effort is
described in Section 4 of this report and represents a "hands-off"
evaluation of the CTOM along with several other dispersion models of
regulatory interest. The results show CTDM to be superior to the
other models in its ability to produce concentration estimates that
are in better agreement with observations. In a broad sense, the
model evaluation effort also included tests quantifying the
sensitivity of the model predictions to changes in input parameters or
conditions. Model performance is shown to decline when degraded
meteorological data are used, as might be expected. The results of
the sensitivity tests are presented in Sections 4 and 5 of this
document.
11
-------
SECTION 3
DESCRIPTION OF THE COMPLEX TERRAIN DISPERSION MODEL (CTDM)
This section provides a detailed description of CTDM, including
its mathematical derivation. As an introduction to the model,
sub-section 3.1 explains in qualitative terms how the Gaussian plume
framework is modified to include the effects of terrain, and how the
modifications differ from adjustments made for terrain in models
presently used in regulatory applications. Sub-section 3.2 describes
the meteorological and terrain variables needed by the model, and how
these are obtained from the data provided by the meteorological and
terrain preprocessors. Finally, sub-section 3.3 contains the
derivation of the dispersion model formulations. Operational aspects
of CTDM and its preprocessors are discussed by Mills et al, (1987),
Paine (1987), and Paine et al. (1987).
3,1 Qualitative Overview
CTDM is a point-source Gaussian plume dispersion model designed
to estimate hourly-averaged concentrations of plume material at
receptors near an isolated hill or near a well-defined segment of'an
array of hills. The Gaussian plume model for simulating the
dispersion of pollutants from a continuous point-source describes a
plume by its average properties as a function of distance along the
flow downwind of the point of release. The concentration of material
in the plume is described by a Gaussian distribution in a .plane
perpendicular to the flow. The vertical distribution in this plane
has & length scale denoted as sigma-z, and the lateral distribution
has a scale denoted as sigma-y. Complete reflection of the plume at
the ground assures that no plume material disappears from the
atmosphere. The concentration of plume material at any point downwind
of the source is determined by the size of the plume (sigma-y and
sigma-z) 8 the wind speed, the strength of the source, and the distance
of the sampling point from the axis of the plume. For example, if the
plume is narrow in the vertical (sigma-z is substantially less than
the height of the axis of the plume), the peak concentration is found
at the center of the plume. Sampling points away from the center
would record smaller concentrations. Over level ground, peak
ground-level concentrations are found when sigma-z is of the same
order as the height of the plume.
When a hill is present, the changed flow alters the way in which
plume material can reach the surface. Obviously, the path of the
plume can change as the flow spreads over or around the hill so that
there is a shift in the relative position of a receptor and the center
of the plume. There are also mechanisms for changing the rate at
which the material diffuses toward the surface, as well as allowing
the center of the plume to impinge on the surface of the hill. These
mechanisms are genera 11-y responsible for increasing peak
12
-------
concentrations expected on terrain beyond those concentrations that
would have been expected for the same meteorological conditions on
level terrain.
In the absence of stratification, all streamlines in the flow
pass over a hill. The centerline of a plume in this flow follows the
streamline that passes through the source of that plume. As the plume
grows in the vertical and horizontal directions (in the plane
perpendicular to the flow), plume material diffuses across adjacent
streamlines, eventually reaching the set of streamlines that marks the
surface of the terrain. Distortions in the flow which are induced by
the hill change the position and relative spacing of the streamlines
from their initial distribution over level terrain, and therefore
change the shape of the plume as it passes over the hill.
Ground-level concentrations (GLC's) of plume material also change, and
this happens in two ways. The first and most obvious change in the
GLC's is the shift in the distribution on the surface of the hill,
arising from the change in the shape of the plume. Typically, the
plume stretches in the horizontal as it passes over the crest of a
simple three-dimensional hill and.this stretching produces a wider
footprint over the hill. The second change in the GLC's is the change
in magnitude, arising from the effect of the distortion on the rate of
diffusion of plume material across streamlines. Typically, spacing
between streamlines is reduced in the vertical and expanded in the
horizontal, while the speed of the flow increases over the crest.
These changes tend to increase the diffusion in the vertical and
reduce it somewhat in the horizontal, thereby altering the magnitude
of the GLC's on the hill. If the diffusion were not altered by the
distortion in the flow, the peak GLC would not change from that
obtained' on level .terrain.
The nature of the flow changes dramatically when the flow is very
stably stratified. A two-layer structure develops in which the flow
in the lower layer primarily deflects around the hill, while the flow
in the upper layer travels over the top of the hill. A critical
height He defines the boundary of these two layers in CTOH. This
concept was suggested by theoretical arguments of Drazin (1961) and
Sheppard (1956) and was demonstrated through laboratory experiments by
Riley et al. (1976), Brighton (1978), Hunt and Snyder (1980), Snyder
et al. (1980), and Snyder and Hunt (1984). In the layer above He,
the approach flow has sufficient kinetic energy to transport a fluid
parcel up and over the hill against the density gradient of the
ambient stratification. In the layer below He, the approach flow
has insufficient kinetic energy to push the parcel over the hill, so
that the flow below Hc is restricted to lie in a nearly horizontal
plane, allowing little motion in the vertical. Consequently, plume
material below Hc travels along and around the terrain rather than
over it.
Above Hc, the flow is similar to that just described above
although the degree of distortion depends on the stratification.
Below Hc, the flow is approximated as an ideal, steady,
two-dimensional flow. Within this flow, only one streamline at each
elevation touches and follows the surface of a hill, and is referred
to as the stagnation streamline. Plume material reaches the surface
of the hill only if it reaches the stagnation streamline. If the
13
-------
plume center-line lies along the stagnation streamline and if it also
lies below Hc, the center of the plume impinges on the hill. But if
it lies to one side of the stagnation streamline, the centerline will
pass to one side of the hill.
The position of Hc and the stagnation streamline relative to
the centerline of the plume dominates the degree to which a hill in
stratified flow is able to alter the peak ground-level concentration
obtained in the absence of the hill. Figure 1 illustrates this. In
the model, the Hc surface slices the plume into two pieces as the
hill is encountered. Plume material now residing below Hc is sliced
once again by the stagnation streamline. Concentrations on the
surface of the hill above Hc are determined by the cut made by the
Hc surface because this now coincides with the bottom of the plume,
which is in contact with the surface of the hill. Concentrations on
the surface of the hill below Hc are determined by the cut made by
the plane of the stagnation streamlines because this cut coincides
with the sides of the plume segments that are in contact with the
surface of the hill. As illustrated, receptors on the hill record
concentrations that are much nearer the center of the plume than do
receptors in the absence of the hill. Figures 2 and 3 provide
further insight into how the plume is modeled in CTDM.
Figure 2 addresses plume material above Hc. The upper portion
illustrates what the plume may actually look like in vertical
cross-section as it travels along the surface. Material below Hc is
removed at so, and the remaining material is distorted in the flow
and reflected from the surface of the hill. The size of the plume in
the vertical at s depends on the amount of distortion in the shape of
the plume as well as the amount of additional growth of the plume
caused by changes to the rate of diffusion. The lower portion of
Figure 2 illustrates how the model actually treats the plume.
Reflection of plume material from the surface 2=0 is allowed from the
source to so, and reflection of plume material above He is allowed
from the surface z-Hc beyond so. Furthermore, the distortion in
the flow (and the plume) beyond so is scaled out, leaving only the
effect of the distortion on the diffusivity in what is termed the
effective sigma-z (<*ze). As illustrated, oze exceeds oz,
the plume size in the absence of the hill, because the diffusion in
the vertical across streamlines is increased by the contraction in the
vertical spacing of the streamlines.
Figure 3 addresses plume material below Hc. The diagram on the
left illustrates the plume in horizontal cross- section as it splits
and flows around a hill. In this case, plume material has crossed the
stagnation streamline before the hill is encountered, so plume
material is found on both sides of the hill. Once the plume wraps
around the leading edge of the hill, the stagnation streamline (which
forms the boundary of the hill) becomes a reflecting surface in
addition to the plane z=0, and material cannot diffuse from the
segment of the plume on one aide of the hill to the segment on the
other side of the hill. The diagram on the right side of Figure 3
illustrates how the model treats this flow. The actual surface of the
hill is replaced by a line in the plume which corresponds to the
14
-------
PLUME CROSS-SECTION
LOOKING DOWNWIND
C • constant
PLUME CENTERL1NE
X. PEAK GLC
(HILL)
He
STAGNATION STREAMLINE
I
/ /
PEAK GLC
(FLAT)
S / S /^r / S^ S S S '
SURFACE UPWIND OF HILL
Figure 1. Idealized picture of how the dividing-streamline (Hc)
plane and the plane of stagnation streamlines "cut" into a
plume, allowing material nearer the center of a plume to
contact the surface of a hill.
15
-------
i PHYSICAL PICTURE: material
_! 'I hR (R«c»Ptor> above Hc rides up and over hill
-—^ in a distorted flow. Material
below He passes round the side.
UFT CALCULATION: net effect of flow
distortion is te increase th«
effective rct« of plume growth
absence of the hill
Figure 2. Illustration of terrain effect on the vertical distribution of
plume material above Hc as modeled in LIFT.
4
§
16
-------
Impenetrable Boundary
Stagnation
Streamline
Plume Centerline
Figure 3. Depiction of plume behavior in CTDM as it is deflected
around a hill as seen from above: the stagnation streamline
in the left figure separates flow going around the right and
left sides of the hill. In CTDM, the hill is treated as
being collapsed into an impenetrable wall that separates the
flow and plume material going on either side of the hill,
with an effective crosswind distance and O.
17
-------
stagnation streamline cut in Figure 1. Ail distortion is scaled out
as was done for the flow above Hc, but the effect of the distortion
on lateral diffusion is ignored as a second order effect. The lateral
position of all receptors below Hc essentially collapses onto the
stagnation streamline, as depicted by the points labeled A and B, so
that each is the same distance from the plume center-line. However,
concentrations along one side of the line differ from those on the
other side because diffusion through the Line is not allowed.
Adjusting receptor positions while keeping the trajectory of the
plume a straight line simplifies the mathematics of CTDM a great
deal. Rather than keeping track of the actual boundary of a hill and
the deformed trajectory of each of the segments of the plume as in the
left portion of Figure 3, concentrations are computed at receptor
points A and B for a plume geometry like that of the right portion of
Figure 3, which is only slightly more complicated than that for flat
terrain. A similar adjustment of receptor positions is employed for
receptors above Hc, as illustrated in Figure 4. The upper portion
of the figure shows how a plume in the flow above Hc distorts over a
hill as viewed from above. Three streamlines are marked, the
centerline of the plume, and streamlines passing through receptors A
and B. When the deflection of each streamline is removed, and the
distortion in the plume is scaled out, an equivalent plume-receptor
geometry is obtained, as illustrated in the lower portion of the
figure.
Many of the concepts contained in CTDM are not present in complex
terrain screening models currently in use for regulatory assessments.
Partitioning of plume material about Hc in the vertical and about
the stagnation streamline in the horizontal is unique to CTDM. This
partitioning is fundamental to describing the transport of plume
material in the flow field around hills. Furthermore, the treatment
of the effect of the hill on the dispersion process for material above
Hc avoids the use of the plume height correction factor found in
other models. This factor is typically applied as a function of
stability and receptor height only, and it leads to an inconsistent
treatment of reflection of plume material from the lower boundary.
Essentially, the height of the plume above the ground is constant all
of the way from the source to a receptor, but this height varies from
receptor to receptor. Hence, adjacent receptors at unequal terrain
elevations are modeled with two very different plumes. If an
impingement computation is invoked, this treatment produces a
concentration equal to twice that at the center of the plume in the
absence of terrain. In CTDM, the impingement concentration is equal
to that at the center of the plume.
The method used to specify the rate of plume growth also differs
from the other models. Both Oy and oz functions depend on the
turbulence intensity, rather than stability class. In the case of
oz, the function describing the rate of growth with time also
depends on the scale of the mixing processes, which depends on the
elevation of the plume above the surface, and on the stratification
and turbulence near this elevation. In contrast, the other models
incorporate a fixed rate-of-growth function for each stability class,
and do not contain the influence of processes at plume height.
18
-------
r- A. PLUME CENTERLINE
WITH DEFLECTION * DISTORTION
PLUME CENTERLINE
B
DEFLECTION * DISTORTION REMOVED
Figure 4. Depiction of plume behavior in CTDM as it passes over a hill
as seen from ahove: upper figure shows actual deflection
and distortion of the plume, bottom figure shows treatment
within CTDM with source-receptor spacing equal to spacing
between streamlines upwind of hill and appropriate
adjustment of effective ov.
19
-------
3.2 Use of Meteorological and Terrain Information
The preprocessors for meteorological data and terrain data
provide CTDM with the information needed to compute concentrations,
but several assumptions are made within the model to convert this
information into specific variables used in the computations. This
subsection describes those assumptions.
3.2.1 Meteorological Data
The meteorological preprocessor provides CTDM with meteorological
data at several heights above the ground (corresponding to measurement
heights) , and it also provides CTDM with surface boundary layer
parameters which allow the model to compute profiles of wind,
temperature, and turbulence within the surface layer, when needed.
Whenever CTDM needs meteorological data at a certain elevation, linear
interpolation between measurement levels is employed. If the
elevation exceeds the uppermost measurement level, the measurements at
the uppermost level are used if this height is above the mixing
height; otherwise the value at the uppermost height is scaled upward.
The most extensive use of the profile data is made in the course
of computing the dividing streamline height He and the Froude number
for the flow above He. Hc is computed for each hill by locating
the lowest height at which the kinetic energy of the approach flow
just balances the potential energy attained in elevating a fluid
parcel from this height to the top of the hill. The statement that
defines this balance is:
| u2(H ) = /H N2(z)(H-z) dz (1)
2 c H
where u(Hc) is the wind speed at z = Hc, H is the elevation of the
top of the hill, and H(z) is the Brunt-Vaisala frequency at height z.
In practice, the value of Hc is obtained by rewriting the integral
on the right-hand side (1HS) of Equation 1 as a series of sums over
layers of constant H. For each layer, say the itn layer,
™
where z^ denotes the mean height of the layer, 0.5 (z^ + zt_i) .
The layer that contains Hc is found by comparing the LHS of Equation
1 at each measurement height with the accumulated HHS^ for all of
the layers between that measurement height and the top of the hill.
If the LHS exceeds the accumulated RHS^, then Hc must lie below
that measurement level, and so the process is repeated until the
lowest level is found for which the LHS becomes less than the RHS.
This then identifies the layer that contains Hc.
20
-------
Hc is computed within this Layer by assuming that the wind
speed follows a linear profile. Denote this as layer j, where the
elevations at the top and bottom of the layer are z\ and Zj_^,
respectively. Denote u(z) in the layer as
u(z) = a. -t- b. z
•J J
then equation 1 becomes
. . 2 . . .
2 J J c j J c j c i
(a. + b. H ) = N (H - l/2[z. -h H ]) (z. - H ) + I RHS. (3)
where the last term, I RHS^, denotes a sum of the RHS^ of all
layers between Zj and the top of the hill. Equation 3 is quadratic
in He, and is readily solved for Hc.
Once Hc is computed for a hill, the Froude number above Hc is
computed as
(A)
where um and Ng, are average values over a layer of depth 1.5
(H-HC), above Hc. This Froude number characterizes the degree of
stratification of the flow above Hc. Note that t^ is computed
from the temperature difference across the layer.
The wind speed shear in the flow above Hc is computed between
the plume height and Hc. However, if the plume height less Hc is
less than one tenth of the hill height, the shear is computed between
the top of the hill and He. This wind speed shear is used in
computing the flow over the hill (see Section 3.3.6).
3.2.2 Terrain Data
The terrain preprocessor provides CTDM with the location, size,
shape, and orientation of each hill or segment of a hill identified by
the user. This information is tabulated as a series of ellipses which
make up a family of horizontal cross-sections of the hill, and a
series of variables describing inverse polynomial (bell-shaped)
profiles which approximate the portion of the hill above each of the
ellipses.
For non-zero He, CTDM uses an ellipse to characterize the shape
of the hill below Hc. The ellipse used for plume material below
Hc is found at the minimum of Hc and the height of the plume. The
center of this ellipse, its orientation, and the axis lengths are
obtained by linear interpolation between the ellipses provided by the
preprocessor at elevations above and below the target elevation.
21
-------
Above Hc, CTDM uses the orientation, height, and Length scales
of the inverse polynomial shape used to approximate the portion of the
hill that lies above the elevation of Hc. However, the current
version of CTDM assumes a Gaussian shape in the formulation of the
flow algorithm rather than the inverse polynomial function. To
convert from the inverse polynomial description to the Gaussian
description, the length scale used in the former shape is divided by
/T75 to convert it to the length scale of the Gaussian shape.
This factor had actually been derived to convert from an inverse
polynomial shape to an equivalent elliptical shape used in an earlier
version of the model. The length scale in the inverse polynomial
description is the half-length of the hill at one-half of the height
of the hill. By demanding that the equivalent ellipse shape coincide
with this profile at the top of the hill and at the point at half of
the height (See Figure 5), the length of the axis of the ellipse is
found to be equal to the length scale at half of the height divided by
/, 75. For the Gaussian profile, the length scale is half the length
of the hill at an elevation of 1/e of the height of the hill. This
length scale is approximately 4% less than the axis length of the
ellipse (conversion factor = /.693), so the earlier conversion
factor was retained.
3.3 Derivation of Concentration Equations
3.3.1 Plume Rise Calculations
Momentum Rise
Momentum rise is used only if the plume rise due to buoyancy flux
is zero (stack temperature not greater than ambient). The following
formulas are used (Briggs, 1975) for momentum rise:
neutral/unstable Ah = 3 dws/us (5)
2 d2 T\ 1/3
stable Ah = 1.5 * a) 3~l/6 (6)
s s/
where d = stack diameter
w3 = stack gas exit velocity
us = stack top wind speed
Ta = ambient temperature
Ts 3 stack gas temperature
s = H2 => (g/Ta) (de/dz).
If L > 0 (stable), the minimum plume rise from equation (5) and
(6) is used.
An iterative technique is used for calculating buoyancy rise,
since the plume rise is assumed to be a function of the wind speed and
temperature gradient at a height halfway between the stack top and
22
-------
L(e)-
• Gaussian profile
i elliptical profile
inverse polynomial profile
Figure 5. Relationship among length scales used to specify inverse
polynomial Cip). Gaussian (G). and elliptical (e) profiles
Note that the equivalence of L(G) and L(e) is approximate.
23
-------
final plume height. For some plume rise formulas, this method does
not converge for certain profiles of wind speed and/or d8/dz.
Therefore iteration is stopped after five tries at convergence
(defined by less than 1% change between successive iterations). After
each iteration (and the fifth one, if no convergence), the plume rise
guess for the next iteration (or the final rise, if no convergence) is
the average of the plume rise estimate for the previous two iterations.
neutral/Unstable Buoyant Final Rise
These formulas from Briggs (1975) apply for plumes within the
mixed layer:
2 1/3
Final transitional rise: Ah = 1.6 (Fxf ) /u (7a)
where xf = 119 F2/5 for F > 55m4 s~3 and (7b)
xf = 49 F5/8 for F < 55m4 s~3 (7c)
Unstable breakup rise: Ah = 4.3 (F/u)3/5H~2/5 (8)
2h 2
p 3
Touchdown plume rise: Ah = 1.0 — r (1 + ~rr) (9)
2 Ah
h 2
Neutral breakup rise: Ah - 1.3 -~ (1 + TT) (10)
2 An
where u a wind speed at height hg +• Ah/2,
F » buoyancy flux, 0.25 wgd2g (Ts-Ta)/Ts,
3
H a -u^/(0.4L) is the surface heat flux,
Wd s. 0.4 W*.
The final neutral/unstable plume rise is the minimum of Equations
7 through 10.
Neutral/Stable Buoyant Final Rise
There are several final plume rise formulas available for stable
conditions, depending on whether winds are nearly calm or not, and
depending on whether conditions are close to neutrality. The final
neutral/stable rise that is used in CTDM is the minimum of those
calculated by means of Equations 7, 10, and the following equations
(11 and 12):
Neutral high wind rise:
2 2/3 1/3
Wh = 1.54 [F/(uu/)] h (11)
Bent-over stable: Ah = 2.6 (F/(us))1/3) (12a)
24
-------
Calm stable: Ah = 4 F1/4 s~3/8 (12b)
In stable conditions, the distance to final rise, Xf, is given
by
Xf = 2.07 u s-1/2. (12c)
3.3.2 Dispersion Parameters
Derivation of or
The formulation for az is described in Venkatram et al.
(1984). It is based on the form of Taylor's (1921) theorem of
diffusion for very short and very long times of travel:
a = a t ; t«TT (13a)
z w L
1/2
a = (2K t) ; t»T. (13b)
z z LI
where ow is the standard deviation of the vertical velocity
fluctuations, t is the travel-time, T^ is the Lagrangian time-scale,
and K£ is the eddy diffusivity. Using mixing-length arguments, Kz
is defined as
K =» a I ; I = o TT (14)
z w w L
so that in the limit t»TL, Equation 13b becomes
oz = ow (2t TL)1/2. (15)
An interpolation formula used by other authors (Deardorff and Willis,
1975) is employed to span the gap between small and large times of
travel:
0=0 t/(l + t/2T_)1/2. (16)
z w L
In Equation 16, ow is measured directly, but TL must be
related to other measurements before oz can be calculated in the
model. This is accomplished by using the empirical flux-profile
relationships of surface similarity theory (Businger, 1973) to
estimate the mixing length, I. The derivation relies on the
appropriateness of surface similarity theory and on the assumption
that aw is proportional to the friction velocity, u*. If ow
is due mainly to turbulent fluctuations, then the formulation is
appropriate.
Assume that passive material diffuses in the same way as heat in
a turbulent flow so that = , and
25
-------
u ,kz o kz
K^ = * « W (17)
"'h a^ is the non-dimensional
potential temperature gradient, and a is the constant of
proportionality between ow and u* applicable to the stable
boundary layer:
u* . ( 18 )
From Equation 14, the length scale, I, can be written as
K
K s z = kz = kz _ (19)
a aq>. a(
-------
I = ^2ow/N = ls. (24)
When N is nearly zero and z is finite, the Length scale for the mixing
process is proportional to height above the surface
I = TZ = ln. (25)
The quantities SLS and ln are introduced to distinguish the
mixing lengths for the stable and neutral limits.
Equation 24 states that the length scale for turbulent mixing in
the stable limit is proportional to ow/N, where Y2 is the
constant of proportionality. This is consistent with the notion that
a fluid element in this limit must overcome a stable potential
temperature gradient in order to be displaced vertically. Given that
the velocity scale for vertical dispersion is cw, the length scale
that naturally follows is proportional to ow/N. In the neutral
limit, the size of the turbulent eddies is restricted by the height
above the surface, so that the mixing length should be proportional to
z, where F is the constant of proportionality. Note that
y = .52
(26)
r = .36
for the choices 8 a 4.7, a = 1.3, and k = .35.
Equation 22 may be viewed as in weighting function for I
between the limits- ls and ln. A simpler weighting function is
actually used in CTDM:
•-*-•
It produces values of 8. which are within 20% of those produced by
Equation 22. With this expression for I, TL is computed as
I/GW (Equation 14). and oz is computed from Equation 16 as a
function of the time-of -travel, including source effects.
Derivation of Qy
An equation similar to that used for oz is used to compute
Oy as a function of the time of travel and the turbulence velocity
scale for lateral fluctuations, ov:
a
y
ovt/(l + t/2TL')1/2. (28)
The departure of this expression from that for oz arises in
specifying the functional form of T^', which is the Lagrangian
time-scale of the transverse correlogram. In this case, TL' cannot
be derived from measurements of the mean flow and its statistics.
27
-------
Attempts were made to relate TL' to the difference among the
statistics of lateral fluctuations over periods of five minutes and
those for periods of one hour, but these types of data are typically
not available for routine applications of CTDM. Because no clear
improvement in the performance of the model could be associated with
the use of these data in a formulation of TL', that formulation was
not incorporated in CTDM.
Instead, TL' is set equal to the time-of-travel required to
cover a distance of 10 km. This choice reduces
-------
2T- a , 1/2
1 + [1 + 4 (-J*— ) ]
t = - ~ — ~ - - - (31)
The virtual time of travel (tv) is found by setting 04 equal
to the plume size caused by source-induced effects (0j,s) and
setting t equal to ts + tv, where ts is the time-of-travel to
the point where ambient turbulence dominates source- induced turbulence:
1/2
[i
t = -t + - r - r- - (32)
V S \ <— )
u O«
Sis
As implemented, 09,3 is just the size of the plume resulting
from buoyant rise (buoyancy-enhanced dispersion):
°ls = °lb " Ah/3.5 (33)
where Ah is the plume rise. Similarly, ts is the time-of-travel
to final rise (see Section 3.3.1). Ambient turbulence is assumed to
dominate source-induced turbulence quickly during neutral and unstable
conditions, so that ts is set to zero for these conditions.
Depending on the relative values of o (ambient turbulence) and
-------
from the source, see Figure 2). If Hc is zero, then this zone
extends from the source to the base of the hill, although it
conceptually could extend to any point where the hill is thought to
exert a significant influence on the flow. Beyond so, the plume
material below Hc is disregarded by the LIFT component, and the
evolution of the remaining material is modeled as if the terrain were
flat, and the lower boundary were Hc (with full reflection).
However, the rate of plume spread and the position of the plume
centerline relative to the receptor are modified to reflect the net
alternation of these properties between s0 and s (where s is the
distance from the source to the receptor) induced by the presence of
the hill. The simplicity of the Gaussian plume solution is retained
in this way, while the full dilution of the plume from the source to
the hill (s0) as well as the effects of the hill on both flow and
dispersion beyond so are explicitly incorporated.
The terrain effect as modeled in LIFT includes re-initializing
the flow at a distance so downwind of the release. This
re-initialization can be illustrated first for flat terrain and
uniform flow. The concentration at a receptor downwind of SQ is
composed of contributions from the entire concentration distribution
at s0. Conceptually, the flux of plume material through the plane x
= xs + so (note that the x-axis lies along the flow direction, and
the plume is released at xg, yg, zs) can be thought of as a
distribution of point sources. If we. track the plume material in
terms of the distance downwind of the source, s = x - xg, then the
source strength of one of these point source elements located at the
point (so, y, z) is given by:
dQ(s0, y, z) = C(s0, y, z) u dy dz. (34)
Because the flow beyond so is considered to be uniform, the
influence of each of these sources follows the Gaussian plume solution
to the advective diffusion equation so that the contribution of the
source element at the point (s0, y, z) to the concentration at the
point (s, I, h) is;
dC(s.t,h;so)
dQ(s ,y,z)
<35)
where ay*t az* denote the plume spread statistics for each
point source element over the interval s - s0 (see Figure 6). The
total concentration at (s.l.h) is found by integrating Equation 35
over all point source elements, so that
30
-------
3410017
u
0(80.20)
/I ////// I 7 i I I I
C(S. K SQ,
C(s, z) = I C (s,
0
•2
where a
Figure 6. Illustration of the relationship between the crosswind-average
concentration profiles at s0 and s, and the plume from one
of many point-source elements representing the flux of
material across the plane at so. The total concentration at
a particular point G(s,z) is constructed by summing the
contribution C(s,z; so,zo) from each point-source element
Q(s0,z0).
31
-------
C(s y.z) 5(2=1)2 z-h 2
e °-5<« *> °"5 (*
o -•*• y z
z+h.2
e~°-5(o *; ] dy dz. (36)
2s
The plume spread statistics Oy* and oz* for the interval
s-s0 are specified by the requirement that Equation 36 for flat
terrain reduces to the expression obtained for the original point
source located at s = 0 (i.e., Equation 35 with so = 0 and a* =
o(s)). Equating these two expressions for C, with h = 0, we obtain
V2 = °z2(3) - °«2(V = °z2 - °zo2 (37a)
V2 - °y2(s) - °y2(°o) E V ' V*' (3?b>
Equations 36 and 37 illustrate the re-initialization technique
for the limiting case of flat terrain and uniform flow. Overcamp
(1983) has developed a similar technique for replacing a simple image
source in a general treatment of the lower boundary condition for the
case of non-Fickian diffusion.
Terrain influences are incorporated by altering the rate of
diffusion within the interval s-so, and by changing the' position of
the receptor relative to the centerline of the deflected plume.
Furthermore, because no plume material below Hc travels over the
hill and because Hc defines the lower boundary over the hill beyond
so (see Figure 3), the integration in the vertical in Equation 36 is
performed over the domain z=Hc to z=» and material from each point
source element is reflected from the boundary z=Hc rather than z^=0.
Denote the receptor height above ground relative to the
terrain-altered plume as hs' (see the lower part of Figure 2) and
its lateral position relative to the terrain-altered plume centerline
as yR' . Further denote the height of a point source element above
the ground at so as hg + Hc, and the altered growth rates of the
plume as oz*' and Oy*" . Then the contribution of the element
at the point (so, y. hg) to the concentration at the point (s,
yK', hRf) is:
dQ(s ,y,h )
h -hh '
_0 5<-JLjR_)2
* V' ]- (38)
32
-------
The total concentration at the receptor (s, y^' , hR') is found by
integrating Equation 38 over all point source elements above Hc (at
so) , so that:
- -H. C(s y,h ) g/"7*')2 5(W }2
vy ' 2«« **o *• e v [e °r *
Hc -« y z
y 2
(39)
where the concentration profile at s0 is expressed in terms
of hg:
ys~y 2 Zs"hs~Hc 2
-0.5( - ), -0.5(— -
e ° [e
o yo
yo zo
WH
(40)
for a plume released at (0,ys,zg).
The result of doing the integrals in Equation 39 is simplified by
defining the effective plume spread parameters oze and 0ve.
First, define terrain-effect factors Tz and T» so that
o *• 3 o-*/T
z z z
o *' = o */T . (41)
y y y
Then the effective plume spread, accounting for the effect of strain
in the flow on the rate at which material diffuses across streamlines,
is defined as
o 2 - o 2 + (o */T )2
ze zo z z
2< (42)
Equation 40 can be substituted into Equation 39, and after isolating
terms containing y and hs, we find that the exponential functions
containing either y or hs can be combined and isolated. By
completing the squares in the argument of these functions, and using
the definitions of Equation 42, the concentration at the receptor can
be written as:
33
-------
(43a)
where
Fz
'-l-z -H
« C)
(43b)
and
A = (/ o o o */T )~. (43c)
2 zo ze z z
Equation 43 provides the framework for estimating concentrations
due to plume material that travels up and over a hill. It shows how
.the influence of the terrain affects the magnitude and the
distribution of GLC's. The most complicated part of Equation 43 is
the expression for Fz> the vertical distribution factor. It
contains four terms because it applies to an elevated receptor, so the
image source contribution is not equal to the contribution from the
primary source (hence two terms are needed rather than one) . And it
also applies to a plume segment above Hc rather than an entire plume
profile, so that an image source contribution at so must be
explicitly maintained (hence, two more terms). The error functions
also arise from treating only the portion of plume material that lies
above Hc at so. If Hc is zero, then Equation 43 becomes:
-0.5(
o ze
ye ze
34
-------
L!V
a ; ] (44)
which, with the exception of alterations in the plume size and in the
distance between the plume center-line and the receptor, is the
familiar Gaussian plume equation for the concentration at an elevated
receptor.
The receptor position in Equation 43 is denoted as (hg',
yg'), where the primes indicate that the distance of the receptor
from the centerline of the plume has been altered by the presence of
the hill. Upwind of the hill, streamlines are straight and parallel.
Over the hill, the spacing is variable and the streamlines are subject
to deflections. Because concentrations are computed for a parallel
flow in which the influence of terrain is manifested in the
Hc-partition of plume material, in the altered rates of diffusion,
and in the altered streamline that, passes through the receptor, the
chief task in computing (hg'.yg1) is in identifying the streamline
that passes through the source, and the streamline that passes through
the receptor. The position of these two streamlines in the
undisturbed flow upwind of the hill defines the effective receptor
location relative to the centerline of the plume. These streamlines
are obtained from the flow model described in subsection 3.3.6.
The design of the.flow model is particularly well-suited to
obtaining hg' and yg' because it is formulated as a
"backwards-looking" solution. It is designed to compute the
deflection experienced by a streamline that passes through a given
point over the hill. If the given point is a receptor, then the model
will compute the deflection experienced by the streamline that passes
through that receptor. Knowing the actual position of the receptor,
the deflections allow hg' and yg' to be computed.
The LIFT equation for concentration, Equation 43, is applied to
all receptors on a hill that lie above He, and are downwind of the
point of impingement or the upwind base of the hill (for He = 0).
Given the way the model must deal with a highly idealized description
of the terrain, there are times when a receptor may lie above He,
but may be positioned upwind of the point of impingement. For
example, the receptor may be on a mast which places it above Hc, or
the terrain on which a ground-level receptor is placed may exceed
Hg. Concentrations at both of these receptors would be estimated
with Equation 44 with two changes:
1) no terrain-effects would be included in computing
-------
the height of the "pole". In this way, the flow above Hc
rides up and over any portion of the hill that lies above
Hc.
3.3.4 Terrain Factors for LIFT: Tz and Ty
For axi-symmetric strain, the theory of Hunt and Mulhearn (1973)
shows that
B (t) o
B2(t')K(t')dt' (45)
where
-------
Streamline 2
Streamline 1 -—-:
'
T .
ha
o
Z2 - Zl
AZ
Az
Figure 7. Derivation of the factor T^ in finite-difference form. The
dashed lines represent the path of two adjacent streamlines,
as they are deflected by the presence of terrain.
37
-------
With this mean distortion, the height n(z,t) of a streamline above
the surface of the hill at time t is approximately
n(z.t) (50)
where z is the height of streamline well upwind of the hill. Because
of the form of Equation 50, Th is equivalent to a "height-correction
factor" for a plume.
To evaluate Tz, consider the argument of the exponential
function in the vertical distribution factor of Equation 44 when the
elevation of the receptor above the surface is zero. The primary
quantity in this argument is the ratio zg/oze, where «?ze is
given in Equation 42. The corresponding ratio in the approach of Hunt
and Mulhearn is n(zg,t)/
-------
Using Equations 48 and 49, the strain function can be rewritten
in terms of T^ as:
-(T.-1)
flz
-------
where Tow is the ratio
-------
process. Each "guess" at where the streamline may be at time t
produces a computed position of that streamline in the incident flow.
New guesses are dependent on the difference between the last computed
position, and the position of the targeted streamline.
3.3.5 Internal Mixing Layer for LIFT
A central assumption involved in the use of the LIFT equations is
that the turbulence field is homogeneous. While this assumption is
violated in nearly all applications of the Gaussian plume model, there
is one situation in modeling the dispersion of plume material over a
hill where such a violation is severe.
An elevated plume in a very stably-stratified flow may reside in
a region of nearly laminar flow, with very little turbulence. The
turbulent boundary layer beneath this region can be very shallow.
When Hc is greater than the depth of the turbulent flow, the entire
laminar region may flow over the terrain, and an internal boundary
layer must form at the bottom of this layer. If this internal
boundary layer were ignored, then the diffusivity of plume material
would remain virtually zero, and the effect of the terrain would only
be seen in the Hc-partition of plume material.
To provide a more realistic treatment of the very stable limit, a
simple mixing layer is included. Within the mixing layer, the
vertical distribution of plume material is uniform, and the
concentration is obtained by sampling the vertical distribution in the
absence of the mixing layer, and taking the average value over the
depth of the layer.
The depth of the mixing layer is estimated by focusing on the
initial stages in the development of such a layer, when turbulent
mixing is very strong. Assume that the turbulence in the layer
produces a layer of constant potential temperature (see Figure 8). As
a result of the rapid mixing in the layer, assume that buoyancy
effects are negligible in the layer, so that the wind speed takes on a
logarithmic structure in the vertical. Wind speed is continuous at
the top of the layer, but its gradient is not.
Csanady (1974) discusses a theory for estimating the evolution of
the depth of such a surface layer. He uses the result of laboratory
simulations by Kato and Phillips (1969) in which a layer is thickening
into a fluid with constant Brunt-Vaisala frequency N, as a result of a
constant stress applied at the bottom of the layer:
1_ dh 2'5 "* 9m
u^ dt ' g A6 h * (61)
Here, h is the thickness of the layer, u* is the surface friction
velocity, g* is the acceleration due to gravity, 6m is the
potential temperature of the mixed layer, and A6 is the jump in
-------
potential temperature at the top of the Layer. This result is
consistent with the notion that the rate of thickening of the layer is
a function of a buoyancy parameter,
b = gAe/em, (62)
the surface stress (expressed as u*) , and the thickness of the layer
so that the entrainment velocity, when scaled by the friction
velocity, is a function of the dimensionless group bh/ui:
). (63)
In essence, the entrainment depends on the turbulence generated at the
surface of the layer, and not on any assumptions about shear-generated
turbulence at the interface.
Referring to Figure 8,
6 =9 +46 = 6 +-j^ £ (64)
mo o dz 2
so that
dh 2"5u*3 290
1)- (65)
If we assume that the layer depths will generally be small enough that
e0 » h de/dz, or that 8m « 6O, then the subsequent
evaluation of the integral is greatly simplified.
The friction velocity in Equation 65 is given by
U* = (66)
where u is the mean wind speed outside of the layer, k is von Karman's
constant, and zo is the roughness length for the surface of the
hill. Rewriting Equation 65 with 8O = 0m, we find
<* 5eo kV
dx=g de/dz h2[ln(h/z 3 • <67)
o
Noting that 15 k3 = 1, an implicit equation for h is obtained:
-------
= Constant
W/w////////////^^^
Potential Temperature Profile
^
Wind Speed Profile
Figure 8. Illustration of the structure assumed for the developing
internal mixing layer over the hill above Hc.
43
-------
. .
0 O
X - X
- — 1£-' (68)
u/N
This implicit equation has two undesirable traits. The growth of
the mixing layer is extremely rapid for very small times-of-travel
(downwind distance) from the point where the flow first encounters the
hill above Hc, and the estimates of mixing depths "blow up" for weak
stratification. The first of these traits is removed by demanding
that the rate-of-growth not exceed unity (slope of the interface
equals 45*). The second trait is removed by limiting how large u/N
can be in Equation 68.
For convenience, u/N is set equal to the minimum of u/N and
H<.. This is done because we wish to turn on the mixing layer
primarily for those cases in which turbulence at plume height is
decoupled from the surface. As Hc becomes less than half the hill
height, there is a strong likelihood that the turbulent boundary layer
in the approach flow encompasses most of the depth of the flow over
the hill. Hence, in using the minimum of u/N and Hc, the stability
is artificially increased for Hc less than H/2 so that the depth of
the mixing layer will decrease. As Hc goes to zero, the mixing
layer is completely absent, as desired.
3.3.6 Flow Model for LIFT
Model Development
CTDM simplifies the treatment of stratified flow over a hill by
separating the flow into two regimes: a lower portion, below Hc,
which is either blocked (in the ease of a long ridge) or flows around
the hill, and an upper portion, above He, which has sufficient
kinetic energy to flow up and over the hill. This simplification
means that for a flow of constant speed, u, which lifts up and over
the hill, one always has the property that (H-Hc)M/u < 1; or
equivalently, that the Froude number for this portion of the flow
exceeds unity and the flow is not "strongly" stratified. This fact,
coupled with the assumption that the hill is not too steep (i.e., less
than about 15") enables one to use the linearized equations of motion
for steady-state Boussinesq flow (Smith, 1980):
3u' 3p«
V aT ' - aT (69a)
o a
aw1
V aT ' ' aT - p'8 <69e)
3u' 3V 3w'
* + = ° (69d)
-------
and
3pn
P' = - (-) n (69e)
These equations, in which x, y, and z indicate downstream,
cross-stream, and vertical coordinates respectively, relate the
perturbation velocities u', V and w' to the perturbation density,
P* , pressure p', and vertical fluid displacement, n = n(x,y,z),
and to the unperturbed initial velocity u and density po. It
should be noted that n(x, y, z) is the vertical displacement that a
parcel of air at the point (x, y, z) has experienced. Adding the
kinematic condition for steady flow in a shear-free approach flow gives
W = u Jj . (70)
Equation 69 can be reduced to the single partial differential equation;
3-r
-------
where
m = n (1 4- L 2/L 2)1/2. (73b)
X y
It has the simple particular solution
n = - |£ (74)
oZ
with G = cos(mR)/R and R2 = x2 + y2 + z2.
This is the Green's function solution for a delta function hill
(i.e., a mathematically narrow but high hill of unit volume). It is
extended to a general hill shape h(x,y) via a two-dimensional
integration combining the hill shape function and the Green's funtion
(i.e, the convolution theorem) and the lower boundary condition that
the flow at the surface follow the hill shape. Thus, for an arbitrary
hill
I = // dx'dy h(x-x", y-y') G(x',y',ze) (75)
and
n - - fj. (76.)
where use of the height above terrain, z' , instead of absolute z
reflects the fact that the lower boundary condition has been
linearized; that is, the transformation to terrain-following
coordinates assumes very shallow terrain.
The other quantities needed for a complete description of the
flow are obtained by using Equation 76a and going back to the basic
equations of motion, yielding:
2
u'/u = - [— (I) + n I] = p'/(p u2) (76b)
dx
V'/U = - [fxfy-(I) + "2 _/X fy- dX'] (76e)
32I
and w'/u » - — -. (76d)
It should be noted that lateral perturbation velocity, v* , and the
subsequent definition of lateral deflection, 6, as
4 = _JX dx' (V/u) - - l + ^XX> dx» dx']
make the additional assumption that streamline deflections are small
enough that integrating along the x-axis at y, z' from x = -» to x
is equivalent to integrating along the streamline from x = -».
46
-------
Equation 76e also provides another example of how higher-order terms
are neglected, because a geometrical picture of lateral deflection
would indicate that the denominator factor u should actually be the
true x-component of velocity, or approximately u + u1. Note that the
deflections n and £ are those experienced by a streamline that
passes through the point (x.y.z1). They tell us where that streamline
originates in the incident flow.
The local strain factors Tj, and Tn used in CTDM can also be
related to I through these quantities:
Tn -(1 -1?1 = (76f)
oZ
2 t 2
_ ._ oo. ™X ._ o X ^ fX » X o X „ „ « . v ~ X * ^ .* »
TB = (1 - —) = (1 + —T + n / I —T dx"dx') . (76g)
1 a* 3y2 -» -- 3y2
Tn and T^ are factors that relate the spacing between
adjacent streamlines in the deformed flow to the spacing in the
incident flow. The gradient of the deflection (either 3n/3z or
34/3y) measures the difference in the deflection experienced by
adjacent streamlines and hence, the degree to which the streamlines
converge or diverge. In this backwards-looking formulation, the
deflection is a function of the streamline position (coordinates)
after deformation (see Figure 9). As detailed in the figure, the
finite-difference form of Equation 76f (and 76g) is readily obtained
from the definitions of streamline positions before and after the
deformation. Note that 3n/9z = 1 is a singular point. This
condition would result if streamlines that pass through z^« and
Z2« were to originate at the same height .in the incident flow. Also
note that Equation 76f differs from Equation 49 because the deflection
in Equation 49 is that for a forward-looking formulation.
Before evaluating Equations 75 and 76 for a realistic hill shape,
it is worthwhile to look back to the solution, 6 = cos(mR)/R, given by
Equation 74, as properties of the delta function hill solution will
show up for finite hills as well. One positive aspect is that as
stratification disappears (i.e., H, n, m go to zero and cos(mR) equals
one), the exact neutral result 6 =» 1/R is recovered. This exact
neutral result of 1/R is the well-known solution of Laplace's
equation, V26 = 0 (i.e., except at x=y=0), that is incorporated
into neutral flow solvers such as that of Hess and Smith (1962).
Thus, the approximation expressed by Equation 72b will not affect
neutral flow but only the modifications to the neutral flow solution
created by stratification. That these stratification influences might
not be too severe can be anticipated by noting that expansion of the
cosine term, as cos(mR) = 1 - 1/2 m2R2, indicates that changes
to the flow are second-order in m (i.e., m2) rather than
first-order. However, a disturbing aspect of this cos(mR) dependence
47
-------
Streamline 2
Streamline 1 ^<""'
T
h
(Z2°n2} " (Zl~
Az'
Az1
Figure 9. Derivation of the factor T^ in finite-difference form for
the "backwards-looking" formulation. The dashed lines
represent the path of two adjacent streamlines, as they are
deflected by the presence of terrain.
-------
is its iso tropic nature; that is, z has no special significance in the
equation, despite the fact that the density stratification and thus
the atmosphere's "springiness" is a z-oriented phenomenon. Exact
solutions to Equation 71 do not display this isotropic behavior.
Thus, one physically significant ramification of the approximation in
Equation 72b becomes apparent. Such isotropic and thus lateral
springiness does, however, occur in the aforementioned infinite field
of cosine hills problem but steps must be taken to suppress this
effect in the single isolated hill problem.
Finally, the connection to the neutral limit solution, G = 1/R,
suggests a way to inject wind speed shear (with height) back into this
shear-free solution. Crapper (1959) points out that the corresponding
vertical deflection in a neutral, shear flow is just
• *'/R3 ("a)
whereas our current neutral form, G = 1/R, yields
n = z'/R3 (77b)
in the shear-free case. Unfortunately, even this simple factor is
awkward to superpose back onto G exactly. The approximation chosen
for this model yields a final Green's function of
1/2
G - [u(0)/u(z')r cos(mR)/R. (78)
Use of the shear exponent of 1/2 represents a compromise between
including shear effects and avoiding negative impacts on the flow
prediction generated by the inexact nature of the approximate solution
given by Equation 78. Any remaining unwanted residual consequences of
this approximation are suppressed by enforcing the constraint that the
deflection of the streamline at the ground equal the rise of the
surface of the hill: -Iz> = n = h(x,y) at z' = 0. This ensures
that no stremalines pass beneath the surface of the hill.
Algorithm Development
Referring back to Equation 76, one notes that all quantities of
interest depend on integrals and derivatives of the basic quantity I,
given by Equation 75 and with the use of the approximate shear
solution G expressed by Equation 78. Evaluation of I requires that
the shape of the hill be chosen carefully. Selection of an
appropriate hill shape was governed by
i) the desire to treat as general a shape and orientation as
possible, and
ii) the necessity for performing the integrations in Equation 75
analytically.
49
-------
A Gaussian-shaped hill is used because it contains the overall
orientation and scales of the "cut-off" hill, and because it has a
tractable mathematical form. The most general Gaussian shape that is
considered involves a hill of height h, having elliptical contours
with major axis La, and a minor axis L^, oriented at an angle *
which is counter-clockwise with respect to the flow direction. For a
coordinate system with the x-axis aligned with the direction of flow,
the hill is prescribed by
h(x,y) = h exp{-[x2/L 2 + y2/L 2 +2rxy]} (79)
x y
where
.
1 f cos Or sin 0?
1 eos * sin
and
Y » [—~ - —~] cos* sin* .
V V
These equations are generated by first writing the hill shape
function, h(x", y") a h exp{-[x"2/Lb 2 +• y"2/La 2]}, in the hill's
natural coordinate system (x'% y"), where x" and y" are distances
along the minor and major axes, respectively, from the hill center,
and then inserting the rotational relationships, x" = x cos* 4- y sin*
and y" = y cos* - x sin*, to eliminate the appearance of the
coordinates (x", y").
For the symmetric hill with La a Lb =» L, the expression
simplifies greatly, as !*% = Ly = L and j - 0. In fact, once the
convoluted form of the hill function (i.e., h(x,y) is rewritten as
h(x-x',y-y')) and is substituted into the integral for I, one finds
that the exact analytic integration can be accomplished for a receptor
position (x.y.z') at the crest (and just off the crest) of such a
symmetric Gaussian hill. However, this analytic result, expressed in
terms of the real and imaginary parts of the complex error function,
can be greatly simplified without substantial loss of accuracy via the
approximation
exp(Z2) erfc(Z) = 1/(1 + 2Z/V?),
where 2 is a complex number having a positive real part. The
expression finally arrived at is
50
-------
h(x,y) L
i = ; • [^^rfri^m+z'/L ) • cos(mz-)
u+ij XI+Z'/L r uu ; n
o n
- a •sin(mz')/m-a »(x /L ) [m-cos(mz')+(— + - -rr) 'sinCmz') ]}
1 Z ID X L 2
n
where bQ = B^/ * ,
Ln=2 *Lz •
a is from the shear equation u(z') = u(0) + oz' , and
. 2 ._ la
a, = b /L - ~ "TT. ,
1 o n 2 u(0) '
(80)
• L • (1 + L _/L
and
all represent convenient clustering or redefinition of terms.
The two parameters, Bo and R^, arose because there was some
conflict between small and large argument, low-order expansions of the
complex error function and an ambiguity in the definition of Lz,
respectively. The parameter B0 was set to (*/2)1/2, the
geometric mean of the small and large argument limit values, whereas
RI was tuned to ln(2) = .0.693 based on optimization studies using
the tow tank data described by Snyder et al. (1986).
The expression for I given above is best described as an
admixture of analytic results for the symmetric hill and empirical
extensions for the asymmetric hill subject to the constraints that the
surface boundary condition be obeyed and that known results be
recovered for the 2-d ridge limit of Ly-*» and 4r=0. Higher-order
terms (e.g., (y/Ly)^ or other such terms arising far from the
crest) are neglected. In addition, we took the liberty of
switching-off the (X^/LX) term upwind of the hill (i.e., XQ < 0),
as it seemed to create excessive early streamline rise as the hill was
approached and because appropriate damping terms in x£ had
already been neglected in the denominator.
Evaluation of the flow variables in Equation 76 involve
straightforward integrations and differentiations of the basic
quantity I, given by Equation 80. Two exceptions to this involve:
51
-------
(i) Replacement of {1 -t- ERFCx,^!^)} with
{1 - ERFlXjn/Lx)}. This appears in terms involving
integrations and double integrations of I with respect to x,
and has the major effect of allowing streamlines to return
laterally to their initial upwind positions more rapidly
after passing by the hill. This relaxation adjustment can
be rationalized on the basis that Equation 80 is most valid
near the crest of the hill.
(ii) A "fully- implicit" style computation of the lateral
deflection. That is, instead of assuming that the
streamline spent most of its* time at its current lateral
displacement y, we conjecture that most of its time was
spent at its undefleeted position, y«o, evaluated back at
x » -«. This enables one to write down the Taylor series
relation 4* = * * (di/dy)(y« - y) between the &
from Equation 76e and the new corrected lateral deflection,
5*. Recognizing that y-y^ is just the corrected
deflection, 5*, one may rearrange terms and solve for 5*
as 4* a 6/[l + (di/dy)]. Such an approach is called
"implicit" because y^ is never explicitly computed and
"fully" implicit because the streamline is assumed to spend
"all" of its time at lateral position y^ before rapidly
moving out to its deflected position y. This compensates
for the fact that lateral deflections can be large and
therefore badly violate the "small deflection" assumption
invoked for Equation 76e. The resulting lateral deflections
are unfortunately smaller than before the correction, but
streamline crossover situations (which create a severe
problem for an iterative streamline solver) for large
lateral aspect ratio hills are greatly suppressed. This
suppression did not quite achieve the required elimination
of crossovers. The above described procedure had to be
carried to second order, i.e., 5* = 6 - (dS/dy) 5* +
1/2 (d2i/dy2) S*2 and solved, and the absolute
value of (d$/dy) taken (to account for a somewhat spurious
sign flip far from the center of the hill)8 before
streamline crossover could be completely blocked for high
lateral aspect-ratio hills (e.g., La = 10
These adjustments to the model, as well as the adjustment of some
of the parameters discussed in association with Equation 80, represent
empirical tuning of the flow algorithm to improve model performance
while avoiding pathologies (e.g., streamline crossover) which create
havoc for an iterative algorithm. It should be noted that these,
sometimes conflicting, considerations led us to consider several other
forms for the Green's function, G, other than the basic G = eos(mR)/R,
later augmented for shear via Equation 78. Some of these trial
Green's functions represented solutions to the governing Equation 73a;
whereas others, such as G = [l-sin(mS) ]/R, were designed to inject a
first-order dependence on stratification (i.e., m rather than the
leading m2 dependence from the cosine) into the solution. In
general, Green's functions incorporating a first-order dependence on m
52
-------
were much more successful in reproducing the tow tank data for 3-d
hills (to be discussed in the following section), but were also more
difficult to tame in terms of pathologies, especially those developing
far from the hill. Damping of these Green's functions at large R
provided some relief from these problems, but the need for an
operational algorithm totally free of pathologies ultimately dictated
that the more conservative cos(mR)/R form, with its relatively gentle
leading m^ stratification dependence, be adopted. Thus, there
remains opportunity for improvement in this field.
Algorithm Evaluation
The behavior of streamlines produced by the flow model is
qualitatively illustrated in the following 7 figures.
Figure 10 shows the paths followed by five streamlines (initially
released at 20, 40, 60, 80, and 100 m above the surface well upwind of
the hill) as they pass over the crest of a symmetric 100 m tall
Gaussian hill of aspect ratio 1. The aspect ratio is defined to be
the ratio of the half-length of the hill at one-half the height of the
hill, to the height of the hill. Therefore, a Gaussian hill of aspect
ratio 1, and height equal to 100 m, has a length scale L=120 m. The
streamline patterns in the figure span hill Froude numbers of 1, 2, 4
and infinity. Qualitative features apparent in these patterns include:
• compression of streamlines near the crest, with the degree
of compression increasing at greater stratifications (lower
Froude number); and
• greater upwind/downwind asymmetry of the streamlines with
increasing stratification such that the point of closest
approach of the streamline to the surface of the hill shifts
from the crest to the downwind side of the hill.
The same information is repeated in Figures 11 and 12 for a
symmetric Gaussian hill of aspect ratio 2 (L = 240 m), and an
asymmetric hill with an alongwind aspect ratio of 2 (L =» 240m) 'and a
crosswind aspect ratio of 10 (L = 1200 m). The flow over the latter
hill is taken to represent nearly two-dimensional flow in the vicinity
of the plane y=0, which is the plane that contains the streamlines
illustrated in the figure. The compression of streamlines over the
crest of the hill of aspect ratio 2 is weaker than that over the crest
of the hill of aspect ratio 1, but the upwind/downwind asymmetry is
stronger. Approaching the limit of the two-dimensional ridge, with an
alongwind aspect ratio of 2, the compression over the crest becomes
somewhat weaker while the upwind/downwind asymmetry is similar to that
for the symmetric hill of aspect ratio 2. These results are in
qualitative agreement with laboratory studies.
Figures 13 and 14 illustrate the lateral deflections produced by
the algorithm for a matrix of streamlines. All streamlines lie
originally at the same elevations used in the previous figures, but
they are arranged in a series of eight bands across the flow. The
53
-------
FLOW TRAJECTORIES
Aap*ct Ratio - 1; Fr «= 1
FLOW TRAJECTORIES
Aspect Ratio - 1; Fr - 4
13O
11O -
100 -
60 -
20 -
o -
-1
80
. •
-^
S
,
s
-1
—-•
•'"
^
f
X"
'
40
, '
"
X
X
'/
'
— 1
^
X
/
f
' /
7
00
^
x
/
//
V
-
^
_/
^
/f
W>
'^
X
s~
//
_ •
-
--
,^~-
~~~
2O
~--^_
^
--.
~— ,
"~~
2
X
X
X
x
^N
o
's.
[N.
x^
>\>
6
N
X
Vs
^s
\
o
-^
-v
X
\\
V
-^
X.
\
\v
\
100
~-^
\
s^
^
1'
-^
v
N
10
u
150 -
140 -
130 -
120 -
110 -
1OO -
00
80
70 -
60
50
4O
30 -
20 -
1O
O
FLOW TRAJECTORIES
Aspect Ratio « 1; Fr » 2
^
\
\x
180 -1-4O —1OO -6O -2O 2O 6O 1OO 14O ISO
X (m)
50 -
30 -
20 -
10 -
0 -
-ieo
150
140 -
110 -
120 -
110-
100 -
GO -
BO -
70 -
60 -
50 -
4O -
30 -
20 -
10 -
0 -
-
--
-^
"*^
^
'
-"
•^
^
^
^
^ f
/
-140
— -
•**
'
^
•
^
f--
/
/
j
/
^/
/
/
/
' '/
f
^
^
/
/
/
/
/•
/
/
/
/{
/
/
/
/,
//
s^
/r
/
/f
//
-*~~~
~*
^— -
-*°
~~-
^~
~
'
\
^
x^
\
N^
x
X
N
NN
x^
11
V
X
\
\
\
^^
v
\
\
\l
^^"^
^
X
N
\
\
\
s>
-100 -eo -20 20 eo 100
X (m)
FLOW TRAJECTORIES
Asptct Ratio «= 1; Fr •= Infinity
^S
s
/
/
' /
'
^
^
/
/
/
/
/>
'
/
/ 1
//
/
^
'
/•
/
/^
/
^
^
^
-*
"^"^
^~~~'
_^~\
— ~-
" •
\
^
X
X
x^
\
N
s\
x\
x
x
N
\
N
\
X
x
\
\
\
\
\_
X
\
\
\
\
\
---
"^
\
N
^
\
. .
— -
--
"^
\
x
140
~^-
~^
x
X
N
\
•
- —
-~-^
*'^,
\
''^
18
-18O -14O -1OO -6O -2O 20 6O 1OO 14O 180
X (m)
Side wlew of computed streamlines of flow passing from left
to Eri-g/S^t- oveur- ^ ^ ^nmTi^ \. ^ n. c Gaioss j_-air?t ^ii.il of aspect fatio 1.
-------
15O
140
1JO
120
110
too
eo
eo
70
6O
50
40
30
20
10
-4OO
FLOW TRAJECTORIES
Aapcct Ratio =• 2; fr = 1
\
-20O
200
4OO
FLOW TRAJECTORIES
A»p»ct Ratio » 2; Fr = -4
30 -
'
-
--
•*"
x
/
-4OO
^x
""
//
^/
/
S /
/ /
' /
/
/
/ /
/ y
/ /
/ /
/ /
/ / /
/ /
/
-200
/
x-~
/ s~^
S/*^
/
^"\
--^X
^s\
N
O
X
\
s N
\\
\\
\\\
\\
\
\
^^~
X^
\
N
\ '
\ X
\\
\ x
\
200
-.
-
•*•
^
K^
4O
FLOW TRAJECTORIES
A«p*ct Ratio - 2; Fr - 2
^
-^
•^
•^
^,
/
-4OO
^
-^
^/
•^
^
/
/ /
/ /
/ /
/
s
/
' /
s /
/ /
v/
/ //
//
/
-20O
^~-
. — —
/ \^-
L/'^^ *~
/s^
//^
/
•^^
^ X
~^\
^^\
•^x
x\
N
O
•>. (m)
x
\ \
\x
. N N
\\\
\\\
\\
\
NS»^_
^
"^^-^
X
\ "^
\x
\x
\ \
x
20O
.
^'
>-
^
s.
4C
FLOW TRAJECTORIES
Aspect Ratio *> 2: Fr " Infinity
13O -
00 -
BO -
7O -
60 -
5O -
4O -
30 -
20 -
10 -
O -
--
-""
^
S
/
-400
^
•*""^
^/
/
^/ _
/
/ /
/ /
' /
/
/ /
/ ;
' /
/ / .
/ /
' //
/ /
/
-200
/
/~^
/ '
s~ ^--»-*
s
^ /^"
/
X.
^~\
"^-\
-*^^^ ^
•~~^XN
"~^N
N
O
\
N \
s \
,\ v
,\\
\\
\\x
\\
\
2(
x^
"""•
x^
\
\^
\
\ x
\ \
\ x
x
)0
~~
~~
x
X
4C
figure 11. Side view of computed streamlines of flow passing from left
to right over a symmetric Gaussian hill of aspect ratio 2.
-------
FLOW TRAJECTORIES
2-O HIU; FT - 1
130 -
ao
5O
30
20
-
"
"
S
/
-4OO
^
^^
^/
s
s/
/
,/ /
/ /
' /
/
/
•/ /
/ /
/ /
' / /
/ / /
/ /
/
-200
•
^ .
/ ' ^^
/,
/x"
s
— ^ X
-^^\
~ — >N.
XX
\
0
K
\
s\
\\ v
\\x
\\\
\\\
\\
\
X^
v
^^^
\
\ '<^-
s\
s\ '
\ V
\
200
^
•^
\
40
FLOW TRAJECTORIES
2-D HIU; Fr - 4
13U -
N /O "
"
**
**"
X
X
-400
X
" /
X
/
s ,
/ /
/ /
' /
/
x
x
/ x
X /
f / /
/ / ,
/ /
' //
//
-200
--
X
x^
/-~~
-------
Calculated Deflections at x = 0
A»p«ct Ratio » (1.1)
(
100 -i
M_
i
•K
o
1KA .
1^ft .
~
*-..
•
X
t
t
A
X
L
^
,
v^
A
X
X
4
s
A
*A
"V
X
*
A
A
X
A
^
X
h
"V
».
X
4
X
A
*
X
"vj
h
^
1
i
i
X,
^
^
^
1
&
4
^_
^
^
i=
0.2 0.4 0,8 0.6 1 1.2 1.4 1.8
Y/(A>p*ct Rotlo*H1ll Height)
4 Fr™f x Fr»lnflnity
Streamer Positions Upwind
s
&
*
*ta
1.
0.4 0.8 0.8 1 1^
Y/(A«p«et Ratto*HtH Height)
1.4 1.8
Figure 13. End-on view of streamlines at x=0 over a symmetric Gaussian
hill of aspect ratio 1, and at x=-<» (i.e., far upwind of
the hill).
-------
Calculated Deflections at x = O
A*p«ct Ratio - (1.1)
ISO j
1 "Vi -
m'
100 H
?
o -
M
~~-J
— *.
X
4
^
&
*
A
X
^
L^
"X
^
&
*
x
<
s
A
A
^
»
a
X
A
X
A
L :
^
k
**.
^
»
X
i
X
4
X
N,
^
^
[
A
A
^
^
^
^
K
&
4
KL
"*«l
g
&
^
0.2 0.4 0.6 O.fl 1 1.2
V/(A«p«ct Rotto»H)ll H«ig
* Fr™f x Tr™ Infinity
1.4
1.8
1,8
Calculated Deflections at x=0
100 -
30 -
20 -
10 -
0-
Nn:
— «.
(
\
X A
SL ||
^fc^
fc
V
^>.
K*
™ Ji
X
x
kl
A
A»
^
X
.
X
*x
k
&
t
1 i
XT
A
X
s
A
A
SJ
3g
ft
X
^
X
•*^^
A
-z-
k,
X*
A
3
2
***.
X
X
«
>
ft
A
0.2
Oc4 0.8
A Fr-
0.8 1 1.2 1.4
1.8 1.8
H«lght)
Fr» Infinity '
Figure 14. Comparison of Lateral positions of streamlines in the plane
x=0 over symmetric Gaussian hills of aspect ratio 1 and 2.
58
-------
position of each streamline well upwind of the hill is shown in the
lower portion of Figure 13, and the position of each over the hill is
also marked in the plane x=0, where the x-axis lies along the flow and
the origin lies at the center of the hill. Froude numbers of 1 and
infinity are simulated for symmetric Gaussian hills of aspect ratio 1
and 2. The compression of streamlines in the vertical is most evident
for the streamlines that pass nearest the crest. The greatest lateral
spread in streamlines in seen in the first two streamer bands beyond
the band that travels directly over the crest. As expected, the
deformation in the flow is greatest near the top of the hill. In the
region of greatest lateral deformation, the scaled deformation
decreases with an increase in the aspect ratio of the hill. This is
especially pronounced for neutral flow (Froude number very large), but
not for highly stratified flow (Froude number equals one).
Results for hills that are axi-symmetric are shown in Figure 15.
In the upper panel, the cross-wind aspect ratio is 1, and the
along-wind aspect ratio is 2, so that the flow is along the longer
axis. In the lower panel, the hill is rotated 90 degrees so that the
flow is along the shorter axis. The streamline positions over the
crest lie approximately midway between those for axi-symmetric hills
of aspect ratio 1 and 2. Away from the crest, lateral deflections are
larger when the flow is along the shorter side of the hill, and the
effect of stratification on the streamlines is also greater.
Predictions of streamline position in the plane x=0 are compared
with streamline positions observed in laboratory simulations in Figure
16. The laboratory simulations (Snyder et al. 1986) were performed
with an axi-symmetric hill, described by a fourth order inverse
polynomial function, immersed in a towing tank. Streamline positions
were measured optically for a number of release positions upwind of
the hill, and for various degrees of fluid stratification. The
results displayed in Figure 16 include no stratification (neutral),
Froude number equal to 2, and Froude number equal to 1.
The shape of the hill used in CTDM is restricted to Gaussian
shapes, so that the hill used in the laboratory simulations is not
represented exactly in the model. The Gaussian shape chosen to
represent the inverse polynomial shape matches at the crest, and at
the height contour approximately one-half the height of the crest. In
the figure, the Gaussian shape is a heavy solid line, while the
polynomial shape is a heavy dashed line. As illustrated, the
polynomial hill exhibits a flatter top, and steeper sides. This
difference must be recognized when one compares the modeled streamline
positions with those observed, especially with regard to the height of
a streamline above the surface of the hill.
The upper panel in Figure 16 illustrates the correspondence
between the laboratory simulation and modeled results for streamlines
in the limit of neutral flow. Over the crest, there is a bias towards
underestimating streamline height above the surface. Away from the
crest, the streamline heights above the surface (not the absolute
heights) are in generally good agreement. The agreement in the
59
-------
Calculated Deflections at x = 0
Aspect Ratio = (1.2)
ISO -
1
i
100 •!
i
i
H— •
t
i
^
i
h
^
x
A
A
A.
X
X
X
%
<
,
i
)
N
f
&
A
\
^
v
s
A
X
A
4
X^
X
\
A
A
x
^
^
^
^-s.
2
X
4
%
--^,
x
if
$
&
*
"•^
0-2 0.4 0.6 0.8 1 1.2
Y/(A*p»et RatJo*f-HII Height)
& Fr™f x Fr» Infinity
1.4
1.6
1.8
Calculated Deflections at x = 0
A«p«ct Ratio m (2.1)
i^rt —
170
c
!
1
1
"ti~«,
"«»
> 0^
K
X
Y
^
s
0.4
4
L
\
A
J
1^,
X
'
,.
N^
x
X
"*
X
A
^
^
1
(
A
X
s
X
ft
AX
"
A
s,
4
-i
K
(•
X
A
A X
A:
^
A
4
A
^
ML
0.6 0.8 1 1^ 1.4
Y/(A*p«et R«rtto*HIII H«lqht)
Fn«1 x FrsslnfTnfty
it
X
X
^
X
1.
^
A
A
"
tla
6
x .
X !
X
X
X
»«w
A
A
*S3
1.8
2
Figure 15. Illustration of streamline positions in the plane x=0 for
an asymmetric hill of aspect ratio 1 and 2. The notation
(1,2) indicates that the longer side (2) is aligned with
the flow.
60
-------
Streamline Positions (x=O)
200
19O
ISO
17O
16O
ISO
140
130
120
110
100
SO
8O
70
6O
SO
40
3O
20
1O •
N«utral Flow
200
19O
100
170
1(*O
Frm2
Legend
-o.i
0.9
1.1
1.3
O— Q Tow Tank
Y/(A»p«ct Ratlo*H)
— Polynomial Hill (Tow Tank)
— — Gaussian Hill (CTDM)
Figure 16 Comparison of streamline positions in the plane x=0
observed in tow-tank simulations by Snyder et al. (1985)/
with those obtained from CTDM. Positions of streamlines
measured in the tow-tank are marked by squares, while
those modeled by CTDM are marked as triangles. Stream-
lines that originated at the same lateral position are
joined by either a thin line (tow-tank) or a thin dashed
line (CTDM). Note that the polynomial hill used in the
tow-tank (heavy solid line) differs from the Ganssian
hill used by CTDM (heavy dashed line).
61
-------
lateral position of the streamline is quite good. Perhaps the bias
towards greater compression of streamlines in the vertical over the
crest is caused by the difference in the degree of "flatness" between
the two hills. Near the crest, the polynomial hill used in laboratory
simulations, is broader than the Gaussian hill. A, broader hill
should, on the basis of the aspect ratios studied above, promote a
smaller degree of compression over the crest.
The center panel in Figure 16 illustrates the results for
moderately stratified flow with a Froude number of 2. Deformation of
streamlines in the vertical over the crest is modeled very well.
However, away from the crest, the streamlines are observed to deflect
more laterally than is predicted by the model. This is especially the
case for streamlines released well below the top of the hill. In
spite of the bias toward underestimating the lateral deflections, the
streamline heights above the surface are modeled fairly well.
The lack of good agreement in modeling the lateral deflections is
consistent with the use of the linearized theory in the limit of small
Froude number. Snyder et al. (1986) report results of computations
that make use of the Fast Fourier Transform (FFT) technique in solving
the linearized equations of Smith (1980).
The use of the FFT, although much more computer-intensive, allows
the actual hill shape to be incorporated, and it precludes the need
for many of the approximations adopted in developing the CTOU
algorithm. A comparison of the modeled lateral positions of the
lowest streamline in Figure 16 (center) is presented in Table 1. The
deflections obtained by Synder et al. are virtually matched by those
obtained from CTDM for streamlines that originate within about
one-half the lateral length scale of the hill (from the plane y=0).
Beyond this, out to streamlines that originate near y =
(aspect ratio) *H, the deflection obtained by CTDM becomes smaller
than that obtained by Snyder et al. This would appear to arise from
approximations contained in the CTDM algorithm for sampling points
away from the center of the hill.
The lower panel in Figure 16 illustrates the results for strongly
stratified flow with a Froude number of 1. The observed lateral
deflections near the surface of the hill become much larger than those
obtained from CTDM, and the compression of streamlines in the vertical
over the crest is stronger. Essentially, the linear theory
substantially underestimates the effects of strong stratification.
This is true of the FFT results also. In fact, because the FFT
solution was formulated as a forward-looking algorithm, streamlines
modeled by Snyder et al. actually pass through the surface of the hill
in the lee, even for streamlines whose initial height upwind of the
hill is equal to the height of the hill. The CTDM algorithm contains
no such pathologies, and can therefore be utilized for Froude numbers
as small as unity. The results of CTDM calculations for a Froude
number equal to 1.0 should therefore be viewed as representative of
significantly stratified flow, but not strongly stratified flow.
62
-------
TABLE 1
COMPARISON OF THE LATERAL POSITION* OF
STREAMLINES MODELED BY SNYDER ET AL. (1985)
USING FFT'S AND THE CTDM ALGORITHM
FOR A FRQUDE NUMBER OF 2
Lateral Position (y/[aspect ratio * H])
Initial Position (x= -
0
.192
.385
.577
.769
.962
-«) Observed (x=0)
0
.40
.77
.97
1.14
1.34
FFT (x=0)
0
.30
.56
.82
1.06
1.27
0
.28
.56
.80
.97
1.14
*The streamlines are sampled in the plane x=0, which
corresponds to a plane perpendicular to the flow and passes
through the center of the hill. All streamlines were
initially at an elevation equal to one-quarter the hight of
the hill. The aspect ratio of the hill is 2.6. Note that the
product of H * (aspect ratio) is just the length scale of the
hill at one-half the hill height.
63
-------
3.3.7 The WRAP Component
A particle in a steady two-dimensional flow around an obstacle
will experience both accelerations and decelerations as it passes by.
The magnitude of these changes in speed depends upon how close the
particle is to the stagnation streamline of the flow. Maximum changes
occur for particles on the stagnation streamline. Furthermore, the
spacing between adjacent streamlines varies in inverse proportion to
these changes in the speed along streamlines. Figure 17 is a
representation of a typical streamline pattern for flow around an
ellipse when the incident flow is at an angle to the axes of the
ellipse.
A plume in this steady flow (with some small-scale turbulence)
will follow the streamline patterns, spreading slowly across adjacent
streamlines. However, as streamlines spread apart (or contract) the
plume size in the horizontal will expand (or shrink) to the same
extent. In the absence of diffusion, these kinematic changes in the
horizontal size of the plume will not alter the concentration of
material within the plume. Changes to the horizontal scale of the
plume are balanced by changes in the flow speed so that the flux of
material is unchanged. With the addition of small-scale diffusion,
the rate of plume growth in the horizontal can be altered by changes
in streamline spacing (Hunt and Hulhearn, 1973). However, based on
the observations at CCB and Tracy we choose to ignore the effects of
small-scale diffusion on concentrations in the WRAP component of
CTDM. The observations suggest that low frequency
turbulence—meanders—control crosswind plume growth over hourly
averaging times.
To simulate ground-level concentrations due to dispersion of
releases below Hc in complex terrain settings, CTDM must approximate
the key features of steady two-dimensional flow around an ellipse that
were described above. Two key approximations in the WRAP component
are (1) lateral diffusion is insensitive to accelerations in the flow
(i.e., the kinematic deformation of the plume has no effect on the
diffusion rate), and (2) the mean flow for the averaging period (one
hour) is considered steady, while all of the variability in the flow
over the period, including that due to meandering, is considered
"turbulence."
A primary difference between WRAP and LIFT formulations arises
from the location of solid boundaries and the relationship between the
position of these boundaries and the wind direction fluctuations. The
terrain effect is modeled in WRAP by re-initializing the flow at the
distance s0 downwind of the source (see Figure 3). The concentration
at a receptor downwind of so is composed of concentrations from that
part of the concentration distribution at so that lies below Hc,
and that also lies on the same side of the stagnation streamline as
the receptor (see Figure 18). Reflection of plume material in the
vertical is allowed from the plane z = 0 over the entire distance s,
and reflection in the horizontal is also allowed from the hillside
beyond s0. Note that the stagnation streamline forms the boundary
of the hill surface in horizontal cross section.
-------
Stagnation Zone
Stagnation Streamline
Stagnation Zone
Figure 17. Typical streamline patterns in two-dimensional flow around
an elliptical cylinder.
65
-------
For a receptor located on the hillside at a distance s (see
Figure 18) and a height ZR above the plane z = 0, the concentration
due to one elemental point source located at (so,y,z) in the plume
is given by
dC(s,0,zR;so)
-.
+ e z 3
Equation 81 assumes that the x-axis of the coordinate system points
along the stagnation streamline, and that the source is located at
(xg,ys,zs). The total concentration at the receptor contains
contributions from those elements below Hc and on the same side of
the stagnation streamline as ys:
H (») C(s ,y,z) -O.
C(s,0,zR;so) = ; c / ° 2e
0(0) y z
- [e z + e z ] dydz (82)
where
C(Vy>z) * 2wuoQ o S "ys [e "Z°
(83)
These expressions are analogous to Equations 38 through 40 of the LIFT
component. The integral for dy has the limits (0) and (»). This is
meant to denote integrating from 0 to -H» if the receptor lies on the
"positive" side of the stagnation streamline, and integrate from -»
to 0 if the receptor lies on the "negative" side. Note that material
is allowed to diffuse upwards through Hc so that receptors which lie
above Hc, and which are downwind of so, can receive a contribution
from the elemental point-sources below Hc at so. The total
concentration at such receptors is the sum of both the LIFT and WRAP
contributions.
66
-------
Stagnation Streamline
Figure 18. Top view of a plume in two-dimensional flow around a hill.
The shape of the hill in eroasection ai the receptor height,
is assumed to be invariant with height so that the
deformation of the entire plume around the hill is a function
of the receptor location.
-------
The integrals in Equation 82 are evaluated (using the same
methods as those employed in the LIFT section) to obtain:
y z
--
e y (1 + sisn(yR} ERF (
Zs~ ZH 2
2
(84)
Most of the notation here has already been encountered in Section
3.3.3. The factor sign(yR) denotes the sign of the receptor
position in the coordinate system with x-axis aligned with the flow,
and it results from the choice of integrating over the "positive" or
"negative" portion of the flow in Equation 82. The factors B^ and
82 are given by
b -b -b Wb3
B m ERF ( * * °) + ERF ( . )
1 bo bo
bl"b2+b3 bl+-b2~b3
B, = ERF ( x ) -I- ERF ( . ) (85)
2 bo bo
where
bO a /2 «z °Zo V
b. » H o2
1 c z
2
b_ =» z_ o
2 R zo
*2
b3 - Zs tfz '
The subscript R denotes the receptor location, and the subscript s
denotes the source location (see Figure 19). The distance from the
stagnation streamline associated with the mean wind direction to the
centerline of the plume is denoted as d, the total sigma-y (for
horizontal spread of the mean plume) as
-------
Stagnation
Streamline
f
y
Stagnation Point
Figure 19. Sketch of the flow around an ideal cylinder of elliptical
cross-section. The section shown is taken either at the
elevation of the centerline of the plume (zs), or at Hc,
depending on which is smaller, and it indicates the
relationship between the streamline through the source
(¥s), the stagnation streamline (¥o=0), and the
coordinate system with x-axis aligned with the mean wind
direction.
69
-------
.222
a * =o - c
z z zo
o *2 = a 2 - c 2. (86)
y y yo
The ellipse that is used to estimate the flow below Hc is taken
from the horizontal cross-section of the hill at the minimum of the
following two elevations: either the elevation of the eenterline of
the plume, or the elevation of Hc. In this way the shape of the
hill selected is associated with the peak concentration of plume
material found within the layer of fluid below He.
Concentrations at receptors located below Hc just upwind of the
stagnation point are estimated as if the receptor sits on a pole of
height equal to the receptor elevation above the base of the stack.
Furthermore, the lateral distance between the plume eenterline and the
receptor is set equal to d, so that concentrations at all of these
receptors are controlled by the amount of material on the stagnation
streamline (see Equation 102 for d) . In this way plume material below
Hc follows streamlines around the hill, and only material which
diffuses onto the stagnation streamline impinges on the hill. The
equation for estimating these concentrations is:
y z
exp(-0.5(d/o )) [exp(-0.5
y
RS 2
+ exp(-0.5 <-*— *)')!. (87)
a
z
Equations 87 and 84 provide estimates of ground-level
concentrations of plume material before and after plume material above
Hg is "removed," respectively. That is, the upper and lower
portions of the flow do not become distinct in GTDM until the
impingement or stagnation point is reached.
3.3.8 Model for Streamlines
The central features of WRAP are the distance between the
stagnation streamline and the streamline that passes through the
source and the relative locations of the source, the receptor, and the
point of impingement. These require a flow model to obtain
streamlines. Because the flow is two-dimensional in this strongly
stratified limit, potential flow solutions are used to obtain the
streamlines.
The hill below Hc is represented as a cylinder of elliptical
cross-section, set on end. This shape is chosen because it contains
the overall scale and orientation of the hill, and it is simple enough
that streamline patterns can be expressed analytically. As already
70
-------
stated, the ellipse used for the entire flow below Hc is the result
of fitting an ellipse to the height-contour that corresponds to the
minimum of Hc and the plume height. The potential flow solution is
expressed in elliptical coordinates following the notation of
Batchelor (1970).
Let the x-axis be aligned with the major axis of ellipse, let a
be the length of the semi-major axis, and b be the length of the
semi-minor axis. Then the elliptical coordinates (y,v) are
related to the cartesian coordinates by the relations
x2 = (a2-b2) cosh2(y+y ) eos2(v)
o
y2 = (a2-b2) sinh2(y+y ) sin2(v). (88)
o
Note that y = y'-yo, where yo is the value of y' along the boundary of
the ellipse (y is constant along ellipses in the family of confocal
ellipses of which the boundary of the hill is a member). Using the
elliptical coordinates, a streamline in the flow is given by
4r » - S
-------
y/j
Impingement Point
ft
Figure 20. Definition of modeling variables, illustrating in particular
the coordinate system in which the xg-axis is aligned with
the tangent to the stagnation streamline at the impingement
point (the 6-coordinate system). The coordinates along the
xg-axis of the source are denoted by XSQ, xofl, and
xrg, respectively.
72
-------
Similarly, the slope of the tangent to the streamline is expressed in
terms of the x-y coordinate system by forming the quantity dy/dx by
differentiating Equation 88, and evaluating this at the position of
the tower:
. (b tanh(jj_) + a) tan(v_) (dy/dv) | + (b + a tanh(n_))
HZ I -- 1 - 1 - 1 - 1 - ,
dx 'T~ (a tanh(jiT) + b) (dv/dv)|T - (a + b tanh(nT)) tan (VT>
Noting that tan ($7) =» ( dy/dx) |T and setting a/b = r,
Equation 92 can now be solved for tan(ow) , so that
- tan (a ) =
[tan(vT)/cosh2(vT)+tan(<|>T)tanh(vT)/cos2(vT)]+r tan(<|»T) [tan2(
(tan2(vT)tanh2(vT)-H)+r[tanh(wT)/cos2(vT)+tan( s S W
73
-------
The stagnation point can also be calculated. Along the stagnation
streamline, t|»o=0, so that Equation 94 becomes
0 = -S (a+b) sinh(y) sin(v+« ) . (98)
co W
Because this must be satisfied for all y, v must be equal to
-ow along the stagnation streamline. Therefore, the stagnation
point is at (0, -<*«), because u equals zero on the boundary of
the ellipse.
Distance is tracked along the xg-axis, which is parallel to the
stagnation streamline at the stagnation point. This coordinate system
is needed to provide a convenient Cartesian coordinate system that
allows the streamline through the source to be a single-valued
function of x for all o^. At the stagnation point, the stagnation
streamline meets the boundary of the ellipse at ah angle of 90°. The
tangent to an ellipse is given by
tan (Y) = I = -() x/y. (99)
3X ellipse S
The coordinates of the impingement point, (x^.y^) are given by
Equation 93 with (ji,v) = (0, - <*w) • so that
x^ = a cos(ow)
(100)
Because the stagnation streamline is perpendicular to the tangent to
the ellipse at (x^, y^), the tangent of the rotation angle, B,
must be -l/tan(f) .
tan (B) = - J tan(aw) . (101)
The distance between the streamline through the source (v|/s)
and the stagnation streamline dl»o=0) far from the hill is related
to the value of HTS and the wind speed at infinity, S«,. Because
the speed of the flow equals the gradient of the stream function far
from the hill, S^ » («jrs - Hro)/d, or
d = *g/Sa,. (102)
However, because ov may be measured closer to the hill, the speed
at the source is substituted for S» to estimate d near the source.
3.3.9 Receptors not Influenced by Hills
In theory, the main subroutines of CTDM give results which are
identical to flat-terrain results in the limit that the hill height
goes to zero. However, the code in the model is not designed to check
74
-------
for hill height of zero before executing statements that may require
division by the hill height; and if it were, many extensive
calculations would be needlessly executed, with the model returning no
terrain-effects. Furthermore, the structure of the model requires
that receptors be associated with specific hills wherein each hill
requires extensive information. A flat terrain algorithm is included
in the model to avoid such numerical problems and extra input
requirements.
The flat terrain algorithm simply performs a Gaussian plume
computation which assumes that there is no mixing lid, that all
receptors lie on a single ground-plane, and that plumes travel in
straight lines. All plume rise and growth algorithms match those used
in the other sections of the model. For a plume at height zs above
the surface, released from y = ys, the concentration of plume
material at a receptor placed at a height hs above the surface, and
set at y = yR, is given for a time of travel t from the virtual
source (including a virtual time) by
o
s 2
}.. (103)
Note that the time of travel appears in the expressions for oy and
oz (Equation 29) .
75
-------
SECTION 4
GTDM EVALUATION ANALYSIS
An assessment of the performance of CTDM is described in this
section. CTDM and two other complex terrain models used for rural
applications, COMPLEX I and RTDM (Paine and Egan, 1987), were
evaluated at five sites. These sites included the three CTMD sites
(CCB, HBR, FSPS) as well as two other sites with conventional S02
data obtained over a one-year period (the Westvaco Lake paper mill and
the Widows Creek steam generating station). The models that were
evaluated and the data bases used are discussed further in Sections
4.1 and 4.2, respectively.
A large part of the CTMD data base (CCB, HBR, FSPS) was used in
the development of CTDM, so the overall results for CTDM at these
three sites do not represent an unbiased test of CTDM versus COMPLEX I
and RTDM. The CCB data base for both SF6 and CF3Br was used
extensively in the development of CTDM. Model development use of the
HBR data base concentrated on only a subset of the CF3&r data (35
hours; see Strimaitis et al., 1985) and none of the SFg data (for
which lidar-derived plume heights were not obtained until the
evaluation task). The use of the Tracy Power Plant data for model
development was focused upon hours involving plume travel toward
Beacon Hill and Target Mountain; hours involving impacts on samplers
on other hills were not used. In addition, the model development used
meteorological data from the 150-m tower only, while the model
evaluation data base also included tethersonde and doppler sodar
data. Withheld data that are available for future model evaluations
include 20 hours from HBR (see Lavery et al., 1983, page 184) and all
of the data from the preliminary Tracy experiment in the fall of 1983.
A series of statistical tests (described in Section 4.3) were run
on the models for data sets both paired or unpaired in time and/or
space. These tests examined the models' overprediction or
underprediction bias as well as the root-mean-square (RMS) error, and
the percentage of predictions within a factor of two of observations.
Results are described in Section 4.4.
Another aspect of the evaluation analysis involved an examination
of the spatial distribution and magnitude of CTDM concentrations for
each hour at the three CTDM tracer sites. CTDM performance in the
LIFT and WRAP components was assessed by examining the behavior of
hourly patterns of predicted and observed concentrations. Results are
given in Section 4.5.
76
-------
4.1 Models Evaluated
At the CTMD tracer sites, CTDM was evaluated using the plume
height set to the tracer release height for near-neutrally buoyant
releases. In some cases with buoyant tracer plumes (for SF6 at HBR
and FSPS), CTDM was run using a final plume height based upon lidar
observations of the plume. At FSPS, CTDM runs for SF6 releases were
done for both observed and calculated plume heights. The "observed"
plume height at the Tracy Power Plant was obtained by examining the
lidar observations. The first two lidar cross-sections through the
plume were both used independently to define two final observed plume
height. The first cross-section occurred at a downwind distance of
roughly 500 meters, while the second typically occurred at a distance
between 1000 and 1500 meters. The resulting two alternative observed
plume heights were tested using CTDM (results are shown in
Section 4.4). The calculated plume heights at FSPS were derived from
conventional stack gas exit parameters (temperature, velocity, stack
diameter). For the conventional S0£ data bases, plume heights were
calculated for each hour.
Other rural complex terrain models that were evaluated included
COMPLEX I and RTDM, which was run in both "default" and "on-site"
modes. RTDM in default mode does not employ vertical and horizontal
turbulence intensity information nor vertical temperature gradient
data. The on-site mode does use these meteorological variables, and
in so doing, is similar to CTDM in taking advantage of these available
measurements. COMPLEX I and RTDM were run using observed plume
heights where applicable, except that only calculated plume heights
were evaluated at FSPS.
Tables 2, 3, and 4 list several important features of these
models, including plume transport, plume dispersion and stability
determination, plume rise and terrain impingement, and limits to
vertical mixing. Of the three models, COMPLEX I is the least refined
because it requires no terrain profile or shape information and does
not consider a critical dividing streamline height. The number of
meteorological input variables available to COMPLEX I and to RTDM in
default mode is limited; no hourly temperature gradient or OQ
and
-------
TABLE 2
FEATURES OF COMPLEX TERRAIN MODELS
USED IN THE EVALUATION: CTDM
Plume Transport And Dilution
• Wind speed and direction interpolated from multiple
levels of input data or scaled from the highest
available level to final plume height.
• Plume deflection around (or over) terrain obstacles
accounted for.
• Plume height wind speed used in Gaussian equation.
Plume Dispersion/Stability
• Meteorological preprocessor provides Monin-Obukhov
length, a continuous stability parameter.
• Plume a , o determined from measured
OA (or o ), o data and estimated
Lagrangian time scales.
• Off-centerline (rather than sector averaging) used in
concentration calculations.
• Buoyancy-enhanced vertical and horizontal dispersion.
• Model does not calculate concentrations for unstable
conditions.
Plume Rise/Terrain Impaction
• Plume lifting (terrain adjustment) over terrain
varies hourly as a function of distance to hill, hill
shape, and critical dividing streamline height.
• Ho minimum terrain approach; direct plume impingement
is possible.
• Briggs final plume rise used; meteorology for plume
rise calculations obtained halfway between stack top
and final plume height.
78
-------
TABLE 2 (Continued)
• Measured values of dO/dz used for stable rise.
Limits to Vertical Mixing
• Full reflection at the ground, unlimited growth above
• Local internal boundary layer used in vicinity of
hills.
• No mixing lid restriction used in stable conditions,
but mixing height governs profiles of wind and
turbulence within the surface layer.
• Nocturnal lid determined from diagnostic boundary
layer depth formula.
Terrain Depiction
• Digitized terrain contours as read by terrain
preprocessor.
• Mathematical description of each hill is used by CTDM.
79
-------
TABLE 3
FEATURES OF COMPLEX TERRAIN MODELS
USED IN THE EVALUATION: RTDM
Plume Transport and Dilution
• Wind speed scaled to stack-top height for plume rise,
to final plume height for use in Gaussian equation.
• Wind direction as input, not scaled with height.
• No plume deflection around hills (in the horizontal).
Plume Dispersion/Stability
• Discrete Pasquill-Gifford stability categories (A-F).
• In default mode, plume a growth determined from
Briggs (1973) rural dispersion coefficients.
• In on-site mode, plume o growth determined by
observed
-------
TABLE 3 (Continued)
Limits to Vertical Mixing
• Full reflection at ground and mixing lid, but
unlimited mixing height for stable conditions.
Terrain Depiction
• Terrain profiles (elevation versus distance from
source) specified in 10° angular intervals;
appropriate profile is chosen each hour based upon
wind direction.
81
-------
TABLE 4
FEATURES OF COMPLEX TERRAIN MODELS
USED IN THE EVALUATION: COMPLEX I
Plume Transport and Dilution
« Wind speed scaled to stack-top height for plume rise.
• Wind direction as input, not scaled with height.
• No plume deflection around hills (in the horizontal).
• Stack-top height wind speed used in Gaussian equation.
Plume Dispersion/Stability
• Discrete Pasquill-Gifford (P-G) stability categories
(A-F).
• Plume a , a growth using P-G dispersion
coefficients (rural).
• 22.5° horizontal sector averaging for all stability
classes.
• Buoyancy-enhanced vertical and horizontal dispersion.
Plume Rise/Terrain Impaction
• Terrain adjustments = .5 for stabilities A-D, 0 for E
and F.
• Closest plume centerline approach to terrain = 10
meters.
• Briggs final rise used, calculated from stack-top
meteorology.
• Stability-dependent values of de/dz used for stable
rise.
• Critical dividing streamline height is not accounted
for.
82
-------
TABLE 4 (Continued)
Limits to Vertical Mixing
• Full or partial reflection at ground and mixing lid,
but unlimited mixing height for stable conditions.
Terrain Depiction
• Receptor heights given only; no hill shape or terrain
profile information used.
83
-------
4.2 Data Sets Used for Evaluation
CTDM, COMPLEX I, and RTDM were evaluated with data from the three
CTMD sites: Cinder Cone Butte (CCB), the Hogback Ridge (HBR), and the
Tracy Power Plant (FSPS). For these model runs, the emissions,
meteorological and receptor data were reformatted to be compatible
with the operational version of CTDM. A special provision was made to
allow these models to read observed plume height data, which varied
hourly at the Hogback Ridge and at Tracy.
To supplement the CTMD sites, two other data bases were selected
for the model evaluation: the Westvaeo pulp and paper mill in Luke,
Maryland (Wackter and Londergan, 1984) and the Widows Creek Steam
Plant (TVA) in northeastern Alabama (Egan et al., 1985).
At Cinder Cone Butte, an isolated hill, SF6 and CF38r
releases were modeled. There were 107 hours of SFg releases and 51
hours of CF3Br releases modeled at 93 receptors (See Figure 21).
The tracers were released from cranes, which were repositioned several
times during the experiment. Each new release position and/or height
was considered as a separate source. Therefore, there were 55 SF6
sources and 22 CF3Br sources modeled in this evaluation.
The Hogback Ridge (Figure 22) was viewed as one long hill for the
CTDM evaluation. A total of 99 hours were modeled at 106 receptors
from 36 release heights/locations for SFg and CF3Br.
At the Tracy Power Plant, the terrain was divided into 18
individual hills (Figure 23) for the GTDM evaluations. SF6 was
released from the Tracy stack and CF3Br was released from 3 levels
of the 150-m tower during the experiment. A total of 128 hours were
modeled at 110 receptors. Each receptor was identified as being on
one of the 18 hills.
Three sets of meteorological inputs were prepared for each site:
a preferred data set using all available data and two versions of
degraded meteorological data. The degraded meteorological data
consisted of either only one or two levels of wind, temperature, and
turbulence data. At GCB, 8 levels of data were used in the profile
(2, 10, 40, 60, 80, 100, 150 m), plus a level obtained from the
Modelers Data Archives (MDA's) for the meteorological data at plume
height for each hour. The first degraded data set at CCB consisted of
10-m and 80-m tower data and the second degradation contained only
10-m tower data.
The preferred meteorological data set at HBR consisted of 10
levels from the 150-m tower (2, 5, 10, 30, 40, 60, 80, 100, 150 m).
The first degraded data set had only the 10-m and 100-m levels. The
second degradation used only the 10-m tower data.
At FSPS, 16 levels of meteorological data were contained in the
preferred data set. The composite profile was obtained from the 150-m
tower, the sodar and the tethersonde. Wind data was extended above
the top of the 150-m tower in 25-m increments up to about 400 m with
84
-------
00
Ul
Fixed
Northwest Flow* Only
Southeast Flows Only
. EPA COMPLEX
TERRAIN PROGRAM
-CINDER CONE BUTTE-
Figure 21. Tracer gas sampler locations on Cinder Cone Built;.
-------
r-Walker
\Mire
LEGEND
A 500' Tower A
A Towor 8
O Tower C
1 Tracer Release Pt. No. R-30
2 Tracer Release Pt. No. 203
• 3 Tracer Release Pt. No. 215
» 4 Tracer Release Pt. No. 216
5 Tracer Release Pt. No. 111
Figure 22. Field experiment layout in the vicinity, of the Hogback Ridge.
36
-------
CO
1300m
Contour Interval 200
Figure 23. Terrain features surrounding the Tracy Power Plant as
modeled by CTDM.
-------
the use of sodar data. Inconsistencies between the tower and sodar
data at 150 m were smoothed out between the levels of 100 and 200
meters. The difference for any hour was handled by linearly
interpolating between the tower level at 100 m and the sodar level at
200 m. Temperatures were extended above 150 m through the use of the
tethersonde temperature gradient above 150 m. In the input data
degradation tests, only the tower data was used. The bottom (10-m)
and top (150-m) levels were used in the first degraded data set and
the second degradation consisted only of the bottom level.
Four hills were identified for modeling with CTDM in the vicinity
of Westvaco's Luke Hill (Figure 24). A single stack was modeled for
11 receptors for a one-year period (December 1980-November 1981).
CTDM was run sequentially for the entire year, but did not calculate
concentrations when the plume was in an unstable layer (Monin-Obukhov
length is negative and the plume is within the mixed layer).
The meteorological profile was obtained from the 30-m Luke Tower
(near site #2, north of the stack) and also from the 100-tn
meteorological tower on the hill to the southeast (at site #1). The
use of a temperature gradient obtained between the tops of these two
towers was expected to result in better CTDM performance than the use
of a single tower with the lower temperature measurement from 10
meters. (A similar result had previously been obtained independently
for RTDM at Widows Creek, where balloon soundings verified this
finding (Egan et al. 1985)). The "full" profile of meteorological
data involved 5 levels: 190, 210, 326, 366, and 416 meters above the
base of the 190-m stack. The first degrade involved only the 190- and
210-m levels (Luke Tower 10- and 30-m levels) and the second degrade
used only the 190-m level.
A total of five hills were used in the CTDM modeling of the Widow
Creek Steam Plant (Figure 25). Three stacks were modeled for 14
receptors for calendar year 1980. The meteorological data profile was
assembled from two 61-m towers, one in the Tennessee River valley and
one on the top of Sand Mountain (neas- Station 3). The full
meteorological profile included data at 10, 61, and 312 meters above
stack base (all of the valley tower and the top of the mountain
tower). One data degrade employed only the valley tower (10- and 61-m
levels) while a second degrade used the 10-m level of the mountain
tower (261 m above stack base).
For both the Westvaco Lake and Widows Creek data base, the
background concentration was determined for each hour as the lowest
monitored concentration in the network. The hourly background
concentrations were subtracted from the total monitored concentrations
to obtain a residual concentration attributable to the source in
question. The use of the lowest monitored observation always resulted
in a non-negative concentration from the source in question.
4.3 Statistical Tests and Case-Study Analyses
Hourly concentration predictions were obtained for each of the
models described in Section 4.1. Although RTDM and COMPLEX I
88
-------
\
Mile
1 Kilometer
Figure 24. Terrain features and monitors in the vicinity of the
Westvaco Luke Mill.
-------
Widows Creek
Steam Plant
442m
467
475m
467m
479m
479m
479m
463m
544m
218m
478m
209m
196m
463m
191 m
4.0km
3.Skm
8.8km
3.Skm
3.2km
3.2km
10.8km
8.7km
30.4km
7.4km
10.5 km
4.3km
1.8 fern
7.3km
Figure 25. 1980 Widows Creek monitoring network with outlines of
terrain features used in CTDM.
90
-------
calculate concentrations for every hour, CTDM does not give
predictions for unstable hours if the plume is within the surface
layer. Therefore, it was necessary to restrict the evaluation results
to only those hours for which CTDM results were available. In
addition, comparisons had to be restricted to only those sampler sites
that were operational for any given hour. Analyses were performed for
1-hour averages for the tracer experiments (CCB, HBR, FSPS) and for
1-hour and 3-hour averages for the conventional S02 data bases
(Westvaco Luke and Widows Creek).
Specific statistical tests and data subsets to which they were
applied are listed in Table 5. For tracer data such as the CTMD
sites, primary tests involve data sets paired in time or paired in
time and space. The excellent spatial resolution in these tracer
experiments maximizes the likelihood that the peak observed and
predicted concentrations will be found somewhere in the network. The
occasional movement of the sources and the monitoring network,
especially for CCB and HBR, causes a paired-in-space test to be less
important for some of the tracer data. The data set unpaired in time
and space is similarly difficult to interpret and sometimes misleading
for these tracer data bases because the sources and network are not
fixed in location. For these reasons, results for the test unpaired
in time and space are not presented for the CCB or HBR data sets.
For the conventional data bases, tests that are unpaired in time
are important because of the long time record but relatively poor
spatial resolution. The number of S(>2 monitors is, however, large
enough to allow testing with data sets paired in time and paired in
time and space.
The model bias test involves the difference of the averages of
the hourly peak model predictions and observations. For each of the
five data sets analyzed, the highest concentration only was used to
compile this statistic. An additional data base employed at the
tracer sites consisted of the average of the highest 5 predictions and
observations each hour. This data base produced results less
dependent upon extreme concentrations and was feasible because of the
high spatial resolution of the tracer samplers. The computed model
bias is the average observation minus the average prediction, so a
negative bias denotes a model overprediction.
The root-mean-square (RMS) error and normalized mean square error
(NMSE) statistics show a measure of model scatter and also incorporate
the model bias. The mean square error is represented as (Co -
where Co is an observed concentration, Cp is a predicted
concentration and an overbar denotes an average over all samples.
While the RMS error is simply the square root of the mean square
error, the NMSE error is represented here in two forms:
(or Ml) = (Co - Cp)2/C0 2
NMSE2 (or M2) = (Co - Cp)2/[C0Cp]
-------
TABLE 5
DATA SUBSETS AND EVALUATION TESTS FOR TRACER (CTMD)
AND CONVENTIONAL DATA BASES
1) Unpaired in Time and Space
(Model bias only for measures listed below)
a) Highest concentration
b) Second-highest
c) Highest Second-Highest
d) Average of Top N values*
2) Paired in Time, Not in Space (peak hourly values)
a) Model bias
b) Root-mean-square (RMS) error
c) Normalized mean square error (NMSE)
d) % of predicted peaks within
a factor of 2 of observations
3) Paired in Space (peak values and Top 10 values at each
monitor)
a) Model bias
b) RMS error
c) NMS error
d) % within factor of 2
4) Paired in Time and Space (each site, each hour)
a) Model bias
b) RMS error
c) NMS error
d) % within factor of 2
* N = 5 for tracer experiments, 10 for conventional data bases.
92
-------
The first of these formulations is desirable because of its simplicity
(no use of the average predicted concentration), while the second one
results in equal treatment for under- and overpredicting models.
The percentage of cases with predictions within a factor of 2 of
observations is listed to summarize the results of a scatter plot of
predictions versus observations. This statistic is also a
rule-of-thumb indicator as to how accurate the model being evaluated
is on a case-by-case basis.
The bias, RMS error, NMSE, and % within a factor of 2 statistics
were also prepared for concentrations paired in time for several
classes of meteorological conditions. These conditions included
stability classes D, E, and F (based upon near-surface conditions) and
3 categories of release-height wind speeds: 0-1, 1-3, and >3 meters
per second.
In addition to the tabulated results described above, scatter
plots of predicted versus observed 1-hour and 3-hour average
concentrations were prepared for peak concentrations paired in time.
The results of the statistical tables and scatter plots for the five
evaluation sites are discussed in Section 4.4.
Case-study analyses involving plots of predicted (CTDM only) and
observed concentrations were also performed at the tracer sites. The
concentrations were plotted on pairs of maps for each hour: one for
predicted and one for observed concentrations (See Figure 26a and
26b). The plots include a listing of plume height, Hc and Froude
number values as well as the location of the source and the positions
of the highest predicted and observed concentrations. Results of the
analyses of these maps are discussed in Section 4.5.
4.4 Results of Statistical Evaluation (All Models)
Excerpts of the results of the statistical evaluation of CTDM,
COMPLEX I, and RTDM are presented in this section, including results
for data sets unpaired in time and space, and paired in time but not
in space. Additional results are listed elsewhere:
• complete statistics for observed and predicted
concentrations paired in time, but not in space - Appendix B;
• statistics for concentrations paired in space, not in time
(conventional data bases only) - Appendix C;
• statistics for concentrations paired in time and space -
Appendix D;
• statistics based upon meteorological category for
concentrations paired in time, not in space - Appendix E.
• scatter plots of peak hourly predictions and observations
and residual plots predicted/observed ratios versus distance
(Appendix F).
93
-------
En »NO P9EDICTEO 3F6 CO^CEMTR«T1113 (gS/-«««3J 0*
JULI»M O'Y *10 HOURl 299S »
M»P OF PREDICTED CONCENTH«riONS» (MJI.TIPLT «[SPL*rEO V»LUE3 9* 1.* TO CtT-THS »CTU»I.
I
100 m
II
V»I.U£SJ
neloase crane locations usod
dining Ihe er.peiimonn
A Position ol Hie 150m iov.-er
»IMO OIBt
WI'lO SPIEP
SISH»»« (H/SJt
VPT6
53.
l.Oi
.7*
.28
.8047
Figure 26a. Sample map showing the distribution of CTDM predictions at
tracer sample locations for one of the Hogback Ridge
experiments; plume height variables are listed in an inset
on the mat).
-------
OF OMSFRVCf) CO*CEWTR»TION8t
DTS»(.»»Er> V4LUE3 8V 1.0 TO SET THf 4CTUAL CONCEMT»»Ttnm )
6
neloase crana locations usod
dining Hit erpeiimenn
Position ol ihc 150in lov.-ei
089 » 37. »S ,
tVC**eE 0^ TOP S CONCENTR«TION3l 089 •
8 • (089 -
, P*C • 39.24
(DBS - PRE) s -4.580
Figure 26b.
Sample map showing observed tracer concentrations
corresponding to the CTDM predictions in Figure 26a;
summary case statistics are also included.
95
-------
Tables 6 through 13 contain the following results:
Unpaired in Paired in
Time and Space Time Only
CCB (SF6) Table 9a
CCB (CF3Br) Table 9b
HBR (SF6) • Table lOa
HER (CF3Br) Table lOb
FSPS (SF6) Table 6a Table lla
FSPS (CF3Br) Table 6b Table lib
Westvaco (S02) Table 7a,b Table 12a,b
Widows Creek (S02) Table 8a,b Table 13a,b
For each evaluation site, CTDM results for all available tower
levels (see Section 4.2), just two levels, and a single level of
meteorological input data are listed. COMPLEX I and RTDM were run
with a single level of meteorology close to stack top (See Section
4.2).
At the Tracy Power Plant, CTDM with all tower levels showed an
underproduction tendency in the unpaired extreme value tests (Table 6,
a and b), while use of fewer levels resulted in higher
concentrations. COMPLEX I overpredicted for both tracers and RTDM
(on-site mode) underpredicted for both tracers. RTDM in default mode
exhibited a modest overpredietion tendency.
Table 7 (a and b) shows 1-hour and 3-hour average evaluation
results for the unpaired data sets-at Westvaco. CTDM shows an
overprediction tendency for the three combinations of meteorological
input data. Use of two tower levels gives poor results - the d6/dz
values are probably too large because one temperature level is too
close to the ground. COMPLEX I shows severe overprediction problems,
while RTDM exhibits a moderate overprediction tendency. The on-site
run of RTDM gives better results than the default mode run.
As shown in Table 8 (a and b), all models overpredict at Widows
Creek. Once again, the CTDM run with two tower levels shows higher
concentrations than the other CTDM runs, again probably due to the
high de/dz values. COMPLEX I shows the highest overpredictions,
while RTDM (default) moderately overpredicts and RTDM (on-site)
slightly overpredicts.
The CCB evaluation results for the paired in time, unpaired in
space data set (Table 9, a and b) show a significant decline in CTDM
performance toward underprediction for both SF$ and CF3Br as the
meteorological input is degraded. CTDM results for the full
meteorological data set are quite good, however. COMPLEX I
overpredicts foe both tracers, but by less than a factor of 2. RTDM
in on-site mode shows a modest underprediction tendency, while in
default mode it overpredicts for SF$ and underpredicts for CF3Br.
96
-------
TABLE 6a
SUMMARY STATISTICS FOR DATA UNPAIRED IN TIME AND SPACE
SITE: TRACY POWER PLANT TRACER: SF6 UNITS: uS/m**3
1-HOUR AVERAGES, 111 HOURS
DATA SUBSET
HIGHEST
SECOND-
HIGHEST
HIGHEST
SECOND-
HIGHEST
AVG OF
TOP 5
OBSERVED
10.227
8.241
6.452
8.092
CTDM, SEVERAL
TOWER LEVELS 7.551
CTDM, SEVERAL
TOWER LEVELS 6.581
(ALT. PLUME HT 1)*
CTDM, SEVERAL
TOWER LEVELS 6.725
(ALT. PLUME HT 2)*
CTDM, TWO
TOWER LEVELS 8.067
CTDM, ONE
TOWER LEVEL 11.755
6.314
6.314
6.661
6.617
11.679
5.356
5.356
5.356
3.970
11.194
6.100
5.648
5.677
6.374
10.014
COMPLEX I
22.605
22.587
22.587
21.686
RTDM, DEFAULT
MODE 12.944
RTDM, FULL
ONSITE MODE 6.180
11.407
3.966
11.407
2.832
11.097
3.877
* Alternative plume height #1 was obtained from lidar measurements
at the first cross section downwind from the source. Plume
height #2 was obtained from the second lidar cross section.
97
-------
TABLE 6b
SUMMARY STATISTICS FOR DATA UNPAIRED IN TIME AND SPACE
SITE: TRACY POWER PLANT TRACER: CF3Br UNITS: uS/m**3
1-HOUR AVERAGES, 111 HOURS
DATA SUBSET
OBSERVED
COMPLEX I
RTDM, DEFAULT
MODE
HIGHEST
19,463
CTDM, SEVERAL
TOWER LEVELS 8.875
CTDM, TWO
TOWER LEVELS 31.598
CTDM, ONE
TOWER LEVEL 36.402
31.715
21.994
RTDM, FULL
ONSITE MODE 7.917
SECOND-
HIGHEST
13.351
8.433
16.630
23.554
30.304.
20.943
6.502
HIGHEST
SECOND-
HIGHEST
7.482
8.433
9.530
11,641
27.107
20.943
5.560
AVG OF
TOP 5
12.900
7.960
15.889
17.320
27.312
19.339
5.834
98
-------
TABLE 7a
SUMMARY STATISTICS FOR DATA UNPAIRED IN TIME AND SPACE
SITE: WESTVACO LUKE TRACER: SO2
1-HOUR AVERAGES
UNITS: uS/m**3
DATA SUBSET
HIGHEST
SECOND-
HIGHEST
HIGHEST
SECOND-
HIGHEST
AVG OF
TOP -10
OBSERVED
7.227
6.911
5.755
5.725
CTDM, SEVERAL
TOWER LEVELS 13.593
CTDM, TWO
TOWER LEVELS 24.838
CTDM, ONE
TOWER LEVEL 9.641
11.076
24.308
9.556
11.076
23.214
8.353
9.356
22.325
8.515
COMPLEX I
50.647
49.478
49.382
47.503
RTDM, DEFAULT
MODE 15.369
RTDM, FULL
ONSITE MODE 15.460
14.250
8.843
13.953
8.519
13.619
7.879
99
-------
TABLE 7b
SUMMARY STATISTICS FOR DATA UNPAIRED IN TIME AND SPACE
SITE: WESTVACO LUKE TRACER; S02
3-HOUR AVERAGES
UNITS: uS/m**3
DATA SUBSET
HIGHEST
SECOND-
HIGHEST
HIGHEST
SECOND-
HIGHEST
AVG OF
TOP 10
OBSERVED
5.036
4.405
4.405
3.43-4
CTDM, SEVERAL
TOWER LEVELS 7.011
CTDM, TWO
TOWER LEVELS 19.556
CTDM, ONE
TOWER LEVEL 7.003
5.728
18.621
6.905
5.728
17.036
6.461
4.951
15.754
6.053
COMPLEX I
42.780
36,749
36.749
32.384
RTDM, DEFAULT
MODE
11.745
RTDM, FULL
ONSITE MODE 7.945
10.274
6.839
10.274
3.769
9.175
3.864
100
-------
TABLE 8a
SUMMARY STATISTICS FOR DATA UNPAIRED IN TIME AND SPACE
SITE: WIDOWS CREEK TRACER: S02
1-HOUR AVERAGES
UNITS: ug/m**3
DATA SUBSET
HIGHEST
SECOND-
HIGHEST
HIGHEST
SECOND-
HIGHEST
AVG OF
TOP 10
OBSERVED
4609
3850
2776
2627
CTDM, SEVERAL
TOWER LEVELS 6857
CTDM, TWO
TOWER LEVELS 8866
CTDM, ONE
TOWER LEVEL 6704
COMPLEX I 12453
6220
8634
6464
11474
5283
7097
6335
11255
5435
6892
5303
10857
RTDM, DEFAULT
MODE 4773
RTDM, FULL
ONSITE MODE 9555
4517
7739
4517
4435
4420
5373
101
-------
TABLE 8b
SUMMARY STATISTICS FOR DATA UNPAIRED IN TIME AND SPACE
SITE: WIDOWS CREEK TRACER: S02
3-HOUR AVERAGES
UNITS! ug/m**3
DATA SUBSET
HIGHEST
SECOND-
HIGHEST
HIGHEST
SECOND-
HIGHEST
AVG OF
TOP 10
OBSERVED
2374
1955
1049
CTDM, SEVERAL
TOWER LEVELS 2670
CTDM, TWO
TOWER LEVELS 3042
CTDM, ONE
TOWER LEVEL 3915
2286
2588
2992
1798
2588
2805
1868
2326
2782
COMPLEX I
6871
6412
5440
5446
RTDM, DEFAULT
MODE 3470
RTDM, FULL
ONSITE MODE 3293
3403
2349
3403
1478
2757
1603
102
-------
TABLE 9a
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: CINDER CONE BUTTE TRACER: SF6 UNITS: uS/m**3
(HIGHEST VALUE FROM EACH HOUR USED), 100 HOURS
DATA SUBSET
AVERAGE
HOURLY
PEAK VALUE
RATIO OF
PRE/OBS
RMS
ERROR
THRESHOLD: 0.00
% CASES:
0.5 <
PRE/OBS
< 2.0
OBSERVED
27.96
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE
TOWER LEVEL
COMPLEX I
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
26.16
0.94
28.66
12.19
6.58
42.18
36.35
21.50
0.44
0.24
1.51
1.30
0.77
36.92
37.54
45.55
58.33
26.10
37
21
17
38
27
38
103
-------
TABLE 9b
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: CINDER CONE BUTTE TRACER: CF3Br UNITS: uS/m**3
(HIGHEST VALUE FROM EACH HOUR USED), 44 HOURS
DATA SUBSET
OBSERVED
AVERAGE
HOURLY
PEAK VALUE
15.01
RATIO OF
PRE/OBS
RMS
ERROR
THRESHOLD: 0,
% CASES:
0.5 <
PRE/OBS
< 2.0
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE
TOWER LEVEL
COMPLEX I
20.95
7.51
4.57
24.45
1.40
Oo50
0.30
1.62
29.21
19.92
20.94
28.68
23
14
34
46
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
11.04
13.68
0.74
0.91
24.29
19.25
50
32
104
-------
The HBR results (Table 10, a and b) show CTDM over-predictions for
SFg (except for the run with one tower level), but underproductions
for CF3Br. COMPLEX I shows a large overprediction for SFg
(released usually above Hc) and a moderate overprediction for
CF3Br (usually released below Hc)- These differences reveal a
critical weakness in COMPLEX I: the inability to distinguish between
flow regimes passing over rather than around a terrain obstacle. RTDM
results foe the on-site mode are very good for both tracers, showing a
modest overprediction tendency. The default mode exhibits a large
overprediction tendency, a result that is not unexpected due to the
large default values of d6/dz that are used. Note that CTDM's
relatively good performance with the HBR SF$ data set (not used in
the model development) is consistent with its performance at other
CTMD sites. This result suggests that CTDM's relatively good
evaluation results at the CTMD sites are due to theoretically sound
design rather than from "tuning" and calibration.
Table 11, a and b (FSPS), shows very good results for CTDM, which
has predicted-to-observed ratios close to. 1.0 and a high number of
cases with predictions within a factor of 2 of observations. COMPLEX
I shows large over-predictions while RTDM (default) over-predictions are
more modest. For this site, RTDM (on-site mode) shows a moderate
underprediction tendency. Results for the models are consistent
between the two tracers.
For the Westvaco conventional data base, results are shown for
1-hour averages (Table 12a) and for 3-hour averages (Table 12b).
Results from Table 12 show good performance for both CTDM (all tower
levels) and RTDM (on-site mode). With fewer tower levels, CTDM
over-predicts significantly, probably due to the use of
unrepresentative or default d6/dz values that are too high.
Over-predictions are quite apparent for RTDM default (d6/dz too high)
and especially COMPLEX I (d6/dz too high and no plume lifting over
terrain in stable conditions).
Table 13 (a and b) gives results for Widow Creek for 1- and
3-hour averages, respectively. As is the case for Westvaco, the CTDM
results show more over-predictions as the meteorological data input is
degraded. The best results are attained by CTDM (all tower levels)
and RTDM (on-site onsite), while RTDM (default) and COMPLEX I
overpredict for the same reasons as stated above for Westvaco.
Statistics for observed and predicted concentrations paired in
space but unpaired in time are presented in Appendix C. The results
for this test are consistent in general with those for the data set
paired in time only.
In Appendix D, results are given for concentrations paired in
time and space for the five sites. For this test, results are
presented both for all data points as well as predicted-observed
concentration pairs that are both above a nominal value of 0.01
microseconds per cubic meter at tracer sites or 1 microgram per cubic
meter at conventional sites. There are no results that are markedly
105
-------
TABLE 10a
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: HOGBACK RIDGE TRACER: SF6 UNITS: uS/M**3
(HIGHEST VALUE FROM EACH HOUR USED), 59 HOURS
DATA SUBSET
OBSERVED
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE
TOWER LEVEL
COMPLEX I
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
AVERAGE
HOURLY
PEAK VALUE
23.48
47.84
30.09
10.82
117.50
74.66
32.60
RATIO OF
PRE/OBS
1.:
0.46
5.
3.18
1.39
RMS
ERROR
58.29
44.25
29.42
125.10
159.86
69.33
THRESHOLD: 0.
% CASES:
0.5 <
PRE/OBS
< 2.0
37
29
17
14
54
106
-------
TABLE lOb
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: HOGBACK RIDGE TRACER: CF3Br UNITS: uS/m**3
(HIGHEST VALUE FROM EACH HOUR USED), 61 HOURS
DATA SUBSET
AVERAGE
HOURLY
PEAK VALUE
RATIO OF
PRE/OBS
RMS
ERROR
THRESHOLD: 0.01
% CASES:
0.5 <
PRE/OBS
< 2.0
OBSERVED
104.32
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE
TOWER LEVEL
COMPLEX I
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
49.12
47.33
0.47
0,45
93.05
144.72
58.53
173.67
444.08
142.24
0.56
1.66
4.26
1.36
101.89
125.62
678.30
572.34
57
38
38
40
13
25
107
-------
TABLE lla
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: TRACY POWER PLANT TRACER: SF6 UNITS: uS/m**3 THRESHOLD: 0.0(
(HIGHEST VALUE FROM EACH HOUR USED), 111 HOURS
DATA SUBSET
AVERAGE
HOURLY
PEAK VALUE
RATIO OF
PRE/OBS
RMS
ERROR
% CASES:
0.5 <
PRE/OBS
< 2.0
OBSERVED
1.96
CTDM, SEVERAL
TOWER LEVELS 1.94
CTDM, SEVERAL
TOWER LEVELS 1.92
(ALT. PLUME HT 1)*
CTDM, SEVERAL
TOWER LEVELS 1.77
(ALT. PLUME HT 2)*
CTDM, TWO
TOWER LEVELS 2.07
CTDM, ONE
TOWER LEVEL 2.41
0.99
0.98
0.90
1.07
1.24
2.11
1.
1.
2.24
3.08
61
68
69
49
38
COMPLEX I
6.14
3.13
7.27
19
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
3.05
1.18
1.56
0.60
3.38
2.22
48
34
* Alternative plume height #1 was obtained from lidar measurements
at_the first cross section downwind from the source. Plume
height #2 was obtained from the second lidar cross section.
108
-------
TABLE lib
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: TRACY POWER PLANT TRACER: CF3Br UNITS: uS/Itl**3
(HIGHEST VALUE FROM EACH HOUR USED), 111 HOURS
DATA SUBSET
AVERAGE
HOURLY
PEAK VALUE
RATIO OF
PRE/OBS
RMS
ERROR
THRESHOLD: 0.00
% CASES I
0.5 <
PRE/OBS
< 2.0
OBSERVED
2.84
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE'
TOWER LEVEL
2.46
3.02
3.04
0.87
1.06
1.07
2.94
4.01
4.99
60
53
31
COMPLEX I
8.54
3.01
9.14
23
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
3.76
1.95
1.32
0.69
4.79
3.17
39
52
109
-------
TABLE 12a
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: WESTVACO LUKE TRACER: S02 UNITS: uS/M**3 1-HOUR AVGS
(HIGHEST VALUE FROM EACH HOUR USED)
AVERAGE
% CASES;
0.5 <
HOURLY
lATA SUBSET PEAK VALUE
OBSERVED 0.33
CTDM, SEVERAL
TOWER LEVELS 0»29
CTDM, TWO
TOWER LEVELS 1.60
CTDM, ONE
TOWER LEVEL 1,70
COMPLEX I 4.12
RTDM, DEFAULT
MODE 0.98
RTDM, FULL
ONSITE MODE 0.21
RATIO OF
PRE/OBS
— ,«,
0.88
4.85
5.15
12.48
2.97
0.64
RMS
ERROR
-„_
0.83
3.30
2.00
9.44
2.46
0.82
PRE/O:
< 2.0
14
-19
15
14
16
15
* All hours used with stable conditions
no
-------
TABLE 12b
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: WESTVACO LUKE TRACER: S02 UNITS: uS/m**3 3-HOUR AVGS
(HIGHEST VALUE FROM EACH HOUR USED)
AVERAGE
% CASES:
0.5 <
DATA SUBSET
OBSERVED
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE
TOWER LEVEL
COMPLEX I
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
3 -HOURLY
PEAK VALUE
0.30
0.27
1.57
1.65
3.98
0.95
0.21
RATIO OF
PRE/OBS
.90
5.23
5.16
13.27
3.17
0.70
RMS
ERROR
.
0.62
2.89
1.82
7.96
2.03
0.64
PRE/O:
< 2.0
17
18
12
9
11
18
All stable hours used for which both the highest predicted and
highest observed concentration was at least 0.01 uS/m**3
ill
-------
TABLE 13a
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITES WIDOWS CREEK TRACER: SO2 UNITS: ug/m**3 1-HOUR AVGS
(HIGHEST VALUE FROM EACH HOUR USED)
AVERAGE
% CASES;
0.5 <
DATA SUBSET
OBSERVED
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE
TOWER LEVEL
COMPLEX I
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
HOURLY
PEAK VALUE
71.87
141.78
151.91
390.43
412.78
322.34
47.56
RATIO OF
PRE/OBS
-„_
1.97
2.11
5.43
5.74
4.49
0.66
RMS
ERROR
403.58
575.81
766.83
1239.13
764.58
298.58
PRE/OBS
< 2.0
—
23
12
16
13
7
15
* All hours used with stable conditions
112
-------
TABLE 13b
SUMMARY STATISTICS FOR DATA PAIRED IN TIME, NOT IN SPACE
SITE: WIDOWS CREEK TRACER: S02 UNITS: ug/m**3 3-HOUR AVGS
(HIGHEST VALUE FROM EACH HOUR USED)
AVERAGE
% CASES:
0.5 <
DATA SUBSET
OBSERVED
CTDM, SEVERAL
TOWER LEVELS
CTDM, TWO
TOWER LEVELS
CTDM, ONE
TOWER LEVEL
COMPLEX I
RTDM, DEFAULT
MODE
RTDM, FULL
ONSITE MODE
3 -HOURLY
PEAK VALUE
66.4
112.0
127.0
342.4
349.0
278.6
41.2
RATIO OF
PRE/OBS
1.69
1.91
5.15
5.26
4.20
0.62
RMS
ERROR
___
230.3
323.1
579.2
827.8
521.4
169.2
PRE/OBS
< 2.0
31
17
20
15
9
20
* All stable hours used for which both the highest predicted and
highest observed concentration was at least 1.00 ug/m**3
113
-------
different from those presented for the data set paired in time,
unpaired in space. However, the inclusion of many zero versus zero
comparisons for all data pairs causes some sharp differences between
results with and without a minimum concentration threshold.
Test results involving the data set paired in time but unpaired
in space with results given as a function of wind speed and stability
class are shown in Appendix E. CTDM results (for all tower levels)
are discussed here.
At CCB, an overall slight underproduction for SFg was comprised
of underpredictions for neutral, windy conditions and overpredictions
for stable, light wind hours. For CF3Br, slight underpredictions
for stability D and E conditions were offset by overpredictions for
stability F.
Hogback Ridge experiments were dominated by stability F
conditions, so no model performance differences among stability
classes can be determined for that site. CTDM peak hourly predictions
were, on average, quite good at the Tracy Power Plant. Model
performance was consistently good over stability classes D, E, and F.
For the Westvaco data base, CTDM's predicted average peak hourly
value was close to the observed average value. This good agreement
was not consistent over all meteorological conditions. CTDM showed a
significant underpredietion tendency for neutral, windy conditions and
underpredicted somewhat for near-calm, stability F conditions. These
underpredictions were offset by slight to moderate overpredictions for
low-wind stability D and E conditions.
CTDM showed an overall over-prediction bias (by about a factor of
2) for the peak hourly SC>2 concentrations at Widows Creek. Among
the meteorological categories examined, none showed underpredictions
at Widows Creek. The very slight overpredietion for neutral windy
conditions is in contrast to the underpredietion tendency at
Westvaco. However, this difference may be due to the fact that at
Westvaco, about 25% of the hours have wind speeds greater than 8 m/sec
(at the top of the tower at site #1), while only about 3% of the hours
at Widows Creek are as windy (at the top of the mountain tower). The
highest overpredietion biases for CTDM occurred for low speeds (less
than 3 m/sec at release height) for all stability classes.
In an attempt to assess the overall skill of CTDM relative* to
COMPLEX I and RTDM, we have assigned arbitrary skill scores to
selected model results. This scoring scheme was not established in
advance and does not reflect a conclusive means of rating the models'
comparative performances. However, this exercise does serve to
condense the large array of statistical results to a more manageable
level for evaluation purposes.
For model bias, an ideal value for the ratio of predicted to
observed concentrations is 1.0. A scoring scale from 1 to 5 has been
established for this exercise, with maximum skill assigned a score of
1 and minimum skill assigned a score of 5. We have devised the
114
-------
following scoring scheme for ratios of predicted to observed
concentrations :
Score = 1 if ratio is between 0.8 and 1.25 (geometrically
centered at 1.0) .
= 2 if ratio is outside bounds listed above, but between
0.67 and 1.50,
3 3 if ratio is outside bounds listed above, but between
0.50 and 2.00,
3 4 if ratio is outside bounds listed above, but between
0.33 and 3.00,
» 5 if ratio is outside bounds listed above.
Measures of model scatter include normalized mean square error,
or "M" values, such as
Ml « MSE/C2, or M2 = MSE/(CO • Cp)
where
MSE is the mean square error of model predictions about
the observations,
Co is the mean observed value, and
is the mean predicted value.
The use of Cp instead of Co in M2 is meant to provide the same
skill scores for models that overpredict and underpredict by the same
ratio. A perfect model would have a mean square error of zero, but
the best available models have M values of order 1. Models with poor
skill have high M values.
The arbitrary scoring method to assess measures of model
prediction scatter about observations involves computing both Ml and
M2. For each data set, the lowest M value among all models evaluated
is first identified. Then a skill score ranging from 1 (most skill)
to 5 (least skill) is assigned as follows, based upon a model's M
value divided by the minimum M value over all models:
If ratio is less than 1.2, skill score = 1;
If ratio is between 1.2 and 1.5, skill score = 2;
If ratio is between 1.5 and 2.0, skill score = 3;
If ratio is between 2.0 and 5.0, skill score = 4;
If ratio exceeds 5.0, skill score » 5.
The results of our attempt to assess skill levels of the complex
terrain models are listed in Tables 14 through 18. Note that the
lowest scores are associated with the model with the best
performance . In each table, a skill score is given for each data base
and each model. For data unpaired in time and space, the highest
concentration (Table 14) and the average of the top 5 or 10 values,
depending upon the site (Table 15), are analyzed. For data paired in
time, the average peak hourly values are examined: ratio of mean
predicted to mean observed (Table 16), model scatter measure Ml (Table
17), and model scatter measure M2 (Table 18).
115
-------
TABLE 14
SUMMARY OF MODEL EVALUATION RESULTS (SKILL SCORES*)
DATA SUBSET:HIGHEST VALUE UNPAIRED IN TIME AND SPACE
CTDM
RTDM >
DATA BASE
TRACY:
SF6
CF3BR
SEVERAL TWO ONE
TOWER TOWER TOWER
LEVELS LEVELS LEVEL
2
4
2
3
1
3
COMPLEX I
4
3
DEFAULT
MODE
FULL
ONSITE
MODE
3
4
WESTVACO
S02
WIDOWS
CREEK
S02
TOTAL
SKILL
SCORE
13
14
16
8
15
* FOR RATIO OF PREDICTED/OBSERVED, THE FOLLOWING ARBITRARY
SKILL SCORES ARE ASSIGNED:
PRE/OBS RATIO
0,80 - 1.25
0.67 - 1.50
0.50 - 2.00
0.33 - 3.00
< 0.33 - > 3.
SCORE
1
2
3
THIS SCHEME ASSIGNS THE LOWEST SCORE TO THE BEST PERFORMANCE.
116
-------
TABLE 15
SUMMARY OF MODEL EVALUATION RESULTS (SKILL SCORES*)
DATA SUBSET:AVERAGE OF TOP N VALUES UNPAIRED IN TIME AND SPACE
(N=5 FOR CCB, HER, TRACY; N=10 FOR WESTVACO AND WIDOWS CREEK)
DATA BASE
TRACY:
SF6
CF3BR
CTDM
SEVERAL TWO ONE
TOWER TOWER TOWER
LEVELS LEVELS LEVEL
2
3
2
1
1
2
COMPLEX I
4
4
~ RTDM >
DEFAULT
MODE
2
2
FULL
ONSITE
MODE
4
4
WESTVACO
S02
WIDOWS
CREEK
S02
TOTAL
SKILL
SCORE
12
13
18
11
14
* FOR RATIO OF PREDICTED/OBSERVED, THE FOLLOWING ARBITRARY
SKILL SCORES ARE ASSIGNED:
PRE/OBS RATIO
0.80
0.67
0.50
0.33
< 0.33
1.25
1.50
2.00
3.00
> 3.
SCORE
1
2
3
4
5
THIS SCHEME ASSIGNS THE LOWEST SCORE TO THE BEST PERFORMANCE,
117
-------
TABLE 16
SUMMARY OF MODEL EVALUATION RESULTS (SKILL SCORES*)
DATA SUBSET:AVERAGE PEAK HOURLY VALUE, PAIRED IN TIME, UNPAIRED IN SPACE
CTDM
<—- RTDM >
DATA BASE
CCBs
SF6
CF3BR
SEVERAL TWO ONE
TOWER TOWER TOWER
LEVELS LEVELS LEVEL
1
3
4
4
5
4
COMPLEX I
3
3
DEFAULT
MODE
2
2
FULL
ONSITE
MODE
2
1
HER:
SF6
CF3BR
3
4
5
3
5
5
2
2
TRACY:
SF6
CF3BR
1
1
1
1
1
1
5
5
3
2
3
2
WESTVACO
S02
WIDOWS
CREEK
S02
TOTAL
SKILL
SCORE
18
24
29
34
28
18
* FOR RATIO OF PREDICTED/OBSERVED, THE FOLLOWING ARBITRARY
SKILL SCORES ARE ASSIGNED:
PRE/OBS RATIO
0.30
0.67
0.50
0.33
1.25
1,
2,
3,
50
00
00
< 0.33 - > 3.
SCORE
1
2
3
4
5
THIS SCHEME ASSIGNS THE LOWEST SCORE TO THE BEST PERFORMANCE.
118
-------
TABLE 17
SUMMARY OF MODEL EVALUATION RESULTS (SKILL SCORES*)
DATA SUBSET:AVERAGE PEAK HOURLY VALUE, PAIRED IN TIME,
UNPAIRED IN SPACE, (RMS/OBS)**2
DATA BASE
CCB:
SF6
CF3BR
CTDM
SEVERAL TWO ONE
TOWER TOWER TOWER
LEVELS LEVELS LEVEL
2
4
3
1
4
1
COMPLEX I
4
4
<—~ RTDM >
DEFAULT
MODE
4
3
FULL
ONSITE
MODE
1
1
HBR:
SF6
CF3BR
4
1
4
2
1
2
5
2
5
5
5
5
TRACY:
SF6
CF3BR
1
1
1
3
4
4
5
5
4
4
1
1
WESTVACO
SO2
WIDOWS
CREEK
SO2
TOTAL
SKILL
SCORE
17
23
26
35
35
16
* FOR MODEL PERFORMANCE MEASURES INVOLVING VARIANCE, THE FOLLOWING
ARBITRARY SKILL SCORES ARE ASSIGNED FOR M = (RMS/OBS)**2
(MODEL M)/(LOWEST MODEL M)
1.00 - 1.20
1.20 - 1.50
1.50 - 2.00
2.00 - 5.00
> 5.00
SCORE
1
2
3
4
5
THIS SCHEME ASSIGNS THE LOWEST SCORE TO THE BEST PERFORMANCE.
119
-------
TABLE 18
SUMMARY OF MODEL EVALUATION RESULTS (SKILL SCORES*)
DATA SUBSET:AVERAGE PEAK HOURLY VALUE, PAIRED IN TIME,
UNPAIRED IN SPACE, (RMS*RMS)/(OBS*PRE)
DATA BASE
CCB:
SF6
CF3BR
CTDM ——
SEVERAL TWO ONE
TOWER TOWER TOWER
LEVELS LEVELS LEVEL
1
3
4
3
5
4
COMPLEX I
3
2
<—- RTDM >
DEFAULT
MODE
4
3
FULL
ONSITE
MODE
1
1
HBR:
SF6
CF3BR
1
4
1
4
2
4
4
1
5
5
4
5
TRACY:
SF6
CF3BR
1
1
1
3
3
4
4
4
3
3
3
2
WESTVACO
S02
WIDOWS
CREEK
SO2
TOTAL
SKILL
SCORE
13
23
25
27
30
21
* FOR MODEL PERFORMANCE MEASURES INVOLVING VARIANCE, THE FOLLOWING
ARBITRARY SKILL SCORES ARE ASSIGNED FOR M = (RMS*RMS)/(OBS*PRE)
(MODEL M)/(LOWEST MODEL M)
1.00 - 1.20
1.20 - 1.50
1.50 - 2.00
2.00 - 5.00
> 5.00
SCORE
1
2
3
4
5
THIS SCHEME ASSIGNS THE LOWEST SCORE TO THE BEST PERFORMANCE.
120
-------
Results in Table 14 for the highest concentrations are based upon
three sites. For the limited period represented by Tracy Power Plant
data base (111 hours), CTDM moderately underpredicted except for the
run using only one tower level. Similarly, RTDM in on-site mode
underpredicted, while RTDM (default) slightly overpredicted. For the
Westvaco data base, CTDM overpredicted for each set of meteorological
data used, but showed the closest agreement (on an unpaired basis) for
the run using the single tower level. RTDM (on-site) overpredicted
less at Westvaco than did RTDM (default), but the reverse was true for
the Widows Creek data base. CTDM run with all available tower levels
showed results similar to that of CTDM run with one tower level at
Widows Creek. In summary, the single-tower-level runs of CTDM and
RTDM (default) showed somewhat better results for this unpaired
statistic over the three data sets examined than the model runs that
used all available meteorological data. Table 15, which is based upon
a larger sample size, shows similar results. This outcome, showing a
favorable result with the use of minimal on-site data for the
comparisons unpaired in time and space, may be fortuitous in light of
the results for tests paired in time to be discussed below. However,
the unexpected favorable results for unpaired data occurs for both
CTDM and RTDM, and imply that the default choices for vertical
potential temperature gradient and wind profiles for these models give
acceptable results for the highest concentrations unpaired in time and
space.
Table 16 is of critical importance, for it evaluates model skill
on an event-by-event basis. The skill levels of the models are
clearly distinguishable here, with CTDM using several tower levels
(CTDM-S) and RTDM in on-site mode (RTDM-0) clearly superior to the
others. The models currently designated or proposed for screening
purposes in complex terrain, COMPLEX I and RTDM in default mode
(RTDM-D), show considerably less skill.
Tables 17 and 18 show results for the measures of model scatter.
Once again, CTDM-S and RTDM-0 show more skill than the other models.
The use of M2, a measure preferred to Ml by statisticians (although
not as simple as Ml), gives a distinct advantage to CTDM-S. (Note,
however, that RTDM-0 is mostly penalized by poor performance at HBR.)
CTDM-S and RTDM-0 show more skill than the other models
evaluated, for reasons that include the following model features:
• plume growth is determined directly from turbulence
measurements;
• hourly values of de/dz are used;
• superior estimates of the critical dividing streamline
height are used.
The results clearly show that the use of meteorological data with good
resolution in the vertical is necessary to assure good model
performance by CTDM on an event-by-event basis.
121
-------
Although CTDM-S and RTDM-0 show comparable skill, CTDM is
preferable to RTDM for the following reasons:
• RTDM performed quite poorly at HER, with large
overpredictions, which are associated with centerline
impacts at short distances from plumes modeled below Hc
with low dilution wind speeds and small Oy and az
values. CTDM avoids these very large overpcedictions
because it considers deflection of streamlines by the ridge
and so avoids a direct impingement of plumes on the ridge at
short distances. A major component incorporated into CTDM
that RTDM lacks is the modeling of streamline flow around a
hill; RTDM does not allow deflections in the horizontal
plane.
• RTDM-0 underpredicts concentrations more than CTDM at the
Tracy Power Plant site, which features a relatively long
travel distance to terrain features of interest (at least 3
km). The RTDM values of ay and oz for very stable
conditions are evidently too large at these distances,
causing an overestimate in the plume cross-sectional area
and therefore an underestimate of the centerline
concentration. CTDM, on the other hand, calculates oz
growth based upon the wind speed and the Brunt-Vaisala
frequency, not just the stability class. This refined
treatment results in more accurate predictions of plume
size, especially az, at large distances.
• CTDM is better able to use meteorological data at plume
height because multiple levels of data can be input to the
model. RTDM, on the other hand, can accept only one level
of data, and therefore cannot readily compute plume height
values of wind and turbulence. Only wind speed profiles are
considered in RTDM; all other meteorological parameters are
assumed to be constant with height, a less sophisticated
treatment.
The tests unpaired in time and space generally show that CTDM
does not underpredict the peak concentrations that would be important
for regulatory application. Exceptions are CF3Br at Hogback Ridge
and SFg at Tracy. Of course, the application of the unpaired test
at tracer sites with intermittently operating monitors results in an
incomplete test. Therefore, the minor underprediction at Tracy of the
highest second-highest 1-hour concentration (5.36 ys/m3 predicted,
6.45 ys/nr observed) is not cause for concern, especially since
tests paired in time show good results (Table 11, a and b).
A more serious underprediction at HBR (129.5 ys/ra3 predicted,
390.0 ys/m3 observed) is likely caused by many hours for which the
plume was blown toward the ridge, but for which the wind direction
input to the model indicated otherwise. The 42 case hours involving
CTDM predictions of CF3Br at the Hogback Ridge were segregated into
two groups: one involving hours with releases at Tower A, where the
meteorological data were taken, and the other groups involving
122
-------
releases much closer to the ridge. These releases sites are shown in
Figure 26(a,b). Over all these cases, the predicted/observed ratio
for the average of the top 5 concentrations from each hour was 0.67, a
significant underprediction. However, the 19 cases involving releases
from tower A showed a slight over-prediction ratio of 1.04; other
remaining cases had a more serious underprediction ratio of 0.36.
This case classification shows that the modeling of release very close
to the ridge using meteorological data a considerable distance away
(in terms of the ridge width length scale) resulted in poor CTDM
performance. This result may not have occurred so dramatically for
other terrain shapes for which small changes in wind direction cause
smaller shifts in plume displacement.
4.5 Results of Case Study Evaluation (CTDM Only)
An in-depth study of the behavior of CTDM predictions relative to
observations was conducted for the three tracer sites. For the three
sites, a total of 255 hourly patterns and 191 CF3Br hourly
patterns were analyzed, with an observed and predicted pair of maps
for each.
For each hour, several characteristics of the observed and
predicted concentration fields were noted:
• the location of the plume height relative to Hc (above,
below, or close);
• plume height wind speed category (0-1, 1-3, 3-6, >6 m/sec);
• the average of the top five predictions and observations and
the ratio.of these averages;
• a categorization of the ratio discussed above (<0.2,
0.2-0.5, 0.5-1.0, 1.0-2.0, 2.0-5.0, > 5.0);
• the locations of the peak predicted and the peak observed
concentrations relative to Hc;
• the comments about the general comparison of predicted and
observed concentrations patterns.
The detailed tabulation of these characteristics for each hour is
included in Appendix G. A summary of CTDM performance for categories
of wind speed and plume height relative to Hc is also included in
Appendix G. These results are discussed below.
A summary of the findings from the case studies is given in
Table 19.. This table lists the distribution of case hours by ratio of
peak predictions to peak observations as a function of plume height
relative to Hc. Each site and tracers are listed individually, and
are discussed separately below.
For Cinder Cone Butte SFg release hours, the overall outcome
shows an unbiased model because nearly equal numbers of cases have
predicted-to-observed ratios both above and below 1.0. Cases with
plumes below Hc exhibit a slight overproduction while those with
plumes above Hc show an overall underprediction bias. Windy
conditions are associated with the underpredictions in the latter
123
-------
TABLE 19
SUMMARY OF CTDM CASE STUDIES:
PREDICTION BIAS AS A FUNCTION OF PLUME LOCATION
RELATIVE TO E
Distribution of Hours by
CCB,CF3Br
HBR.SFg
HBR,CF3Br
FSPS.SFg
Plume Height
Relative to He** « <
Below
Near
Above
Total
Below
Near
Above
Total
Below
Near
Above
Total
Below
Near
Above
Total
Below
Near
Above
Total
Below
Near
Above
Total
4
0
8
12
0
1
2
3
0
0
0
0
1
0
0
1
0
0
2
2
4
0
0
4
4
2
8
14
0
0
5
5
1
0
1
2
12
0
0
12
8
2
9
19
12
1
0
13
<=
5
7
9
21
I
2
4
7
2
0
0
2
8
4
0
12
12
2
7
21
24
1
14
39
>=
10
1
9
20
0
0
4
4
5
0
6
11
9
0
0
9
11
6
24
41
11
2
21
34
>
6
4
3
13
0
3
3
6
13
7
6
26
3
0
1
4
10
1
14
25
8
3
8
19
»
4
5
5
14
1
4
9
14
2
3
S
10
3
1
0
4
2
0
0
2
1
0
0
1
r— -w— —
Total
Hours
33
19
42
94
2
10
27
39
23
10
18
51
36
S
1
42
43
11
56
110
60
7
43
110
Grand Total
22 65 102 119 93
45
446
* Ratio categories:
1.0-2.0 is ">=",
<0.2 is "«
0.2-0.5 is "<*
>5.0 is "»"
0.5-1.0 is "<=•
** "Near Hc" was within 5 meters at CCB and HER, and within 10
meters at FSPS.
124
-------
category, an outcome consistent with that found at Westvaco. In many
of these cases, the observed maximum is on the near side of the hill
and the predicted maximum is on the far side of the hill.
The CCB CF3Br hours show an overall over-prediction bias,
contributed to in large part by the LIFT component of the model. The
hours where overpredictions occur often feature a predicted maximum on
the near side of the hill and an observed maximum farther up the hill
or on the far side of the hill. There are also fewer windy hours
associated with the CF3Br LIFT cases than with the SFg LIFT cases,
Which may partly explain the difference in the overall outcome between
the SFg and CF3&r LIFT cases at CCB.
At Hogback Ridge, a sharp distinction is evident between the
SF6 results (over-prediction bias) and the CF3BR results
(underprediction bias). The SFg plumes were released higher than
the CF3&r plumes were. The SFg predicted maxima were often
located on the near side of the ridge closer to the source than the
observed peak were located. The predicted CF3Br peaks were almost
always on the near side of the ridge, but displaced laterally from the
observed peaks for hours in which CTDM underpredicted. The hours of
poor CTDM predictions of CF3Br, while being associated with releases
very close to the ridge, also featured large wind direction
variability, causing a poorly defined mean wind direction and a large
effective hourly plume oy.
CTDM predictions at FSFS show an overall modest overprediction
tendency for SFg and a nearly unbiased overall prediction for
CF3Br. Results for LIFT predictions in windy conditions show little
overall bias.
125
-------
SECTION 5
SENSITIVITY TESTS
This section describes a sensitivity study that was performed to
illustrate how the magnitude of the terrain effect in CTDM varies with
meteorology, and with the slope and orientation of the hill. Because
CTDM is a modified Gaussian plume model, many of the characteristics
are already familiar, such as its sensitivity to wind direction.
However, many of the modifications introduce new and complex responses
to variations in meteorology, and these also change depending on the
scale and shape of the terrain that is being modeled. It is the
illustration of these responses that is the subject of this study.
Specific recommendations for analysis of sensitivity had been
made at the CTMD Workshop held in February, 1986. These are discussed
in Section 5.1. The rationale adopted in the sensitivity study, and a
description of how the matrix of model-runs was constructed is
contained in Section 5.2. The results of the study are discussed in
Section 5.3.
Also described is an operational test of the sensitivity of CTDM
to the manner in which the shape of a hill is specified. In this
test, model performance at the Widows Creek site was evaluated for two
alternatives in defining the shape parameters of one of the hills..
This test is discussed in Section 5.4.
5.1 Workshop Recommendations for Sensitivity Analyses
The CTDM Workshop held in February, 1986 made several
recommendations for testing the sensitivity of CTDM. These are:
1. Sensitivity to wind direction, noting whether the effects of
errors are to move the location of the peak only, or to
change its value as well;
2. Sensitivity to vertical dispersion, both as specified by the
sigma-w input, and the formulation of the model itself;
3. Sensitivity to the definition of the hill shape and
orientation;
4. Sensitivity to plume height relative to the dividing
streamline height;
5. Sensitivity to the potential temperature gradient through
its effect on plume rise, dividing streamline height,
sigma-z, and Froude number; and
6. Sensitivity to horizontal diffusion over a range of travel
times.
Many of these are addressed through the matrix of input data discussed
in Section 5.2, but several are better addressed analytically here.
126
-------
The spread of the plume in the vertical direction is the subject
of recommendation 2 as well as part of recommendation 5. As described
in Section 3.3.2, sigma-z in the absence of any hill effects is given
by
o,
= owt/[l -H t/2TL]1/2 (104)
so that the initial growth of sigma-z is linearly proportional to
sigma-w. When the time of travel (this includes a virtual time which
incorporates spread due to source-effects) is large compared to the
Lagrangian time-scale, sigma-z grows as the square root of the time,
oz = ow [2tTL]1/2 .(105)
and is proportional to OW/TL- The Lagrangian time-scale is
given by
TL = !/[<*„ (2.8/z + 3.7 N/ow)] (106)
where N is the Brunt-Vaisala frequency, proportional to the square
root of the potential temperature gradient, and z is the elevation of
the plume. For larger sources which may have plume heights on the
order of 100 m, TL is approximately equal to .27/N whenever a
non-zero value of H is measured, so that
oz = «w [.54t/N]1/2- (107)
When stratification is absent, TL equals .36z/ow and
a,
z
[.72ztow]1/2. (108)
This analysis indicates that, in the absence of significant
source-effects, sigma-z is generally most sensitive to the
turbulence. The potential temperature gradient and plume height are
most important in changing the dependence of plume growth on time from
a linear to a square root trend. Once the square root trend is
established, sigma-z is only weakly dependent on changes in these
because the gradient enters as a one-fourth power, and the plume
height enters as a square root. But note that the jump between the
z/ow-scaling and a 1/N-scaling can be abrupt for high plumes
because of the limit in the ability of temperature observations to
resolve very weak temperature gradients.
Source-effects can dominate sigma-z when the virtual travel-time
is large compared to TL because additional growth occurs at a rate
(3oz/3t) that is inversely proportional to the square root of
the travel-time. As the virtual time exceeds TL several-fold (due
to source-effects), the rate of additional growth diminishes so that
sigma-z can remain nearly equal to its initial value. This is
particularly evident when the stratification is strong and the
turbulence is weak.
The spread of the plume in the lateral direction is the subject
of recommendation 6. Unlike sigma-z, the Lagrangian time-scale for
sigma-y is not computed within the model. It is set to the time it
127
-------
takes the plume to travel 10 km. This scale is chosen because there
appears to be no method available for estimating it, and evidence
suggests that a linear growth is generally seen over distances up to
10 km. Consequently, sigma-y is solely determined in the model by the
wind speed and the turbulence, and is linearly proportional to changes
in sigma-v.
Recommendation 5 was addressed in part in the discussion of
sigma-z, and the sensitivity of CTDM to the difference between plume
height and Hc (recommendation 4) will be discussed in Section 5.3.
Beyond these issues, the intent of recommendation 5 is to assess the
overall sensitivity of the model to the resolution of the temperature
profile measurements. This has been done to some extent at the five
sites discussed in the evaluation results in Section 4. In general,
such a sensitivity study is possible to do only in a very
site-specific way. Hc is a function of the hill height and the wind
speed profi-le as well as the temperature profile. The height of the
plume depends on the wind speed and temperature profiles between the
top of the stack and plume height, and also on the stack height and
buoyancy flux. Furthermore, the Froude number (above Hc) depends on
the bulk speed and stratification between Hc and the top of the
hill. It seems that estimates of a matrix of plume heights, Froude
numbers, and Hc values could be obtained for a range of temperature
gradients and wind speed profiles for a specific source and hill. The
sensitivity of the hill effect in CTDM for each cell in the matrix
might then be obtained from the information in Section 5.3, and an
assessment could be made of the ability of a monitoring system to
resolve that range of gradients for which the model is most sensitive.
5.2 Test Design
The intent of this study is to illustrate how the magnitude of
the terrain effect in CTDH varies with meteorology for various
hill-shapes and orientations, so that the sensitivity of the model to
these aspects of the input data can be discussed. The measure of the
terrain effect used here was obtained by taking the peak concentration
for a simulation, and dividing by the concentration that is predicted
in the center of the plume for the same travel time, but in the
absence of the hill. This "flat/centerline" concentration contains
all of the same dispersion formulations that are contained in CTDM, it
includes complete reflection of plume material from the surface on
which the hill "sits", and it applies to the same downwind distance or
time of travel. Being a centerline concentration, it also represents
what may be thought of as a peak impingement concentration for that
travel time. What it does not include are any of the terrain-specific
features of CTDM, such as changes to the rate of diffusion (above
Hc), steering of the plume away from the hill (below Hc), or
trapping of plume material against the hill (reflection from Hc for
that portion of the hill above Hc, or reflection from the stagnation
streamline for that portion of the hill below Hc). As a result, the
terrain effect is a measure of how close (or far) the sampling "cuts''
through the plume (see Figure 1) along Hc and the stagnation
streamline push the receptor toward (or away from) the center of the
plume, combined with how much the alterations in the diffusion and
reflection processes have altered concentrations within the plume.
128
-------
Focusing on the predicted terrain effect rather than predicted
concentrations is .deliberate. CTDM remains a Gaussian plume model, as
discussed in Section 3, and shares many properties with other widely
used plume models. Instead of demonstrating all of the features
common to this type of model such as sensitivity to wind direction and
source height (relative to the plume spread), we have chosen to
illustrate how the unique features of this model respond to various
input conditions. Also, in exploring various shapes and orientations
of hills, and various wind directions, we have found that resolving
peak concentrations becomes dependent on having a very dense array of
receptors all over the hill, regardless of how large the hill may
become. The terrain effect measure reduces the need for such a dense
array of receptors and it is a concise indicator of how the terrain is
either fostering impingement or avoidance behavior.
The matrix of data used as input in this study was based in large
part on the scale of the setting at the Tracy power plant, the site of
the FSPS. A hill that is 300 m tall was placed 4 km from the source.
Several aspect ratios and source positions were specified. The aspect
ratio of a hill was defined for each of its axes as the length of the
axis at one half of the hill height, divided by the height of the
hill. The longer of the axes was oriented north-south (0°), and the
source was placed at several azimuths between 0° and 90" relative to
the center of the hill:
Aspect Ratios (Major Minor) Source Positions (degrees)
2-2 ' 90
3-2 90 80 70 45 20 10 0
5-2 90 80 70 45 20 10 0
10-2 90 80 70
Note that the hill with aspect ratios of 2-2 is a symmetric hill so
that only one source-position was required. The longest hill, 10-2,
was modeled only for azimuths of 90°, 80°, and 70° to keep the source
upwind of the hill — recall that the source is 4 km from the center
of the hill and a hill of aspect ratio 10 and height 300 m has a
length of 3 km at half its height.
Receptors were placed on each of the hills along azimuths at
intervals of 30°, starting at 0°. Along each azimuth, the receptors
were placed every 30 m in elevation between 30 m and 270 m. One
receptor was also placed at the top of the hill at 300 m elevation.
Figure 27 illustrates the relative positions of the hill, the source
locations, and the receptor azimuths for the hill of aspect ratios 3-2,
Meteorological data were chosen to be consistent with the range
of conditions observed during the experiments. The data obtained at
the elevation of the Freon release from the 150-m meteorological tower
indicate that wind speed and
-------
10°
45e
70C
90°
Figure 27. Illustration of receptor radials, source locations, and wind
directions used in the sensitivity analysis of hill 3-2.
130
-------
as a quarter of the wind speed. This choice removes the contributions
to the scatter in Figure 28 that are apparently due to meandering at
low wind speed. The stratification (Brunt-Vaisala frequency, N) is
largely independent of the other variables. Figure 28 presents
scatterplots of ow, ov, and N versus the wind speed for these
data. On the basis of this observed behavior, the following matrix of
meteorological data was used in the sensitivity modeling:
Met. Variable Values Chosen
wind speed (m/s) 1.0, 1.5, 2.5, 4.0, 7.0
N (1/s) .01, .02, .03
sigma-w (m/s) 0.1 • wind speed
sigma-v (m/s) 0.25 - wind speed
direction (deg) 0, +/-10, W-20
The direction indicated above is not simply the wind direction. It is
the difference between the wind direction and the direction from the
center of the hill to the source. For a direction of 0°, the flow
direction is from the source, wherever it may be located, to the
center of the hill.
All other parameters needed in the model were computed from these
values and the assumption that wind speed and stratification are
constant with height. Values computed for Hc were 0, 50, 67, 100,
150, 167, 175, 200, 217, 225, 250, and 267 m. Given this range, the
plume height was fixed at 225 m to supply an adequate number of cases
in which the plume is marginally above Hc. In practice, the height
of the plume will vary with the meteorology as well, but it was kept
constant in these runs to simplify the assessment of model
sensitivity, recognizing that sources typically operate at variable
loads which also alters plume height.
This combination of values for wind speed, direction, and
stratification combine to give 75 simulations for each source
position, for each hill. The peak concentration obtained for each
simulation was saved as was the terrain-effect parameter. The results
allow us to see how much the terrain-effect changes with the shape of
the hill, with the orientation of the hill to the flow, with small
changes in wind direction (+/-200), and with changes in wind speed and
stratification.
5.3 Test Results
5.3.1 Sensitivity to He and Hill Shape
Figure 29 displays the trend in the terrain effect as a function
of Hc for the case of a source located at 90° and a wind direction
deviation of 0° (the source is directly upwind of the center of the
hill). All of the hills used in the analysis are shown on this plot,
as indicated by the legend. Note that the aspect ratios that identify
each hill place the aspect ratio of the axis that is oriented along 0°
131
-------
CO
e
5
1
O :
<;-) :
• •
He
• a
*"§"'« "
Q
tt
J« .
•
•
1
a
a « •
•
•
••
*
» *
• ' «
•
•
.
WIND SPEED (m/s)
CO :
> 1 50 •
en
o.oo -
e
«
as
- * ¥
* • ' J
c- •
•
a
• 9
'
„
« *
0 *
e •
-
B
» •
[» * •
* _ • •
e *
WIND SPEED (m/s)
0.0*
0.03
O>
C/5
>>SM3.02
0.01
C.OO 2.00 *.QO 6.00 3.00 10.00
WIND SPEED (m/s)
Figure 28. Scatter plots of sigma-w, sigma-v, and Brunt-Vaisala
frequency (n) versus wind speed. Data are taken from
the MDA for Freon releases at FSPS.
132
-------
.8
§
1.4
1.3
0.0
CTDM [H = 300 Hp = 225]
WD:90 Fr:1 SOURCE.-90
2-2
SO
3-2
12O
» 5-2
16O 20O
He [ml
A TO-2
240
2-3
280
2-5
S
u
o
1.4
0.0
CTDM [H = 300 Hp=225]
WO;90 Fr4 SOURC&9O
2-2
Figure 29. Terrain-effect factor versus Hc for wind and source aligned
perpendicular to the center of the hill. Symbols denote hill
shape by aspect ratios perpendicular to and parallel to the
wind.
133
-------
first so that the hill designated 3-2 is modeled for a plume that
passes over its shorter side, while the hill designated 2-3 is modeled
for a plume that passes over its longer side. (Results for hill 2-3
with a source at 90° are equivalent to the results for hill 3-2 with a
source at 0° so long as the wind takes the plume directly over the
hill- in each case.) The results in the upper portion of this figure
are for a Froude number equaL to 1.0 above Hc, which is consistent
with the assumption that the wind speed and the stratification are
constant with height. The Froude number above He is set equal to
4.0 in the lower portion to simulate more weakly stratified conditions
above Hc .
o
Greater Than or Equal to Plume Height (225 tn)
This class contains peak concentrations which are primarily
associated with impingement conditions. The terrain effect factor
lies between .8 and 1.0 and the results are nearly identical for a] I
hill-shapes. Although a factor of 1.0 may have been anticipated for
this class , lower values result because peak concentrations are found
at receptors upwind of the point at which the centeriine of the plume
meets the hill, due to smaller values of crz and cry at these
distances. The single difference between the upper and lower portions
of the figure in this class occurs for Hc equal to plume height. In
this case, plume growth rates differ and the peak concentration is
actually found at a receptor above Hc at 240 m for all hills except
2-5 for a Froude number of 1,0. When the Froude number is 4.0, the
peak concentration occurs at a receptor below Hc at 210 m. Note
that the results for all hills with the same along-wind aspect ratio
and for the same Froude number are identical.
H^ Less Than Plume Height (225 m)
Three features dominate the behavior of the terrain-effect factor
in this class. (1) For Hc just below the height of the plume (Hc
equals 217 m and the plume is at 225 m for these simulations) , the
factor equals approximately 0.99 for all of the hill- shapes. Peak
concentrations are found as the plume material above Hc just begins
to move up the hill, and the scale or shape of the hill has little
influence on the magnitude of the terrain-effect. This is nearly an
impingement situation even though the center of the plume lies above
Hc. (2) For smaller values of Hc, as small as 140 m, the factor
may increase or decrease depending on the Froude number and the shape
of the hill. In this region, the effects of strain in the flow and
reflection from the surface of the hill determine the magnitude of the
hill-effect, which may exceed unity. (3) Beyond this range, at ever
smaller values of Hc, the terrain-effect diminishes for all of the
hills because the bulk of the plume lies far above the hill and
receptors remain far from the centeriine of the plume. Over the
latter two regimes, the magnitude of the terrain-effect increases with
the degree of strain and the length of time that the plume experiences
the strain. The terrain-effect is generally weakest for flow across a
ridge, and strongest for flow along an elongated hill.
134
-------
5.3.2 Sensitivity to Source Position and Wind Direction:
Symmetric Hill and 2D Hill
Figure 30 illustrates the sensitivity of the model to wind
direction and the azimuthal position of a source for the extremes of a
three-dimensional hill and a nearly two-dimensional ridge. As
indicated by the legend, the various curves on each of the plots
correspond to differences in wind direction of 0°, +/-100, and +/-200
from the direction that aligns the center of the hill and the source.
Note that the Froude number above Hc is equal to 1.0 in all of the
remaining figures in this section.
Hc Greater Than Plume Height (225 m)
The sensitivity of the model to wind direction is extreme for
this class. When the source is located at 90°, the terrain-effect
drops from .85 or .9 to .6 for a shift in wind direction of 10°, and
it drops to a factor as small as .2 for an additional change of 10°.
This behavior is the result of steering the plume away from the hill
when it is well within the stable layer below Hc. The primary
difference between hill 2-2 and hill 10-2 occurs for a shift of 20°
with Hc equal to 250 m in which case the peak concentration at hill
10-2 occurs above Hc on the north side of the hill, while that at
hill 2-2 remains below Hc. When the source moves more to the north
(azimuths 80° and 70°) the symmetry for differences of +/-100 and 20°
disappears, as expected, and the range in the terrain-effect factors
extends to smaller values.
Hc Less Than or Equal to Plume Height (225 m)
The sensitivity of the model to wind direction remains
substantial for hi 1.1 2-2, but not for hill 10-2. This results from
the length of hill 10-2 in the cross-wind direction compared to hill
2-2. The plumes in this analysis always pass over a substantial
portion of hill 10-2, but pass more to the side of hill 2-2 with
increasing shifts in the wind direction. Differences among the three
plots for hill 10-2 occur primarily for Hc between 160 m and 200 m.
Over this range, changes in the wind direction and the orientation of
the source produce changes in both the strain in the flow and in the
length of the path over the hill, and this fosters relatively minor
changes in the terrain effect. Some of the variability is also
associated with the spacing between receptors in that the center!ine
of the plumes may pass nearer a receptor for one of the combinations
of source location and wind direction.
5.3.3 Sensitivity to Source .Position and Wind Direction:
Asymmetric Hills
Figure 31 illustrates the sensitivity of the model to wind
direction and the azimuthal position of the source for hills
intermediate in scale between the extremes of a symmetric hill and A
ridge. The extreme sensitivity to wind direction when the plume is
well below Hc remains evident in each of the plots. At smaller
135
-------
1.3
1.2
1.1
r-i 1.0
4 o.,
^ 0.3
•s.
r—»
JL
O
1
o
0.7
0.8
0.5
0.4
0.3
0.2
0.1
0.0
CTDM [H = 300 Hp = 225]
HILL: 2-2 SOURCE «90
*
\
40 SO
Q 20
120
16O
200
1O
He [m]
o 0 A -10
24O
-2O
280
1.4
1.3
1.2
1.1
r-, 1.0
U
o
0.9
1 0.3
0.7
0.6
Jb OJ-
0.4
0^
0.1
Q.Q
CTDM [H = 300 Hp=225]
HILL:10-2 SOURCE «90
\
\\
\
40 8O
S 20
12O
18O
2OO
Ho [m]
> 10 o 0 A -10
240
x -20
280
Figure 30. Terrain-effect factor versus Hc for symmetric 3-D hill and
a long ridge. Symbols denote deviation of wind direction
from that which places the source upwind of the center of the
hill.
136
-------
o
o
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.8
0.4
0 ,3
0.2
0.1
0.0
40
o
CTDM [H = 300 Hp = 225]
HILL:10-2 SOURCE 980
x
\
ao
120
160
200
20
10
He [m]
o 0 A -10
240
x -20
280
CTDM [H = 300 Hp = 225]
HIUU10-2 SOURCE «70
He [m]
« 0 A -1O
280
Figure 30. (Continued).
137
-------
1.4-
1.3
1.2
1.1
r-, 1.0
o
\
r*-l
~0
1
U
0.9
e
u
& 0.8
0.7
0.8
0.5
0.4
0.3
0.2
0.1
0.0
40
Q
CTDM [H = 300 Hp = 225]
HILL; 3-2 SOURCE «90
\
V
\
\
\
80
1 20
20
10
1 6O
He [m]
oO A
20O
-10
24O
-20
280
1.4
1.3
1.2
1.1
r-. 1.0
O
0.9
0.8
0.7
0.6
Ji 0-5
0.4
0.3
0.2
0.1
0.0
CTDM [H = 300 Hp-225]
HILU 5-2 SOURCE
\
\
\
\
40 8O
a 20
120
160
20O
10
He [m]
oO A-10
24O
x -20
280
Figure 31. Terrain-effect factor versus Hc for hills of aspect ratios
3-2 and 5-2. Source positions are aligned with both the
major and minor axes, and 45° in between.
138
-------
1.4
1.3
• 1.2
1.1
ri 1.0 •
•£ 0.9
u
^ 0.8
" 0.7
0.6
Jb 0.5
-------
1.4
1.3
1.2
1.1
r-, 1.0 •
0
o
I
I_J
o
0.9
^ 0.8
0.7
0.6
0.5
0.4
0 J
0.2
0.1
0.0
CTDM [H = 300 Hp = 225]
HILL: 3-2 SOURCE ®00
S
\
40 80
020
120
16O
20O
10
He [m]
o 0 A -10
24O
-20
280
O
1.4
U
1.2
1.1
— 0.8
o
x 0.7
? °'S
Ji OJ
0.4
OJ
0^
0.1
0.0
CTDM [H = 300 Hp = 225]
HILL: 5-2 SOURCE «00
\
X
V
\
\L
4-0 80
Q 20
120
1 SO
200
10
He [m]
oO A -10
240
x -20
280
Figure 31. (Continued)
140
-------
values of Hc, sensitivity to wind direction is smallest for plumes
crossing an elongated hill over its shorter side, and greatest for
plumes crossing the same hill along its longer side. Intermediate
results are obtained for sources midway, at 45°. This behavior is
consistent with the results already discussed.
All of these results indicate the importance of Hc and the
stagnation streamline in allowing the center of the plume to either
impact a receptor on a hill, or avoid it. Regardless of the overall
shape of the hill, the stagnation streamline is directly related to
the wind direction and the direction from the center of the hill to
the source. Peak concentrations are obtained for plumes well below
Hc only when the source lies on the stagnation streamline, and
concentrations drop rapidly as the wind direction deviates from this
condition. The terrain-effect factor for plumes above Hc may exceed
that of direct impingement, depending on the shape and orientation of
the hill, but peak concentrations are generally less than those
resulting from impingement, because the increase in travel time leads
to an increase plume dilution.
5.4 Operational test on Hill Shape Sensitivity
Terrain features in the vicinity of the Widows Creek St«am Plant
presented an opportunity to test the sensitivity of CTDH to
uncertainties in specifying the shape of the hill. Sand Mountain,
which lies to the southeast of the power plant, is actually a broad
plateau with a sharp rise from the Tennessee River valley to the
plateau level. Because this hill is not an isolated feature, it is
unclear how to digitize it for the terrain preprocessor and. how to
specify the center. For this exercise, two hill centers were chosen
after a sufficient amount of the hill was digitized (Sec Figure 32).
One center was positioned to create a rather narrow hill (Figure 33),
while a second choice caused a wider hill to be created for input to
CTDH (Figure 34). CTDH was then run for a full year of meteorological
data, using all available tower levels, to test the sensitivity of the
model to these alternative choices.
Peak concentrations at the 7 monitors on the hill (positioned as
shown in Figure 32) are listed in Table 20. The location of the peak
concentrations change, but their magnitude is nearly the same in this
case. The orientations of the two fitted hills differ by about 10°,
resulting in a different deflection of the plume being modeled. In
this case, it is likely that if adequate receptor coverage ia given.
the peak concentrations will not be very sensitive to changes in the
terrain input.
141
-------
Stack Location
Figure 32. Digitized contours of Sand Mountain, located southeast of
the Widow's Greek steam station (see triangle), with
positions of 302 monitors indicated by site number.
L42
-------
Stack
Location
Figure 33. Elliptical fits to contours with hill "center" (see heavy
circle) positioned relatively close to the edge of the
plateau (narrow hill).
143
-------
Stack
Location
Figure 34. Elliptical fits to contours with hill "center" (see heavy
circle) located far from the edge of the plateau (wide hill)
144
-------
TABLE 20
PEAK '1-HOUR SO2 CONCENTRATIONS (|Jg/m3)
PREDICTED BY CTDH FOR SAND MOUNTAIN MONITORS USING
TWO DIFFERENT HILL CONFIGURATIONS
(WIDOWS CREEK 1980 DATA)
Narrow Hill Wide Hill
Highest
tank Site // Concentration
1
2
3
4
5
6
7
6
25
10
3
11
9
24
6857
6219
5437
5394
5330
4882
1607
Second Highest
Site // Concentration
6
3
9
10
25
11
24
5283
5034
4725
4579
4531
2374
1345
Highest
Second Highest
Site // Concentration Site //
25
6
9
3
10
11
24
6847
6303
5957
5491
5328
3111
2008
25
6
3
10
9
11
24
Concentration
4942
4896
4828
4816
4794
2421
718
-------
SECTION 6
MODEL APPLICABILITY AND LIMITATIONS
The question of applicability requires a discussion of the
theoretical limitations inherent in the algorithms of CTDM as well as
a discussion of how well the model performed for the various sites and
meteorological data sets reported in Section 5. Theoretical
limitations are addressed first.
The focus of much of the model development activity that has
culminated in CTDM is the stable plume impingement problem. A plume
is emitted into a stably-stratified flow which carries it toward
elevated terrain. The growth rate of the plume in the vertical is
small compared to that typically found in non-stably-stratified
flows. As the plume encounters the terrain, concentrations of plume
material on the hill are much greater than those that would have
occurred if the plume had diffused to the surface in the absence of
the terrain. This focus places primary emphasis on the interaction of
a plume with one terrain feature, in a flow that is documented by
measurements of wind, turbulence, and temperature stratification at
many levels in the vertical near the source. During impingement, peak
concentrations are expected along the windward face of the terrain, or
in the case of near-impingement, near the crest of the hill.
As a result, CTDM is most applicable for periods of stable
stratification at sites in which nearby terrain exceeds plume height
and at which the terrain elements can be isolated. Meteorological
data should provide adequate definition of the vertical structure of
the approach flow to the terrain^ and generally should be obtained
near the source. The degree of stratification that can be
accommodated in CTDM includes near-neutral conditions as well as
strongly stratified conditions.
Implicit in this statement of applicability are several
restrictions and assumptions that were adopted in the overall design
of the model:
1. CTDM contains no wake algorithms for simulating the mixing
and recirculation found in cavity zones in the lee of a
hill. Therefore, sources within the lee of terrain features
are not treated in the model and estimates of concentrations
at receptors in the lee may not be reliable when such zones
are present.
2. CTDM contains no global flow calculation that accounts for
the presence of many hills. The path taken by a plume
through an array of hills cannot be simulated by the model.
It relies on measurements of the flow taken in the
neighborhood of the source to define the incident flow field
146
-------
for each of the hills or terrain segments independently. If
there is a strong channeling of the flow due to large-scale
terrain features (e.g. a valley setting), then this will be
reflected in the modeling only insofar as it is contained in
the measurements.
3. All hills that are explicitly modeled are done so in
isolation; any changes to the plume size caused by one hill
are not carried forward to subsequent simulations downwind.
4. CTDH assumes that the meteorological data are representative
of the entire 1-hour averaging period, and apply to the
entire spatial domain. Spatial and temporal variability
that may be resolved by an array of meteorological towers
cannot be used directly in the model.
5. As an outgrowth of 4 in combination with the Gaussian plume
formulation, unsteady conditions which foster recirculation
of plume material are not treated in CTDM.
Other limitations arise from assumptions adopted in formulating
algorithms for phenomena included in the model. The flow-field
solutions used both above and below Hc demand simple models for the
shape of the hill. Below Hc, the model is a cylinder of elliptical
cross-section. Above Hc, the model is a hill of Gaussian profile in
the vertical and elliptical cross-section in the horizontal. The
choice of length scales for these shapes becomes less apparent as the
complexity of the terrain increases. Typical problems encountered in
selecting terrain attributes are discussed in the terrain preprocessor
user manual (Mills et al., 1987). But beyond the problem of
representing the terrain is the question of whether a terrain feature
can be modeled at all by CTDM. Above H<., the flow model is
formulated for hills of low-to-moderate slope. Steep-walled buttes
and mesas violate the low-slope assumption, and the LIFT computation
is clearly inappropriate. However, the formulation for the flow below
Kg is quite appropriate for such steep-walled features, provided
that a suitable ellipse can be used. Hence, CTDM may be considered
for use in modeling such features in the limit of very stable
stratification.
Lest these restrictions on the application of CTDM appear too
severe, we point out that CTDM is an extension to the level-terrain
Gaussian plume model. It is designed for use in the near-field of a
source where the steady-state formulation is most appropriate. It
simulates the effect of actual terrain on flow and dispersion by using
simplified terrain elements as a surrogate for the actual terrain
features. The surrogate features reflect the overall scale and
orientation of the actual terrain. Within this context, the lack of a
global treatment of transport and diffusion among an array of terrain
features is not a crippling deficiency.
A potential limitation of CTDM involves its neglect of drainage
flows. Certainly, large-scale drainage flows would be resolved by
on-site meteorological measurements, and a drainage flow would in some
circumstances be the transport flow for the plume calculation. But in
a complex array of tributary valleys, the combined effect of these
147
-------
large-scale flows on the elevation and transport of a plume would not
be modeled. Local drainage flows are another matter. Although these
too are neglected, it is not clear that this represents a limitation.
Shallow, local drainage flows were observed during the field
experiments conducted during the course of this program, but they
never had an observable effect on the location or magnitude of the
peak concentrations. The effect of the local flows appears to have
been limited to the transport of diluted plume material into the lower
basins.
The limits on the vertical growth of a plume trapped in an
elevated layer are qualitatively understood, but could not be easily
incorporated into the operational version of CTDM. The depth of such
a layer involves detailed sodar observations which are often difficult
to interpret. Underpredictions can occur, therefore, when the modeled
crz growth exceeds the thickness of the elevated layer occupied by
the plume.
Application of CTDM in the model evaluation tasks raises several
additional issues regarding limitations to the model. At the Westvaco
site, GTDM shows a tendency toward underpredicting peak concentrations
under high wind speed (neutral) conditions. For these conditions for
SFg predictions at CCB, the peak observed concentrations are
typically found on the windward face of the hill, while the model
typically places them on the leeward side. This suggests that the
rate of plume growth in the vertical in the presence of the hill is
underestimated for this condition. This may be attributable to the
formulation of sigma-z in the neutral limit for high plumes. The
evaluations at Westvaco and Widows Creek also underscore the need for
temperature gradient measurements that properly resolve the
stratification of the flow. When a single delta-T is measured, the
elevation of the lower measurement must not be too close to the
surface. If it is too close, the degree of stratification is
overestimated and the performance of CTDM suffers. In fact, all of
the tests of the model that involved the use of less than the full set
of on-site measurements showed a degradation in model performance.
This trend should not be viewed as a limitation of the model, but
rather as a guide to what can be expected from it in certain
applications with poor resolution in the vertical structure of the
flow.
A special note of caution in applying CTDM to sources very close
to a ridge may be read into its performance at the Hogback Ridge
site. When all of the cases in which the plume is below Hc are
grouped together, CTDM shows a strong bias toward underestimating peak
concentrations. But in many of these cases, the plume was released at
the foot of the ridge, while the meteorology was measured further
away. Upon removing these cases, the bias is largely removed. Most
of the cases remaining in this data set involved the release of the
plume from the main meteorological tower. Hence, it appears that one
should strive to capture the properties of the flow as close to the
source as is practical, especially for sources in the vicinity of
large, two-dimensional hills that have a substantial impact on the
flow near the source.
L48
-------
SECTION 7
CONCLUSIONS AND RECOMMENDATIONS
The Complex Terrain Model Development program objectives have
been met. The Complex Terrain Dispersion Model, CTDM, is the primary
product of the effort. This model displays considerable improvement
over the models that EPA has been using in regulatory practice,
especially on an event-by-event basis. It also shows improved
performance over RTDM, a model EPA is adopting as a third-level
screening model and which benefited from the early findings of the
CTMD program on the importance of the dividing streamline concept to
understanding stable flows.
Before describing more specific conclusions about CTDM, it is
appropriate to identify some of the other products and contributions
which have resulted from this effort.
The four field programs have produced a wealth of data for others
to use. Although the CTMD field programs were designed to focus on
specific model development needs, the data bases contain information
which should be of interest to future researchers in a number of
areas. Examples include information for further development of
dispersion models for unstable conditions and lee side effects. The
density of the sampling arrays provides sufficient coverage for
statistical analyses of monitoring plan efficiencies. The
meteorological data is extensive. . The multiple towers and
supplementary remote sensing and sounding data provide detailed
information on time and spatial variations in wind and temperature
fields of special interest to micrometeorologists and, as well, to
those interested in measurement technology issues.
One of the key technical concepts in the CTMD program was the
complementary use of field experiment data together with fluid
modeling experiment data in the development and testing of
mathematical modeling concepts. This program represents, to our
knowledge, the largest endeavor in the area of dispersion modeling to
effectively utilize this approach. The fluid modeling efforts
assisted in the design of the field experiments, in the verification
of some of the field experiment findings, and in exploring technical
areas of uncertainty in the late stages of the mathematical model
development. The success of our use of these complementary approaches
will, hopefully, encourage others to consider similar use in future
model development efforts.
CTDM Attributes and Limitations
CTDM is an improved and versatile refined air quality model for
use with elevated point sources in high terrain settings during stable
conditions. Its improvements over the screening models currently used
in complex terrain applications can be attributed to several factors:
149
-------
e its ability to use observed vertical profiles of
meteorological data (rather than just one level) to obtain
plume height estimates of these variables;
• computation of plume dispersion parameters, 0y and
oz, directly from turbulence measurements rather than
indirectly from discrete stability classes.
Despite these advances, CTDM still contains several limitations:
• Its framework is a steady-state Gaussian model. It is not
designed for extreme light-wind conditions with highly
variable wind directions.
• The mathematical depiction of terrain shapes is simplified
from actual shapes.
• Flow interactions among different terrain features are not
explicitly accounted for.
• Heteorological data can be input to the model for only one
location.
• Flow deformation in the LIFT module is treated with
linearized equations of motion for steady-state Boussinesq
flow, with higher order terms neglected. These assumptions
are not valid for applications involving steep terrain
(greater than about 15°) or strongly stable flow (Froude
number of order 1).
CTDM can be used for regulatory applications involving a long
series (e.g., a full year) of model simulations. Several of its
limitations are related to the desire to keep the computer execution
time reasonable.
An operational limitation of the current version of CTDM is that
it provides concentration estimates only for stable hours. For
averages of concentrations over several hours, including nonstable
conditions, a second model must be run to augment the GTDM
predictions. CTDM also presents operational challenges to the user.
Detailed terrain and meteorological data must be provided. "Isolated"
terrain elements need to be defined, and this task can be complicated
by superimposed and/or interconnected features. The considerable
demands for meteorological input, while necessary, represent a
significant increase over those for current models that use a single
level of data.
The CTDM user must be careful in obtaining the proper
meteorological data for the model. As has been stated in Section 5,
CTDM can be very sensitive to errors in wind direction, for example.
Plume
-------
resolution in the vertical will degrade the performance of CTDM (on an
event-by-event basis, at least). The use of tall towers or doppler
acoustic sounders will be necessary to obtain representative wind and
turbulence data. The capability for accurate remote temperature
sensing is still being developed, but representative AT measurements
are essential for obtaining accurate concentration estimates. Such
measurements can be obtained from two levels on a tall tower or from
two separate (but electronically linked) shorter towers (one on a
hill) if instruments are placed well away from the ground (e.g., 50
meters or higher) on each tower.
Performance Assessment
The accuracy of CTDM has been assessed in the model evaluation
analysis described in Section 4. Various statistical measures used in
the evaluation include model bias, model "scatter," and the percentage
of model predictions within a factor of 2 of the observations. A
particularly relevant statistic for model evaluation (Hanna and
Heinold, 1985) is the normalized mean square error (version 2 as used
in Section 4):
M value
This parameter is chosen because it contains no arbitrary
weighting and accounts for both model bias and random variances in the
model predictions. To simplify comparisons among data sets, the mean
square error is made dimensionless by dividing it by the product of
the mean observed and predicted concentrations. Low values of M are
associated with good models. High concentrations are strongly
weighted in this scheme because the difference, Cp-C0, is likely
to be large for high concentrations. In general, a very good model
has an M value of the order 1 or less, while models with little skill
have on M value of about 5 or more (see Hanna and Heinold, 1985).
The M values from the evaluation results reported in Section 4
are summarized in Table 21. For the tracer experiments, with high
spatial resolution, results are shown for the data sets paired in
time, not space. For the conventional S02 networks with low spatial
resolution but a long monitoring record, results are reported for the
data subset paired in space, not time. The CTDM results for all tower
levels are quite good, with most M values between 1 and 2. A
deterioration in performance is evident for CTDM using the degraded
data. RTDM (on-site) shows good performance except for HBR; reasons
for its problems at HBR have been discussed in Section 4. The benefit
of on-site meteorological data is evident for both CTDM (all tower
levels) and RTDM (on-site).
It is useful to compare CTDM's M values with those of EPA refined
models as listed in Appendix A of the Guideline on Air Quality Models
(Revised). 1986. CRSTER has been tested at tracer sites in Illinois
(flat site) and Tennessee (moderately hilly site; see Hanna et al,
1986). These experiments, sponsored by the Electric Power Research
151
-------
TABLE 21
SUMMARY OF M VALUES FROM THE COMPLEX TERRAIN EVALUATION DATA BASES*
H
u-.
M
CCB, SF
6
CCB. CF Br
HBR. Sf
6
HBH, CF Br
FSPS, SF,
6
FSPS, CF_Br
Westvaco, SO.
Widows Creek, SO,
CTDM, All
Tower Levels
1.12
2.71
3.03
1.92
1.17
1.24
0.27
1.27
CTDM, 2 Tower
Levels
4
3
2
3
1
1
1
1
.00
.52
.77
.75
.25
.88
.58
.82
CTDM, 1 Tower
Level
7
6
3
1
2
2
0
1
.66
.39
.41
.94
.03
.88
.16
.30
Complex I
1
2
5
0
4
3
5
3
.76
.22
.59
.82
.41
.44
.58
.02
RTDM.
3.
3.
14.
9.
1.
2.
0.
0.
Default
35
56
58
37
92
15
64
98
RTDM. Oi
1.13
1.80
6.28
22.52
2.13
1.82
0.36
2.22
*Data subset paired in time, not space was used for the tracer experiments (CCB, HBR, FSPS);
data subset paired in space not time was used for the SO. sites, 1-hour averages
(Uestvaco, Widows Creek).
-------
Institute, featured several weeks of data collection at a network of
150-200 tracer samples. ISC was tested by Hanna and Schulman (1985)
with tracer data bases collected by the American Gas Association (AGA)
at two natural gas compressor stations. These tests featured movable
arrays of some 40 tracer samples that were located in the wake zone of
a building; the aerodynamic building downwash algorithm in ISC was
tested.
The H values from these evaluation results are summarized in
Table 22. It is evident that CTDM's performance at the CTMD tracer
sites is comparable to those of EPA-designated refined models in
similar test environments.••
CTDM, while showing good performance at the evaluation sites,
also exhibits an overprediction tendency for most of the data bases;
this is important for regulators who are interested in protecting air
quality through the use of analytical modeling techniques. The most
serious underproduction result, at Hogback Ridge (CF3Br), is
associated with mobile crane tracer releases close to the ridge, while
using meteorological data from the main tower farther from the ridge.
This supports the concept that the location as well as the vertical
resolution of the meteorological data must be designed with care for
CTDM use.
Recommendations
The additional number of meteorological and terrain input
variables requires more care on the part of the user. The terrain
must be specified for each receptor; the best model performance is
realized when the terrain feature most local to each receptor is
specified (see terrain preprocessor user guide. Mills et al., 1987).
Receptor coverage on each terrain feature should be extensive to
assure the identification of the highest concentrations.
Meteorological measurements made close to the release point's
horizontal and vertical positions are essential for good model
results. Some testing on the sensitivity of CTDM to less ideal input
to the model has been discussed in Sections 4 and 5. We recommend
more testing of CTDM by the user community in real-world applications
to provide additional information on model sensitivity. The terrain
input requirements are new to complex terrain modelers and feedback on
actual experience will be valuable. Requirements for meteorological
data are quite demanding. Situations to avoid due to poor resulting
model performance need to be further defined, such as the HBR CF3Br
releases very close to the ridge that were accompanied by, perhaps,
misrepresentative meteorological input. Special attention should be
given to model performance for two-dimensional ridge or mountain
valley situations. In addition, cases involving plume transport for
several kilometers before terrain is encountered (such as buttes or
mesas in the western U.S.) need to be tested.
CTDM does not predict concentrations for hours when the modeled
plumes are in a convective boundary layer. We see the need for
further model development to close this gap and provide a "complete"
153
-------
TABLE 22
SUMMARY OF M VALUES FROM TRACER EXPERIMENTS*
FOR CTDM AND OTHER REFINED AIR QUALITY MODELS
Data Base Model M Value
GCB, SF CTDM 1.12
•
-------
model for regulatory applications requiring sequential use of
meteorological data. Some of the CTHD experiments did cover the
period of inversion breakup and subsequent convective activity. These
data could be used as part of the recommended further development.
Future work on complex terrain models should also seek to improve
the estimates of meteorological data at release height for cases where
observations are not ideally designed. A model preprocessor that
takes into account terrain interactions and thermal stratification as
well as multiple sources for meteorological data could be used to
refine the input data to CTOH.
The CTMD program was designed to focus upon plume impingement
cases on nearby terrain features. This project has resulted in a
model that has been applied to a larger variety of plume interactions
with terrain. For example, concentration estimates at receptors on
the lee side of the hills have been attempted for the CTMD sites, even
though CTDM was not specifically designed and developed for that
application. Synder, 1987 (Appendix H) reports upon the ratio of
maximum ground-level concentrations measured in the Fluid Modeling
Facility tow tank in the presence of hills versus those with the hill
removed. This ratio, called the terrain amplification factor, can be
highest on the lee side of the hill in some cases. Therefore, we
recommend more investigation into the phenomena of lee-side effects as
well as plume behavior on terrain features beyond those adjacent to a
source.
The data analyses performed during this program effort support
the concept that there are inherent limits to our ability to predict
measured or observed air quality concentrations. Improvements to
models, such as those accomplished in this effort, establish
confidence that a model is properly accounting for the physical
phenomena involved, and is therefore "fair" in its application to
different situations. It is especially noteworthy in this regard that
CTDM consistently performed well with all of the data sets used, in
contrast to the other models tested. Nevertheless, the effort has not
resulted in a "breakthrough" in reducing statistical uncertainty
associated with individual predictions versus observations. The use
of a high resolution profile of meteorology measurements with height
resulted in improvements to CTDM's performance. It is apparent from
our case-study analyses that further model performance improvements
would emerge from an increase in the information on horizontal as well
as vertical variations in meteorological data (i.e., better local,
geographic coverage).
155
-------
REFERENCES
Batchelor, G.K. 1970. An Introduction to Fluid Dynamics. Cambridge
University Press, London NW18 England.
Briggs, G.A. 1973. Diffusion Estimation for Small Emissions. ATDL
Contribution File No. 79, Atmospheric Turbulence and Diffusion
Laboratory.
Briggs, G.A. 1975. Plume Rise Predictions. Lectures on Air Pollution
and Environmental Impact Analyses. AMS, Boston.
Brighton, P.W.M. 1978. Strongly Stratified Flow Past Three-
Dimensional Obstacles. Quarterly Journal of the Royal
Meteorological Society. 104; 289-307.
Britter, R,E., J.C.R. Hunt and K.J. Richards 1981. Airflow Over a
Two-Dimensional Hill: Studies of Velocity Speed-Up, Roughness
Effects and Turbulence. Quart. J.R. Met. Soe.. 107; 91-110.
Burt, E.W. 1977. Valley Model User's Guide. EPA-450/2-77-018.
U.S. EPA, Office of Air Quality Planning and Standards, Research
Triangle Park, HC.
Businger, J.A. 1973. Turbulent Transfer in the Atmospheric
Surface Layer. Chapter 2 in Workshop on Micrometeoroiogy. D.A.
Haugen (ed.). American Meteorological Society, Boston, MA.
Carson, D.J. 1973. The Development of a Dry Inversion - Capped
Convectively Unstable Boundary Layer. Quart. J.R. Meteorol.
Soe.. 99.: 450-467.
Crapper, G.D., 1959. A Three-Dimensional Solution for Waves in the
Lee of Mountains, J. Fluid Mech.. 6_: 51-76.
Csanady, G.T. 1974. Equilibrium Theory of the Planetary Boundary
Layer with an Inversion Lid. Boundary Layer Meteor.. 6_: 63-79.
Deardorff, J.W. and G.E. Willis 1975. A Parameterization of Diffusion
into the Mixed Layer. J. Appl. Met.. 14: 1451-1458.
DiCristofaro, D.C. 1986. EPA Complex Terrain Model Development FSPS
Modelers' Data Archive - 1986. U.S. EPA, Atmospheric Sciences
Research Laboratory, Research Triangle Park, NC.
DiCristofaro, D.C., D.G. Strimaitis, B.R. Greene, R.J. Yamartino,
A. Venkatram, D.A. Godden, T.F. Lavery, and B.A..Egan 1986. EPA
Complex Terrain Model Development Program: Fifth Milestone
Report - 1985. SPA/600/3-85/069, U.S. Environmental Protection
Agency, Research Triangle Park, HC.
156
-------
REFERENCES (Continued)
Drazin, P.6. 1961. On The Steady Flow of a Fluid of Variable Density
Past an Obstacle. Tellus. 13: 239-251.
Eberhard, W.L. 1986. Contributions by Wave Propagation Laboratory to
EPA's Complex Terrain Model Development Project. NOAA Technical
Memorandum ERL WPL-143. Wave Propagation Laboratory, Boulder, CO,
Egan, B.A., R.J. Paine, P.E. Flaherty and J.E. Pleim 1985. Evaluation
of COMPLEX I and RTDM Using 1979-1980 Data from the TVA Widows
Creek Monitoring Network. ERT Document PD523-AOO. Available
from Hunton & Williams (UARG), .Washington, D.C.
Greene, B.R. 1985. Complex Terrain Model Development Quality
Assurance Project for Small Hill Impaction Study No. 2. ERT
Document P-B876-350. Prepared for U.S. EPA, Research Triangle
Park, NC.
Greene, B.R. 1986. Complex Terrain Model Development: Quality
Assurance Project Report for Full-Scale Plume Study. ERT
Document P-B876-725. Prepared for U.S. EPA, Research Triangle
Park, NC.
Greene, B.R. and S. Heisler 1982. EPA CTMD Quality Assurance Project
Report for SHIS #1. ERT Document P-B348-350. Prepared for U.S.
EPA, Research Triangle Park, NC.
Guldberg, P.H., J.P. Myers, K.W. Wiltsee, and P. Morganstern,
1977- Handbook for the Single Source (CRSTER) Model.
EPA-450/2-77-013. EPA Office of Research and Development,
Research Triangle Park, NC.
Hanna, S.R., 1983. Lateral Turbulence Intensity and Plume Meandering
During Stable Conditions. J. Glim, and Appl. Meteor.. 22;
1424-1430.
Hanna, S.R. and D.W. Heinold 1985. Simple Methods for Comparative
Evaluation of Air Quality Models, in Proceedings of the 15th
International Technical Meeting on Air Pollution Modeling and Its
Applications. NATO/CCMS.
Hanna, S.R., J.C. Weil, and R.J. Paine, 1986. Plume Model
Development and Evaluation - Hybrid Approach EPRI Contract No.
RP-1616-27. Prepared for Electric Power Research Institute, Palo
Alto, CA.
Hess, J.L. and A.M.D. Smith, 1962. Calculation of Non-Lifting
Potential Flow About Arbitrary Three-Dimensional Bodies.
McDonnell-Douglas Report E.S. 40622.
157
-------
REFERENCES (Continued)
Holzworth, G.C. 1980. The EPA Program for Dispersion Model
Development for Sources in Complex Terrain. Second Joint
Conference on Applications of Air Pollution Meteorology, New
Orleans, LA. AMS, Boston.
Hovind, E.L., M.W. Edelstein, and V.C. Sutherland, 1979. Workshop on
Atmospheric Dispersion Models in Complex Terrain.
EPA-600/9-79-041. U.S. EPA. Research Triangle Park, N.C,
Hunt, J.C.R., and R.J. Mulhearn 1973. Turbulent Dispersion from
Sources Near Two-Dimensional Obstacles. J. Fluid Mech.. £1:
245-274.
Hunt, J.C.R. and W.H. Snyder 1980. Experiments on Stably and
Neutrally Stratified Flow Over a Model Three-Dimensional Hill.
J. Fluid Mech.. 9.6.: 671-704.
Kato H., O.M. Phillips 1969. On the Penetration of a Turbulent
Layer Into Stratified Fluid, J. Fluid Mech.. 37,: 643-655.
Lavery, T.F., A. Bass, D.G. Strimaitis, A. Venkatram, B.R. Greene,
P.J. Drivas, and B.A. Egan, 1982. EPA Complex Terrain Model
Development Program; First Milestone Report - 1981.
EPA-600/3-82/036, U.S. Environmental Protection Agency, Research
Triangle Park, NC.
Lavery, T.F., D.G. Strimaitis, A. Venkatram, B.R. Greene,. D,,C.
DiCristofaeo, and B.A. Egan, 1983. EPA Complex Terrain Model
Development Program: Third Milestone Report - 1983.
EPA-600/3-83/101, U.S. Environmental Protection Agency. Research
Triangle Park, NC.
Lavery, T.F., D.G. Strimaitis, and B.A. Egan 1986. A Workshop Report
on the Complex Terrain Model Development Project (February 4-6,
1986). Prepared for U.S. EPA, Atmospheric Sciences Research
Laboratory, Research Triangle Park, NC.
Mills, M.T., R.J. Paine, E.M. Insley, and B.A. Egan 1987. The
Complex Terrain Dispersion Model (CTDM) Terrain Preprocessor
System1 - User Guide and Program Descriptions. Prepared for U.S.
EPA, Atmospheric Sciences Research Laboratory, Research Triangle
Park, NC.
Overcamp, T.J. 1983. A Surface-Corrected Gaussian Model for Elevated
Sources. J. Glim, and Appl. Met.. 22: 111-1115.
Paine, R.J. 1987. User's Guide to the CTDM Meteorological
Preprocessor (METPRO) Program. Prepared for U.S. EPA, Atmospheric
Sciences Research Laboratory, Research Triangle Park, NC.
158
-------
REFERENCES (Continued)
Paine, R.J. and B.A. Egan 1987- User's Guide to the Rough Terrain
Diffusion Model (RTDM) - Revision 3.20. ERT Document
PD-535-585. ERT, Inc., 696 Virginia Road, Concord, MA 01742.
Paine, R.J., D.6. Strimaitis, M.G. Dennis, R.J. Yamartino, M.T. Hills,
and E.M. Insley 1987. User's Guide to the Complex Terrain
Dispersion Model. Prepared for U.S. EPA, Atmospheric Sciences
Research Laboratory, Research Triangle Park, NC.
Riley, J.J., Liu, H.T. and Geller, E.W. 1976. A Numerical and
Experimental Study of Stably Stratified Flow Around Complex
Terrain. EPA Report Mo. EPA-600/4-76-021, Res. Tri. Pk., NC, Alp.
Schulman, L.L. and S.R. Hanna 1985. Evaluation of Downwash
Modifications to the Industrial Source Complex Model. JAPCA. 36:
258-264.
Sheppard, P.A. 1956. Airflow Over Mountains. Quart. J. R. Meteor.
Soc.. 82: 528-529.
Smith, R.B., 1980. Linear Theory of Stratified Hydrostatic Flow
Past an Isolated Mountain, Tellus. 32: 348-64.
Snyder W.H. 1987. Contributions of the Fluid Modeling Facility to
EPA's Complex Terrain Model Development Program. May 1987 EPA
Report. Atmospheric Sciences Research Laboratory. Research
.Triangle Park, NC.
Snyder, W.H., R.E. Britter and J.C.R. Hunt 1980. A Fluid Modeling
Study of the Flow Structure and Plume Impingement on a
Three-Dimensional Hill in Stably Stratified Flow. Proc. Fifth
Int. Conf. on Wind Engr. (J.E. Cermak, ed.). 1: 319-329, Pergamon
Press, NY,NY.
Snyder, W.H. and J.C.R. Hunt 1984. Turbulent Diffusion from a Point
Source in Stratified and Neutral Flows Around a Three-Dimensional
Hill; Part II - Laboratory Measurement of Surface Concentrations.
Atanos. Envir.. 18: 1969-2002.
Snyder, W.H., R.S. Thompson, M.S. Shipman 1986. Streamline
Trajectories in Neutral and Stratified Flow Over a
Three-Dimensional Hill. Appendix to EPA Complex Terrain Model
Development; Fifth Milestone Report - 1985. EPA/600/3-85/069,
January, 1986.
Strimaitis, D.6., A. Venkatram, B.R. Greene, S. Hanna, S. Heisler,
T.F. Lavery, A. Bass, and B.A. Egan, 1983. EPA Complex Terrain
Model Development Program; Second Milestone Report - 1982.
EPA-600/1-83/015, U.S. Environmental Protection Agency. Research
Triangle Park, NC.
159
-------
REFERENCES (Continued)
Strimaitis, D.G., T.F. Lavery, A. Venkatam, D.C. Dicristofaro,
B.R. Greene, and B.A. Egan, 1984. EPA Complex Terrain Model
Development Program; Fourth Milestone Report - 1984.
EPA-600/3-84/110, U.S. Environmental Protection Agency, Research
Triangle Park, NC.
Taylor, G.I. 1921. Diffusion by Continuous Movements. Proc. London
Math. Soe. Ser. 2.. 20: 196.
Truppi, L.E. and G.C. Holzworth. 1983. EPA Complex Terrain Model
Development: Description of a Computer Data Base from Small Hill
Impaction Study #1, Cinder Cone Butte, Idaho. EPA/600/3-84/008.
U.S. Environmental Protection Agency, Research Triangle Park, NC.
98 pp.
Truppi, L.E. 1985. EPA Complex Terrain Model Development:
Description of a Computer Data Base from Small Hill Impaction
Study #2, Hogback Ridge, New Mexico. EPA/600/3-84-038. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
87 pp.
Truppi9 L.E. 1986. EPA Complex Terrain Model Development:
Description of a Computer Data Base from the Full Scale Plume
Study, Tracy Power Plant, Nevada. EPA/600/3-86/068. U.S.
Environmental Protection Agency, Research Triangle Park, NC.
105 pp.
Truppi, L.E. 1987. EPA Complex Terrain Terrain Model Development:
Description of a Computer Data base of SAS™ System Data Sets
from Tracer Field Studies at CCB, Idaho, Hogback Ridge, New
Mexico, and the Tracy Power Plant, Nevada. U.S. Environmental
Protection Agency, Research Triangle Park, NC (in press).
U.S. Environmental Protection Agency 1986. Guidelines on Air
Quality Models (Revised). EPA-450/2-78/027R. July 1986. Office
of Air Quality Planning and Standards. Research Triangle Park, NC
Venkatram, A., Do Strimaitis, and D. DiCristofaro 1984. A
Semiempirical Model to Estimate Vertical Dispersion of Elevated
Releases in the Stable Boundary Layer. Atmos. Environ.. 18:
923-928.
Wackter, D.J. and R.J. Londergan 1984. Evaluation of Complex Terrain
Air Quality Simulation Models. EPA-450/4-84-017, Office of Air
Quality Planning and Standards, EPA, Research Triangle Park, NC.
243 pp.
Yamartino, R.J. 1987. Exact Solutions to the Linearized Equation
for Stratified Flow Over Terrain in Multi-Dimensional Space. To
be published, (see Appendix A)
160
-------
APPENDIX A
EXACT SOLUTIONS TO THE LINEARIZED EQUATION
FOR STRATIFIED FLOW OVER TERRAIN IN
MULTIDIMENSIONAL SPACE
161
-------
Exact Solutions to the Linearized Equation for Stratified
Flow Over Terrain in Multidimensional Space
Robert J. Yamartino
Sigma Research Corporation
1. Introduction
In an attempt to improve the theoretical basis of the U.S. EPA's Complex
Terrain Dispersion Model (CTDM) and enhance its predictive power under a
range of stratification (and shear) conditions, a basic investigation of
the linearized, partial differential governing equations was undertaken.
A number of researchers have examined various aspects of the problem of
stratified flow over two- and three-dimensional obstacles. Queney
(1947) considered hydrostatic (i.e., highly stratified) flow past a 2-d
ridge and Smith (1980) extended this to 3-d symmetric hills. The
hydrostatic assumption makes it difficult, however, to connect these
solutions to those appropriate for neutral flows. Hunt, Leibovich, and Lumley
(1981) and Hunt and Richards (1984) suggest interpolative methods for
connecting these regimes. Other researchers have focused on the nature of the
lee waves far downwind of the hill. Wurtele (1957) and Crapper (1959)
examined the vertical velocity field of these far field waves, whereas
Janowitz (1984) provides a complete description of all flow quantities in the
far field of a dipole (i.e., Dirac delta function) obstacle. Berkshire (1985)
and Bois (1984) provide detailed analyses of far field lee waves in two
dimensions.
As the EPA is concerned with estimating pollutant concentrations in hilly
terrain, it becomes necessary to predict the path followed by pollutant plumes
in the near vicinity of such terrain. Thus, far field solutions are of
limited usefulness and efforts must be focused on the near field (or complete)
solutions. In addition, the moderately stratified conditions that often exist
in nature, correspond to hill Froude numbers of order unity: a regime that
satisfies neither the near neutral assumption of F » 1 nor the hydrostatic
assumption of Fr« 1. Hence , effort must be directed to eliminating
approximations in several areas. Finally, the results must be easy to use and
inexpensive to compute; thus, eliminating integral formulations (e.g.,
162
-------
Trubnikov, 1959) or numerical methods requiring repeated application of Fast
Fourier Transform (FFT) techniques. In this paper we report on the pure
theoretical developments associated with this effort.
Section 2 describes the basis for the partial differential equation
governing the three-dimensional (3-d) problem and the integral formulation of
this problem. Section 3 presents the exact solution for the 3-d problem,
whereas solutions in fewer dimensional space and their interrelations are
discussed in Section 4. Finally, Section 5 summarizes the results of this
paper.
2. The Mathematical Model
We begin with the linearized equation of motion for steady-state
flow of a Boussinesq fluid (Smith, 1980):
POUU; = -PX da)
p Uv< = -P' (lb)
Ko x y
Pa Uwx = -P; - P'g (10
u' + v' + w' » 0 (Id)
x y z
(le)
These equations, in which subscripts x, y, and z indicate derivatives with
respect to downstream, cross-stream, and vertical coordinates respectively,
relate the perturbation velocities, u' , v' , and w' , to the perturbation
density, p' , and vertical fluid displacement, TJ * 7)(x,y,z),and to the
unperturbed initial velocity U and density p . Adding the kinematic condition
for steady flow in a shear-free flow,
w' = Un (2)
J\
Eqs.(l) can be reduced to the single partial differential equation (PDE) for
163
-------
t^'xx * nV , = 0 (3)
, o 2 s 2
, _2 a a
where v~3 —^ + —-,
H ax2 ay2
n = N/U, and the Brunt Vaisala frequency, N, is defined as
N = -(g/PQ)
dz
Representing T)(x,y,z) as the 2-d Fourier transform
00
D a XX dk dlT7 exp{i(kx + ty}} (4)
-OB
reduces Eq.(3) to the equation
\z * "^ = ° (5)
where m2 = n2 (k2 + ^)/k2 - (k2 + ^), (6)
and the outgoing wave, or radiation, condition dictates that the exp(+imz)
solution choice be made. Smith (1980) chose to work with the solution of
Eq. (5), that is, T)(k,£,z) = 7)(k,£,0) exp(imz); however, this is cumbersome
as m involves a branch cut. Instead, we utilize the relation
f "
»& I 1 I*
exp(imz) a — J - dq exp(iqz)
az I * J —2—2
v. -a, Q -m
(7)
to convert the singularity to a simple pole and re-express the problem,
including the linearized terrain boundary condition that flow at the ground
follow the ground or i)(x,y,z' = 0) = h(x,y), (where h(x,y) is the terrain
height function and z' represents height above terrain), in terms of the
convolution
(8a)
where I = Re ^ JJ dx'dy'hfx-x', y-y') G(x',y',z'), (8b)
164
-------
Re denotes the real part, and G is the Green's function
09
G(x'.y'.z') = -^ JJJ dk dl dq exp(i(k x'+£ y'+q z'))/(q2-m2) (9)
—00
For the Dirac delta function hill, normalized such that
JJdx'dy' h(x'.y') = 1, we see that Kx.y.z') - G(x,y,z')/(27t), as
one expects from the definition of a Green's function solution.
In Appendix A, one solution of Eq.(9) is determined to be
G(x',y',z') = sin(nRz'/d)/R , (lOa)
2222 222
where Rax' + y' + z' and day' + z' ; however, this solution does not
obey the surface boundary condition that the vertical deflection,
TJ =» -I / = - G_//(27t), vanish at the surface away from the origin (x'=y'=0).
z z
Yamartino and Pavelle (1987) show by direct substitution, using the algebraic
processor MACSYMA, that both Eq.(lOa) and
G(x',y',z') = cos(nRz'/d)/R (lOb)
are solutions to the equation
(v2!)^ + nVl = 0 , (11)
for the delta function hill and therefore are solutions to the basic PDE
given by Eq.(3).
The solution given by Eq. (lOb) obeys the aforementioned surface boundary
condition and also has the desirable property that G goes to the well known
solution G » 1/R in the neutral (n = 0) limit. However, Eqs.(lOa) and (lOb)
are not the only solutions. Any linear combination of these solutions and any
spatial derivative of these solutions is also a solution. Hence, after
struggling to obtain one solution, we now have the ambiguity of an infinity of
solutions, that can only be resolved by obtaining the correct particular
solution via proper evaluation of the integral in Eq.(9). Thus, there is no
guarantee that Eq.(lOb) is the correct choice; nevertheless, it is one of the
165
-------
simplest.
In Appendix AB, the Eq. (lOb) solution is applied to the problem of the
infinite field of cosine hills and is found to yield a solution that differs
2 2 1/2
from the correct solution by a factor of (n /k - 1) , where k is the
M ri
x-wavenumber of the hills. This implies that Eq. (lOb) is not the correct
full form of the Green's function and that other combinations should be
checked (e.g., derivatives and/or linear combinations); however, very few of
the possible candidates can even be checked via the direct integration
approach of Appendix AB, as the integrals are not presently doable. This, of
course, raises the practical consideration of the value of expending
substantial effort to find the correct particular Green's function solution if
integrals convolving it (e.g., I, Ix, I30*) together with practical hill
functions can not be easily evaluated. Nevertheless, it would be of
considerable theoretical interest to have the correct form of the solution.
Janowitz (1984) has isolated the leading term of the far-field portion
(i.e., nR » 1) of the problem and his solution has the same argument of the
trigonometric functions as in Eq.(10), but does not appear computable as
simple derivatives of either Eq.(lOa) or (lOb). This is not surprising, as
the stationary phase technique he employed projects out the leading dependence
and not necessarily a complete or exact solution. Nevertheless, Janowitz's
solution represents an important theoretical milestone.
The closest we can presently come to matching his far field expression
Thus, (ignoring
involves taking ^p- of the expression presented in Eq. (lOb).
the primes on x and y) TJ
(i.e., decays slowest in R) of this expression is Just
- G ,/2n and the longest lived, or wave, part
V - • T •
W 2n d3
f 4
[d
2 21 1/2
+ x y j ' •
r-« 22f/2
I y J
Rd
cos(nRz'/d),
In the limit x » y » z', one finds that { } •» 1 and, with the exception of a
factor n/2, Janowitz's result is obtained. Hence, Eq.(lOb) cannot be too far
from the desired result.
The correct particular solution could also possibly involve an
166
-------
integration e.g., [ dx cos(nRz'/d)/R however, this would lead us into the
—on
realm of the generalized cosine-integral functions, about which little seems
to have been done analytically since the studies of Aiken (1949).
Several other issues also emerge when the exact particular solution is
sought. The first of these involves the related differential equation
governing the conjugate , potential function variable. If the formal
solution for TJ as TJ = -I is substituted back through the Eq. (1) system
of governing equations, one obtains
u'/U = - [1 + n2!] = P'/U2) (12a)
v'/U = - [I + n2!*] (12b)
xy y
and w'/U = -I , (12c)
xz
as a complete description of the flow. If one now integrates Eq. (11)
twice with respect to x and assumes that no f (x) .quantities appear on the
right hand side, one has
T2! + n2 fl + I**!- 0 , (13)
I yyJ
which for n -> 0 is identical 'to the 7 = 0 PDE for the potential function .
Hence, if I is reset to , such that Eq. (12a) becomes
u'/U = - Ux + nV]
for example, then the divergence-free relation (Id) leads immediately to
Janowitz (1984) shows that the delta function hill problem is without
swirl and therefore describable in terms of a potential.
167
-------
+ n2 L + ^1 =0 (14)
Thus, the identical nature of the PDE's for $ and I as given by Eqs.(13)
and (14) could suggest that the solutions given by Eq.(10) might be more
correctly called $ solutions, with I being subsequently computed as I = x.
Finally, the richness of the solution possibilities offered by a
fourth-order PDE should not be underestimated. One approach to evaluating the
integral, Eq.(9), for the simplified case of x' = y' =» 0 leads to the
particular solution.
GCO.O.z') - 5 /I - aJ (a) - 5| fj^ajH (a) - J (a) H.(a)l, (15)
R^ o 2 |_ 1 o o 1J
where a = nz' and whe J and H indicate Bessel and Struve functions
respectively. While quite cumbersome to work with, Eq.(15) has the
2
interesting feature that it has a lowest-order n (rather than n ) dependence.
Such a stronger dependence on the stratification variable n appeared desirable
during efforts to reproduce laboratory, studies.
3. A Related 3-d Problem
The twice integrated PDE given by Eq.(13) seen in a Fourier transform
2. 2 A
sense converts the rightmost term into (rYk )I. In the case of the infinite
field of cosine hills problem (Appendix B), only the wave numbers
(k t) associated with the hill shape survive. In the language of hill
22
length scales (L^, L ) this term becomes (LT/L )I and Eq. (13) can be
rewritten as the Helmholtz equation
V2! + n/2I = 0 (16a)
where n'= n (1 + L2/L2)1/2 , . (16b)
x y
which has the known Green's function solution
168
-------
G = cos(n'R)/R (17)
(plus other solutions) in, three dimensions. As anticipated from the foregoing
discussion, Eq. (17) correctly solves the field of cosine hills problem
and the necessary integrals are discussed in Appendix B. This solution is
quite useful because it is easier to deal with in subsequent integrations and
because it provides a valuable bridge between the three- and two-dimensional
problems. However, a clear shortcoming is that its simple R dependence
suggests isotropy; that is, z has no special significance in the equation,
despite the fact that the density stratification and thus the atmosphere's
"springiness" is a z-oriented phenomenon. Such isotropic "springiness" is,
however, expected for a repetitive field of cosine hills.
It should be noted that for use in Eq. (8b), the Green's function
given by Eq. (17) represents only the real part. For completeness, and
to satisfy the outgoing wave energy constraint, the full complex Green's
function,
G = exp(i n' R)/R , (17a)
should be used In Eq. (8b). The real part is then taken after the
convolution process is complete. The same argument holds in the fewer
dimension problems that follow.
4. Solutions in Fewer Dimensions
As our 3-d hill starts to spread in the crosswind (i.e., ± y) direction,
Q
bility in the y direction drops (i.e., 3— terms get smaller, except
near the hill's y boundaries) until the situation of the infinite crosswind
a2
— •
8y
Q
variability in the y direction drops (i.e., 3— terms get smaller, except
i
a a2
ridge is achieved. In this case all g- and — •= terms vanish and the PDE for
shear-free flow given by Eq. (13) becomes
V?, I + n2! = 0 , (18)
3 *2 a2
, _2 o o
where vl » — - + — ~
2 ax2 3z2
169
-------
Thus, the most troublesome I** term has disappeared and, along with it, any
differences between the Eq. (17) solution, with n' = n (as L -> «), and
rf
whatever the correct particular solution of Eq.(3) or (13) should be.
Eq.(18) is known to have zeroth-order Bessel function solutions, J (nr) and
222
Y (nr) where r = x + z , but it should be possible to obtain the correct 2-d
o
Green's function, G-, by integrating the 3-d Green's function, G^, over y.
It is here where the "dimensional bridge" solution provided by Eq.(17) proves
particularly useful. Using it, one finds
= Jdy G3 » J dy cos(nR)/R
cos(nrp) = -ir Y (nr)
(19)
via the aid of the transformation p = R/r and an integral representation of
Yy found in Abramowitz and Stegun (1972) (pg. 360, Eq. 9.1.24). Writing G_ in
its full complex form as
G_ = i n H(1)(nr) (20)
£ o
where H (x) = J (x) + i Y (x) is the Hankel function, then enables one
o o o
t© proceed easily to the 1-d problem. Integrating over x from -« to +®, one
obtains the one-dimensional Green's function
which corresponds to upward travelling waves resulting from a displacement of
the entire x-y plane.
5. Discussion
Several solutions to the fourth-order PDE for linearized fluid flow over
a 3-d obstacle, as expressed by Eqs.(3) and (11), are found. These solutions
are given by Eq.(10) and have been verified to be exact solutions by the
algebraic processor MACSYMA. Two steps that aided In determining these
170
-------
Green's function solutions involved,
• working with the integrated quantity I, such that T? = - I ,, and
• converting the branch cut to a contour integral over a pole.
The first of these steps actually goes a long way toward restructuring the
integral into a form more closely related to known tabulated integrals.
The candidate solution, Eq.(lOb), is used in connection with the problem
of an infinite 2-d field of cosine hills and found not to be the exact
particular solution. However, comparison with the far-field, wave solution of
Janowitz (1984) suggests that Eq.(lOb) (or actually its derivative with
respect to x) may be appropriate. Unfortunately, the needed integrals to test
this hypothesis are not presently tractable.
A related Helmholtz PDE, Eq.(16), is developed along dimensional
arguments and its solution, Eq. (17), is also found to yield the correct
solution to the above-mentioned, field of cosine hills problem. In addition,
this solution, Eq.(17), provides a convenient bridge to the Green's function
solutions to the stratified flow problem in fewer (i.e., 2- and 1-d)
dimensions.
Acknowledgment
This work was supported as part of U.S. EPA Contract 68-02-3421 to
Environmental Research and Technology, Inc. The author wishes to acknowledge
the project officer, Peter Finkelstein, for his encouragement to pursue this
topic.
171
-------
References
Abramowitz, M. and I.A. Stegun, 1972. Handbook of Mathematical Functions,
National Bureau of Standards, Applied Mathematics Series 55,
Washington, DC, Tenth printing with corrections.
Aiken, H.H. (editor), 1949. Tables of Generalized Sine- and Cosine-Integral
Functions: Part I. Harvard University Press.
•v
Berkshire, F.H., 1985. Two-dimensional linear lee wave modes for models
including a stratosphere, Quart. J.R. Met. Soc., 101. 259-266.
Bois, P.A., 1984. Asymptotic theory of lee waves in an unbounded atmosphere.
Geophys. Astropys. Fluid Dynamics, 29, 267-303.
Crapper, G.D., 1959. A three-dimensional solution for waves in the lee of
mountains. J. Fluid Mech., 6, 51-76.
Gradshteyn, I.S. and I.M. Ryzhik, 1965. Tables of Integrals, Series, and
Products, Fourth Edition. Academic Press, New York.
Hunt, J.C.R., S. Leibovich, andJ.L. Lumley, 1981. Prediction Methods for
the Dispersal of Atmospheric Pollutants in Complex Terrain.
Flow Analysis Associates. Report P85-81-04, Ithaca, NY 14850.
Hunt, J.C.R. and K.J. Richards, 1984. Stratified airflow over one or two
hills. Boundary-Layer Met., 3£, 223-259.
Janowitz, G.S., 1984. Lee waves in three-dimensional stratified flow.
J. Fluid Mech.. 148. 97-108.
Queney, P., 1947. Theory of Perturbations In Stratified Currents with
Applications to Airflow Over Mountain Barriers. Dept. of Met..
Univ. of Chicago Report No. 23, Univ. of Chicago Press. Also
described in Dynamic Meteorology, P. Morel (ed.), D. Reidel Publishing
Co. , Boston, MA 622 pp., 1970.
172
-------
Smith, R. B., 1980. Linear theory of stratified hydrostatic flow past
an isolated mountain. Tellus, 32, 348-364.
Trubnikov, B.N., 1959. The three-dimensional problem of the flow over a
barrier of an air current unbounded at the top. Dokl. of the Acad. of
the U.S.S.R, 129. 4, 781-3 (English translation pgs. 1136-1138).
Wurtele, M., 1957. The three-dimensional lee wave. Beitr. Phys. frel Atmos.,
29, 242-252.
Yamartino, R. J. and R. Pavelle, 1987. An application of MACSYMA to the
fourth-order partial differential equation for linearized, stratified
fluid flow in three-dimensions. Submitted to J. Symbolic Computation.
173
-------
Appendix AA: The 3-d Green's Function Integral
Eq.(9) can be put in a more compact form by first defining
the vectors Q = (k,£,q) and R = (x'.y'.z') and noting that
kx'+£y'+qz'=Q«R =* QR cose'. In spherical coordinates, Eq.(9) then
becomes
00 2tt 1
'.y'.z')- ^ JdQQ2L|d(cose)eiQRcOS9' (A-l)
2* J J \ Q2-n2/cos20
o o —1
where k = Q sine cos# x'= R sin8D cos^0
fl K
t = Q sine sin# y' = R sine_
ii
q = Q cose " z'~ R eos6
-|c
cose' = cose-cosO- + sin6-sin6_ term prevents one from beneficially rotating the
system so that e becomes e' (i.e., this simplicity is offset by the resulting
complexity in the rotated cos# term).
Next, we non-dimensionalize the Q integration by defining a a QR and
a a nr, and then alter the usual spherical integration limits to yield
G = G'/R (A-2a)
w 27t +1E/2
fr r
d
-------
+TT/2 2ir
G' = fe k_de sine e j
471
-71/2 O
The 0 integration is made possible by substituting the complex form
for sine and extending a result given in Groebner and Hofreiter (1973,
pg. 337, 9b) via the analytic continuation I.(ix)=iJ.(x), where I., and
J. are Bessel functions of order 1. The result is
it/2
G' = - | cose. d* • Jl |cosT I (A-4)
* K Jcos*
-it/2 J'
i1/2
where V a -Icos^O,, + sin&e.
f 2 2 2 1
-jcos 0_ + sin 0_ • cos (#-#„) >•
However, because a number of square roots, and thus sign ambiguities, are
involved in the intermediate calculation leading up to Eq. (A-4), it is not
clear that (A-4) represents the only solution for G' .
The <£ integration could not be accomplished analytically, but was
2
instead evaluated to a high degree of accuracy for hundreds of
different values of a = nR, 6 , and #R. In all cases the computations were
consistent with the result
G' = -sin(nRzVd) (A-5)
with d2 a y/2 + z'2, so that
G = -sin(nRzVd)/R (A-6)
becomes the needed result.
2
The integrand was sampled at 50,000 points, judiciously avoiding the
end points at #=±ir/., and the results appeared to be accurate to within
C»,
a few percent.
1/5
-------
Appendix AB: Solution for a Field of Cosine Hills
Consider the case of the infinitely repetitive grouping of cosine hills,
given by the hill function
h(x,y) = h cos(k x)-cosU y), (B-l)
H H
where k = 2it/X and t - 2n/X
H H H y
relate the hill's x,y wavelengths X , X , respectively, to equivalent
x y
wave numbers.
The fact that the actual peak-to-trough height of the hills is 2h is not
of particular concern; however, the relationship between the Fourier transform
and the Dirac delta function proves particularly useful. That is,
ee
1 ,. . -ikx
=- dx cos(k x)e
<£iL I H
eix(kH-k) + e-lx(kH+k)
* -I 5(k -k)
n
(B-2)
where the last line in Eq.(B-2) reflects the fact that there are no
differences between plus and minus k properties of the hill function.
H
the Fourier transform of the hill shape function is simply
) = h 3(k -k)-3U -t) (B-3)
M H
In the sections which follow, the vertical deflection, •»}, will be
computed from both the wavenumber space (i.e., Eq. (B-3)) and coordinate space
(i.e., Eq.(B-l)) representations of the hill.
176
-------
Wavenumber Space Approach
The fundamental defintion of the Dirac delta function as
00
fdk f(k)5(k -k) a f(k), (B-4)
J MM
—00
coupled with the outgoing wave solution of Eq. (5) as
7»(k,£,z) = n(k,i,0) exp(+imz). (B-5)
and the Fourier transform of the linearized (i.e. , z •» z' = height above
terrain) surface boundary condition
), (B-6)
enables one to solve directly for TJ. Guided by Eq. (4), one may then write
09
,£) exp[im(k,£)z'l expJKkx + ty (B-7a)
H,iH)z'l-exp|
= h expim(k,i)z'-expi(kX + t (B-7b)
H
where m(kH^H) = (* + )/k - (k + ) from Eq.(6)
After taking the appropriate real parts, one obtains
Tj(x,y,zy)
h cosm(k,*)z' + kx • cos(«y) f or
k exp/- |m(k ,«u)z'll • cos(kx) • coa(ty) for 0
1 I H H I H M
(B-7c)
The two solutions given by (B-7c) are equivalent for n=k and correspond to
H
reasonant driving of the entire fluid in a z-independent (i.e., as m=0) mode.
Coordinate Space Approach
The simplicity of Eq.(8) seems to suggest that this may be the simpler
177
-------
route, but that does not turn out to be the case. First, it is necessary to
consider the hill shape function in the convolution form
h(x-x', y-y') = h cos|kH (x-x')| • cos^Cy-y') j; (B-8)
cos-U (x-x'H = cos(k x) • cos(k x') + slnCk x) • sin(k x')8
H H H H
however, noting that
3{k (x-x')l
I M J
and that integration from x' = -OB to + w will kill off terms odd in x'
(assuming that the Green's function is purely even in x'), one obtains
D(x,y,z) = -I (B-9a)
with
» Re- |- -cosCk x) • cosU y) • (B-9b)
bTE H H
oa ce
2fdy'cosU y') • 2fdx'cos(k x' ) G(x',y',z')
j H J H
O O
Before proceeding further, a candidate Green's function is required. Using
the exact solution expressed by Eq. (lOb) and noting that (e.g. sea Gradshteyn
and Ryzhik, 1965, pg. 472)
f
x'+d
for 0 < k < p,
where p = nz'/d and d2 = y'2 + z'2,
one obtains,
09
I—Re h(x,y) dy'cosU y') Y f v£i2z'2-k2d2 1. (B-lOa)
j H °l H J
o
Rewriting the argument of the Y Bessel function as
o
k Vz' (n -k2)/k2 - y'2- and referring to the above cited reference (pg. 737,
f 11
involving the Hankel function H '), one finally obtains
178
-------
1/2 f „ „> 1/2 -i r „ „ ^ 1/2
I = -Re h(x.y) • sin{z< [n2-k2] [k2 +
-------
APPENDIX B
EVALUATION RESULTS FOR CONCENTRATIONS
PAIRED IN TIME, UNPAIRED IN SPACE
180
-------
APPENDIX- B
EVALUATION RESULTS FOR CONCENTRATIONS
PAIRED IN TIME, UNPAIRED IN SPACE
Evaluation statistics for data subsets paired in time, not in
space, are presented in tabular form in this appendix. A guide to the
tables is given below. The statistical tests and their results are
discussed further in Section 4.
Table # Description
B-l Evaluation results for SF6 at CCB
B-2 Evaluation results for CF3Br at CCB
B-3 Evaluation results for SF6 at HBR
B-4 Evaluation results for CF3Br at HBR
B-5 Evaluation results for SF$ at FSPS
B-6 Evaluation results for CF3Br at FSPS
B-7 Evaluation results for SO2 at Westvaco,
1-hour averages
B-8 Evaluation results for S02 at Westvaco,
3-hour averages
B-9 Evaluation results for S02 at Widows
Creek, 1-hour averages
B-10 Evaluation results for S02 at Widows
•Creek, 3-hour averages
For the CTMD (tracer) sites, the average of the top 5
concentrations is considered as well as the peak hourly value to
provide a larger sample size. In nearly all cases, all of the top 5
values represent significant plume impacts.
181
-------
TABLE B-l
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
HIGHEST 1-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 100 28.0 26.2 0.94 28.7 1.05 1.12 37
TWO TOWER LEVELS 100 28.0 12.2 0.44 36.9 1.74 4.00 21
ONE TOWER LEVEL 100 28.0 6.6 0.24 37.5 1.80 7.66' 17
COMPLEX I 100 28.0 42.2 1.51 45.6 2.65 1.76 38
RTDM (DEFAULT) 100 28.0 36.4 1.30 58.3 4.35 3.35 27
RTDM (ONSITE) 100 28.0 21.5 0.77 26.1 0.87 1.13 38
AVERAGE OF THE
TOP 5 VALUES
FROM EACH HOUR:
CTDM, SEVERAL
TOWER LEVELS 100 18.4 19.5 1.06 21.4 1.36 1.28 38
TWO TOWER LEVELS 100 18.4 8.5 0.46 24.9 1.83 3.96 19
ONE TOWER LEVEL 100 18.4 5.2 0.28 25=0 1.84 6.49 16
COMPLEX I 100 18.4 33.3 1.81 38.0 4.27 2.36 39
RTDM (DEFAULT) 100 18.4 25.1 1,37 34.1 3.45 2.53 28
RTDM (.ONSITE) 100 18.4 15.3 0.36 17.6 0.92 1.07 36
* Threshold for both observed and predicted concentrations = .00 uS/M**3
182
-------
TABLE B-2
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES TRACER: CF3BR SITE: CINDER CONE BUTTE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
# PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
HIGHEST 1-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 44 15.0 21.0 1.4.0 29.2 3.79 2.71 2-3
TWO TOWER LEVELS 44 15.0 7.5 0.50 19.9 1.76 3.52 14
, ONE TOWER LEVEL 44 15.0 4.6 0.30 20.9 1..95 6.39 .34
COMPLEX I 50 15.1 24.5 1.62 28.7 3.59 2.22 46
RTDM (DEFAULT) 44 15.0 11.0 0.74 24.3 2.62 3.56 50
RTDM (ONSITE) 44 15.0 13.7 0.91 19.3 1.65 1.80 32
AVERAGE OF THE
TOP 5 VALUES
FROM EACH HOUR:
CTDM SEVERAL
TOWER LEVELS 44 8.0 13.8 1.72 20.0 6.19 3.60 16
TWO TOWER LEVELS 44 8.0 4.9 0.62 12.8 2.54 4.12 14
ONE TOWER LEVEL 44 8.0 2.7 0.34 12.6 2.46 7.20 30
COMPLEX I 50 8.4 19.5 2.32 23.4 7.84 3.38 28
RTDM (DEFAULT) 44 8.0 8.7 1.09 17.6 4.79 4.42 45
RTDM (ONSITE) 44 8.0 9.0 1.12 11.9 2.20 1.97 36
* Threshold for both observed and predicted concentrations = .00 uS/M**3
183
-------
TABLE B-3
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES TRACER: SF6 SITE: HOGBACK RIDGE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0..5 <
PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.1
HIGHEST 1-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 59 23.5 47.8 2.04 58.3 6.16 3.03 37
TWO TOWER LEVELS 59 23.5 30.1 1.28 44.3 3.55 2.77 29
ONE TOWER LEVEL 59 23.5 10.8 0.46 29.4 1.57 3.41 17
COMPLEX I 60 23.8 117.5 4,94 12S.1 27.61 5.59 2
RTDM (DEFAULT) 59 23.5 74.7 3.18 159.9 46.35 14.58 14
RTDM (ONSITE) 59 23.5 32.6 1.39 69.3 8.72 6,28 54
AVERAGE OF THE
TOP 5 VALUES
FROM EACH HOUR:
CTDM, SEVERAL
TOWER LEVELS 59 18.3 36.4 1.99 37.6 4.22 2.12 25
TWO TOWER LEVELS 59 18.3 22.6 1.23 29.6 2.60 2.11 27
ONE TOWER LEVEL 59 18.3 9.0 0.49 23.7 1.67 3.40 14
COMPLEX I 60 18.6 101.6 5.47 109«5 34.68 6.34 2
RTDM (DEFAULT) 59 18.3 46.6 2.54 77.0 17.66 6.94 IS
RTDM (ONSITE) 59 18.3 22.3 1.22 31.7 3.00 2.46 54
* Threshold for both observed and predicted concentrations = .00 uS/M**3
184
-------
TABLE B-4
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
HIGHEST 1-HOUR
VALUES:
CTDM, 'SEVERAL
TOWER LEVELS 42 92.0 49.1 0.53 93.1 1.02 1.92 57
TWO TOWER LEVELS 24 118.2 47.3 0.40 144.7 1.50 3.75 38
ONE TOWER LEVEL 13 91.6 58.5 0.64 101.9 1.24 1.94 38
COMPLEX I 40 111.3 173.7 1.56 125.6 1.27 0.82 40
RTDM (DEFAULT) 38 110.6 444.1 4.02 678.3 37.63 9.37 13
RTDM (ONSITE) 61 102.3 142.2 1.39 572.3 31.33 22.52 25
AVERAGE OF THE
TOP 5 VALUES
FROM EACH HOUR:
CTDM, SEVERAL
TOWER LEVELS 42 59.9 40.2 0.67 48.5 0.65 0.98 50
TWO TOWER LEVELS 24 75.1 39.8 0.53 77.4 1.06 2.01 38
ONE TOWER LEVEL 13 58.0 51.7 0.89 52.2 0.81 0.91 62
COMPLEX I 40 73.0 151,0 2.07 112.3 2.37 1.14 40
RTDM (DEFAULT) 37 67.7 179.7 2.65 218.9 10.45 3.94 3
RTDM (ONSITE) 61 63.7 52.9 0.83 145.1 5.18 6.24 20
* Threshold for both observed and predicted concentrations » .01 uS/M**3
185
-------
TABLE B-5
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES TRACER: SF6 SITE: TRACY POWER PLANT
(Concentrations* given in units of microseconds per cubic meter.)
% CASES;
0.5 <
# _ _ PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
HIGHEST 1-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 110 1.96 1,94 0.99 2.11 1.16 1.17 61
(ALT. PLUME HT 1)** 110 1.96 1.92 0.98 1.84 0.88 0.90 68
(ALT. PLUME HT 2)** 110 1.96 1.77 0.90 1.88 0.92 1.02 69
TWO TOWER LEVELS 109 1.94 2.07 1.07 2.24 1.33 1.25 49
ONE TOWER LEVEL 105 1.94 2.41 1.24 3.08 2.52 2.03 38
COMPLEX I 111 1.95 6.14 3.15 7.27 13.90 4.41 19
RTDM (DEFAULT) 111 1.95 3.05 1.56 3.38 3.00 1.92 48
RTDM (ONSITE) 110 1.96 1.18 0.60 2.22 1.28 2.13 34
AVERAGE OF THE
TOP 5 VALUES
FROM EACH HOUR:
CTDM, SEVERAL
TOWER LEVELS 110 1.29 1.36 1.05 1.15 0.80 0.75 55
TWO TOWER LEVELS 109 1.28 1.36 1.06 1.33 1.08 1.02 49
ONE TOWER LEVEL 105 1.27 1.36 1.07 1.95 2.36 2.20 37
COMPLEX I HI 1.28 4ol4 3.23 4.83 14.24 4.40 16
RTDM (DEFAULT) 111 1.28 1.87 1.46 1.86 2.11 1.45 44
RTDM (ONSITE) 110 1.29 0.76 0.59 1.33 1.06 1.80 37
* Threshold for both observed and predicted concentrations = .00 uS/M**3
** Alternative plume height #1 was obtained from lidar measurements
at the first cross section downwind from the source. Plume
height #2 was obtained from the second lidar cross section.
186
-------
TABLE B-6
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES TRACER: CF3BR SITE: TRACY POWER PLANT
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
# PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
HIGHEST 1-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 110 2.84 2.46 0.87 2.94 1.07 1.24 60
TWO TOWER LEVELS 109 2.84 3.02 1.06 4.01 1.99 1.88 53
ONE TOWER LEVEL . 105 2.84 3.04 1.07 4.99 3.09 2.88 31
COMPLEX I 111 2.34 8.54 3.01 9.14 10.36 3.44 23
RTDM (DEFAULT) 111 2.84 3.76 1.32 4.79 2.85 2.15 43
RTDM (ONSITE) 110 2.84 1.95 0.69 3.17 1.25 1.82 52
AVERAGE OF THE
TOP 5 VALUES
FROM EACH HOUR:
CTDM, SEVERAL
TOWER LEVELS 110 1.74 1.30 0.75 1.45 0.69 0.93 66
TWO TOWER LEVELS 109 1.73 1.24 0.72 1.54 0.79 1.11 57
ONE TOWER LEVEL 105 1.73 1.02 0.59 2.04 1.39 2.36 32
COMPLEX I 111 1.73 5.27 3.05 . 5.55 10.2? 3.38 30
RTDM (DEFAULT) 111 1.73 1.74 1.01 1.83 1.12 1.11 46
RTDM (ONSITE) 110 1.74 1.15 0.66 1.55 0.79 1.20 57
Threshold for both observed and predicted concentrations = .00 uS/M**3
187
-------
TABLE B-7
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES TRACER: SO2 SITE: WESTVACO LUKE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
HIGHEST 1-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 4687 0.33 Q.29 0.88 -0.83 6.3 7,2 14
TWO TOWER LEVELS 4702 0.33 1.60 4.85 3.30 100.0 20.6 19
ONE TOWER LEVEL 4022 0.34 1.70 5.00 2.00 34.6 6.9 15
COMPLEX I 4687 0.33 4.12 12.48 9.44 818.3 65.5 14
RTDM (DEFAULT) 4687 0.33 0.98 2.97 2.46 55.6 18.7 16
RTDM (ONSITE) 4687 0.33 0.21 0.64 0.82 6.2 9.7 15
* Threshold for both observed and predicted concentrations - .00 uS/M**3
188
-------
TABLE B-8
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
3-HOUR AVERAGES TRACER: SO2 SITE: WESTVACO LUKE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
# PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
HIGHEST 3-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 1318 0.30 0.27 0.90 0.62 4.3 4.7 17
TWO TOWER LEVELS 1322 0.30 1.57 5.23 2.89 92.8 17.7 18
ONE TOWER LEVEL 1136 0.32 1.65 5.16 1.82 32.3 6.3 12
COMPLEX I 1318 0.30 3.98 13.27 7.96 704.0 53.1 9
RTDM (DEFAULT) 1318 0.30 0.95 3.17 2.03 45.8 14.5 11
RTDM (ONSITE) 1318 0.30 0.21 0.70 0.64 4.6 6.5 18
* Threshold for both observed and predicted concentrations - .00 uS/M**3
189
-------
TABLE B-9
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
1-HOUR AVERAGES, 1980 TRACER: SO2 SITE: WIDOWS CREEK
(Concentrations* given in units of micrograms per cubic meter.)
% CASES:
0.5 <
# ____ ____ _ PRE/OBS
HOURS OBS -PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
HIGHEST 1-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 4942 71.9 142 1.97 404 31.5. 16.0 23
TWO TOWER LEVELS 4809 70=8 152 2.15 576 66.1 30.8 12
ONE TOWER LEVEL 4131 70.3 390 5.55 767 118.9 21.4 16
COMPLEX I 5065 71.1 413 5.80 1239 303.4 52.3 13
RTDM (DEFAULT) 5065 71.1 322 4.53 765 115.5 25.5 7
RTDM (ONSITE) 5065 71.1 48 0.67 299 17.6 26.3 15
* Threshold for both observed and predicted concentrations = .00 uG/M**3
190
-------
TABLE B-10
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
3-HOUR AVERAGES, 1980 TRACER: SO2 SITE: WIDOWS CREEK
(Concentrations* given in units of micrograms per cubic meter.)
% CASES:
0.5 <
# PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
HIGHEST 3-HOUR
VALUES:
CTDM, SEVERAL
TOWER LEVELS 1370 66.4 112. 1.69 230. 12.0 7.1 31
TWO TOWER LEVELS 1321 65.4 127. 1.94 323. 24.4 12.6 17
ONE TOWER LEVEL 1089 64.3 342. 5.33 579. 81.2 15,2 20
COMPLEX I 1423 65.4 349. 5.34 828. 160.3 30.0 15
RTDM (DEFAULT) 1423 65.4 279. 4.26 521. 63.6 14.9 9
RTDM (ONSITE) 1423 65.4 41. 0.63 169. 6.7 10.6 20
* Threshold for both observed and predicted concentrations - .00 uS/M**3
191
-------
APPENDIX C
EVALUATION RESULTS FOR CONCENTRATIONS
PAIRED IN SPACE, UNPAIRED IN TIME
192
-------
APPENDIX C
EVALUATION RESULTS FOR CONCENTRATIONS
PAIRED IN SPACE, UNPAIRED IN TIME
Statistics for the evaluation data subset paired in space, not in
time are presented here. These tests and their results are discussed
in more detail in Section 4. A guide to the tables in this appendix
is given below.
Table it Description
C-l Evaluation results for SF6 at CCB
C-2 Evaluation results for CF3Br at CCB
C-3 Evaluation results for SF6 at HBR
C-4 Evaluation results for CF3Br at HBR
C-5 Evaluation results for SFg at FSPS
C-6 Evaluation results for CF3&r at
FSPS
• C-7 Evaluation results for S02 at
Westvaco, 1-hour averages
C-8 Evaluation results for S02 at
Westvaco, 3-hour averages
C-9 Evaluation results for S02 at
Widows Creek, 1-hour averages
C-10 Evaluation results for S02 at
Widows Creek, 3-hour averages
The number of monitoring sites in the evaluation sample are
listed in each table. The average over the top 5 events for the
tracer sites and the top 10 events at the S(>2 sites were computed to
provide a larger evaluation sample then just the highest concentration
event at each receptor.
193
-------
TABLE C-l
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES TRACER? SF6 SITE; CINDER CONE BUTTE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES;
0.5 <
# _ _ _ PRE/OBS
SITES OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 93 46.37 51.08 1.10 37.11 0.64 0.58 67
TWO TOWER LEVELS 93 46.37 30.47 0.66 29.92 0.42 0.63 65
ONE TOWER LEVEL 93 46.37 19.03 0.41 39.18 0.71 1.74 33
COMPLEX I 93 46.37 84.34 1.82 51.80 1.25 0.69 51
RTDM (DEFAULT) 93 46.37 77.37 1.67 56.03 1.46 0.88 56
RTDM (ONSITE) 93 46.37 36.87 0.80 27.54 0.35 0.44 72
AVERAGE OF THE
TOP 5 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 93 27.26 28.52 1.05 17.68 0.42 0.40 71
TWO TOWER LEVELS 93 27.26 14.42 0.53 18.65 0.47 0.89 57
ONE TOWER LEVEL 93 27.26 9.58 0.35 23.39 0.74 2.10 23
COMPLEX I 93 27.26 55.36 2.03 35.59 1.71 0.84 39
RTDM (DEFAULT) 93 27.26 43.23 1.59 31.22 1.31 0.83 67
RTDM (ONSITE) 93 27.26 2.1,20 0.78 13.38 '0.24 0.31 76
Threshold for both observed and predicted concentrations = .00 uS/M**3
194
-------
TABLE C-2
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES TRACER: CF3BR SITE: CINDER CONE BUTTE
(Concentrations* given in units of microseconds per cubic meter.)
_ __ _ _ _
SITES OBS PRE PRE/OBS RMS V/Co2 V/(CoCp)
CASES:
0.5 <
PRE/OBS
< 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
AVERAGE OF THE
TOP 5 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
80 15.18 27.03 1.78 29.50 3.78
80 15. .18 11.65 0.77 17.89 1.39
80 15.18 7.13 0.47 18.31 1.46
81 15.59 35.02 2.25 36.24 5.40
80 15.18 18.69 1.23 27.86 3.37
80 15.18 14.74 0.97 19.37 1.63
93 4.53 8.50
93 4.53 3.31
93 4.53 1.78
93 5.18 13.12
93 4.53 5.41
93 4.53 5.09
1.88 10.43 5.30
0.73 5.68 1.57
0.39 6.05 1.78
2.53 13.04 6.34
1.19 7.36 2.64
1.12 6.35 1.97
2.12
1.81
3.10
2.41
2.74
1.68
2.83
2.15
4.54
2.50
2.21
1.75
25
33
24
32
34
36
34
45
34
34
46
47
* Threshold for both observed and predicted concentrations = .00 uS/M**3
195
-------
TABLE C-3
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES TRACER? SF6 SITES HOGBACK RIDGE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
Oo5 <
# ____ __ __ _ PRE/OBS
SITES OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS. 106 35,86 61.65 1.72 62.77 3.06 1.78 58
TWO TOWER LEVELS 106 35.86 49.84 1.39 41.19 1.32 0.95 67
ONE TOWER LEVEL 106 35.86 32,17 0.90 25.52 0.51 0.57 61
COMPLEX I 106 36.04 127.85 3.55 132.13 13.44 3.79 14
RTDM (DEFAULT) 106 35.86 111.08 3.10 157.67 19.33 6.24 32
RTDM (ONSITE) 106 35.86 36.13 1.01 57.68 2.59 2.57 43
AVERAGE OF THE
TOP 5 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 106 22.17 36.22 1.63 28.64 1.67 1.02 61
TWO TOWER LEVELS 106 22.17 29.10 1.31 21.83 0.97 0.74 59
ONE TOWER LEVEL 106 22.17 15.75 0.71 16.34-0.54 0.77 53
COMPLEX I 106 22.44 79.19 3.53 82.59 13.55 3.84 9
RTDM (DEFAULT) 106 22.17 40.01 1.81 46.29 4.36 2.42 36
RTDM (ONSITE) 106 22.17 19.09 0.36 17.26 0.61 0.70 53
Threshold for both observed and predicted concentrations = .00 uS/M**3
196
-------
TABLE C-4
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES TRACER: CF3BR SITE; HOGBACK RIDGE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
# _ _ _ __ _ __ PRE/O'BS
SITES OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
AVERAGE OF THE
TOP 5 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
106 122.9 43.4 0.35 104.6 0.72
106 122.9 39.0 0.32 111.2 0.82
105 121.4 28.8 0.24 114.0 0.88
106 122.9 190.6 1.55 118.1 0.92
104 123.9 271.1 2.19 381.4 9.48
2.05
2.53
3.72
0.60
4.33
106 122.9 120.8 0.98 400.7 10.63 10.32
106 73.2 27-7 0.38 55.82 0.58 1.54
106 73.2 21.9 0.30 62.79 0.74 2.46
105 -72.8 14.9 0.21 68.22 0.88 4.28
106 73.2 123.3 1.68 85.32 1.36 0.81
102 74.6 91.4 1.23 114.05 2.34 1.91
106 73.2 38.6 0.53 98.73 1.82 3.45
14
22
15
41
32
19
19
12
11
38
49
25
* Threshold for both observed and predicted concentrations = .01 uS/M**3
197
-------
TABLE C-5
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES _ TRACER: SF6 SITE: TRACY POWER PLANT
(Concentrations* given in units of microseconds per cubic meter.)
% CASESs
0.5 <
# _ _ __ PRE/OBS
SITES OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 106 2.19 1.80 0.82 2.25 1.06 1.28 40
(ALT. PLUME HT 1)* 106 2.19 1.67 0.76 2.02 1.46 1.12 42
(ALT. PLUME HT 2)* 106 2.19 1.68 0.77 2.09 1.55 1.19 43
TWO TOWER LEVELS 106 2.18 1.72 0.79 2.36 1.17 1.49 37
ONE TOWER LEVEL 106 2.15 2.08 0.97 3.57 2.76 2.85 25
#
COMPLEX I 106 2.19 5.95 2=72 7.10 10.51 3.87 23
RTDM (DEFAULT) ioe 2.19 2.80 i.as 3.50 2.55 2.00 30
RTDM (ONSITE) 106 2.19 1.25 0.57 2.17 0.98 1.72 43
AVERAGE OF THE
TOP 5 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 106 1.32 1.05 0.80 1.04 0.62 0.78 41
TWO TOWER LEVELS 106 1.30 1.03 0.79 1.20 0.85 1.08 36
ONE TOWER LEVEL 106 1.27 1.11 0.87 1.97 2.41 2.75 21
COMPLEX I 106 1.32 4.02 3.05 4.83 13.39 4«40 13
RTDM (DEFAULT) 106 1.32 1.66 1.26 1.89 2.05 1.63 31
RTDM (ONSITE) 106 1.32 0.76 0.58 1.09 0.68 1.18 47
Threshold for both observed and predicted concentrations =» .00 uS/M**3
* Alternative plume height #1 was obtained from lidar measurements
at the first cross section downwind from the source. Plume
height #2 was obtained from the second lidar cross section.
198
-------
TABLE C-6
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES TRACER: CF3BR SITE: TRACY POWER PLANT
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
# _ _ PRE/OBS
SITES OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 106 2.70 1.49 0.55 2.93 1.18 2.13 41
TWO TOWER LEVELS 106 2.70 1.64 0.61 4.21 2.43 4.00 32
ONE TOWER LEVEL 106 2.69 1.66 0.62 4.99 3.44 5.58 20
COMPLEX I 106 2.70 6.89 2.55 7.50 7.72 3.02 13
RTDM (DEFAULT) 106 2.70 2.40 0.89 3.47 1.65 1.86 37
RTDM (ONSITE) 106 2.70 1.58 0.59 2.86 1.12 1.92 53
AVERAGE OF THE
TOP 5 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 106 1.57 0.93 0.59 1.34 0.73 1.23 37
TWO TOWER LEVELS 106 1.57 0.89 0.57 1.83 1.36 2.40 25
ONE TOWER LEVEL 106 1.55 0.78 0.50 2.15 1.92 3.82 15
COMPLEX I 106 1.58 4.62 2.92 5.52 12.21 4.17 11
RTDM (DEFAULT) 106 1.58 1.36 0.86 1.99 1.59 1.84 35
'RTDM (ONSITE) 106 1.57 1.01 0.64 1.22 0.60 0.94 58
* Threshold for both observed and predicted concentrations » .00 uS/M**3
199
-------
TABLE C-7
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES TRACER: S02 SITE: WESTVACO LUKE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
# PRE/OBS
SITES OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 11 4.77
TWO TOWER LEVELS 11 4«77
ONE TOWER LEVEL 11 4*61
COMPLEX I 11 4.77
RTDM (DEFAULT) 11 4.77
RTDM (ONSITE) 11 4.77
AVERAGE OF THE
TOP 10 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 11 3.05
TWO TOWER LEVELS 11 3.05
ONE TOWER LEVEL 11 2.95
COMPLEX I 11 3.05
RTDM (DEFAULT) 11 3.05
RTDM (ONSITE) 11 3.05
6.48 1.36 2.86
10.09 2.12 8.71
5.71 1.24 2.03
0.36
3.33
0.19
23.97 5.03 25.26 28.04
7.44 1.56 4.78 1.00
5.80 1.22 3.16 0.44
3.97 1.30 1.56 0.26
7.20 2.36 7.63 6.26
4.15 1.41 1.97 0.45
20.13 6.60 22.51 54.47
6.04 1.98 4.67 2.34
3.10 1.02 1.17 0.15
0.27
1.58
0.16
5.58
0.64
0.36
0.20
2.65
0.32
8.25
Iol8
0.15
64
45
82
9
45
73
73
45
73
9
55
91
* Threshold for both observed and predicted concentrations = .00 uS/M**3
200
-------
TABLE C-8
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
3-HOUR AVERAGES TRACER: SO2 SITE: WESTVACO LUKE
(Concentrations* given in units of microseconds per cubic meter.)
% CASES:
0.5 <
# PRE/OBS
SITES OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 11 2.76 3.27 1.19 1.41 0.26 0.22 64
TWO TOWER LEVELS 11 2.76 6.03 2.19 7.19 6.79 3.11 55
ONE TOWER LEVEL .11 2.76 3.34 1.21 1.61 0.34 0.28 73
COMPLEX I 11 2.76 17.38 6.30 19.34 49.10 7.80 0
RTDM (DEFAULT) 11 2.76 5.07 1.84 3.84 1.94 1.05 45
RTDM (ONSITE) 11 2.76 3.09 1.12 1.79 0.42 0.38 73
AVERAGE OF THE
TOP 10 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS 11 1.63 1.69 1.04 0.69 0.18 0.17 73
TWO TOWER LEVELS 11 1.63 4.07 2.-50 5.18 10.10 4.05 55
ONE TOWER LEVEL 11 1.62 2.57 1.59 1.57 0.94 0.59 65
COMPLEX I 11 1.63 12.17 7.47 14.59 8Q.12 10.73 0
RTDM (DEFAULT) 11 1.63 3.63 2.23 3.30 4.10 1.84 45
RTDM (ONSITE) 11 1.63 1.48 0.91 0.67 0.17 0.19 82
* Threshold for both observed and predicted concentrations = .00 uS/M**3
201
-------
TABLE C-9
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
1-HOUR AVERAGES, 1980 TRACERS SO2 SITE: WIDOWS CREEK
(Concentrations* given in units of micrograms per cubic meter.)
SITES OBS PRE PRE/OBS RMS V/CO2 V/(CoCp)
k CASES:
0.5 <
PRE/OBS
< 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER- LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
AVERAGE OF THE
TOP 10 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
14 1695 3720
14 1695 4142
14 1609 2822
14 169S 4953
14 1695 2189
14 1695 3700
2.20 2830 2.79
2.44 3577 4.46
1.75 2433 2.29
2.92 5033 3.82
1.29 1911 1.27
2.18 3727 4<,84
1.27
1.82
1.30
3,02
0.93
2.22
43
43
43
36
36
43
14
14
14
14
14
14
740
724
645
740
740
740
1886
2456
2079
3919
1914
994
2.
3.
3.
5.
2.
1.
55
39
22
30
59
34
1548
2304
1940
4582
1762
748
4.
10.
9.
38.
5.
1.
38
14
04
36
67
02
1
2
2
7
2
0
.72
.99
.80
.24
o!9
.76
50
14
14
14
36
50
* Threshold for both observed and predicted concentrations = .00 uG/M**3
202
-------
TABLE C-10
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN SPACE
3-HOUR AVERAGES, 1980 TRACER: SO2 SITE: WIDOWS CREEK
(Concentrations* given in units of micrograms per cubic meter.)
SITES OBS PRE PRE/OBS RMS V/Co2 V/(CoCp)
It CASES:
0.5 <
PRE/OBS
< 2.0
AVERAGE OF
HIGHEST VALUES
FOR EACH MONITOR:
CTDM) SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
AVERAGE OF THE
TOP 10 VALUES
OBTAINED FOR
EACH MONITOR:
CTDM, SEVERAL
TOWER LEVELS
TWO TOWER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
14
14
14
14
14
14
840
840
725
840
840
840
1454
1288
1629
2859
1476
1137
1.
1.
2.
3.
1.
1.
73
53
25
40
76
35
1058
843
1409
3113
1296
1239
1.
1.
3.
13.
2.
2.
59
01
77
74
38
18
0.
0.
1.
4.
1.
1.
92
66
68
04
36
61
50
43
29
36
50
43
14
14
14
14
14
14
324
319
279
325
325
325
719
752
1139
1755
1001
341
2,
2.
4.
5.
3.
1.
22
36
09
40
08
05
559
637
1130
2089
992
294
2
3
16
41
9
.98
.98
.48
.38
.34
.82
1
1
4
7
3
.34
.69
.03
.66
.03
.78
57
36
7
14
36
43
* Threshold for both observed and predicted concentrations = .00 uG/M**3
203
-------
APPENDIX 0
EVALUATION RESULTS FOR CONCENTRATIONS
PAIRED IN TIME AND SPACE
204
-------
APPENDIX D
EVALUATION RESULTS FOR CONCENTRATIONS
PAIRED IN TIME AND SPACE
Evaluation statistics for data for all hours and monitoring sites
- paired in time and space - are given in this appendix. Further
discussions can be found in Section 4. A guide to the table showing
the results is listed below.
Table # Description
D-l Evaluation results for SFg at CCB
D-2 Evaluation results for CF3Br at CCB
D-3 Evaluation results for SF6 at HBR
D-4 Evaluation results for CF3Br at HBR
D-5 Evaluation results for SF6 at FSPS
D-6 Evaluation results for CF3Br at FSPS
D-7 Evaluation results for S02 at Westvaco,
1-hour averages
D-8 Evaluation results for SO2 at Westvaco,
3-hour averages
D-9 Evaluation results for S02 at Widows
Creek, 1-hour averages
D-10 Evaluation results for S02 at Widows
Creek, 3-hour averages
These tables contain results for two concentration
thresholds. The zero threshold retains all cases, while the nonzero
threshold must be exceeded by both the prediction and the observation
to be included in the statistics. The nonzero threshold effectively
deletes the uninteresting zero versus zero matchup.
205
-------
TABLE D-l
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
(Concentrations given in units of microseconds per cubic meter.)
% CASES:
| 0.5 <
SITE- __ ____ _ PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD -
0.0 US/M**3
CTDM,SEVERAL
TOWER LEVELS 3683 5.87 5.74 0.98 14.2 5.83 5.96 20
TWO TOWER LEVELS 3683 5.87 2.52 0.45 14.1 5.73 12.74 14
ONE TOWER LEVEL 3683 5.87 1.62 0.28 13.2 5*07 18.38 15
COMPLEX I 3683 5.87 11.96 2.04 -23.7 16.26 7.98 25
RTDM (DEFAULT) 3683 5.87 8.60 1.47 21.7 13.65 9.32 25
RTDM (ONSITE) 3683 5.87 4.95 0.84 12.2 4.33 5.13 25
2) THRESHOLD -
0.01 US/M**3
CTDM, SEVERAL
TOWER LEVELS 2079 9.34 8.99 0.96 17.2 3.38 3.51 30
TWO TOWER LEVELS 1584 9.44 4.89 0.52 16.3 2.99 5-77 19
ONE TOWER LEVEL 1214 9,,71 4.19 0.43 15.3 2.49 5.76 19
COMPLEX I 1351 10.83 27.57 2.55 33.5 9.59 3.77 29
RTDM (DEFAULT) 1329 10.42 20.35 1.95 31.1 8.91 4.56 28
RTDM (ONSITE) 2203 8.82 7.05 0.80 14.2 2.59 3.24 29
206
-------
TABLE D-2
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR AVERAGES TRACER: CF3BR SITE: CINDER CONE BUTTE
(Concentrations given in units of microseconds per cubic meter.)
f
SITE-
HOURS OBS
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD -
0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS 566 3.90
TWO TOWER LEVELS 566 3.90
ONE TOWER LEVEL 566 3.90
PRE PRE/OBS RMS
% CASES:
0.5 <
_ PRE/OBS
V/C02 V/(CoCp) < 2.0
7.39 1.90 16.3 17.49 9.23
2.77 0.71 10.5 7.29 10.26
1.48 0.38 10.0 6.55 17.26
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
565 3.90 11.30 2.90 18.9 23.49 8.10
566 3.90 4.87 1.25 13.4 11.34 9.48
566 3.90 4.69 1.20 11.2 8.25 6.86
9
13
29
40
42
22
2) THRESHOLD -.
0.01 US/M**3
CTDM, SEVERAL
TOWER LEVELS 208 8.79 12.16 1.38 22.1
TWO TOWER LEVELS 164 9.69 5.36 0.55 15.9
ONE TOWER LEVEL 123 9.90 3.63 0.37 16.3
COMPLEX I 221 8.55 18.08 2.12 21.7
RTDM (DEFAULT) 182 8.51 8.65 1.02 18.9
RTDM (ONSITE) 225 8.41 6.37 0.76 13.8
6.30
2.71
2.70
6.41
4.91
2.70
4.56
4.89
7.37
.3.03
4.83
3.56
21
18
28
32
41
26
207
-------
TABLE D-3
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR AVERAGES TRACER: SF6 SITE5 HOGBACK RIDGE
(Concentrations given in units of microseconds per cubic meter.)
#
SITE-
HOURS OBS
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD -
0.0 uS/M**3
CTDM, SEVERAL
TOWER LEVELS 4714 5.19
TWO TOWER LEVELS 4714 5.19
ONE TOWER LEVEL 4714 -5.19
PRE PRE/OBS RMS
% CASES:
0.5 <
PRE/OBS
V/C02 V/(CoCp) < 2.0
8o4§ 1,63 17.9 11.86 7.28
4.60 0.39 14.8 S.18 9.22
1.94 0.37 11.8 5.20 13.90
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
4792 S.25 12.83 2.44 37.5 50.94 20,84
4714 5.19 4.75 0.92 29.0 31.29 34.19
4714 5.19 4.30 0.83 13.2 6.47 7.81
19
9
4
2
4
2) THRESHOLD »
0.01 uS/M**3
CTDM, SEVERAL
TOWER LEVELS 3271 5.66 12.17 2.15 20.7 13.31 6.19 26
TWO TOWER LEVELS 1823 7.14 11.87 1.66 21.6 9.18 5.52 20
ONE TOWER LEVEL 957 4.62 9.54 2.07 16.4 12.63 6,12 17
COMPLEX I 903 7.64 68.08 8.91 84.3 121.61 13.65 6
RTDM (DEFAULT) 743 7.91 30.13 3.81 70.1 78.56 20.63 18
RTDM (ONSITE) 3162 5.98 6.39 1.07 14.8 6.14 5.75 29
208
-------
TABLE D-4
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR 'AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
(Concentrations given in units of microseconds per cubic meter.)
#
SITE-
HOURS OBS
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD *
0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS 5026 15.31
TWO TOWER LEVELS 5026 15.31
ONE TOWER LEVEL 5026 15.31
PRE PRE/OBS RMS
% CASES:
0.5 <
PRE/OBS
V/C02 V/(CoCp) < 2.0
6.55 0.43 32.1
3.22 0.21 33.0
1.86 0.12 34.7
4.39 10.27
4.64 22.05
5.15 42.38
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
5175 14.96 14.68 0.98 53.9 13.00 13.25
5175 14.96 9.16 0.61 80.3 23.78 47.00
5026 15.31 5.61 0.37 72.7 22.55 61.54
22
15
13
12
12
21
2) THRESHOLD -
0.01 US/M**3
CTDM, SEVERAL
TOWER LEVELS 2729 16.34 11.51 0.68 26.6 2.50
TWO TOWER LEVELS 1366 21.02 11.53 0.55 38.9 3.42
ONE TOWER LEVEL 613 15.96 14.64 0.92 32.2 4.08
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
557 19.29 134.60 6.98 129.8 45.26
3.66
6.24
4.44
6.49
367 17.83 128.44 7.20 274.1 236.36 32.81
2333 16.56 11.80 0.71 99.4 36.05 50.59
32
26
21
6
12
29
209
-------
TABLE D-5
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR AVERAGES TRACER? SF6 SITESTRACY POWER PLANT
(Concentrations given in units of. microseconds per cubic meter.)
% CASES;
* 0.5 <
SITE- _ PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD - 0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS 9713 0.19 0.13 0.68 0.59 9.64 14.09 7
(ALT. PLUME HT 1)* 9713 0.19 0.13 0.68 0.58 9.32 13.62 7
(ALT. PLUME HT 2)* 9713 0.19 0..12 0.63 0.58 9.32 14.75 7
TWO TOWER LEVELS 9633 0.19 0.12 0.63 0.61 10.31 16.32 5
ONE TOWER LEVEL 9291 0.19 0.09 0.47 0.75 15.58 32.90 2
COMPLEX I 9806 0.19 0.36 1.89 1.61 71.80 37.90 2
RTDM (DEFAULT) 9806 0.19 0.15 0.79 0.82 18.63 23.59 3
RTDM (ONSITE) '9713 0.19 0.08 0.42 0.53 7-78 18.48 6
2) THRESHOLD -0.01 US/M**3
CTDM, SEVERAL
TOWER LEVELS 2366 0.40 0.54 1.35 0.98 6.00 4.45 27
TWO TOWER LEVELS 2010 0.35 0.56 1.60 1.03 8.66 5.41 25
ONE TOWER LEVEL 918 0.33 0.94 2.85 1.89 32.80 11.52 19
COMPLEX I 1309 0.33 2.66 8.06 4.21 162.76 20.19 18
RTDM (DEFAULT) 1126-0.34 1.30 3.82 2.00 34.60 9.05 25
RTDM (ONSITE) 2329 0.37 0.34 0.92 0.76 4.22 4.59 25
* Alternative plume height #1 was obtained from lidar measurements
at the first cross section downwind from the source. Plume
height #2 was obtained from the second lidar cross section.
210
-------
TABLE D-6
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR AVERAGES TRACER: CF3BR SITE:TRACY POWER PLANT
(Concentrations given in units of microseconds per cubic meter.)
% CASES:
# 0.5 <
SITE- _ _ PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD -
0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS 9713 0.23 0.11 0.39 0.72 6.61 16.83 6
TWO TOWER LEVELS 9633 0.27 0.09 0.33 0.82 9.22 27.67 5
ONE TOWER LEVEL 9291 0.27 0.06 0.22 0.89 10.87 48.90 1
COMPLEX I 9806 0.28 0.42 1.50 1.93 47.51 31.67 2
RTDM (DEFAULT) 9806 0.28 0.12 0.43 0.94 11.27 26.30 2
RTDM (ONSITE) 9713 0.28 0.11 0.39 0.70 6.25 15.91 6
2) THRESHOLD -
0.01 US/M**3
CTDM, SEVERAL
TOWER LEVELS 2067 0.59 0.52 0.88 1.27 4.63 5.26 30
TWO TOWER LEVELS 1524 0.42 0.57 1.36 1.47 12.25 9.03 29
ONE TOWER LEVEL 556 0.48 1.02 2.13 2.62 29.79 14.02 20
COMPLEX I 1256 0.53 3.27 6.17 5.15 94.42 15.30 18
RTDM .(DEFAULT) 1036 0.55 1.17 2.13 2.22 16.29 7.66 23
RTDM (ONSITE) 2269 0.53 0.48 0.91 1.15 4.71 5.20 26
211
-------
TABLE D-7
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR AVERAGES TRACER: SO2 SITE? WESTVACO LUKE
(Concentrations given in units of microseconds per cubic meter.)
% CASES:
# 0.5 <
SITE- PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD -
0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS 46092 0.09 0=05. 0.56 0.38 17.8 32.1 40
TWO TOWER LEVELS46283 0.09 0.24 0.67 1.20 177.8 66.7 38
ONE TOWER LEVEL 39838 0.10 0.32 3.20 0.86 74.0 23.1 32
COMPLEX I 46092 0.09 0.52 5.78 3.34 1377.2 238.4 55
RTDM (DEFAULT) 46092 0.09 0.13 1.44 0*92 104.5 72.3 55
RTDM (ONSITE) 46092- 0.09 0.04 0.44. 0.36 16.0 36.0 45
2) THRESHOLD -
0.01 US/M**3
CTDM, SEVERAL
TOWER LEVELS 2865 0.53 0.60 1.13 1.18 5.0 4.4 26
TWO TOWER LEVELS 5119 0.35 1.39 3.97 2.93 70.1 17.6 23
ONE TOWER LEVEL 6796 0.28 1.10 3.93 1,44 26.4 6.7 21
COMPLEX I 1006 0.34 16.94 49.82 18.90 3090=1 62.0 1
RTDM (DEFAULT) 1668 0.36 2.49 6.92 3.86 115.0 16.6 7
RTDM (ONSITE) 3477 0.34 0.35 1.03 0.83 6.0 5.3 26
212
-------
TABLE D-8
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
3-HOUR AVERAGES TRACER: S02 SITE: WESTVACO LUKE
(Concentrations given in units of microseconds per cubic meter.)
% CASES:
# 0.5 <
SITE- , _ PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/Co2 V/(CoCp) < 2.0
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD -
0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS • 12969 0.09 0.05 0.56 0.28 9.7 17.4 32
TWO TOWER LEVELS13020 0.09 0.26 2.89 1.05 136.1 47.1 30
ONE TOWER LEVEL 11252 0.10 0.33 3.30 0.78 60.8 18.4 25
COMPLEX I 12969 0.09 0.55 6.11 2.81 974.8 159.5 43
RTDM (DEFAULT) 12969 0.09 0.14 1.56 0.76 71.3 45.8 44
RTDM (ONSITE) 12969 0.09 0.04 0.44 0.30 11.1 25.0 35
2) THRESHOLD »
0.01 US/M**3
CTDM, SEVERAL
TOWER LEVELS 1344 0.39 0.40 1.03 0.69 3.1 3.1 32
TWO TOWER LEVELS 2231 0.26 1.14 4.39 2.27* 76.2 17.4 25
ONE TOWER LEVEL 2753 0.22 0.95 4.32 1.25 32.3 7.5 20
COMPLEX I 597' 0.27 10.24 37.93 12.18 2035.0 53.7 2
RTDM (DEFAULT) 765 0.29 1.98 6.83 2.79 92.6 13.6 9
RTDM (ONSITE) 1542 0.27 0.27 1.00 0.59 4.8 4-. 8 29
213
-------
TABLE D-9
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
1-HOUR AVERAGES, 1980 TRACER: S02 SITE? WIDOWS CREEK
(Ceneentrations given in units of micrograms per cubic meter.)
% CASES;
I 0.5 <
SITE- _____ _____ _ _ _ PRE/OBS
HOURS OBS PRE PRE/OBS RMS V/CO2 V/(CoCp) < 2.0
DATA FROM ALL
PERIODS AND
MONITORS:
THRESHOLD -
0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS 6S762 14.76 13.28
TWO TOWER LEVELS 67900 14.71 16.95
ONE TOWER LEVEL 57876 14.97 56.91
COMPLEX I 71736 14.59 35.04
RTDM (DEFAULT) 71736 14.59 26.73
RTDM (ONSITE) 71736 14.59 §.24
0.90 132.8 81.0 90.0 47
1.15 186.8 161.3 140.0 47
3.80 278.0 344.9 90.7 44
2.40 366.5 631.0 262.8 56
1.83 227.§ 243.2 132.7 56
0.36 96.5 43.8 121.9 49
2) THRESHOLD »
1.00 UG/M**3
CTDM, SEVERAL
TOWER LEVELS
TWO TOER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT-)
RTDM (ONSITE)
3682 65.8 142
2184 50.3 265
4144 57.1 414
1886 81.5 786
1330 66.9 742
3133 67.5 66
2.15 358 29.7 13.8 31
5.27 744 219.2 41.6 22
7.25 753 174.0 24.0 20
9.64 1756 463.9 48.1 21
11.10 1160 300.7 27.1 8
0.97 318 22.1 22.8 25
214
-------
TABLE D-10
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME AND SPACE
3-HOUR AVERAGES, 1980 TRACER: S02 SITE: WIDOWS CREEK
(Concentrations given in units of micrograms per cubic meter.)
SITE-
HOURS OBS
PRE PRE/OBS RMS
% CASES:
0.5 <
PRE/O'BS
V/Co2 V/(CoCp) < 2.0
DATA FROM ALL
PERIODS AND
MONITORS:
1) THRESHOLD *
0.0 US/M**3
CTDM, SEVERAL
TOWER LEVELS 18021 15.93 13.51
TWO TOWER LEVELS 17377 15.88 16.15
ONE TOWER LEVEL 14449 15.85 55.52
COMPLEX I 18695 15.71 36.58
RTDM (DEFAULT) 18695 15.71 29.98
RTDM (ONSITE) 18695 15.71 5.42
0.85 93.2 34.2 40.3 37
1.02 112.0 49.8 48.9 34
3.50 218.9 190.8 54.5 33
2.33 266.4 287.6 123.5 46
1.91 170.7 118.1 61.9 45
0.35 67.5 18.4 53.5 38
2) THRESHOLD -
1.00 UG/M**3
CTDM, SEVERAL
TOWER LEVELS
TWO TOER LEVELS
ONE TOWER LEVEL
COMPLEX I
RTDM (DEFAULT)
RTDM (ONSITE)
1824 47.9 97
1370 38.5 136
1926 42.3 295
1129 57.5 449
927 53.0 402
1704 51.7 42
2
3
6
7
7
0
.02
.53
.97
.82
.59
.80
213
305
516
932
619
181
19
62
149
263
136
12
.8
.8
.1
.0
.2
.3
9
17
21
33
18
15
.8
.8
.4
.6
.0
.3
30
24
20
19
13
27
215
-------
APPENDIX E
EVALUATION RESULTS BY METEOROLOGICAL CATEGORY
FOR CONCENTRATIONS PAIRED IN TIME, UNPAIRED IN SPACE
216
-------
APPENDIX E
EVALUATION RESULTS BY METEOROLOGICAL CATEGORY
FOR CONCENTRATIONS PAIRED IN TIME, UNPAIRED IN SPACE
In this appendix, the results presented in Appendix B for
concentrations paired in time, not space, are extended to
meteorological categories. These categories include three stability
classes (D,E,F) and three wind speed categories (0-1, 1-3 and >3 m/sec
at release height). More discussion can be found in Section 4. A
guide to these tables is listed below. Note that the six model runs
(CTDM for several, two and one tower level(s), COMPLEX I, and RTDH
(default and on-site) are grouped together for each site and tracer.
Table it
B-l through E-6
E-7 through E-12
E-13 through E-18
E-19 thorugh E-24
E-25 through E-30
E-31 through E-36
E-37 through E-42
E-43 through E-48
Description
Evaluation results for SF6 at CCB
Evaluation results for CF3&r at CCB
Evaluation results for SF6 at HBR
Evaluation results for CF3Br at HBR
Evaluation results for SFg at FSPS
Evaluation results for CF3&r at FSPS
Evaluation results for S02 at Westvaco,
1-hour averages
Evaluation results for S0£ at Widows
Creek, 1-hour averages
217
-------
TABLE E-l
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
MODEL: CTDM THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/CQ2
V/(CoCp)
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02 —
V/(CoCp)
% WITHIN FACTOR 2
0-1 1-3 GT 3
0 1 27
— — 21.73
-__ ___ H.25
, .52
_~_ 17.51
. gs
— — 1.26
2 26.
E TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
— 27
— 15.05
— - — 3 . 02
— — r . 53
— — 11.18
0-1 1-3 GT
3 9
— 13.31 30
• 14.69 24
— 1.10
- — 14.93 19
— Io26
1.14
22.
EACH 1-HOUR PERIODS
STABILITY E
0-1 1-3 GT
9
8.18 21
— 9.48 16
1.16
— 11.45 14
3
16
.31
.17
.80
.54
.42
.52
50.
3
16
.91
.83
.77
.59
0-1 1-3
2 30
— 33.58
— 39.72
— 1.18
— 36.10
— 1.16
— .98
— 47 .
STABILITY
0-1 1-3
30
21.83
— 30.83
1.41
28.41
GT 3
12
36.35
37.71
1.04
39.58
1.19
1.14
42.
F
GT 3
12
23.76
28.44
1.20
31,00
-1.04
33.
1,96 .44
1.69 .58
11. 50.
1.69 1.70
1.20 1.42
43. 50.
* Statistics are not presented for cases with less than 6 data pairs
218
-------
TABLE E-2
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
MODEL: CTDM (DEGR1) THRESHOLD: .00 uS/M**3 ~
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
0-1 1-3 GT 3
0 1 27
21.73
11.43
. 53
• 17 . 15
. 62
— 1.18
2 33.
E TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 1 27
15.05
— — 8.78
. 58
11.21
. 55
.95
2 30.
0-1 1-3
3 9
13.31
22.78
1.71
38.26
8 . 27
4.83
22.
GT 3
16
30.31
8.74
.29
32.48
1.15
3.98
19.
0-1 1-3
2 30
— 33.58
9.83
.29-
49.83
2.20
7.52
10.
GT 3
12
36.35
19.15
.53
33.83
.87
1.64
33.
EACH 1-HOUR PERIOD:
STABILITY
0-1 1-3
3 9
8 . 18
12.28
1.50
22 . 87
7.83
5.21
22.
E
GT 3
16
21.91
5.56
.25
25.42
1.35
5.30
19.
STABILITY
0-1 1-3
2 30
— 21.83
7 . 00
.32
34.71
2.53
7.89
10 .
F
GT 3
12
23.76
14.50
.61
22.25
.88
1.44
25.
WIND SPEED(M/S)
I OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
Statistics are not presented for cases with less than 6 data pairs.
219
-------
TABLE E-3
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
MODEL: CTDM (DEGR2) THRESHOLDS .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/Co2
V/(CoCp)
0-1 1-3 GT 3
0 1 27
— - 21.73
7 . 09
.33
— • 20.53
.89
2.73
2 — 33.
E TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 1 27
15 . 05
.40
.40
— 14.29
_._ >90
-- - 2.28
2 26.
0-1 1-3 GT
3 9
— 13.31 30
— 3.16 8
.24
— 18.37 35
1.90 1
8.03 4
0.
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT
3 9
80 18 21
— .28
.28
13.10 26
2 . 57 1
9 . 18 4
0.
3
16
.31
.46
.28
.07
.34
.79
19.
3
16
.91
.32
.32
.03
.41
.44
25.
0-1 1-3
2 30
— 33.58
— 4.35
-— . 13
— 51.08
2.31
— 17.87
1 — 10.
STABILITY
0-1 1-3
2 30
— 21.83
— . 14
. 14
— 34.16
— 2.45
17,01
10.
GT 3
12
36.35
10.70
.29
39.00
1.15
3.91
17.
F
GT 3
12
23.76
.37
.37
26.05
1.20
3.21
17.
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/-(CoCp)
* Statistics are not presented for cases with less than 6 data pairs.
220
-------
TABLE E-4
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
MODEL: COMPLEX I THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
^m
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
0-1 1-3 GT 3
0 1 27
21.73
10.45
.48
19.54
— — ,— ___ 0 1
™~ • ox
1 . 63
52.
TOP 5 VALUES FROM EA(
STABILITY D
0-1 1-3 GT 3
0 1 27
15.05
9.15
.61
13.30
.78
1.28
59.
0-1 1-3 GT
3 9
13.31 30
42.11 25
.3.16
— - 51.38 23
14.91
4.71
— 11.
:H 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT
3 9
8.18 21
31.15 20
3.81
41.33 20
25.56
6.71
11.
3
16
.31
.50
.84
.61
.61
.72
56.
3
16
.91
.76
.95
.34
.86
.91
56.
0-1 1-3
2 30
33.58
68.63
2 . 04
58.68
3 . 05
1.49
20.
STABILITY
0-1 1-3
2 30
21.83
52.96
2.43
47.46
• 4.73
1.95
23 .
GT 3
12
36.35
48.31
1.33
28.73
.62
.47
58.
F
GT 3
12
23.76
41.23
1.74
26.97
1.29
.74
50.
* Statistics are not presented for cases with less than 6 data pairs
221
-------
TABLE E-5
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
MODEL: RTDM (DEFAULT) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/Co2
V/(CGCp)
0-1 1-3 GT 3
0 1 27
— 21.73
— — 8 . 15
— .38
— 20.61
— - ..90
> 2.40
2 — - 26.
E TOP 5 VALUES FROM EACH
STABILITY D
0-1 1-3 GT 3
0 1 27
_„_ 15.Q5
6.72
— — .45
— 14 . 03
— — .87
. . Ia95
2 ' 30.
0-1 1-3 GT
3 9
— 13.31 30
— 27.23 13
— 2.05
— 32.32 27
— i.90
— 2.88 1
— 22 .
1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT
3 9
8.18 21
— 20.57 11
— 2 . 52
— 27.61 20
— 11.41
— 4.53 1
22.
3
16
.31
.67
.45
.10
.80
.77
44.
3
16
.91
.91
.54 .
.78
.90
.65
44.
0-1 1-3
2 30
— 33.58
~- 75.15
— 2.24
— 94.98
— 8.00
OK>aiB ^ ^ OO
— 13 .
STABILITY
0-1 ' 1-3
2 30
• 21.83
46.75
2 . 14
50.12
— 5.27
— 2.46
17.
GT 3
12
36.35
23.24
.64
28.41
.61
.96
50.
F
GT 3
12
23.76
21.95
.92
16.15
.46
.50
50.
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
* Statistics are not presented for cases with less than 6 data pairs
222
-------
TABLE E-6
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: CINDER CONE BUTTE
MODEL: RTDM (ONSITE) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CQCp)
% WITHIN FACTOR 2
0-1 1-3 GT 3
0 1 27
21.73
14.46
-67
18.97
0.76
1.15
0. 100. 48.
TOP 5 VALUES FROM EACH
STABILITY D '
0-1 ' 1-3 GT 3
0 1 27
15.05
11.46
.76
13 . 64
0.82
1.08
48.
0-1 1-3
3 9
13.31
20.09
1.51
22.46
— .- 2.85
1.89
0. 22.
GT 3
16
30.31
23.72
.78
18.25
0.36
0.46
44.
0-1 1-3
2 30
33.58
26.15
.78
33.21
0.98
1.26
50. 33.
GT 3
12
36.35
25.68
.71
30.65
.71'
1.01
33.
1-HOUR PERIOD:
STABILITY
0-1 1-3
3 9
8 . 18
14.48
1.77
16.24
3 . 94
2.23
11.
E-
GT 3
16
21.91
18.27
.83
14.08
0.41
0.50
56.
STABILITY
0-1 1-3
2 30
21.83
17 . 63
.81
20.78
0.91
' 1 . 12
27.
F
GT 3
12
23.76
20.37
.86
23.45
0.97
1.14
25.
* Statistics are not presented for cases with less than 6 data pairs
223
-------
TABLE E-7
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: CINDER CONE BUTTE
MODEL: CTDM THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/Co2
V/(CoCp)
0-1 1-3 GT 3
0 0 11
17.04
— 15.95
— — . 94
— -__ 30.37
— — 3 . 18
. 3.40
2 9.
tE TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 0 11
10.51
— - 10 . 15
— — .97
— 20.11
— _ 3 „ QQ
3o79
2 9-.
0-1 1-3 GT 3
0 2 11
— — 23.43
17.86
___ .76
-— — 27.81
— — 1.41
— i.85
— — 27 .
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
0 2 11
12.32
— 12.51
1.02
— - 18.96
— 2.37
— 2.33
9 .
0-1 1-3 GT 3
0 4 16
___ „_„ 9,53
— — - 18.49
— — 1.94
— -— 24.58
g> f e
-------
TABLE E-8
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: CINDER CONE BUTTE
MODEL: CTDM (DEGR1) THRESHOLD: .00 US/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR DBS
MEAN FOR.PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/CCoCp)
% WITHIN FACTOR 2
0-1 1-3 GT 3 0-1 1-3 GT
0 0 11 02
17.04 — - 23.
6.39 11.
.38
25.08 23.
— ^wm «•» 2 « I/ ••— •«• 9
___ _«_ K ^Q ••« ««« 1
••» ««• 3 • / O ••«• »««• ^ 9
••« ••• ^g 9 •••• •«••
TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
0-1 1-3 GT 3 0-1 1-3 GT
0 0 11 02
10.51 — 12.
4.47 5.
.43
— 16.79 - — 14.
2.55 1.
5.99 2.
• ^^
3
11
43
61
50
19
98
98
9.
3
11
32
88
48
39
37
86
9.
0-1 1-3 GT 3_
0 4 16
9.53
7 . 16
.75
13 . 64
2 . 05
2.73
6 .
STABILITY F
0-1 1-3 GT 3
0 4 16
4.95
5.92
1.20
9.95
— - 4 . 05
3.38
13 .
* Statistics are not presented for cases with less than 6 data pairs
225
-------
TABLE E-9
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: GSNDER CONE BUTTE
MODEL: CTDM (DEGR2) THRESHOLD; .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
STABILITY D STABILITY E
0-1 1-3 GT 3 0-1 1-3 GT
0 0 11 02
_- ... „„_ 17.04 — — 23
— — _ 6.83 — — 6
__» _„_ .40 — —
— — 27.22 — — 25
— _ 2.55 — — 1
6.37 • — 4
— ._- 45. „._ _„_
TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
0-1 1-3 GT 3 0-1 1-3 GT
0 • 0 11 0 2
10.51 12
— — 5.07 - — — 2
— _«,_ .48 .._
— 18 .50 — — 14
— — 3.10 — — i
— 6.42 — - 6
45.
3
11
.43
.11
.26
.66 '
.20
.60
36.
3
11
.32
.80
.23
.47
.38
.07
27-
STABILITY F
0-1 1-3 GT 3
0 4 16
— • • 9.53
— 3.49
(MKXBB «0>*» » J /
a a « Q
Jm b f 13
-~ — - 1.53
WIHUSO fa A 1 Q
*w • J.S
— — 25.
STABILITY F
0-1 1-3 GT 3
0 4 16
. 4.95
___ .__ 2 . 05
_„_ .41
„__ __. 6.60
— — 1.78
— ___ 4.29
— — 19.
* Statistics are not presented for cases with less than 6 data pairs
226
-------
TABLE E-10
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: CINDER CONE BUTTE
MODEL: COMPLEX I THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
0-1 1-3 GT 3
0 0 14
16.21
12 . 01
.74
19.86
1.50
2 . 03
64 .
TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 0 14
10.41
10.27
___ — — — QQ
™ «•«• € yy
12 . 37
1.41
- — 1.43
— — : 29 .
0-1 1-3 GT 3
0 2 13
22.37
20.90
_•__ _«« Q ^
MM •«»«• » J J
— ** — ~m • 22*82
1.04
1.11
54 .
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
0 2 13
12 . 10
14 . 54
1.20
13.36
1.22
1.01
54 .
0-1 1-3 GT 3
0 4 16
9.53
25.60
2 . 68
22.21
5.43
2 . 02
25.
STABILITY F
0-1 1-3 GT 3
0 4 16
4.95
20.94
4.23
19.53
15.59
3 . 68
13 .
* Statistics are not presented for cases with less than 6 data pairs
227
-------
TABLE E-ll
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES ~ TRACERS CF3BR SITES CINDER CONE BUTTE
MODEL: RTDM (DEFAULT) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITIES A-D STABILITY E STABILITY F
WIND SPEED (M/S)
#. OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
0-1 1-3 GT 3 0-1 1-3 GT
0 0 11 02
17.04 — — 23
8.33 12
___ .49 . — _
— — 22.40 — — 20
_„- . i.73 _„„
— — 3.53 , !
— _ 64. ... _„_
TOP 5 VALUES FROM EACH 1-HOUR PERIODS
STABILITY D STABILITY E
0-1 1-3 GT 3 0-1 1-3 GT
0 0 11 02
10.51 — 12
6.50 „.. 8
.62
14.63 — 11
— 1.94
. ___ 3ol3 ___ i
55.
• ^^
3
11
.43
.39
.53
.00
.73
.38
€4.
3
11
.32
.61
.70
.00
.80
.14
64.
0-1 1-3 GT 3
0 4 16
— — 9 . 53
— — 4.25
aascMB-™ «»,aic«=> A^
• *B 3
— - — 10.84
__„ __„ 1,29
— — 2.90
M Q
**""* «5 O 0
STABILITY F
0-1 1-3 GT 3
0 4 16
4.95
3 . 65
.74
— ,_ 7oo6
. _ — 2.04
2.76
31.
* Statistics are not presented for cases with less than 6 data pairs
228
-------
TABLE E-12
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: CINDER CONE BUTTE
MODEL: RTDM (ONSITE) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3
# OF DATA PAIRS 0 0
MEAN FOR OBS
MEAN FOR PRE — -
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
* WITHIN FACTOR 2
AVERAGE OF THE TOP 5 VALUES
STABILITY
WIND SPEED (M/S) 0-1 1-3
# OF DATA PAIRS 0 0
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CQCp)
% WITHIN FACTOR 2
GT 3
11
17.04
17.50
1.03
22.01
1.67
1.63
27.
FROM
D
GT 3
11
10.51
11.29
1.07
14.72
1.96
1.83
45.
0-1 1-3 GT 3
0 2 11
23.43
10.00
.43
25.01
1.14
2 . 67
— 45 .
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
0 2 11
12.32
7 . Q7
. 57
13.88
1.27
2.21
.45.
0-1 1-3 GT 3
0 4 16
9.53
11.49
— 1.20
12.24
, i.65
L37
31.
STABILITY F
0-1 1-3 GT 3
0 4 16
, 4.95
7 . lx
1.44
7 . 04
2 . 03
1.41
31.
* Statistics are not presented for cases with less than 6 data pairs
229
-------
TABLE E-13
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACERS SF6 SITE: HOGBACK RIDGE
MODEL: CTDM THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 Ir3 GT 3
# OF DATA PAIRS Oil 001
MEAN FOR OBS - — — — — - •
MEAN FOR PRE — — -
BIAS (PRE/OBS) — — — •
RMS ERROR — — — — — —
V/C02 — — — -
V/(CoCp)
% WITHIN FACTOR 2 — — -
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS Oil 001
MEAN FOR OBS
MEAN FOR PRE —
BIAS (PRE/OBS) —
RMS ERROR — — —
V/C02 — —
V/(CoCp) — . —
% WITHIN FACTOR 2
0-1 1-3
2 43
— 26.69
— 52.44
1.96
— 63.96
— 5.74
2.92
35.
STABILITY
0-1 1-3
2 43
20.62
39.36
1.91
• 41.03
— 3.96
2 . 07
23 .
GT 3
11
11.92
22.94
1.93
12.25
1.06
0.55
55o
F
GT 3
11
9.93
21.50
2,17
12.40
1.56
0.72
36.
* Statistics are not presented for cases with less than 6 data pairs
230
-------
TABLE E-14
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: HOGBACK RIDGE
MODEL: CTDM. (DEGR1) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS Oil 001
MEAN FOR OBS —
MEAN FOR PRE —
BIAS (PRE/OBS)
RMS ERROR — —
V/C02 —
V/(CoCp)
% WITHIN FACTOR 2 ---
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
0-1 1-3
2 43
26.69
30.80
1 . 15
35.96
1.81
1.57
30.
STABILITY
0-1 1-3
GT 3
11
11.92
14.58
1.22
17.54
2.17
1.77
36.
F
GT 3
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(.CoCp)
% WITHIN FACTOR 2 —
* Statistics are not presented for cases with less than 6 data pairs
2 43 11
- 20.62 9.93
- 24.15 11.15
1.17 1.12
- 26.31 11.23
1.63 1.28
1.39 1.14
30. 27.
231
-------
TABLE E-15
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITES HOGBACK RIDGE
MODEL: CTDM (DEGR2) THRESHOLDS .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS Oil 001
MEAN FOR OBS — - — — - — - — —
MEAN FOR PRE — — — —-
BIAS (PRE/OBS) — - — —
RMS ERROR — - — — - —
V/Co2 — •' — — —
V/(CoCp)
% WITHIN FACTOR 2 — — - —
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS 01 1 0 0 1
MEAN FOR OBS — -
MEAN FOR PRE — — —
BIAS (PRE/OBS) — —
RMS ERROR —
V/C02
V/(CoCp)
% WITHIN FACTOR 2 — —
0-1 1-3
2 43
— 26.69
— 12.38
— .46
— 31.78
— 1 = 42
— 3.06
— 14 .
STABILITY
0-1 1-3
2 43
— 20.62
10.31
.50
25.49
Io53
— • 3.06
9.
GT 3
11
11.92
9.62
.81
11.12
.87
1.08
36.
F
GT 3
11
9.93
3.05
.81
9.69
.95
1.17
36.
* Statistics are not presented for cases with less than 6 data pairs
232
-------
TABLE E-16
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: HOGBACK RIDGE
MODEL: COMPLEX I THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
• «•«»•• WW •»•«»•«••••••• •«••••>«»•«••••«••«• W««4»
-------
TABLE E-17
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: HOGBACK RIDGE
MODEL: RTDM (DEFAULT) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
1
1
V/GO2 <«»«•«> <»«£«=,
V/ (COCp) •»«•«> •««• «,«>«» • OH,™ oeiBxn. «,«,«,
% WITHIN FACTOR 2 —- —- — —• — —
AVERAGE OF THE TOP § VALUES FROM EACH 1-HOUR PERIODS
STABILITY D STABILITY E
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
Oil 001
2 44 11
— 27.05 11.92
— 78.77 1.74
— 2.91 .15
—- 121.17 11.70
— 20.07 .96
— 6.89 6.58
— 20. 0.
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/-(CoCp)
% WITHIN FACTOR 2
* Statistics are not presented for cases with less than 6 data pairs
STABILITY F
1 1-3 GT 3
2 44 11
— 20.93 9.93.
— 58.75
2.81
— 84.86
— 16.43
5.85
23 .
1.52
.15
9.85
.98
6.41
0.
234
-------
TABLE E-18
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: HOGBACK RIDGE
MODEL: RTDM (ONSITE) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CQCp)
% WITHIN FACTOR 2
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3
Oil 001 2 43
— _ 26.69
. — 38.70
1.45
— 80.10
— — 9.00
6.21
2 53.
IE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E STABILITY
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3
GT 3
11
11.92
13.65
1.14
4. .53
.14
.13
82.
F
GT 3
WIND SPEED(M/S)
I OF DATA PAIRS 0110
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS) •
RMS ERROR —
V/Co2
V/.(CQCp)
%'WITHIN FACTOR 2
* Statistics are not presented for cases with less than 6 data pairs
2 43 11
20.62 9.93
.25.88 11.75
1.25 1.18
35.70 4.33
3.00 .19
2.39 .16
56. 73.
235
-------
TABLE E-19
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
MODEL: CTDM THRESHOLD: .01 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY S STABILITY
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3
# OF DATA PAIRS 010 0 1 0 14 26
MEAN FOR OBS — — — — — — 101.11 93.86
MEAN FOR PRE —- — — — — 60.69 45.29
BIAS (PRE/OBS) — — — — .60 .48
T51WTC B*T3OrtO ^^^ -I. oomon anamn Q "7 ^A QO "5 C
•saro'E) O / a J ** 9 7 e «£ W
V/CO2 — —«»— — ««_ =,=«» ... 075 1.12
V/(CoCp) — — — — — — 1.24 2.32
% WITHIN FACTOR. 2 — — — — — — 64. 54.
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
010 010
GT 3
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS) —
RMS ERROR
V/C02 —
V/.(CoCp)
% WITHIN FACTOR 2
STABILITY
0-1 1-3
14 26
59.17 64.72
48.90 37.52
.83 .58
32.34 56.61
.30 .76
.36 1.32
57. 46.
GT 3
* Statistics are not presented for cases with less than 6 data pairs
236
-------
TABLE E-20
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
• (SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
MODEL: CTDM (DEGR1) THRESHOLD: .01 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT
000 0 0 0 10 14
— • 113.28 121.63
— 41.28 51.66
— 141.69 146.85
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS —
MEAN FOR PRE —
BIAS (PRE/OBS) :
RMS ERROR —-
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
1.56
4.29
30.
1.46
3.43
43.
0
f OF DATA PAIRS 0 0
MEAN FOR OBS .
MEAN FOR PRE
BIAS (PRE/OBS) —
RMS ERROR —
V/C02
V/(CQCp)
% WITHIN FACTOR 2
* Statistics are not presented for cases with less than 6 data pairs
STABILITY F
0-1 1-3 GT
10 14
63.96 82.98
36.06 42.43
.56 .51
65.44 84.86
1.05 1.05
1.86 2.05
30. 43.
237,
-------
TABLE E-21
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
MODEL: CTDM (DEGR2) THRESHOLD: .01 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3
000 010 67
__. «._ , „— _„_ 114.67 71.74
___ —_ ___ „— 70.64 48.14
—_ „__ —. ___ .62 .67
_„_ ___ —_ .._ 146.08 31.43
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS — — • -— —
MEAN FOR PRE —- — —- —
BIAS (PRE/OBS) —- — — — —
RMS ERROR • — — —- —
V/C02 —- — — — —
V/(CoCp) — — — — —
% WITHIN FACTOR 2 — —- . —- —
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
^^••^^•^ «M^ «><••<«> ^<»^^^i» <»^^^a»«» ^^^^^>» ^^^^^^
000 010
GT 3
1.62 .19
2.63 .29
17- 57-
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE —
BIAS (PRE/OBS)
RMS ERROR
V/C02 —
V/CCoCp)
% WITHIN FACTOR 2
STABILITY F
0-1 1-3 GT 3
6 7 C
60.97 55.45
61.80 43.05
1.01 .78
71.95 25.03
1.39 .20
1.37 .26
33. 36.
* Statistics are not presented for cases with less than 6 data pairs
238
-------
TABLE E-22
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
MODEL: COMPLEX I THRESHOLD: .01 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3
WIND SPEED(M/S)
# OF DATA PAIRS 001 QIC
MEAN FOR OBS
MEAN FOR PRE ---.-
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CQCp)
% WITHIN FACTOR 2
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
GT 3
# OF DATA PAIRS 001 01
MEAN FOR'OBS
MEAN FOR PRE •
BIAS (PRE/OBS)
RMS ERROR
V/(CoCp)
% WITHIN FACTOR 2 —
* Statistics are not presented for cases with less than 6 data pairs
12 26 C
129.13 111.31
21S.96 163.69
1.67 1.47
142.59 121.39
1.22 1.19
.73 .81
42. 42.
STABILITY F
0-1 1-3 GT 3
12 26 (
74.80 77.51
192.12 141.72
2.57 1.83
135.55 104.26
3.28 1.81
1.28 .99
33. 46.
239
-------
TABLE E-23
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
* MODEL: RTDM (DEFAULT) THRESHOLD: .01 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3
1
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS — — •
MEAN FOR PRE -— —-
BIAS (PRE/OBS) — -— —
RMS ERROR — —— — —— — —
V/C02 -— —- —- —
V/(CoCp) . — —
% WITHIN FACTOR 2 •
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD;
STABILITY D STABILITY E
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
GT 3
0 12 24
— 129.13 110.15
— 528.66 434.83
4.09 3.95
— 729,26 680.00
31.90 38.11
7.79 9.65
17. 13.
# OF DATA PAIRS 0
MEAN FOR OBS
MEAN FOR PRE •
BIAS (PRE/OBS)
RMS ERROR
V/C02 —
V/(CoCp)
% WITHIN FACTOR 2
STABILITY F
0-1 1-3 GT 3
12 23 C
74.80 69.67
198.39 183.54
2.65 2.63
242.62 215.28
10.52 9.55
3.97 3.62
0. 4.
* Statistics are not presented for cases with less than 6 data pairs
240
-------
TABLE E-24
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: HOGBACK RIDGE
MODEL: RTDM (ONSITE) THRESHOLD: .01 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS 001 0 1 (
MEAN FOR OBS —
MEAN FOR PRE
BIAS (PRE/OBS) • —
RMS ERROR. :
V/C02
V/(CoCp)
% WITHIN FACTOR 2 —
AVERAGE OF THE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/Co2
% WITHIN FACTOR 2 ---
16 34 C
126.66 92.76
127.40 58.63
1.01 .63
274.65 146.70
4.70 2.50
4.68 3.96
25. 18.
STABILITY F
0-1 1-3 GT 3
16 34 (
72.31 63.50
58.44 42.37
.81 .67
79.22 104.00
1.20 2.68
1.49 4.02
25. 24.
* Statistics are not presented for cases with less than 6 data pairs
241
-------
TABLE E-25
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: TRACY POWER PLANT
MODEL: CTDM (MODEL) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/Co2
V/(CQCp)
% WITHIN FACTOR
AVERAGE OF TI
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
STABILITY D
0-1 1-3 GT 3
0 1 23
— — 1.03
.96
— — - — .92
— — .79
— _ „ 58
— - . 63
2 - — — 61.
£ TOP 5 VALUES FROM EACH
STABILITY D
0-1 1-3 GT 3
0 1 23
,71
.71
1.01
-•==- .51
, .._ . 52
.52
2 52.
STABILITY E
0-1 1-3 GT 3
2 2 13
„__ ___ 1,02
— — 1.45
!.42
— 1.42
1.94
!.37
69.
1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
2 2 13
— . 67
.90
— 1.35
— — . §5
— .S3
— .69
— 54.
STAI
0-1
10
3.24
2.32
.87
2.14
.44
.50
80.
STAI
0-1
10
2.27
1.77
.78
1.19
.27
.35
30.
JILITY
1-3
42
2.33
2.24
.96
2.34
1.00
1.04
60.
JILITY
1-3
42
1.53
1.57
1.03
1.38
.82
.30
55.
F
GT 3
17
1.84
1.93
1.05
2.84
2.38
2.27
53.
F
GT 3
17
1.06
1.36
1.29
1.09
Io06
.82
53.
* Statistics are not presented for cases with less than 6 data pairs
242
-------
TABLE E-26
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: TRACY POWER PLANT
MODEL: CTDM (DEGR1) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
J OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
STABILITY D
0-1 1-3 GT 3
0 1 23
1.03
f — 1 . 11
1.07
.94
.33
.78
61.
TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 1 23
.71
.76
1.07
.62
.77
.72
61.
STABILITY E
0-1 1-3 GT 3
1 2 13
1.02
1.18
1.16
.88
.75
. ss
69 .
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
1 2 13
. 67
.78
1.17
•«» ••• , 53
1.03
.88
54 .
STAI
0-1
10
3.24
3.02
.93
3.18
.97
1.04
20.
STAI
0-1
10
2.27
1.99
.88
2.04
.81
.92
40.
3ILITY
1-3
42
2.33
2.38
1.02
2.20
.39
.87
43.
3ILITY
1-3
42
1.53
1.55
1.02
1.49
.96
.94
48.
F
GT 3
17
1.84
2.45
1.33
3.33
3.27
2.46
47.
F
GT 3
17
1.06
1.58
1.50
1.29
1.50
1.00
35.
* Statistics are not presented for cases with less than 6 data pairs
243
-------
TABLE E-27
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: TRACY POWER PLANT
MODEL: CTDM (DEGR2) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CQCp)
STABILITY D
0-1 1-3 GT 3
0 1 23
— 1.03
1.35
1.30
— 1 . 13
— 1.20
.92
2 48.
E TOP 5 VALUES FROM EACH
STABILITY D
0-1 1-3 GT 3
0 1 23
.71
.73
1.04
.56
— . 63
— — . 61
2 — - 48.
STABILITY E
0-1 1-3 GT 3
1 2 13
„__ 1.02
_„- _-_ 1.84
___ __. i.ai
— — - 2 . 33
— 5.23
___ 2,90
— 38 .
1-HOUR PERIODS
STABILITY E
0-1 1-3 GT 3
1 2 13
___ . §7
.81
1.21
1.19
— — 3 . 15
2.61
54.
STAI
0-1
10
3.24
2.30
.71
3.09
.91
1.28
40.
STAI
0-1
10
2.27
1.62
.71
2.37
1.09
1.53
20.
JILITY
1-3
39
2.36
3.55
1.50
3.99
2.83
1.89
36.
3ILITY
1-3
39
1.54
2.06
1.33
2.68
3.01
2.26
26.
F
GT 3
16
1.84
1.85
1.01
2.76
2.26
2.24
38.
F
GT 3
16
1.02
.93
.91
,74
.53
.59
56.
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(.CoCp)
* Statistics are not presented for cases with less than 6 data pairs
244
-------
TABLE E-28
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: TRACY POWER PLANT
MODEL: COMPLEX I THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
WIND SPEED(M/S)
I OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
STABILITY D
0-1 1-3 GT 3
0 1 23
— 1 . 03
.45
, 44
. 1.09
1.10
2.51
2 35.
£ TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 1 23
.71
.37
*•«•*• MW • 53
.70
.98
1.84
2 39.
STABILITY E
0-1 1-3 GT 3
2 2 14
. 98
— 2.43
2.48
— 1.72
3 . 06
!.23
29.
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
2 2 14
— > . 65
1.76
2.72
1.26
3 . 79
1.39
14 .
0
3
10
3
10
11
3
0
2
6
3
8
12
4
STA
-1
10
.24
.32
.18
.39
.31
.55
0.
STA
-1
10
.27
.91
.04
.03
.50
.11
10.
BIL
1
2
10
•4
10
13
4
BIL
1
1
6
4
6
18
4
ITY
-3
42
.33
.47
.48
.12
.79
.19
12.
ITY
-3
42
.53
.32
.46
.49
.04
.04
7.
F
GT
1
4
2
3
4
1
F
GT
1
2
2
2
4
1
3
17
.84
.06
.21
.77
.20
.90
12.
3
17
.06
.97
.81
.30
.73
.68
6.
WIND SPEED(M/S)
f OF 'DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/fCoCp)
% WITHIN FACTOR
* Statistics are not presented for cases with less than 6 data pairs
245
-------
TABLE E-29
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER; SF6 SITE: TRACY POWER PLANT
MODEL: RTDM (DEFAULT) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
STABILITY D STABILITY E
0-1 1-3 GT 3 0-1 1-3 GT
0 1 23 22
— — 1.03 — - —
— — .59 — — 1
.»_ — .57 — __. i
— — LOS
___ -„„ i.03 — -
__„ „_. 1.79 — . „__
„„„ _ — 43. _-_ _~_
TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
0-1 1-3 GT 3 0-1 1-3 GT
0 1 23 22
.71
— .49 _._ — .
— .69 — — 1
_„. .70 —
«>^ • 79 ^^^ • » «•
1.44
39.
»«•••
3
14
o98
.21
.23
.94
.92
.74
43.
3
14
.65
.39
.38
.55
.72
.52
57.
STAE
0-1
10
3.28
5.44
1.68
S.99
3.42
2.04
20.
STAI
0-1
10
2.27
3.43
1.51
3o70
2.65
1.75
10.
JILITY
1-3
42
2.33
5.11
2.19
4ol8
3.21
1.47
45.
SILITY
1-3
42
1.53
2.87
1.88
2.20
2.08
1.11
40.
F
GT 3
17
1.84
1.49
.81
2,49
1=83
2.27
82.
F
GT 3
17
1.06
1.07
1.01
.82
.SO
.59
65.
* Statistics are not presented for cases with less than 6 data pairs
246
-------
TABLE E-30
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SF6 SITE: TRACY POWER PLANT
MODEL: RTDM (ONSITE) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HQUR VALUES, UNPAIRED IN SPACE:
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
STABILITY D
0-1 1-3 GT 3
0 1 23
1.03
1.14
1.10
..... ___ 1 RQ
•««• ••• i • O3J
«. ___ ^ ^ fl
*** •" •«• * « JO
___ — — — *9 IK
••• ••• ^ % j^y
22.
TOP 5 VALUES FROM EACH
STABILITY D
0-1 1-3 GT 3
0 1 23
___ ___ T1
• / i
.75
1.06
.94
1.77
1.67
17 -
STABILITY E
0-1 1-3 GT 3
2 2 13
1.02
1.40
1.37
1.30
~*" ••^ 1 • 6 2
1.18
46.
1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
2 2 13
. 67
.81
1.22
. 69
1.07
•*•• «••«» .88
62 .
STAI
0-1
10
3.24
.97
.30
2.86
.78
2.60
10.
STAI
0-1
10
2.27
.61
.27
2.08
.84
3.14
30.
3ILITY
1-3
42
2.33
1.07
.46
2.40
1.05
2.31
33.
JILITY
1-3
42
1.53
.72
.47
1.50
.96
2.05
3.3.
F
GT 3
17
1.84
1.32
.72
2.64
2.06
2.86
47.
F
GT 3
17
1.06
.83
.78
.90
.72
.92
53.
* Statistics are not presented for cases with less than 6 data pairs
247
-------
TABLE E-31
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER; CF3BR SITE: TRACY POWER PLANT
MODEL: CTDM THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
0-1 1-3 GT 3
0 1 23
— - — 1,11
— — !• 1.49
— — 1.34
— — - .98
— — .,78
— • . 58
2 — 83.
E TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 1 23
— .72
.84
• 1.16
— .42
.35
— — .30
2 74.
0-1 1-3 GT 3
2 3 12
-__ _._ 2 . 02
— 2 . 04
— • 1.01
— — 2 . 44
— --- 1.45
— — 1.44
tn «ja S?rt
3 W »
EACH 1-HOUR PERIODS
STABILITY E
0-1 1-3 GT 3
2 3 12
— — 1.21
— — 1.06
— .88
— 1.56
.„_ 3. >67
— — 1.91
< . 75.
0-1
11
4.35
2.87
.59
5.03
1.08
1.82
64.
STAI
0-1
11
2.78
1.61
.58
2.26
.66
1.14
55.
1-3
41
3.21
2,81
.38
2.86
.79
.91
51.
3ILITY
1-3
41
2.08
1.37
.66
1.63
.62
.94
61.
GT 3
17
3,56
2.42
.68
3.59
1.01
1.49
53.
F
GT 3
17
1.89
1.42
.75
1.18
.39
.52
65.
WIND SPEED(M/S)
I OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/.(CQCp)
* Statistics are not presented for cases with less than 6 data pairs
248
-------
TABLE E-32
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: TRACY POWER PLANT
MODEL: CTDM(DEGRl) THRESHOLD: .00 US/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(H/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
WIND SPEED(M/S)
I OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
0-1 1-3 GT 3 0-1 1-3 GT
0 1 23 13
l.H , 2
1.30 l
1.17
••*» •*••• ^ m Qg «M <•••> 2
.91 2
.78 2
2 — 57.
IE TOP 5 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
0-1 1-3 GT 3 0-1 1-3 GT
0 1 23 13
.72 — 1
.84
— 1.17
.51 1
.51 2
— .43 3
2 70.
3
12
.02
.44
.71
.90
.06
.89
33.
3
12
.21
.80
.66
.84
.32
.49
67.
0-1
11
4.85
3.13
.64
2.64
.30
.46
45.
STA1
0-1
11
2.78
1.43
.51
1.82
.43
.84
45.
1-3
41
3.21
3.84
1.19
3.15
.96
.81
59.
3ILITY
1-3
41
2.08
1.42
.68
1.62
.61
.89
51.
GT 3
17
3.56
3.96
1.11
' 8.07
5.14
4.62
41.
F
GT 3
17
1.89
1.49
.79
1.97
1.09
1.38
47.
Statistics are not presented for cases with less than 6 data pairs
249
-------
TABLE E-33
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: TRACY POWER PLANT
MODEL: CTDM(DEGR2) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR
AVERAGE OF TI
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR
0-1 1-3 GT 3
0 1 23
— — loll
— 1.85
— — 1.67
— — ' 1.66
— 2.24
— — 1.34
2 — — 57.
[E TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 1 23
. - — .72
— . 82
— 1.14
— — . 63
— . 77
— — . 67
2 57.
0-1 1-3 GT 3
1 3 ' 12
— — 2 . 02
„__ 2 . 13
— _ i.os
„„- — _ i.2S
_-_ — _ .38
— — .36
— — 50 =
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
1 3 12
— - 1.21
— .36
— > .47
— 1.22
— 1.02
— - 2 . 19
42.
0-1
11
4«85
4.16
.86
3.72
.59
.69
18.
STAI
0-1
11
2.78
1.36
.49
1.89
.46
.94
36.
1-3
38
3.30
3.27
.99
7.04
4.55
4.60
16.
3ILITY
1-3
38
2.13
1.01
.47
2.88
1.82
3.83
13.
GT 3
16
3.49
4.09
1.17
5.31
2.32
1.98
25.
F
GT 3
16
1.82
1.41
.78
1.61
.78
1.01
38.
* Statistics are not presented for cases with less than 6 data pairs
250
-------
TABLE E-34
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: TRACY POWER PLANT
MODEL: COMPLEX I THRESHOLD: .00 US/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3
# OF DATA PAIRS 0 1
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02 '
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE TOP 5 VALUES
STABILITY
WIND SPEED (M/S) 0-1 1-3
# OF DATA PAIRS 0 1
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS) - —
RMS ERROR
V/C02 '
v/ccocp)
% WITHIN FACTOR 2
GT 3
23
1.11
.97
.88
1.01
.83
.95
30.
FROM
D
GT 3
23
.72
.67
.93
.63
.76
.81
39.
0-1 1-3 GT 3
2 3 13
2 . 05
3.20
1.56
1.71
.70
.45
54 .
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
2 3 13
1.18
2 . 14
1.81
!..22
1.06
.59
46.
0-1
11
4.35
15.82
3.26
15.23
9.87
3.02
18.
STA
0-1
11
2.78
9.41
3.38
8.72
9.86
2.91
18.
1-3
41
3.21
13.78
4.29
12 . 01
13.99
3.26
10.
BILITY
1-3
41
2.08
8.34
4.01
7.55
13.20
3.29
20.
GT 3
17
3.56
5.60
1.57
4.39
1.52
.97
29.
F
GT 3
17
1.39
3.77
2.00
2.32
1.51
.75
47-
* Statistics are not presented for cases with less than 6 data pairs
251
-------
TABLE E-35
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITES TRACY POWER PLANT
MODEL: RTDM (DEFAULT) THRESHOLD; .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS.
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
0-1 1-3 GT 3
0 1 23
. i . n
— 1.17
— _ 1.05
— 1.19
1.16
!.10
2 39.
£ TOP 5 VALUES FROM
STABILITY D
0-1 1-3 GT 3
0 1 23
.72
.79
1 . 10
. 34
— — 1.36
— — 1.23
2 48.
0-1 1-3 GT 3
2 3 13
— 2.05
— — 1.53
„**« «,«•>«, f A
« / 'B
— — 1.90
A e
e 0 3
— — ~ 1 . IS
— 54 .
EACH 1-HOUR PERIODS
STABILITY E
0-1 1-3 GT 3
2 3 13
!.18
1 . 09
.92
1.19
. Io02
lolo
54 .
0-1
11
4.85
8.66
1.79
10.35
4.56
2.55
18.
STAI
0-1
11
2.78
2.. 87
1.03
3.33
1.44
1.39
36.
1-3
41
3.21
4.92
1.53
4.77
2e21
1.44
37.
JILITY
1-3
41
2.08
2.08
1.00
1.96
.88
.38
41.
GT 3
17
3.56
2.82
.79
4.08
1.31
1.65
47.
F
GT 3
17
1.89
1.63
.86
1.59
.71
.82
53.
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
* Statistics are not presented for cases with less than 6 data pairs
252
-------
TABLE E-36
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: CF3BR SITE: TRACY POWER PLANT
MODEL: RTDM (ONSITE) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3
f OF DATA PAIRS 0 1
MEAN FOR OBS
MEAN FOR PRE —
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE TOP 5 VALUES
STABILITY
WIND SPEED (M/S) 0-1 1-3
# OF DATA PAIRS 0 1
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
GT 3
23
1.11
1.78
1.61
1.88
2.88
1.79
35.
FROM
D
GT 3
23
.72
1.02
1.42
1.07
2.22
1.56
43.
0-1 1-3 GT 3
2 3 12
2.02
— 2 . 19
— 1.08
2 . 55
— 1.59
. 1.46
— 42 .
EACH 1-HOUR PERIOD:
STABILITY E
0-1 1-3 GT 3
2 3 12
1.21
1.20
. 99
1.42
• • « • •• 1 • 3 8
1.40
58.
0-1
11
4.85
1.35
.28
5.90
1.48
5.33
36.
STAI
0-1
11
2.78
1.03
.37
2.52
.82
2.22
36.
1-3
41
3.21
1.87
.58
2.88
.80
1.38
63.
JILITY
1-3
41
2.08
1.17
.56
1.64
.62
1.10
61.
GT 3
17
3.56
2.65
.74
3.41
.92
1.24
59.
F
GT 3
17
1.89
1.35
.72
1.12
.35
.49
82.
* Statistics are not presented for cases with less than 6 data pairs
253
-------
TABLE E-37
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACERS S02 SITE; WESTVACO LUKE
MODEL: CTDM THRESHOLDS .00 uS/m**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/GS2
V/(CoCp)
% WITHIN FACTOR 2
AVERAGE OF THE
0-1
98
.36
.48
1.33
.80
4.89
3.68
27.
TOP 10
1-3
704
.31
.40
1.29
.96
9.39
7.30
21.
VALUES
GT 3
1180
.30
.06
.19
.44
2.12
11.04
8.
FROM
STABILITY D
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
0-1
98
.11
.11
.95
.22
3.76
3.94
30.
1-3
704
.10
.08
.79
.24
5.82
7.33
21.
GT 3
1180
.07
.01
.10
.10
1.73
17.59
5.
STABILITY E
0-1
33
.19
.19
1.02
,58
9.70
io'48
30.
1-3 GT 3
271
.48
.57
1.20
1.06
4.93 8
4.10 9
24.
639
.30
.27
.88
.87
.22
.38
8.
STABILITY
0-1
75
.18
.09
.50
.43
5.93
11.76
17.
1-3
497
.39
.42
1.07
1.00
6.41
6.00
19.
F
GT 3
845
.37
.40
1.07
.97
6. .85
6.42
10.
EACH 1-HOUR PERIOD:
STABILITY E
0-1
33
.07
.03
.31
.16
5.89
11.59
24.
1-3 GT
271
.15
.12
.80
.28
3.73 3
4,64 7
21.
3
639
.08
.03
.40
.14
.22
.96
6.
STABILITY
0-1
75
.07
.02
.27
.19
7.37
27.75
15.
1-3
497
.12
.08
.67
.25
3.95
5.91
18.
F
GT 3
845
.11
,05
.49
.23
4.14
8.51
12.
254
-------
0-1
100
.36
1.88
5.28
2.55
1-3
732
.31
1.46
4.78
2.88
GT 3
1167
.30
.34
1.13
1.23
0-1
34
.18
2.15
11.65
4.08
1-3
270
.47
3.14
6.62
4.97
GT 3
640
.30
1.60
5.27
3.35
0-1
71
.18
1.45
8.12
2.64
1-3
497
.40
3.09
7.82
4.99
GT 3
846
.37
2.10
5.66
3.79
TABLE E-38
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: S02 SITE: WESTVACO LUKE
MODEL: CTDM (DE6R1) THRESHOLD: .00 uS/m**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR QBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02 51.11 88.38 16.20 487.96 109.67 121.80 213.91 159.48 104.00
V/(CoCp) 9.69 18.50 14.39 41.90 16.57. 23.11 26.97 20.39 18.38
% WITHIN FACTOR 2 10. 22. 17. 15. 19. 19. 7. 19. 19.
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S) 0-1
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS) 5.72
RMS ERROR
V/C02 73.43 31.49 5.42 278.26 27.10 29.77 93.96 35.81 19.83
V/(CoCp) 12.83 11.51 8.73 29.57 8.12 10.76 18.06 9.42 8.12
% WITHIN FACTOR 2 11. 22. 12. 12. 23. 19. 8. 21. 18.
255
1-3
0 732
1 .10
3 .27
2 2.74
4 .55
GT 3
1167
.07
.05
.62
.17
0-1
34
.07
.62
9.41
1.09
.1-3
270
.15
.49
3.34
.76
GT 3
640
.08
.21
2.77
.42
0-1
71
.07
.36
5.20
.68
1-3
497
.12
.47
3.80
.74
GT 3
846
.11
.27
2.44
.49
-------
TABLE E-39
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SO2 SITE; WESTVACO LUKE
* MODEL: CTDM (DEGR2) THRESHOLD: .00 uS/m**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
STABILITY F
0-1
87
.39
2.23
5.69
2.35
1-3
678
.31
2.03
6.55
2.58
GT 3
1075
.30
1.20
3.97
1.31
0-
*
2.
11.
2.
•i
28
19
18
22
50
1-3
214
.52
2.17
4.20
2.43
GT
1
4
1
3
539
.31
.40
.46
.65
0-
«
2.
9.
2.
1
57
22
03
35
32
1-3
388
.45
2.35
5.26
2.52
GT 3
717
.40
1.59
4.00
1.71
35.83 69.27 18.72 165.95 22.12 27.65 114.76 31.91 18.60
6.30 10.57 4.72 14.79 5.27 6.20 12.28 6.07 4.65
% WITHIN FACTOR 2 10.
9.
22.
4.
14.
17.
11.
11.
15.
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
87 678 1075 28 214 539
.12 .10 ,,07 .07 .16 .08
.77 .45 .22 .64 .42 .25
6.40 4.63 2.97 8.97 2.60 3.13
.89 .60 .25 .80 .53 .31
54.75 37.94 12.10 124.42 10.44 15.18
8.56 3.20 4.08 13.36 4.01 4.34
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2 13.
10.
30.
7.
16.
24.
STABILITY F
0-1 1-3 GT 3
57 388 717
.14
.44
3.11
.50
78.09 12.51
11.35 4.02
9. 17.
.09
.59
6.S8
.76
.12
.26
2.17
.36
9.05
4.17
23.
256
-------
TABLE E-40
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: S02 SITE: WESTVACO LUKE
MODEL: COMPLEX I THRESHOLD: .00 uS/m**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY
WIND
f OF
MEAN
MEAN
BIAS
RMS
V/C02
SPEED (M/S)
DATA PAIRS
FOR OBS
FOR PRE
(PRE/OBS)
ERROR'
0-1
98
.36
.00
.00
.69
3.63
V/(CoCp) 1104.7
% WITHIN FACTOR 2
AVERAGE OF THE
13.
TOP
1-3
704
.31
.00
.01
.71
5.03
895.7
16.
D
GT 3
1180
.30
.00
.00
.42
1.96
630.4
11.
10 VALUES FROM
STABILITY
WIND
f OF
MEAN
MEAN
BIAS
SPEED (M/S)
DATA PAIRS
FOR OBS
FOR PRE
(PRE/OBS)
RMS ERROR
V/C02
0-1
98
.11
.00
.00
.20
3.32
1-3
704
.10
.00
.00
.23
5.41
D
GT 3
1180
.07
.00
.00
.10
1.88
STABILITY
0-1
33
.19
2.08
11.12
8.18
1914.6
172.3
30.
1-3
271
.48
5.63
11.81
11.11
542.4
45.9
18.
E
GT 3
639
.30
6.68
22.01
9.34
1051.9
47.8
12.
STABILITY
0-1
75
.18
.83
4.71
5.23
881.3
187.0
27-
1-3
7
19
14
497
.39
.74
.62
.82
1410.3
71.9
17-
F
GT 3
845
.37
7.73
20.80
12.12
1063.9
51.1
12.
EACH 1-HOUR PERIOD:
STABILITY
0-1
33
.07
.21
3.11
.81
145.44
1-3
271
.15
.78
5.31
1.59
116.94
E
GT 3
639
.08
.82
10.68
1.25
263.10
STABILITY
0-1
75
.07
.11
1.65
.77
126.20
1-3
1
8
2
280
497
.12
.06
.59
.07
-.09
F
GT 3
345
.11
.94
8.47
1.54
192.48
V/(CoCp)
2149.0 2389.5 1310.7
% WITHIN FACTOR 2 13.
16.
11.
46.72 22.01 24.64
27. 20. 14.
76.61 32.60 22.73
24. 18. 13.
257
-------
TABLE E-41
•EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SO2 SITE: WESTVACO LUKE
MODEL: RTDM(DEFAULT) THRESHOLD: .00 uS/M**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND
1 OF
MEAN
MEAN
BIAS
RMS
V/C02
SPEED (M/S)
DATA PAIRS
FOR OBS
FOR PRE
(PRE/OBS)
ERROR
V/(CoCp)
% WITHIN FACTOR
0-1
98
.36
.01
.02
.63
3.60
180.11
2 15.
AVERAGE OF THE TOP
1-3
704
.31
.01
.02
.70
5.00
214.57
17.
10 VALUES
GT 3
1180
.30
.01
.04
.41
1.85
46.08
11.
FROM
STABILITY D
WIND
# OF
MEAN
MEAN
BIAS
RMS
V/C02
SPEED (M/S)
DATA PAIRS
FOR OBS
FOR PRE
(PRE/OBS)
ERROR
V/(CoCp)
% WITHIN FACTOR
0-1
98
.11
.00
.01
.20
3.31
331.97
2 15.
1-3
704
.10
.00
.01
.23
5.39
576.49
17.
GT 3
1180
.07
.00
.02
.10
1.83
97.32
11.
0-1
33
.19
.84
4.50
3.29
309.10
68.67
30.
EACH 1-
1-3 GT
1
3
3
39
12
HOUR
271
.48
.52
.20 2
.00 1
.61 25
.39 a
21.
PERIOD
3
639
.30
.86
.83
.52
.12
.89
20.
£
STABILITY E
0-1
33
.07
.08
1.26
.33
24.16
19.16
27-
1-3 GT
1
10
7
271
.15
.21
.43 1
.48
.85 8
.56 5
19.
3
639
.08
.11
.42
.22
.06
.66
22.
0-1
75
.18
.35
1.96
2.06
136.16
69.50
27.
1-3
497
.39
2.22
5.63
4.11
108.73
19.33
19.
STABILITY
0-1
75
.07
.05
.75
.37
28.95
38.69
25.
1-3
497
.12
.31
2.47
.59
22.45
9.09
19.
GT 3
845
.37
1.91
5.13
3. IS
72.03
14.04
14.
F
GT 3
345
.11
.24
2.13
.45
16.61
7.81
15.
258
-------
TABLE E-42
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SO2 SITE: WESTVACO LUKE
MODEL: RTDM(ONSITE) THRESHOLD: .00 uS/m**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY
WIND SPEED (M/S) '
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CQCp)
% WITHIN FACTOR 2
AVERAGE OF THE
0-1
98
.36
.30
.83
1.01
7.89
9.53
12.
TOP
1-3
704
.31
.19
.61
1.02
.10.47
17.28
13.
D
GT 3
1180
.30
.11
.36
.39
1.66
4.62
16.
10 VALUES FROM
STABILITY
WIND SPEED (M/S)
f OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
0-1
98
.11
.06
.53
.20
3.10
5.80
17.
1-3
704
.10
.04
.39
.26
6.70
17.16
11.
D
GT 3
1180
.07
.02
.23
.09
1.60
7.02
11.
STABILITY E
0-1
33
.19
.02
.10
.30
2.61
26.75
6.
1-3 GT
271
.48
.25
.52
1.00
4.36 6
8.32 7
13.
3
639
.30
.26
.84
.75
.09
.21
16.
STABILITY
0-1
75
.18
.05
.30
.35
3.87
12.74
16.
1-3
497
.39
.22
.56
.80
4.15
7.44
14.
F
GT 3
845
.37
.41
1.09
1.15
" 9.58
8.75
19.
EACH 1-HOUR PERIOD:
STABILITY E
0-1
33
.07
.00
.05
.12
3.11
64.43
3.
1-3 GT
271
.15
.05
.35
.30
4.09 4
11.76 7
11.
3
639
.08
.04
.55
.16
.09
.41
13.
STABILITY
0-1
75
.07
.01
.16
.17
6.03
37.17
13.
1-3
497
.12
.05
.39
.24
3.74
9.58
13.
F
GT 3
845
.11
.07
.64
.30
7.15
11.19
13.
259
-------
TABLE E-43
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SO2 SITES WIDOWS CREEK
MODEL: CTDM THRESHOLD? .00 ug/m**3
(Concentrations given in units of micrograms per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
28 333 2051 25 372 1396 50 367 320
165.5 71.0 57.8 407.5 93.3 64.9 161.8 84.6 105.6
601.4 186c9 63.7 917.1 290.9 96.4
3.6 2.6 1.1 2.3 3.1 1.5
897.9 457.6 149.8 1273,9 628.0 312.3
29.4 41.5 6.7 9.3 45',3 23*2
8.1 15.8 6.1 4.3 14.5 15.6
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2 14.
23.
27.
28.
20.
20.
459.8 407.5 165.0
2.8 4.8 1.6
832.1 842.6 387.2
26.4 99.1 13.5
9.3 20.6 8.6
12. 16. 22,
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIOD:
WIND SPEED (M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2
STABIj
0-1 !
28
33.58
69.39
2.07
79.90
5.66
2.74
21.
LITY
L-3
333
22.
23.
1.
72.
10.
9.
D
GT 3
2051
26 17.
98 8.
08
82 25.
70 1.
93 4.
28.
0-
97
44
47
27
98
21
20.
STABI
-1
25
74
134
1
236
10
5
LITY
1-3
372
.58
.51
.80
.26
.03
.56
32.
E
GT 3
1396
26.76
37.65
1.41
89.00
11.06
7.86
24.
STAB:
0-1
50
19.40
11.81
.61
39.11
4.06
6.68
17.
ELITY
1-3
367
33
63
2
138
17
3
F
GT
.10
.61
.07
.92
.61
.50
13.
1 3
320
2'
5
11
2
1
260
-------
TABLE E-44
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: S02 SITE: WIDOWS CREEK
MODEL: CTDM (DEGR1) THRESHOLD: .00 ug/m**3
(Concentrations given in units of micrograms per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS 27 321 2012 25 360 1364
MEAN FOR .OBS 157.1 72.1 57.9 407.5 92.7 65.1
276.3 199.5 135.3 120.9 202.1 132.7
1.8 2.8 2. .3 2.2 2.0
887.2 579.9 532.6 1212.4 636.6 537.1
31.9 64.7 84.5 8.9 47.2 68.0
18.1 23.4 36.2 29.9 21.6 33.4
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2 15.
13.
13.
28.
11.
9.
48 353 299
166.9 85.9 36.3
214.1 191.4 174.4
1.3 2.2 2.0
834.1 631.4 694.3
25.0 54.0 64.9
19.5 24.2 32.1
13. 7. • 14.
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIOD:
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CGCp)
% WITHIN FACTOR 2 15.
STA
0-1
27
32.5
29.8
.92
126.4
15.1
16.5
: 15.
BILITY 1
1-3
321
22.6
28.7
1.27
108.7
23.2
13.3
17-
D
GT 3
2012
17.9
20.5
1.15
88.9
24.6
21.5
11.
STA
0-1
25
74.6
17.0
.23
198.7
7.1
31.2
20.
BILITY
1-3
360
27.0
27.3
1.01
101.2
14.1
13.9
13.
E
GT 3
1364
19.5
21.0
1.08
128.3
43.4
40.4
9.
STA
0-1
48
33.8
30.4
.90
115.8
11.7
13.0
15.
BILITY
1-3
353
20.7
31.4
1.52
126.3
37.2
24.6
10.
F
GT 3
299
20.3
29.3
1.45
176.5
76.0
52.5
16.
261
-------
TABLE E-45
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: S02 SITE: WIDOWS CREEK^
MODEL: CTDM (DEGR2) THRESHOLD: .00 ug/m**3
(Concentrations given in units of micrograms per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
WIND SPEED(M/S)
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/Co2
V/(CoCp)
% WITHIN FACTOR 2
9 267 1906
84.4 79.0 58.2
494.2 430.1 369.3
5.89 5.45 6.34
898«9 775.0 '705o9
113o5 96.4 147.0
19.4 17.7 23.2
0* 10. 22.
19 319 1302
519.6 89.8 66.0
334.3 460.0 404.4
.64 5.12 6.13
1589.5 875.7 812.0
9.4 95.1 1S1.4
14.6 18.6 24.7
5. 7. 13.
7 143 159
160.9 77.7 132.4
254.9 334.2 381.3
1.58 4.30 2.83
445.8 813.5 637.2
7.7 109.8 23.2
4.9 25,5 8.0
14. 11. 16.
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIODS
STABILITY D STABILITY E
^^^^•^«»C»^«>^«»«»^^«l<»^«»«>t« H>W>^^^^^^^^^^^^M»«B^^«
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
9 267 1906 19 319 1302
19.5 24.5 17.9 90.3 26.7 19.6
121.6 99.0 63.9 81.9 115.8 77.5
6.24 4.05 3«57 .91 4.34 3.95
268.0 212.8 136.7 346.6 276.7 182.3
139.0
30.3
% WITHIN FACTOR 2 22.
75.7 58.3 14.7 107.7 86.3
18.7 16.3 16.2 24.8 21.8
16. 23. 5. 12. 16.
STABILITY F
0-1 1-3 GT 3
B^VP^^^ «>^^^ «=«i»*»«D«E
7 143 159
54.6 21.9 25.9
73.2 111.8 76.6
1.34 5.12 2o96
140.1 325.3 151.9
6.6 221.7 34.4
4.9 43.3 11.6
14. 3. 17.
262
-------
TABLE E-46
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES 'TRACER: SO2 SITE: WIDOWS CREEK
MODEL: COMPLEX I THRESHOLD: .00 ug/m**3
(Concentrations given in units of micrograms per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 • 1-3 GT 3
31 341 2118 26 335 1422 50 369 323
158.8 70.2 57.1 392.8 92.9 64.4 161.8 84.4 105.
.38 3.30 18.6 1729.8 1108.3 470.3 1637.8 1735.6 581.
.00 .05 .33 4.4 11.9 7.3 10.1 20.6 5.5
84.6 2014.5 1944.2 1090.2 2959.5 3146.0 1113.
2.2 26.3 438.5 286.9 334.4 1390.1 112.6
6.8 6.0 36.7 39.3 33.0 67.6 20.4
6. 19. 15. 6. 10. 14. 3. 11.
WIND SPEED(M/S)
* OF DATA PAIRS
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
516.3 184.9
10.6 6.9
4391.0 147.7
% WITHIN FACTOR 2 16.
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS 31 341 2118 26 385 1422
32.57 22.19 17.71 71.81 26.72 19.30
.04 .37 2.09 201.81 125.92 54.67
.00 .02 .12 2.81 4.71 2.83
100.9 55.4 22.6 262.6 223.8 131.0
9.60 6.24 1.63 13.37 70.14 46.07
8176 376.6 13.82 4.76 14.89 16.27
MEAN FOR OBS
MEAN FOR PRE
BIAS (PRE/OBS)
RMS ERROR
V/C02
V/(CoCp)
% WITHIN FACTOR 2 16.
2.
12.
3.
11.
14.
STABILITY F
0-1 1-3 GT 3
50 369 323
33.10 20.48 21.93
176.47 203.58 70.85
5.33 9.94 3.23
328.7 392.0 143.4
98.63 366.3 42.74
13.50 36.85 13.23
10. 5. 15.
263
-------
TABLE E-47
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SO2 SITE: WIDOW CREEK
•»
MODEL: RTDM (DEFAULT) THRESHOLD: =00 ug/m**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS 31 341 2118 26 385 1422 50 369 323
MEAN FOR OBS 158.8 70.2 57.1 392.8 92.9 - 64.4 161.8- 84.4 105.0
MEAN FOR PRE 191.1 315.2 229.2 602.6 502.2 297.9 421.5 634.2 452,5
BIAS (PRE/OBS) 1.20 4.49 4.02 1.53 5.41 4.63 2.60 7.52 4.31
RMS ERROR 896.9 799.4 622.6 1721.0 950.9 731.1 1006.2 1121.0 782.4
V/C02 31.89 129.7 119.0 19.20 104.9 129.0 38.7 176.5 55,5
V/(CoCp) 26.50 28.9 29.7 12.51 19.4 27.9 14.8 23.5 12.9
% WITHIN FACTOR 2 16. 3. 8. 4. 4. 6. 12= 2. 10,
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS 31 341 2118 26 385 1422 50 369 323
MEAN FOR OBS 32.57 22.19 17.71 71.81 26.72 19.30 33.10 20.48 21.93
MEAN FOR PRE 19.11 34.75 25.56 60.26 57.84 33.83 47.64 74.29 51.71
BIAS (PRE/OBS) .59 1.57 1.44 .84 2.16 1.75 1.44 3.63 2.36
RMS ERROR 123.2 94.3 72.4 223.5 115.9 87.4 127.8 134.6 94.5
V/C02 14.31 18.05 16.69 9.69 18.79 20.50 14.91 43.20 18.56
V/(CoCp) 24.39 11.53 11.57 11.54 3.68 11.69- 10.36 11.91 7.37
% WITHIN FACTOR 2 19. 7. 9. 8. 9. 12. 14. 9. 17.
264
-------
TABLE E-48
EVALUATION STATISTICS FOR DATA SUBSET PAIRED IN TIME
(SUBDIVIDED INTO STABILITY AND WIND SPEED CLASSES)
1-HOUR AVERAGES TRACER: SO2 SITE: WIDOWS CREEK
MODEL: RTDM (ONSITE) THRESHOLD: .00 ug/m**3
(Concentrations given in units of microseconds per cubic meter.)
HIGHEST 1-HOUR VALUES, UNPAIRED IN SPACE:
STABILITY D STABILITY E STABILITY F
WIND SPEED(M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
# OF DATA PAIRS 31 341 2118 26 385 1422 50 369 323
MEAN FOR OBS 158.81 70.19 57.05 392.80 92.85 64.36 161.84 84.38 104.99
MEAN FOR PRE .62 .41 5.09 344.30 112.15 36.20 639.39 209.26 53.20
BIAS (PRE/OBS) .00 .01 .09 .88 1.21 .56 3.95 2.48 .51
RMS ERROR 516.28 186.66 93.23 803.85 261.53 134.00 1810.39 660.08 246.69
7/C02 10.57 7.07 2.67 4.19 7.93 4.33 125.14 61.19 5.52
V(CoCp) 2723.5 1220.3 29.94 4.78 6.57 7.71 31.67 24.68 10.90
% WITHIN FACTOR 2 16. 1. 9. 42. 31. 13. 22. 21. 21.
AVERAGE OF THE TOP 10 VALUES FROM EACH 1-HOUR PERIOD:
STABILITY D STABILITY E STABILITY F
WIND SPEED (M/S) 0-1 1-3 GT 3 0-1 1-3 GT 3 0-1 1-3 GT 3
I OF DATA PAIRS 31 341 2118 26 385 1422 50 369 323
MEAN FOR OBS 32.57 22.19 17-71 71.81 26.72 19.30 33.10 20.48 21.93
MEAN FOR PRE .16 .07 .66 73.80 22.46 5.10 79.11 29.37 7.59
BIAS (PRE/OBS) .00 .00 .04 1.03 .84 .26 2.39 1.43 .35
RMS ERROR 100.88 55.57 23.74 132.4 62.38 27.00 203.74 74.39 32.44
V/C02 9.59 6.27 1.80 3.40 5.45 1.96 37.89 13.19 2.19
V/(CoCp) 1981.3 1971.2 48.16 3.31 6.48 7.40 15.35 9.20 6.32
% WITHIN FACTOR 2 16. 1. 5. 38. 30. 13. 24. 20. 17.
265
-------
APPENDIX F
SCATTER PLOTS OF PEAK HOURLY MODEL
PREDICTIONS AMD OBSER¥AT10NS
266
-------
APPENDIX F
SCATTER PLOTS OF PEAK HOURLY MODEL
PREDICTIONS AND OBSERVATIONS
Scatter plots of modeled predictions versus observations were
generated to augment the statistic tables of model evaluation results
presented elsewhere in this report. For the FSPS, Westvaco, and
Widows Creek sites, plots of the predicted-to-observed ratio of
peak-concentrations versus plume travel distance are presented as
residual plots. In these plots, each point represents the average of
the top 10 concentrations at each monitor. A guide to the figures in
this appendix is given below.
Figures
F-l through F-6
F-7 through F-12
F-13 through F-18
F-19 through F-24
F-25 through F-32
F-33 through F-38
F-39 through F-44
F-4S thorugh F-50
F-51 through F-56
F-57 through F-62
F-63 thorugh F-68
F-69 through F-74
Description
Predicted vs. observed peak 1-hour SF6
concentrations at CCB
Predicted vs. observed peak 1-hour
CF3Br concentrations at CCB
Predicted vs. observed peak 1-hour SF6
concentrations at HBR
Predicted vs observed peak 1-hour CF3Br
concentrations at HBR
Predicted vs. observed peak 1-hour SFg
concentrations at FSPS
Predicted vs. observed peak 1-hour
CF3Br concentrations at FSPS
Predicted observed ratio vs. distance for
SF6 concentrations at FSPS
Predicted observed ratio vs. distance for
CF3Br concentrations at FSPS
Predicted vs. observed peak 1-hour S02
concentrations at Westvaco
Predicted vs. observed peak 3-hour S02
concentrations at Westvaco
Predicted/observed ratio vs. distance for
1-hour S02 concentrations at Westvaco
Predicted vs. observed peak 1-hour S02
concentrations at Widows Creek
267
-------
FiRures Description
F-75 through F-80 Predicted vs. observed peak 3-hour S(>2
concentrations at Widows Creek
F-81 through F-86 Predicted/observed ratio vs. distance for
1-hour SOj concentrations at Widows Creek
268
-------
256-
2M
in
IM
•H*
OBSERVED VS. PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODEL I CTOM TRACERt SF6
(AVERAGE DBS- £6.0 AVERAGE PRE- SB .8 RHBE- Z8 .7 « OF HOURS-1991
IM IM 2M 26*
OBSERVED CONCENTRATIONS (US/MXK3)
Figure F-l
-------
to
S
n
1
o
2M
17S-
ISC
I2S-
IM
7S-
2S-
08SERVED VS . PREDICTED CONCENTRATIONS AT CINDER CONE 0UTTE
1-HOUR AVERAGES MODELS CTDM(DEGRi) TRACER» SF6
(AVERAGE OBS- 20.0 AVERAGE PRE- 26 .2 RM8E- 20.7 M OF HOURS-10O!)
>• J **
• i
<%* **:***
» * »
-*-
25 58 76 IM 92G I6« 176
OBSERVED CONCENTRATIONS (US/M»13)
Figure F- 2
-------
§
p
CJ
2M
175-
150
128-
IM
g 76
26-
OBSERVED VS. PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODEL: CTDM(DEGR2) TRACERS SF6
(AVERAGE OBS- 26.0 AVERAGE PRE- 6.6 RM8E- 37.9 * OF HOURS-1001
*
*,
•*•
* +
V *
25 58 76 IM 126 IM 176
OBSERVED CONCENTRATIONS (US/M1M3)
Figure F-3
-------
aer
96*
IM*
OBSERVED VS. PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODELi COMPLEX I TRACERS SF6
(AVERAGE 088- C0 .0 AVERAGE PRE- 4« 4t RH8E- 49^ • HOURS- 100)
**
* *
*
—r~
IM
—r™
IM
T
2M
4H
OBSERVED CONCENTRATIONS CUS/MJKX3}
Figure F-4
-------
4M-
IS*
OBSERVED VS . PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
I-HOUR AVERAGES MODELi RTOM (DEFAULT) TRACERS SF6
(AVERAGE O88- C0 .0 AVERAGE PRE-
RH8E- 58.8 • HOURS* 1001
IM 2M 2M 9M 9M 4M
OBSERVED CONCENTRATIONS (US/M113)
Figure F-5 .
4M
-------
4M
86i-
IM-
st-
OBSERVEO VS. PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODELS RTDM CONSITE) TRACERI SF6
(AVERAGE O88- £0.0 AVERAGE (P«E- £S «« RM8E« £0.1 « HOURS- 1001
>***
<»* *
IM 1S8 2M 2St SM 8S« 4»B
OBSERVED CONCENTRATIONS CUS/MX13)
Figure F-6
-------
Ni
^-1
Ln
m
g
LJ
156
135-
120-
105-
75-
60-
45-
30'
16-
OBSERVED VS . PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODEL: CTDM TRACER: CF3BR
I AVERAGE OBS- 13.0 AVERAGE PRE- 81 .0 RMSE- 28.8 * OF HOURS-
* *
* *
,
16 30 46 00 76 00 106 120 136
OBSERVED CONCENTRATIONS (US/M**3)
Figure F-7
150
-------
1*0
oe
ro
CJ
I
4«H
38H
OBSERVED VS . PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODELS CTDM(DEGRl) TRACER: CF3BR
(AVERAGE DBS" 19.0 AVERAGE PRE- 7,5
RM8E- 19.9 • OF HOURS"
*
*
*
+
+
20 30 . 40 S0 80 70 M
OBSERVED CONCENTRATIONS (US/MX13 )
Figure F-8
-------
m
I
108
80-
00-
70-
60-
60-
40-
30-
20
10-
0T
OBSERVED VS . PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODEL: CTDM(DEGRS) TRACER« CF3BR
I AVERAGE DBS- 19.0 AVERAGE PRE-
RMSE- 20 .9 M OF HOURS-
:* *
I**!"1—' M
10 20
30 40 E0 60 70 80
OBSERVED CONCENTRATIONS (US/M**3)
Figure F-9
00
100
-------
tx)
•~J
CO
i si
135-
120-
m i«5-
UJ
7S-
68-
4S-
S6-
OBSERVED VS . PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODELS COMPLEX I TRACERt CF3BR
(AVERAGE DBS- 19 .t AVERAGE PRE- g<« ,9 RM8E- 28.7 » OF HOURS- 90)
* *
*
* *
**
+•»
85
45 88 7S 08 186 12*
OBSERVED CONCENTRATIONS (US/MJKX3)
Figure F-10
836
-------
IM
i as-
12*
IK
4K
1C
OBSERVED VS. PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODELt RTDM (DEFAULT) TRACERS CF3BR
I AVERAGE OB8- 13.0 AVERAGE PRE- 11.* RM8E- 84.8 • HOURB- 44)
IB
i
4K
l
M
7S M I*B l» 186
OBSERVED CONCENTRATIONS (US/MXX3)
Figure F-ll
-------
160
136
186-
ho
CO
o
76-
88-
46-
16-
OBSERVEO VS. PREDICTED CONCENTRATIONS AT CINDER CONE BUTTE
1-HOUR AVERAGES MODELt RTDH IONSITE) TRACERS CF3BR
C AVERAGE DBS- 19.8 AVERAGE PRE- 18.7 RH8E- 18.2 • HOURS-
+
."** +
* +
16
46 M 76 M tW 12*
OBSERVED CONCENTRATIONS (US/MXK3)
figure F-12
1S5
16*
-------
00
m
4M
350-
380
250-
200
150-
100
OBSERVED VS. PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELt CTDM TRACERi SFS
fAVERAGE O8S-23 .3 AVERAGE PRE- 47.8 RM8E- 98.9 « OF HOUR8-S0)
'« • I
100 IM 200 2S0 M0 W0
OBSERVED CONCENTRATIONS (US/MXK3)
B'igure F-13
-------
00
NJ
380-
„ 2S0-
rn
200
IS0-
LJ
OBSERVED VS . PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODEL tCTDM! DEGRU TRACER 8 SF6
(AVERAGE OBS-23 .3 AVERAGE PRE» 3® .8 «M8g» *W .3 M OF HOUR8-381
*
* «
* *
* *
MHI I •
j — —J • , T
100 IS0 200 250
OBSERVED CONCENTRATIONS (US/M1M3J
Figure F-14
-------
oo
CJ
ro
»*
M<
Q
•H
a
150
120-
00-
68-
30-
OBSERVED VS . PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODEL:CTDM(DEGR2) TRACER: SF6
fAVERAGE OBS-83 .3 AVERAGE PRE-18 .8 RMSE- 29 .4 • OF HOURS-SS)
0 Jim mttt HiiiHMi—i)—i »» i 1 ,-
30 60 00 120
OBSERVED CONCENTRATIONS (US/M**3)
Figure F-15
150
-------
00
!3
9M
268
SW-
OBSERVED VS. PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELi COMPLEX I TRACERS SF6
(AVERAGE 088-83.0 AVERAGE PRE-117.8 RM8E»1C9.1 « OF HOUft>«0»
* *
<*
•»
tsc Mi 2ft ate is*
OBSERVED CONCENTRATIONS (US/MXK3J
Figure F-16
4M
-------
00
Cn
7M
6M
3M-
2M
OBSERVED VS. PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELS RTDM CDEFAULT) TRACERt SF6
CAVERAGE 068-88.9 AVERAGE PRE-74 .7 RMSE-158 .8 • OF HOUR8-S81
IM 2M M0 4M CM Ml 7M *M
OBSERVED CONCENTRATIONS (US/M1I3)
Figure F-17
-------
I2M
iaei
eae
see-
7ea-
OBSERVED VS. PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODEL« RTOM CONSITEJ TRACER? SF6
(AVERAGE OBS-S3.3 AVERAGE PRE-92M RM8E-88 ,9 » OF HOUR8-38 J
M
00
S00-
LJ
n
Q
900-
20e-
tae-
100 200 300 400 S00 S00 700 000 000 1000
OBSERVED CONCENTRATIONS (US/M1X3)
Figure F-18
1100 1200
-------
00
s0a
4S0H
480
m 3501
J*
see
HI
or
L.)
250
200
150-
108
50
OBSERVED VS . PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELj CTDM TRACER:CF3BR
(AVERAGE DBS- 98.0 AVERAGE PRE» 43.1 RMSE" 33.1 • OF
SO 100 IE0 288 250 300 350 400
OBSERVED CONCENTRATIONS (US/MMM3)
Figure F-19
450
508
-------
see
00
co
m
>*
Xt
LT>
ft
n
I—
a
ft.
t~
o
a
LU
CJ
n
O
zee
sse-
i ea-
OBSERVED VS . PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELsCTDMfDEGRi) TRACER»CF3BR
(AVERAGE 083-118,2 AVERAGE PRE- 47.3 RMSE-J^.7 « OF HOURS»a<4)
* *
150 200 258 398 3SB 408
OBSERVED CONCENTRATIONS 1US/MMM3)
Figure F-20
-------
see
00
m
t*.
1.0
o
i -i
cc
H-
UJ
CJ
O
LJ
358
300-
288
i ea-
OBSERVED VS . PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELsCTDM(DEGRE) TRACERsCFSBR
(AVERAGE DBS- 91.6 AVERAGE PRE- 98 .5 RMSE- 53.3 « OF HOURS-
+ *
•»•
•» ^
se tee tee zee 2ee see 358
-------
tv)
vD
o
45*
2M-
!§•-
IM-
OBSERVED VS . PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELS COMPLEX £ TRACERiCF3BR
f AVERAGE OBS-iU.* AVERAGE W«"l7i
# OF
-8-
1M 2M 2M SM ISA 4M 4S« M
OBSERVED CONCENTRATIONS (US/MXI3)
figure F- 22
KM CM
-------
ISM
OBSERVED VS. PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODEL iRTDM (DEFAULT) TRACER ICF3BR
(AVERAGE OM-M«.a AVERAGE
.1 AMSE-A7* .S • OF HOUKS-8*)
IMC ISM 2Mt 2CM
OBSERVED CONCENTRATIONS (US/MXK3)
Figure F-23
-------
M
seee
4598
asoe-
to
3888-
2500-
UJ 2000-
CJ
isee-
1800-
OBSERVED VS . PREDICTED CONCENTRATIONS AT HOGBACK RIDGE
1-HOUR AVERAGES MODELsRTDM (ONSITE ) TRACER:CF3BR
(AVERAGE OBS-108.3 AVERAGE PRE-l^a .3 RMSE-37E ,3 » OF HOUR8-81>
s00 1000
!
3000
2000 2600 3000 3600 4000 4600 6000
OBSERVED CONCENTRATIONS JUS/M**3)
B'igure F-24
CTSKt
-------
m
i*
i-U
L.)
a
t_>
W
o
12
II-
10
7"
*•
I-
OBSERVED VS . PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODEL:CTDM TRACER: SF6 (MODEL)
(AVERAGE DBS-1 .36 AVERAGE PRE-1 .84 RMSE-2.11 * OF HOURS-110)
* +
+
+ •»
*
•»
•»• •»
****
»'«.•*
I.T* . * * +
.
*
T
2
4 5 6 7 0 0 10
OBSERVED CONCENTRATIONS (US/MM13)
Figure F-25
II
12
-------
z:
\
lO
o
i- t
Q:
LU
CJ
S
c_)
o
£
CJ
•—(
Q
12
II-
18
6-
4-
3-
2-
I-
OBSERVED VS . PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODELsCTDMtDEGRi ) TRACERS SF6
tAVERAGE OBS-J .9^ AVERAGE PRE-2 .07 RHSE-2 .84 • OF HOURS-109»
* +
•»•
+ *
* \
* *
*.+ **"* **
* * *
+
+
12
OBSERVED CONCENTRATIONS {US/MMM3)
Figure F-26
-------
10
m
( J
I—<
Q
16-
14-
12-
10
OBSERVED VS. PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODEL:CTDM(DEGR2) • TRACER> SF6
I AVERAGE OBS-t .94 AVERAGE PRE»2 .
RM8E-3.ee • OF HOURS-test
* +
* + *
+ r *
a 1*1* *** ,<
6 • 18 12 14 16
OBSERVED CONCENTRATIONS (US/MKX3)
Figure F-27.
ia
-------
9
it
7-1
s-
3-
2-
OBSERVED VS» PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODEL. tCTDM TRACER t SF6 CMOA)
CAVERACE 088-1.88 AVERAGE PRE-1 .77 RH8C»t .88 * OF HOURS-lie I
** +
. . * *•. *
* * ** *
^. *
199
0 i 2 a
f o
II 12
OBSERVED CONCENTRATIONS {US/MM13)
Figure F-28
-------
12
7H
OBSERVED VS . PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODELiCTDM TRACERt SF6 (RISE)
(AVERAGE OB8-1 .88 AVERAGE PRE-lM RM8E-1 .84 • OF HOURS-110)
V * *
* +*
+ * *
** ^
* * * *
2 3 4 68 7 8 8 18 II 12
OBSERVED CONCENTRATIONS (US/MKK3)
B'igure F-29
-------
2S
IS-
OBSERVED VS . PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODEL^COMPLEX I TRACER« SF6 CMODEL)
(AVERAGE 088-1 .89 AVERAGE PRE-8 „&<« RMSE-7 .37 « OF HOURC-UI)
*
* *
+ %* * * +
*
* * *
fr *
** I
* **
t* *
+
***** *
• Mi i «»
2f
OBSERVED CONCENTRATIONS (US/M1X3)
Figure F-30
-------
2t
li-
n-
14-
12
li
4-
OBSERVED VS . PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODELiRTDM (DEFAULT) TRACERiSFStMODEL»
(AVERAGE OBS-1 .89 AVERAGE PRE-9 .03 RH8E-3 .38 * OF HOUR8-111I
*
4
* * *
+ »
+ * +
* J
+
*
M «> », •—M-
• • !• 12 M 1C
OBSERVED CONCENTRATIONS (US/MKK3)
Figure F-31
-------
CJ
O
O
CO
s
12
II
It
7-
6-
2-
OBSERVED VS . PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODELiRTDM IONSITE ) TRACERiSF6 CMODEL)
(AVERAGE OBS-i .86 AVERAGE PRE-1.1* RM8E-S.ee * OF HOURS-U»)
*** *
A
r B i B I B c
4 S S 7 • 0 It
OBSERVED CONCENTRATIONS (US/M**3J
Figure F-32
12
-------
CJ
O
25
a
en
g
I
IS-
10
OBSERVED VS . PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODEL? CTDM TRACER:CF3BR
C AVERAGE QBS-2 .81 AVERAGE PRE-3 .MB RUSE-a .
* OF HOURS-110)
IB IS 20
OBSERVED CONCENTRATIONS (US/MXX3)
Figure F-33
2K
-------
CO
u)
O
to
H
Q
28-
IS
18-
6-
OBSERVED VS. PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODEL 8 CTDM CDEGRS3 TRACERiCFSBR
(AVERAGE OBS-S .8*4 AVERAGE PRE-8 MS RM8E-^4 .91 « OF HOURS-1081
+
**
+ *
* it •«
£
•» * *
****
»
1C IS 2« 26 M 16
OBSERVED CONCENTRATIONS (US/M**3)
Figure F-34
-------
U)
o
M
45-
40'
— 30
IS
10
OBSERVED VS. PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODELiCTDM (DEGR2) TRACERICF3BR
I AVERAGE OBS-a.84 AVERAGE PRE-3 .04 RM8E-4 .88 * OF HOURS-J831
:**.
****
**
1
ft
—r
16
—r~
36
20 26 30 36 40
OBSERVED CONCENTRATIONS (US/M**3)
Figure F-35
—i—
45
60
-------
28-
is-
it-
OBSERVED VS « PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODELt COMPLEX I TRACER1CF3BR
C AVERAGE OB*-£
AVEMACK PRE-» .8^4 MMSE-S.lt • OF HOURS-lllI
*
+ 4
*****
y**
*
>* *
i/**
^+* *
•
r-f
K
!•
e
18
29 26
t
M
M
4f
OBSERVED CONCENTRATIONS (US/MKX3)
Figure F-36
-------
CJ
o
Ul
S-
OBSERVED VS. PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODELt RTDM (DEFAULT) TRACERICF3BR
I AVERAGE 000-8.04 AVERAGE PRS-0 .70 RM0E-4.70 • OF HOUR0»tU>
**
* V
10 IS M 26
OBSERVED CONCENTRATIONS (US/MMK3)
Figure F-37
-------
18
18
*"
4-
OBSERVED VS. PREDICTED CONCENTRATIONS AT TRACY POWER PLANT
1-HOUR AVERAGES MODEL» RTDM IONSITE) TRACERICF3BR
(AVERAGE OB8»e.«4 AVERAGE PRE-1 .93 RM8E-9 .17 • OF HOURS-110)
* + * *
***
****
'-*v *V/* .
3&*. . •*:
THr""TTTT|
8 8 90 12 M 18
OBSERVED CONCENTRATIONS (US/MXX3)
figure F-38
18
-------
o
~J
21. Ml
It. NT
.2M
.IM
.•11-
PRE/OBS RATIO VS. DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HOUR VALUES MODELiCTDM + TRACERiSF6
* » ;* * **
*
+ »
* * * * ***** +
+
+
DISTANCE (M)
Figure F-39
-------
CJ
o
oo
1
M.tM
ia.ee*
S.888-
2.888-
.588
.288-
.•10-
PRE/OBS RATIO VS. DISTANCE AT TRACY POWER PLANT
AVC OF TOP 10 1-HOUR VALUES MODELtCTDM(DEGRi) TRACER|SF6
«• *
+ *
4
* + * 4. + **
* **
*$ *
* + + + +
^ 1 ******' +
* * +
* *
* +
* •»
TAAfl AAAA
/ ^^^p ^^^»^
DISTANCE (MS
Figure F-40
-------
CJ
o
vO
21.1
It.I
fi.
2.1
.SM
.2M
.IM
PRE/OBS RATIO VS . DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HOUR VALUES MODEL ICTDM(DEGR2) TRACERsSF6
+ +
* +
4000 6000 8000 7000 0000
DISTANCE (M)
Figure F-41
-------
?08. aea
,aa.aaa
sa.aoa
aa.aee
ta.aaa
s.ee0-
2.000-
1.000-
g •5fle
ca 2aa-
M V? *
JH .108-
.050-
.010-
.005-
aai-
PRE/OBS RATIO VS . DISTANCE AT TRACY PO^^eR PLANT
AVG OF TOP 40 i-HOUR VALUES MODEL iCOMPLEXI TRACER «SF6
** * *
* * **
*****
.* * * u * *
' ;'v -,*:•;, ;*'*;..*
*******
•«•
* t .**
****** I
* *
* * *
+ *
*
*
2000 3000 4000 5000
DISTANCE CM)
E'igure F-42
-------
60.000
20.000-
10.000
S.000
2.000-
1.000
.600
.200-
.100
.060
.010-
.005
.001
PRE/OBS RATIO VS . DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HOUR VALUES MODEL IRTDMCDEFAULTJ TRACERiSF6
* *
+
* +
* +
+
+
+ *
+ +
*+ *
****
1000 2000
4000 6000
0000 7000
DISTANCE (M)
Figure F-43
0000 1
-------
N)
20.000
10.000
S.000-
2.000-
.S00-
.200-
.050-
.010-
.001
PRE/OBS RATIO VS . DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HOUR VALUES MODEL|RTOM(ONSITE) TRACERiSF6
* *
*
•fr*
* * *
* .. • *. J
• * . 7 *. •
* * fr* + *
leee 20m aaae 4000 5000 6000 7000 0000
DISTANCE (M)
Figure F-44
-------
CO
H
Ul
M.tM
C.MT
2.
I.
.2M
. IM
PRE/OBS RATIO VS. DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HOUR VALUES + MODELiCTOM TRACERtCF3BR
* * +
.* * .v** «•
:<*
•*
%'. *
***
MM 7M»
DISTANCE CM)
Figure F-45
-------
S.VM
g
.SM-
.29*-
.»&•-
.•18-
PRE/OBS RATIO VS . DISTANCE AT TRACY POWER PLANT
AVG OF TOP 18 1-HOUR VALUES MOQELsCTDMlDEGRlJ TRACERICF38R
* * ******
* **
+* * * *«• «
******* *
* + * * * + '
* *
* ***/
** * *
+ •»
law 2M8
MM 9998 7M0 MM MM
DISTANCE CM)
Figure F-46
-------
OJ
H
Cn
M.tM
It.
K.
2.
.2M
.•10
PRE/OBS RATIO VS . DISTANCE AT TRACY POUER PLANT
AVC OF TOP 10 1-HOUR VALUES MODELiCTOMfOECR8) TRACERICF38R
* +* *
DISTANCE (M)
Figure F-47
-------
.100.000
50.000
20.eea
10.000
S.000
2.000
1.000
.S00H
g
g .200-|
£
.100-
.050-
.818-1
.005-1
.001
PRE/OBS RATIO VS . DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HOUR VALUES MODEL ICOMPLEXI TRACER|CF3BR
* +
* **
* \ ,;+M* +* >*. *** *
* !****. . *
+ % *
» + *
+
•»•
..*
1000 2000 3000 4000
5009 6800
DISTANCE CM!
Figure F-48
7000
0000 900i
-------
OJ
H
-•J
60.000
20.000-
10.000-
s.oeo-
2.000-
1.00
.500-
.200-
.100
.060H
.010
.085-1
.001
PRE/OBS RATIO VS. DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HR VALUES MODELiRTDMCDEFAULT) TRACERICF3BR
'
* + **
X + * * +
\ *>'. • .
* *
* +
+
1000 2000 9000 4000
S000 8000 7000
DISTANCE (M)
Figure F-49
-------
oo
it.
S.000-
2.000
.500-
.200-
.IM-
.060-
.018-
.ees-
.00f
PRE/OBS RATIO VS . DISTANCE AT TRACY POWER PLANT
AVG OF TOP 10 1-HR VALUES MODELsRTDMfONSITE) TRACER:CF3BR
* * +
* * * *
* **
~ , . -v
* * **+*+*«.
* *** *****: *** i* **
4+ * * *
** * *^
*+
*
* + *
* *
+ *
8000 2000 M00 4008
S000
DISTANCE
0000 90000
Figure F-50
-------
vfl
M-
12-
OBSERVED VS. PREDICTED CONCENTRATIONS AT MESTVACO LUKE
1-HOUR AVERAGES MODEL I CTDM TRACER I SO2
f AVERAGE O88- 9.99 AVERAGE PME* • . .«• MMK- 9.99 9 OF HOLMS-4M7I
v*% *
r£2 *
* v>> , + * *
4w£r * *
f F " "T f T •
• • !• 12 14 IS
OBSERVED CONCENTRATIONS (US/M1I3)
Figure F-51
-------
45-
2S-
OBSERVEO VS. PREDICTED CONCENTRATIONS AT UESTVACO UKE
1-HOUR AVERAGES MODEL I CTDMf OEGR 1 I TRACER I SOS
(AVERAGE 08«- 9.99 AVEMACE
MMW- 9 .»• • OF HOUM««47tt>
U)
M
O
is m s» M
OBSERVED CONCENTRATIONS
Figure F-52
-------
It-
M-
12-
«•
OBSERVED VS . PREDICTED CONCENTRATIONS AT MESTVACO LIKE
1-HOUR AVERAGES MODEL|CTDM(DEGR2J TRACERI SO£
r AVERAGE OBt- • .•4 AVERAGE PRE* 1.79 RHtE- t*9 • OF HOUR»»4M9I
t t It 12 M
OBSERVED CONCENTRATIONS (US/MMQK31
Figure F-53
If
It
-------
M
56-
OBSERVED VS . PREDICTED CONCENTRATIONS AT MESTVACO LUKE
1-HOUR AVERAGES MODELS COMPLEX I TRACERS SO2
r AVERAGE OB8- 0.89 AVERACE PRE»
MH8E
• OF
46-
96-
M26Ma64t4f
OBSERVED CONCENTRATIONS CUS/MKM3)
Figure F-54
-------
It-
14-
OBSERVEO VS . PREDICTED CONCENTRATIONS AT MESTVACO ULKE
1-HOUR AVERAGES MODEL tRTOU DEFAULT ) TRACERS SO2
I AVERAGE 00t» • .»» AVERAGE M»E- •
+
RH8E- « .IS • Of HOUMS^IMTI
• • It I* 14
OBSERVED CONCENTRATIONS (US/MW3)
Figure F-55 .
It
It
-------
U)
M
It-
16
14-
12-
OBSERVED VS . PREDICTED CONCENTRATIONS AT WESTVACO LUKE
1-HOUR AVERAGES MODEL iRTDMC QNSITE > TRACER t SO2
(AVERAGE OB8- • .»3 AVERAGE PRE» »,
*
RMSE» • ,•£ tt OF HOURS-«*»7)
* *
+
•f
*
.**
• 9i II 14
OBSERVED CONCENTRATIONS CUS/MWK3)
Figure F-56
It
It
-------
4"
2-
OBSERVED VS. PREDICTED CONCENTRATIONS AT WESTVACO LUKE
3-HOUR AVERAGES MODEL I CTDM TRACER I SOS
I AVERAGE OB8- • .•• AVERAGE W»E- • X9 RHM- • .M • Of HOUR»«|«1»)
• * * «
•
*•*
*»
. .
<*.*«
*.*
+
+
I
«
9 4 C • 7 •
OBSERVED CONCENTRATIONS (US/MXK3)
Figure F-57
!•
-------
u>
M
2S
2«
OBSERVED VS . PREDICTED CONCENTRATIONS AT WESTVACO LIKE
3-HOUR AVERAGES MODEL sCTDMCDEGRU TRACER I SOS
(AVERAGE OB8- • .30 AVERAGE PREo 1.97 RMBE- £ ,•• « Of HOUM»*1SCC}
i» si M as
OBSERVED CONCENTRATIONS (US/MKK3)
Figure F-58
-------
LJ
CO
•-4
II-
7-
OBSERVEO VS. PREDICTED CONCENTRATIONS AT UESTVACO LUKE
3-HOUR AVERAGES MODELtCTDMCDECR2) TRACERS SO3
IAVERAGE OB8- • .88 AVERAGE PRE* I .«9 RHSE- 1 .•£ • OF HOURS-11*7)
II II
OBSERVED CONCENTRATIONS (US/MKK3)
Figure F-59'
-------
M
4S-
OBSERVEO VS. PREDICTED CONCENTRATIONS AT WESTVACO LUKE
3-HOUR AVERAGES MODELi COMPLEX I TRACER I SQ2
(AVERAGE DBS- 9.80 AVERAGE PRE- 9.99 RHSE» 7.99 • OF HOURS-
LO
M
00
18-
* *
18 IS 28 X m 8B 48
OBSERVED CONCENTRATIONS CUS/MJUK3)
Figure F-60
4C
-------
It
CO
II
18-
OBSERVED VS. PREDICTED CONCENTRATIONS AT HESTVACO LUKE
1-HOUR AVERAGES MODELtRTOMfDEFAULT) TRACERt SQ2
tAVERAGE 088* 8.88 AVERAGE P8E- 8.85 RM8E- C .88 • OF HOUR8-1818)
4C8 788 18 II
OBSERVED CONCENTRATIONS (US/MKK3)
Figure F-61
12 II
-------
OBSERVED VS . PREDICTED CONCENTRATIONS AT WESTVACO LUKE
3-HOUR AVERAGES MODELtRTDMCONSITE 1 TRACERS SO2
t AVERAGE OB8- • .•• AVERAGE PME-
RMtS* 0
• OF HOURS- 1*181
7'
A.
I I ! I I I
B 4 S • 7 •
OBSERVED CONCENTRATIONS CUS/MH13)
Figure F-62
it
-------
5.M
2.M-
I.M-
.68-
.20-
PRE/OBS RATIO VS. DISTANCE AT WESTVACO LUKE
AVC OF TOP 10 1-HOUR VALUES MODELtCTDM TRACERiSOS
i ma
IBM
DISTANCE CM)
Figure F-63
-------
2«.M
S.M
2.M-
I '
.18-
PRE/OBS RATIO VS. DISTANCE AT WESTVACO LUKE
AVG OF TOP 10 1-HOUR VALUES MODELsCTOMtDEGRl) TRACERtSO2
SM
(KM
2SM
DISTANCE (M)
Figure F-64
-------
20.00
10.00-
6.00-
2.00
u>
w
CO
.50-
.20-
.18
.•5
.01
PRE/OBS RATIO VS. DISTANCE AT UESTVACO LUKE
AVC OF TOP 10 1-HOUR VALUES MODEL |CTDM(DEGR2) TRACERiS02
* + *
DISTANCE (M)
Figure F-65
-------
M.
Zt.
II.
.2M-
.0S0-
.8i§-
PRE/OBS RATIO VS . DISTAh4CE AT WESTVACO LUKE
AVG OF TCDP 10 I -HOUR VALUES MODEL sCQMPLEXI
OF
+* +
TRACERiS02
ISM
26M
DISTANCE CM)
Figure F-66
-------
U)
w
Ul
6.8M
2.1
I.I
.5M
.2M
.120
.885
PRE/06S RATIO VS . DISTANCE AT UESTVACO LUKE
AVC OF TOP 10 1-HOUR VALUES MODELjRTDM(DEFAULT) TRACERtSO2
ISM 28M
DISTANCE (M)
Figure F-67
-------
10.0
S.0-
2.0-
1.0-
.s-
.2-
PRE/CBS RATIO VS. DISTANCE AT NESTVACO LUKE
AVG OF TOP 10 1-HOUR VALUES MODEL8RTDMCONSITE) TRACER 8SQ3
500
9600 2«M 2SM
DISTANCE CM
Figure F-68
8600
-------
MMT
I2MT
OBSERVED VS . PREDICTED CONCENTRATIONS AT WIDOWS CREEK
1-HOUR AVERAGES MODEL. I CTDM TRACER I SOS
I AVERAGE OB8- 71 .• AVERAGE PRE»14I .0 RM«E*4»»* • OF
OBSERVED CONCENTRATIONS (UC/MMK3)
Figure F-69
-------
OBSERVED VS. PREDICTED CONCENTRATIONS AT WIDOWS CREEK
1-HOUR AVERAGES MODEL I CTOMCDEGRU TRACER t SOS
f AVERAGE OSS- 7« .• AVERAGE l*ftE«19t .» ftMM-979 .• • Of HOUM««4CM)
I2M«
OJ
CO
00
OBSERVED CONCENTRATIONS (UG/MKK3)
Figure F-70
-------
LO
7«M
OBSERVED VS . PREDICTED CONCENTRATIONS AT WIDOWS CREEK
1-HOUR AVERAGES MODELt CTDM(DEGRS) TRACERi SO2
(AVERAGE OB8- 79 .9 AVERAGE PNE-8C0 .4 RMCE-7M .• II OF HOURt»
+ +
OBSERVED CONCENTRATIONS (UC/MJKX3)
Figure F-71
-------
UJ
«•
o
OBSERVED VS . PREDICTED CONCENTRATIONS AT WIDOWS CREEK
1-HOUR AVERAGES MODEL« COMPLEX I TRACERS SOS
(AVERAGE OB8« 71.1 AVERAGE PftE»41S .• MM8E»te8ft
• OF HOUR«-B(Ma)
2flM
I
OBSERVED CONCENTRATIONS CUG/MXK3)
Figure F-72
-------
OJ
fe-
46M
OBSERVED VS. PREDICTED CONCENTRATIONS AT WIDOWS CREEK
1-HOUR AVERAGES MODELiRTDMCDEFAULT) TRACERt SOS
(AVERAGE Oqp- 71.1 AVERAGE FRE-9CC .» RHBE-7B4 .8 • OF HOUH8-3««a)
!>•*.** * *
-------
111
OBSERVED VS. PREDICTED CONCENTRATIONS AT WIDOWS CREEK
1-HOUR AVERAGES MODEL »RTDMC ONSITE ) TRACERS SO2
t< AVERAGE 088- 71 .1 AVERAGE PRE- <47
RM»E»CM
• OF HOURS-MO9)
7SM
OBSERVED CONCENTRATIONS (UG/MKK3)
Figure F-74
IM
-------
3SM
3M0-
2SM
2M0
ISM
OBSERVED VS . PREDICTED CONCENTRATIONS AT WIDOWS CREEK
3-HOUR AVERAGES MODEL I CTDM TRACER I SOS
(AVERAGE DBS* 88 .* AVERAGE PME-IIC.0 RH8E-Caa .8 • OF HOURS-18701
ISM
OBSERVED CONCENTRATIONS (UG/MHX3)
Figure F-75
-------
S000
4S00
4000
3500-
O8SERVED VS. PREDICTED CONCENTRATIONS AT WIDOWS CREEK
3-HOUR AVERAGES MODELt CTDM(DEGR1I TRACERS SO3
I AVERAGE DBS- 69.4 AVERAGE PRE-127 .» RM8E-3e3 .1
OF HOURS-iaeit
-
2580
zaea
* * * *
1600 2000 2600 S000 3600 4000
OBSERVED CONCENTRATIONS (UG/MMIM3J
Figure F-76
4600
-------
CJ
b-
Ln
46M
3SM
2SM-
OBSERVED VS . PREDICTED CONCENTRATIONS AT WIDOWS CREEK
3-HOUR AVERAGES MODEL I CTDM(DEGRS) TRACERS SO2
CAVERAGE DBS- 04.0 AVERAGE PRE-84C .4 RM8E-978 .£ • OF HOURS*a80S I
I EM 2M0 2KM MM KM 4MI 4SM
OBSERVED CONCENTRATIONS (UG/M1M3)
Figure F-77
MM KKM
-------
7M0
4006
3800
OBSERVED VS. PREDICTED CONCENTRATIONS AT WIDOWS CREEK
3-HOUR AVERAGES HODELs COMPLEX I TRACER 8 SO2
I AVERAGE OB8- 69.4 AVERAGE PRE-818.* RH8E-8e7 .7 » OF HOUR«-I*4«3>
OBSERVED CONCENTRATIONS (UG/M*I3)
Figure F-78
-------
2SM
OBSERVED VS. PREDICTED CONCENTRATIONS AT WIDOWS CREEK
3-HOUR AVERAGES MODEL|RTDM(DEFAULT 1 TRACERt SO2
IAVERAGE OB8- 89.4 AVERAGE PHE-278 .• MM8E-981 .9 * OF HOURS-
LO
fc-
ISM
OBSERVED CONCENTRATIONS (UG/MMK3)
Figure F-79
-------
fe-
00
OBSERVED VS . PREDICTED CONCENTRATIONS AT WIDOWS CREEK
3-HOUR AVERAGES MODEL IRTOM(ONSITE1 TRACER 8 SO2
(AVERAGE DBS- 89.4 AVERAGE PHE- 41.8 RH8E«iB8 .£ » OF
2SM
ISM
IMS-
ItM 28M 26M MM
OBSERVED CONCENTRATIONS (UG/MK13)
Figure F-80
-------
M.t
.5-
.2-
PRE/O8S RATIO VS. DISTANCE AT WIDOWS CREEK
AVG OF TOP 10 1-HOUR VALUES MOOELtCTOM
TRACER|S02
!• 12 14 It !• 2*
DISTANCE (KM)
Figure F-81
22 24
12
-------
Ln
O
M.t
28.6-
6.8
CD
£ 2.8-
.8-
.2-
PRE/OBS RATIO VS. DISTANCE AT WIDOWS CREEK
AVG OF TOP 10 1-HR VALUES MODELiCTDMCDEGRl
TRACER IS03
2 4 8 8 It
l
92
I I 9 1
14 98 18 28
DISTANCE CKM5
Figure F-82
22 24 28 28 S8 12
-------
28.8
11.8
Ul
2.8H
i.a
.6-
PRE/OBS RATIO VS. DISTANCE AT WIDOWS CREEK
AVG OF TOP 10 1-HR VALUES MODELiCTDMtDEGR2) TRACERSSOS
!• 12 14 18 II M 22 24 28 2« M 92
DISTANCE (KM)
Figure F-83
-------
LO
Ln
laa.eea
sa.aaa
2a.aee-
s.eae-
2.eae-
i.eae-
.saa-
-------
U)
50.00
20.00-
10.00-
s.00-
2.00-
.00
.68-
.20
.18
.05-
.02-
.01
PRE/OBS RATIO VS. DISTANCE AT WIDOWS CREEK
AVG OF TOP 10 1-HR VALUES MODELtRTDMCDEFAULT) TRACERISO2
* *
10 12 14 10 10 20 22 24 20 20 30 12
DISTANCE (KM)
Figure F-85
-------
CO
u\
tt-
IS.t
2.8
1 '-
.6"
.2"
PRE/08S RATIO VS. DISTANCE AT WIDOHS CREEK
AVG OF TOP 10 1-HR VALUES MODELsRTOMCONSITEI TRACERiSO2
* *•
I I
9 8
• 2 4 8 • !•
12 14 tf It 2t
DISTANCE (KM)
Figure F-86
24 28
M i2
-------
APPENDIX G
SUMMARY OF CASE-STUDY ANALYSES OF
CTDM PREDICTIONS AT THE TRACER SITES
355
-------
APPENDIX G
SUMMARY OF CASE-STUDY ANALYSES OF
CTDM PREDICTIONS AT THE TRACER SITES
Patterns of CTDM predictions and observed concentrations at the
SFg and CF3Br sampler sites have been analyzed for the CCB, HER,
and FSPS experiments. The examination of results for each hour
involved a comparison of the average of the top 5 predicted and
observed concentrations. The height of the plume as well as peak
predicted and observed concentrations relative to Hc were noted, as
well as a plume-height wind speed category. Comments about the
locations of the peak predicted and observed concentrations were also
logged. Case-by-case results are given in Tables G-l through G-6.
Summary statistics by wind speed category and by the plume height
relative to Hc are listed in Tables G-7 through G-12. An index to
these tables is given below.
Table Description
G-l Individual case results for tracer SFg at CCB
G-2 Individual case results for tracer CFjBr at CCB
G-3 Individual case results for tracer SF6 at HBK
G-4 Individual case results for tracer CF£Br at HBR
G-5 Individual case results for tracer SF6 at FSPS
G-6 Individual case results for tracer CF3Br at FSPS
G-7 Summary statistics for case-hour categories for
SF at CCB 6
G-8 Summary statistics for case-hour categories for
CF Br at CCB 3
G-9 Summary statistics for case-hour categories for
SF at HBR 6
C-10 Summary statistics for ease-hour categories for
CF Br at HBR 3
G-ll Summary statistics for case-hour categories for
SF at FSPS 6
Summary statistics for case-hour categories for
CF Br at FSPS •»
356
-------
TABLE G-l
SUMMARY OF PREDICTED AND OBSERVED DATA FOR SF6
TRACER AT CINDER CONE BUTTE SITE*
PLUME WIND
JUL. HEIGHT SPD
DAY HR VS He CAT
.TOP 5 AVE CONG DATA ,(uS/M**3)
TOP 5 TOP 5 PRE/OBS RATIO
AVE PRE AVE OBS RATIO CAT
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He COMMENTS
290
290
290
290
290
291
291
291
291
291
291
294
294
294
294
294
294
295
295
295
295
297
297
297
297
297
297
297
298
298
298
298
298
298
298
299
299
299
301
301
301
302 .
302
302
18
19
21
22
23
18
19
20
21
22
23
2
3 .
5 .
6
7
8
1
2
6
7
2
3
4
5
6
7
8
2
3
4
5
6
7
8
2
3
4
20
21
23
1
18
19
A
A
A
N
N
A
A
A
A
A
A
B
B
A
B
N
A
B
B
B
B
A
A
A
A
A
B
N
A
A
A
A
N
N
N
A
A
A
A
A
A
N
A
A
4
3
3
3
3
4
4
4
4
4
4
3
2
3
2
2
3
2
2
2
2
4
4
4
4
4
2
2
4
4
3
3
2
2
2
3
3
3
4
4
3
3
3
3
10.
6.6
90.
63.
51.
4.2
1..8
21.
30..
9.9
.98E-07
.46E-01
23.
63.
25.
82.
87.
33.
5.4
19.
32.
.35E-01
1.8
16.
17.
14.
42.
. 10E+03
1.3
.97
62.
85.
. 14E+03
.16E+03
. 11E+03
13.
4.7
16.
13.
5.8
11.
34.
13.
26.
24.
21.
17.
19.
22.
18.
20.
42.
27.
23.
11.
4.8
2.1
2.7
19.
16.
30.
8.2
9.9
9.5
7.0
1.9
5.7
19.
18.
17.
9.2
12.
21.
14.
46.
33.
86.
67.
. 11E+03
3.2
2.9
10.
7.9
4.7
.26
4.3
1.6
4.7
.431 <
.311 <
5.167 »
3.311 >
2.321 >
.229 <
.089 «
.503 <=
1.134 >»
.428 <
.000 «
.010 «
10.914 »
23.231 »
1.262 >=»
5.154 »
2.873 >
4.061 >
.546 <-
1.965 >-
4.553 >
.019 «
.309 <
.847 <-
.911 <-
.804 «-
4.563 >
8.532 »
.065 «
.072 «
1.347 >-
2.586 >
1.673 >-
2.373 >
.972 <-
1.623 >-
1.615 >-
1.519 >-
1.636 >-
1.236 >-
43.446 »
7.824 »
8.006 »
5.609 »
A/A
A/A
A/A
N/A
N/B
A/A
A/A
A/A
A/A
A/A
?/A
N/B
N/A
A/A
B/B
N/B
A/B
B/B
B/B
B/N
B/N
N/A
A/A
A/A
A/B
A/N
B/B
A/N
A/A
B/B
A/A
A/A
N/B
N/N
N/B
A/N
N/A
A/A
A/A
A/A
N/N
N/B
A/A
A/A
C,F
C,F
C,F
C,F
C,F
C,F
C,F
J
D,F
D,F
D,F
D,F
D,E
C,F
C,F
D,F
C,E
C,E
C,F
D,F
D,E
D,F
C,F
A,C,E
B,C,E
D,E
D,F
A,D,F
D,F
D,F
D,F
D,F
C,F
B,C,E
D,E
D,F
B,C,E
C,E
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
357
-------
TABLE G-l (Page 2 of 4)
SUMMARY OF PREDICTED AND OBSERVED DATA FOR SF6
TRACER AT CINDER CONE BUTTE SITE*
PLUME WIND
JUL. HEIGHT SPD
DAY HR VS He CAT
TOP 5 AVE CONG DATA (uS/M**3)
*********************A********
TOP 5 TOP 5 PRE/OBS RATIO
AVE PRE AVE OBS RATIO CAT
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He COMMENTS
302
302
303
304
304
304
304
304
304
305
305
305
305
305
305
305
309
309
309
309
309
310
310
310
310
311
311
311
311.
311
311
314
314
314
314
314
314
314
314
315
315
315
315
.315
_U -J_ _i_ _i-
20
24
1
1
2
3
4
6
7
1
2
3
4
5
6
7
4
5
6
7
8
4
5
6
7
1
2
3
4
5
6
1
2
3
4
5
6
7
3
3
4
5
7
8
N
B
B
B
A
A
A
A
A
B
N
N
B
B
N
N
B
B
B
B
B
B
B
N
A
B
N
B
B
N
B
B
B
B
B
B
3
B
B
N
N
N
A
A
2
2
2
2
3
4
4
4
4
2
3
2
2
2
3
2
2
2
2
2
2
2
'2
3
3
2
3
2
2
3
2
3
2
3
3
3
2
2
2
3
3
3
4
4
44.
68.
73.
4.3
5.4
.80E-02
.98E-03
.68E-03
2.5
24.
19.
19.
8.2
52.
16.
3.9
29.
41.
17.
14.
40.
36.
19 *
41.
31.
25,
1.1
5.7
21.
22.
14.
.97
9.0
3.6
4.1
3.1
.20
4.2
1.2
25«
43.
50.
21.
12.
4.7
6.1
15.
.91
14.
8.1
5.4
3.2
3.5
10.
5.0
2.9
37.
74.
35.
6.6
.83
29.
14.
11.
4.2
22.
52.
63.
9.7
26.
3.5
4.5
18.
35.
7.6
20.
28.
57.
18.
8.6
2.3
5.4
.66
25.
69.
59.
16.
29.
9.
11.
4.
4.
.
.
,
.
.
2.
3.
6.
Q
o
.
*
35.
1.
1.
1.
9.
1.
.
.
3.
.
0
lo
1.
c
1.
0
*
.
9
9
^
.
1.
o
•
v
1.
.
422 »
184 »
802 >
700 >
399 <
001 «
000 «
000 «
713 <=
338 >
886 >
571 »
218 <
698 <=-
456 <
596 <=
328 »
413 >-
262 >-
271 >-
474 »
626 >-
359 <
650 <»
170 >
965 <=-
324 <
263 >*•
185 >=
642 <-
887 >-
047 «
319 <
150 «
234 <
940 <»
085 «
772 <-
876 >-
992 <=»
624 <-
845 <=•
270 >-
404 <
N/B
B/B
N/B
N/B
N/B
?/B
?/B
?/A
A/A
B/B
N/A
N/N
B/B
B/A
N/A
A/N
N/B
B/B
N/B
B/B
B/B
B/B
B/B
A/N
A/N
N/B
A/A
B/N
N/B
N/N
B/B
N/B
B/N
B/N
A/A
A/B
B/B
B/B
B/B
A/N
A/A
A/A
A/N
A/A
D,
c,
c,
c,
J
J
J
c,
D,
c,
D,
B,
C,
c,
D,
D,
D,
A,
B,
D,
D,
A,
D,
c,
D,
D,
D!
c,
c,
D,
B
c,
C,
D,
D,
D,
c,
A,
E
F
F
F
E
F
F
F
c,
E
F
F
F
F
D,
D,
F
F
D,
F
F
F
F
F
F
F
F
F
F
F
F
F
F
E
c,
E
F
F
F
E
*SEE INTERPRETATION OF CODES AT END OF TABLE
3.58
-------
TABLE G-l (Page 3 of 4)
SUMMARY OF PREDICTED AND OBSERVED DATA FOR SF6
TRACER AT CINDER CONE BUTTE SITE*
JUL.
DAY
317
317
317
317
317
317
1
I
HR ^
3
4
5
6
7
8
?LUME
IEIGH1
TS HC
A
A
A
A
A
A
WIND
: SPD
CAT
3
4
4
4
4
4
*******
TOP 5 AVE CONG DATA (uS/M**3)
******************************
TOP 5 TOP 5 PRE/OBS RATIO
AVE PRE AVE OBS RATIO CAT
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
COMMENTS
12.
7.2
7.9
4.2
6.3
8.1
7.1
21.
16.
7.5
13.
15.
1.662
.343
.505
.560
.505
.547
>•
<
<»
<=
<=
<=»
A/A
A/A
A/A
A/A
A/A
A/A
C,F
C,F
C,F
C,F
*SEE INTERPRETATION OF CODES AT END OF TABLE
359
-------
TABLE G-l (Page 4 of 4)
INTERPRETATION OF CODES:
TABLE ITEM
PLUME HEIGHT VS He
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
COMMENTS
CODE
N
A
B
1
2
3
4
«
<
>
»
N
A
A
B
C
D
E
F
G
H
I
J
K
MEANING
Plume height within
Plume height > He +
Plume height < He -
5 m of He
5 m
5 m
Wind speed less* than 1 m/sac
Wind speed between 1 and 3 m/sec
Wind gpeed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between -S and 1.0
Ratio is between 1=0 and 2.0
Ratio is between 2 . 0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m of He
Elevation ©f maximum > He + 5m
Elevation of maximum < He - 5m
The location of the predicted maximum coincides
with or is at the closest adjacent receptor
to the location of the observed maximum
The location of the predicted maximum is at a
receptor close to the location of the observe
maximum (with no more than 1 or -2 receptors
closer to the location of the predicted
. maximum)
The predicted maximum concentration is on
the far side of the hill/ridge
The predicted maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration is on
the far side of the hill/ridge
The observed maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration occurs
north of the predicted maximum concentration
The predicted maximum concentration occurs
south of the predicted maximum concentration
The angle formed by the intersection of the
stack-predicted maximum concentration recepto
and the stack-observed maximum concentration
receptor lines is more than 90 degrees
Predicted map is all zeroes
Observed map is all zeroes
360
-------
TABLE G-2
SUMMARY OF PREDICTED AND OBSERVED DATA FOR CF3BR
TRACER AT CINDER CONE BUTTE SITE*
JUL.
DAY
301
301
301
302
304
304
304
304
304
305
305
305
305
305
305
309
309
309
309
310
310
310
311
311
311
311
314
314
314
314
314
314
315
315
317
317
317
317
317
PLUME WIND
HEIGHT SPD
HR VS He CAT
20
21
23
1
1
3
4
6
7
3
4
5
6
7
8
5
6
7
8
4
5
6
1
2
5
6
1
2
3
4
5
6
7
3
4
5
6
7
8
A
A
N
N
N
N
A
A
A
B
A
A
A
A
A
A
A
A
A
A
N
A
A
A
A
A
N
N
N
A
A
B
N
N
A
A
A
A
A
4
4
3
3
3
3
3
3
3
2
3
3
4
4
3
3
3
2
3
2
2
3
3
3
4
3
3
3
3
3
3
3
3
3
4
4
4
4
4
TOP 5 AVE CONG DATA (US/M**3) ELEV OF PRE
****************************** MAX VS He/
TOP 5 TOP 5 PRE/OBS RATIO ELEV OF OBS
AVE PRE AVE OBS RATIO CAT MAX VS He COMMENTS
26.
14.
17.
37.
.27E-01
63.
53.
8.9
14.
30.
7.2
3.0
.22
.33
6.7
.64
12.
21.
8.6
89.
40.
11.
6.6
1.5
7.0
21.
45.
44.
52.
61.
. 11E+03
4.3
22.
25.
2.3
2.2
.79
1.8
2.1
2.6
9.0
2.3
2.5
1.6
20.
6.7
35.
47.
2.5
1.2
.95
1.0
7.0
4.8
7.7
24.
8.6
.27
11.
.79
6.7
1.5
.26
11.
12.
5.9
11.
15.
1.5
6.9
8.0
35.
29.
6.0
2.3
.63E-01
2.3
2.5
9.924 »
1.571 >=
7.269 »
15.059 »
.017 «
3.178 >
7.845 »
.256 <
.287 <
11.880 »
6.081 »
3.192 >
.211 <
.047 «
1.384 >=
.083 «
.497 <
2.428 >
32.416 - »
8.208 »
51.224 »
1.660 >-
4.519 >
5.657 »
.628 <-
1.790 >-
7.652 »
3.918 >
3.557 >
40.851 »
15.455 »
.529 <»
.641 <-
.364 <=
.384 <
.935 <=
12.615 »
.643 <=
.850 <-
A/A
A/A
N/N
N/B
B/B
N/A
N/A
A/A
N/A
B/B
N/A
A/N
A/A
A/A
N/A
N/B
N/A
N/A
N/?
A/A
A/A
A/A
N/A
A/A
A/A
N/N
N/A
N/B
N/N
N/A
N/A
B/A
A/N
A/N
A/A
A/A
A/A
A/A
A/A
B
B
D,E
D,F
B,D,F
B,C,E
B,C,E
D,E
C,E
B,C,E
C,F
D,F
B,D,F
D,F
B,C,E
C,E
C,F
K
D,F
D,F
D,F
C,E
A,D,F
A,C,E
D,E
D,E
D,E
D,F
D,F
C,F
B,C,E
C,E '
D,E
D,E
D,E
D,E
D,E
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
361
-------
TABLE G-2 (Continued)
INTERPRETATION OF CODES:
TABLE ITEM CODE
PLUME HEIGHT VS He
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
N
A
3
4
ELEVATION 'OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
COMMENTS
N
A
E
F
G
H
I
J
K
MEANING
Plume height within 5 m
Plume height > He H- 5m
Plume height < He - 5m
of He
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind.speed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between .5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m of He
Elevation of maximum > He + 5m
Elevation of maximum < He - 5m
The location of the predicted maximum coincides
with or is at the closest adjacent receptor
to the location of the observed maximum
The location of the predicted maximum is at a
receptor close to the location of the observ*
maximum (with no more than 1 or 2 receptors
closer to the location of the predicted
maximum)
The predicted maximum concentration is on
the far side of the hill/ridge
The predicted maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration is on
the far side of the hill/ridge
The observed maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration occurs
north of the predicted maximum concentration
The predicted maximum concentration occurs
south of the predicted maximum concentration
The angle formed by the intersection of the
stack-predicted maximum concentration receptu
and the stack-observed maximum concentration
receptor lines is more than 90 degrees
Predicted map is all zeroes
Observed map is all zeroes
362
-------
TABLE G-3
SUMMARY OF PREDICTED AND OBSERVED DATA FOR SF6
TRACER AT HOG BACK RIDGE SITE*
TOP 5 AVE CONG DATA (uS/M**3) ELEV OF PRE
PLUME WIND ****************************** MAX VS He/
JUL. HEIGHT SPD TOP 5 TOP 5 PRE/OBS RATIO ELEV OF OBS
DAY HR VS He CAT AVE PRE AVE OBS RATIO CAT MAX VS He COMMENTS
284
284
285
285
286
286
286
236
287
287
287
287
287
287
238
288
288
288
288
295
295
295
295
295
295
295
296
296
296
296
296
296
297
297
297
297
297
297
299
299
299
299
299
299
3
5
2
3
2
5
7
8
2
3
4
5
6
7
4
5
6
7
8
2
3
4
5
6
7
8
1
2
3
5
6
7
2
3
4
5
7
9
3
4
6
7
8
9
B
A
B
B
A
N
N
B
B
B
B
B
B
B
B
B
B
B
B
B
N
B
N
A
A
A
N
B
B
A
A
A
A
A
A
A
A
A
N
A
A
A
N
A
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
. 2
2
2
2
3
2
2
2
2
2
2
3
2
2
2
2
3
3
2
2
2
2
2
2
22.
14.
34.
27.
17.
21.
38.
32.
43.
39. .
46.
.26E+03
89.
. 13E+03
31.
47.
33.
24.
36.
76.
36.
16.
21.
16.
21.
28.
. 12E+03
19.
37.
17.
21.
35.
8.8
8.9
23.
7.6
1.3
13.
76.
32.
40.
31.
57.
39.
44.
33.
16.
29.
3.4
9.0
11.
27.
7.5
18.
33.
62.
29.
63.
13.
17.
19.
28.
15.
35.
9.4 .
13.
7.4
2.4
3.9
12.
12.
14.
9.4
7.9
11.
18.
.12
1.5
8.1
3.5
.41
16.
17.
19.
21.
13.
17-
35.
.5dt> <
.416 <
2.157 >
.932 <=
5.010 »
2.378 >
3.316 >
1.163 >=
5.776 »
2.211 >
1.395 >=
4.156 >
3.104 >
2.115 >
2.490 >
2.694 >
1.748 >-
.843 <=
2.413 >
2.185 >
3.815 >
1.238 >-
2.873 >
6.465 »
&. 471 »
2.236 >
9.681 »
1.381 >-
3.899 >
2.141 >
1.938 >-
1.934 >-
73.928 »
5.359 »
2.812 >
2.163 >
3.077 >
1.125 >-
4.594 >
1.746 >-
1.877 >»
2.451 >
3.407 >
1.132 >=
B/A
B/A
B/A
B/B
A/A
N/N
N/A
B/A
B/B
B/A
B/A
B/A
B/A
B/A
B/B
B/N
N/A
N/A
B/A
B/A
N/B
B/A
A/A
A/A
A/A
N/A
B/N
B/A
B/N
A/A
N/A
N/A
A/N
A/N
B/A
A/A
A/A
N/A
B/A
N/A
B/B
N/B
N/B
N/A
D,F
D,F,
D,E
D,F
D,F
D,F
D,E
D,F
D,F
D,E
D,E
D,F
D,E
D,E
D,F
D,E
D,F
D,E
D,F
D,E
B
D,E
D,E
C,F
B,D,
D,F
D,E
D,F
D,F
D,E
D,F
D,E
C,F
D,F
D,F
D,F
D,F
D,F,
D,F
D,F
H
F
H
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
363
-------
TABLE G-3 (Continued)
SUMMARY OF PREDICTED AND OBSERVED DATA FOR SF6
TRACER AT HOG BACK RIDGE SITE*
PLUME WIND
JUL. HEIGHT SPD
DAY HR VS He CAT
TOP 5 AVE CONC DATA (uS/M**3)
******************************
TOP 5 TOP 5 PRE/OBS RATIO
AVE PRE AVE OBS RATIO CAT
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
COMMENTS
302
302
302
302
302
302
302
1
2
3
4
5
7
8
B
B
B
N
B
N
N
2
2
2
2
2
2
2
50.
99.
.13E+03
76.
.13E+03
29.
39.
19
22
61
7.
5<,
7.
7.
"«
•
0
2
0
8
6
2
4
2
10
25
3
5
.655 >
.584 >
.216 >
.579 »
.754 »
.714 >
.136 »
B/A
B/N
B/B
N/N
N/?
N/B
B/A
D,
D,
D,
D,
K
H
F
F
F
E
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
364
-------
TABLE G-3 (Continued)
INTERPRETATION OF CODES:
TABLE ITEM
PLUME HEIGHT VS He
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
COMMENTS
CODE
N
A
B
1
2
3
4
«
<
>
»
N
A
B
B
C
D
E
F
G
H
I
J
K
MEANING
Plume height within
Plume height > He +
Plume height < He -
5 m of He
5 m
5 m
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind -speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between -5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m of He
Elevation of maximum > He + 5m
Elevation of maximum < He - 5m
The location of the predicted maximum coincides
with or is at the closest adjacent receptor
to the location of the observed maximum
The location of the predicted maximum is at a.
receptor close to the location of the observed
maximum (with no more than 1 or 2 receptors
closer to the location of the predicted
•maviinimi)
The predicted maximum concentration is on
the far side of the hill/ridge
The predicted maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration is on
the far side of the hill/ridge
The observed maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration occurs
north of the predicted maximum concentration
The predicted maximum concentration occurs
south of the predicted maximum concentration
The angle formed by the intersection of the
stack-predicted maximum concentration receptor
and the stack-observed maximum concentration
receptor lines is more than 90 degrees
Predicted map is all zeroes
Observed map is all zeroes
365
-------
TABLE G-4
SUMMARY OF PREDICTED AND OBSERVED DATA FOR CF3BR
TRACER AT HOG BACK RIDGE SITE*
PLUME WIND
JUL. HEIGHT SPD
DAY HR VS He CAT
TOP 5 AVE CONG DATA (US/M**3)
******************************
TOP 5 TOP 5 PRE/OBS RATIO
AVE PRE AVE OBS RATIO CAT
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He COMMENTS
285
285
285
285
286
286
286
287
287
287
287
287
287
287
288
238
283
288
295
295
295
295
296
296
297
297
297
297
297
299
299
299
299
299
299
302
302
302
302
302
302
302
1
2
3
24
5
7
8
2
3
4
5
7
8
24
5
6
7
8
3
6
7
24
6
7
2
3
4
. 6
7
3
4
5
6
8
9
2
3
4
5
7
8
9
N
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
A
B
N
B
B
B
B
B
N
B
N
B
B
B
N
B
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
47.
38.
33.
19.
20.
57.
34.
59.
52.
.10E+03
.15E+03
.18E+03
. 17E+03
63.
35.
32.
24.
50.
IS.
30.
18.
.14E+03
16.
36.
6.5
IS.
23.
8.9
3.2
73.
61.
52.
44.
10.
7.3
18.
23.
20.
15.
20.
24.
14.
58.
98.
67.
65.
63.
.
.13E+03
.26E+03
.13E+03
.12E+03
.10E+03
.141+03
.21E+03
.16E+03
18.
49.
66.
40.
50.
13.
33.
38.
78.
€.3
43.
1.2
15.
8.3
1.1
3.4
29.
96.
.11E+03
56.
5.6
7.8
3.2
1.5
15.
43.
19.
25.
30.
.807
.338
.499
.295
.316
.445
.129
.473
.438
.971
1.084
.836
1.044
3.498
.726
.490
.592
1.009
1.115
.909
,469
1.763
2.613
.342
5.549
1.228
2.739
7.968
.927
2.522
.634
.480
.779
1.326
.936
5.503
18.549
1.371
.316
1.090
.943
.464
<=
<
<
<
<
<
«
<
<
<=
>=s
<=
>=*
>
<=
<
<=»
>a
>ai
<=
.<
>=«
>
<=«
»
>«»
>
»
<=»
>
»
a
<
>SB
-------
TABLE G-4 (Continued)
INTERPRETATION OF CODES:
TABLE ITEM
PLDME HEIGHT VS He
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
CODE
N
A
B
1
2
3
4
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
COMMENTS
N
A
B
B
C
D
E
F
G
H
I
J
K
MEANING
Plume height within 5 m of He
Plume height > He + 5m
Plume height < He - 5m
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between -5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m of He
Elevation of maximum > He + 5m
Elevation of maximum < He - 5m
The location of the predicted maximum coincides
with or is at the closest adjacent receptor
to the location of the observed maximum
The location of the predicted maximum is at a
receptor close to the location of the observed
maximum (with no more than 1 or 2 receptors
closer to the location of the predicted
maximum)
The predicted maximum concentration is on
the far side of the hill/ridge
The predicted maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration is on
the far side of the hill/ridge •
The observed maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration occurs
north of the predicted maximum concentration
The predicted maximum concentration occurs
south of the predicted maximum concentration
The angle formed by the intersection of the
stack-predicted maximum concentration receptor
and the stack-observed maximum concentration
receptor lines is more than 90 degrees
Predicted map is all zeroes
Observed map is all zeroes
367
-------
TABLE G-5
SUMMARY OF PREDICTED AND OBSERVED DATA' FOR SF6
TRACER AT TRACY POWER PLANT SITE*
TOP 5 AVE CONG DATA (US/M**3) ELEV OF PRE
PLUME WIND ****************************** MAX VS He/
JUL. HEIGHT SPD TOP 5 TOP 5 PRE/OBS RATIO ELEV OF OBS
DAY HR VS HC CAT AVE PRE AVE OBS RATIO CAT MAX VS He COMMENTS
219
219
219
219
220
220
220
220
222
222
222
222
223
223
223
223
223
223
223
223
223
224
224
224
224
224
224
224
224
224
224
225
225
225
225
225
225
228
228
229
229
229
229
229
.it _k .v _i_ .
4
5
6
7
4
5
6
7
21
22
23
24
1
2
3
4
5
6
21
22
24
1
2
3
4
5
6
21
22
23
24
1
2
3
4
5
6
23
24
1
2
3
4
5
j. -j- _L.
N
N
A
N
N
B
B
B
A
A
A
A
A
A
A
A
B
N
A
A
A
A
B
B
B
B
B
A
A
A
A
A
N
B
B
B
B
A
A
A
A
A
A
B
3
3
2
2
3
2
2
2
4
4
4
3
2
2
3
3
2
3
3
3
3
3
3
2
2
2
2
4
3
3
3
3
2
2
2
2
2
3
4
4
3
4
4
3
.74
.78
2.9
2.6
2.9
1.8
2.7
2.8
.29
.52E-01
.56
.78
.59
1.0
2.3
3.3
1.2
.63
.53
.60
.92
1.3
2.6
.54
2.8
2.5
4.3
.23
.19
.47
.66
1.4
3 = 0
1.9
1.3
1.9
1.4
• .51
.52
.44
.52
.29
.37
1.8
.64
1.3
2.2
1.8
•2.5
5.2
2.8
3.7
.23
.19
.19
.37
.38
.57
.62
.86
.84
1*8
.72
.39
.54
.66
.54
.32
.83
.71
2.6
.25
2.2
.36
.59
.70
.80
1.1
4.3
4.8
3.3
1.6
1.0
1.2
1.2
.35
1.0
2.2
1.170
.602
1.331
1.445
1.175
.348
.962
.746
1.260
.270
2.938
2.111
1.530
1.773
3.678
3.850
1.409
.343
= 739
. 1.508
1.700-
2.014
4.895
.651
3.319
3.553
1.637
.918
.087
1.304
1.122
2.040
3.759
1.754
.416
.403
.428
.324
.513
.359
.437
.337
.352
.806
>«•
<=«
>«
>ss
>«
<
<=
<=
>=
<
>
>
>-
>*
>
>
>™
<
<=•
>=
>a=
>
>
<=»
>
>
>=
<=»
«
>™
>»
>
>
>=
<
<
<
<
<=»
<
<
<
<
<=
A/B
A/A
A/A
A/A
A/A
A/A
B/B
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/A
A/B
A/A
A/A
?/A
A/B
B/A
A/A
A/N
A/A
A/A
A/A
A/A
A/N
A/A
A/A
A/B
A/A
A/B
A/A
B
A
A
I
B
B
I
I
A
I
I
J
I
I
A
I
A
*SEE INTERPRETATION OF CODES AT END OF TABLE
368
-------
TABLE G-5 (Page 2 of 4)
SUMMARY OF PREDICTED AND OBSERVED DATA FOR SF6
TRACER AT TRACY POWER PLANT SITE*
JUL.
DAY
229
229
229
229
230
230
230
230
230
230
230
230
230
231
231
231
231
231
231
231
233
233
234
234
234
234
234
234
234
234
235
235
235
235
235
235
235
235
236
236
236
236
236
236
PLUME WIND
HEIGHT SPD
HR VS He CAT
6
7
23
24
1
2
3
4
5
6
7
23
24
1
2
3
4
5
6
7
23
24
1
2
3
4
5
6
7
24-
1
2
3
4
5
6
7
24
1
2
3
4
5
6
A
A
A
A
A
A
A
A
B
B
A
A
A
A
A
A
A
N
N
N
A
A
A
A
A
A
A
B
N
A
A
A
B
B
B
B
B
A
A
A
B
B
B
B
3
2
3
3
3
3
4
3
3
2
2
4
4
4
4
3
3
3
3
3
3
3
3
4
4
4
4
3
3
3
2
2
2
2
2
2
2
3
3
2
2
2
2
2
TOP 5 AVE CONG DATA (uS/M**3) ELEV OF PRE
****************************** WAX VS He/
TOP 5 TOP 5 PRE/OBS RATIO ELEV OF OBS
AVE PRE AVE OBS RATIO CAT MAX VS He COMMENTS
.34
.53
.46
.44
.63
.75
1.2
1.2
1.4
2.3
2.5
.47
.69
.59
1.1
1.6
2.1
1.2
1.3
2.1
.76
2.5
1.1
.98
1.3
1.9
1.2
.77
2.0
.49
1.7
.51
2.9
4.4
2.1
2.5
4.8
.73
1.1
.52
.67
2.3
3.7
1.'3
1.8
2.3
.24
.48
.37
.36
.30
.94
1.7
2.7
4.5
.35
.38
.57
.73
.74
.42
.72
2.0
1.7
.58
.55
.44
.73
1.3
1.1
.79
1.7
1.1
.43
.49
.46
.73
2.9
4.7
2.6
1.6
.54
.85
.45
.34
.26
2.1
3.7
.186
.232
1.909
.930
1.715
2.057
3.887
1.242
.823
.863
.564
1.346
1.800
1.039
1.464
2.091
4.998
1.723
.628
1.279
1.309
4.481
2.489
1.340
.973
1.703
1.569
.445
1.762
1.140
3.533
1.121
4.049
1.536
.444
.957
2.937
1.340
1.331
1.176
1.962
8.805
1.809
.347
«
<
>*
<=»
>«
>
>
>»
<=•
m
>m
>—
>—
>
>
>«
<-
>»
>•
>
>
>•
<»
>«•
>=»
<
>«
>«
>
>«
>
>=»
<
>-
>»
>-
>»
»
>»
<
A/A
A/A
A/A
A/N
A/B
A/A
A/A
A/B
A/A
B/A
A/A
A/A
A/A
A/N
A/B
A/A
A/A
B/A
A/A
A/A
N/A
B/N
A/A
A/A
A/B
A/A
B/A
B/A
A/A
A/A
A/A
A/A
A/A
A/A
N/A
B/B
A/A
A/B
A/A
A/B
A/B
A/B
A/A
A/A
B
B
B
A
B
A
I
I
I
I
I
I
B
I
I
I
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
369
-------
TABLE G-5 (Page 3 of 4)
SUMMARY OF PREDICTED AND OBSERVED DATA FOR SF6
TRACER AT TRACY POWER PLANT SITE*
TOP 5 AVE CONG DATA (uS/M**3) ELEV OF PRE
PLUME WIND ****************************** MAX VS He/
JUL. HEIGHT SPD TOP 5 TOP 5 PRE/OBS RATIO ELEV OF OBS
DAY HR VS He CAT AVE PRE AVE OBS RATIO CAT MAX VS He COMMENTS
236
238
238
238
238
238
238
238
239
239
239
239
239
239
239
240
240
240
240
240
240
240
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
B
B
B
A
B
N
B
B
B
B
B
3
B
B
B
B
A
A
B
A
B
B
2
3
3
2
2
2
2
2
2
2
2
2
2
2
2
3
3
2
2
2
2
2
.72
.12
.24
.15
4.5
.56
1.8
.91
1.2
2.1
3.3
1.9
3.1
2.1
2.0
.79
1.3
.67
.64
1.9
2.0
3.1
.98
.11
.20
.30
.27
1.4
2.1
1.2
.25
.52
1.8
1.5
.81
1.9
2.8
.19
.31
1.4
1.6
3.2
2.1
1.4
.730
1.118
1.175
.488
16.536
.400
.856
.737
4.753
4.119
1.846
1.294
3.889
1.117
.732
4.201
4.227
.465
.391
.608
.945
2.156
<»
>=
>a*
<
»
<
<=*
<=
>
>
>«
>s
>
>=
>
<
<
<=
<»
>
A/A I
A/A A
A/A
A/A
B/A B
A/A
A/A
A/A
A/A
A/A
A/A
B/B
A/A
A/A
A/A
A/A
A/B
A/A
B/B
A/A
A/A
A/B
*SEE INTERPRETATION OF CODES AT END OF TABLE
370
-------
TABLE G-5 (Page 4 of 4)
INTERPRETATION OF CODES:
TABLE ITEM CODE
PLUME HEIGHT VS He N
A
B
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
COMMENTS
1
2
3
4
«
<
<=
>«
>
»
N
A
B
B
C
D
E
F
G
H
r
MEANING
Plume height within 10 m of He
Plume height > He + 10 m
Plume height < He - 10 m
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between .5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 10 m of He
Elevation of maximum > He + 10.m
Elevation of maximum < He - 10 m
The location of the predicted maximum coincides
with or is at the closest adjacent receptor
to the location of the observed maximum
The location of the predicted maximum is at a
receptor close to the location of the observed
maximum (with no more than 1 or 2 receptors
closer to the location of the predicted
maximum)
The predicted maximum concentration is on
the far side of the hill/ridge
The predicted maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration is on
the far side of the hill/ridge
The observed maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration occurs
north of the predicted maximum concentration
The predicted maximum concentration occurs
south of the predicted maximum concentration
The angle formed by the intersection of the
stack-predicted maximum concentration receptor
and the stack-observed maximum concentration
receptor lines is more than 90 degrees
Predicted map is all zeroes
Observed map is all zeroes
371
-------
TABLE G-6
SUMMARY OF PREDICTED AND OBSERVED DATA FOR CF3BR
TRACER AT TRACY POWER PLANT SITE*
TOP 5 AVE CONG DATA (uS/M**3) ELEV OF PRE
PLUME WIND ****************************** MAX VS He/
JUL. HEIGHT SPD TOP 5 TOP 5 PRE/OBS RATIO ELEV OF OBS
DAY HR VS He CAT AVE PRE AVE OBS RATIO CAT MAX VS He COMMENTS
219
219
219
219
220
220
220
220
222
222
222
222
223
223
223
223
223
223
223
223
223
224
224
224
224
224
224
224
224
224
224
225
225
225
225
225
225
228
228
229
229
229
229
229
4
5
6
7
4
5
6
7
21
22
23
24
1
2
3
4
5
6
21
22
24
1
2
3
4
5
6
21
22
23
24
1
2
3
4
5
6
23
24
1
2
3
4
5
B
B
B
3
B
B
B
B
A
B
A
A
N
B
B
N
B
B
A
A
A
A
A
B
B
B
B
A
A
A
A
A
N
B
B
B
B
A
A
A
A
A
A
A
2
2
2
2
2
2
2
2
4
4
4
3
2
2
3
3
2
2
3
3
3
3
3
2
2
2
2
4
3
3
2
3
2
2
2
2
2
4
4
4
4
4
4
3
1.6
2.1
4.3
2.2
4.6
2.7
2.1
1.9
.17
.61E-01
.79
1.5
2.5
1.4
1.1
3.0
1.7
1.0
.52
.60
.93
1.2
2.1
.99
2.2
.76
4.8
.20
.oil
.45
.78
1.5
2.9
3.6
2.3
1.8
1.8
.94
.77
.59
.93
.56
1.2
2.8
.52
1.6
2.2
1.8
1.5
3.7
2.9
7.8
.31
.15
.55
.49
.60
.66
1.3
1.5
1.5
4.0
.51
1.1
.57
.78
3.9
1.1
1.0
4.0
4.5
.15
.20
.68
.75
.79
1.1
1.3
3-8
2.9
.79
1.5
1.1
1.0
.87
.41
1.2
2.3
3.083
1.351
1.954
1.206
2.948
.722
.735
.248
.550
.393
1.453
2.968
4.208
2.076
.874
2.007
1.125
.247
1.013
.572
1.630
1.594
.537
.903
2.193
.188
1.051
1.340
.556
.659
1.038
1.876
2.614
2.642
.612
.634
2.263
.611
.679
.582
1.068
1..353
.967
1.256
>
>a=
>=s
>=•
>
<=
<=
<
<=
<
>=*
>
>
>
<=•
>
>w
<
>«
<»
>»
>»
<»
<=
>
«
>»
>»
»
>aa
>
>
<=
<=
>
<=
<»
<3S
><=
>a
<=a
>=
B/B
B/A
B/B
B/B
B/A
B/B
A/B
B/B
A/A
A/A
A/N
A/B
B/A
A/B
A/A
A/A
B/B
A/B
A/B
A/B
A/B
A/B
A/A
B/B
B/B
B/A
B/A
A/N
A/A
A/A
B/A
B/A
A/A
A/A
A/A
B/B
A/A
A/B
A/A
A/A
A/A
A/B
A/B
B/A
A
A
B
A
B
A
B
A
I
I
A
A
A
I
B
*SEE INTERPRETATION OF CODES AT END OF TABLE
372
-------
TABLE G-6 (Page 2 of 4)
SUMMARY OF PREDICTED AND OBSERVED DATA FOR CF3BR
TRACER AT TRACY POWER PLANT SITE*
JUL.
DAY
229
229
229
229
230
230
230
230
230
230
230
230
230
231
231
231
-231-
231
231
231
233
233
234
234
234
234
234
234
234
234
235
235
235
235
235
235
235
235
236
236
236
236
236
236
PLUME WIND
HEIGHT SPD
HR VS He CAT
6
7
23
24
1
2
3
4
5
6
7
23
24
1
2
3
4
5
6
7
23
24
1
. 2
3
4
5
6
7
24
1
2
3
4
5
6
7
24
1
2
3
4
5
6
A
B
A
A
A
A
A
B
B
B
B
A
A
A
A
A
A
N
N
N
A
A
A
A
A
A
A
B
B
A
B
B
B
B
B
B
B
A
B
N
B
B
B
B
3
2
3
3
4
4
4
3
2
3
2
4
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
3
3
2
2
2
2
2
2
2.
2
3
2
2
2
2
2
2
TOP 5 AVE CONG DATA (uS/M**3) ELEV OF PRE '
****************************** MAX VS He/
TOP 5 TOP 5 PRE/OBS RATIO ELEV OF OBS
AVE PRE AVE OBS RATIO CAT MAX VS He COMMENTS
1.5
1.1
1.2
.78
.93
1.3
2.0
1.1
1.7
.96
2.7
.45
.58
.54
.65
.94
1.2
1.3
1.3
1.5
1.2
3.7
1.9
1.5
1.3
2.0
1.8
.76
3.0
1.1
1.1
1.1
.24
1.2
1.4
1.0
2.0
1.1
1.2
1.2
.69
2.9
1.7
.68
2.1
1.7
.38
.46
.36
.87
.83
1.4
3.1
6.1
3.1
.31
.39
.32
.43
.60
.40
1.0
3.2
3.0
.68
1.3
.84
2.4
1.4
1.8
2.6
2.3
2.4
.81
.95
.85
1.3
1.9
2.3
4.9
3.0
1.1
1.3
1.2
2.6
2.0
2.7
2.8
.732
.640
3.062
1.693
2.607
1.517
2.362
.783
.553
.157
.869
1.453
1.482
1.679
1.506
1.565
2.941
1.282
.397
.507
1.822
2.948
2.309
.617
.956
1.141
.685
.327
1.246
1.391
1.154
• 1.347
.177
.610
.630
.210
.665
1.011
.912
1.004
.263
1.454
.649
.246
<=
<=»
>
>»
>
>=
>
<=
<-
«
<=
>«•
>-
>*
>»
>«
>
>—
<
<-
>-
>
>
<»
<»
>«
<»
<
>«
>-
>-
>-
«
-
<«
>-
<
>-
<=
<
B/A
B/A
B/B
A/A
N/A
A/B
A/A
A/B
A/A
B/A
B/B
A/A
A/A
A/N
A/A
A/N
A/B
A/B
A/B
A/B
B/B
B/A
A/A
A/A
A/A
A/A
B/N
B/A
B/A
A/A
B/B
A/A
A/B
B/B
A/B
A/A
B/B
A/B
B/A
A/A
A/A
B/A
B/B
A/A
A
A
B
B
A
B
B
B
A
A
A
I
I
I
B
I
A
I
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
373
-------
TABLE G-6 (Page 3 of 4)
SUMMARY OF PREDICTED AND OBSERVED DATA FOR CF3BR
TRACER AT TRACY POWER PLANT SITE*
JUL.
DAY
236
238
238
238
238
238
238
238
239
239
239
239
239
239
239
240
240
240
240
240
240
240
TOP 5 AVE CONC DATA (US/M**3)
PLUME WIND ******************************
HEIGHT SPD TOP 5 TOP 5 PRE/OBS RATIO
HR VS He CAT AVE PRE AVE OBS RATIO CAT
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
B
A
B
B
B
B
B
B
B
B
B
B
B
B
B
A
B
B
B
B
B
B
2
3
2
2
2
2 .
2
2
2
2
2
2
2
2
2
3
3
2
2
2
2
2
1.1
.67
.66
1.3
.78
.41
1.3
.79
3.3
2.7
1.1
,87
2.0
1.9
3.0
1.5
.74
1.5
1.0
1.5
1.7
Io3
1.7
.88
.60
.61
1.1
1.2
1.1
3.1
.26
.80
6.0
3.4
2.6
2.3
3.1
.66
1.3
2.3
2.8
2.6
1.9
4.2
.620 <=
.765 <=
1.096 >=
2.189 >
.704 <=
.340 <
1.171 >=»
.257 <
12=458 »
3.381 >
.188 «
.252 <
.760 <=
.843 <=-
.946 <=»
2.262 >
.573 <=•
.652 <-
.369 <
.561 <=*
.906 <-
.305 <
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He COMMENTS
B/A
A/A
A/A
B/B
A/A
A/A
A/A
B/A
B/A
B/A
B/B
A/B
B/B
B/B
B/B
B/B
N/B
B/B
N/N
A/A
A/A
A/B
A
A
I
I
I
A
B
B
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
374
-------
TABLE G-6 (Page 4 of 4)
INTERPRETATION OF CODES:
TABLE ITEM
PLUME HEIGHT VS He
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
COMMENTS
CODE
N
A
B
1
2
3
4
«
<
>
»
N
A
B
B
C
D
E
F
G
H
I
J
K
MEANING
Plume height within 10 m of He
Plume height > He + 10 m
Plume height < He - 10 m
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind .speed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between .*5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 10 m of He
Elevation of maximum > He + 10 m
Elevation of maximum < He - 10 m
The location of the predicted maximum coincides
with or is at the closest adjacent receptor
to the location of the observed maximum
The location of the predicted maximum is at a
receptor close to the location of the observed
maximum (with no more than 1 or 2 receptors
closer to the location of the predicted
maximum)
The predicted maximum concentration is on
the far side of the hill/ridge
The predicted maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration is on
the far side of the hill/ridge
The observed maximum concentration is on
the near side of the hill/ridge
The observed maximum concentration occurs
north of the predicted maximum concentration
The predicted maximum concentration occurs
south of the predicted maximum concentration
The angle formed by the intersection of the
stack-predicted maximum concentration receptor
and the stack-observed maximum concentration
receptor lines is more than 90 degrees
Predicted map is all zeroes
Observed map is all zeroes
375
-------
TABLE G-7
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = CINDER CONE BUTTE TRACER = SF6
CASE 1: WIND SPEED CATEGORY » 1, PLUME HEIGHT BELOW He: NO CASES
CASE 2: WIND SPEED CATEGORY » 1, PLUME HEIGHT NEAR HCJ NO CASES
CASE 3: WIND SPEED CATEGORY =» 1, PLUME HEIGHT ABOVE He: NO CASES
CASE 4: WIND SPEED CATEGORY = 2, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <= >» > » HOURS
16
4
1
6
0
1
0
0
0
28
1
0
0
0
0
0
0
0
0
2
1
0
0
0
0
0
0
0
2
0
1
1
0
0
0
0
0
6
2
0
2
0
0
0
0
0
3
1
0
2
0
0
0
0
0
2
0
0
1
0
1
0
0
0
10
4
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 5: WIND SPEED CATEGORY - 2, PLUME HEIGHT NEAR He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
>»
»
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
2
1
0
0
1
0
TOTAL
HOURS
0
0
0
4
2
0
0
2
0
376
-------
TABLE G-7 (Page 2 of 4)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = CINDER CONE BUTTE TRACER = SF6
CASE 6: WIND SPEED CATEGORY - 2, PLUME HEIGHT ABOVE He: NO CASES
CASE 7: WIND SPEED CATEGORY - 3, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <- >- > » HOURS
0
1
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
2
0
0
1
0
1
0
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 8: WIND SPEED CATEGORY - 3, PLUME HEIGHT NEAR He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
»
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
1
0
0
2
2
0
0
0
0
0
0
0
0
0
0
0
0
1
0
2
0
0
0
0
0
0
1
0
0
0
0
0
• TOTAL
HOURS
0
0
0
2
1
3
0
2
3
11
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
377
-------
TABLE G-7 (Page 3 of 4)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE - CINDER CONE BUTTE TRACER =• SF6
CASE 9: WIND SPEED CATEGORY =• 3, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >= - > » HOURS
0
0
0
1
1
1
1
2
9
15
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
3
0
0
0
0
0
0
1
1
1
0
0
0
0
1
0
0
0
4
5
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 10: WIND SPEED CATEGORY - 4, PLUME HEIGHT BELOW He; NO CASES
CASE 11: WIND SPEED CATEGORY - 4, PLUME HEIGHT NEAR He: NO CASES
CASE 12S WIND SPEED CATEGORY =- 4, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS HC
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RA'TIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >- > » HOURS
1
0
0
3
0
1
0
0
3
0
0
0
0
0
0
0
0
6
0
0
0
0
0
0
1
1
7
0
0
0
0
0
0
0
1
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
3
0
1
1
2
19
27
*SEE INTERPRETATION OF CODES AT END OF TABLE
378
-------
TABLE G-7 (Page 4 of 4)
INTERPRETATION OF CODES:
TABLE ITEM CODE
WIND SPEED CATEGORY.
'RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
1
2
3
4
MEANING
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between
Ratio is between
.2 and .5
.5 and 1.0
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
N
A
B
Ratio is between 1.0 and 2.0
Ratio is between-2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m' of He
Elevation of maximum > He + 5m
Elevation of maximum < He - 5m
379
-------
TABLE G-8
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = CINDER CONE BUTTE TRACER = CF3BR
CASE 1: WIND SPEED CATEGORY =» 1, PLUME HEIGHT BELOW He; NO CASES
CASE 2: WIND SPEED CATEGORY » 1, PLUME HEIGHT NEAR He: NO CASES
CASE 3s WIND SPEED CATEGORY - 1, PLUME HEIGHT ABOVE He: NO CASES
CASE 4: WIND SPEED CATEGORY - 2, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <=» >« > » HOURS
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
******* A
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 5: WIND SPEED CATEGORY » 2, PLUME HEIGHT NEAR He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <- >- > » HOURS
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
380
-------
TABLE G-8 (Page 2 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE - CINDER CONE BUTTE TRACER = CF3BR
CASE 6: WIND SPEED CATEGORY - 2, PLUME HEIGHT ABOVE He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
>-
»
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 7: WIND SPEED CATEGORY - 3, PLUME HEIGHT BELOW He*
0
0
0
0
0 -
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
1
TOTAL
HOURS
0
0
0
0
0
1
0
0
1
DISTRIBUTION OF HOURS BY RATIO
.CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP '5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <- >- > » HOURS
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
381
-------
TABLE G-8 (Page 3 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE =» CINDER CONE BUTTE TRACER * CF3BR
CASE 8: WIND SPEED CATEGORY =» 3, PLUME HEIGHT NEAR He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < » > » HOURS
1
0
0
2
2
2
0
2
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
0
0
1
1
1
0
0
0
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 9: WIND SPEED CATEGORY - 3, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >» > » HOURS
0
0
0
1
1
9
0
1
3
15
0
0
0
1
0
0
0
0
0
0
0
0
0
0
2
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
1
0
0
0
0
0
1
0
1
0
0
0
0
0
0
5
0
0
1
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
382
-------
TABLE G-8 (Page 4 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE - CINDER CONE BUTTE TRACER = CF3BR
CASE 10: WIND SPEED CATEGORY - 4, PLUME HEIGHT BELOW He: NO CASES
CASE 11: WIND SPEED CATEGORY - 4, PLUME HEIGHT NEAR He: NO CASES
CASE 12: WIND SPEED CATEGORY - 4, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS HC/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >- > » HOURS
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
10
10
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
383
-------
TABLE G-8 (Page 5 of 5)
INTERPRETATION OF CODES:
TABLE ITEM CODE
WIND- SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
1
2
3
4
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
N
A
B
MEANING
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind-speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between .5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m of He
Elevation of maximum > He + 5m
Elevation of maximum < He - 5m
384
-------
TABLE G-9
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE » HOG BACK RIDGE TRACER = SF6
CASE 1: WIND SPEED CATEGORY - 1, PLUME HEIGHT BELOW He: NO CASES
CASE 2: WIND SPEED CATEGORY - 1, PLUME HEIGHT NEAR He: NO CASES
CASE 3s WIND SPEED CATEGORY " 1,* PLUME HEIGHT ABOVE He: NO CASES
CASE 4: WIND SPEED CATEGORY - 2, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS HC
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
<«
»
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
0
4
0
0
1
0
0
0
2
3
8
0
0
0
0
0
0
1
0
0
0
1
0
•0
0
0
13
TOTAL
HOURS
4
3
13'
0
1
2
0
0
0
23
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 5: WIND SPEED CATEGORY - 2, PLUME HEIGHT NEAR He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <« >» > » HOURS
ELEV OF PRE
MAX VS HC/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
3
1
1
0
0
1
0
1
1
0
1
0
0
0
0
0
1
2
3
2
1
0
0
1
10
385
-------
TABLE G-9 (Page 2 of 4)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = HOG BACK RIDGE TRACER » SF6
CASE 6: WIND SPEED CATEGORY - 2, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY Of TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
- >
»
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
3
0
0
0
0
0
1
1
0
1
0
0
2
0
0
0
0
0
0
0
2
2
TOTAL
HOURS
1
0
2
I
2
4
14
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 7: WIND SPEED CATEGORY - 3, PLUME HEIGHT BELOW He: NO CASES
CASE 8; WIND SPEED CATEGORY - 3, PLUME HEIGHT NEAR He: NO CASES
CASE 9: WIND SPEED CATEGORY - 3, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >« > » HOURS
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
2
0
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
386
-------
TABLE G-9 (Page 3 of 4)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = HOG BACK RIDGE TRACER = SF6
CASE 10: WIND SPEED CATEGORY =- 4, PLUME HEIGHT BELOW He: NO CASES
CASE^ll: WIND SPEED CATEGORY - 4, PLUME HEIGHT NEAR He: NO CASES
CASE 12: WIND SPEED CATEGORY =» 4, PLUME HEIGHT ABOVE He: NO CASES
387
-------
TABLE G-9 (Page 4 of 4)
INTERPRETATION OF CODES:
TABLE ITEM CODE
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
1
2
3
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
N
A
B
MEANING
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between .5 and 1,0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5. 0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m of He
Elevation of maximum > He + 5m
Elevation of maximum < He - 5m
388
-------
TABLE G-10
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE - HOG BACK RIDGE TRACER = CF3BR
CASE 1: WIND SPEED CATEGORY - 1, PLUME HEIGHT BELOW He: NO CASES
CASE 2: WIND SPEED CATEGORY » 1, PLUME HEIGHT NEAR He: NO CASES
CASE 3: WIND SPEED CATEGORY - 1, PLUME HEIGHT ABOVE He: NO CASES
CASE 4: WIND SPEED CATEGORY = 2, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
***********************************. TOTAL
« < <- >- > » HOURS
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
1
0
0
0
0
0
0
0
0
6
2
2
0
0
2
0
0
0
2
2
2
0
1
0
0
0
1
5
3
0
0
0
0
0
0
1
3
0
0
0
0
0
0
0
0
1
0
1
1
0
0
0
0
0
12
8
18
7
5
1
1
2
0
0
2
36
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 5: WIND SPEED CATEGORY - 2, PLUME HEIGHT NEAR He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >» > » HOURS
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
.0
0
0
0
0
0
0
o-
0
0
0
0
0
1
1
0
0
1
0
0
0
0
3
389
-------
TABLE G-10 (Continued)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE - HOG BACK RIDGE TRACER = CF3BR
CASE 6: WIND SPEED CATEGORY = 2, PLUME HEIGHT ABOVE He*
DISTRIBUTION OF HOURS BY RATIO
ELEV OF PRE CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >= - > » HOURS
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 7; WIND SPEED CATEGORY - 3, PLUME HEIGHT BELOW He; NO CASES
CASE 8: WIND SPEED CATEGORY » 3, PLUME HEIGHT NEAR He: NO CASES
CASE 9: WIND SPEED CATEGORY =» 3, PLUME HEIGHT ABOVE He: NO CASES
CASE 10: WIND SPEED CATEGORY - 4, PLUME HEIGHT BELOW He: NO CASES
CASE .11: WIND SPEED CATEGORY » 4, PLUME HEIGHT NEAR He: NO CASES
CASE 12: WIND SPEED CATEGORY =• 4, PLUME HEIGHT ABOVE He: NO CASES
390
-------
TABLE G-10 (Continued)
INTERPRETATION OF CODES:
TABLE ITEM CODE
WIND SPEED CATEGORY
•RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
1
2
3
4
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
N
A
B
MEANING
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind speed -greater than or equal to 6 m/sec
Ratio .< .2
Ratio is between .2 and .5
Ratio is between .5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 5 m of He
Elevation of maximum > He + 5m
Elevation of maximum < He - 5m
391
-------
TABLE G-ll
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = TRACY POWER PLANT TRACER = SF6
CASE 1: WIND SPEED CATEGORY = 1, PLUME HEIGHT BELOW He: NO CASES
CASE 2: WIND SPEED CATEGORY =» 1, PLUME HEIGHT NEAR He: NO CASES
CASE 3: WIND SPEED CATEGORY = 1, PLUME HEIGHT ABOVE Hcs NO CASES
CASE 4: WIND SPEED CATEGORY =» 2, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >a> > » HOURS
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
5
2
0
1
0
0
0
0
0
7
1
0
0
0
0
0
1
0
7
0
0
0
0
0
0
2
0
6
0
0
1
0
0
0
1
0
0
10
8
0
0
1
4
0
25
36
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE Si WIND SPEED CATEGORY =» 2, PLUME HEIGHT NEAR He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >« > » HOURS
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
"0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
392
-------
TABLE G-ll (Page 2 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = TRACY POWER PLANT TRACER = SF6
CASE 6: WIND SPEED CATEGORY = 2, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
»
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
1
0
4
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
TOTAL
HOURS
0
0
0
0
0
0
1
0
10
11
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 7: WIND SPEED CATEGORY - 3, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <- >«• > » HOURS
'•«•
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
6
*SEE INTERPRETATION OF CODES AT END OF TABLE
393
-------
TABLE G-ll (Page 3 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = TRACY POWER PLANT TRACER = SF6
CASE 8: WIND SPEED CATEGORY =» 3, PLUME HEIGHT NEAR He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <=» >» - > » HOURS
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
2
0
0
1
0
0.
0
1
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 9: WIND SPEED CATEGORY - 3> PLUME HEIGHT ABOVE He*
0
0
1
0
0
0
1
0
6
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE* PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < .<* >m > » HOURS
0
1
1
0
0
1
6
2
17
28
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
1
1
0
0
1
0
0
1
4
0
5
0
1
0
0
0
0
1
0
9
0
0
0
0
0
0
0
0
0
11 11
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 10: WIND SPEED CATEGORY = 4, PLUME HEIGHT BELOW He: NO CASES
CASE 11: WIND SPEED CATEGORY = 4, PLUME HEIGHT NEAR He: NO CASES
394
-------
TABLE G-ll (Page 4 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = TRACY POWER PLANT TRACER = SF6
CASE 12: WIND SPEED CATEGORY = 4, PLUME HEIGHT ABOVE He*
DISTRIBUTION OF HOURS BY RATIO
ELEV OF PRE CATEGORY OF TOP 5 AVE. PREDICTED TO
MAX VS HC/ TOP 5 AVE. OBSERVED CONCENTRATIONS
ELEV OF OBS *********************************** TOTAL
MAX VS He « < <»>.-> » HOURS
B/B 000000 0
B/N 000000 0
B/A 000100 1
N/B 000000 0
N/N 000000 0
N/A 000000 0
A/B 011100 3
A/N 000100 1
A/A 032520 12
TOTALS 043820 17
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
395
-------
TABLE G-ll (Page 5 of 5)
INTERPRETATION OF CODES:
TABLE ITEM
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
CODE MEANING
1 Wind speed less than 1 m/sec
2 Wind speed between 1 and 3 m/sec
3 Wind speed between 3 and € m/sec
4 Wind speed greater than or equal to 6 m/sec
« Ratio < .2
< Ratio is between .2 and .5
<« Ratio is between .5 and 1.0
>» Ratio is between 1.0 and 2.0
> Ratio is between 2.0 and 5.0
» Ratio is greater than or equal to 5.0
N Elevation of maximum is within 10 m of He
A Elevation of maximum > He + 10 m
B Elevation of maximum < He - 10 m
396
-------
TABLE G-12
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE - TRACY POWER PLANT TRACER = CF3BR
CASE 1: WIND SPEED CATEGORY » 1, PLUME HEIGHT BELOW He: NO CASES
CASE 2: WIND SPEED CATEGORY - 1, PLUME HEIGHT NEAR He: NO CASES
CASE 3: WIND SPEED CATEGORY - 1, PLUME HEIGHT ABOVE He: NO CASES
CASE 4: WIND SPEED CATEGORY = 2, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS HC/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
>»
»
1
0
1
0
0
0
1
0
0
1
0
1
0
1
0
3
0
4
11
0
3
0
0
0
2
0
5
4
0
3
0
0
0
0
0
3
3
0
2
0
0
0
1
0
2
0
0
1
0
0
0
0
0
0
10
21
10
TOTAL
HOURS
20
0
11
"0
1
0
7
0
14
53
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 5: WIND SPEED CATEGORY - 2, PLUME HEIGHT NEAR He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < m > » HOURS
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0 .
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
2
397
-------
TABLE G-12 (Page 2 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE =» TRACY POWER PLANT TRACER =» CF3BR
CASE 6: WIND SPEED CATEGORY » 2, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE» OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <= >= - > » HOURS
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
0
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 7: WIND SPEED CATEGORY - 3, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <- >« > » HOURS
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
1
0
1
a
0
i
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
1
0
0
1
0
1
0
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
398
-------
TABLE G-12 (Page 3 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE = TRACY POWER PLANT TRACER = CF3BR
CASE 8: WIND SPEED CATEGORY » 3, PLUME HEIGHT NEAR He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
«
»
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
TOTAL
HOURS
0
0
0
0
0
0
3
0
1
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE 9: WIND SPEED CATEGORY * 3, PLUME HEIGHT ABOVE He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS HC
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS .
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <- >- > » HOURS
3
0
4
0
0
0
7
2
8
24
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
• 1
0
4
1
0
2
0
0
0
4
2
3
2
0
1
0
0
0
2
0
1
0
0
0
0
0
0
0
0
0
12
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
399
-------
TABLE G-12 (Page 4 of 5)
SUMMARY STATISTICS OF CTDM MODEL EVALUATION
SITE - TRACY POWER PLANT TRACER = CF3BR
CASE 10: WIND SPEED CATEGORY - 4, PLUME HEIGHT BELOW He*
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <=» >» " > » HOURS
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
CASE lls WIND SPEED CATEGORY - 4, PLUME HEIGHT NEAR He: NO CASES
CASE 12: WIND SPEED CATEGORY » 4, PLUME HEIGHT ABOVE He*
DISTRIBUTION OF HOURS BY RATIO
CATEGORY OF TOP 5 AVE. PREDICTED TO
TOP 5 AVE. OBSERVED CONCENTRATIONS
*********************************** TOTAL
« < <» >- > » HOURS
ELEV OF PRE
MAX VS He/
ELEV OF OBS
MAX VS He
B/B
B/N
B/A
N/B
N/N
N/A
A/B
A/N
A/A
TOTALS
*******
*SEE INTERPRETATION OF CODES AT END OF TABLE
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
2
0
5
0
0
0
0
0
0
2
2
3
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
4
2
9
17
400
-------
TABLE G-12 (Page 5 of 5)
INTERPRETATION OF CODES:
TABLE ITEM CODE
WIND SPEED CATEGORY
RATIO CATEGORY OF
PREDICTED TO OBSERVED
AVERAGES OF HIGHEST
5 CONCENTRATIONS
ELEVATION OF MAXIMUM
PREDICTED OR OBSERVED
CONCENTRATION VS He
1
2
3
4
«
<
>
»
N
A
B
MEANING
Wind speed less than 1 m/sec
Wind speed between 1 and 3 m/sec
Wind speed between 3 and 6 m/sec
Wind speed greater than or equal to 6 m/sec
Ratio < .2
Ratio is between .2 and .5
Ratio is between .5 and 1.0
Ratio is between 1.0 and 2.0
Ratio is between 2.0 and 5.0
Ratio is greater than or equal to 5.0
Elevation of maximum is within 10 m of He
Elevation of maximum > He + 10 m
Elevation of maximum < He - 10 m
401.
-------
APPENDIX H
CONTRIBUTIONS OF THE FLUID MODELING FACILITY TO EPA'S
COMPLEX TERRAIN MODEL DEVELOPMENT PROGRAM
402
-------
EPA Report Number
May 1987
CONTRIBUTIONS OF THE FLUID MODELING FACILITY
TOEPA'3 COMPLEX TERRAIN MODEL DEVELOPMENT PROGRAM
by
WILLIAM H. SNYDER
Meteorology and Assessment Division
Atmospheric ScienceaTResearch Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
May 1987
Atmospheric Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
403
-------
NOTICE
This information in this document has been funded by the United States
Environment^ Protection Agency. It has been subject to the Agency's peer and
administrative review, and it has been approved for publication as an EPA
document Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
The author, William H. Snyder, is a physical scientist in the Meteorology
and Assessment Division, Atmospheric Sciences Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, NC. He is on
assignment from the National Oceanic and Atmospheric Administration, U.S.
Department of Commerce.
404
-------
FORWARD
The Atmospheric Sciences Research Laboratory (ASRL) conducts intramural
and extramural research programs in the physical sciences to detect, define,
and quantify air pollution and its effects on urban, regional, and global
atmospheres and the subsequent impact on water quality and land use. The
Laboratory is responsible for planning, implementing, and managing research
and development programs designed to quantify the relationships between
emissions of pollutants for all types of sources with air quality and
atmospheric effects, and to uncover and characterize hitherto unidentified air
pollution problems. Information from ASRL programs and from the programs of
other government agencies, private industry, and the academic community are
integrated by the Laboratory to develop the technical basis for air pollution
control strategies for various pollutants.
The Complex Terrain Model Development Program (CTMDP) is designed to
develop reliable atmospheric dispersion models that are applicable to large
pollutant sources located in complex terrain. The major field studies of this
six-year program were conducted during 1980 at Cinder Cone Butte near Boise,
Idaho, during 1982 at Hogback Ridge near Farmington, New Mexico, and during
1983-84 at the Tracy Power Plant near Reno, Nevada Data from these field
studies along with measurements of fluid modeling simulations performed in the
EPA Fluid Modeling Facility are being used to quantify the effects of terrain
obstacles on stable plume dispersion. A series of annual milestone reports
has been issued to describe the development of the Complex Terrain Dispersion
Model (CTDM) and to contrast the performance evaluation of the CTDM against
existing complex terrain dispersion models. This report describes the
contributions of the Fluid Modeling Facility to the Complex Terrain Model
Development Program.
A.H. Ellison
Director
Atmospheric Sciences Research Laboratory
405
-------
ABSTRACT
The contributions of the EPA Fluid Modeling Facility (FMF) to the Complex
Terrain Model Development Program (CTMDP) are described. These contributions
included a wide range of laboratory studies and a limited amount of numerical
modeling of flow and diffusion in neutral and stably stratified conditions in
complex terrain. The goai of the CTMDP is the development of a dispersion
model valid in complex terrain, with emphasis on plume impaclion on nearby
hills during nighttime stable conditions. Work at the FMF prior to the
inception of the program provided the basic framework for the model - the
dividing-streamline concept - and the focal point around which the field
program was designed. Throughout the course of the CTMDP, the FMF interacted
vigorously with the model developers by providing support in various ways.
Early work provided direct support as an aid to planning the details and
strategies of the field experiments and testing the limits of applicability of
the dividing-streamline concept Later work included exercises of "filling in
the gaps' in the field data, furthering the understanding of the physical
mechanisms important to plume impaction in complex terrain and in stably
stratified flows in general, testing various modeling assumptions, providing
data for 'calibration* of various modeling parameters, and testing the ability
of the laboratory models to simulate full-scale conditions. Simultaneously,
the FMF responded to the needs of the regulatory arm of EPA, the Office of Air
Quality Planning and Standards (OAQPS), by providing guidance concerning
expected terrain effects and by conducting demonstration studies. Finally,
several supplemental studies were conducted, broadening and expanding upon the
specific requests of the model developers and the OAQPS.
406
-------
CONTENTS
Forward ///
Abstract iv
list of Rgures " vi
List of Tables. vii
List of Symbols and Abbreviations viii
Acknowledgements ix
1. Introduction 1
2. Background 3
3. Description of Experiments and Results 8
3.1 Direct Interactions with the Model Developers 8
The period 1960 through 1981 8
The period 1982 through 1983 19
The period 1984 through 1985 26
77)0 period 1986 through present 32
3.2 Supplemental Modeling of Complex Terrain 37
Neutral-Flow Wind Tunnel Studies 37
Stably Stratified Towing-Tank Studies 42
4. Summary 45
References 46
407
-------
UST OF FIGURES
Number Title
1 Oblique view of dye streamers released from a horizontal rake 11
upwind of the CCB model at z/h = 0.3 under strongly stratified
conditions (F=0.2). Flow is from the left
2 Top view of dye streamers impinging on CCB under strongly 11
stratified conditions (2/7?=0.3, F=0.4).
3 Vortex rollup and eddy-shedding in the lee of CCB under strongly 12
stratified conditions (z//7=0.6, F=0.2).
4 Oblique view of impinging streamers on CCB. Middle dye streamer 14
is released at the dividing-streamline height; others at ±1cm (±6m
full scale).
5 Comparison of predicted dividing-streamline heights with 15
observations as functions of towing speed. Open symbols:
predictions using integral formula; closed symbols: observations.
6 Concentration distributions measured during individual tows of CCB 17
with H3/rt=0.31 and H0//7=0.38; wind direction: — 117°, —
122s.
7 Scatter diagram comparing superposition of concentration 18
distributions from series of 18 tows of CCB mode! with field
distributions. Dotted lines denote factor of two on either side of
perfect fit
8 Deformation of vertical dye line by upstream columnar disturbances. 23
Dye line was formed at a location 16m upstream of starting position
of fence, at time when fence was at x=12.5m (18.6/1 upstream of
fence). Photograph was taken when fence was at x=» 13.8m (11.6/j
upstream of fence). Fence is out of photograph, approaching from
top left
9 Concentration distributions measured on the hiil surface with 28
H0//j»0.5 and H,//?a0.6. Top: fully submerged; bottom: half
submerged. Dotted circle indicates half the hill height
10 Scatter plot comparing concentrations on fully immersed hill with 29
those on half-immersed hill on a port by port basis. H.//J=0.6,
HD//I-Q.S.
408
-------
11 Terrain amplification factors measured upwind of axisymmetric CCB 31
model. Heavy lines divide the regions into areas where the source.
produced the maximum gic upwind of the hilltop, between the hilltop
and the separation point, and downwind of the hill. Note that the
vertical scale is exaggerated by a factor of 3.
12 Plume cross sections measured in presence ( - ) and in absence 36
( -- — ) of axisymmetric CCB model at x=0 (hill center).
-6,
13 Contours of constant terrain amplification factors over (a) 41
axisymmetric hill and (b) two-dimensional ridge. Note that
vertical scale is exaggerated by a factor of 3.
UST OF FIGURES
Number Title Page
1 Summary of Terrain Amplification Factors for Sources in the 39
vlcinrty of Hills in Neutral Flow.
409
-------
UST OF SYMBOLS AND ABBREVIATIONS
Symbols
A Terrain amplification factor, XmiJ'X.^ol
F Froude number, UJNh
g Acceleration due to gravity
h Hill height
/70 Height of density interface from surface
HD Dividing-streamline height
H, Source height
L Length of ridge
N Brunt-Vaisaia frequency, [-(g/p)dp/dzj
U^ Towing speed or free-stream velocity
xs Source position in along-wind direction (origin at hill center)
ys Source position in crosswind direction (origin at hill center) •
Ap Density difference across interface
p Fluid density
pt Density of fluid between interface and surface
-------
ACKNOWLEDGEMENTS
Many people have contributed to the work described in this report I am
particularly grateful to R.E Lawson for his untiring efforts in the
day-to-day operations of the laboratory, to FLS. Thompson for his unfailing
support and many enlightening discussions, to J.C.R. Hunt for his continual
encouragement, unending infusion of new ideas, and enduring patience in
teaching me so much about stratified flow over obstacles, to R.E. Britter and
I.P. Castro for their many contributions, to G.L Marsh for his dogged
persistence in operating the towing-tank experiments, to M.S. Shipman for his
quiet but solid computer support, to J.C. Smith for his many hours at the
filling station, to G.C. Holzworth for his insistence upon FMF involvement in
the CTMO Program and his acceptance of different viewpoints, and to F.A.
Schiermeier for letting us 'do our thing*. Finally, I wish to express thanks
to the entire FMF staff, past and present, who do the real work day in and day
out and whose efforts too often go unrecognized and unrewarded.
411
-------
1. INTRODUCTION
In the late -\97Os the Office of Air Quality Planning and Standards
(OAQPS) of the Environmental Protection Agency (EPA) identified a crucial need
to develop a mathematical model that dealt with plume impaction from large
sources located in mountainous terrain under stable flow conditions, with
demonstrated reliability. A workshop was convened (Hovind et al., 1979) to
focus on complex terrain modeling problems and to develop recommendations to
EPA with respect to the design of a program of experiments and model
development efforts. Subsequently, Holzworth (1980) outlined the EPA plan to
achieve the objective through an integrated program of model development,
fluid modeling experiments and field studies of plume-terrain interactions on
hills of progressively increasing size and complexity. This multi-year,
multi-faceted program is known as the Complex Terrain Model Development
Program (CTMDP). The" prime contractor for this effort is Environmental
Research and Technology (ERT), which has produced a comprehensive series of
annual reports, called Milestone Reports, that describe ail phases of the
research program. The specific references are: (1) Lavery st al (1982), (2)
Strimartis et al (1983), (3) Lavery ef al (1983), (4) Strimaitis et al (1985),
and (5) DiCristofaro ef al (1986); a final report is to be completed in 1987.
The Ruid Modeling Facility (FMF) interacted vigorously with various
subgroups participating in the CTMDP, and provided direct support and guidance
in many different ways. Whereas the field work and model development effort
up to the present time has been specifically focused on plume impaction under
stable conditions, the work at the FMF has taken a much broader view. The FMF
research program has ranged from the development of broad guidelines (e.g.,
terrain amplification factors) and physical concepts (e.g.,
dividing-streamline height) to specific site studies (e.g., Cinder Cone Butte)
and regulatory applications (e.g., good-engineering-practice stack height).
The FMF has provided laboratory data to "fill in the gaps" in the field data
(e.g., measurements of plume deformations over hills) and tested the validity
of convenient modeling assumptions (e.g., cut-off hill approach).
412
-------
This report summarizes the contributions, both direct and indirect, of
the FMF to the CTMDP. The discussion provides a historical perspective and a
comprehensive list of FMPs accomplishments with respect to furthering the
physical understanding of flow and diffusion in complex terrain. In many
cases the early research results were first pubished as internal documents or
project reports or presented at workshops or conferences in order to speed the
flow of information to the model developers. In most cases, these results
have been published in peer-reviewed journals (which took, in one extreme
case, 8 years to appear in print). For completeness and to provide the proper
perspective, both references are cited at first mention in the text that
follows; thereafter, only the journal publication is cited.
413
-------
5 BACKGROUND
Research work conducted at the FMF prior to the inception of the CTMDP
had a strong influence on the directions to be taken in the field work and on
the type of model (I.e., physical concepts) to be developed. The stratified
towing tank was commissioned in 1976 (Thompson and Snyder, 1976) and rather
fundamental studies were begun immediately on the structure of stably
stratified flow over idealized three-dimensional hills and on diffusion from a
point source within a stably stratified field of turbulence.
The first published reports on this work (Hunt er a/, 1978; Hunt and
Snyder, 1980) described the flow structure observed over a bell-shaped hill
under neutral and stably stratified conditions. Earlier theoretical work by
Drazin (1961), model experiments by Brighton (1978) and Riley ef al (1976),
and observations (e.g., Queney er a/, 1960) all indicated that, when the
stratification is strong enough, the air flows in approximately horizontal
planes around the topography. And this observation had been used by EPA in
estimating the surface concentrations on hills caused by upwind sources of
pollution (Burt and Slater, 1977). Up to that time, however, there had been
littie firm laboratory or field data as to how strong the stratification must
be for any given streamline starting below the hill top to pass round the side
rather than over the top of the hilt The Hunt and Snyder (1980) paper
suggested a criterion for this change-over to occur on the basis of the
tow-Froude-number theory of Drazin (1961), and confirmed that criterion with
experimental data
The Drazin (1961) theory is applicable to strongly stratified flows
around three-dimensional hills; indeed, it is asympotically valid at
zero-Froude-number. In simplistic terms, the theory suggests that the
stratification inhibits vertical motions, so that fluid parcels are
constrained to move in horizontal planes. Hence, the flow may be described in
terms of two-dimensional flow around a cylinder which is not necessarily
circular but in fact, has the cross-sectional shape of the intersection of a
horizontal plane with the three-dimensional hiil. Hunt and Snyder (1980)
414
-------
verified that, for a bell-shaped hill, a linearly stratified environment, and
an effectively uniform approach-flow velocity profile, Drazin's theory was
applicable in the range F<0.4, where F is the Froude number (=UJNh, U^ being
the towing speed, N the Brunt-Vaisala frequency, and h the hill height).
More importantly, Hunt and Snyder (1980) showed evidence for a dividing
streamline (on the centerplane determined by the flow and the axis of the
axisymmetric hill) of height H9 such that streamlines below H, would impinge
on the hill surface and follow the surface around the sides, whereas
streamlines above H9 would go over the top. They suggested the simple formula
H9 - h (1 - F) (1)
as the criterion to determine whether a plume embedded in the flow approaching
the hill would impact on the surface or surmount the top, for 0
-------
and resulting surface concentrations were amplified by Hunt et al (1979).
This primarily theoretical work described two conceptual models for dealing
with the problem of plume impingement The first model was applicable to
strongly stratified flow around three-dimensional hills, where vertical motion
and vertical diffusion is negligible. The advective-diffusive equation around
a three-dimensional hill which is axisymmetric about a vertical axis was
solved (using an eddy diffusivity) to show how source positions on and off the
centerline affect *ie trajectories and splitting of impinging plumes and the
value and position of the maximum surface concentration on the hill. The
results showed that the plume behavior is very sensitive to quite small
changes in wind direction away from the direction that transports the plume
onto a stagnation point, and the model provided a simple way to estimate the
effect of these changes. This model also allowed the computation of
concentrations within the separated, horizontally recirculating wake of the
hill (source upwind of hill).
In the second model, a plume in a neutrally stable potential flow around
a hemisphere was analyzed, also using the diffusion equation. The solutions
showed how, because streamlines approach the surface of a three-dimensional
hill much more dosely than that of a two-dimensional hill, the maximum
surface concentration on the hill can become very much greater than in the
absence of the hill (but only for a limited range of source heights).
Prior to the inception of the CTMDP, another complex terrain model was
developed by ERT under contract to EPA. The algorithm developed at that stage
was generally applicable to plume behavior in stability conditions ranging
from neutral to slightly stable. The general approach followed the theory of
turbulent plumes embedded In potential flow fields as developed by Hunt and
Mulheam (1973), Snyder and Hunt (1984 - original manuscript made available to
ERT in 1978), and Hunt et al (1979). This theory was applied to the
calculation of ground-level concentrations using a Gaussian form of solution
to the diffusion equation. Stream functions appropriate to the potential flow
over a cylinder (aspect ratio, /7/L-«) and to the potential flow over a sphere
(h/L-1) form the cornerstones of the model. These solutions were extended to
describe flows over terrain features of intermediate crosswind aspect ratio by
a weighting of the two limiting stream functions. The derivation of this
weighing scheme relied heavily on wind-tunnel experiments of flows over hills
of various aspect ratios (Snyder and Britter, 1987; data reports made
416
-------
available to ERT in 1979). Strictly speaking, this algorithm was applicable
to neutral flows, but an empirical approximation was included to define
streamline lowering caused by an imposed stable stratification. This
empirical scheme was derived on the basis of the stratified towing-tank
experiments of Hunt and Snyder (1980). Extensive comparisons of model
predictions with FMF laboratory data were made for both neutral and weakly
stable conditions. A full account of these model-development efforts and the
essential physics of the model are provided by Bass et al (1981). These
algorithms were subsequently incorporated into a routine operational model
called COMPLEX/PFM (Potential Row Model; Strimaitis et al, 1982). In
COMPLEX/PPM, potential flow calculations are performed whenever the plume lies
above the dividing-streamline height and the stability is between neutral and
slightly stable; when the plume is below the dividing-streamline height, the
model reverts to the standard COMPLEX I computation (see Wackter and
Londergan, 1984). The COMPLEX I computation makes the level-plume assumption,
with an effective doubling of surface concentration above the plume centeriine
concentration (Burt and Slater, 1977). This particular aspect of the complex
terrain diffusion problem was one of the hotly contested issues that provided
the impetus for theCTMDP.
A •strawman' was proposed by Holzworth and Snyder (1979) for discussion
at the 1979 workshop convened by EPA to make recommendations with regard to
the directions to be taken under the CTMDP. This strawman was hotly debated
at the workshop and, in the end, was largely accepted by the workshop
participants (Hovind et al, 1979). The plan that emerged (Holzworth, 1980)
called for an enlargement of some of the major concepts arising from the
previous work at the FMF, and for a verification of these concepts through the
conduct of a series of field studies on hills of progressively increasing size
and complexity.
Prior to the request for bids on the CTMD contract, a preliminary
one-week field study of the nighttime flow patterns at Cinder Cone Butte was
organized and conducted primarily by FMF personnel (Snyder er al, 1980); the
primary purpose was to assess the suitability of Cinder Cone Butte as the site
for the first small hill study (identified in the ERT Milestone reports as
Small Hill Impaction Study # 1). Numerous observations were made of the flow
structure and plume behavior around the hiil, including (1) plumes spread
broadly in the lateral direction but very thinly in the vertical direction
4L7
-------
over the hill in neutral conditions, (2) lee-side separation under
high-wind-speed, neutral conditions, (3) plume impingement under light wind,
strongly stable approach flows, and (4) katabatic winds under light-wind,
neutral approach flow conditions with clear night sky. Cinder Cone Butte was
judged as ideally suited for the first small hill study in several respects.
Finally, numerous suggestions were offered as an aid to the design and conduct
of future field studies at this site (most of which were adopted in the later
studies).
To recap the "state of the science" immediately prior to the contract
award, the dividing-streamline concept had been shown to be a useful
conceptual framework to use in describing the structure of strongly stratified
flow around three-dimensional hills. It had only been shown to be valid,
however, for quite a limited number of hill shapes, all of which were
axisymmetric. It had only been verified under uniform stratification (linear
density gradient) or under a step inversion (sharp density interface), under a
uniform approach-flow velocity profile, and, of course, only under
steady-state, small-scale laboratory conditions (although the preliminary
field study provided reassurances of the validity of the concept).
418
-------
3. DESCRIPTIONS OF EXPERIMENTS AND RESULTS
3.1 Direct Interactions with the Model Developers
/
The period 1980 through 1981
The major CTMDP contract was awarded in June, 1980, and the work plan
called for the small hill impaction study to begin at Cinder Cone Butte (CCB)
in September. Almost immediately, the FMF was called upon to conduct
towing-tank experiments to aid in the detafled planning and design of the
field experiments. The first request was to provide quidance with regard to
the location of the main meteorological tower. The second request was to
provide guidance for smoke- and tracer-release strategies, for preselecting
locations for samplers and cameras, and for choosing in advance several
different sampler strategies to account for variations in flow regimes and
wind fields. The third request was to test the validity of an integral
formula for predicting the dividing-streamline height
At an early July meeting at ERT headquarters in Boston, MA, the question
arose as to how to predict the dividing-streamline height when the wind
profile was not uniform and the density gradient was not linear. This was of
paramount Importance in planning the release scenarios, as the release
locations and heights were to be chosen in real time during the field study
based upon the incoming real-time meteorological data J.C.R. Hunt
immediately sketched the now well-known integral formula (on the back of an
envelope!) as
*) [- 1|
^
pt (",) - g *) - cfc . (3)
This formula had, in fact, been published 24 years earlier by Sheppard
(1956) as a small note, actually in answer to his own question which arose at
a meeting of the Royal Meteorological Society, although Shepparcfs note was
virtually unknown to the modeling community at that time. This integral
formula is based upon simple energy arguments. Sheppard asked the question:
in a strongly stratified flow approaching a hill, does a particular fluid
parcel at. some height upstream possess sufficient kinetic energy to overcome
419
-------
the potential energy required to lift itself through the potential density
gradient from its upstream elevation to the hill top? The left-hand side may
be interpreted as the kinetic energy of the parcel far upstream at elevation
Hv and the right-hand side as the potential energy gained by the parcel in
being lifted from the dividing-streamline height H9 to the hill top h through
the density gradient dp/dz. This integral formula was presumably applicable
to a fluid with any shape of stable density profile and, presumably, with any
shape of approach-flow velocity profile. In practice, it must be solved
iteratively, because the unknown H3 is the lower limit of integration; the
formula can easily be reduced to the simpler formulae (1) and (2) by using the
boundary conditions applicable to those special cases. The third request to
the FMF was thus to verify this integral formula
Three studies were conducted in the summer of 1980, and three reports
were prepared in response to these requests. In the first study (Snyder,
1980a), twenty six separate tows of a mode! of COB were made through the tank
In a two-week period. The objective was to assess the suitability of the
particular site chosen for the main (150m) meteorological tower, /.e, was it
close enough to COB to be representative of the flow approaching the hill, yet
far enough away that the measurements were unaffected by the hill itself? It
was impossible, of course, to meet both these criteria for all wind
directions; the question addressed in the towing-tank studies, then, was
whether the flow field at the proposed tower site would be perturbed by the
hill, given the climatological ranges of prevailing wind directions for light,
nighttime winds. Measurements were made of surface flow patterns, deformation
of material lines, velocity profiles and streamline patterns over a model of
CC8, and these measurements were compared favorably with predictions of
potential flow theory. The findings from the study suggested that no
significant perturbations to the approach wind field were to be expected due
to the presence of the hill when the wind direction was outside the range of
the prevailing wind directions. Nevertheless, a shorter (20m) tower was
recommended, to be erected at the trailer site (3km ESE of the hill). This
recommendation was indeed implemented in the field study.
This was, in fact, the first real-terrain model {i.e., non-idealized
shape) to be studied in the stratified towing tank. As a side benefit,
therefore, the study provided reassurances that the basic flow features and,
of course, the same physical principles applied to more realistically shaped
420
-------
hiils. Some examples are provided here to illustrate the point. Rgure 1
shows an oblique view of the hill, where neutrafy buoyant dye streamers were
released from a horizontal rake upstream of the hill under strongly stratified
conditions. The horizontal nature of the flow field is immediately obvious.
Rgure 2 shows a top or plan view of the CCB model under roughly similar
conditions. One of the dye streamers obviously impinged on the upwind
shoulder of CCB. Row separation in the lee of the hill is dramatically
illustrated in Rgure 3, where vortex roll-up and eddy-shedding in the lee are
quite vivid. The Karman vortex street was a common occurrence at low Froude
number; a street appeared to form at all elevations below the
dividing-streamline height (at least for smaJ Froude numbers), but the
shedding frequency seemed to vary with elevation and motions at different
elevations were seemingly uncorrelated with one another.
In the second phase of this summer series, eleven tows of the CCB model
were made during which two of the model developers from ERT participated as
observers. In this case, vertical rakes of tubes emitted neutrally buoyant
dye at up to 6 elevations, with different colors of dye being emitted at the
different levels. Each tow was filmed from the side using a camera that moved
with the hill, and from directly below using a fixed camera pointed upward at
the (inverted) model hill. The films were viewed with an analyst's projector,
and the plume paths and envelopes were sketched. These results corroborated
the previous results of Hunt and Snyder (1980) on idealized hills, i.e., that
plumes below the dividing-streamline height Hs and on a stagnation streamline
would impinge on the upwind side of the butte and flow around the sides, and
that plumes released just above H9 may produce maximum ground-level
concentrations on the upwind side as they pass over the top. The results
further emphasized that plumes travelling in a direction only slightly away
from that of the stagnation streamline would tend to pass around CCB without
significant impact, and that plumes released somewhat higher above H, may be
caught in strong downslope flows and produce maximum ground-level
concentrations on the lee side of the hill. The results were also used, of
course, for the originally intended purpose as a guide for planning of release
and sampler strategies and selection of sampler and camera locations. The
results are described by Bass (1980).
In the third phase of this summer series, the goal was to test the
validity of the integral formula for the height of the dividing streamline
421
-------
Figure 1. Oblique view of dye streamers released from a horizontal rake
upwind of the CCB model at z/r?=0.3 under strongly stratified conditions (F =
0.2). Row is from the left
\
Rgure 2. Top view of dye streamers impinging on CCB under strongly
stratified conditions (z/h^Q.3, F = 0.4).
422
-------
Figure 3. Vortex rollup and eddy-shedding In the lee of CCB under strongly
stratified conditions (///»=0,6, f = 0.2).
-------
under density profiles more typical of those expected at CCB. A typical
nighttime temperature profile in the Snake River Basin (site of CCB) was found
to consist of a strong, surface-based immersion of depth 50 to 100m and a
weaker inversion above extending to several hill heights. Hence, the
stratified towing tank was filled with a strong density gradient near the
surface and a weaker gradient abova (In actuality, the weaker gradient was
below the stronger gradient, but, as the model is towed upside down, we will,
for ""clarity, describe the behavior as if the model were right-side-up.) The
break-point between the two gradients was * initially slightly above the crest
of CCB (at 1.25/7). A vertical rake of 3 tubes was positioned well upwind of
the hill, with vertical spacing between the tubes of 1cm, which is equivalent
to 6.4m full scale or 0.064/7. Neutrally buoyant dye was emitted from each
tube. For each tow, a particular stack height (center tube) was chosen and
the general formula was integrated numericaiy using the measured density
profile to predict the towing speed required such that the center streamer
would rise to the elevation of the saddle point of CCB, La., the minimum
height of the draw between the two peaks. If the formula were correct, then,
the lower streamer should go around the side of the hill, the upper streamer
should go over the top, and the center one should split The height of the
break-point between the two gradients was then reduced and the process
repeated. In all, twelve tows were made, varying the height of the
break-point or the dividing-streamline height (release height) each time.
Rgure 4 shows a side view of the impinging streamers during a typical
tow, i.e., the upper streamer going through the draw, the lower streamer going
round the side, and the middle one splitting. Rgure 5 shows the results in
quantitative fashion. The density profiles were integrated in accordance with
Equation (3) to find the dividing-streamline heights (based on the height of
the saddle point) as functions of the towing speed. These predictions are
shown in Rgure 5 as the continuous lines. The observations of the
dividing-streamline heights made during the twelve tows are ateo plotted in
the figure; the agreement between the predictions and observations is regarded
as excellent The error bars result because of some fluctuating behavior of
the streamers, especially at the higher speeds; occasionally, an intermittent
vortex at the top windward side of the hill would engulf all three streamers
and they would all go round the sides temporarily; on other occasions, parts
of the lower streamer could be observed passing through the draw. The results
424
-------
' / I \ V
Rgure 4. Oblique view of impinging streamers on CC8. Middle dye streamer is
released at the dividing-streamline height; others at ± 1cm (±6m full scale).
-------
o
•*
H
X
O
UJ
UJ
UJ
cc
I-
OT
O
z
o
5 10 15 20 25
TOWING SPEED, CM/S
30
Rgure 5. Comparison of predicted dividing-streamline heights with
observations as functions of towing speed. Open symbols: predictions using
integral formula; closed symbols: observations.
426
-------
of this set of experiments (Snyder, 1980b) provided confidence in the validity
of the general integral formula for predicting the height of the dividing
streamline for a wide range of shapes of stable density profiles.
During the six-week field study at COB, detailed measurements were made
of wind, turbulence, and temperature profiles in the approach flow and at
other positions on the hill. Sulfur hexafluoride (as a tracer) and smoke (for
flow visualization) were released from a platform suspended from a mobile
crane that allowed flexibility in positioning the source (height and
location). One hundred samplers on the hill collected data on surface
concentrations, and lidar was used to obtain plume trajectories and
dimensions.
One particular hour from the field study was selected for simulation in
the towing tank (Snyder and Lawson, 1981). That hour was 0500 to 0600, 24
October 1980 (Case 206), which may be characterized as very stable, i.e.,
light winds and strong stable temperature gradients. Measurements made during
the towing-tank experiments included ground-level concentrations under various
stabilities and wind directions, vertical distributions of concentration at
selected points, plume distributions in the absence of the hill, and visual
observations of plume characteristics and trajectories.
This series of tows showed that the surface-concentration distributions
were extremely sensitive to changes in wind direction. For example, Rgure 6
shows that the distribution shifted from the north side of the hill to the
south side with a shift of only 5° in wind direction. Comparisons of
individual distributions with field results showed very much larger maximum
surface concentrations and much narrower distributions in the model results.
To account for the large variability in the winds measured during the hour, a
matrix of 18 tows (three wind directions x six wind speeds) was conducted, and
the concentration patterns were superimposed. A scatter plot of superimposed
model concentrations versus field concentrations (Figure 7) shows a marked
Improvement over the single-tow comparisons. The largest model concentrations
were within a factor of two of the highest field values, and 70% of the model
concentrations were within a factor of two of the observed field values.
It was interesting to team that, whereas the location of the maximum
shifted dramatically with small shifts in wind direction, the value of the
maximum changed very little with changes in wind direction or wind speed.
Maximum surface concentrations approached those at the plume centeriine in the
427
-------
600 M
SCALE
SOURCE
200M
Rgure 6. Concentration distributions measured during individual tows of COS
with H,//)»0.31 and HD/ft = 0.38; wind direction : 117°, 122°
428
-------
10000
1000
z
111
o
o
o
111
o
o
100
10
10
100
1000 10000
FIELD CONCENTRATION
Rgure 7. Scatter diagram comparing superposition of concentration
distributions measured over Cinder Cone Butte with field distributions.
Dotted lines denote factor of two on either side of perfect fit.
429
-------
absence of the hill during individual tows, but because of the extreme
sensitivity of the location to wind direction, the plume was 'smeared' broadly
across the hill surface as the wind direction changed through only a few
degrees. Therefore, short-term averages (*5min) in the field may be expected
to approach plume-centerline concentrations; longer-term averages («1h) may be
expected to be reduced by factors of five to ten (or more depending upon the
magnitudes of the fluctuations in wind speed and direction).
The period 1982 through 1983
Although considerable work had been done concerning the validity and
limits of applicability of the dividing-streamline concept, several questions
still remained. One question concerned the effects of shear in the
approach-flow velocity profile. Another concerned the effects of the aspect
ratio (ratio of crosswind length of the hill to its height) and, in
particular, its applicability in strongly stable flows to a truly
two-dimensional ridge. A third questioned the effects of the slope of the
hill, and a fourth, the effect of wind angle on a long ridge.
A few other studies had shed light on some of these problems. Baines
(1979), for example, had conducted towing-tank studies of low-Froude-number
flows around a barrier with a gap. His results suggested
HJh - 1-2F (4)
for barriers with very small gaps, tending toward Hs/ri = 1-^ (Equation 1) for
those with wider gaps. Weil et a/ (1981) conducted similar towing-tank
studies, extending the work of Baines, and found quite similar results.
However, data from a field study by Rowe et a/ (1982) of stable air flow over
a long' ridge showed much better agreement with the data for axisymmetric
hills (Equation 1) than for ridges with gaps (Equation 4).
In the eariy 1980*3, a series of experiments was done by numerous
investigators at the FMF and for a variety of different purposes. The overall
objective was to gain fundamental understanding of flow and diffusion under
stably stratified conditions in complex terrain, but the individual projects
were designed with very specific and limited objectives in mind.
Nevertheless, one aspect of each of the projects was to examine the concept of
the dividing-streamline height, as it obviously had very important
consequences with respect to the CTMDP. The results of most of these projects
were published separately and independently, as will be referenced below, but
430
-------
the results concerning the validity and applicability of the
dividing-streamline concept were extracted and published as an appendix
(Snyder et al, 1983) to the Second Milestone Report (Strimaitis et al, 1983)
in order to provide timely support and guidance to (1) the mathematical
modelers attempting to expand their models to include a wide variety of
terrain shapes and approach flows and (2) planners of the Second Small Hill
Impaction Study, which was to take place at the Hogback Ridge in northwestern
New Mexico. This paper was subsequently published in a journal (Snyder ef al,
1985). The individual laboratory experiments included:
1. Towing-tank studies on truncated, steep-sided ridges of various '
crosswind aspect ratios. These included examination of upstream
•blockage* regions, surface flow patterns and lee-wave structure
and were reported by Castro ef al (1983); those aspects dealing
specifically with the dh/iding-streamfine concept were reported by
Snyder ef al (1983) and Snyder et al (1985).
2. Stratified wind-tunnel studies (in Japan) on shear flow over
vertical fences of various crosswind aspect ratios and over a model
of Cinder Cone Butte. (Snyder and Ogawa, 1982; Snyder ef al,
1985).
3. Towing-tank studies on a truncated sinusoidal ridge with a maximum
slope of 40° positioned perpendicular and at other angles to the
approach wind direction (Lee ef al, 1984a, 1984b).
4. Towing-tank studies on an 'infinite' triangular ridge and a long
sinusoidal ridge to test the validity of the 'steady-state*
assumption of flow upwind of an obstacle under strongly stratified
conditions.
The conclusion from the studies with truncated triangular and sinusoidal
ridges perpendicular to the wind was that the aspect ratio per se, does not
have a significant influence on the dividing-streamline height Hy Deviations
from the H,//7=1-f rule were attributed to the combination of shear in the
approach flow and the very steep slope of the triangular ridges, which
resulted in the formation of an upwind vortex with downward flow on the front
faces of the ridges. The *1-P rule was validated for the sinusoidal ridge
with a length-to-height ratio greater from 16:1; in this case, the shear in
the approach flow was much less pronounced, and the upwind slope was
substantially smaller. Note that these deviations to the "\-F rule did not
invalidate Sheppartfs concept, but required a reinterpretation of the rule as
a necessary but not sufficient condition, i.e., a fluid parcel may possess
431
-------
sufficient kinetic energy to surmount a hill, but it does not necessarily do
so.
In the stratified wind-tunnel studies, reasonably strong shear layers
with depths more than twice the hill heights were developed in conjunction
with strong stable temperature gradients. These approach flows provided
dividing-streamline heights as large as 0.75/7. In the vertical fence studies
with a stratified approach flow, the shear was found to have an overwhelming
influence. The conclusions were: (a) as in the triangular ridge studies, the
aspect ratio was relatively unimportant; the basic flow structure was
independent of aspect ratio; (b) the shear, in conjunction with the steep
slope, created an upwind vortex such that plumes were downwashed on the front
faces; and (c) under strong enough stratification, there was a limit to the
downward penetration of elevated streamlines; the extent of this penetration
appeared to be predictable as a balance between kinetic and potential
energies. However, when these same shear flows approached the much lower
sloped CCB model, there was no evidence of upwind vortex formation. Limited
concentration measurements on the CCB model suggested that Sheppard's integral
formula correctly predicted the height of the dividing streamline.
From the sinusoidal ridge studies with wind angles at other than 90°, it
was concluded that the effect of deviations in wind direction (from 90°) are
relatively insignificant until the wind direction is in the vicinity of 4Se to
the ridge axis. At 30°, significant departures from the "1-P rule were
observed; the fluid had sufficient kinetic energy to surmount the ridge, but
found a path requiring less potential energy round the end of the ridge. When
the dye streamers were moved closer to the upstream stagnation streamline
(upwind of the upstream end of the ridge), they behaved according to the "1-P
rule.
The two-dimensional ridge studies showed that steady-state conditions are
not established in strongly stratified flows (say F<1). Two different
physical mechanisms give rise to this unsteadiness; one is called 'squashing",
the other, upstream wave propagation. Brief explanations will be given here;
the interested reader should consult the cited references.
The squashing phenomenon is most easily described in terms of the simple
energy arguments as used in deriving Sheppard's formula (Equation 3). As
discussed there, a fluid parcel with insufficient kinetic energy to overcome
the potential energy requirement to surmount the hill must pass round the
432
-------
sides of the hill. But a two-dimensional hill has no sides around which to
pass (in a towing-tank, a two-dimensional' hill is one that spans the entire
width of the tank). Hence, the fluid parcel must be brought to rest In the
towing tank, since the fluid is generally at rest and the hill is towed, this
means that the fluid ahead (upstream) of the hill must be pushed ahead of the
hill, instead of being allowed to surmount the hill top. However, the
upstream endwall of the towing tank, of course, prohibits this fluid from
being pushed. Hence, the fluid between the hill and the upstream endwall is
•squashed" as the hill approaches the endwall; because the fluid is
incompressible, it must rise and spill over* the top of the hill, just as the
water in a bucket will rise and spill over the top when the sides are
•squashed".
This squashing phenomenon seems to have no counterpart in the atmosphere.
If true blocking occurred upwind of an "infinite' ridge in the atmosphere, it
seems that the flow would be blocked to infinity upwind (i.e., there is no
'endwall" forcing the flow toward the ridge). In more practical terms,
•blocking" upstream of a very long ridge would imply 'upstream influence* to
very large distances, possibly through an upstream-propagating front which
would imply non-steady-state behavior. From another viewpoint there are no
infinite ridges in the real world, so that fluid parcels can always be
diverted around the obstacles without changing their elevation.
The results leading to the '1-2P formula (Equation 4) by Baines (1979)
and Weil ef. a/ (1981) for two-dimensional ridges and ridges with gaps were
surprising because they suggested that fluid parcels could surmount the hills
even though they had insufficient kinetic energy to do so. Snyder et al
(1983, 1985) suggested that these earlier results were erroneous; that they
were largely due to the squashing phenomenon, i.e., the gaps in their ridges
were insufficiently large to allow a "relief valve* to avoid the squashing.
Upstream wave propagation is also possible in stratified flows. The
introduction of an obstacle in a stratified flow on which lee waves can form
will result in 'columnar* disturbances extending upstream (see Turner, 1973);
if such motions are present they will modify the approaching flow. These
columnar disturbances take a sinusoidal form in the vertical, with the 'mode*
(number of oscillations) being dependent upon the Froude number based on the
depth of the tank. An example of an upstream columnar disturbance is shown in
Figure 8. Dye crystals were dropped into the stratified tank at a position
433
-------
Figure a Deformation of vertical dye tine by upstream columnar disturbances.
Dye line was formed at a location 16m upstream of starting position of fence,
at time when fence was at x-12.5m (18.6/j upstream of fence). Photograph was
taken when fence was at x- 13.8m (11.6/7 upstream of fence). Fence is out of
photograph, approaching from top left
434
-------
16m upstream of the starting position of an obstacle (in this case, a vertical
fence) which was to be towed along the water surface. These crystals
dissolved as they sank to the bottom, leaving behind a vertical dye line. The
dye line was formed after the commencement of the tow, and the photograph was
taken well before the obstacle reached the dye-line position, i.e., the dye
line was deformed by the upstream columnar disturbance into the sinuous curve
shown in Rgure 8.
These columnar disturbances, unlike the squashing phenomenon, do have
counterparts in the real atmosphere. They result in 'blocking" and 'upstream
influence*. However, in the laboratory tank, these upstream waves are
reflected from the upstream endwall of the tank and return to modify the flow
locally around the model hill; this reflection from the upstream endwall does
nor have a counterpart in the real atmosphere. Baines (1979) argued that
valid observations could be made of the flow over and around the obstacle in
isolation (in the absence of end effects) by making the observations after
steady state was reached (estimated by direct observation), but before
reflected upstream motions arrived. Evidently, he believed that a local
steady state was achieved in that, at some not-too-distant point upstream of
the obstacle, steady-state velocity and density profiles were established
before the reflected motions returned to modify them.
Snyder ef al (1983; 1985) showed that steady-state conditions are not
established in strongly stratified flows (say F<1) over two-dimensional
ridges. The squashing phenomenon and reflections of upstream columnar
disturbances continuously changed the shapes of the 'approach flow" velocity
and density profiles. Thus, these experiments have no analogue in the real
atmosphere. Further, because long ridges cut by periodic small gaps require
very long tow distances in order for steady state to be established, Snyder ef
a/ concluded that the previous laboratory studies were not valid models of
atmospheric flows; specifically, the H,//7»1-2F formula proposed for flow about
ridges with small gaps is not expected to apply to the real atmosphere.
Further work was done to better understand the nature and causes of these
upstream motions and lee waves by Thompson and Snyder (1984), Castro (1987)
and Castro and Snyder (1987b, 1987c), but the interpretation of these results
is somewhat controversial. More work-is required to establish the precise
relationships between model size and shape, stability, and tank size, shape
and configuration in order to determine the limits of applicability of fluid
435
-------
modeling and ranges of transferability to the atmosphere.
The second Small Hill Impaction Study was conducted during October 1982
at Hogback Ridge (HER) near Farmington, NM. In providing input to the
experimental design, the FMF conducted a series of wind-tunnel and towing-tank
flow-visualization experiments prior to the field study. The laboratory
studies were designed to investigate
• plume height above the surface over the hill crest and at
the upwind edge of the hiil,
• apparent size of any plume deformation upwind of the hill,
• lee wave importance and structure, and
• sensitivity of the plume trajectory to "wind angle".
This information was subsequently used by the field designers to guide the
design of the smoke and tracer-gas release protocols at HBR, and to help
select sampler and camera locations.
Two tests were made in the wind tunnel. One test was done with the ridge
perpendicular to the flow, the other with the ridge rotated by 30°. These
tests suggested that in neutral conditions the streamline patterns were
similar to those expected from potential flow theory; a plume released at a
given height upwind of the ridge should traverse the crest at an elevation of
one-half its initial height The test with the ridge at an angle to the flow
showed only a very small (<4°) deflection of the plume path as the plume
traversed the ridge.
Eight individual tows of the HBR model were done in the stratified towing
tank, varying the Froude number and wind direction, and each time releasing
dye at eleven different elevations upstream. Heights of these dye streamers
were measured at the upstream base and at the crest of the ridge. These
experiments snowed that, during weakly stratified conditions, plumes rose near
the upwind base and fell over the crest to near or slightly lower than their
upstream heights. Low-level releases experienced extensive mixing. More
detailed results are contained in the Third Milestone Report (Lavery ef a/,
1983, p. 117-123).
Around this same time period, the FMF undertook two separate laboratory
experiments ttiat attempted to simulate two specific one-hour periods as
observed in the field at Cinder Cone Sutte. The first simulated a neutral
stability period in the Meteorological Wind Tunnel (Thompson et a/, 1983).
436
-------
The second simulated a moderately stable period in the stratified towing tank
(Eskridge et al, 1983). [Recall that the simulation of a strongly stable
period was described earlier (Snyder and Lawson, 1981).]
In the summer of 1983, Ben Greene and coteagues from ERT, in cooperation
with the FMF staff, conducted experiments in the Meteorological Wind Tunnel to
characterize the response of the Climatronics UVW propeller anemometers. The
primary objectives of the experiments were to determine the calibration curves
and the non-cosine response corrections, especially at low wind speeds. The
results of these tests are contained in the Fourth Milestone Report
(Strimaitis et al, 1985, p. 85-93). The cafibration factors and non-cosine
response correction factors were applied to the HBR data base in forming the
Modeler's Data Archive.
In late summer of 1983, discussions were held with ERT concerning
possible contributions of FMF to the Full-Scale Plume Study planned for the
following year at the Tracy Power Plant near Reno, NV. Considerations of
scaling the site for towing-tank studies revealed that, at any reasonable
scale, the model would appear as a two-dimensional ridge with a small gap
(river valley) running through it Recent work at the FMF as discussed above
had shown that this situation could not be modeled under strongly stable
conditions. Hence, specific site modeling at the Tracy Power Plant was not
undertaken at the FMF. Instead, other studies in direct support of the model-
development effort were undertaken as described below.
The period 1984 through 1985
In September 1983, A. Venkatram, D. Strimaitis and R. Britter from ERT
requested that the FMF conduct two studies in support of their
model-development efforts. The first study attempted to shed light on the
question of the validity of the assumption of a flat dividing-streamline
surface, a key assumption in the model under development The second study
was to provide a complete set of data on neutral flow and diffusion around a
three-dimensional hill with a shape and slope approximating that of Cinder
Cone Butte. These data were to help ERT to evaluate the separate effects of
plume deformation kinematics and those of increased turbulence around the
hill.
The first study, testing the validity of the flat dividing-streamline
assumption, consisted of a series of 26 tows of a model hill in the stratified
437
-------
towing tank. The model hill was the fourth-order polynomial (45° maximum
slope) used by Hunt et ai (1978), except that it was in this case instrumented
with 100 sampling ports located along 8 radial lines. The density gradient
was linear and the dividing-streamline height was fixed at half the hill
height Effluent was released at three elevations above the
dividing-streamline height Pairs of tows were made such that, in one tow,
the hill (upside down) was fully immersed in the water and the towing speed
was adjusted to provide a 'natural' dividing-streamline surface. In the
second tow of the pair, the model (baseplate, hill, and source, as a unit) was
raised out of the water to the point where only the top half of the hill was
immersed, thus, forcing a flat dividing-streamline surface, while all other
conditions remained identical Concentration distributions were measured on
the hill surface (and in the absence of the hill). Concentration
distributions from each pair of tows were compared to ascertain any
differences between the "natural" dividing-streamline surface and the (forced)
flat dividing-streamline surface. A comparison of surface-concentration
patterns from a typical pair of tows is shown in Figure 9, and a scatter plot
comparing concentrations on a port by port basis is shown in Rgure 10. These
results showed that the assumption of a flat dividing-streamline surface is a
reasonable assumption to make, at least with regard to predicting the
locations and values of the maximum surface concentrations and areas of
coverage on the windward side of the hill. The results are contained in an
appendix to the Fourth Milestone Report (Snyder and Lawson, 1985a) and were
presented at the Third International Symposium on Stratified Rows (Snyder and
Lawson, 1987).
The second study, providing a relatively complete set of data on flow and
diffusion around a three-dimensional hill, was conducted in the Meteorological
Wind Tunnel. The primary objective was to determine the influence of the hill
on the maximum ground-level concentration (glc) and to locate the source
positions where this influence was greatest All measurements were made with
an approach flow that simulated the neutral atmospheric boundary layer
measured at Cinder Cone Butte. However, the nearly axisymmetric CCB shape was
replaced by a truly axisymmetric hill represented by a simple mathematical
formula, and having a maximum slope of 24° (the same as CCB).
The measure of the hilPs influence on the maximum glc was the terrain
amplification factor A. This factor is defined as the ratio of the maximum
438
-------
V
\
\
Figure 9. Concentration distributions measured on the hill surface with
H0//J-0.5 and Hy/h = 0.6. Top: fully submerged; bottom: half submerged. Dotted
cirde Indicates half the hill height
439
-------
100
ec
UJ
10
u.
an
z
o
UJ
o
z
o
o
.1
.1
10
100
CONCENTRATION, HALF SUBMERGED
Figure 10. Scatter plot comparing concentrations on fully immersed hill with
those on half-immersed hill on a port by port basis. H3/rt = 0.6, H0/h = Q.5.
440
-------
glc observed in the presence of the hill to the maximum observed in the
absence of the hill. The locations of the maxima are not considered in this
evaluation; the maxima may be found at entirely different places in the
presence and in the absence of the hill
A matrix of source locations was used covering the range from 4 to 16
hill heights (h) upstream of the hill center and to 1.25/7 in the vertical. A
map at terrain amplification factors is shown in Rgure 11. The presence of
the hill was found to influence the transport and dispersion of the plume and
to increase the maximum glc in three, ways. For low sources at moderate
distances from the hill, the reduction in mean wind speed and increase in
turbulence allow the plume to reach the ground surface closer to the source,
thus producing higher concentrations than in the absence of the hill. Plumes
from higher sources may be thought of as being intercepted by the hill, that
is, the hill penetrates the plume to where the concentrations are greater than
those that would occur at ground-level farther downstream over flat terrain.
For yet higher sources, the streamline convergence over the hill top and the
corresponding downward flow and much enhanced turbulence in the lee of the
hill again bring the plume to the ground more rapidly than over flat terrain.
Terrain amplification factors ranged from near 1.0 to 3.63, and the range of
source locations that produced an amplification factor greater than 1.4
extended to an upwind distance of 14 hill heights. These results were
reported in an appendix to the Fourth Milestone Report (Thompson and Snyder,
1985b).
In the fall of 1984, ERT requested a list of data sets available from
previous complex terrain studies that had been conducted at the FMF. A report
was prepared by Thompson ef a/ (1985) listing 24 separate complex terrain
studies. Each project was synopsized with a brief description of the project,
the name of the principal investigators), the facilities used, types of data
collected, names of data reports available, major conclusions reached, listing
of published results from the project, and a listing and description of the
data files available.
An earlier request (prior to summer 1983) from the modelers at ERT had
been to provide data on streamline trajectories in neutral and stratified flow
over a three-dimensional hill, i.e., to provide data to use in developing
algorithms for predicting lateral and vertical streamline displacements over a
hill as functions of source location and stratification. Earlier work on this
441
-------
z
h
U
MAX DOWNWIND
M
1IAX UPWIND
x/h
Figure 11. Terrain amplification factors measured upwind of axisymmetric CCB
model. Heavy lines divide the region into areas where the source produced the
maximum glc upwind of the hill top, between the hill top and the separation
point, and downwind of the hill Note that the vertical scale is exaggerated
by a factor of 3.
-------
project had been set aside because of experimental difficulties and because of
the more urgent requests described immediately above. Having completed those
studies, work commenced again on the streamline trajectories.
In attempting to predict the maximum glc from a source upwind of a hill,
the most important feature of the flow is the displacement of the mean
streamline through the source, because that displacement determines how near
to the surface the 'centeriine* of the plume will reach. The exact path taken
by the plume in circumventing the hill and the plume's closeness of approach
to the hill surface are critical in determining the location and magnitude of
the gic's. These displacements are known to be strongly affected by the hill
shape and especially by the stratification in the approach flow. The purpose
of this study was thus to characterize the effects of stability on the
horizontal and vertical deflections around an isolated hill. A large set of
streamline trajectories over the axisymmetric CCB model was measured using the
stratified towing tank. Three-dimensional coordinates of the streamlines (86
independent trajectories) were determined through stereographic analysis of
photographs of dye streak lines released at a matrix of source positions
(heights and lateral offsets from the hill/flow centeriine), and at
stabilities ranging from strongly stable to neutral (Froude numbers of 0.6,
1.0, 2.0, and 09). These measurements provided a relatively complete data set
for testing mathematical models and algorithms of the detailed structure of
stratified flow over hills. The results were presented in an appendix to the
Fifth Milestone Report (Snyder et al, 1986).
As an example use of the data set, a particular mathematical model using
linear theory and a Fast Fourier Transform (FFT) technique to predict these
streamline trajectories was evaluated and described in the above appendix by
Snyder et al (1986) and, with some additional work and computations, by
Thompson and Shipman (1986). The calculated results agreed well with the
experimental results for neutral flow. In the stable flow (Fr=»2.0), however,
lateral deflections were underpredicted and vertical deflections were
overpredicted using the FFT model.
The period 1986 through present
In February 1986, ERT conducted a Complex Terrain Workshop at Research
Triangle Park, NC (Lavery et al, 1986). Each participant was beforehand
provided a diskette containing the Complex Terrain Dispersion Model (CTDM)
443
-------
code and a draft User's Guide and was asked to exercise the model to assess
its overall effectiveness and validity in whatever way he chose. The purpose
of the workshop, then, was to exchange information on the results of these
exercises and to make recommendations to the model developers concerning
further refinements of the CTDM model.
The present author exercised the CTDM by comparing is predictions with
previous laboratory measurements of flow and diffusion over hills made in the
FMF. This was accomplished in four phases. In phase 1, CTDM calculations
were compared with wind-tunnel simulations of plumes released upwind of two-
and three-dimensional hills in a neutral atmospheric boundary layer. Terrain
amplification factors were compared for a matrix of source locations upwind of
the hills. This phase was intended to test the LIFT module of CTDM, where the
stratification was neutral and the potential flow calculations of LIFT should
be most applicable. In phase 2, CTDM calculations were compared with stably
stratified towing-tank observations, where plumes were released above the
dividing-streamline height upwind of a three-dimensional hill. This phase was
intended to again test the LIFT module, but this time under strongly
stratified conditions. In phase 3, CTDM calculations were compared with
strongly stratified towing-tank observations wherein plumes were released
below the dividing-streamline height upwind of the Cinder Cone Butte model.
This phase was intended to test the WRAP module exclusively. In phase 4, CTDM
calculations were made for one selected hour of field conditions, and were
compared with results of towing-tank observations. This phase was intended to
exercise both the LIFT and WRAP modules of CTDM.
The results, made available in a detailed report that was distributed to
the workshop participants (Snyder, 1986), may be summarized as follows:
1. From the neutral flow simulations (phase 1), the hill effects (as
exemplified through computations of terrain amplification factors)
appeared to be much too small. Reasons speculated for this
discrepancy included: (a) plume trajectories were too far from the
hill surface, (b) potential flow calculations did not properly
handle the deep boundary-layer flow approaching the hill, or, more
likely (c) the plume centeriine did approach the hill surface
closely enough, but the plume did not mix to the the surface through
the hill-surface boundary layer.
2. From the stable flow simulations with releases above the
dividing-streamline height (phase 2), it appeared that the plume
trajectories were again too far from the hill surface. Vertical
deflections of streamlines appeared to be strongly overestimated and
lateral deflections appeared to be strongly underestimated. In the
444
-------
towing tank, plumes released slightly above the dividing-streamline
height spread broadly but thinly to cover the entire hill surface
above the dividing-streamline height whereas the CTDM plume was
apparently deformed only slightly - it lung together in going over
the top of the hill. An apparent shortcoming of UFT at that time
was its lack of appropriate treatment of the stratification effects
in the flow that surmounted the hill, i.e., streamline (hence,
plume) deformations under quite strongly stratified flows (Fr»1)
were treated the same as those for neutral flow (Fr=»).
3. From the stable flow simulations with releases below the
dividing-streamline height (phase 3), the WRAP module yielded rather
poor results except when the input parameters (primarily
-------
an upstream source with height such as to obtain a plume that just "grazed"
the hill top. The postulate here was that the roughness on the surface would
maximize the effects of a rapid-mixing layer near the hill surface, thus
mixing material from this elevated plume to the surface, whereas the smooth
surface would minimize the effects of this mixing layer. The results showed
that the postulate of an 'inner hill-surface boundary layer" was untenable;
extremely steep concentration gradients remained near the hill surface, even
when the hill was roughened, so that rapid mixing was not induced by the
hill-surface boundary layer.
To satisfy the second goal, a series of measurements was made of plume
characteristics in flat terrain and over a three-dimensional hill. Effluent
was released at a number of elevations, upwind distances, and positions
laterally offset from the cerrterplane determined by the wind direction and the
center of the hill. Sufficient concentration measurements were made to enable
the construction of plume cross sections at the downwind position of the hill
center and, in a few cases, at the upwind base of the hill. These data were
analyzed to provide the desired information on horizontal and vertical plume
deflections and deformations effected by the hill. One of the more dramatic
examples is shown 'in Rgure 12 In this case, the source was on the
cerrterplane at ground level, 6 hill heights upwind of the hill center (the
skirt of the hill extended to 5h). Plume cross sections measured at the
position of the center of the hill, both in the presence and in the absence of
the hill, are shown. The hill effected a 91% increase in the lateral plume
width. In this case, the maximum surface concentration (at the same downwind
distance) was decreased by a factor of 2 but, of course, the area of coverage
by large concentrations was greatly increased. Detailed data reports were
provided to ERT in March 1986, and the results were published by Snyder and
Lawson(1986).
Subsequent to the CTMD Workshop (and as a result of the rather poor
comparisons of the CTDM predictions of terrain amplification factors with
wind-tunnel data), refinements were made to CTDM. Specifically, the strain
inferred or measured over the crests of two- and three-dimensional hills in
the wind tunnel were used In the calculations, i.e., the T-factors in the
model were adjusted in accordance with wind-tunnel data Substantial
improvements in the CTDM predictions of terrain amplification factors were
obtained, as described by Strimaitis and Snyder (1986).
446
-------
H
05
HILL-
-2
-1
Y/H
Rgure 12. Plume cross sections measured in presence ( ) and in absence
( —!—) of axisymmetric CCB model at x-0 (hill center). H,/h»0, x,//7=»-6,
447
-------
3.2 Supplemental Modeling of Complex Terrain
In addition to the modeling done in direct support of the CTMDP, numerous
other complex terrain studies were conducted at the FMF, primarily in response
to envisioned needs of and direct requests from the OAQPS, the regulatory arm
of EPA. These ranged from generic studies attempting to understand the
fundamental physics of now and diffusion in neutral and stable environments
to a practical demonstration to determine the good-engineering-practJee stack
height for a specific power plant located in complex terrain. Whereas the
studies done in direct support of the CTDMP were primarily concerned with
plume impingement from upwind sources and focussed primarily on strongly
stable conditions, the supplemental studies were broader ranging, for example,
including sources on the tops and lee sides of hills,, perhaps a broader range
in stability from strongly stable to neutral, and investigation of similarity
criteria - rules to ensure that the behavior of the flow in the laboratory
simulates that in the real world, An example of the latter is the Guideline
for Fluid Modeling of Atmospheric Diffusion, prepared by Snyder (1981) in
response to a request from the OAQPS. In several cases, studies that were
initiated through the CTOM developers were subsequently enlarged upon and
expanded so as to be useful to the modeling community at large. Hence, in
many cases, studies could have been described as supplemental (this section)
or in direct support of the CTMDP (Section 3.1). The choices have been
somewhat arbitrary.
One of the important overall goals in this effort was to ascertain what
circumstances lead to the largest ground-level concentrations, i.e., are
larger glt^s expected when the plume from an upwind source impinges on a hill
or when the source is downwind of that hill such that the plume is caught in a
recirculation region and downwashed to the surface? Which are likely to lead
to larger glc's, two-dimensional or three-dimensional hills? Stable
conditions or neutral conditions? In each of these circumstances, what order
of magnitude of surface concentrations may be expected?
NeutraJ-Flow Wind-Tunnel Studies
A simple method used to intercompare effects of terrain on the maximum
gic and to determine worst-case conditions is through the terrain
amplification factor, as mentioned in Section 3.1. Again, the terrain
448
-------
amplification factor, A, is defined as the ratio of the maximum ground-level
concentration occurring in the presence of the terrain feature, A:^ to the
maximum that would occur from the same source located in flat terrain, X^,
i.e., A=*XmJ'X%rx. This definition is useful only for elevated sources, of
course, because for ground-level sources, the maximum surface concentration
occurs at the source itself.
Numerous neutral-flow wind-tunnel studfes have been conducted at the FMF
on diffusion over two-dimensional terrain features: (a) a ramp with a slope
of 14° followed by a plateau (Snyder and Pendergrass, 1980; Pendergrass and
An/a, 1983; Pendergrass and Snyder, 1987), (b) a bell-shaped hill with a
maximum slope of 12° (Courtney, 1979; Courtney and Arya, 1980), (c) a steep
triangular ridge with a slope of 63° (Arya and Shipman, 1981; Arya et al,
1981), (d) a series of smooth shaped hills of various slopes (Khurshudyan er
a/, 1981; Capuano, 1983; and Lawson and Snyder, 1985, 1987) and (e) a valley
formed between two ridges of sinusoidal cross section (Lee et al, 1981).
Three studies have been performed to determine the effects of the crosswind
aspect ratio of a triangular ridge on dispersion from nearby sources. As
mentioned in Section 2, Snyder and Britter (1987) investigated surface
concentrations on the ridges from upwind sources. (Note that the work was
done in 1979, much earlier than the publication date, so that the results were
available, indeed, used in the development of a forerunner to CTDM.) Castro
and Snyder (1982) extended the study by measuring the sizes and shapes of the
recircuiation regions downwind of these hills of various crosswind aspect
ratio, and by measuring the concentration fields resulting from sources placed
at various downwind locations. Recently, Castro and Snyder (1987a) have
further extended this work to include the case when the approaching wind is
not perpendicular to the long axis of the hill. This allows one to use the
wind-tunnel data to estimate the effects of long-time-scale wind meander.
Other generic three-dimensional hill studies included: (a) conical hills with
slopes of 26.5° and 17.5° with sources located at the hill top or at the
downwind base (Gadfyaram, 1984; Arya and Gadiyaram, 1986) and (b) the
axisymmetric CCB model with downwind sources (Lawson and Snyder, 1985,1987).
These various studies were summarized through publications at various stages
by Thompson and Snyder (1981, proceedings published 1985a) and Snyder (1983a,
1983b, 1984). Only a broad overview and a few typical results will be
presented here.
449
-------
Table .1 shows the terrain amplification factors for the cases listed
above, in order of decreasing -4. From the standpoint of a fixed stack height,
the worst location for a source appears to be just downwind of a
two-dimensional ridge. Downwind sources generally result in larger glc's
because of the excess turbulence generated by the hills and because the
effluent is generally emitted into a low speed region where the streamlines
are descending toward the surface. Maximum As are considerably larger than
those downwind of three-dimensional hills. A probable cause of this effect is
that, in three-dimensional flows, lateral and vertical turbulence intensities
are enhanced by roughly equal factors, whereas in two-dimensional flows, the
lateral turbulence intensities are not enhanced as much as are the vertical
turbulence intensities (because of the two-dimensionality). Since the maximum
glc depends upon the ration
-------
They do not provide practical estimates for use by, say, an air pollution
meteorologist in determining the maximum glc resulting from a particular power
plant or for determining the best location for that plant For that purpose,
the concept of a •window* of excess concentrations, as Introduced by Hunt et
al (1979) is more useful For any given plant location, (say, upwind of the
hill), there is a limited range of stack heights H3 for which a significant
amplification of the glc will occur. (For sake of argument, we will here
define significant as a factor of 2.) This amplification can occur only if
the position of the maximum glc lies on or near the hill surface. For small
Hv *mx wTO occur upwind of the hill and thus be little influenced by the
hill, so that A (^^^ will approach unity. If H, is too large (for
example, H^h, the hill height), Xm will lie well beyond the hill and A will
again approach unity. In either case, there is little amplification. These
•windows' of critical H, values have been measured by Lawson and Snyder (1985,
1987) for two typical hill shapes that might be found in the real world, one
axisymmetric, the other two-dimensional. The results are shown in Figure 13.
The 1.4-window, for example, extends to about 14/7 upstream, 10/7 downstream,
and as high as 1.8/7 in the vertical for the axisymmetric hill. For the
two-dimensional hill, this 1.4-window extends about 8/7 upstream, 15/r
downstream, and as high as 2.2/7 in the vertical.
Such contour maps as provided in Figure 13 can be very useful for the
practitioner. Once an acceptable terrain amplification factor (or 'excess
concentration11) is decided upon, it is a simple matter to trace the window on
the contour map to determine the area (plant location and/or stack height) to
be avoided. Conversely, from such maps, the likely maximum glc for a
potential site and stack height can be estimated. The use of terrain
amplification factors simplifies the application of these data to full-scale
situations. The expected maximum glc in flat terrain is calculated (from
mathematical models or standard curves), then the concentration in the
presence of the hill is simply the product of this quantity and the TAF. This
study was initiated through a request from the EPA Office of Air Quality
Planning and Standards (OAQPS) to aid in the decision-making process with
regard to the promulgation of the Stack Height Regulations under the Clean Air
Act, and the data were provided to OAQPS much earlier than the publication
dates shown.
451
-------
-15 -10
Ln
N>
x/h
Figure 13. Contours of constant terrain amplification (actors over (a)
axisymmatrlc hill and (b) two-dimensional ridge. Note that vertical scale is
exaggerated by a factor of 3.
-------
Subsequent to the idealized study described above, the OAQPS requested
the FMF to conduct a study demonstrating the application of the fluid modeling
approach to the determination of good-engjieering-practice (GEP) stack height
for a power plant in complex terrain, i.e., to provide an example study/report
for industry to follow in the conduct of a GEP determination. The site chosen
for this demonstration was the Clinch Rwer Power Plant in southwestern
Virginia, and a 1:1920 scale model of the surrounding terrain was constructed.
Measurements were presented (Snyder and Lawson, 1985b) that described the
simulated atmospheric boundary layer structure, plume-dispersion
characteristics in that boundary layer, and the maximum glc of effluent
downstream from the plant, both in ttie presence of all significant terrain
surrounding the plant and in the absence of 'nearby upwind terrain. Analysis
of the maximum glc showed that, in this case, a stack height of 326m met the
GEP criteria under 50% load conditions, i.e., the nearby upwind terrain
effected an increase of 40% in the maximum ground-level concentration. This
study followed the general guidance set forth in the Guideline for Fluid
Modeling of Atmospheric Diffusion (Snyder, 1981) and the specific
recommendations set forth in the Guideline for Use of Fluid Modeling to
Determine Good Engineering Practice Stack Height (EPA, 1981) and the Guideline
for Determination of Good Engineering Practice Stack Height (Technical Support
Document for the Stack Height Regulations, Revised Draft) (EPA, 1985).
Staoly Stratified Towing-Tank Studies
Lamb and Britter (1984) conducted a combined numerical and laboratory
study of so-named shallow water flow over an isolated hill They showed how
certain geometrical and flow parameters affect the tendency of a fluid to flow
around rather than over an obstacle in the case of a homogeneous single layer
fluid, i.e., simulating the atmospheric condition of an elevated step
Inversion. A series of numerical experiments was conducted using a
finite-difference model. Measures were suggested for quantitative assessment
of the tendency of the fluid to flow around the obstacle as a function of the
relative hffl height and the Froude number. The laboratory experiments
examined the motions of two superposed homogeneous layers of fluid past a
conical hill in the towing tank. The resulting motions were found to agree
with the results of the numerical experiments and extended the understanding
gained from them. Row visualization techniques were used to demonstrate the
453
-------
impingement of the interface on the obstacle, and its dependence on flow speed
and hill height
Another numerical model was acquired and implemented at the FMF for
comparison with laboratory results. The numerical model was originally
developed by Mason and Sykes (1979); it integrates the Navier-Stokes equations
for incompressible stratified flow using a finite-difference schema Direct
comparisons were made between the results of this mode! and laboratory
experiments for density-stratified flow around the idealized axisymmetric CCB
model by Rottman et al (1987) for three specific experimental arrangements.
First, a small towing tank was used in which both the Reynolds number and
Froude number were matched exactly with the numerical model. This provided an
overall assessment of the accuracy of the approximations made in the numerical
model. Second, the large towing tank was used in which mean plume
trajectories were measured and compared with particle paths computed through
the numerical model. Third, some comparisons were made with wind-tunnel
measurements of the flow structure over the hill. In general, the numerical
model qualitatively reproduced the experimental results on the flow structure,
but there were some substantial differences, particularly near the hill
surface and in the wake and at the larger values of the Reynolds number.
Whereas the following area of investigation is not directly related to
complex terrain, it is included here because it played an important (and
somewhat controversial) role In the CTDM formulation. This is the area of
describing the effects of stable stratification on turbulent diffusion or, put
another way, estimating vertical plume growth in the nighttime stable boundary
layer. Experiments were conducted (Britter ef al, 1983) in which a grid was
towed horizontally along the stratified towing tank. The vertical velocity
fluctuations produced near the grid were reduced under strong stratification
by up to 30%, but the decay rates of the turbulent velocity fluctuations were
found to be unaffected by the stratification over a considerable distance
downstream. Turbulent diffusion from a point source located downstream of the
grid was also measured. The lateral plume widths were found to be largely
unaffected by the stratification and grew with the 1/2-power of time. The
vertical plume growth, however, was found to reach an asymptotic limit These
results were largely in agreement with the theoretical models of Csanady
(1964) and Pearson er al (1983), but in contradiction to the theory and
limited data of Venkatram et al (1984). The latter data suggest a continuous
-------
vertical plume growth (for large times), but the measurements did not, in
fact, extend very far downwind (maximum downwind distance of about 1 km).
Further grid-turbulence studies were done in the towing tank with the aim
of investigating internal wave effects and providing guidance on the
partitioning of wave and turbulence energies in stably stratified flows
(Rottman and Britter, 1986). The results suggested that the mixing efficiency
increases monotonically with increasing stability, with some indication that
it approaches a constant as the flow becomes strongly stable.
A cooperative project was completed with the Los Alamos National
Laboratory to examine the conditions under which flushing of a valley between
two ridges will occur, i.e., to answer the question of when a stable crosswind
win sweep the valley dean and when the flow will separate from the top lee
side of the first ridge, reattach at the top windward side of the second
ridge, and thus form a nearly stagnant region in the valley beneath. In this
series of towing-tank studies, three experimental parameters were varied: the
steepness of the ridge/valley slopes (40°, .27° and 13°), the separation
distance between the ridges, and the Froude number that characterizes the
stability of the crosswind. In broad terms, the characteristics of the flow
between the ridges may be explained using criteria for boundary-layer
separation from the lee side of a single ridge, the downstream ridge appears
to induce separation from the lee side of the upstream ridge only when it is
steep-sided (Lee et a/, 1984a,b, 1986, 1987). As an offshoot of this work,
the conditions conducive to the onset of severe downslope winds on the lee
sides of mountains was investigated (Rottman and Smith, 1987). The results
snowed that an intrusion (breaking wave - associated with severe downslope
winds) existed when the Froude number based on the ridge height was in the
range 0.2 s F s 0.6 for a steep-sloped ridge (maximum slope 40°) and 0.2 s F s
1.1 for a low-sloped ridge (13°).
An overview of fluid modeling of pollutant transport and diffusion in
stably stratified flows over complex terrain was provided for Annual Review of
Fluid Mechanics by Snyder (1985).
455
-------
4. SUMMARY
The EPA Fluid Modeling Facility has conducted a wide range of laboratory
studies and a limited amount of numerical modeling of flow and diffusion in
association with the Complex Terrain Model Development Program. The goal of
the CTMDP is the development of a dispersion model valid in complex terrain,
with emphasis on plume impaction on nearby hills during nighttime stable
conditions. Work at the FMF prior to the inception of the program provided
the basic framework for the model - the dividing-streamline concept - and the
focal point around which to design the field program.
Throughout the course of the CTMDP, the FMF interacted vigorously with
the model developers by providing support in various ways. Early work
provided direct support in planning the details and strategies of the field
experiments and solidifying and testing the limits of applicability of the
dividing-streamline concept Later, work included exercises of filling in the
gaps' in the field data, furthering the understanding of the physical
mechanisms important to plume impaction in complex terrain and in stably
stratified flows in general, and testing the ability of the laboratory models
to simulate full-scale field conditions. And, as the needs arose, the FMF
tested various modeling assumptions, concepts, and hypotheses and provided
data for 'calibration' of various parameters within the CTDM model.
Simultaneously, the FMF responded to the needs of the regulatory arm of
EPA, the Office of Air Quality Planning and Standards, by providing guidance
concerning expected terrain effects and by providing a demonstration study -
an example for industries to follow in conducting good-engineering-practice
stack height determinations in complex terrain. Also, a broad range of
supplemental studies was conducted, expanding and enlarging upon the specific
requests of the OAQPS and the CTDM model developers to provide information of
general use to the scientific and air pollution modeling communities. Many of
the data sets generated in the course of this program have been provided to
and used by various groups (nationally and internationally) in the
development, testing and evaluation of complex terrain dispersion models.
456
-------
REFERENCES
* Arya, S.P.S. & Gadlyaram, P.S. 1986 An Experimental Study of Row and
Dispersion in the Wakes of Three-Dimensional Low Hills. Atmos. Envir.,
20,729-40.
* An/a, S.P.S. & Shipman, M.S. 1981 An Experimental Investigation of Row and
Diffusion in the Disturbed Boundary Layer over a Ridge, Part I: Mean
Row and Turbulence Structure. Atmos. Envir., 15, 1173-84.
* Arya, S.P.S., Shipman, M.S. & Courtney, LY. 1981 An Experimental
Investigation of Row and Diffusion in the Disturbed Boundary Layer over
a Ridge, Part II: Diffusion from a Continuous Point Source. Atmos.
finv/X, 15, 1185-94.
Baines, P.O. 1979 Observations of Stratified Row Past Three-Dimensional
Barriers. J. Geophys, Res., 84, no. C12, 7834-8.
* Bass, A. 1980 Towing Tank Studies in Support of Reid Experiments at Cinder
. Cone Butte, Idaho, Part II: Plume Behavior with Froude Number and
Incident Wind Direction. Rpt by Envir. Res. & Tech. on cooperative
work with Ruid Mod. Facility, Envir. Prat Agcy., Res. Tri. Pk., NC.
Bass, A., Strimaitis, D.G. & Egan, B.A. 1981 Potential Row Model for
Gaussian Plume Interaction with Simple Terrain Features. Rpt under
Contract No. 68-02-2759, Envir. Prat Agcy., Res. Tri. Pk., NC, 201 p.
Brighton, P.W.M. 1978 Strongly Stratified Row Past Three-Dimensional
Obstacles. Quart. J. Roy. Meteorol. Soc., 104, 289-307.
* Britter, RE, Hunt, J.C.R., Marsh, G.L & Snyder, W.H. 1983 The Effects of
Stable Stratification on Turbulent Diffusion and the Decay of Grid
Turbulence. J. Fluid Mech., 127, 27-44.
But, EW. & Slater, H.H. 1977 Evaluation of the Valley Model. AMS-APCA
Joint Conf. on AppL of Air PolL Meteorol, Salt Lake City, UT, Amer.
Meteorol Soc., Boston, MA.
* Capuano, M.E 1983 The Effects of Hill Slope on Row and Dispersion over
Two-dimensional Hills - A Wind Tunnel Study. M.S. Thesis, Dept
Marine, Earth, Atmos. ScL, NC State Univ., Raleigh, NC, 153p.
* Castro, I.P . 1987 A Note on Lee Wave Structures in Stratified Row over
Three-Dimensional Obstacles. Tellus, 39A, 72-81.
* Castro, I.P. & Snyder, W.H. 1982 A Wind Tunnel Study of Dispersion from
Sources Downwind of Three-Dimensional Hills. Atmos. Envir., 16,
1869-87.
* Castro, I.P. & Snyder, W.H. 1987a Wind Direction Effects on Dispersion from
Sources Downwind of Steep Hills. Atmos. Envir. (to be submitted).
* Publications generated from research conducted within the Fluid Modeling
Facility.
457
-------
* Castro, I.P. & Snyder, W.H. 1987b Obstacle Drag and Upstream Motions in
Stratified Row. Proc. Third Int. Symp. Stratified Flows, Cal. Inst
Tech., Pasadena, CA, Feb. 3-5 (general session).
* Castro, I.P. & Snyder, W.H. 1987c Upstream Motions in Stratified Row. J.
Fluid Mech. (submitted).
* Castro, IP., Snyder, W.H. & Marsh, G.L 1983 Stratified Flow over Three-
Dimensional Ridges. J. Fluid Mech., 13S, 261-82.
* Courtney, LY. 1979 A Wind Tunnel Study of Row and Diffusions over a Two-
Dimensional Low Hill. M.S. Thesis, Dept of Meteorol., NC State Univ.,
Raleigh, NC, 134p.
* Courtney, LY. & An/a, S.P.S. 1980 Boundary Layer Row and Diffusion over a
Two-dimensional Low Hill. Preprints Vol., 2nd Jt Conf. Appl. Air Poll.
Meteorol., Mar. 24-28, New Orleans, LA, 551-8. Amer. Meteorol. Soc.,
Boston, MA.
Csanady, G.T. 1964 Turbulent Diffusion in a Stratified Ruid. Atmos. Sci.,
21, 439-47.
DiCristofaro, D.C., Strimaitis, D.G., Greene, B.R., Yamartino, R.J.,
Venkatram A., Godden, DA, Lavery, T.F. & Egan, B.A. 1986 EPA Complex
Terrain Model Development Fifth Milestone Report - 1985. Rpt. No.
EPA/600/3-85/069, Envir. Prot Agcy., Res. Tri. Pk., NC, 277p.
Drazin, P.O. 1961 On the Steady Row of a Ruid of Variable Density Past an
Obstacle. Tellus, 13, 239-51.
EPA 1981 Guideline for Use of Ruid Modeling to Determine Good Engineering
Practice Stack Height Rpt No. EPA-450/4-81-003, Envir. Prot. Agcy.,
Res, Tri. Pk., NC, 47p.
EPA 1985 Guideline for Determination of Good Engineering Practice Stack
Height (Technical Support Document for the Stack Height Regulations).
Rpt No. EPA-450/4-80-023R (Revised June 1985), Envir. Prot Agcy.,
Res. Tri. Pk., NC, 102p.
* Eskridge, R.E, Lawson, R.E Jr. & Marsh, G.L 1983 Simulation of an
Atmospheric Tracer Experiment in Complex Terrain Using a Stratified
Towing Tank: A Case Study. 6th Symp. Turb. & Diffusion, Boston, MA,
Mar. 22-25, Amer. Meteorol. Soc., Boston, MA.
* Gadiyaram, P.S. 1984 Row and Dispersion over Three-Oimensional
Axlsymmetric Hills: A Wind Tunnel Study. M.S. Thesis, Dept Marine,
Earth, Atmos. Sci, NC State Univ., Raleigh, NC, 126p.
Hoteworth, G.C. 1980 The EPA Program for Dispersion Model Development for
Sources in Complex Terrain. 2nd Jt Conf. Appl. Air Poll. Meteorol.,
March 24-27, New Orleans, LA, Amer. Meteorol. Soc., Boston, MA.
* Holzworth, G.C. & Snyder, W.H. 1979 Program Plan for Development of a
Mathematical Air Quality Assessment System for Use in Complex Terrain.
Rpt No. EPA-600/9-79-041, Workshop on Atmos. Disp. Models in Complex
Terrain, 137-50. Envir. Prot Agcy., Res. Tri. Pk., NC.
Hovind. E.L, Edelstein, M.W. & Sutherland, V.C. 1979 Workshop on
Atmospheric Dispersion Models in Complex Terrain. Rpt No.
EPA-600/9-79-041, Envir. Prot Agcy., Res. Tri. Pk., NC, 213p.
458
-------
Hunt, J.C.R. & Mulheam, P.J. 1973 Turbulent Dispersion from Sources Near
Two-Dimensional Obstacles. J. Fluid Mech., 61, 245-74.
* Hunt, J.C.R, Puttock, J.S. & Snyder, W.H. 1979 Turbulent Diffusion from a
Point Source In Stratified and Neutral Rows around a Three-Oimensional
Hill: Part I: Diffusion Equation Analysis. Atmos. Envir., 13, 1227-39.
* Hunt, J.C.R & Snyder, W.H. 1980 Experiments on Stably and Neutrally
Stratified Flow over a Model Three-Oimensional Hill. J. Fluid Mech.,
96, 671- 704.
* Hunt. J.C.R, Snyder, W.H. & Lawson, RE Jr. 1978 Row Structure and
Turbulent Diffusion around a Three-Dimensional Hill: Ruid Modeling
Study on Effects of Stratification; Part I: Row Structure. Rpt No.
EPA-600/4-78-041, Envir. Prot Agcy., Res. Tri. Pk., NC..
* Khursnudyan, LH., Snyder, W.H. & Nekrasov, I.V. 1981 Row and Dispersion
of Pollutants over Two-Dimensional Hills: Summary Report on Joint
Soviet-American Study. Rpt No. EPA-600/4-81-067, Envir. Prot. Agcy.,
Res. Tri Pk., NC, 143p.
* Lamb, V.R & Britter, RE 1984 Shallow Water Row over an Isolated
Obstacle. J. Fluid Mech., 147, 291-313.
Lavery, T.F., Bass, A., Strimaitis, D.G., Venkatram, A., Greene, B.R.,
Drtvas, P.J. & Egan, BA 1982 EPA Complex Terrain Modeling Program:
First Milestone Report - 1981. Rpt No. EPA-600/3-82-036, Envir. Prot
Agcy., Res. Tri. Pk., NC, 304p.
Lavery, T.F., Strimaitis, D.G. & Egan, BA 1986 A Workshop Report on the
Complex Terrain Model Development Project (February 4-6, 1986). Rpt
under Contract 68-02-3421, Envir. Prot Agcy., Res. Tri. Pk., NC, 7Sp.
Lavery, T.F., Strimaitis, D.G., Venkatram, A., Greene, B.R., DiCristofaro,
D.C. and Egan, BA 1983 EPA Complex Terrain Model Development: Third
Milestone Report - 1983. Rpt No. EPA-600/3-83-101, Envir. Prot
Agcy., Res, Tri. Pk., NC, 271 p.
* Lawson, R.E Jr. & Snyder, W.H. 1985 Stack Heights and Locations in Complex
Terrain. Preprints Vofc 7th Symp. Turb. Diff., Nov. 12-15, Boulder, CO,
223-6. Amer. MeteoroL Soc., Boston, MA.
* Lawson, RE Jr. & Snyder, W.H. 1987 Estimation of Pollutant Concentration
from Sources Near Complex Terrain in Neutral Row. Atmos. Emir, (to be
submitted).
* Lee. J.T., Barr, S., Lawson, RE, Jr., Snyder, W.H. & Marsh, G.L 1984a
Towing Tank Studies of Stratified Row over Ridges and Valleys. Rpt
No. LA-UR-84-1314, Los Alamos National Laboratory, Los Alamos, NM, 29p.
* Lee. J.T.. Barr. S.. Lawson, RE, Jr., Snyder, W.H. & Marsh, G.L 1984b
Towing Tank Studies of Stratified Row over Ridges and Valleys.
Preprints VoL 3rd Conf. Mtn. MeteoroL, Portland, OR, 37-41. Amer.
MeteoroL Soc., Boston, MA
* Lee, J.T., Barr, S., Snyder, W.H. & Lawson, R.E Jr. 1981 Wind Tunnel
Studies of Row Channeling in Valleys. Preprint Vol. 2nd Conf. Mtn.
MeteoroL, Nov. 9-12, Steamboat Springs. CO, Amer. MeteoroL Soc.,
Boston, MA.
-------
* Lee, J.T., Lawson, R.E Jr. & Marsh, G.L 1986 Flow Visualization
Experiments on Stably Stratified Row over Ridges and Valleys. Proc.
3rd Int. Workshop on Wind and Water Tunnel Modeling of Atmos. Row and
Dispersion, Sept, Lausanne, Switzerland.
* Lee, J.T., Lawson, R.E Jr. & Marsh, G.L 1987 Row Visualization
Experiments on Stably Stratified Row Over Ridges and Valleys: Final
Report Rpt No. LA-UR-87-127, Los Alamos National Laboratory, Los
Alamos, NM.
Mason, P.J. & Sykes, R.I. 1979 Three-Oimensional Numerical Integrations of
the Navier-Stokes Equations for Row Over Surface-Mounted Obstacles. J.
Fluid Mech., 91, 433-50.
Pasquill, F. 1974 Atmospheric Diffusion. 2nd Ed., Chichester, Ellis
Horwood Ltd., John Wiley & Sons, NY, NY, 429p.
Pearson, H.J., Puttock, J.S. & Hunt, J.C.R. 1983 A Statistical Model of
Ruid-Element Motions and Vertical Diffusion in a Homogeneous Stratified
Turbulent Row. J. Fluid Mech., 129, 219-49.
* Pendergrass, W.R & Arya, S.P.S. 1983 Vortex Development in Boundary Layer
Rows over Two-Dimensional Ramps. Preprint Vo!. 6th Symp. Turb. &
Diff., Mar. 22-25, Boston, MA, Amer. Meteorol. Soc., Boston, MA.
* Pendergrass, W.R. & Snyder, W.H. 1987 Wind Tunnel Measurements of Terrain
Amplification Factors for Sources Upwind of Two-Oimensional Ramps of
Various Slopes. Atmos. Envir. (to be submitted).
Queney, P., Corby, G.A., Gerbier, N., Koschmieder, H. & Zlerep, J. 1960 The
Airflow over Mountains. World Meteorol. Org., Tech. Note No. 34.
Geneva, Swftz.
Riley, J.J., Liu, H.T. & Geller, EW. 1976 A Numerical and Experimental
Study of Stably Stratified Row Around Complex Terrain. Rpt No.
EPA-600/4-76-021, Envir. Prot Agcy., Res. Tri. Pk., NC, 41 p.
* Rottman, J.W. & Britter, R.E. 1986 The Mixing Efficiency and Decay of Grid-
Generated Turbulence in Stably Stratified Ruids. Proc. 9th
Australasian Fluid Mech. Conf., Dec. 8-12, Univ. Auckland, Auckland, New
Zealand
* Rottman, J.W., Lawson, RE Jr. & Snyder, W.H. 1987 A Comparison of
Numerical and Laboratory Experiments on Density-Stratified Rows around
a Three-Dimensional Hill Proc. Third Int. Symp. Stratified Flows, Cal.
Inst Tech., Pasadena, CA, Feb. 3-5.
* Rottman, J.W. & Smith, R.B. 1987 Tow-Tank Simulations of the Severe
Downslope Wind. Proc. Third Int. Symp. on Stratified Flows, Cal. Inst
Tech., Pasadena, CA, Feb. 3-5.
Rowe, R.D., Benjamin, S.F., Chung, K.P., Havlena, JJ. & Lee, C.Z 1982
Reid Studies of Stable Air Row over and around a Ridge. Atmos.
Envir., 16, 643- 53.
Sheppard, PA 1956 Airflow over Mountains. Quart. J. Roy. Meteorol. Soc.,
82, 528-9.
460
-------
* Snyder, W.H. 1980a Towing Tank Studies in Support of Reid Experiments at
Cinder Cone Butte, Idaho, Phase I: Influence of Hill on Wind Reid at
the Meteorological Tower Site. Ruid Modeling Facility Internal Rpt,
July 30, 20p. Envir. Prot Agcy., Res. Tri. Pk., NC.
* Snyder, W.H. 1980b Towing Tank Studies in Support of Reid Experiments at
Cinder Cone Butte, Idaho, Phase III: Verification of Formula for
Prediction of Dividing Streamline Height Ruid Modeling Facility
Internal Rpt, Aug. 29, Envir. Prot Agcy., Res. Tri. Pk., NC, 12p.
* Snyder, W.H. 1981 Guideline for Ruid Modeling of Atmospheric Diffusion.
Rpt No. EPA-600/8-81-009, Envir. Prot Agcy., Res. Tri. Pk., NC, 200p.
* Snyder, W.H. 1983a Ruid Modeling of Terrain Aerodynamics and Plume
Dispersion - A Perspective View. Invited Presentation, AMS Workshop on
Dispersion in Complex Terrain, Keystone, CO, May 17-20.
* Snyder, W.H. 1983b Ruid Modeling of Terrain Aerodynamics and Plume
Dispersion - A Perspective View. Preprint Vol. 6th Symp. Turb. & Diff.,
March 22-25, Boston, MA, 317-20. Amer. Meteorol. Soc., Boston, MA.
* Snyder, W.H. 1984 Terrain Aerodynamics and Plume Dispersion: A Perspective
View Gained from Ruid Modeling Studies. Proc. Symp. Tibetan Plateau &
Mtn. Meteorol., Beijing, P.R.C., March.
* Snyder, W.H. 1985 Ruid Modeling of Pollutant Transport and Diffusion in
Stably Stratified Rows over Complex Terrain. Ann. Rev. Fluid Mech.,
17,239-66.
* Snyder, W.H. 1986 Comparisons of CTDM Calculations with Ruid Modeling
Observations. Complex Terrain Workshop, Research Triangle Park, NC,
Feb. 4-6, 45p.
* Snyder. W.H. & Britter, R.E 1987 A Wind Tunnel Study of the Row Structure
and Dispersion from Sources Upwind of Three-Dimensional Hills. Atmos.
Envir., 21, 735.
* Snyder, W.H., Britter, R.E & Hunt J.C.R. 1980 A Ruid Modeling Study of
the Row Structure and Plume Impingement on a Three-Dimensional Hill in
Stably Stratified Row. Proc. Rfth Int Corrf. on Wind Engr. (J.E.
Cermak, ed), 1, 319-29. Pergamon Press, NY, NY.
* Snyder, W.H. & Hurt, J.C.R. 1984 Turbulent Diffusion from a Point Source in
Stratified and Neutral Rows around a Three-Dimensional Hill, Part II:
Laboratory Measurements of Surface Concentrations. Atmos. Envir., 18,
1969-2002.
* Snyder, W.H. & Lawson, R.E Jr. 1981 Laboratory Simulation of Stable Plume
Dispersion over Cinder Cone Butte: Comparison with Reid Data
Appendix: EPA Complex Terrain Model Development Rrst Milestone Report -
1982, Rpt No. EPA-600/3-82-036, p. 250-304. Envir. Prot Agcy., Res.
Tri. Pk., NC.
* Snyder, W.H. & Lawson, R.E Jr. 1985a Stable Plume Dispersion over an
Isolated Hill: Releases above the Dividing-Streamline Height Appendix
A: EPA Complex Terrain Model Development Fourth Milestone Report -1984,
Rpt No. EPA/600/3-84/110, 233-68. Envir. Prot. Agcy., Res. Tri. Pk, NC.
461
-------
* Snyder, W.H. & Lawson, R.E Jr. 1985b Ruid Modeling Demonstration of Good-
Engineering-Practice Stack Height in Complex Terrain. Rpt. No.
EPA-600/3-85/022, Envir. Prat. Agcy., Res. Tri. Pk., NC, 89p.
* Snyder, W.H. & Lawson, R.E Jr. 1986 Laboratory Observations of Plume
Deformations in Neutral Flow over a Three-Dimensional Hill. Preprint
VoL AMS 5th Jt. Conf. Appl. Air Poll. Meteorol. with APCA, Nov., Chapel
Hill, NC, Amer. Meteorol. Soc., Boston, MA,
* Snyder, W.H. & Lawson, R.E Jr. 1987 Stable Plume Dispersion over an
Isolated Hill: Releases above the Dividing-Streamline Height Proc.
Third Int. Symp. Stratified Flows, Cai. Inst. Tech., Pasadena, CA, Feb.
3-5.
* Snyder, W.H., Lawson, R.E Jr., Thompson, R.S. & Holzworth, G.C. 1980
Observations of Ftow around Cinder Cone Butte, Idaho. Rpt. No.
EPA-600/7-80-150, Envir. Prot Agcy., Res. Tri. Pk., NC, 30p.
• Snyder, W.H. & Ogawa, Y. 1982 Simulation of Row and Diffusion over Cinder
Cone Butte In a Stratified Wind Tunnel. Data Report National Inst.
for Envir. Studies, Tsukuba, Japan.
* Snyder, W.H. & Pendergrass, W.R. Ill 1980 Ramp Study: Idealized Widows
Creek. Unpublished Data Rpt., Ruid Modeling Facility, Envir. Prot.
Agcy., Res. Tri. Pk., NC.
* Snyder, W.H., Thompson, R.S., Eskridge, R.E., Lawson, R.E., Jr., Castro,
I.P., Lee, J.T., Hunt J.C.R. & Ogawa, Y. 1983 The Structure of
Strongly Stratified Row over Hills: Dividing-Streamline Concept.
Appendix: EPA Complex Terrain Model Development Second Milestone Report
- 1982, Rpt No. EPA-600/3-83-Q15, p. 319-75. Envir. Prot Agcy., Res.
Tri. Pk., NC.
* Snyder, W.H., Thompson, R.S., Eskridge, R.E, Lawson, R.E, Jr., Castro,
I.P., Lee, J.T., Hurt, J.C.R & Ogawa, Y. 1985 The Structure of
Strongly Stratified Row over Hills: Dividing-Streamline Concept. J.
Fluid Mech., 152, 249-88.
* Snyder, W.H., Thompson, R.S. & Shipman, M.S. 1986 Streamline Trajectories
in Neutral and Stratified Row over a Three-Dimensional Hill. Appendix:
Rpt No. EPA/600/3-85/069, EPA Complex Terrain Model Development Fifth
Milestone Report - 1985, 240-277. Envir. Prot Agcy., Res. Tri. Pk.,
NC.
Strimaitis, D.G., Lavery, T.F., Venkatram, A., DiCristofaro, D.C., Greene,
8.R. & Egan, BA 1985 EPA Complex Terrain Model Development: Fourth
Milestone Report - 1984. Rpt No. EPA-600/3-84-110, Envir. Prot
Agcy., Res. Tri. Pk., NC.
Strimaitis, D.G., Scire, J.S. & Bass, A. 1982 COMPLEX/PFM Air Quality Model
User's Guide. Rpt (awaiting printing), Envir. Prot Agcy., Res. Tri.
Pk., NC, 114p.
* Strimaitis, D.G. & Snyder, W.H. 1986 An Evaluation of the Complex Terrain
Dispersion Model Against Laboratory Observations: Neutral Row over 2-0
and 3-0 Hills. Preprint Vol. AMS 5th Jt. Conf. Appl. Air Poll.
Meteorol. with APCA, Nov., Chapel Hill, NC, Amer. Meteorol. Soc.,
Boston, MA.
462
-------
Slrimaitis, D.G., Venkatram, A., Greene, B.R., Hanna, S., Heisler, S.,
Lavery, T.F., Bass, A., & Egan, B.A. 1983 EPA Complex Terrain Model
Development: Second Milestone Report -1982. Rpt No. EPA-600/3-83-015,
Envir. Prot Agcy., Res. Tri. Pk., NC. .
* Thompson, R.S. & Shipman, M.S. 1986 Streamlines in Stratified Row over a
Three-Oimensional Hill. Preprint Vol. AMS 5th Jt Conf. Appl. Air Poll.
Meteorol. with APCA, Nov., Chapel Hill, NC, Amer. Meteorol. Soc.,
Boston, MA.
* Thompson, R.S., Shipman, M.S. & Snyder, W.H. 1985 Synopses of FMF Projects
In Complex Terrain. Rpt to ERT on CTMD Program, Envir. Prot Agcy.,
Res. Tri Pk., NC, 24p.
* Thompson, R.S. & Snyder, W.H. 1976 EPA Fluid Modeling Facility. Proc.
Conf. on Modeling & Simulation, Rpt No. EPA-600/9-76-016, Envir. Prot.
Agcy., Wash. D.C., July.
* Thompson, R.S. & Snyder, W.H. 1981^ Air Pollution and Terrain Aerodynamics:
A Review of Ruid Modeling Studies at the EPA Ruid Modeling Facility.
ASCE Fall Conv., St Louis, MO, Oct
* Thompson, R.S. & Snyder, W.H. 1984 Ruid Modeling of Blocking and Upstream
Influences of Stable Row over Two-Oimensionai Hills. Proc. 2nd
Workshop Wind/Water Tunnel Dispersion Modeling, Oxford, England, Sept.
26-28, C3.1-3.7.
* Thompson, R.S. & Snyder, W.H. 1985a Air Pollution and Terrain Aerodynamics:
A Review of Ruid Modeling Studies at the EPA Ruid Modeling Facility.
J. Wind Engr. & Indus. Aerodyn., 21, 1-19.
* Thompson, R.S. & Snyder, W.H. 1985b Dispersion from a Source Upwind of a
Three-Oimensional Hill of Moderate Slope. Appendix B: EPA Complex
Terrain Model Development Fourth Milestone Report - 1984, Rpt No.
EPA/600/3-84/110, 269-86. Envir. Prot Agcy., Res. Tri. Pk., NC.
* Thompson, R.S., Snyder, W.H. & Lawson, RE Jr. 1983 Laboratory Simulation
of Neutral Plume Dispersion over Cinder Cone Butte: Comparison with
Reid Data Appendix: EPA Complex Terrain Model Development Third
Milestone Report - 1983, Rpt No. EPA-600/3-83-101, p. 212-51. Envir.
Prot Agcy., Res. Tri. Pk., NC,
Turner, J.S. 1973 Buoyancy Effects in Fluids. Cambridge Univ. Press,
Cambridge, England, 368p.
Venkatram, A., Strimaitis, D. & DiCristofaro, D. 1984 A Semiempirical Model
to Estimate Vertical Dispersion of Elevated Releases in the Stable
Boundary Layer. Atmos. Envir., 18, 923-8.
Wackter, D.J. & Londergan, RJ. 1984 Evaluation of Complex Terrain Air
Quality Models. Rpt under Contract No. 68-02-3514, Envir. Prot
Agcy., Res. Tri. Pk., NC, 233p.
Weil, J.C., Traugott, S.C. & Wong, O.K. 1981 Stack Plume Interaction and
Row Characteristics for a Notched Ridge. Rpt No. PPRP-61, Martin
Marietta Corp., Baltimore, MD, 92p.
463
------- |