United States Office of A ir Quality ' EPA-450/3-78-110a
Environmental Protection Planning and Standards September 1978
Agency Research Triangle Park NC 27711
Air ~~
The Development of
Mathematical Models for
the Prediction of
Anthropogenic Visibility
Impairment
Volume I
-------
EPA-450/3/78-110a
Volume I
THE DEVELOPMENT OF MATHEMATICAL MODELS
FOR THE PREDICTION OF ANTHROPOGENIC
VISIBILITY IMPAIRMENT
by
Douglas A. Latimer, Robert W. Bergstrom, Stanley R. Hayes
Mei-Kao Liu, John H. Seinfeld, Gary Z. Whitten
Michael A. Wojcik, Martin J. Hillyer
Systems Application, Incorporated
San Rafael, California 94903
Contract 68-01-3947
EPA Project Officers: John Butler, David Shaver, James Dicke
Prepared for
U. S. ENVIRONMENTAL PROTECTION AGENCY
Office of Planning and Evaluation Office of Air Noise and Radiation
401 M Street, SW Office of Air Quality Planning and
Washington, DC 20460 Standards
Research Triangle Park, NC 27711
September 1978
-------
DISCLAIMER
This report has been reviewed by the Office of Air Quality
Planning and Standards and the Office of Planning and Evaluation,
U.S. Environmental Protection Agency and approved for publication.
Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
-------
m
EXECUTIVE SUMMARY
This three-volume report describes a nine-month study performed by
Systems Applications, Incorporated for the Environmental Protection Agency
to recommend and develop models that predict the contribution of manmade
air pollution to visibility impairment in Class I areas. Two models were
developed: One is a near-source plume model, based on a Gaussian formula-
tion, that is designed to compute the impact of a plume on visual range and
atmospheric coloration. The other is a regional model designed to calculate
pollutant concentrations and visibility impairment resulting from emissions
from multiple sources within a region with a spatial scale of 1000 km on a
temporal scale of several days.
The objective of this effort was to develop models that are useful pre-
dictive tools for making policy and regulatory decisions, for evaluating the
impacts of proposed new sources, and for determining the amount of emissions
reduction required from existing sources, as mandated by the Clean Air Act
Amendments of 1977.
Both visibility models (plume and regional) are based on atmospheric
dispersion models that account for the transport, diffusion, and surface
deposition of emissions from point sources. Concentrations of nitrogen
dioxide (N0?) are computed using a modified steady-state relationship;
sulfate and nitrate concentrations are calculated from S09 and NO emis-
I- A
sions using user-inputed pseudo-first-order rate constants. The aerosol
size distribution is assumed to consist of four log-normally distributed
modes: background submicron (accumulation), background coarse, plume
coarse (primary particulate), and plume submicron (secondary particulate).
The effect of relative humidity on the mass of liquid water associated with
submicron aerosol is included in the calculation of aerosol volume. The
-------
IV
spectral light intensities at 39 wavelengths of light in the visible spec-
trum are calculated for given lines of sight through the plume and the
background atmosphere by treating the plume as a homogeneous layer and the
background atmosphere as two homogeneous layers. The plume optical thick-
ness is calculated taking into account the geometry of the plume and the
observer's line of sight.
For a given line of sight, the spectral light intensities are used to
compute parameters that characterize visibility impairment, including vis-
ual range, luminance, chromaticity coordinates, contrast between plume and
background at various wavelengths, the blue-red luminance ratio, and a
parameter that characterizes plume perceptibility due to changes in both
light intensity and color. These quantitative specifications of discolor-
ation can be translated to Munsell color notation so that a representative
color paint chip can be selected, thereby providing a subjective understand-
ing of the computed color. Finally, with these color chips and a computer
graphics capability that displays perspective views of plumes and background
terrain, color illustrations of calculated atmospheric discoloration and
plume impact can be prepared.
The plume model was applied to the hypothetical case of a 2250 Mwe
coal-fired power plant emitting primary particulate, S09, and NO at the
C. A
maximum rates permitted by EPA's New Source Performance Standards. Ambient
conditions typical of the Southwest were used in these calculations,
including a background visual range of 130 km (80 miles). For an assumed
sulfate formation rate of 0.5 percent per hour, the plume's impact on
visual range was small near the source, but it increased with distance
from the source as sulfate aerosol was formed. For neutral stability and
for sight paths perpendicular to the plume centerline, the calculated
visual range was reduced approximately 5 percent from the background value
at distances 200 to 300 km downwind of the power plant. Except at short
downwind distances, where scattering is dominated by primary particulate
(fly ash), sulfate formed in the atmosphere was the principal cause of
reductions in visual range. The calculations, however, indicate that plume
-------
discoloration was caused principally by N02. Yellow-brown plume discolor-
ation was most prominent during stable conditions at downwind distances of
40 to 100 km. Light scattered by primary particulate and sulfate aerosol
tended to mask the discoloration due to N02; thus, control of particulate
and S02 emissions would increase discoloration. Sulfate aerosol, by itself,
tended to cause a white plume (in forward scatter) or a grey plume (in back
scatter) when viewed against the horizon sky.
The regional model was applied to point source emissions of SO in
/\
the Northern Great Plains for two cases: estimated 1975 emissions of 430
tons/day and projected 1986 emissions of 1990 tons/day. For a stagnation
episode characterized by light, variable winds, the maximum reduction in
visual range was calculated to be 10 percent for the 1975 emissions and
25 percent for the 1986 emissions. The most significant reductions in
visual range occurred hundreds of kilometers from the emissions sources.
For comparison, the impact of a hypothetical complex of copper smelters
with a total SO emissions rate of 6000 tons/day (equivalent to the 1972
/\
emissions from copper smelters in Arizona) was evaluated with the Northern
Great Plains regional visibility model. The maximum reduction in visual
range (for stagnation conditions) was calculated to be 50 percent and to
occur hundreds of kilometers from the sources. Calculated maximum sulfate
concentrations were in rough agreement with maximum sulfate concentrations
measured in nonurban locations in Arizona in 1972-1973.
National Weather Service visual range and meteorological data for the
period 1948-1976 at 18 locations in the western United States were analyzed.
Trends in visibility in Phoenix and Tucson were found to be correlated with
trends in SO emissions from copper smelters. During periods of copper
A
strikes when smelter SO emissions were eliminated, and during the period
X
1973-1976 when SO emissions from copper smelters were reduced from 6000 to
X
3000 tons/day, visibility improved. The effect of smelter SO emissions on
/\
visual range at distant nonurban locations was evaluated by comparing visual
range frequency distributions at Prescott, Winslow, and Farmington during
the 1967-1968 copper strike with other periods. Significant improvements
-------
in visibility during the copper strike were particularly associated with
winds that would transport smelter emissions directly to the given loca-
tions.
-------
VII
CONTENTS
DISCLAIMER ii
EXECUTIVE SUMMARY 111
LIST OF ILLUSTRATIONS ix
LIST OF TABLES xiii
LIST OF EXHIBITS xiii
ACKNOWLEDGMENTS xiv
DEDICATION xv
I INTRODUCTION 1
II THE NATURE OF VISIBILITY IMPAIRMENT 8
A. Definition of Visibility Impairment 8
B. Fundamental Causes of Visibility Impairment 17
C. Visibility Impairment in the Western United States 24
III THE ELEMENTS OF VISIBILITY MODELS 35
A. Pollutant Transport, Diffusion, and Removal 35
1. Initial Dilution in a Buoyant Plume 37
2. Gaussian Plume Diffusion 39
3. Observer-Plume Orientation 40
4. Limited Mixing 42
5. Plume Trajectory Box Model 42
6. Regional Transport and Diffusion 46
B. Atmospheric Chemistry 48
1. Conversion of NO to N02 48
2. Conversion of Gases to Particles 53
C. Aerosol Size Distribution 57
D. Atmospheric Optics • 65
1. Calculation of the Scattering and
Absorption Properties 65
2. Calculation of Light Intensity 71
-------
III THE ELEMENTS OF VISIBILITY MODELS (continued)
E. Quantifying Visibility Impairment 75
1. Visual Range 76
2, Contrast of Haze Layers and Plumes 78
3. Color 81
4. Color Changes 86
IV THE OUTPUT OF VISIBILITY MODELS 89
A. The Plume Visibility Model 90
B. Plume/Terrain Perspective Model 110
C. Color Display Techniques 113
1. Color Illustration 130
2. Color Video Display 136
D. The Regional Visibility Model ..... 138
V RECOMMENDATIONS FOR FUTURE WORK 155
A. Impact Analysis in Support of Regulation Development . . 156
B. Model Refinement and Testing 157
1. Model Testing 158
2. Gas-to-Particle Conversion and Aerosol Growth .... 160
3. Assessment of Color Impact Thresholds 161
4. Refinement of Color Display 161
C. Model Validation 163
1. The Type of Measurement Program 163
2. Specific Measurements 165
3. Data Analysis, Assessment of Model Performance,
and Model Refinement 170
D. Further Data Analysis 172
E. Development of a Southwest Regional Visibility Model . . 174
REFERENCES 176
GLOSSARY 187
FORM 2220-1 193
NOTE 194
-------
ILLUSTRATIONS
1 Outline of SAI's Study To Develop Models for Predicting
Visibility Impairment 4
2 Elements and Potential Uses of Visibility Models 6
3 An Example of Reduced Visual Range: Marble Canyon, Arizona .... 10
4 An Example of Near-Source Visibility Impairment: Plume
Discoloration Downwind of the Navajo Generating Station
in Northern Arizona 13
5 An Example of Visibility Impairment: A Uniform Haze
Layer Visible from Bryce Canyon National Park 14
6 Natural Causes of Visibility Impairment 16
7 Effect of an Atmosphere on the Perceived Light
Intensity of Objects 21
8 Map of the Western United States Showing the Locations of
Large Point Sources, Mandatory Federal Class I Areas, and
NWS Stations Where Visibility Observations Are Made 26
9 Frequency Distributions of Extinction Coefficients Based
on Visual Range Observations at 13 Western U.S. Locations
on Days Without Precipitation or Fog in 1976 27
10 Dependence of Visual Range on Relative Humidity at Four
Locations in the Southwest 28
11 Historical Trends in Visibility in Phoenix, Arizona 30
12 Percentage of Daylight Observations with RH < 60 Percent
for Which Visual Ranqe Exceeded 121 km, as a Function of
Wind Direction, at Farmington, New Mexico, 1949-1976 . 31
13 Schematic Logic Flow Diagram of the Visibility Models 36
14 Gaussian Plume Visual Impact Model: Observer-Plume Geometry ... 41
15 Plume Trajectory Box Model 43
16 Sensitivity of NO to N02 Conversion in Power Plant Plumes
to the Rate of Plume Dilution, Background Ozone Concen-
tration, and Solar Radiation 52
-------
17 Comparison of Measured N02/N0x Mole Ratios (Circled Points)
in the Plume Centerline Downwind of a Coal-Fired Power
Plant with Computer-Calculated Values (Solid Lines) Using
Standard Pasquill and Fitted or 54
18 Schematic of an Atmospheric Aerosol Surface Area Distri-
bution Showing the Principal Modes, Sources of Mass
for Each Mode, Process Involved in Inserting Mass in
Each Mode, and Removal Mechanisms 58
19 Average Urban Model Aerosol Distribution Plotted in
Five Different Ways 59
20 Scattering-to-Volume Ratios for Various Size Distributions ... 63
21 Ratio of Light Scattering to Mass as a Function of
Relative Humidity 70
22 Light Scattering and Absorption in the Atmosphere 72
23 An Example of Plume Visual Impact 79
24 Chromaticity Diagram 82
25 Spectral Tristimulus Values x~(x), y(A), z"(x) 84
26 Representation of a Color Solid 85
27 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering
Angle of 45° and Stability Class C 100
28 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering
Angle of 45° and Stability Class D 101
29 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering
Angle of 45° and Stability Class E 102
30 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering
Angle of 90° and Stability Class C 103
31 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering
Angle of 90° and Stability Class D 104
32 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering
Angle of 90° and Stability Class E 105
-------
33 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering
Angle of 180° and Stability Class C 106
34 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering Angle
of 180° and Stability Class D 107
35 Calculated Plume Visibility Impairment for a Hypothetical
2250 Mwe Coal-Fired Power Plant with a Light Scattering Angle
of 180° and Stability Class E 108
36 Observer Locations for Plume-Terrain Perspective Views 112
37 View from Location 1 114
38 View from Location 2 115
39 View from Location 3 116
40 View from Location 4 117
41 View from Location 5 118
42 View from Location 6 119
43 View from Location 7 120
44 View from Location 8 121
45 View from Location 9 122
46 View from Location 10 123
47 View from Location 11 124
48 View from Location 12 125
49 View from Location 13 126
50 View from Location 14 127
51 View from Location 15 128
52 Schematic Showing Color Display Techniques 129
53 Photocopy of the Power Plant Plume in Northern.Arizona
Selected for Simulation 132
-------
xii
54 Photocopy of the Perspective Terrain View with Calculated
Munsell Color Notation and Corresponding Color Chips 134
55 Photocopy of Color Illustration Created by an
Artist from Indicated Munsell Color Chips 135
56 Photocopy of a Color Video Display of Plume Visual Impact - . . . 137
O
57 Predicted Regional Three-Hour-Average SO? Concentrations (pg/m )
Based on 1986 Emissions in the Northern Great Plains for the
Meteorological Conditions of 1400-1700 MST on 6 April 1976. ... 140
58 Predicted Regional Three-Hour-Average N0£ Concentrations (ppb)
Based on 1986 Emissions in the Northern Great Plains for the
Meteorological Conditions of 1400-1700 MST on 6 April 1976 ... 141
59 Predicted Regional Three-Hour-Average Sulfate Concentrations
(pg/mj), Assuming a 0.5 Percent per Hour Sulfate Formation Rate,
Based on 1986 Emissions in the Northern Great Plains for the
Meteorological Conditions of 1400-1700 MST on 6 April 1976 ... 142
60 Predicted Regional Visual Range (km), Assuming a Background
Visual Range of 130 km and a 0.5 Percent per Hour Sulfate
Formation Rate, Based on 1986 Emissions in the Northern Great
Plains for the Meteorological Conditions of 1400-1700
MST on 6 April 1976 143
61 Predicted Regional Three-Hour-Average S02 Concentrations
(yg/m3) Based on Hypothetical Copper Smelter Emissions
for the Meteorological Conditions of 1400-1700 MST on
6 April 1976 in the Northern Great Plains 145
62 Predicted Regional Three-Hour-Average Sulfate Concentrations
(yg/m^), Assuming a 0.5 Percent per Hour Sulfate Formation
Rate, Based on Hypothetical Copper Smelter Emissions for
Meteorological Conditions of 1400-1700 MST on 6 April 1976
in the Northern Great Plains 146
63 Predicted Regional Visual Range (km), Assuming a Background
Visual Range of 130 km and a 0.5 Percent per Hour Sulfate
Formation Rate, Based on Hypothetical Copper Smelter Emissions
for Meteorological Conditions of 1400-1700 MST on 6 April 1976
in the Northern Great Plains 147
64 Predicted Regional Visual Range (km), Assuming a Background
Visual Range of 130 km and a 0.3 Percent per Hour Sulfate
Formation Rate, Based on Hypothetical Copper Smelter Emissions
for Meteorological Conditions of 1400-1700 MST on 6 April 1976
in the Northern Great Plains 149
65 Predicted Regional Visual Range (km), Assuming a Background
Visual Range of 130 km and a 1 Percent per Hour Sulfate Formation
Formation Rate, Based on Hypothetical Copper Smelter Emissions
for Meteorological Conditions of 1400-1700 MST on 6 April 1976
in the Northern Great Plains 150
-------
xiii
TABLES
1 Main Milestones in Visibility Protection Regulation 2
2 Percent Reduction in Visual Range by Three Levels
of S0? Emissions 45
3 Measured Size Distributions of Atmospheric Aerosol 62
4 Estimates of Extinction Coefficients per Unit Mass 67
5 Mean and Maximum 24-Hour-Average Sulfate Concentrations
Measured in Arizona in 1972-1973 151
EXHIBITS
1 Sample Plume Model Output: Input Emissions
and Ambient Conditions , 91
2 Sample Plume Model Output: Pollutant Concentrations 93
3 Sample Plume Model Output: Visual Effects for
Horizontal Sight Paths 95
4 Sample Plume Model Output: Visual Effects for
Nonhorizontal Sight Paths 97
5 Sample Plume Model Output: Visual Effects for
White, Grey, and Black Backgrounds 98
6 Example of Visual Effects at a Given Location in a
Region with 2 ug/nr S0= and 30 ug/m3 Coarse Particulate 153
-------
XIV
ACKNOWLEDGMENTS
Many people inside and outside of SAI have assisted us in this
work. At SAI we particularly want to thank Gary Lundberg for his
excellent computer work and Shep Burton for his guidance. The support
of John Butler, David Shaver, James Dicke, John Bachmann, Steve Eigsti,
Joseph Tikvart, Terry Thoem, and Donald Henderson of the EPA is greatly
appreciated. We thank Donald LaBash for his color illustrations and
Ellen Leonard and Michael Williams of Los Alamos Scientific Laboratory
for their help with the color video display techniques. Also, we grate-
fully acknowledge the conversations and helpful suggestions of George
Hidy, Robert Charlson, John Trijonis, Thomas Peterson, William Wagner,
and Edwin Roberts.
-------
XV
DEDICATION
We dedicate this report to the memory of Terry N. Jerskey.
-------
1 INTRODUCTION
The Clean Air Act was amended in August 1977 to contain a section
requiring the restoration and protection of visibility in national parks,
wilderness areas, and forests that have been classified as mandatory Class I
federal areas.* In that section Congress boldly declared "as a national
goal the prevention of any future, and the remedying of any existing
[our emphasis], impairment of visibility in mandatory Class I Federal
areas which impairment results from man-made air pollution." Furthermore,
Congress defined "impairment of visibility" as a reduction in visual range,
an atmospheric discoloration, or both.
Although many sources of air pollution cause or contribute to vis-
ibility impairment, including vehicular emissions from urban areas and
sulfur dioxide emissions from copper smelters, Congress was particularly
concerned with the alleged impact of power plant emissions on the magnif-
icent vistas in national parks such as the Grand Canyon and Bryce Canyon.
The 1977 Clean Air Act Amendments require that the foil owing action be
taken (see Table 1):
> The Department of the Interior, in conjunction with the
Environmental Protection Agency (EPA), must identify all
mandatory Class I areas where visibility is deemed to be
aesthetically important.
> The EPA must report to Congress on the emissions sources
and the air pollutants that cause or contribute to visibil-
ity impairment, on recommended methods of measuring
* Section 128 of Public Law 95-95 amends Part C of Title I of the Clean
Air Act by adding Section 169A concerning "visibility protection for
Federal Class I areas."
-------
TABLE 1
Date
August 1977
February 1978
August 1978
February 1979
August 1979
MAIN MILESTONES IN VISIBILITY PROTECTION REGULATION
Milestone
Congress passed the 1977 Clean Air Act Amendments
containing Section 169A on visibility protection:
"Congress hereby declares as a national goal the
prevention of any future, and the remedying of any
existing, impairment of visibility [defined as a
'reduction in visual range1 or an 'atmospheric
discoloration'] in mandatory Class I Federal areas
which impairment results from manmade air pollution"
Department of Interior is required to identify all
mandatory Class I areas where visibility is an impor-
tant aesthetic value
EPA is required to promulgate a list of mandatory
Class I federal areas where visibility is an important
aesthetic value
EPA is required to report to Congress recommending
methods to characterize visibility in Class I areas,
modeling techniques (or other methods) for determining
the impact of man-made air pollution on visibility, and
methods to prevent and remedy air pollution
EPA is required to promulgate regulations requiring
states to "make reasonable progress" toward preventing
and remedying visibility impairment, including use of
the best available retrofit technology for point sources
less than 15 years old and development of a 10 to 15
year long-term strategy
-------
visibility impairment in Class I areas, on recommended
modeling techniques for determining the contribution of
man-made air pollution to visibility impairment in Class I
areas, and on methods for preventing and remedying such air
pollution.
> The EPA must promulgate regulations that ensure reasonable
progress toward meeting the national visibility goal.
It is clear that Congress intended that existing sources be controlled
so that visibility would not be significantly impaired in Class I areas.
Each existing major stationary source less than 15 years old that emits
an air pollutant that causes or contributes to any Class I area visi-
bility impairment is required to procure, install, and operate the "best
available retrofit technology" for controlling air pollutant emissions
"as expeditiously as possible." However, the Environmental Protection
Agency may exempt a source from this retrofit requirement on the grounds
that the source does not emit pollutants that cause or contribute to
significant visibility impairment in Class I areas. Although the main
focus of the visibility protection section (No. 169A) is on existing
sources, other sections of the 1977 Clean Air Act Amendments require
that new-source reviews demonstrate that the emissions from a proposed
major facility do not cause "an adverse impact on the air quality related
values (including visibility)" of Class I areas.
This report describes a nine-month study performed by Systems
Applications, Incorporated (SAI) for the Environmental Protection Agency
to recommend and develop models that predict the contribution of man-
made air pollution to visibility impairment in Class I areas. As shown
in Figure 1, this study was divided into three main phases, consisting
of seven tasks:
> Model formulation
- Recommendations for modeling approaches.
- Collection and analysis of data to characterize
existing visibility impairment.
-------
oo
§
crt LU
_l V)
Q
ii
o z
< o
O. —
zi-
^i
CO O
>
0<
OZ
Q>—
I UJ
zo
oz
O UJ O
Ul< XZ
— t— t—
0:0 <
o:
uj o r
Q CO *— — I O
< —
a
z
2 3Z
o a- r>
o •>
o tea
o o o
co o cr co cr
Ul —CO 19 Z 19
I- >- z :*o *
< cr o o— o
cr < — cr
o >- m >- \- t-
ei o o
iC UJ
(_> u. -5
< Oca
CO O
CO CC to
oui
z_l
o
CO
< cr uj LU ui
ZH- —o —
a: a.
t- i
to o
— z to<
>- to O — CT
_! O
Q- — I- _l
O-Z — ZO
< UJO O O
ui a. a < o
Ul O Ul LU CO
— LU— X<
> 19 > OO
t
Q_
O
I I—
cr cr
Ocoo
I UJ CO
•iz J .
< UJ
z o
— co
o
to a:
Ul UJ
< —
— ouj
> IT Z
UJ— ~
K
o ui z
z x o
— i- o
~ to uj
CO OQ
3CO
O UJ
LU I— CO
a
< cr i_
UJ Q
Z X<
< h-:
QUJ<
C£
H^
a.
co
h—t
00
coo —
O U- QQ
_
Ul <
Q a-
O Z
to o
— co
> i
i z
_iui
o
o
cr
I
o
o
Qa:
zo
oco
o.
tu. U. H-
cr u. i
o
cr —I
u.o
co oo
to
< ZO
£ o —
i —cr
o i-ui
i- < x
i -lo-
K 3UI —
QQ> COZ
UJ
QE:
CD
I—I
u_
-------
> Model development
- Near-source plume visibility model.
- Regional, multiple source visibility model.
> Model applications
- Documentation of models.
- Application of models to typical emission, meteoro-
logical, and ambient conditions.
- Recommendations for further work.
- Information for the Report to Congress.
Two basic types of models were developed for the prediction of
visibility impairment caused by anthropogenic pollutant emissions. One
is a near-source plume model, applicable to a wide range of emission,
meteorological, and ambient conditions, that is based on a Gaussian
formulation. This model is designed to estimate plume concentrations
and visual effects on a spatial scale of up to 350 km. The other is a
regional-scale model designed to estimate pollutant concentrations and
visual effects in the Northern Great Plains over a spatial scale of 1000 km.
Both models calculate the reduction in visual range and the atmospheric dis-
coloration resulting from directly emitted primary particulate matter and
from nitrogen dioxide, sulfates, and nitrates formed in the atmosphere
from pollutant precursors.
The goal of this development is models that are useful predictive
tools for makinq policy and regulatory decisions, for evaluating the
impact of proposed new sources, and for determining the amount of emissions
reduction required from existing sources. Figure 2 shows the elements
and potential uses of visibility models. The critical decisions that
must be made in the future regarding the definition of "significant
visibility impairment" cannot be made without a basic understanding of
the implications for enforcement in new and existing source reviews, and,
in particular, the impact on energy development in the western United
States. We believe that visibility issues will have a significant
influence on the siting and design of new coal-fired power plants in
------- An error occurred while trying to OCR this image.
-------
the West and on the evaluation of retrofit requirements for existing
plants. Thus, modeling of visibility impairment is likely to become an
integral part of the environmental assessment of new and existing sources,
providing a basis for major siting and pollution control decisions in
the future.
This report contains five chapters and seven appendices. Chapter II
discusses the nature of existing visibility impairment in Class I areas,
with particular emphasis on the western United States. The elements of the
visibility models are discussed in Chapter III, and a summary of the plume
and regional models and their output is given in Chapter IV. Chapter V
discusses our recommendations for future work. Appendix A presents the
details of our analysis of visibility data in the western United States,
and Appendix B, the details of the atmospheric optics calculations.
Appendix C discusses sulfate formation in the atmosphere, Appendix D
describes the plume model, and Appendix E shows plume model calculations
and the results of a sensitivity analysis. Appendices F and G describe
the Northern Great Plains regional model and calculations for seven
scenarios. The appendices are bound separately in Volume II. Volume III
presents a series of case studies of power plant plume visual impact for
a number of emission, meteorological, and ambient background scenarios.
-------
8
II THE NATURE OF VISIBILITY IMPAIRMENT
"Visibility impairment" must be defined before discussing the components
of the mathematical models used to predict anthropogenic visibility impair-
ment and the specific plume and regional visibility models developed as a part
of this study. This chapter defines and classifies "visibility impairment"
by type, magnitude, spatial and temporal extent, and cause, and it provides
examples of visibility impairment. The fundamental physical concepts of
visibility impairment are also briefly discussed, and the analysis of visual
range data from the Southwest and the Northern Great Plains is summarized;
this material is presented in more detail in Appendix A.
A. DEFINITION OF VISIBILITY IMPAIRMENT
"Visibility impairment" has been defined generally by Congress in the
Clean Air Act Amendments of 1977 as a "reduction in visual range" or an
"atmospheric discoloration," but the term must be defined more precisely and
illustrated by examples before we discuss the models and quantitative speci-
fications of visibility impairment.
Visibility impairment can be defined and classified according to:
> Type (e.g., the appearance of distant objects, general hazi-
ness, yellow-brown or grey discoloration).
> Magnitude (e.g., visual range, degree of coloration, contrast,
"any" or "significant" impairment of visibility in the termin-
ology of the Clean Air Act Amendments).
> Spatial extent (e.g., localized plume appearance, uniform haze,
distance downwind of source).
> Temporal extent (frequency of occurrence of reduced visual
range or of discoloration).
-------
> Location relative to Class I areas (impact on a vista from
a Class I area or on a vista looking into a Class I area).
> Cause (natural or man-made aerosols of coarse particulate
matter, sulfate, nitrate, organics, soot, or nitrogen
dioxide gas).
The term "visibility" is generally used synonymously with "visual
range," meaning the farthest distance at which one can see a large, black
object against the sky at the horizon. One can make subjective evaluations
of "visibility" every time he views objects outdoors. Although large
black objects are not generally available for observing and evaluating
visual range, dark objects such as buildings, TV towers, hills, or mountains
can be viewed against the horizon sky.
Even if no distant objects are within view, subjective judgments about
visual range can be made by noting the coloration and light intensity of the
sky and nearby objects. For example, one perceives reduced visual range if
a distant mountain that is usually visible cannot be seen, if nearby objects
look "hazy" or have diminished contrast, or if the sky is white, grey,
yellow, or brown rather than blue.
Figure 3 shows an example of reduced visual range in Marble Canyon in
northern Arizona. As shown by this photograph, reduced visual range is
detectable because the distant walls of the canyon are difficult to distin-
guish. The contrast between the given object (part of the canyon) and the
background (the horizon or a more distant terrain feature) is reduced by
light scattered from particles in the intervening atmosphere. Also, even
if terrain features were not visible, the intensity and the yellowish
coloration of the scattered light would degrade the aesthetic quality of the
atmosphere. Many of the Class I areas (e.g., national parks, national
forests, wilderness areas) were so designated because of their scenic views
of such distant terrain features as mountains, canyon walls, plateaus, and
buttes. Indeed, in the western United States, where most of the Class I
areas are located, spectacular scenery is enhanced by generally excellent
visibility, which makes the colorful terrain features stand out with great
-------
10
o
o
>-
•a:
o
UJ
_l
CO
cc:
UJ
C3
1
_i
<
oo
i—i
>
o
LU
O
O
LU
C£
U_
O
LU
I
0_
-------
11
clarity. However, even in flat areas (e.g., the Big Sky Country of the
Northern Great Plains), a slight reduction in visual range or a slight
atmospheric discoloration can change what originally appeared to be an
"infinite" horizon to a less desirable white, yellow, grey, or brown
horizon.
The magnitude of impairment can be characterized by the reduction in
visual range from some reference value, by a reduction in contrast between
an object and the horizon sky at a known distance from the observer, or by
a shift in coloration or light intensity of the sky or distant objects,
such as clouds or terrain features, compared with what is perceived on a
"clear" day. In all cases, the magnitude of visibility impairment can be
characterized by the change in light intensity or coloration of an object
(or part of the sky) compared with that of some reference object. For
example, a distant mountain is visible because the intensity and coloration
of light from the mountain is different from that of the horizon sky.
Another example is a plume or haze layer seen against the background sky
or terrain features. The pollution is visible only if the light intensity
or coloration of the plume contrasts with that of the surrounding sky or
terrain.
The most subjective impairment observation occurs when an observer
compares the appearance of the sky or distant terrain features on a hazy
day with recollections of what it was on a clear day in the past. Exam-
ples of such subjective comparisons are those of old-timers who recall
that visibility used to be much better. Although these observations can-
not be discounted entirely, it is possible that such judgments may be the
result of nostalgia or poor memory.
The spatial extent of visibility impairment is important to both the
perception and the significance of impairment to observers in Class I
areas. The sensitivity of an observer to brightness and color differences
between two objects depends on the geometric relationship between the
objects. If each of the objects is uniformly colored and there is a sharp
-------
12
line of demarcation between the objects, such as when a mountain is viewed
against a horizon sky, a smaller change in light intensity or color can
be perceived than if the boundary between the two objects is vague, as in
the case of a plume viewed against the horizon sky. If the observer is
located in a uniformly colored atmosphere, atmospheric discoloration is
perceived, not by comparison of two colored fields, but by comparison
with his recollection of a clear atmosphere.
Figure 4 provides an example of near-source visibility impairment.
The spatial extent of visibility impairment is defined by the dimensions
of the plume. The plume is visible because the light intensity and color
of the plume are different from those of the clouds in the background.
Because of the resultant relatively sharp boundary between the plume and
the background, the visual impact on the observer is dramatic. Figure 5
shows another example of the importance of the spatial extent of impair-
ment on its perceptibility. In that photograph, the haze layer is clearly
visible because of the sharp demarcation line between it and the layer of
clean air above it.
The temporal extent (or the frequency of occurrence) of visibility
impairment) is of great importance in determining the acceptability of air
pollution levels. The frequency of occurrence of impairment could be char-
acterized by stating the number of days or hours in a year that the magni-
tude of visibility impairment is greater than some standard. Using these
data, one might state that at a given location it is acceptable for visual
range to be less than y km for x percent of the daylight hours.
The location of visibility impairment is extremely important in terms
of visibility protection legislation because the law states that only the
visibility in Class I areas is to be protected and restored. We assume that
this definition can include impairment caused by pollution outside of a
Class I area that is visible within a Class I area. In areas with excel-
lent background visibility, visual degradation perceived by an observer in
a Class I area could be caused by pollution many kilometers away. It is not
-------
13
LU Z
SO
0-j.C
oo
l—l CtL
- o
•~2
S$
i—i ^
CO Z
I—I
OO LU
i—i I
> I—
LU U_
O O
•=> c*
o z
00 H-I
i 3 <;
S3 O
O M
Z Q i-t
OH
Lu Z ct
O O
LU £ o:
_J
-------
14
a:
OL.
O
2: o
H- o
z >-
o: o
I— I
c£ LU
D- U
s: >-
i— i OL
CO
CO
i— i LU
(/) _J
1-1 CQ
Q- >-
S «C
< —I
X
LU LU
M
LO
LU
o:
=5
C3
-------
15
clear whether Congress meant to protect the visibility in Class I areas
from such distant pollutants.
Finally, perhaps the most important classification of visibility
impairment is by cause, in particular, whether the cause is natural or
anthropogenic sources. Clearly, Congress has been concerned only with
anthropogenic visibility impairment. Reductions in visual range caused
by precipitation, fog, clouds, windblown dust, sand, snow, or "natural"
aerosol are natural occurrences and cannot be controlled by man. Indeed,
some natural visibility impairment may contribute to the enjoyment of
Class I areas. Examples of such phenomena are the blue haze of the Great
Smoky Mountains and the fog and hazes along the California and Oregon
coast.
Assessment of anthropogenic contributions to visibility impairment
can be difficult when background visibility varies spatially and temporally
and when natural atmospheric constituents interact with anthropogenic
emissions to create a combined effect, such as that of the haze formed
when anthropogenically emitted hygroscopic particles absorb liquid water
in the atmosphere. For the purpose of discussion only, the Venn diagram
in Figure 6 shows the frequency of occurrence of visibility impairment,
which is defined here as visual range less than 80 km (50 miles). Some
natural causes of visibility impairment, such as windblown dust, precip-
itation, fog, and cloud cover, are represented by circles whose total
areas and areas of overlap represent the frequency of occurrence of the
given phenomenon and the associations among phenomena. Note in Figure 6
that fog always causes visibility impairment and precipitation usually
does. In this highly schematic representation, the diagonally lined area
represents the fraction of time that man-made emissions cause or contribute
to visibility "impairment. In actual situations, it is difficult to sep-
arate the relative magnitude's of natural and man-made contributions to
-------
16
oo
z
o
*— i
«=c
a;
CD
O
CQ
»— H
oo
o
-------
17
visibility impairment. For example, as noted above, during humid condi-
tions, natural and anthropogenic hygroscopic particles may absorb water,
thereby causing visibility impairment. Because the atmosphere is a complex
system consisting of both natural and anthropogenic aerosol, the contribu-
tion of man-made pollution to visibility impairment may be difficult to
characterize by measurements. The causes of visibility impairment are dis-
cussed in further detail in the sections below from both fundamental and
phenomenological viewpoints.
B. FUNDAMENTAL CAUSES OF VISIBILITY IMPAIRMENT
Visibility impairment is caused by the following interactions in the
atmosphere:
> Light scattering
- By molecules of air
- By particles
> Light absorption
- By gases
- By particles.
Light scattering by gaseous molecules of air (Rayleigh scattering), which
causes the blue color of the atmosphere, is dominant when the air is rela-
tively free of aerosols and light-absorbing gases. Light scattering by
particles is the most important mechanism causing reductions in visual
range. Fine solid or liquid particulates whose diameters range from 0.1
to 1.0 ytn (the most effective size per unit mass in scattering light)
account for most of atmospheric light scattering. Light absorption by
gases is particularly important in the discussion of anthropogenic visi-
bility impairment because nitrogen dioxide, a major constituent of power
plant plumes, absorbs light. Nitrogen dioxide is reddish-brown because
it absorbs strongly at the blue end of the visible spectrum while allowing
light at the red end to pass through. Light absorption by particles is
-------
18
important when black soot (finely divided carbon) is present. However,
most atmospheric particles are not generally considered to be light
absorbers.
Anthropogenic contributions to visibility impairment result from the
emission of primary particulate matter (such as fly ash, acid, or water
droplets, soot, and fugitive dust) and of pollutant precursors that are
converted in the atmosphere to the following secondary species:
> Nitrogen dioxide (NCL) gas, from emissions of nitric
oxide (NO).
> Sulfate (SOp particles, from SOX emissions.
> Nitrate (NO:;) particles, from NOV emissions.
O X
> Organic particles, from hydrocarbon and NO emissions.
X
Before particulate control technology was commonly employed, primary par-
ticulate matter, such as smoke, windblown dust, and soot, was a major con-
tributor to visibility impairment. Coal-fired power plants emit primary
particles of fly ash and combustion-generated particulates to the atmo-
sphere. If such plants are equipped with efficient precipitators or other
abatement equipment, the emission rate of primary particles may be small.
However, some emissions escape the control equipment and do contribute
to the ambient particulate concentration and hence to general visibility
impairment. If the emission rate of primary particulates is sufficiently
large, the plume itself may be visible.
In the past, many of the older coal-fired power plants generated
conspicuous, visible plumes resulting from the large emission rates of
primary particulate matter. Old plants still in operation and newer plants
have benefited from more efficient particulate abatement equipment and a
state-of-the-art that has reached the point where particulate removal
efficiencies in excess of 99 percent are commonly specified and achieved.
In addition, with the installation of flue gas desulfurization systems
(scrubbers) and with boiler combustion modifications, sulfur dioxide and
-------
19
nitrogen oxide emissions have also been reduced. As a result, the visual
impact of power plant plumes has been sharply reduced, as evidenced by
the nearly invisible plumes of modern coal-fired power plants. Unfor-
tunately, however, the contribution to visibility impairment of the
secondary pollutants—nitrogen dioxide gas and sulfate, nitrate, and
organic aerosol--is now becoming increasingly evident and is of growing
concern.
Since nitrogen dioxide absorbs light selectively, it can discolor
the atmosphere, causing a yellow or brown plume when present in sufficient
concentrations. Almost all of the nitrogen oxide emitted from power plant
stacks is nitric oxide, a colorless gas. But chemical reactions in the
atmosphere can oxidize a substantial portion of the colorless NO to the
reddish-brown N(L.
Secondary sulfate, nitrate, and organic particles have a dominating
effect on visual range in many situations because these particles range in
size between 0.1 and 1.0 ym in diameter, which is the most efficient size
per unit mass for light scattering. As is noted later, submicron aerosol
(with diameters in the range from 0.1 to 1.0 ym) is 10 times more effec-
tive in light scattering than the same mass of coarse (> 1 ym) aerosol.
Also, because secondary aerosol forms slowly in the clean atmospheres
typical of Class I areas in the western United States, maximum aerosol
concentrations and associated visibility impairment may occur at large
distances from emissions sources. We discuss this problem further in
Section C, which concerns visual range observations in the West, and in
the chapters that describe the models.
The effect of the intervening atmosphere on the visibility and colora-
tion of a viewed object (e.g., the horizon sky, a mountain, a cloud) can
be calculated by solving the radiation transfer equation along the line of
sight. As we noted in Section A, visibility impairment can be quantified
by comparing the intensity or the coloration of two objects (e.g., a distant
-------
20
mountain against the horizon sky). The effect of the intervening atmosphere
on the light intensity, as a function of wavelength, of the viewed object can
be determined if the concentration and characteristics of air molecules,
aerosol, and nitrogen dioxide are known along the line of sight.
The change in spectral light intensity I (A) as a function of distance
along the sight path at any point in the atmosphere, neglecting multiple
scattering (see Figure 7), can be calculated as:
where
r = the distance along the sight path from the object
to the observer,
p(0) = the scattering distribution function for scattering
angle 0 [see glossary and Figure 7(a) for definitions],
2
F = the solar flux (watts/m ) incident on the atmosphere,
bp,. = the sum of the Rayleigh scattering (due to air
sea L
molecules) and the scattering due to particles:
bscat(x> = bR(x) and bsp(x) • {2)
b . = the sum of the scattering and absorption coefficients:
bext'x) ' bscat + babs • (3)
On the right-hand side of Eq. (1), the first term represents light
absorbed or scattered out of the line of sight; the second term represents
light scattered into the line of sight. The values of bscat and bext can
be evaluated if the aerosol and NOp concentrations and such characteristics
as the refractive index and the size distribution of the aerosol are known.
Except in the cleanest atmospheres, b • . is dominated by b : also, unless
SCuU Sp
soot is present, b . is dominated by the absorption coefficient due to
-------
SUN
vZENITH
.ANGLE
SCATTERING
ANGLE e
OBJECT
ELEMENTAL VOLUME
.(CONTAINING AIR,
/PARTICLES, AND N02)
/INE OF SIGHT
+ dl
OBSERVER
(a) Geometry
LIGHT INTENSITY OF HORIZON
Object-Observer Distance r
(b) Visual Range r (Homogeneous Atmosphere)
FIGURE 7. EFFECT OF AN ATMOSPHERE ON THE PERCEIVED. LIGHT
INTENSITY OF OBJECTS
-------
22
N02- Scattering and absorption are wavelength-dependent, with effects
being largest at the blue end (A = 0.4 pm) of the visible spectrum
(0.4 < x < 0.7 pm). The Rayleigh scattering coefficient bD is propor-
-4
tional to A ; the scattering coefficient caused by particles is generally
proportional to A~n, where 1 < n < 2. Also, N02 absorption is greatest at
the blue end. This wavelength dependence causes the coloration of the
atmosphere.
For a uniform atmosphere, without inhomogeneities caused by plumes
(where b . and b t do not vary with r), Eq. (1) can be solved to find
the intensity and coloration of the horizon sky:
(4)
The perceived intensity of distant bright and dark objects will approach
this intensity as an asymptote, as illustrated by Figure 7(b).
The visual range rv is the distance at which a black object is barely
perceptible against the horizon sky, which occurs when the perceived light
intensity of the black object is (1 - C -n)Iu» where C. is the liminal
(barely perceptible) contrast, commonly assumed to be 0.02. When Eq. (1)
is solved for r , for a uniform atmosphere, ry is independent of p(0) and
FS(A) and can be calculated using Koschmieder's equation:
r .
'
where bext(A) ">s evaluated at the middle of the visible spectrum (to which
the human eye is most sensitive) and where A = 0.55 ym. The visual range
for a nonuniform atmosphere must be calculated by evaluating Eq. (1) for
the appropriate conditions of the given situation.
Atmospheric coloration is determined by the wavelength-dependent
scattering and absorption in the atmosphere. The spectral distribution
-------
23
of I(A) for x over the visible spectrum determines the perceived color
and light intensity of the viewed object. The relative contributions of
scattering (aerosols plus Rayleigh) and absorption (NO,-,) to coloration can
be illustrated by rearranging Eq. (1):
TOT dr
= b
scat
(A)
(6)
Note from Eq. (4) that when light absorption is negligible com-
pared with light scattering the clear horizon intensity is simply:
P(A,G) FS(A)
(7)
We now can rewrite Eq. (6):
dlix]
dr
I (A)
- 1
(8)
Equation (8) is thus an expression relating the effects of light scat-
tering and light absorption to the change in spectral light intensity with
distance along a sight path. On the right-hand side of Eq. (8) the first
term is the effect of light scattering (aerosol plus Rayleigh), and the
second term is the effect of light absorption (NOp). As noted previously,
since b t and bgbs (due to N02) are strong functions of wavelength and
are larger at the blue end (A = 0.4 ym) of the spectrum than the red end
(A = 0.7 ym), coloration can result.
Equation (8) makes clear that NOp always tends to cause a decrease in
light intensity and a yellow-brown coloration by preferentially absorbing
blue light, whereas particles may cause a blue-white or a yellow-brown
coloration, depending on the value of the quantity in the brackets. If,
at a given point along the sight path, I(A) is greater than the "clean"
horizon sky intensity Ino(A), then the quantity in brackets in the first
term on the right-hand side of Eq. (8) will be negative, which means that
the net effect of scattering will be to remove predominantly blue light
-------
24
from the line of sight. This effect would occur when a bright, white
cloud or distant snowbank was observed through an aerosol not containing
N02; scattering would cause a yellow-brown coloration. If, however, I(x)
is less than Ino(A), then the quantity in brackets in Eq. (8) will be pos-
itive, which means that the net effect of scattering will be to add pre-
dominantly blue light into the line of sight. This effect would occur when
a distant, dark mountain was observed through an aerosol not containing N0n;
scattering would cause the mountain to appear lighter and bluish. Only light
absorption can cause IuU) to be less than Ino(x), and whenever I(A) < InoU)>
scattering will add light to the sight path, thereby masking the coloration
caused by NCL light absorption.
Our visibility models are simply mathematical expressions for the solu-
tion to Eq. (1) for different boundary conditions and for different values
of bscat' bext' p(0)' and Fs5 as the>y are a"ffectecj b-y natural and man-made
light scatterers and absorbers. We turn now from this fundamental viewpoint
to a phenomenological viewpoint, with a summary of the results of an
analysis of visual range in the western United States, before describing the
elements and outputs of the plume and regional visibility models in Chapters
III and IV.
C. VISIBILITY IMPAIRMENT IN THE WESTERN UNITED STATES
In support of the development and validation of visibility models, we
analyzed visibility data, including National Weather Service and National
Park Service visibility observations, available nephelometer and telepho-
tometer measurements, and photographs of power plant plumes. The overall
objectives of the data analysis were:
> To determine the magnitude, temporal and spatial
variations, and natural and source-related causes of
visibility impairment in the western United States.
> To identify the meteorological and geographical con-
ditions associated with visibility impairment.
-------
25
Figure 8 shows the locations in the western United States of large
point sources, mandatory federal Class I areas, and National Weather
Service (NWS) stations from which meteorological data (including visual
range) were analyzed. Three-hourly, daytime visual range observations
for each of the 18 stations over the period from 1948 to 1976 were analyzed
and were stratified by associated meteorological conditions — precipitation,
surface wind speed and direction, atmospheric pressure, relative humidity,
total sky cover, ceiling height, season, time of day, mixing depth, mixed
layer wind speed, and ventilation—and by year—season, month, and time of
day. The details and results of this analysis are presented in Appendix A.
Figure 9 displays the frequency distributions of extinction coefficients
calculated from visual range observations at 13 NWS stations in 1976. Only
observations made on days without precipitation or fog were used to compile
these frequency distributions. Extinction coefficients were calculated from
observed visual ranges using the Koschmieder relationship [see Eq. (5)].
Several interesting observations can be made on the basis of Figure 9.
Median visual ranges for nonurban locations are on the order of 120 km or
greater, corresponding to extinction coefficients less than 0.32 x 10~^ m~ .
At nonurban locations the visual range is greater than 100 km nearly three-
fourths of the time, and extinction coefficients appear to be asymptotically
approaching a lower bound on the order of 0.2 x 10 m , which corresponds
to a visual range of 196 km, or 122 miles. The dominant cause of the
shape of the extinction coefficient frequency distribution in nonurban
areas is the strong dependence of the scattering coefficient on relative
humidity. This effect is due to the hygroscopic growth of subm'cron
aerosol particles, thereby adding to aerosol mass, an effect discussed
in greater detail in Appendices A and B.
Figure 10 summarizes the strong dependence of visual range on relative
humidity at four locations in the Southwest. The frequency of occurrence
of visual range greater than the indicated values decreases dramatically
with relative humidity at all locations except Phoenix. In addition,
-------
26
i/l _J
S,-' S >-
OC Z O £
_g 3 2
M o jZ S
IS) C
5SS 5
10 x ac i/i a= o xqc 13 o
< _, M M uj zo« z on
> *t —• wo a. •—
a < t— h- u-
uj uJt— h-l*.U- I-O
5 a: •-* o O *«
1.1 *-t ^H CO l/> UJ UJ tt
ce » z K *- o o 1/1
O > ^ «C t/> LO fcO l/l Z
h- IS) t— > Z Z O
guj l^J t/1 OC H- O O ^~ ^~
CC I- UJ Zl-l- Z
fUJ < UO CO l-l *-" O
x i— :E ^ ooo po
3 to £o a.
I— '
I—l OO LU
Z. UJ OH
13 O LU
a: o
LU oo oo
I— -z.
00 1— O
LU a. H-
a: oo
I— LU
o oo
u_ o; 3
o o. a. S 2
I
2 vO ^^cocr»o^c\jn^tf)\or*-co
-------
27
c; zz
^. o
el
! I—
i Q-
Q ^3
UJ O
00 1C
oo
I— 00
sr >-
UJ «=c
KH Q
O
1 — I zz.
u. o
u.
UJ (/)
o 2:
<_> o
I — I
2: t-
o
Q Z r—
o
>- •— i S
O I— HH
2: <:
UJ > tD
Z3 Q; o
crui u.
uj oo
a: CQ o:
u. o o
en
ui
a;
-------
28
!!!Ǥ
9 *"" 5*1®
(O
C
O
•4-J
O
O
1/5
re
T3
ro
>
O)
to
CD
QJ
I/)
re
>-
I—
I—1
O
I—I
rs
LU
O
•I
•!
O
O
8233
" « in I
O
re
re
,/x
.,-'"'
"x.
•
~x /
§ s s
IPIHH **t MMUMJM4Q IH
^/'/'
^''
1 /
\/' 2 •
S «!•
; ISSSi
/ "i"
i! ;
s s *
*••* |Mf|«
ll^tfl *• «*n»»aj»^
8
*!
•i
s
O
3
re
c
O
IM
-------
29
the trends in visual range over three decades--1948 to 1956, 1957 to
1966, and 1967 to 1976--can be determined from Figure 10. During the
period from 1948 to 1976, the frequency of good visibility decreased
slightly in both Las Vegas and Prescott, it remained relatively constant
in Farmington, and it improved in Phoenix. The significant quantities of
SO emitted by the copper smelters in Arizona (about 6000 tons per day in
A
1973) appear to be the most significant anthropogenic contributions to
visibility impairment in the Southwest.
The improvement in visibility in Phoenix during the decade from 1967
to 1976 can be explained by the reduction in emissions (by a factor of
two) that occurred during the period from 1973 to 1976. As a result of
this pollution control effort, visibility in Phoenix increased signifi-
cantly from 1973 to 1976, as evidenced by Figure 11. The effect of sul-
fate formed in the atmosphere from SO emissions from the copper smelters
A^
on visual range in Phoenix is even more clearly indicated by the increases
in visibility (as shown in Figure 11) that occurred during the years when
smelter emissions were diminished because of strikes (1949, 1959, and 1967
to 1968). The effect of total elimination of copper smelter SO emissions
A
on visual range is also illustrated in Figure 10, which shows the frequency
distribution of visual range as a function of relative humidity for the
copper strike period, which was July 1967 to March 1968. Improvements in
visual range during this-period were observed not only in Phoenix, but
also in remote nonurban areas, such as Farmington and Prescott.
The effect of smelter emissions on visual range in remote nonurban
locations is even more convincingly illustrated by Figure 12, which shows
the frequency of occurrence with which the distant visibility marker (121 km)
was visible in Farmington, New Mexico, stratified by surface wind direction.
Only observations for which relative humidity was less than 60 percent were
used in this figure so as to minimize the effect that any dependence of
relative humidity would have on wind direction. When the frequencies for
the copper strike period are compared with those for the three decades, a
-------
30
> 24 km
> 48 km
> 64 km
97 km
1950
1955
1960
1965
1970
1975
Year
FIGURE 11. HISTORICAL TRENDS IN VISIBILITY IN PHOENIX, ARIZONA
-------
31
0- O Z O>
fi LU •— O •—
O
O LU LU <£
00 O CD
VO OO Z 3 UJ
O1 00
•— «£
O £<>H-
S P2^£
I (_> 5*
O QC
X
LU
fUD
0>
< cn
=> f
00 CT>
>—I r—
>
M
n: o
o o
LU
D.
c
o
LU U
•o
c
a6uey
oo o
•z. uj
o a:
i— 1 1— i
£Q
> o
oo s
oa
O u_
o
z z
CD O
>- O
-------
32
striking improvement in visual range is apparent, particularly for the
winds (south-southeast through west) that would transport smelter emissions
directly toward Farmington (see Figure 8 for the locations of the smelters
and Farmington). This suggests that the reduced visual range in Farmington
associated with southwesterly winds is the result of SO emissions from
/\
smelters more than 400 km away. The spatial extent of anthropogenic visi-
bility impairment suggested by these data is also indicated by plume and
regional visibility model calculations presented in the following chapters.
Additional conclusions obtained through the analysis of visibility
data from the 18 NWS stations in the West are summarized below (see
Appendix A for the details of this analysis):
> Of the 16 NWS stations where long-term visibility
data appear to be valid (data taken at Alamogordo
and Ft. Huachuca were erratic and did not extend over
the entire 29-year period), visibility decreased at 7
locations, remained relatively constant at 8, and
improved at 1 during the period from 1948 through 1970.
Since 1970 visibility has improved at 12 locations,
has stayed relatively constant at 3, and has decreased
at 1. Thus, the data suggest that pollution control
during the 1970s has reversed a downward trend in visi-
bility that was observed in many western locations during
the 1950s and 1960s.
> Visual range decreases significantly with increasing
relative humidity at all 18 NWS stations except Phoenix;
however, the dependence of visual range on cloud cover is
less dramatic at most locations. The trend of dependence
of visual range on relative humidity agrees with our under-
standing of the influence of hygroscopic particle growth
with increasing relative humidity on the light scattering
coefficient.
-------
33
At most locations reductions in visual range are corre-
lated with low barometric pressure, which probably
results from the occurrence of high relative humidity
during lows. However, at Salt Lake City, Ely, Grand
Junction, and Pueblo, reduced visual range was corre-
lated with high pressure conditions. Since high pres-
sure systems sometimes cause air stagnation, the reduced
visual range during highs at Ely and Salt Lake City
(which are near copper smelters emitting SO ) may be
A
the result of air stagnation.
The stratification of visual range data by ventilation
(the product of mixing depth and average wind speed in
the mixed layer) indicates that reduced visual range is
correlated with reduced ventilation (stagnation conditions)
in Phoenix, Salt Lake City, Tucson, Billings, Cheyenne, Ely,
Ft. Huachuca, Grand Junction, Great Falls, and Rock Springs.
Reductions in visual range were found to be correlated with
high ventilation in Denver and Las Vegas, possibly because
urban pollutants were transported into the sight path between
the city and nearby mountains used as visibility markers.
At 11 of the 18 locations, visual range decreased signifi-
cantly with increasing surface wind speed, suggesting that
windblown dust, particularly at wind speeds greater than
10 m/s, contributes to visibility impairment. This effect
might explain the decrease in visual range with increasing
ventilation observed in several locations, particularly
Denver, Las Vegas, Alamogordo, and Winslow.
At most locations there are no significant seasonal varia-
tions in visual range. Denver, Las Vegas, and Colorado
Springs have minimum seasonal average visibilities in the
summer. At Phoenix, Salt Lake City, Ely, Grand Junction,
and Rock Springs, minimum visibilities occur during the
winter.
-------
34
> Variation in visual range with time of day is negligible
at most locations. However, at some locations, such as
Billings, Cheyenne, Farmington, Great Falls, and Rock
Springs, visibilities at midday (1000 to 1400 hrs) were
slightly greater than at other times.
> Visual range varies significantly with wind direction at
Denver, Las Vegas, Phoenix, Colorado Springs, Farmington,
Prescott, and Winslow. This relationship can be explained
by the geographical locations of emissions sources relative
to the sight paths of the NWS observers. Reductions in
visual range in both Billings and Great Falls, Montana, are
associated with generally easterly flow; the cause of this
effect is unknown. The significant reductions in visual
range in Prescott, Hinslow, and Farmington that are associ-
ated with winds that transport smelter S0? emissions directly
to the given location were not observed during the copper
strike period of 1967 to 1968. During that period, visual
range increased significantly throughout the Southwest.
To summarize, reduced visual range has been found to be associated with:
> SO emissions from copper smelters
/\
> High relative humidity
> Strong winds (windblown dust)
> Precipitation
> Stagnation conditions (low ventilation).
-------
35
III THE ELEMENTS OF VISIBILITY MODELS
This chapter discusses the elements of visibility modeling that are
common to both the plume and the regional visibility models. The specific
models and their outputs are discussed in Chapter IV. Appendices B and C
discuss the modeling of atmospheric optics and sulfate formation in greater
detail. Appendices D and E describe the plume visibility model and output;
the regional visibility model and output are discussed in Appendices F
and G.
As shown schematically in Figure 13, modeling of visibility impairment
requires mathematical descriptions for the following physical and chemical
atmospheric processes in succession:
> Emissions.
> Atmospheric transport, diffusion, and removal.
> Chemical and physical reactions and transformations
of precursors in the atmosphere.
> Light scattering and absorption characteristics of
the resultant aerosol.
> Radiative transfer through the aerosol along differ-
ent lines of sight.
Because visibility models are based on atmospheric dispersion and
chemistry models, the accuracy of the former depends on that of the latter.
We recognize that future improvements in modeling dispersion—particularly
on the regional scale and in complex terrain—as well as improvements in
secondary aerosol formation will increase the accuracy of visibility models.
A. POLLUTANT TRANSPORT, DIFFUSION, AND REMOVAL
We have developed two visibility models based on two basic types of
dispersion models:
------- An error occurred while trying to OCR this image.
-------
37
> A near-source plume model designed to predict the
incremental impact of one emissions source (such as
a power plant or smelter) at distances from 1 to 350 km
downwind.
> A regional model designed to predict, over time per-
iods of several days, the impact of several emissions
sources within a region with a spatial scale of 1000 km.
Calculation of near-source visual impacts requires a basic plume model
that accurately predicts the spatial distribution of pollutants and the
chemical conversion of NO to N09 and SO and NO to sulfates and nitrates.
£ A A
The plume model must be capable of handling the spatial scale from emis-
sions at the source to at least 100 km downwind. Because the regional-scale
problem may be caused by long-range transport of pollutants over a spatial
scale of 1000 km, an air quality model is needed that can account for mul-
tiple sources and for temporal variations in mixing heights, dispersion
parameters, emission rates, reaction rates, and wind speed and direction.
In the following subsections, we discuss atmospheric dispersion modeling
as it relates to visibility modeling. In addition, we classify the spatial
scale of the models in further detail:
> Initial dilution in a buoyant plume
> Gaussian plume diffusion
> Limitations on mixing
> Plume trajectory box model
> Regional transport and diffusion.
1. Initial Dilution in a Buoyant Plume
Modeling of the initial dilution of a plume from the top of the stack
to the point at which final plume rise is achieved is important when model-
ing the conversion of nitric oxide to nitrogen dioxide in a power plant
-------
38
plume because of the quick quenching of the thermal oxidation of NO. Because
the rate of this reaction is second order with respect to NO concentrations,
the rate is fastest in the initial stages of plume dilution. It is also
important to account for the initial dilution of buoyant releases because
the rate of dilution caused by the turbulent entrainment of ambient air by
the rising plume parcel can be considerably greater than that indicated by
diffusion coefficients based on measurements for nonbuoyant releases (e.g.,
Pasquill-Gifford a , a ).
Briggs (1969) suggested that the characteristic plume radius increases
linearly with height of the plume above the stack and can be represented as
follows:
Rp = 0.5 (Ah) . (9)
Briggs described the plume rise as a function of downwind distance (the "2/3
power law") as follows:
For initial dilution, we can assume that the plume is circular in cross
section:
The concentration of a given species at the centerline of the plume can be
calculated by a modified Gaussian model that can be represented as:
where V is the velocity of the parcel, which has a horizontal component
(the wind speed u) and a vertical component w, which can be calculated
by differentiating Eq. (10) as follows:
-------
39
With this formulation, time-dependent plume temperature and NO concentra-
tions can be calculated for accurate prediction of the thermal oxidation
of NO during plume rise.
2. Gaussian Plume Diffusion
After the plume has achieved its final height (about 1 km downwind),
plume concentrations for uniform wind fields can be adequately predicted
using a Gaussian model if the wind speed u at plume elevation H (h + Ah)
and the rate of diffusion are known for the particular situation so that
dispersion coefficients (o ,o ) can be selected:
Equation (14) is appropriate for a conservative species and can be modi-
fied to be appropriate for a nonconservative species by changing the
source term Q.
It is necessary for calculating plume visual impact to.integrate,
along the line of sight, the primary and secondary particulate and nitro-
gen dioxide concentrations. Equation (14) can be integrated (Ensor, Sparks,
and Pilat, 1973) in the cross-wind direction y from y = -°° to y = +°° to
obtain the optical thickness of the plume:
-------
40
TPy
+ exp - ±
(15)
where Q1 is the integral of the flux of plume extinction coefficient over
the entire plume cross section at downwind distance x. In the vertical
direction z, from z = 0 to z = +°°, the optical thickness is:
3. Observer-Plume Orientation
The magnitude of the visual impact of a plume depends on the orien-
tation of the observer with respect to the plume because the plume optical
thickness will vary depending on this orientation. Figure 14 shows plan
and elevation views of an observer and a plume and indicates that the sight
path distance through the constituents of the plume is a function of angles
a and e. The optical thickness at any angle a and any angle 3 can be deter-
mined as follows:
Figure 14 suggests that plume optical thickness is greater for horizontal
sight paths than vertical ones, particularly during stable conditions when
the plume cross section is flattened.
-------
41
(a) Plan View
%%^^//^^^^
(b) Elevation View
Source: Latimer and Samuelsen (1978),
FIGURE 14. GAUSSIAN PLUME VISUAL IMPACT MODEL:
OBSERVER-PLUME GEOMETRY
-------
42
4. Limited Mixing
When vertical diffusion is limited by a stable capping layer, Eq. (14)
is no longer valid, and a Gaussian formulation with terms for reflection
from the top of the mixed layer (at altitude H ) is more appropriate. In
this instance of limited mixing, the plume concentration becomes uniformly
mixed in the vertical direction for 0 < z < H and:
~ — m
TT
'
• s (tfl
The calculation of plume optical thickness in the y-direction becomes simply:
Equation (19) suggests a simple box model for calculating the reduction in
visual range of a plume after it has become uniformly diffused in the ver-
tical direction, which we discuss in the following subsection.
5. Plume Trajectory Box Model
The reduction in visual range resulting from a plume that is uniformly
mixed in the vertical direction can be evaluated simply, as Figure 15
illustrates. Latimer and Samuelsen (1970) showed that reduced visual range
resulting from pollutants in a plume is given by:
r =
v bext-0
[3.912 - tofyf} - J . (20)
When plume material does not significantly affect the intensity of the
horizon sky, the second term in the brackets drops out. After noting that
the background visual range (without the plume) is:
-------
43
Q
O
X
O
CO
o
o
Ul
'-3
LT>
ex.
C3
-------
44
3 912
=
we can rearrange Eq. (20) as follows:
rv= rvo(] -1*7) ' (22)
The fractional reduction in visual range then becomes:
Q'(x)
3.912 uH
where Q' is the integral of the flux of the plume extinction coefficient
over the total plume cross section at downwind distance x.
To illustrate the use of Eq. (23), we evaluated the percentage reduc-
tion in visual range resulting from three S02 emissions sources, assuming
average afternoon ventilation for Winslow, Arizona (u = 6.7 m/s, Hm = 2613 m),
a sulfate formation rate of 0.5 percent per hour, and a b___^-to-mass ratio
1 3
of 0.04 m /yg/m of S0| and persistent meteorological conditions. Three
S02 emissions rates were selected: 2000 tons per day (representative of
a large copper smelter complex), 200 tons per day (representative of a
large coal-fired power plant without S02 scrubbers), and 20 tons per day
(representative of a large coal-fired power plant with 90 percent S02
control). The results of this calculation using this simple model are
shown in Table 2.
-------
45
TABLE 2. PERCENT REDUCTION IN VISUAL RANGE BY
THREE LEVELS OF S02 EMISSIONS
Downwind
Distance
(km)
50
100
200
500
1000
2000
Tons/Day
1.9%
3.8
7.5
18.1
34.5
200
Tons/Day
0.2%
0.4
0.8
1.8
3.5
20
Tons/ Day
0.0%
0.0
0.1
0.2
0.3
As the sample calculations summarized in Table 2 illustrate, the most
significant reduction in visual range from anthropogenic omissions may not
occur close to the source, but rather, at locations hundreds of kilometers
downwind from the source. However, the values in Table 2 are examples for
average conditions and an assumed sulfate formation rate. Equation (23)
makes clear that visibility impairment is a function of the ventilation and
the total quantity of light scatterers and absorbers in the plume. It should
be noted that these calculations ignore the effect of SO,, and sulfate removal
processes that decrease the sulfate concentrations and thereby reduce visi-
bility impairment. (Such processes are considered in the plume and regional
visibility models.) As demonstrated by these simple calculations, secondary
aerosol formation has a significant effect on the magnitude of impairment
for given emission and meteorological conditions. Equation (23) is a simple
calculation through which a first-order estimate of an emission source's
impact on visual range can be made.
-------
46
6. Regional Transport and Diffusion
The modeling of visibility impairment on a regional scale requires the
use of a model that accounts for time-dependent meteorological conditions,
multisource emissions, and pollutant removal through surface deposition,
washout, rainout, and the like. We selected the SAI Northern Great Plains
grid model (Liu and Durran, 1977) as the basis for our regional visibility
model. Appendix F describes the model in detail.
The model is composed of two interconnected submodels: a mixed layer
model and a surface layer model. The mixed layer model treats transport and
diffusion above the surface. A grid approach is adopted to facilitate the
handling of multiple sources and complex chemistry. The major feature of
this model is the assumption that pollutant distribution is nearly uniform
in the vertical direction. With this assumption, a simplified form of the
general atmospheric diffusion equation can be invoked.
The surface layer model calculates the pollutant fluxes lost to the
ground. The surface layer, a shallow layer immediately above the terrain,
is embedded within the mixing layer. If pollutants originate from either
elevated sources or distant ground-level sources, most of the pollutant mass
is contained in a layer aloft, i.e., in the mixed layer. The removal pro-
cesses consist of the diffusion of the pollutants through the surface layer
to the ground, followed by absorption or adsorption at the atmosphere-ground
interface. A unique feature of the surface layer is its diurnal variation
in surface temperature, which is a result of daytime heating and nighttime
cooling. This variation affects the vertical pollutant distribution through
atmospheric stabilities, and, consequently, the rate of surface uptake of
pollutants.
The regional model assumes vertical homogeneity in the concentration
distribution. One of the reasons for this assumption is that the vertical
diffusion term, based on dimensional analysis, is about 100 times greater
-------
47
than the transport term, and the horizontal diffusion term is only a frac-
tion of the transport term.
Assuming that the concentration distribution in the vertical is nearly
uniform below the base of the temperature inversion, a vertically averaged
concentration of species i may be computed from the following equation:
3C". ZC* 3 C",
3
- (Cj - c0.)-D.c(D) , 1-1. 2,..., N , (24)
where
CQ. = the background concentration of species i,
IT, v = the vertically averaged horizontal wind
components ("u *jf u dz/H, v =/Q v dz/H),
D = the two-dimensional divergence [D = (3i7/ax)
+ (9v/3y)],
?(D) = a step function defined by
n for D > 0
c(D) = (25)
' 0 for D £ 0 ,
S^ R. = rate of change of concentration of species i due
to emission sources and chemical reaction.
In the derivation of Eq. (24), the following assumptions were made:
> Deviations from the average concentration, c~. , in the
vertical direction are small.
> The vertical velocity at the top boundary is approxi-
mately given by:
-------
48
(26>
> The diffusive flux of pollutants at the top boundary is
negligible.
> The following relationships hold for the reaction and
source/sink terms:
R-- (c, ,C£ t - • • »c,.) = R. ic-j »Cp > « • • »c., ) (27)
S.(c.) - S.^) . (28)
B. ATMOSPHERIC CHEMISTRY
The visibility models require calculations of the conversion of pre-
cursors to species that cause visual effects, namely, nitrogen dioxide and
secondary aerosol (sulfate, nitrate, organics).
1 . Conversion of NO to NOp
As previously noted, nitrogen dioxide gas causes a yellow-brown dis-
coloration of the atmosphere. Although some discoloration can be caused
by wavelength-dependent light scattering by submicron aerosol, we demon-
strate in Chapter IV that the dominant colorant of power plant plumes is
N02 and that yellow-brown discoloration may be apparent at significant
distances downwind of large coal-fired power plants, particularly in areas
where the background visual range is excellent.
Very little N02 is emitted directly from power plants. However, color-
less nitric oxide is formed by the thermal oxidation of atmospheric nitrogen
at the high temperatures experienced in the combustion zone (the boiler in a
power plant) and the oxidation of nitrogen that may be present in the fuel.
Chemical reactions in the atmosphere can form sufficient NOn from NO to cause
-------
49
atmospheric discoloration. Available measurements of NO and NCL concen-
trations in power plant plumes in nonurban areas suggest that the conver-
sion of NO to NOp can be calculated from a simple set of three reactions.
The first of these is the thermal oxidation of NO to N02:
2NO + 02 •* 2N02 . (29)
The reaction is termolecular but bimolecular with respect to NO; it is
therefore very fast at high concentrations of NO but slow at the lower con-
centrations that exist in the atmosphere or in a plume. The reaction rate
for Eq. (29) [based on Baulch, Drysdale, and Home (1973)] is:
d[N02]
dt
x 10"12 exp(1^6fl[NO]2[02] ppm/s . (30)
The reaction with ozone also affects the conversion of NO to NO,,:
NO + 03 -* N02 + 02 (31)
The reaction is fast,with a rate (Leighton, 1961; Davis, Smith, and
Klauber, 1974; Niki, 1974) at 25°C of:
d[NO?]
—g^- = 0.44 [N0][03] ppm/s . (32)
This reaction accounts for the ozone depletion measured within power plant
plumes and is important because ozone concentrations can be high even in
nonurban regions. Measured ozone concentrations in nonurban areas of the
western United States range from 0.02 to 0.04 ppm.
Whereas the thermal oxidation rate [Reaction (30)] decreases as the
plume mixes (because the NO concentration decreases), the formation of
-------
50
nitrogen dioxide via Reaction (31) is enhanced as the plume mixes, because
additional ozone from the atmosphere is mixed into the plume, allowing
Reaction (31) to proceed. When there are no reactions converting N02 to
NO (e.g., at night), Reaction (31) proceeds until all of the NO in the plume
is converted to N02 or until the ozone concentration in the plume drops to
zero. Therefore, the rate of conversion of NO to NOp via Reaction (31) is
limited by the rate of plume mixing that provides the necessary atmospheric
ozone.
To complete the set of chemical reaction mechanisms, we must consider
the photodissociation of NOp. When sunlight illuminates a plume containing
nitrogen dioxide, short wavelength light and ultraviolet radiation are
absorbed by the NO,,. As noted above, absorption of the shorter wavelength
light produces the characteristic yellow-brown color associated with N02-
Absorption of the more energetic ultraviolet light (x < 0.4 pm) results in
dissociation of the N02 molecule:
N02 + hv -*• NO + 0 . (33)
0 + 02 + 03 . (34)
Leighton (1961) gave the rate of Reaction (34) as:
[N0] ppm/s , (35)
where K. depends on the amount of light incident on the nitrogen dioxide.
Davis, Smith, and Klauber (1974) gave the following expression for K, as
a function of the solar zenith angle X:
K. = 1 x 10~2 exp(- 7^|) s"1 . (36)
\ /
With this set of chemical reactions, the chemical conversion of NO to
N02 in the atmosphere can be calculated from background pollutant concen-
trations and from plume NOV increments using the technique suggested by
-------
51
Latimer and Samuelsen (1975) and White (1977). Making the steady-state
approximation, we have:
K
[N02] = / [N0][03] , (37)
where
[NO] = [NO ] - [NOJ (38)
A C.
and
[03] = [03]b - ([N02] - [N02]t - [N02]b) . (39)
Substituting Eqs. (38) and (39) into Eq. (37) we can solve for the concen-
tration of N00:
[N02] =0.5
[NOX] + [03]b + [N02]t
K
r
2
1/2
!
J
(40)
Using this formulation to compute NO to N02 conversion in a hypothetical
power plant plume, Latimer and Samuelsen (197G) studied the sensitivity of
N02 formation to the rate of plume dilution, background ozone concentration,
and solar radiation. The results of this analysis are presented in Figure 16.
This figure shows that thermal oxidation (e.g., [03] = 0) converts up to 10
percent of the plume NO to N02> and additional conversion results when ambient
ozone is mixed into the plume. A recent comparison of observations with cal-
culations using Eq. (40) indicates good agreement, particularly if the
-------
52
* * "2
8n ^ •«
Cb
OJ
•4-'
ro ^~^»
O CO
r-.
>> CD
J +> '-
c
a)
t/i
'a!
oo
(O
S-
O)
(O 'I-
I i 1 _\
co (O
OJ
o
13
O
co
U.
O
•a:
a:
:r
h-
O O
^;g
ii 11—«
z: o
a.
o:
cr
UJ
a.
O t—
i—i 2:
CO LU
C£ O
LU z
> o
g°
o uj
CsO
o rvi
z o
o o
o o
z a:
CQ
t-H O
CO
UJ i— i
co o
CD
-------
53
diffusion of the plume is correctly calculated by using fitted dispersion
coefficients based on plume diffusion measurements (see Figure 17).
2. Conversion of Gases to Particles
Although SOp, NO, and N02 gas do not scatter liqht appreciably, they
react in the atmosphere to form secondary sulfate and nitrate particles in
the size range that is most effective at light scattering (0.1 to 1.0 ;im).
In many situations, sulfate and nitrate are the dominant contributors to the
scattering coefficient of the atmospheric aerosol. It is essential, there-
fore, that the effect of these secondary aerosols and the rate at which they
are formed from precursors be included in visibility models.
During this first year of development work, we have concentrated on
developing the atmospheric optics and visibility impairment components of
the plume and regional visibility models. The development of a model to
predict the rate of formation of sulfate and nitrate aerosol based on the
fundamental chemical reactions and physical processes would be a tremendou:.
undertaking. Furthermore, even if such a model were available, we do not
know the background atmospheric concentrations of the species responsible
for the conversion of gases to particles.
We have therefore adopted the approach of calculating gas-to-particle
conversion using measured rates of reaction appropriate for the clean
(Class I) areas of the country. We recommend, however, that consideration
be given in future visibility model development work to the incorporation of
fundamental gas-to-particle reaction mechanisms in the visibility models.
The measured rates of conversion of the gaseous species to particles
vary over several orders of magnitude, depending on the amount of sunlight,
the concentrations of hydrocarbons, ammonia, manganese, iron, and other
chemical species, and the relative humidity. Available data suggest that
-------
54
-------
55
the rate of conversion of SCL to sulfate particles is much greater in
polluted urban atmospheres and in locations where relative humidity is
high. Appendix C summarizes the fundamental reaction mechanisms and
the available measurements of the rate of sulfate formation in the
atmosphere. Other recent reviews of sulfate formation include those
of Calvert et al. (1977), ER&T (1977), and Levy, Drewes, and Hales (1976).
Orel and Seinfeld (1977) have reviewed nitrate as well as sulfate informa-
tion, particularly in polluted urban atmospheres. For our initial visi-
bility model development work, we have modeled gas-to-particle conversion
using pseudo-first-order rate constants typical of clean areas. Levy,
Drewes, and Hales (1976), in their review of SOp oxidation in plumes, sup-
ported such an approach.
We calculate the conversion of S00 and NO to sulfate and nitrate as
2 x
follows:
9[S02]
~
3[NOl
. (42)
The integral of the flux sulfate and nitrate mass contributed by a given
emissions source (Qso?' QNOX) over the entire plume cross section at a given
downwind distance or transit time (t = x/u) can be calculated as follows:
Qso2 1 - e • (43)
where r^ and r2 are simply the ratios of molecular weights of sulfate to S02
(r, = 1.5) and nitrate to N02 (r2 = 1.35). The mass of the ammonium ion
(or other cation) and water associated with the sulfate and nitrate also
contributes to the total aerosol mass concentration.
-------
56
The effect of plume dilution on gas-to-particle conversion can vary
depending on the relative contribution of reactions with plume constituents
(e.g., catalytic oxidation of S0? on particle surfaces or in the liquid
phase by metal ions) and reactions with trace species in the background
atmosphere (e.g., the reaction of S02 and NOX with OH', HOA, HpO, NhL, and
RO^). For heterogeneous catalytic reactions on emitted primary particles,
the rate of aerosol formation decreases with plume dilution. For example,
the fractional rate of S02 conversion to sulfate becomes:
i 3[SO?]
-" CCatalyst3 ' (45)
2-
The concentration C of conservative plume species (e.n., precursor SOX and
NO or primary particulate) decreases with time after emission and can be
described mathematically (Schwartz and Newman, 1978) as:
C « t~n . (46)
Thus, if heterogeneous reactions with primary particles are the sole con-
tributors to secondary aerosol, one would expect a secondary aerosol forma-
tion rate that rapidly decreases with time after emission from the stack;
for example,
afSO^l
(47)
at
Furthermore, if a catalyst were used up, this reaction would decrease
even more rapidly and would eventually be quenched. However, the formation
rate of secondary aerosol by reactions with trace constituents of the back-
ground atmosphere (such as OH*, HOA, MHo) or background aerosol increase with
increasing plume dilution and will dominate at long reaction times. This
occurs because, with increased mixing associated with plume dilution, fresh
ambient air (containing the reactive species) is mixed into the plume. With-
out plume dilution, the reactive species are used up and no further conversion
takes place. The process becomes diffusion-limited in the same way N02 pro-
duction in a plume is limited by the rate at which ambient ozone is mixed into
the plume.
-------
57
At this stage in visibility modeling, we are using pseudo-first-order
rate constants to model the conversion of S00 and NO to sulfates and
2 x
nitrates. In our sensitivity analyses, described in Chapter IV, we selected
sulfate formation rates ranging between 0.3 and 1 percent per hour and, due
to the lack of quantitative data due, in part, to the uncertainty in measured
nitrate concentrations (Spicer and Schumacher, 1977), a nitrate formation
rate of zero.
C. AEROSOL SIZE DISTRIBUTION
To determine the visual effects of aerosols, one must specify or cal-
culate aerosol physical size and composition. This section reviews the
available data on atmospheric aerosol size distributions, particularly the
recent work of Whitby and his coworkers. We then discuss the types of
aerosols that are important for visibility considerations—background,
primary, and secondary aerosols--and how they are currently treated in
the visibility models.
Atmospheric aerosols can be grouped into a trimodal size distribution.
Figure 18 shows the three modes (nuclei, accumulation, and coarse particles)
and the processes that form them. According to Whitby and Sverdrup (1978):
The physical separation of the fine and coarse modes
originates because condensation produces fine parti-
cles while mechanical processes produce mostly coarse
particles. The dynamics of fine particle growth
ordinarily operate to prevent the fine particles from
growing larger than about 1 ym. Thus as a first
approximation, the fine and coarse modes originate
separately, are transformed separately, are removed
separately, and are usually chemically different.
Figure 19 shows the various ways of plotting size distribution infor-
mation. Figure 19(a), a number size distribution, is the plot that Junge
(1963) originally used for his data. This plot led to the aerosol power
law formulation:
-------
58
CHEMICAL CONVERSION
OF GASES TO LOW
VOLATILITY VAPORS
0.002
CONDENSATION GROWTH
OF NUCLEI
WINDBLOWN DUST
+
EMISSIONS
+
SEA SPRAY
+
VOLCANOES
+
PLANT PARTICLES
0.01
1 2
Particle Diameter (pm)
10
100
TRANSIENT NUCLEI OR
AITKEN NUCLEI RANGE
ACCUMULATION
RANGE
FINE PARTICLES
MECHANICALLY GENERATED
~ AEROSOL RANGE
COARSE PARTICLES
Source: Whitby and Sverdrup (1978).
FIGURE 18. SCHEMATIC OF AN ATMOSPHERIC AEROSOL SURFACE AREA
DISTRIBUTION SHOWING THE PRINCIPAL MODES, SOURCES
OF MASS FOR EACH MODE, PROCESSES INVOLVED IN
INSERTING MASS IN EACH MODE, AND REMOVAL MECHANISMS
-------
59
ro
I
C)
o
JO6
I05
I04
I03
I02
10'
ID'1
JO'2
IO'3
ID-
d log Dp
r*«0.99
0.001 0.01 O.I
1.0
10 100
D (UN)
(a) Power Function Fitted to the Number
Distribution over the Size Range
0 to 32 pm
FIGURE 19. AVERAGE URBAN MODEL AEROSOL DISTRIBUTION
PLOTTED IN FIVE DIFFERENT WAYS
-------
60
> N
0) c/i
rO C
-P (O
0)
Q.
999
99
90
70
50
30
10
I
0.1
06VT «1.62/1
-------
61
d log D " cDp ' (48)
where N is the number of particles of diameter Dp. Figure 19(b) shows the
standard lognormal plot. Figures 19(c), 19(d), and 19(e) show the number,
surface area, and volume distributions of an urban aerosol. These last
three plots show the three distinct aerosol modes: nuclei (0.001 to 0.1 ym),
accumulation (0.1 to 1.0 ym), and coarse (>1 ym).
Whitby and Sverdrup (1978) recently computed a table showing typical
aerosol size distributions and concentrations based on measurements in many
locations. Table 3 summarizes the mass median diameters (DG), the geometric
3 3
standard deviations (o), and the volume concentration (V in ym /cm , which
3 3
is equal to the mass concentration in yg/m , divided by density in g/cm ) of
each of the modes for eight classifications of aerosol. Table 3 shows only
a small variability in the accumulation and coarse mode size distributions
for the variety of aerosol types measured (excluding the marine aerosol).
The average specifications for the accumulation mode and the coarse distri-
butions are summarized below:
Mode DG °q
Accumulation 0.29 ± 0.06 ym 2.0 ± 0.1
Coarse 6.3 ± 2.3 ym 2.3 ± 0.2
The data suggest that the average specifications (mass median diameter and
standard deviation) of the accumulation and coarse modes fit a wide range
of atmospheric conditions, a finding that greatly simplifies the calculation
of the scattering properties of the atmosphere.
Figure 20 shows the results of our calculations of the scattering
coefficient-to-volume ratio for a variety of different aerosol size distri-
butions. The calculations are based on Mie theory. The average specifica-
tions and corresponding scattering properties of the accumulation and coarse
modes are also given in Figure 20. As this figure shows, the accumulation
mode aerosol is roughly an order of magnitude more effective per unit volume
than the coarse mode aerosol.
-------
62
o
co
o
o;
ill
LLJ
o
CO
£
> ^ °.
CO CM
^
C7 P*^
o
CM
-H o
CJ3 E 1
Q 3. CM
* 1 •""
^««
(U
T3
g
C
0
•f-
ro
^—
3
3
U
U
**
OO
6 o
E 0
a
x_^
^0
CM*
,-j
o E| ro
Q 3.
^J o
1
! _
Qi
•o
O
£
•r—
"o
3
zn
co
E in
o o
=• --i o
co o
E
3. 0
Hvo
-
1 O^
—J •—
0 C] 0
-fj 0*
r^
O
VI
0 oj
s- o
a> re
«*
fv^
CM
CM
r—
•
CM
Uf)
•
•*
O
•
**t"
cS
vo
CO
•
o
Q}
CM
O
•
0
r—
CO
CM
O
O
•o
CO
CM
O
•
CM
«3
•
Lf>
CM
O
CO
CM
Lf)
CM
•
O
CM
10
»
O
•
r—*
r-.
CM
O
O
•a
C V)
to o
t)
"O >-
C 3
3 O
0,0
CHr—
J^ (O
o o
to o
CQ i—
CO
o
CO
r_
CM
•
CM
r-.
•
UO
«3-
CO
co
CM
CM
CO
•
0
CO
10
o
CO
CO
CO
0
0
(U
en
fO
S-
0)
O
^T
CM
IT)
•
CM
LO
•
LT>
O
CM
r—
CTi
•
00
r—
•
o
r—
•
o
•
p—
IT)
i—
o
o
H^
vo
f*s.
CM
^—
(U
E
3
r>_
Q.
>v-
* •+-
co"
f^
r—
v_
00
T3
C
re
_>>
+->
.c
(U
S
3
o
CO
-------
63
CJ
Q
CO
t—t
o:
co
»—i
a
LU
M
to
OO
Z3
O
t-H
a:
-------
64
If the scattering coefficient corresponding to each of the modes for
the different aerosols sampled by Whitby and Sverdrup (see Table 3) is cal-
culated, the accumulation mode is the dominant scattering mode, with the
coarse mode contributing a small amount and the nuclei mode a negligible
amount. For the clean continental background aerosol (second entry in Table
3), we calculated that the total scattering coefficient (b ,) equaled
_4 _i scat
0.23 x 10 m , which is very close to the minimum scattering coefficient sug-
gested by the data shown in Figure 9 (Chapter II). The accumulation mode
contributes 0.09 x 10 m (39 percent of bc^a.), the coarse mode 0.04 x
A 1 scat . -1
10 m (17 percent of b$cat), and Rayleigh scattering 0.10 x 10 m (43
percent of bsca+)- In the clean background case, the contribution of the
accumulation mode to the extra extinction (b ) above Rayleinh was 69 per-
cent of the total. This clean background aerosol corresponds to a visual
range of 170 km, or 105 miles (3.912/b .). For the average background
" _A -i
aerosol (third entry in Table 3) the computed b is 0.57 x 10" rrf ,
scan
corresponding to a visual range of 69 km. In this case, the accumulation
4 -1
mode contributed 0.27 x 10" m (46 percent of b ,), the coarse mode
4 i scat 4 i
0.21 x 10 m (36 percent of bscat)» and Rayleigh 0.10 x 10 m (17 per-
cent of b,.,.). These calculations indicate that for background conditions
scat
the accumulation mode is a larger contributor to light scattering than the
coarse mode but that the coarse mode is a nonnegligible component of the
scattering coefficient of the background atmosphere. This situation strik-
ingly contrasts with that of polluted urban atmospheres, in which the
accumulation mode causes more than 90 percent of the total scattering coef-
ficient. We should also note that the average background presented in
Table 3 is not as clean as the average background in the nonurban western
United States, as Figure 9 suggests.
In our visibility models, we have used the size distributions given in
Table 3 for specifying ambient background (second and third items in Table 3)
and plume aerosol (eighth item). We then calculated optical properties of
the aerosol (using Mie theory) from the computed concentrations of coarse and
accumulation mode (sulfate, nitrate, and associated cations and water)
aerosol.
-------
65
D. ATMOSPHERIC OPTICS
In the atmospheric optics component of the visibility models, the
light scattering and absorption properties of the aerosol and the resultant
light intensity for various illumination and viewing situations are computed.
The details of these calculations are given in Appendix B; the major points
are summarized in this section to give the reader an overall view of the
process.
1. Calculation of the Scattering and Absorption Properties
After the concentrations of the pollutants are specified by the trans-
port and chemistry subroutines, their radiative properties must be deter-
mined. For N02, the absorption at a particular wavelength is a tabulated
function (Nixon, 1940) multiplied by the concentration. For aerosols, the
procedure is more complicated, however.
In general, a particle's ability to scatter and absorb radiation at
a particular frequency is a function of size, composition, shape, and
relative humidity. Because we wanted to be able to alter the size distri-
bution of both primary and secondary particles, we needed to be able to
compute the effect of particle size on the wavelength dependence of the
extinction coefficient and the scattering distribution function. The only
rational method of making this computation is to use the solution of
Maxwell's equations for scattering by a sphere, the so-called Mie equations.
To verify that these calculations were appropriate for atmospheric aerosols,
we compared them with the empirical correlations of scattering to mass and
found substantial agreement, as discussed below.
The calculations were performed using an IBM subroutine written by
J. V. Dave (Dave, 1970). The required inputs are the particle size param-
eter (ratio of the circumference to the wavelength of radiation), the index
-------
66
of refraction (real and imaginary part), and the number and location of the
scattering angles (between 0° and 180°). The output is the scattering and
absorption cross sections and the Stokes transformation matrix (Van De Hulst,
1957), which can be simply converted to the scattering distribution assuming
randomly polarized light. The scattering and absorption properties per par-
ticle are then summed over the particular size distribution in such a way
that as the size distribution changes so do the radiative properties.
Different types of empirical correlations have been made in recent years
relating particle scattering properties to particle mass. The property mea-
sured has been either the volumetric scattering coefficient, as measured by
an integrating nephelometer, or the extinction coefficient, calculated from
the observed visual range. Among recent discoveries is the conclusion that
the scattering properties of urban atmospheres correlate much better with
the submicron accumulation mode concentration than with total mass. A
second important development is that for most of the United States and
Europe, sulfates are generally a significant fraction of submicron accumu-
lation mode mass.
Many studies of the empirical correlations of scattering coefficients
to sulfate mass have been performed. Table 4 summarizes some of these
measurements. Similar tables have appeared in other reports, such as that
by Trijonis and Yuan (1977). The correlation coefficients for these rela-
tionships have been very high (0.7 to 0.9), supporting the dominant role of
sulfate in scattering. Of particular importance is the use, by many
researchers, of visual range data (e.g., airport visibility) derived from
actual, though somewhat imprecise, visual perception of objects in the atmo-
sphere. Thus, it appears that sulfates play an important role in visibility
impairment.
Calculations of scattering-to-volume ratios (see Figure 20) reveal that
the maximum theoretically possible value is about 0.06, as expressed in the
-------
67
TABLE 4. ESTIMATES OF EXTINCTION COEFFICIENTS PER UNIT MASS
Extinction
Coefficients
COO4 m~V(Mg/m3)]
Source
Regression models
(Trijonis and Yuan, 1978)
Regression models
(Trijonis and Yuan, 1977)
Dust storms
(Hagen and Woodruff, 1973)
Regression model
(White and Roberts, 1975)
Regression model
(Cass, 1976)
Calculations for a model aero-
Location
Chicago
Newark
Cleveland
Lexington
Charlotte
Columbus
Salt Lake City
Phoenix
(county data)
Phoenix
(NASN data)
Great Plains
Los Angeles
Los Angeles
0
Sulfates
0.04
0.03*
(0.02)
0.06*
0.08
0.07*
0.06
0.06*
0.11
0.11*
0.12
0.13*
0.04
0.04*
0.04
0.03
NC
0.07
0.16
0.09*
.05-0.10
Nitrates
(0.00)
(0.00*)
(0.00)
(0.00*)
(0.00)
(0.00*)
(0.00)
(0.04*)
(0.00)
(0.00*)
0.09
(0.06*)
0.13
0.10*
0.05
0.03
NC
0.05
(0.00)
0.05*
NC
Remainder
of TSP
(0.000)
(0.000*)
0.026
0.014*
(0.000)
(0.000*)
(0.000)
0.019*
(0.001)
(0.000*)
(0.000)
(0.001*)
0.004
0.004*
(0.000)
(0.000)
0.001
0.015
0.008
(0.004*)
NC
sol of (NHjpSOjj at 70% RH
(Waggoner fit al., 1976)
Regression model
(Waggoner et al., 1976)
Southern Sweden 0.05
NC
( ) = not significant at the 95 percent confidence level.
NC = not calculated.
*Based on nonlinear RH regression model, with insertion of average RH.
Source: Trijonis and Yuan (1978).
-------
68
*
units used in Table 4. However, the reported empirically determined
values range from about 0.03 to 0.1. This discrepancy was resolved by
including the mass of liquid water associated with the sulfate. Investi-
gators have found better correlations with scattering-to-mass ratios that
depend on relative humidity. These fits are of the form:
(49)
1 - RH
where c represents the scattering-to-mass ratio at 0 percent relative
humidity. The range of these values is between 0.02 and 0.04, which is
in agreement with theoretical values shown in Figure 20."*"
To account for the effects of relative humidity, we simply added the
amount of water absorbed by the sulfate particles:
MaSSSulfate + MassCation + MassWater
Then we used a formula from Winkler (1973) to account for the mass of water
as a function of relative humidity. Finally, we compared the dependence of
the scattering-to-mass ratio on relative humidity determined by Cass and
Trijonis with calculations using the following assumptions:
This was computed from the maximum value of 0.08 x 10"^ m~'/ym3/cm3 in
Figure 20 assuming the sulfate was associated with ammonium ion as
(NH4)2S04:
cm3 132
W
A
l .8 g(NH4)2S04A10 yg 96 gS04
= 0.06 x 10"4 m"1/(yg/m3 SOj)
1 If we. use an accumulation mode bscat/V = 0.06 x 10~4 nH/ym3/cm3, we obtain
bscat/(vg/m3 S04) ranging from 0.034 to 0.046 x 10~4 m~Yyg/m3 depending
on whether the sulfate is H2S04 or (NH4)2S04.
-------
69
> Aerosol with a lognormal size distribution and a mass
median diameter of 0.2 ym and a geometric standard
deviation equal to 2.0 at RH = 0 percent.
> Index of refraction equal to 1.5 - Oi (typical, non-
absorbing aerosol).
> Density equal to 1.8 g/cm .
> Light of 0.55 pro wavelength.
> Sulfate as NH4HS04 (molecular weight of 115).
Figure 21 shows the striking agreement between our calculations and the
dependence of scattering-to-mass ratio on relative humidity observed by
Trijonis in the Southwest. This agreement gives support to our calculation
method for the scattering properties of the secondary aerosol.
The computational procedure used in the visibility models then takes
the input size distribution, assuming that the particles are spherical
with an index of refraction of 1.5 - Oi, and computes the scattering pro-
perties of the aerosol as a function of wavelength from the Mie equations.
The background size distribution properties are taken from the Whitby and
Sverdrup model of clean continental aerosols (see Table 3). The accumula-
tion mode in the plume is assumed to be the size of those measured in the
plume downwind of the Labadie power plant near St. Louis. The properties
of the sulfates and other accumulation mode particles are assumed to change
with relative humidity, as discussed above. A more complete inclusion of
relative humidity would require a modification of the refractive indices and
a recomputation from the Mie equations. This modification could easily be
done later if desired.
The limitations of the process of specifying the radiative properties
of aerosols are the usual ones: uncertainties in the size distribution,
nonspherical particles, and ambiguities in mean refractive indices. How-
ever, the close agreement shown for the sulfate scattering-to-mass values
suggests that the errors are not large.
-------
70
0.14
0.12
0.10
o
GO
CO
I, 0.08
2 0.06
o
CO
0.04
0.02
0.20
0
0 CALCULATED VALUES
*=> TRIJONIS AND YUAN (1977)
I
0.40 0.60
Relative Humidity
0.80
FIGURE 21. RATIO OF LIGHT SCATTERING TO MASS AS A FUNCTION
OF RELATIVE HUMIDITY
-------
71
2. Calculation of Light Intensity
p
The light intensity (watts/m /steradian) at a particular location in
the atmosphere is a function of the direction of observation n and the wave-
length >. Calculation of the light intensity in a medium follows from the
radiative transfer equation. This equation is a conservation of energy
statement that accounts for the light added to the line of sight by scat-
tering and the light lost because of absorption and scattering. Approxi-
mations and solution techniques applicable to planetary atmospheres have
been discussed by Hansen and Travis (1974) and Irvine (1975).
The physical situation that we are concerned with is shown schemati-
cally in Figure 22. To compute the spectral light intensity at the
observer, we sum (integrate) the scattered and absorbed light over the
path, r, associated with the line of sight n. The resultant general
expression for the background sky intensity at a particular wavelength is
'T
e' dT'
where
r
T = the optical depth (T E^ b . dr, where
is the extinction coefficient),
w = the albedo for single scattering (w = bscat/bext
where b . is the scattering coefficient),
-»• n) = the scattering distribution function for the
angle n1 -> jj,
I = the spectral intensity at T' from direct and
diffuse solar radiation.
-------
72
o.
oo
D_
cc
o
oo
5
CJ
OJ
C3
-------
73
Equation (51) is valid for the usual continuum, no refraction, random
polarization assumptions.
The intensity seen by an observer in direction n of an object at dis
tance R is:
WB) -
) dfi1 e"T dT' . (52)
'=4TT
Equations (51) and (52) then completely describe the spectral intensity
of the background and an object. Once these two quantities are known, the
visual effects of the intervening atmosphere can be quantified. In evaluat-
ing Eqs. (51) and (52), we encounter two main difficulties: First, the
quantity in the integral is a fairly complicated function, and accurate
specification is tedious. Second, the atmosphere is inherently inhomogen-
eous, and thus, the radiative properties w, p are somewhat complicated
functions of r and n. Approximations are therefore necessary. Appendix B
outlines in some detail the approximations we have used; we present only a
summary here. The approximations we used are the following:
> Plane parallel atmosphere.
> Two homogeneous layers.
> Average solar flux approximation.
> Average diffuse intensity approximation.
The equation for the background intensity at the surface becomes, for a
direction y, »
-------
74
_ Tdi f /n 0°° \ / r i\
WOD !av U - e ) ' (53)
and for the intensity in the direction of an object in the planetary bound
ary layer,
where
o' , Pnn(o) = the average albedo and phase function
respectively,
Tnn = ^^e °Ptica^ depth of the path in the boundary
layer,
F_ , I = the average solar direct intensity and diffuse
s ,av av
intensity, respectively,
I . , I = the intensities from the upper atmosphere and
S Kjr 0
object, respectively.
The exact definitions of the terms are given in Appendix B.
Thus, the background intensity and the intensity in the direction of an
object at distance R from the observer can be computed given the following
inputs:
> Background radiative properties (e.g., size distribution,
visual range).
> Solar zenith angle.
> Scattering angle.
> Direction of observation, n, 4>-
> Planetary boundary la^er height.
-------
75
The intensities, including the effects of air pollution, are computed
from essentially the same formulae with the radiative effects of the pollu-
tants included in the background atmosphere. In the regional model, the
intensity for a given optical path is calculated from an integration of the
concentration through the cells of the grid model. For an initial approxi-
mation, we used the expressions for a homogeneous atmosphere.
In the plume model, it was necessary to treat the plume as a homogen-
eous layer with an optical depth and mean properties
-------
76
> Coloration of objects.
- Brightness
- Hue and saturation.
> Contrast and color difference between two objects.
- Black object and horizon sky (to calculate
visual range).
- Haze layers.
- Plume and background.
The perception of an object such as a distant mountain results from
changes in light intensity, coloration, or both. Visual range is defined in
terms of differences in light intensity (contrast) between a distant black
object and the horizon sky. Contrast is also a useful concept for charac-
terizing the appearance of plumes and haze layers. However, a plume may
be perceived against a background as the result of a color change unaccom-
panied by a change in light intensity (i.e., with no contrast). Ue there-
fore need a means of characterizing the perception of changes in both the
intensity and the coloration of light. We discuss the different means of
characterizing visibility impairment in the following subsections.
1. Visual Range
Visual range is defined as the farthest distance at which a black
object can be perceived against the horizon sky. As we have noted in
Chapter II, the threshold of perception of differences between the light
intensity of two objects has been characterized by a liminal contrast.
The value of the liminal contrast is commonly taken to be 0.02, as first
suggested by Koschmieder in 1924 (Middleton, 1952). However, the liminal
contrast is a function of the observer and his state of mind (e.g., fatigue,
attentiveness) as well as the intensity of the background lighting. Under
the best conditions, the liminal contrast may be as low as 0.005 (Committee
on Colorimetry, Optical Society of America, 1963). The Federal Aviation
Administration assumes a value of 0.055. Based on an experiment using 10
-------
77
observers and a total of 1000 observation hours, Middleton (1952) reported
a median of 0.03 and a mode of 0.02 for the liminal contrast. For the pur-
poses of standardization, it is reasonable to describe the perception of a
"standard observer" and to select and use a single value for the liminal
contrast. We used the Koschmieder value (0.02) for our calculations.
It is clear then that the observation of distant targets such as moun-
tains is not an accurate measurement of strictly defined visual range, i.e.,
the farthest distance at which a black object is distinguishable from the
horizon sky by a standard observer where liminal contrast is 0.02. This
is true not only because of the variability in the contrast threshold, but
also because distant markers such as mountains are usually not perfectly
black.
The contrast between two objects is defined as:
- I2(x) '
If the two objects are the same color [i.e., I,(A)/Ip(x) is constant
over 0.4 < A <0.7 pm], then the contrast at all wavelengths will be the
same. However, if the objects have different colors, then C is a function
of wavelength. For the calculation of visual range, we evaluate the con-
trast at a wavelength of 0.55 um, which is at the middle of the visible
spectrum and is the wavelength to which the human eye is most sensitive.
The intrinsic contrast of a black object (I, =0) against the horizon sky
(\2 - Ih) is -1; the visual range is the distance at which this contrast
is reduced by the light scatter and absorption of the intervening atmo-
sphere to -0.02. Thus, visual range can be evaluated by computing con-
trast iteratively as a function of distance from the observer until it
drops to -0.02. This approach is necessary if one is dealing with a non-
homogeneous atmosphere.
-------
78
For a homogeneous atmosphere, however, the calculation of visual
range is analytic, using the Koschmieder relationship:
ext
For the computation of visual range through a homogeneous atmosphere con-
taining an optically thin plume, Latimer and Samuelsen (1978) suggested
the following simplified approach:
r =
3.912 - £n -^ - T I . (57)
v bext-0
The second term in the brackets is necessary to account for the effect of
light absorption caused by plume NCL on the contrast between the horizon
sky (IhD) and the black object seen through the plume. It can easily be
shown that the effect of the plume on visual range is significantly less
when the plume is discolored by NOo Uhp/Ih < 1), and greater when the
plume is bright Ohp/Ih > 1). As a corollary, it is also true that with
increasing distance between the observer and the plume, the impact of
plume N02 on visual range increases as plume coloration decreases. This
somewhat surprising result was confirmed in the sensitivity analysis of
the plume visibility code.
2. Contrast of Haze Layers and Plumes
Contrast can be used to characterize the perceptibility of a haze
layer or a plume against a background—the sky, a cloud, or a distant
mountain. A plume would be visible if the absolute value of the contrast
between it and the background were greater than a threshold or liminal
contrast. Figure 23 is a photograph of a plume illustrating plume con-
trast. The plume is clearly visible against the mountain because the
plume light intensity is greater than that of the mountain. Thus, the
contrast of the plume against the mountain can be calculated using Eq. (45):
-------
79
o
<
Q.
U.
o
ro
C\J
-------
80
C . J^L-J^o . (58)
H m
The plume is also visible against the horizon sky, perhaps mainly
because of the color change, but also because of contrast:
(59)
The magnitude and the sign of the contrast of a haze layer or plume against
a background is therefore a useful way to characterize visibility impair-
ment. Positive contrasts refer to plumes brighter than the background,
whereas negative contrasts refer to plumes darker than the background. We
do not have any experimental data for liminal contrast (the barely percep-
tible threshold contrast) in the case of a plume against a background. The
same liminal contrast used to define visual range (0.02) could be used to
define plume visibility. However, it seems likely that the liminal contrast
for plumes is greater than 0.02 because in many cases the boundary between a
plume and the background is not distinct owing to the nature of plume dilu-
tion. It would be useful to carry out some experiments with several
observers and plume views to determine the liminal contrast for an average
observer.
Contrast of plumes can be evaluated at several different wavelengths;
we used 0.55 ym for the evaluation of plume contrast. However, plume con-
trast may be greater at the blue end of the visible spectrum. Latimer and
Samuelsen (1975, 1978) used the ratio of plume to background intensities
at the blue end (x = 0.4 ym) and at the red end (x = 0.7 ym) as a means of
characterizing the wavelength-dependent plume contrast and plume coloration
with respect to the background. This blue-red luminance ratio is defined
as:
I (0.4 ym)/I. (0.4 ym) C, (0.4 ym) + 1
D = JL, _ ' " _ _ P _ (60)
K Ip(0.7 ym)/Ih(0.7 ym) Cp(0.7 ym) +1
-------
81
The use of the luminance ratio in conjunction with the plume contrast at
0.55 ym is a simple way of characterizing plume color. When R > 1, the
plume is more blue than the background; when R < 1, the plume is redder
(or more yellow-brown); when R = 1, with C (0.55 pm) > 0, the plume is a
brighter white than the horizon, and with C (0.55 pm) < 0, the plume is
a darker grey. We discuss more sophisticated methods of quantifying color
in the next subsection.
3. Color
The color associated with a given spectral light intensity distribution
is due to processes occurring in the human eye. The retina has three dif-
ferent frequency sensors that convert signals into color sensations by means
of the brain. The system operates so that an object that reflects half blue
light and half yellow light is identified not as yellow-blue, but rather as
a new color, green. This attribute of the eye-brain system gives rise to
another mode of detecting an object, that of color change or discoloration.
Thus, an object can be perceived because it has a different brightness from
that of the background (contrast) or because it has a different color (so-
called color contrast). Gases and particles in the atmosphere can give rise
to coloration by their scattered light (blue sky or white clouds) or by
altering the color of objects seen through them (brown coloration due to NO?).
The chromaticity diagram was developed to quantify the concept of color.
In such a diagram, the spectral distribution of light is weighted with three
functions corresponding to the detectors in the eye. For any distribution
of light, there are three numbers, which define a point in space. Next, the
projection of the point onto a unit plane (x + y + z = 1) is computed. The
result is a two-dimensional surface called a chromaticity diagram (see
Figure 24). Monochromatic light forms the outside of the surface, and white
light is located in the center. Any color can thus be represented by its
chromaticity coordinates (x,y), which are defined by:
-------
82
0.80 —
0.70 —
.0.555
>,0.4C—
0.00
GREENISH YELLOW
YELLOW
A58^-ORANGE YELLOW
.585
°-49 Y GREENISH
BLUE
0.00 0.10
REDDISH ORANGE
0.70 ym
0.70
FIGURE 24. CHROMATICITY DIAGRAM
-------
83
X=X+Y+Z ' y X+Y+Z
where
X = / I(A) x dA
A
Y =/ I(x) y dA
=J
I(A) Z dA
A
and I (A) is the wavelength distribution of light and x, y, z are the three
weighting functions. The weighting functions (called tristimulus values)
are shown in Figure 25.
Horvath (1971) and Husar and White (1976) computed chromaticity coor-
dinates of atmospheric scattered or transmitted light and showed that the
light would be distinguishable from white light for various sun angles,
aerosol properties, and NfL concentrations. Since the chromaticity dia-
gram does not differentiate between differences in intensity (e.g., between
yellow and brown or between white, grey, and black), chromaticity coordinates
must be used in conjunction with a descriptor of light intensity for a com-
plete specification of color. Thus, if we establish a color solid by taking
the two-dimensional chromaticity diagram and adding a third dimension per-
pendicular to this plane to represent brightness, we have a means of com-
pletely specifying by three coordinates the color and intensity of a color.
Figure 26 is a drawing of such a color solid. The brightness in such
a coordinate system is usually specified by the value of Y [see Eq. (61)]
or by a parameter (L*), which is directly proportional to the subjective
perception of brightness and is related to Y as follows:
-------
84
O)
3
a
3
I/)
•r—
$-
400
500 600
Wavelength, x (nm)
Source: Judd and Wyszecki (1975).
FIGURE 25. SPECTRAL TRISTIMULUS VALUES x(x), y(x), z(x)
-------
85
L* = 25 Y1/3 - 17
(62)
L* is used in quantifying color differences and is simply the parameter
called "value" in the Munsell color system multiplied by 10.
VALUE
(brightness)
CHROMA
(saturation)
Source: Munsell Color Company (1976).
FIGURE 26. REPRESENTATION OF A COLOR SOLID
The Munsell color system is the most widely used means of specifying
colors. In this system, colors are arranged in order by value (brightness),
hue (the shade of color, for example, yellow, red, green, blue), and chroma
or saturation (the degree of departure of a given hue from a neutral grey of
the same value). By specifying a given hue, value, and chroma, one can
obtain a sample color chip from the Munsell Book of Color that corresponds
to the specification. By this means, the objective specification of color
-------
86
(L*,x,y) can be related to the subjective perception of color by visually
examining the color paint chip. ASTM Standard D 1535-68 (American Society
for Testing and Materials, 1974) is the reference method for converting
objective color specifications (L*,x,y) to the Munsell hue, value, and
chroma notations by which a colored paint sample can be selected. We
used this method to convert light intensity (Y or L*) and chromaticity
coordinates (x,y) calculated by the plume and regional visibility models
to Munsell notation to be used by a commercial artist in illustrating
atmospheric discoloration. We discuss this process in Chapter IV.
4. Color Changes
The final step in the quantification of visibility impairment is the
specification of color differences—differences both in chromaticity (x,y)
and brightness (Y). In 1976 the Commission Internationale de 1'Eclairage
(CIE) adopted two color difference formulae by which the perceived magni-
tude of color differences can be calculated. Color differences are speci-
fied by a parameter A£, which is a function of the change in light intensity
or value (AL*) and the change in chromaticity (AX,Ay). AE can be consid-
ered as a distance between two colors in a color space that is transformed
in such a way that equal distances (A£) between any two colors correspond
to equally perceived color changes. This suggests that a threshold (A£Q)
can be found to determine whether a given color change is perceptible.
Since the CIE could not decide between two different proposed formulae
for AE, both were adopted in 1976 as standard means by which color differ-
ences can be specified. These color differences, which are labeled
AE(L*U*V*) and AE(L*a*b*), are calculated as follows:
AE(L*U*V*) * [(AL*)2 + (AU*)2 + (AV*)2]
where
-------
87
L* = 116 (Y/YO)I/S - is,
U* = 13W*(u - U0),
V* = 13W*(v - v0),
and u and v are defined as
- 4X _ 6Y
U / u ', l rw i l-> \ s V
and u~, v0 as
(X + 15Y + 3Z) ' v (X + 15Y + 3Z)
4XQ 6YQ
U0 " (XQ + 15YQ + 3ZQ) ' V0 " (XQ + 15YQ + 3ZQ)
Similarly,
AE(L*a*b*) = [(AL*)2 + (Aa*)2 + (Ab*)2]
where L* is defined as above and
Ky \1/3 , v \l/3-|
— I - (~\
A0/ \rO' J
.0 " (zn)
In these equations, the tristimulus values X0, YO, Zfi define the color of
the nominally white object-color stimulus. In our atmospheric discolora-
tion calculations, we used values of Xfl, Y,,, Z- corresponding to the
reflected intensity from a perfectly diffuse reflector normal to the direct
solar beam. Calculations are normalized such that Yn = 100.
To determine the liminal or threshold (just perceptible) value of AE,
we computed AE for two color fields with identical chromaticities
(AU* = AV* = Aa* = Ab* = 0) and with a contrast of 0.02 (i.e., Y2 =
as:
-------
V3, ,v J/3
- 0.981/3^ -
3) - 0.78 (
Thus, for a bright horizon (say, Y, = TOO), we obtain a threshold or liminal
AE equal to 0.78. This value can be compared with a AE = 10, which is the
difference between two colors having identical Munsell hue and chroma but
with values differing by 1. Thus, AE can be used as indicator of atmo-
spheric discoloration: AE's less than 1 would be imperceptible, those between
1 and 10 would be detected as a discoloration by most people, and the sever-
ity of discoloration would increase with increasing AE. More work is clearly
necessary to determine what the standards of atmospheric discoloration
should be.
-------
89
IV THE OUTPUT OF VISIBILITY MODELS
This chapter discusses the outputs from several sample calculations
using our regional and plume visibility models. The models are described
from the viewpoint of a person who will be faced with regulatory, siting,
and design decisions based on the output of such models. We provide
samples of graphic display alternatives that can be used to translate
quantitative descriptions of visibility impairment into color samples,
perspective views, artist's renderings, and color television video displays,
These display techniques enable the user to understand the meaning of
visibility impairment models. Further details of the models and sample
outouts are given in Appendices D, E, F, and G.
The following models are illustrated by examples:
> Plume visibility model.
- Emissions from a hypothetical 2250 Mwe coal-fired
power plant meeting New Source Performance Standards.
- Emissions from a large copper smelter in Arizona.
- Emissions from a large coal-fired power plant in
Arizona.
> Plume/terrain perspective and color graphic displays--
emissions from a large coal-fired power plant in Arizona.
> Regional visibility model.
- 1976 and 1986 SOV and NOY emissions from sources in
A A
the Northern Great Plains.
1973 SOV emissii
A
and New Mexico.
- 1973 SO emissions from copper smelters in Arizona
A
-------
90
A. THE PLUME VISIBILITY MODEL
The plume visibility model predicts the visibility impairment
resulting from emissions from a single source, such as a power plant or
smelter . The model calculates the reduction in visual range caused by
the plume for several observer locations, and it also calculates plume
color, plume contrast, and color changes to determine whether the plume
can be distinguished by an observer. In this latter regard, the plume
model differs from the regional model, which calculates the visual effects
of a relatively homogeneous atmosphere. The plume model quantifies the
coloration and appearance of a plume in comparison with the homogeneous
background atmosphere and thereby characterizes the perceptibility of the
plume. The logic flow, program structure, and data requirements of the plume
visibility code (PLUVUE) are presented in detail in Appendix D. In this
section, we illustrate and discuss sample outputs of the model.
The user of the model must provide the source emission parameters,
ambient meteorological conditions, ambient air quality, and background
aerosol size distribution parameters. Exhibit 1 lists the parameters of
the sample calculation done for the hypothetical 2250 Mwe coal-fired power
plant, which was assumed to emit particulates, S02> and NOX at the maximum
rates permitted by the EPA's New Source Performance Standards. The user
must also select the dispersion coefficients (oy, az) to be used to com-
pute plume dilution as a function of downwind distance. The code has
subroutines for Pasquill-Gifford and for TVA dispersion coefficients, and
it will also accept values entered by the user.
After computing the initial plume dilution and N0£ formation during
plume rise from the stack to the location of final plume rise (1.2 km
downwind), the code calculates pollutant concentrations within the plume
and parameters characterizing plume visual impact at distances from 1.2
out to 350 km downwind of the source. Exhibit 2 presents an example of
the pollutant concentration parameters that are printed out at each down-
wind distance. In that exhibit, both the plume increments and the total
-------
91
a
E
a
5
SB
74
-s
I I
• i
^ «
g
i !
b. fc
I
5?
i i **
§ 5 S
I
c
g
i
t
i
*
3
B P s
W®
••w
c;
I
a
E S
H- O
rs z
Qu O
i— o
O h-
Q CO
o z:
z: <
LU O
s:^
Z3
-------
92
I
-1
i P
i I
K
S>
I i
T3
-------
93
nt- «"-
**
»h^
O
O
O
o.
ace
n ic
D.
O
UJ
o
»i
e
gg
> >•
> i* ww Wrt tr tj co rs rort $ *~
"&i ** C5C5 »•*- COK1 <"<> V®
1J C4d W P5 C4 W CO C? W W 41 **5
>e coco -- -- -- coco i-
H
E
I
UJ
I g
a »
II
fw-
OC
-a. e£
4» »
«, . . . «,
' & *9 9 N ^ COCO
vW^*0 OiC cOrt
I 55 SS si 9S 55 II
I
I
I
P
CO
I—I
:c
x
13
!>-N3
J* S
5£«im££
JJ-rfeg ^
£^l
nr
«i
S
CL,
>•
-------
94
ambient concentrations are displayed. The mole ratios of sulfate to
total sulfur and NC^ and nitrate to total nitrogen are also displayed, as
are the concentrations of ozone and the plume ozone deficit resulting from
NCL production. The plume increment and total particle scattering coef-
ficient (bsp at 0.55 ym) are printed out, as is the percentage of b con-
tributed by secondary aerosol (SO^, NO^). Note that in the example given
in Exhibit 1, at 10 km downwind, sulfate--which is assumed to form at the
rate of 0.5 percent per hour—contributes 46-1 percent of the plume scat-
tering coefficient, the remainder of the scattering is due to the emitted
primary particulate matter (fly ash).
Exhibit 3 provides the first of the three visual effects printouts
that the user can choose to display for each downwind distance. These
computations are done for sight paths through the plume center and can be
done for ground-level sight paths as well. Visual effects can be dis-
played for scattering angles 0 selected by the user (22°, 45°, 90°, 135°,
and 180°); only 180° (back scatter) calculations are shown in Exhibit 3.
Visual effects are calculated as a function of assumed observer location
relative to the plume. Observer location is specified by the distance
(along the sight path) between the observer and the plume at distances
that are 2, 5, 10, 20, 50, and 80 percent of the background visual range
and at four azimuthal angles with respect to the plume centerline
(a = 30°, 45°, 60°, and 90°).
In Exhibit 3, the first parameters printed out are the visual range
rv and the percentage reduction from background visual range. The follow-
ing two columns are the light intensity parameters Y and L*, described in
Chapter III. The chromaticity coordinates (x,y) are then displayed,
followed by two columns showing the differences in light intensity
(AY, AL*) between the plume and the background sky (without clouds). The
negative values in this example indicate that the plume is darker than
the background sky. The plume contrast (at X = 0.55 ym) is shown next,
succeeded by the blue-red ratio. The change in chromaticity coordinates
between the plume and the background sky is shown next (AX,Ay). Positive
-------
95
§n n P?
. . .
>•
-»*CI ««S« —
-* eo g ^ « ee n N r- •»
-------
96
Ax's and Ay's indicate a shift toward yellow-brown, and negative values
indicate a shift toward blue relative to the horizon. The final two
columns are the CIE color difference values AE(L*U*V*) and AE(L*a*b*).
To understand how the values in Exhibit 3 can be used to characterize
plume color, consider an observer's sight path that is perpendicular to
the plume (a = 90°) at a distance rp/rvQ = 0.02. An L* of 82.23 indicates
that the plume is bright, but not as bright as the horizon because AY,
AL*, and C are all negative. The chromaticity coordinates (0.3181,0.3253)
used in conjunction with L* (which is 10 times the Munsell "value") specify
the Munsell color notation, which in this case is 2.5 Y 8.912/0.6, a
weakly saturated yellow, essentially grey. The blue-red ratio of 0.9194
also indicates a slight, but perhaps not visible, yellow discoloration.
However, the contrast of -0.1611 and the AE values of 10.6 and 6.8 indicate
that the plume would be visible because it is darker than the background
horizon sky.
Exhibit 4 shows the visual effects of the plume for nonhorizontal
sight paths when viewed against a background of blue sky. Note that
visual effects are calculated for the permutations of a (azimuthal angle
relative to the plume centerline) and elevation angle 8 (15°, 30°, 45°,
60°, 75°, and 90°). These calculations indicate that the plume is more
distinctly visible against the blue sky background than it was against
the horizon sky (Exhibit 2). Note also that the plume is much brighter
than the blue sky background because AY, AL*, and C are all positive.
Exhibit 5 completes the characterization of plume visibility at a
given downwind distance by comparing the light intensity of the plume with
white (representative of a white cloud or snowbank), grey, and black back-
grounds at various distances from the observer behind the plume. The
plume appears somewhat darker and bluish in front of the white object
(REFLECT = 1) and brighter than the black objects (REFLECT = 0) at close
distances. The plume appears slightly darker than black objects at long
distances because the apparent light intensity of the black object distant
from the observer approaches that of the horizon.
-------
97
3««e)«te + «aei*n««
®co«. — (8«w + NN + -+4<
c?**p)gioc3egcoi'?M
j
H
A
O
M
o
0.
I
o;
o
u_
oo
v 9 OIO
< i£''#1*
U « CO CO IT) » 9 <) * 03 C5
^ d f> t«- QCO » ••
I/O
S
I
|
g
* V ••* S
u CO «9 «• <» IO * « ^»b- K5 * * C- >
-j E w— ' ' ' '<>i- '-
o
------- An error occurred while trying to OCR this image.
-------
99
The large amount of output required to characterize the visual impact
of a plume at a given downwind distance is necessary because of the large
number of possible observer locations, sight path orientations, and back-
ground objects. We have an optional printout table for plume visual
effects for only one given observer location for user-contributed values
of rp and a. This output option is useful for describing a plume for an
artist's rendering, as discussed in Section B.
We have designed a computer plot package to display plume visual
effects for horizontal sight paths as a function of distance. Examples of
these computer plots are shown in Figures 27 through 35. Four parameters
were selected that most easily characterize (with numbers) the visual im-
pact of a plume:
> Percentage reduction in visual range.
> Blue-red ratio (plume color relative to background).
> Plume contrast (plume light intensity).
> AE (L*a*b*) (plume perceptibility).
We have selected these plots to show the visual effects seen by an observer
situated at distance rp = 0.02 x ryQ and with a horizontal sight path per-
pendicular to the plume center! ine.
Figures 27 through 32 show the results of a sensitivity analysis to
determine the effect on plume visibility impairment of:
> Plume diffusion.
- Distance downwind (1.2 km < x < 350 km).
- Atmospheric stability (Pasquill C, D, and E).
> Scattering angle (0 = 45°, 90°, 180°).
> Sulfate formation (0 and 0.5 percent per hour).
(0 and 0.7 lb/106 Btu NOX emission rate).
The effect on visual range of the 3.4 ton/day primary particulate
emission rate assumed here (less than the 0.1 lb/10 Btu standard in order
-------
100
60.0
SQ.O
40.0
30.0
20.0
10.0
0.0
1.1
B
ii.o
a
j
JJ0.9
j
a
0.8
0.7
0.1
-o.i
U
1-o.a
-0.3
15.0
10.0
S.O
0.0
(1! NBRHRL NBX EMHSiaNS; 0.5 PERCENTER ?;..,_--
(2) MBRMfiL HBX EMI3SI0NS; N« SULFHTE F0RHHTI0N
(3i HB HBV EMISSIONS: o.s PERCENT .'HI? SULFHTE FT
«) NB NBX EHISSIBN3: N8 SULFflTE F£lPMCiTI8U
I ill
10 20 40 60
OfMNHIND DISTANCE (KM)
100
20-
FIGURE 27. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING
ANGLE OF 45° AND STABILITY CLASS C
-------
101
60.0
z 50.0
r°'°
,,,30.0
SIlO.O
0.0
1.1
= 1.0
a
ui
^ 0.9
(11 NBRHSL N0X EHI33I0NS: 0.5 "ERCENT •-is 5!.!,_-;7- -jif|.!;T;ti|,,
(21 N8RMRL NBV EMI5SI0NS; N8 3ULFSTE FBRMPTI«-tj /
13) N« NBS EHIS31BNS: 0.5 PERCENT/HR TilJLffiTE "T -TlflN
(4) MB NST' EMfSSIBN?: KB SULFfiTE
0.8
8:1
-0.0
-0.2
-0.3
JS.O
10.0
ui
a
S.O
0.0
10 20 40
DMNHINO OIBTRNCE (KM)
100
200
FIGURE 28. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 45° AND STABILITY CLASS D
-------
102
60.0
(I) NBPHRL NBX EHI3SIBNS! 0.5 FEfCEfiT.'KS SJAF;,-::
(21 NBRHBL NBX EHISSIBNS: NB SULFflTE FBRMST;?'.!
(3) H8 N8X EHISSIBNS; 0.5 PERCENT/HR SULFflTE ff
(41 NB NBX EHISSIBNS: NB SULFflTE FBPHRTI3N
z 50.0 -
40.0 -
30.0 -
20.0 -
10.0 -
0.0
1.1
1.0
d.9
0.8
-0.0
-0.3
15.0
to.o
5.0
0.0
I I I I i I I
6 10 20 40 60 100
DWNMIND DISTANCE (KM)
FIGURE 29. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 45° AND STABILITY CLASS E
-------
103
60.0
x SO.O
5 20.0
u
2 10.0
0.0
1.1
£ i.o
i 0.9
0.8
8.7
.1
. -0.0
8-o.t
-0.2
(1) MBRMRL NBV EHI33I3N3; 0.5 PERCENT/MR SULFflTE FBRHRTiaH
(2) NBRHRL NflX EHISSI8NS; H9 SULFflTE F0RKHTI0N
(31 N0 N8)t EMISSI8N5: 0.5 PERCENT/HR SULFflTE FBRHSTIBN
It) NB NBX EHISSIBNSi NB SULFflTE FBRHRTIBN
-0.3
1S.O
10.0
6.0
0.0
10 20 40 60
(MMNHtNO DISTANCE (KM;
100
ZOO
FIGURE 30. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 90° AND STABILITY CLASS C
-------
104
60.0
50.0
40.0
SO.O
20<0
UlO.O
0.0
1.1
I 1.0
(1) NBRMflL MBS EHI35IBN5: 0.5 FERCENT/HR SULrfiTE F3f»KJ
(2) NBRMflL NBX EHIS5UNS: MB SULFflTE FBRMflTISN
(3) NB HVt EHISSIBNS: 0.5 PERCENT/HR SULFfiTE FBRHflTiaN
(4i NB NBX EHISSIBNS: NB SULFRTE FBRMSTIBN
0.8
0.8
8:1
-0.0
1-0.1
• -0.2
-0.3
15.0
1 i J I 1^ 1
10.0
I
5.0
0.0
10 20 40 60
MHNHIW OI6TANCE (KH)
100
200
FIGURE 31. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 90° AND STABILITY CLASS D
-------
105
60.0
50.0
£30.0
>
5 20.0
u
Kio.o
0.0
I.I
I
: 1.0
I
, 0.9
II! NBRHflL NBX EHI3SI8NS; 0.5 PERCEHT/HR SULFfl'E FOf rinT
(2) NBRHflL N8J EHISS18NS; N8 3ULFRTE FBRHBTIBH
(3) NB N8X EHI3SI8N5: 0.5 PERCENT/MR 3ULFHTE FBRH^TIBN
(4) NB HBX ENIS3IBNS; MB 3ULFSTE F0RMBTI8N
0.8
-0.0
I -0.1
-0.2
-0.3
15.0
I I I I I I
10.0
u
I
5.0
0.0
10 20 «0
DWNHINO DISTRNCE (KM)
100
200
FIGURE 32.
CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 90° AND STABILITY CLASS E
-------
106
60.0
50.0
10.0
ao.o
20.0
lO.O
0.0
1.1
(1) N0RHHL N0X EHI3SI0N3J 0.5 PERCENT/HI! %F«Tc
(2) N0RMHL NBX EMI3SI8HS: N8 3ULFATE F0RHRTI0N
O) MB NBX EMISSIONS; o.s PERCENT/HI! SULFRTE Fen
(4) M0 NflX EMISSI0NS: N8 SULFRTE FBRHATI0N
S
I'
0.8
8.7
.1
-0.0
! -o.i
• -0.2
10.0
S.O
0.0
10 20 40
OWWHIHD DISTflNCE (KH)
100
200
FIGURE 33. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 180° AND STABILITY CLASS C
-------
1U7
60.0
xSO.O
§40.0
o
£
.,30.0
>
5 20.0
£ to.o
0.0
1.1
§ 1.0
£0.9
0.8
.-0.0
[-0.1
• -0.2
•0.3
15.0
(1) NBRHSL NB» EMISSIBNS! 0.5 FERCENT/HR SULFfiTt FBRHRT18N
(2) HBKHRL N0X EHISSIBN3; Ng 3ULFHTE FBRMBTIBN
(31 Ng N8X EM1SSI8NS! O.S PERCENT/HR SULFflTE FBRHflTIgN
(4) N8 N8X EHISSIBNS; N8 SULFfiTE FBRMBT10N
to.o
ki
I
s.o
0.0
10 tO 40
DMNNIND DISTANCE (KM)
100
JOO
FIGURE 34. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 180° AND STABILITY CLASS D
-------
108
so.o
z SO.O
^ 40.0
^30.0
>
5 20.0
s
£ 10.0
o.o
1.1
I 1.0
(1) NBRHflL N8I EHI33I8N5: 0.5 FERCENT/HR SULFSTe FSifwaT
(21 ttflRHflL N8T EM1S3IBN5: N0 3ULFRTE FBRHATI0N
(3) NB M«X EMISSIBN5: 0.5 PERCENTER 3ULFBTE FPRHflTISH
(4) NB NBX EMISSIONS; NB SULFPTE FBRHRTIBN
g
A
0,8
-0.0
1-0.1
• -O.Z
to.o
5
s
E.O
0.0
10 20 <0
DWMMINO D1STRMCE (KM)
60
100
£00
FIGURE 35. CALCULATED PLUME VISIBILITY IMPAIRMENT
FOR A HYPOTHETICAL 2250 Mwe COAL-FIRED
POWER PLANT WITH A LIGHT SCATTERING ANGLE
OF 180° AND STABILITY CLASS E
-------
109
order to meet the 20 percent opacity standard) is apparent in these figures,
particularly for stable atmospheric conditions (Pasquill E). However, the
significant reduction in visual range caused by primary particulate is
noticed only at short downwind distances. All these effects are for sight
paths through the plume centerline. At large downwind distances, the
effect of sulfate on visual range becomes very Significant, particularly
for stable conditions. Indeed, the reduction in visual range increases
with increasing downwind distance between 60 and 350 km downwind.* The
scattering angle has a small but not negligible effect on visual range:
Visual range increases with increasing scattering angle. In other words,
the reduction in visual range is greatest for forward scattering.
Plume coloration and contrast are indicated by the values of the blue-
red ratio, plume contrast (at A - 0.55 pro), and the CIE color difference
parameter AE(L*a*b*). The effect of NO^ on plume coloration becomes clear
when the curves (Nos. 1 and 2) that assume normal NO emissions are com-
X
pared with those curves (Nos. 3 and 4) that assume no NOX emissions.
Yellow-brown coloration, as indicated by blue-red ratios less than 1.0, is
stronger with N02 than without. Note that the effect of sulfate (Curve 3)
on color is very small for all scattering angles; however, sulfate has a
significant effect on plume contrast, increasing plume brightness at small
scattering angles and decreasing plume brightness at large angles. Compar-
ing Curves 1 and 2, we can see that the pronounced coloration caused by N02
during stable conditions (E stability) is reduced by light scattered by
sulfate. If we use AE(L*a*b*) as an overall indicator of the perceptibility
of the plume (because of both plume contrast and color changes), we find
that N02 has the most pronounced effect on plume visibility at significant
downwind distances. Sulfate has a significant but smaller effect, and'
primary particulate has the least effect. Perhaps the most significant
result of these calculations is that plume visibility impairment (both a
* A cautionary comment is in order here. For this evaluation, the wind
speed was assumed to be 5.0 m/s; thus, almost 20 hours would be re-
quired for emissions to be carried 350 km. It is unlikely that stable
atmospheric conditions will persist that long.
-------
no
reduction in visual range and an atmospheric discoloration) increases with
downwind distance, suggesting that a significant impact could occur hun-
dreds of kilometers from the source.
B. PLUME/TERRAIN PERSPECTIVE MODEL
To supplement the quantitative description of plume visual impact
described in Section A, we developed a Perspective Terrain Viewing Program
(PTVP). Using computer graphics, this program is capable of displaying
views of plumes and background terrain with the perspective of the human
observer or camera situated at user-specified positions. These plume and
terrain perspective scenes can be used in conjunction with the quantifi-
cation of plume visibility impairment discussed in the previous section to
provide an understanding of the subjective impact of the computer predic-
tions. In addition, these scenes, along with computed Munsell color nota-
tion, can be used by a commercial artist to produce color renderings of
the visual impression of the background atmosphere and the plume for
various assumed emission conditions.
To use the PTVP, the user must provide the following information:
> Boundaries of the region in which the facility is situated
or is to be constructed must be identified, and terrain
within it must be digitized. (The U.S. Geological Survey sup-
plies digitized terrain elevations.)
> Design parameters of the facility that affect the effluent
plume characteristics must be determined. Among these are
stack height, flue gas temperature, and flow rate.
> Representative meteorological conditions must be specified;
important parameters include wind speed and direction,
ambient temperature, lapse rate, atmospheric stability
category, and the height of the inversion layer, if one
is expected to exist.
-------
Ill
> Observation parameters must be decided upon. Among these
are the location of the observer with respect to the
facility, the direction in which he is viewing, and the
field of view of the "conceptual camera" he is using to
record the scene. The last of these parameters is required
because the PTVP uses the lens characteristics to reproduce
the optics of a camera, "recording" by means of computer
graphics the appropriate camera film image. Lens "size"
as measured by the cone-angle of the field of view may
range from a few degrees (telescopic lens), to from 40° to
50° (standard lens), to 90° ("wide-angle" lens), to 180°
("fish-eye" lens).
The use of the PTVP can best be illustrated by means of an example.
Digitized terrain elevation data were obtained from the U.S. Geological
Survey for an extensive portion of the Southwest. From this data base,
a 50 x 50 km portion of terrain immediately west of Page, Arizona, was
isolated. A computer-generated plot of the terrain is presented in
Figure 36. Among the prominent geographical features contained within.
that region are the Vermilion Cliffs, the Marble Canyon through which the
Colorado River flows, and the Paria Canyon.
In this sample terrain grid, a plume from a hypothetical power plant
was displayed and viewed from several different observer vantage points.
In the example, the power plant stack is 775 feet high, plume rise was
determined using values typical of a large coal-fired power plant, winds
were light and headed slightly south of due west (compass heading of 2556
and meteorological conditions prevailed that are typical of Pasquill-Giffo'rd
Stability Category E (stable). The plume was assumed to be Gaussian, with
its "envelope" defined by the locus of la dispersion coefficient values.
The hypothetical observer in this example flew around the power plant
observing the power plant plume. Using a "camera" having a wide-angle lens
(with a 90° field of view), the observer took a series of pictures. The
-------
112
00
O
LU
CL.
OO
i
LU
Q_
a:
o
u.
oo
o
oo
ca
o
to
-------
113
location of each picture is shown in Figure 36. The sequence of pictures
is presented in Figures 37 through 51. All views are directed at the
power plant, except for Figure 39, which is aimed west toward the Vermilion
Cliffs.
C. COLOR DISPLAY TECHNIQUES
The most complete and realistic display of predicted visibility impair-
ment, particularly atmospheric discoloration, is a plume-terrain perspective
view in color, with accurately specified and rendered colors calculated from
the plume and regional models. We investigated two methods of displaying
atmospheric discoloration and plume visual imp-set:
> A color illustration, drawn or painted by a commercial
artist, using Munsell specifications for plume and
background color and plume and terrain perspective
views.
> A color video display, based on a photograph of a view
from a vista, computer-enhanced to display a plume or
homogeneous atmospheric discoloration on a color
television.
Both of these methods, presented schematically in Figure 52, use the spec-
tral intensity I(A) calculated by the visibility models for specific lines
of sight as a base.
These two display methods are the most technically difficult, time-
consuming, and expensive output options available for visibility models,
but they may be the only ways of giving the user of models an understand-
ing of the calculated visibility impairment. Without the aid of these
color display techniques, it is very difficult to translate numbers
describing visual impact into an observer's actual visual impression.
-------
114
c_>
c
o
o:
co
UJ
a;
-------
115
5
o
o
cc
CO
ro
UJ
-------
116
CO
o
5
o
o
CsL
LU
O1
CO
uu
01
-------
117
s
o
o
or
o
-------
118
LO
Z
o
o
e;
-------
119
UJ
I—t
>
C\J
*3-
UJ
OL
Z3
CD
-------
120
r-.
o
i—i
h-
o
_i
o
co
UJ
-------
121
CO
2TT
o
O
O
o:
-------
122
-------
123
•a:
O
o
o:
CD
-------
124
5
o
o
C£
Lu
-------
125
CVJ
CJ
o
o
c;
UJ
CO
^j-
UJ
cz
-------
126
co
o
o
01
UJ
o:
-------
127
-------
128
-------
129
d.
CO
a:
o
o
es
o zr en 3E
U-l I-H 2= 0.
CO
tn
CJ
i—i
u_
-------
130
1. Color Illustration
This technique is a synthesis of the regional and plume visibility
model outputs and the Perspective Terrain Viewing Program (PTVP) illus-
trated in Section B. The process requires the following steps:
> The source characteristics and location are selected,
and the digital terrain data are obtained.
> The Perspective Terrain Viewing Program is used to
generate a plume terrain view for a given set of
meteorological conditions (wind speed, wind direction,
stability category).
> The plume visibility model is run for the same set of
source and meteorological conditions. The coloration
of various sight paths through the plume and the back-
ground sky are predicted.
> Specific areas of the plume terrain view are assigned
the appropriate Munsell color notation and associated
color chips.
> A commercial artist colors the plume terrain view.
The color chips are used as a reference check on the
artist's color display.
To illustrate the capabilities and limitations of this technique, we
have constructed a test case that compares our predictions with the actual
visual impact of a power plant plume. The comparison demonstrates the
need for carefully documented studies of the accuracy of the model.
A test case should have the following attributes:
> A large point source with a visible effluent impact.
> Documentation of the source emissions (NO , S09,
}\ £
primary particulates).
> Location near a nonurban area of great aesthetic value.
-------
131
For our test case, we selected a power plant in northern Arizona for
which we had color photographs documenting the visual impact of the plume.
The source emissions and dispersion conditions have been studied, but
these data were not available to us. However, we were able to estimate
the source emission conditions from other similar power plant sources.
The photograph we selected is reproduced in Figure 53, which clearly
shows the brown coloration caused by the power plant plume. We chose this
example because of the clearly apparent brown coloration. However, upon
closer inspection, one notices that the lighting conditions are somewhat
unusual because the photograph was taken very early in the morning. The
sky is less blue than normal and is slightly yellowish at the horizon. The
long shadows of the river canyons are visible, indicating a very low sun
angle. As noted in Chapter III and Appendix B, the diffuse component
(multiple scattered light) of the solar intensity becomes significant
near sunrise and sunset. Since the diffuse component is hard to model
correctly, particularly in extreme situations like this, this photograph
represents a difficult test case.
For our estimation of the source characteristics and meteorological
conditions, the Perspective Terrain Viewing Program was used to generate
a plume terrain view, which is shown in Figure 39. A comparison of Figures
39 and 53 shows that the terrain and plume locations are rather faithfully
reproduced. Although the resolution of the photograph is much greater than
that of the computer graphics algorithm, the resemblance is clear. The
distant mountains on the horizon (dark blue on the right-hand side of the
photograph) are not plotted because they were outside the terrain boundar-
ies of the program for this particular case. The plume boundaries are
plotted at la (a ,a ) concentration values. The vertical extent of the
plume in the photograph is less than it is in the computer plot, suggest-
ing that the actual plume oz was less than a Pasquill E stability. Plume
concentration measurements would be required to substantiate this assump-
tion, but the usefulness of the PTVP is clear.
-------
132
srs:
C-
0£
1-0
O LU
O- 00
I— o
M
U. i— l
Si
O LU
o :n
re o
CO
LO
-------
133
The plume visibility model was then run for our estimated source and
meteorological conditions. The Munsell color notation and associated
color chips for various parts of the plume terrain scene are shown in
Figure 54. After comparing the color chips and the original photograph,
we concluded that the colors are reasonably close. The yellowish color
of the horizon and the brownish color of the plume are reproduced, demon-
strating that the plume model is capable of predicting the brownish color-
ation and displaying it correctly to a user of the model. The sky color
is approximately the correct saturation and brightness.
The results of the final step of having a commercial artist paint in
the correct colors are shown in Figure 55. The artist was never shown
the original photograph; he had to rely on the color predictions from the
plume visibility model. Unfortunately, we are less satisfied with the
results of this step than with the previous two. The problems in this
step appear to be that:
> It is difficult to paint and blend the correct colors
to maintain fidelity to the predictions.
> The spatial resolution necessary to produce a realis-
tic scene is also difficult and requires a large amount
of time.
> The artist has a natural tendency to paint what he thinks
the plume should look like.
Despite these difficulties, we believe the technique has promise and should
be pursued, though more work on this step is needed.
Overall, we are encouraged with the results of the comparison. The
significant findings were that:
> The Perspective Terrain Viewing Program can generate
a realistic plume terrain scene.
-------
134
UJ
"O
Q-
UJ
I— a:
UJ o
(/)
Qi
LU i
is
-------
135
00
CL
l—o
<— J
u3o
0:0
OLU
C/)Q
ctro
02:
a.
o >
Is
cc
^
C3
-------
136
> The plume visibility model can predict the plume
coloration correctly.
It is important to emphasize that these findings are somewhat preliminary,
and testing must continue to verify computer models. In addition, more
information must be gathered so that other tests can be conducted.
2. Color Video Display
In addition to the color illustration described in the preceding
section, we investigated the possibility of using computer-generated color
display facilities. This technique was originally developed for proces-
sing photographic data, particularly satellite data, and it requires
special equipment, including:
> A color densitometer to digitize a color photograph
in three colors.
> Image enhancement software to allow manipulation of
the digitized information.
> A color video display unit and supporting software.
Although these facilities were not directly accessible, we were fortunately
able to acquire the assistance of researchers at Los Alamos Scientific
Laboratory who have used this technique. The Los Alamos personnel utilized
this technique to predict visibility impairment from power plant plumes
(Williams, Wecksung, and Leonard, 1978).
We sent Los Alamos the test case photograph (Figure 53), which was
then digitized into three colors. Then the plume was removed from the
digitized photograph by interpolating the sky intensity from the horizon
below the plume to the sky above it. Next we gave the Los Alamos per-
sonnel the results from our plume model for specific locations in the
plume. These intensities were displayed on the color video screen,
and a photograph was taken. The results are shown in Figure 56. This
-------
137
o
QL
S
i—i
—I
-------
138
figure is then an illustration of the color video display technique. The
photograph can be compared to the test case photograph (Figure 53).
Because of compatibility difficulties between our computer output and the
Los Alamos facility, the results in Figure 56 should be considered as a
qualitative indication of the technique.
The color video display is a very powerful technique that gives the
user a rendition of a photograph with the plume effects superimposed. The
technique is somewhat cumbersome, however. Specifically, the difficulties
are the following:
> The hardware is expensive, available only at specific
locations, and not available on a dedicated basis.
> It is difficult to transfer information from the ori-
ginal film to the final produced photograph without
introducing errors. In other words, a quantitative
measure of the color fidelity of the process is not
possible at present. This difficulty is due mostly to
the problems involved in film processing.
This description of the difficulties of the color video technique is not
meant to be a criticism of the Los Alamos personnel and their work. It is
simply a listing of problems that must be faced in using the technique on
a routine basis.
D. THE REGIONAL VISIBILITY MODEL
We have modified the Northern Great Plains regional grid model (Liu
and Durran, 1976) so that it has the capability to compute regional con-
centrations of N02 and sulfate. We show in this section how these pollu-
tant concentrations can be used both to display visual range isopleths
for the region and to characterize atmospheric discoloration at specific
locations (Class I areas) within the region. We summarize the results of
sample calculations using 1975 and 1986 SO and NO emissions from point
A X
sources in the Northern Great Plains. Also, using 1972 SOp emissions from
-------
139
the copper smelters in Arizona and New Mexico, we constructed a hypotheti-
cal situation by assuming that these sources were located in the Northern
Great Plains. The objective of that task was to study the effect of the
large SO emission rates from the copper smelters (6000 tons per day) on
X
regional visibility using the existing Northern Great Plains regional
model. The significant impact of copper smelter SO emissions on visi-
X
bility in the Southwest is indicated by the results of the data analysis
described in Appendix A and the regional model calculations reported in
this section. This impact suggests the need for a regional visibility
model for the Southwest capable of handling the transport and diffusion
of copper smelter emissions as well as power plant emissions in complex
terrain.
Figures 57 through 60 show the isopleths of S02, N02> and sulfate
concentrations and visual range calculated using the regional grid model.
N0? concentrations were calculated using the technique described in
Chapter III from total NO emissions assuming a background ozone concen-
X
tration of 0.020 ppm. Sulfate concentrations were calculated from SOX
emissions using a pseudo-first-order rate constant of 0.5 percent per
hour and assuming negligible primary sulfate emissions. Visual range was
calculated from the Koschmieder relationship using the following value for
the extinction coefficient:
bext = (°'24 + °-04[s04' in M9/m3])(lO"4 m"1)
With the assumed background SO^ concentration of 1.5 ug/m , this
expression gives bgxt = 0.30 x 10"4 m'1 , which corresponds to a visual
range of 130 km. The bscat-to-mass ratio used here (0.04 x 10"4 nr1/
yg/m3) is appropriate for sulfate aerosol in the accumulation mode at
average relative humidity, and it is the average reported by Trijonis
and Yuan (1977) for the Southwest.
The calculations of visual range (Figure 60) indicate that anthro-
pogenic emissions from point sources within the Northern Great Plains
-------
140
en
£
— I— 10
r--
oo a: CTI
z o •—
o u_
z a:
i— i a.
LU a.
O et O
o LU
o: I—
C3 OO
o z:
oo z:
a: o
LU LU O
O O
Z O
O O
:r z
I •-( c^
LU z:
LU CO O
oo o
— i to z
-------
141
!:;'•
o
o.
0
O)'"
o.
td
0
r*"
o.
03
O.
Ml
O.
O.
o.
«M
O.
""*
or, oo u.'.
„- -"'*"
: "
"
•
.
•
•
P ..•"""<•
^?
,. -•'" ..- ' •.. t
:\ ".-•;""" \ [
.' .• * * " jt
\ *. '•••••.
''•». -( •-„.., /
ft/ f
'.,. , /
*--o: """ - ' •'"
- t— *'• — , .-•*' '
- g <
;g .v-
:'•'*"""""
- cc
z
cc
" ^ .••'""
• x _.-•"
/•' .. •''
V, ..' ..
.* • .' *
' .*' ** •** f "'^t
. **" »*
•>:.!•••""' >
.• *" .•''
*' /*'
*'' .•-'""
• ' .»••••"***
' '.3
00
^y
T
1—
x:
^
0 SE 0
| 111 .W—
1 Ivi-f ^^
a: x:
ZD 1 — U.
-0 °..°
— • 1— 1
•n 00 O
— 1 CO Z
«=C I-H O
Z S 0
O UJ
_ t— < — 1
•2 C3 ^D o r^
H-. O O
a LU uj
uj oo i—
OS
.0 W
"" rs
CJ3
H!
06
QL 08
OS
oe
01
-------
142
06
09
OL
09
OS
ot
OS
OZ
01
o
o. -
o.
O>
o.
ID
O. .
:•-•§
»'tt i
B-!2
i
a
tn
o. .
Ift
-•• ----------
O.
(O
i—i
<: en H- o:
i ra «c Q-
CC O UJ fS.
•=>^&.
O CD VO
:c a:
I LU Z Z
LU a. o: o
LU LU
O
_l C£.
^C I1'
Z O-
o
i—i UO
CD •
LU O
a:
a
LU CD
o
•z. o
o
LU r~-
h- T
o
z o
oo
z: u_
O O
O i— « l~l
a o
LU oo
0£ 00
Q- —
LUt-
-------
143
06
03
o. .
en
o.
CD
o.
<0
o. .
ir>
Oi
9-Hf
09
05
Ob
oe
o?
01
••••;> -on, \
ii
ct
o
- z
*
o
6-
--.__„„.... 12D
021
"T»B •'"T
•ft
IT
CO
bJ
130
06
08
09
OS
a
o
3
s
*o.§
. .o
,4.0
Ok
06
0?
01
o
/:.o
«
. .o
. .a
21!
LU CC
gj
Q
UJ O
o: oo
O. r—
o
10
UJ
Q:
LU •—i
ar o;
fc S=
Q_. *^
o
UJ 2:
re o
oo
00 O
z o
o r^-
f—l r—
OO I
oo o
>-> o
-------
144
cause only a small reduction in visual range, even with the increased
emissions projected for 1986. Maximum visibility impairment (i.e.,
minimum visual range) occurs in only a small part of the region. Visual
range is predicted to be reduced from the background 130 km to about
100 km, approximately a 25 percent reduction in visual range. This
small reduction would be difficult to measure, particularly using visi-
bility target observations, and to separate from visibility impairment
caused by natural sources. The calculated reduction in visual range in
the Northern Great Plains resulting from 1975 emissions is even smaller.
The results of calculations for 1975 emissions are presented in Appendix G.
Anthropogenic visibility impairment is significant in the Southwest,
however, as indicated by the analysis of visual range data presented in
Appendix A and summarized in Appendix B. In southeastern Arizona, copper
smelters emit large quantities of SO ; in 1972, before pollution abatement
A
equipment was installed, copper smelters emitted an average of 6000 tons
of SO per day, which is more than an order of magnitude larger than the
/\
1975 S02 emissions from point sources in the Northern Great Plains (430
tons per day) and three times the projected 1986 SO emissions (1990 tons
A
per day) in the Northern Great Plains. Although we have not developed a
regional model for the Southwest capable of handling the complex flow pat-
terns caused by rugged terrain, we examined the impact of the copper
smelter SO emissions on visual range by locating hypothetical copper
A
smelter emissions sources in the middle of the Northern Great Plains grid.
The hypothetical emission sources had characteristics identical to the
copper smelters located in Arizona and New Mexico that were operating in
1972.
We then ran the grid model for the meteorological conditions of the
Great Plains to evaluate the impact of the high SO emission rates from
A
the copper smelters. We summarize the results of this calculation for a
given time period in one of the test simulations in Figures 61 through 63.
The state boundaries have been removed from these plots to emphasize that
-------
145
c»
*— «~
o
o.
o.
0.
CO
o.
o.
UJ
o.
o.
«*•
o.
OO
o_
C*J
o.
06 08 OL 09 OS Ofr 06 01 01
.... •' .- .-•••"-.. *..'"'.
..-• - ? .', ,' •; \ •'..
'<•' '.'•'''
• •"' .' .' •'' ..'
" ' .' _ .' ..' ' ,, .-'" .. ••
: X" ."••'////>"., •"'" / :
,.-'' • • f <••'' •' ' . '
•: ' i ••'•'/
.' .' / ; / 1 ''
'. . '' /' ..•>' .-' •'.' /'' !
.' .' / ? ,'' '
•' ••' ' t* ' •''
/,',•''
'"//''' '
-•"' •*' / •'* ' * r*"*****! •"" »
' * '•*"** "*"' ' ' WT '
; /*/.•"""" >/ / :
•l • •".*,•* .
.. *' .-••"* ."•'* .•• '' .*".*' '"' •'*
" \\T\'- * -""!."•-• '''"..-'" '''-'? w^ .-' .-.--
--••"""""-*" ^ " • ' ' *fll';^ ''-f I • '"/' t»"T"'i"tt*'
*"....."" T •• " " .•''"<£' ,- ' " " ' j »fc5 .*' "*" "•• •'•'•' ' "5i»*' *
: "**••. •'--"' >/•'•' ••.'•>:-;;' ••-"' ;
' ' f'\^Lr'e t ' • ^"^ .'"•','
' *-' V '• «?*"••;•,:'' "' .-'"'
• " ••''.-'':' -.;.;:•• V '' .-,>'''
- " ' -"_•".' •"" ' •'! " •**' >^"'*
..... .... . .-•''_.,._ - ' ,;, %(•-'." ',.•' • ^ "
.. '" % '•''"' y.:'"- • ',."' '
,»•• '" . • ' ."/'..•"-'*" ,<•*' •' *". * -•' .•"'
* " - \V '' * '' ' ' "'"' ' fe .*" • ' ' •*'
" **-. (M :- •-'"!'*''.- ' . " .-*" t' *
'" ' . ' • .— --"*'" • '*' .* ' .'" •'
*'•-., * "' .-""" • '' • ' • *
"*'* ..,- • - * ,. '*\ '"•',''
,•-' ^'
. .-'•'. v ..
*•?> .-"',-v .-•" ,'' . '''
. ' f --••" ' ,..-*'* >
/'^ . ^ . ' ' •''*
/-'* »
* ''
-.'_.'"
-
,...•''' ,-•-••• " I — .......
/ . ' " ..-• * "" "*'•. ..^
"^ ...-'" ..,. .•-•" "*•— • / v
[ ^ ^ ..--•" ..,..•• * "'""- .. "•""
..-••'
(
"
- / ^
06 08 QL 09 OS 0> 06 OZ 01
~JU
..
o
•^
o
.0
.0
0)
.0
00
.0
r-
.0
— i
OO _J — 1
E •— <
to z
O «3
III ; [ ^^
f n c/o ^t
LU sr i
S> LU t— "
•a: ce
I O£ CX
rv LU O
LU »— i O
LU r—
Q 3C
LU 1 — LU.
t— O O
CJ Q-
t— i >- (/)
C3 DC 2T
LU O
fV^ ~Z^ >-H
Q- O h-
__:
UD
LU
O£.
CD
I-H
U.
-------
146
06 08 Oi 09 OS
Ob
oe
QZ 01
o
— '
o
0-
o_
o.
03
O.
^
0.
'
' / /' .-•'£"
-
* •' I*
/ / /' •'''''
/
' •'
{**
tfr
•'',"':
/
'" ' .' ,!
. P
*v
/
«' *x
''*.'"
.-"".•''•'••
"" ••" " ,.••
-- -- 4 -•""" / N
" .- " • -.. .-' .''' *
•••-._. -
(t) •
/ .-*' • -
W .-•"'' ,•'•'*"''
.'* . '**' *'"
""""-. .* ' .••' .••'"' ,;
„•' ,..'•" .''
{
'.
.,••'
- ... ' ..--"" 'i
\ .'' ..-'*' f
.' -'* •'*' i
''
t"
' ^ W.''
.
-..•'"'
•
.
-
•
i ;
06 08 OL 09 OS Ok 06 OZ 01
r— fV)
x -^ Z
01 O
O f~*
co o o co
^y 1 1 1 121 y^*
O CO O i— i
t-i co co
eC 00 MD
i D: i— i r^-
.0 GZ. rD sr CTI
vo ra o LU i —
o :c
•y C^ 1
i a: LU i— <
LU LU (— C£
LU O. _1 Q.
-O fV LU cf
"* 31 1— 21
1— Z CO U3
LU
_i c_> o: z
1
*^^ <~^/ t^^ 1
1— Z LU O
0 •-< x; o
O ^D O r—
-2 UJ CO D.
°* oi co >- u_
Q_ et 3; O
.0 ^
LU
a:
rs
U-
-------
147
c;
(Zl
o
o.
o.
O)
o.
00
0.
o.
to
o_
in
o.
<**
o.
W>
o.
rg
O.
06 08 Oi 09 OS Ob 0£ 0? Ot
•'.-.'/:- .''"'VV' \X", 'r
•.<-'.', ! ••' .: ' '.• 'c0 r
• ^«! ,' .' . .' \ •*! ,1
. ' .' ' " ' ' • ' / •'
'• •'' •'& •'•'.-'" ...'" ,'"
• ,• . :'Q> •' •' •' .•'" I'
1 $/// '/// o
: "N •' • i .' iS.- .•' <""•
; /••''/ ••'!/' >^/ ""
/ / / i'" •'.•**.•'
.' . * / .• •' •
,• ' * ~ /f ••'
• '•/'/ *
. .- . .-' * ,-' i
,• .• / .' ' j -
.' *s •' f / > 1.
• ' -V / •' '' i
••' 'i^5 • / •' ••' L
". /:'*?/'! i I / ;_
; ^ (j . '...•',.••/ //// i;
..-••"' v '.-•*' •' ..•' -•' • V - !•
-•-.•-.::• - •-• '••-•.:".-" - • \"^r ..--, .-''.••'/ / f
* •:'":: l;;--:-"- ••- -• iX .-- •- "" ,'' /!»* '.
- ...;—•--..-- «v / / _>v "- c5
>"""*... " " % *'-.' ..'' / .-•"' 2
(5j * i *' * /* * ™
. , —,-* •- •*• '* •' . t' Cj J m
"n • . • ' •'' .' ' / •'' ^- (
: "" V''"-''-''/ •••••-""" ^" ' :
***/•' •' i \ ••"'' t ' 1
'1 j' i *. '-• X *•"
1 "•( \ '.' \* -•''
" ' • / / 1-0 . •' ,.-•''
^' i ,-' ••' •o* » .•' ' •*
>' .;' '' ,'4' i''
//'.'.•'
, '^i'-' •' /''
. ' '.^ . ,•'''
'''•''•'
'*•• . ''••''••<&> "
VV' -•' "*
"••' .'
.'
/*'"
^'
•
/
/
.-•''
..." /'"""' *
--•—'*"'' t'' \
.-.^— «*™***-""~" 3 30 '" *"'"" .*' .
06 08 Oi 09 OS Ot 06 OZ 01
-x ;c o
i— 2: Q- OO
< >- 2:
x a: z
o
o -fZ.0
oo o
0 Q CO f—
.0 -^>
^4 ^—
O " U-
O UJ O
o: (—
C3 et OO
^ cc 2:
.0 0 0
O) ef 2: >— '
CQ 0 (—
<£ 1— O
2 ^
•o 2: o; o
•D __ O
s: u i
oo uj o oo
OO I— •— i ^
•° ^ U- O <
^-Z3 o a.
E oo o;
^1 ° *"
.0 3 (— UJ
UJ UJ O UJ CC
CD 3C S C3
cC C£ Di 2t
a: uJ o cc
o U_ \ ' '
"10 i< i— oo ^
Z3 2* 2= CC
OO UJ O O
t— * C-> >— < 2T
> CC OO
0 UJ 00 UJ
•• i ex i— -i re
< 2: H-
z: uo uj
o • zc
i — i O CC i — i
CD UJ
-O I i I CC _J I^~
Q UJ CTi
Q -z. n »-
UJ cC 00
1__ |
r~~ —J
0 o E cc K-I
"AJ '""' -^ Lx* CC
" Q Q. a.
UJ O Q- ct
cc co o
O- r~ C_5 **O
.0
oo
to
UJ
CC
C3
t— •
-------
148
this is a hypothetical test case. We assumed a 0.5 percent per hour sul-
fate formation rate in these calculations and negligible primary sulfate
emissions. The results of a sensitivity analysis using different sulfate
formation rates are summarized later in this section and are presented in
detail in Appendix G.
Figure 61 shows that the maximum Sf^ concentrations occur at short
distances downwind of the hypothetical smelters, which are located in the
center of the region. However, as shown in Figure 62, the maximum S0|
o ^
concentrations (greater than 8 u9/m ) occur hundreds of kilometers down-
wind of the smelters. The impact of this sulfate on visual range is shown
in Figure 63. The worst visibility (less than 60 km) occurs in one small
area in the upper middle portion of the region. The increased sulfate
concentrations and resultant decreases in visual range occur in two direc-
tions from the sources as a result of a change in wind direction that
occurred on the day before the simulation. The calculated concentration
maps for other time periods of this simulation period are presented in
Appendix G.
The impact on visual range of assuming different sulfate formation
rates (0.3 and 1.0 percent per hour) is indicated in Figures 64 and 65.
Note that with the reduced sulfate formation the minimum visual range is
70 km, and with the increased sulfate formation rate it is 40 km, which
compares with the minimum visual range of 60 km computed for the base case
of 0.5 percent per hour.
These sample calculations of reduced visual range cannot be compared
directly with the observational data from the Southwest because these cal-
culations were based on Northern Great Plains meteorological conditions.
However, the results agree qualitatively with some of the conclusions of
the data analysis summarized in Chapter II: namely, SOX emissions from
copper smelters in the Southwest can cause a significant reduction in
visual range even at locations several hundreds of kilometers downwind.
The predicted maximum SO? concentrations in these simulations, ranging from
3
8 to 16 vig/m, agree qualitatively with measured maxima in Arizona. The
-------
149
o
-!'
o
o.
«-*
o.
at
o.
CD
o.
fw
o.
ID
0.
in
o.
«r
o.
Ol
o.
'' .*•-"' i'
•'. // *•*' /'' / o -
.-' / / / '•' m .
/ f '
/ /
; .• .• /
.'* -'* •' f
,' / / .'
.' ,.' •• f
1* .'' / •'' "
'„'*•"*
,' / / /
£, ••" ••' .'
$/ II \
.' /'* / ' "
'" ' •""*'• f^J f *
.-•'' 0^ •'* • ""• *'
.M***" * \ .. ' (X •' l&
-• - — --•"".",-."•• • ""-••"'' ^ / >
'tfiljV-- . .. ,-•"*"" .•' /* /
' • ' OA ** '* ' / /* * f*.
.-.. **"*•'•--.„., / .•' ^ f aC
?' '*. ***** /' /' .'' / *^ "
„ j J •'/'/• t .
/ ' / .- -•**' 1
''" " .'* / , ' •"' •*" / *
: -•---•' ////" 1 •
'• ^ i& .•' / .*' / •
/ c> » ^0 ' ' •*' •
'. i "£ '""'* .-•''' "'S
.••'•.. O \
•, 9 }
• •*
' I f'' . '' "
"•"/•'.'
• / ' ^* •' .' •'*
.• ' .•' .-•*'
,••' .• V ••'"'
- , • ' / ' .'l^^
' ' '.''' "\
'•';''•''
:''.'' '.
p.-
-
f'
/
/
*•''
,.''&{•*.__ . ^- .— '-••" .•••'"'\-
06 08 QL 09 OS Oft 06 02 01
E
I
X
o
-— *
a
.0
•••
.a
o>
o
ID
•g
.O
«
.O
U>
.O
*
.0
w>
,0
0
^
LU
CD
Z U-
«r zr o
c2 o
oo
_J Q Z
=3! LU O
^ 00 i— >
i— i CQ i— <
> 0
-2T
Q LU O
Z t— O 00
~™N —•«• ^^
~^ ^4^ ^—
O C£. 1 t— i
rv «3C eC
CD z: cj _i
:*£ o i— i a.
CJ i— i CD
<: I— O (—
CQ <=C _ J ef
2: 0 LU
=1 C£ Di Di
O O CD
CD U_ LU
'ZL i — z;
i — i iii 1 1_ i rv
^- i 5- iij
^^ P^ ac^ I-*-!
r> «=c n:
oo LI_ a; i —
co _i o cc:
=t rs u_ o
oo 2:
• oo
^--.a; 2: LU
E rs o a:
• — -ac oo
oo z
LU CcT ' — ' > — i
CD LU s;
z: a. LU UD
as H- CK: 01
«Z LU i—
_l LU (—
cj o re i—
i— l CO h- 1
a i— o o
LU Q- O
C£ U_ >- <^-
Q. O 3C r-
•
VO
LU
CD
HH
1 i
-------
150
• -,
%-t _
— <
o
o.
O-
o>
0.
eo
O-
o.
0.
o.
o.
o.
'
'• '••"'''•.',•*? ' j "f
- .' '••'.' : •<£•' ''.' •""' .••' f
' t-' ^ • - "* ' C^ •-' * " *"
/ 'Cvatv" . ' ' ^^ ..-"'" Q
r ' -'*Y' :'. ''•"•', / /' •*
t- f\7 '. / -' " •' '* •'" t-
**• i -' ; •' .* " . •' L
• •• ':.£>• ..•'•>/,§> i-
•' ' • v' •''••'•' /"* t
''.''.•*•' i"
^, /••'•' /"V/ r
*/'"**' **V ,',' V
/••/.-•'.••"•'/ r
: :•;•••'•"•''/•$•.•'/.•' I
.- •'' " > 'C1 .- ' $*•• ' .' '* 'Qi i i
. ... "" '"—'", ->- '1 ••'..*' -' ••/', "V '•— ^ i
"* '•" " 1)}" '•- • "" ' ' .. • '"' • ,• ' , t •' \
- '" 1" * •t?..".,'. •-•- """ "' -' *'/'/ , ' a
' *'-- crfi ., ^' •*••' .'' J A*' .••'"'.• Cf
' t?^ .. * " ". ^J^ •' / ' ' *1^ ,-•'' •* *"
* — . / / ,•' / / ^' ••* *v «'f f
} / • •' '• '••'',•''..-• ' -vl? /' f
•' .•' - - _. ,' .•' ..-•' t-' f
" -•' • ' • w* ' .-'*' -•'* ' . ' t
•* .•"'.'" • .'*' .' •'*' ' . • "'1 ••*' i-
"-••. .—• ,• .V r> •' • ,-'' _.""' ,; ,.' i.
- .''/.''•'•$? ''"" V' •'•'"' l
" • •'/'•' i • *. \ • t\y '*•
' • ,''*)'"'" * ""'*" *' .••' S '..
'.t \ \& '.:."""' /,;<•••' -
t *'""'. } • ' •'*' .'•''
') £"• '' '" •'.'"'
- .••'"" '\ I ^ ••' . .•'.*$••"'' \
: .' : .-' ' . 130 -, ».
iiiiiiiiiiiiiiil'ii'iiiiiiiii -*-'**f'* i I i i i i i t t i i i i i i.* i i i i i
E
O
X
o
. "~«
*^
o
.0
.0
•»
.0
QO
.0
.O
vD
.O
I/)
.O
«•
.O
en
.0
.0
w*
O <-i
UJ UJ o
CO T*
QO
uj o
O CO «rf-
py ^f __
•^> DO
O U_
a: "o
e; uj
«r. §o
cc i— i
^ i
<; o 1-1
i— i r~i
=i<£o
— ^ CJ
I) O _l
co o z
<; uj i— i •— i
"«£ O -J
•^-Li ID-
E-JO
NX "*"^ fV ^.^
CO UJ cj;
1 — UJ
UJ C£. UJ C£
2: o
•=t r: o: z
a: o en
(V 1 i 1 .J
• uj co
-------
151
measured mean and maximum SOT concentrations in 1972 and 1973 in several
locations in Arizona are summarized in Table 5. Note that mean sulfate
3
concentrations ranged from 2.2 to 12.5 ^g/m , and maxima ranged from 5.8 to
41.0 yg/m3.
TABLE 5. MEAN AND MAXIMUM 24-HOUR-AVERAGE
SULFATE CONCENTRATIONS MEASURED
IN ARIZONA IN 1972-1973
[so]
_ Location Mean Maximum
Ajo 12.5 41.0
Douglas 9.1 25.9
Flagstaff 2.6 5.8
Grand Canyon 2.2 6.0
Phoenix 6.5 34.5
Superior 5.2 22.1
Tuba City 2.6 9.7
Tucson 6.4 22.0
The highest concentrations occurred in the immediate vicinity of
smelters (e.g., Ajo, Douglas, and Superior) and in cities (e.g., Phoenix
and Tucson), where, presumably, the rate of sulfate formation would increase
because of the elevated concentrations of reactive species in polluted
urban air. The high sulfate concentrations measured near the smelters may
be the result of primary SOT emissions or the rapid conversion to sulfate
of emitted S02 in the initial stages of plume dilution. The measured
maximum sulfate concentrations in nonurban areas of Arizona distant from
smelters (e.g., Flagstaff, Grand Canyon, and Tuba City) ranged from 6 to
10 pg/m , in qualitative agreement with the regional model calculations.
-------
152
Our approach to the calculation of atmospheric discoloration on the
regional scale has been to compute color parameters for particular loca-
tions (e.g., Class I areas) using the NOp and S07 concentrations obtained
using the regional model. In these calculations, we assumed homogeneous
pollutant concentrations within the mixed layer, and we computed optical
effects for several different sight paths. Exhibit 6 shows the output
from an example of these calculations for a location with background con-
3 3
centrations of 3.4 ug/m of sulfate, 30 yg/m of coarse mode particulate,
and no NCL. For different scattering angles 3 and sight path elevation
angles 6, the following specifications of color and color change (similar
to the parameters characterizing plume impact) are printed:
> Optical thickness (T).
> Light intensity (Y and L*).
> Chromaticity coordinates (x,y).
> Change in light intensity between the given background
atmosphere and a Rayleigh (no particles) reference
atmosphere for the given Q and 6 (AY,AL*)
> Contrast between the given background atmosphere and
the reference at three wavelengths [C(>K ^ = 0-40,
0.55, and 0.70 ym].
> Blue-red ratio between given background and reference.
> Changes in chromaticity coordinates (AX,Ay).
> Color difference parameters [AE(L*l!*V*) and AE(L*a*b*)].
With these parameters, the color of the background sky at a given loca-
tion in a region can be specified and Munsell color notation can be deter-
mined from the values of L*, x, and y. The change in light intensity and
color between the given location's sky and the reference atmosphere is spe-
cified by contrast values AY, At*, AX, Ay, and A£. These differences do
not have the same meaning as the corresponding parameters for plume impact
because the observer compares the light intensity and coloration of the
-------
153
i>:
m
a.
a
1
£ K
OH
sS
«
£ s;
;r*T l-c
SSSSSSS
> I- K * co w •? N ; «• co a co
«• * s * ® I- *
N * aIt-1- * *
«• i- co «• n *
9 0* 4' ff1 s ft f-
* « n tf»»« - * »• n <
ei — »ee t-(• i- «c « — «»«• eo ec
*> *i •»• CO
> Sco-
i
i
— rao — M— N c* * cc co t» » M * j *)« «r es
e - i- + 1-
— i- IT * fc
§« «*
wwwN S? * *>""£:§
— »•
n M
flM
« ^ «. «. S' « -5'
- «• n «cj - -
9 !M M N * M -
CO CO CO CO CO CO « (5 k' M M M N
<- «-*s-NCOt-
? co •• ep c* fr o »p ff- r- w i^ * *
> N N M — — — ® M M C-t M f 1 IS
!•«'««'»»«' *"""""'
.. £jS®*S«<
C&QC^C&t^^^ *^^0>{v
§»«*«* * I- « — N N « rt — * * w t> * *
nowt^i^'c oMft»n»o«ic CO******
i £
5 -
W 1- 1- » ES
Nf- *• r> * i
M 5' N «
ff SO C
1- 0- M 5' N « 4
O CD
05 1* «' N « cc « ^ T:
* N * — N *? l»
«• N » «• «• + «
- S 1- t- » ^6 -0
CM- 0- * M - l-
SR n c c> M M
Nme1-*^ —
* «• «• c- e- »• e-
1
5
»C9<>S*CfJ ^•^h-M*®C-
np*«5*«r5 *«h-*«*-*
3rt^»«o*-« (fc^ — c-*>*o^
>«K5^-*f * — *•*
1- * n N * * *
e- *.**«!- 0»
— » W O 8 N N
*p fr- (6 *O Cj C5 * i *
99>O*9'C&9tM CD*r
SMMNNMS S«
renncin £^e
* b- 9* 9 <» -•
0» N «• S i 0
D 04 JC C* W ^ C4
I ^8:
g
g
i i i i i i i i i i i i i i i i i i i
- * rt S M r-
i CO * ® « « «
i i t i t » i
t 1 1 1 t t
*««**rt« —mneoMMM
•* CO « C5 W
•» M — e-oCD*
— M *9®K9COeO£'
••51 CO^'C^NCOO'f
- 9- «> * * CO CO ft
row® N»»N
— * 4 <> » M N
CO N * J1 ?l S (5
t* o> * •• c- co co
CO N M CJ N C:l M
cjSeo*-* ^ ®
fr CO *
N * n t» »• M —
<> *• *• i* •* i^ 10
»««** + •*
fl 0) 0) P» N N M
»n9>ce4-s
§N N 9- CO N 1-
*>«•*"#•* *
M N M N N «
p «• ce * i- n
J P? »P ^ t^ *t CO
J - ^ W ^ « I*
-«!-
* n co
05*Ne-
»!>•«««»*
* P5 N CO * «
C* (*• *^ *^ ^ *^ ^
?SS!33c3«2
•• w (* ft* t* w w
Sn'I'^aNl-
*««-—-
6r5**«M»l-
*«« — ••••
& .^,
a -c
«««»
- n + «
t^c* ••CO**!*^ -^
LU
o;
««
rs
o
—I Q
LU
> LU
o o:
o
I—
«=C u_
o
oo
I—
o
U.
LU
O
CO
-------
154
plume and the background simultaneously; a relatively sharp line of demarca-
tion separates the plume and the background. For the homogeneous regional,
background atmosphere, the observer compares the given atmosphere with a
recollection of a clean atmosphere. An exception would be an observer on
a mountain looking down on a homogeneous mixed layer from a position where
he can compare the color of the mixed layer with the clean air above it.
These characteristics of the color differences for the homogeneous,
regional background atmosphere must be kept in mind when interpreting the
values of color difference parameters for determining the perceptibility
and the significance of atmospheric discoloration. Further work is neces-
sary to identify the threshold values of the color difference parameters
for the homogeneous atmosphere.
-------
155
V RECOMMENDATIONS FOR FUTURE WORK
As stated in Chapter I, we have followed a pragmatic approach to the
development of models to predict anthropogenic visibility impairment. Our
goal has been to develop predictive tools helpful in:
> Writing the report to Congress on visibility for setting
policy and promulgating regulations.
> Evaluating the impact of proposed sources and making siting
and pollution control decisions.
> Determining the requirements for retrofitting pollution
abatement equipment on existing sources.
This chapter recommends additional work that we believe is necessary
in the near term to refine and test the models, to assess the impact of
proposed visibility regulations, to improve color display techniques, to
develop a regional visibility model for the Southwest and other regions,
and to verify the predictions of the models by comparing them with field
measurements.
Figure 3 in Chapter I illustrates the potential uses for visibility
models in environmental policy and regulatory decisions, emission source
siting, and pollution control. Assessment of the extent of existing or
past visibility impairment can be accomplished through measurements using
such methods as (1) visual range and coloration observation by trained
personnel, (2) photographic documentation of visual range and atmospheric
discoloration, (3) telephotometry, (4) integrating nephelometry, and (5)
transmissometry. However, estimation of the extent of future impairment
(e.g., impairment caused by new sources or by the abatement of existing
sources) requires a scientifically based prediction capability that can
provide estimates of visual range and atmospheric discoloration.
-------
156
Visibility modeling will clearly play an important role in determining
a rational definition of "significant visibility impairment," in setting
environmental policy (regulation promulgation, new source and existing
source retrofit reviews, and long-term goals), and in determining pollution
abatement and siting requirements on a case-by-case basis. It is expected
that visibility impairment rather than ground-level air quality will be-
come the dominant air quality issue and will have a significant influence
on siting and pollution abatement decisions, particularly in the West.
The following sections outline the work that we believe is necessary
to support the EPA's efforts in visibility regulation promulgation. These
recommendations are presented in the order of their urgency. In our view,
further testing of models and analysis of the impacts of visibility regu-
lations should be done as soon as possible.
A. IMPACT ANALYSIS IN SUPPORT OF REGULATION DEVELOPMENT
The most urgent requirement for the application of visibility models
is the development of regulations. Modeling work will be necessary to
determine siting constraints on new sources and requirements for pollution
abatement, both for new and existing sources, that will be imposed by pro-
posed visibility regulations.
We have drawn two conclusions of major regulatory importance in our
initial applications of visibility models:
> The sulfate formed from S02 emitted from such sources as
smelters and power plants may cause significantly reduced
visual range at locations hundreds of kilometers away from
the sources. Indeed, the magnitude of the visibility im-
pairment may increase with increasing distance downwind
from the source, thereby making identification of cause
and effect more difficult.
-------
157
> NO emissions from large coal-fired ppwer plants may cause
X
perceptible yellow-brown plumes and atmospheric discolora-
tion more than 100 km downwind, particularly during stable
atmospheric conditions. Control of participate and S02
emissions will make the discoloration more prominent by
reducing the masking effect due to light scatter.
The implications of these conclusions for siting and control are
obvious. Impacts at large distances from emissions sources must be con-
sidered in siting studies. Although the impact of power plant emissions
on visual range will be reduced by controlling SO emissions, NO control
x x
is needed to reduce the yellow-brown discoloration that is caused by N0?.
In the analysis of the impact of visibility on industry, considera-
tion must be given to:
> The magnitude and the spatial and temporal extent of im-
pairment for various sources, ambient conditions, and
geographical locations.
> The siting constraints imposed on new sources.
> The pollutants that must be controlled.
> The degree of control required to reduce visibility im-
pairment to acceptable levels compared with the capabi-
lity for, feasibility of, and cost of implementation of
various pollutant control technologies.
> The appropriate regulatory policy to deal with visibi-
lity impairment (i.e., emission standards, ambient air
quality standards, or some standard of visual range and
atmospheric coloration).
B. MODEL REFINEMENT AND TESTING
Further work is recommended to test and refine the visibility models
in the near term, including:
-------
158
> Further testing of the models through sensitivity analyses.
> Incorporation of more sophisticated gas-to-particle and
aerosol growth algorithms in the code.
> Further assessment of the subjective visual impact of and
human threshold response to light intensity and color changes.
> Refinement of color display.
We limit our discussion here to work that should be done in the near
term to support the development of visibility regulations. In the future,
when a complete set of measurements are available (e.g., from EPA's VISTTA
program), visibility models should be verified. Measurements needed to
validate models include source emission rates, primary particulate size
distribution, meteorological conditions, plume dimensions, plume and ambi-
ent chemistry, aerosol size distribution and chemical composition, scat-
tering and absorption coefficients, solar direct and diffuse intensity, and
spectral light intensities and color photographs for several lines of sight,
Model validation is discussed in Section C.
1. Model Testing
As we noted in Chapter IV and demonstrated in Appendices E and G, we
have started to test the models by applying them to different emission and
ambient conditions to test their sensitivity to various input parameters,
including:
> Atmospheric stability (rate of dilution)
> Background ozone concentration
> Solar zenith angle
> Scattering angle
> Observer location and sight path orientation
> Background object light intensity and color
> Pollutant emission rate.
We recommend that more sensitivity analyses be performed with the
plume model for a variety of emission sources, meteorology, ambient
-------
159
conditions, and viewing conditions to evaluate the model qualitatively.
The results of this sensitivity analysis could be displayed in graphical
and tabular form so that they could be used by environmental engineers in
regulatory actions, impact analyses, siting studies, and design.
Further parametric analyses should be performed to evaluate the sen-
sitivity of model results to:
> Primary, secondary, and background aerosol size distribution.
> Ratio of diffuse to direct solar flux.
> Ratio of [NOg] to bscat-
> Locations of the background object and the plume relative
to the observer.
In a manner analogous to the ozone isopleth diagram, it may be
possible to characterize on an isopleth diagram the impact on visibility
of a range of combinations of the precursor pollutants. By plotting
contours of constant value for some visibility-related objective function,
these precursor mixtures which lead to the same visibility conditions may
be identified. Among the candidate objective functions are the contrast,
visual range, blue-red ratio, and AE.
Another sort of visibility isopleth diagram might be constructed to
characterize general regional visibility. Instead of HC and NO as
/\
"precursors," sulfate and nitrogen dioxide could be viewed as "indices" of
visibility degradation. By plotting concentrations of SCL and N0? along
the axes, one could determine isopleth lines that correspond to constant
objective function values.
One of the chief values of the ozone isopleth diagram is that it
provides an easily computed estimate of the reduction in precursor emis-
sions from current ambient levels required to reach the NAAQS. If the ob-
jective function chosen for use in the visibility isopleth diagrams were
the same as that employed in setting the federal standard, these diagrams
-------
160
might have a similar use. Required reductions in either SCL/NO or SCL/NO?
C. A *T t.
might be directly estimable. If the development and use of visibility plots
were shown to be both feasible and reliable, they might prove to be impor-
tant tools as promulgation and implementation of a federal visibility
standard occurs. This could be of particular significance for state and
local agencies, having limited resources and expertise, since they are re-
quired to incorporate visibility considerations in State Implementation
Plans.
2. 6as-to-Partic1e Conversion and Aerosol Growth
Currently in the visibility models gas-to-particle conversion (SOp to
SO^ and NO to NO^) is treated in a simple manner through the use of pseudo-
T* X O
first-order rate constants. Secondary aerosol is assumed to form in the
accumulation mode with properties observed by Whitby and Sverdrup (1978),
in the Labadie plume. Although this is a first approximation, it is a
reasonable assumption for modeling purposes.
We recommend that further work be done to identify the reaction
mechanisms effecting the conversion of SOp and NO to sulfates and nitrates
t~ A
and typical concentrations of reactive species in various nonurban areas
(Class I) in the United States. Reactions with the following species should
be considered:
OH'
»
>
and 0^ (in clouds)
I
RO-
HO*
NH3
Through evaluation of the concentrations of reactive species in non-
urban areas and resultant formation rates, appropriate formation rates can
be selected by the user or computed in the code.
-------
161
We recommend that an aerosol growth model be studied for possible
incorporation in the plume model. Such a model would compute the equilib-
rium particle size distribution as sulfate and nitrate form and water con-
denses onto the particle surfaces. We would determine if such a growth
model would improve the existing model sufficiently to justify its use.
3. Assessment of Color Impact Thresholds
We have incorporated into the visibility models the most recent
methods for quantifying color differences developed by the CIE in 1976
[AE(L*U*V*) and AE(L*a*b*)]. More work is necessary to determine what
standards for atmospheric coloration should be used (if any) in the vis-
ibility regulations. The AE's appear to be reasonable parameters to
characterize color changes associated with pollution; however, more work
is needed to determine what AE values mean subjectively in various cases
and what perceptibility-threshold and acceptability-threshold values should
be adopted in the analysis of atmospheric discoloration.
4. Refinement ofColor Display
In few instances is the display of model results so important as it
is in the prediction of visibility impact. A considerable number of sep-
arate lists of information are required in order to characterize a single
scene. However, the human eye and brain together are able to assemble and
integrate all this input, synthesizing it to a final impression of visual
impairment. It is the subjective judgments based on these impressions
that constitute "visual impact" of the most fundamental sort.
Consequently, the practical utility of a model depends on its ability
to collapse its predictions into a similarly simple and usable format.
It is for this reason that model predictions in this study have been ex-
pressed not only by means of specific visibility-related parameters, but
also through artist renderings of entire scenes with colors and intensi-
ties of sky and pollutant determined by model predictions.
-------
162
Several advantages have been achieved as a result. Predictions can
be assimilated more easily. Judgments about visual impairment are faci-
litated. The visibility impact projected to result from construction or
alteration of a facility can be more readily presented to and evaluated
by policy-makers and the general public.
While considerable progress has been made in this study to develop
suitable means for displaying model predictions, the following continued
efforts seem to hold the promise of substantial payoff:
> Additional studies could be conducted of the feasibility of
using artist-produced color illustrations to represent
visibility model predictions.
> A study could be performed of the feasibility of using
artist renderings of plume and atmospheric coloration pre-
dictions "overlayed" onto actual photographs of terrain.
This "photo-montage" technique has been used successfully
by the U.S. Forest Service's MOSAIC land use assessment
program.
> A study could be undertaken of the comparative accuracy
and acceptability of each of the above two display tech-
niques, as well as with the color video approach used by
workers at Los Alamos.
> The Perspective Terrain Viewing Program (PTVP) could be
linked to the plume visibility prediction model (PLUVUE).
By doing so, one could first display the terrain as seen
from a specified location, select a point whose colora-
tion was desired (as expressed in chromaticity coordinates,
perhaps), and calculate directly the visibility predicted
at that point. In this way, use of the visibility model
would be much more tightly integrated conceptually with
terrain views.
-------
163
C. MODEL VALIDATION
In this section we discuss our preliminary thoughts on a model vali-
dation effort, including:
> The type of measurement program that is needed.
> The specific measurements that should be made.
> The type of analysis of measurements and model pre-
dictions needed to assess model performance and to
provide direction for model refinement.
1. The Type of Measurement Program
No attempt should be made to validate SAI's regional visibility model
at this time. Rather, efforts should be aimed at providing a comprehen-
sive set of measurements downwind of a point source so that the plume vis-
ibility model can be validated. Information obtained from the point source
measurement program will be useful later in regional model validation and
refinement.
A large, coal-fired power plant should be selected for the measurement
program. The visibility regulations required by the Clean Air Act Amend-
ments of 1977 are likely to affect power plants, particularly in the wes-
tern United States, more than any other single class of emissions source.
Although copper smelters emit large quantities of SO , which has been shown
A
to significantly affect visual range in the Southwest, most of those sources
are exempted from the requirements of Section 169A on visibility protection
because they are more than 15 years old.
The power plant that is selected for measurement should have the fol-
lowing attributes:
> Pollutant emissions should be easily measurable and
should be relatively constant during the measurement
program.
-------
164
> Participate emissions should be well controlled using
state-of-the-art abatement equipment, such as efficient
electrostatic precipitators or wet scrubbers, so the
plant is representative of modern coal-fired power plants.
A major objective of the measurement program is not to
measure the visibility impairment caused by large emission
rates of primary particulates from older plants, but to
assess the visibility impairment caused by secondary aero-
sols (i.e., sulfates and nitrates) and NCL. Large emission
rates of primary particulate might interfere with the mea-
surement of secondary aerosol generation and light scattering.
> Sulfur dioxide (SO^) emissions should not be controlled by
scrubbers so that a significant amount of SC^ is available
for conversion to sulfate.
> The power plant should be located in the western United
States and should be isolated so that the plume can easily
be identified, tracked, and measured without interference
from plumes from other sources.
The emphasis of the measurement program should be on the visibility
impairment caused at far downwind distances. This contrasts with the
objectives of most air quality monitoring programs, which are designed to
determine the maximum ground-level pollutant concentrations, which gener-
ally occur within 20 to 30 km of the source. Visibility impairment appears
to be a long-range air pollution problem because it is caused by secondary
pollutants (NOp, sulfates, and nitrates) that are formed relatively slowly
in the atmosphere. Preliminary calculations show that the maximum reduc-
tions in visual range occur hundreds of kilometers from power plants and
maximum plume discoloration due to NO^ occurs during stable conditions 40
to 100 km downwind. Visibility impairment at distances 100 km or more
downwind of proposed or existing emissions sources will be the controlling
factor in determining the amount of pollution control equipment that must
be retrofitted on existing sources and in evaluating what the siting and
-------
165
emission constraints on new sources will be. Therefore, long-range impacts
must be measured so that the visibility model can be validated.
A crucial part of the long-range tracking and measurement of plume
visibility impairment will be the measurement of upper air transport winds
at frequent intervals and several locations so that accurate, real-time
plume trajectories can be calculated. These trajectories will be needed
to help track the plume and to identify the plume location relative to
fixed observation locations. Also, the upper air wind data can be used
later in conjunction with National Weather Service measurements to calcu-
late back trajectories and air parcel histories, so that potential sources
of background aerosol and trace gases can be identified. For example,
during a measurement program at a point source in a nonurban area of the
Southwest, there might be several days when air originating from an urban
area or from a copper smelter complex would get transported to the measure-
ment area. One could thus, with little additional expenditure, supplement
the information regarding point-source plume impact with information
regarding the regional impact of distant sources.
An attempt should be made to measure plume visibility impairment dur-
ing stable meteorological conditions. Greatest visibility impairment,
according to model calculations, occurs during stable conditions (e.g.,
Pasquill E or F). However, even greater impacts might occur during stag-
nant conditions or in locations where there are flow reversals (e.g.,
drainage flows) that could cause a build-up of pollutants in a confined
area. A study of climatological records could be carried out prior to
the measurement program so that periods of the year most likely to have
stable or stagnant conditions could be selected. For example, in the
Southwest stable conditions occur most frequently in the winter.
2. Specific Measurements
A large number of measurements will be required to validate the plume
visibility model. The necessity of each measurement can be appreciated
by examining Figure 13 (Chapter III), which shows the schematic logic flow
-------
166
diagram of the visibility models. Most air quality measurement programs
and models are concerned only with the first two elements of the visibility
model—the emissions and the atmospheric transport, diffusion, and removal
processes. The desired measurements and output are time-averaged pollutant
concentrations at given ground-level locations. However, in the visibility
model and in measurements to validate visibility models, the desired result
is a light intensity, perceived by an observer at a given ground-level
location, which is affected by air pollutants some distance away from the
observer.
Thus, the most important single measurement necessary for the valida-
tion of visibility models is of the spectral light intensity I(A) for spe-
cific observer locations and lines of sight. The spectral light intensity,
a strictly physical parameter, can be translated to visibility-related
psychophysical parameters, such as luminance (Y), chromaticity (x, y),
contrast (C), and the color difference parameter (A£), by weighting the
light intensity by the spectral response characteristics of the three
different light sensors of the human eye. These psychophysical parameters
are directly related to what an observer sees and are necessary and suffi-
cient for quantitying visual range and atmospheric discoloration. The
multiwavelength telephotometer is the only instrument with which we are
familiar that can directly measure these psychophysical parameters. By
equipping the instrument with color filters corresponding to the spectral
response of the three light receptors of the human eye, one can measure
tristimulus values (X, Y, and Z) for a given line of sight and calculate
Y, x, y, C, and A£. The telephotometer can also be used to determine
visual range by measuring the contrast between a distant mountain and
the horizon sky. Since the light intensity of several lines of sight can
be measured with a single telephotometer at one location, three or four
ground-based telephotometer measurement stations might be sufficient for
a measurement program at a single point source. Station locations might
be in a preferred transport direction (in stable conditions) at distances
from the source of 20, 50, 100, and 150 km. If possible, some of these
-------
167
stations might be at vista points in Class I areas. Telephotometer sta-
tions can be moved fairly easily, depending on the transport of the plume,
to the most strategic locations and could be located on high terrain where
views in many directions are possible. It would be valuable to have sta-
tions on opposite sides of a plume so that the same sight path could be
sampled from different angles to study the effect of the scattering angle e.
One telephotometer could be constantly moved to obtain views of the plume
from several angles and distances to characterize fully the effects of
plume-observer geometry. For example, measurements of the plume could be
made at several locations along a road or highway to evaluate the effect
of the plume-observer distance and the angles between the line of sight
and the plume centerline and between the line of sight and the horizon.
Each telephotometer operator should take color photographs of the
scenes that he is measuring with the telephotometer for later documentation
of contrast, atmospheric coloration, and the positions of the plume and the
sampling aircraft. At the edge of the camera's field of view in each photo-
graph, a color test strip should be placed (in direct sunlight, if possible)
so that the quality of the development of the color film can be controlled
and checked. In cases of forward scatter (e < 90°) where the sun is in
front of the camera, the color test strip cannot be placed both in the field
of view and in direct sunlight. In such cases, the test strip can be photo-
graphed separately at some interval. It may be possible to maintain and
check color film development quality by calibration using the color test
strip only once per role of film. The feasibility of cross-checking the
color photographs with the multiwavelength telephotometer measurements
should be evaluated. Color time-lapse movies from strategic vista points
could also be taken.
To link the measured spectral intensity to air pollution, one must
know the aerosol and NOp concentrations along the specific sight paths.
Thus, airborne measurements of NOp concentration, scattering coefficient
(b ,), and aerosol size distribution will be necessary. Attempts should
-------
168
be made to make aircraft traverses of the plume as close to the lines of
sight used in the telephotometer measurements as possible. It is essen-
tial to measure the location of the plume relative to telephotometer
stations since coloration will depend on the proximity of the plume to
the telephotometer. Airborne measurements could be supplemented and
checked with correlation spectrometer measurements of N0? burden. Since
the plume optical depth due to NCL is directly proportional to the N0~
burden, this measurement would be valuable. The correlation spectro-
iiieter should be considered as an optional supplement to airborne
monitoring.
Direct and diffuse solar flux should be measured using a pyrheliometer/
pyranometer combination. The occurrence of cloud cover should be docu-
mented. The location of the sun should either be measured or calculated
so that the solar zenith angle and the scattering angle for all telephoto-
meter and photographic lines of sight can be calculated later. Alterna-
tively, scattering angles could be measured at the time of measurement.
It is imperative that the line of sight of each light intensity measure-
ment be specified and recorded; measurements of the scattering angle,
solar zenith angle, sight path azimuth, sight path elevation angle,
observer location, plume location, and plume dimensions fully describe
each line of sight.
An important part of the measurement program should be the determi-
nation of the production site, chemical composition, size distribution,
and causes of secondary aerosol production, particularly at large down-
wind distances. The secondary aerosol production rate could be determined
by calculating [SO=]/[S02], [NO^/LNO^, [bscat]/[S02], and [bscat]/[NOx].
Each of these ratios will increase with secondary aerosol formation. The
measured aerosol size distribution, chemical composition, mass concentra-
tion, and scattering coefficient should be cross-checked using Mie theory
and accounting for the cations and liquid water associated with sulfates
and nitrates. Hypotheses regarding the mechanisms of secondary aerosol
formation should be tested by looking at the time-dependent rate of
-------
169
aerosol formation. If possible, measurements of background and plume
ammonia and radical concentrations should be made to identify possible
fundamental reactions effecting gas-to-particle conversion. Some attempt
should be made to measure the refractive indices, including the imaginary
(absorption) components, of the background, fly ash, and secondary aerosols.
The conversion of NO to NCL in the plume should be measured. Ozone
concentrations in the background air and in the plume and ultraviolet radi-
ation should be measured to test the validity of the steady-state assump-
tion used in calculating plume N02 production.
Plume dispersion parameters (a , o2) should be calculated from the
measurements of peak plume concentrations. The wind field at the plume
centerline should be measured by pibal releases from several locations at
hourly or three-hourly intervals. The plume position and transit times
based on measured wind speed and direction at plume height should be com-
puted on a real-time basis and should be compared with actual plume position.
Real-time calculations of plume position in the field would be used by the
pilot and ground-level observers to determine the location of the plume,
and to direct the airborne plume measurements at night and at far down-
wind distances. These calculations could also be used in conjunction with
weather forecasts to relocate ground-based stations to optimize plume
impact measurement. Vertical temperature gradients should be measured
during the aircraft flights.
Finally, the emissions from the power plant must be measured accur-
ately. If possible, the following measurements should be made throughout
the measurement program: mass emission rates or flue gas concentrations
of S02, NO, N02» and fly ash, in-stack opacity, flue gas volumetric flow
rate, flue gas temperature, and flue gas oxygen concentration. It would
be desirable in simplifying the measurement program if the power plant
operated at constant capacity throughout the measurement program.
-------
170
3. Data Analysis, Assessment of Model Performance, and Model Refinement
Data collected during the measurement program must be reduced and com-
piled in a format useful for providing input data for the visibility computer
simulation models. After the plume visibility model is run, the calculated
SC>2> NO, N02, 0^, SO^, and NO" concentrations, scattering coefficients,
spectral light intensities, and visual effects (reduction in visual range,
plume perceptibility, and atmospheric discoloration) should be compared with
measurements.
Model calculations should be made based on the measured values of:
> Emission rates
> Upper air wind speed and direction
> Plume dilution (a , a )
> Secondary aerosol formation rates
> Aerosol size distributions
> Ambient conditions
> Geometry of sun, plume, and observer.
Calculations and measurements of the following parameters could be
compared:
> Pollutant concentrations
> [N02]/[NOX]
> bscat
> bscat/mass ratios
> Visual range
> Luminance (Y)
> Chromaticity (x, y)
> Perceptibility (A£)
> Contrast.
-------
171
In addition, SAI's plume perspective and color display techniques could be
used to create color renderings of certain vistas for comparison with color
photographs.
In comparing model calculations and measurements necessary for model
validation it is important to consider the errors that can occur in:
> Measurement of input parameters
> Measurement of output parameters
> Model formulation.
Since we want to test and validate the model formulation, it is essen-
tial that errors in measurement of input and output data be minimized, error
bounds established, and that all parameters necessary for defining model
input and output be measured.
Model performance can be evaluated using:
> Correlation coefficients.
> Differences between measured and calculated values: Either
mean or root-mean-square, and either absolute or relative
differences.
> Ratio of measured to calculated values.
> Regression statistics.
> Qualitative comparisons.
A thorough model evaluation may identify directions for model refinement.
Limits on model applicability and accuracy may also be established. After
further model refinement and development based on the information gained from
the comparison with measurements, it could be advantageous to test the model
again using another set of measurements, possibly from another emissions
source, and allowing no intermediate fine tuning of the model input parameters,
-------
172
D. FURTHER DATA ANALYSIS
We have obtained an extensive data base with which further analyses
of anthropogenic visibility impairment can be made, including:
> Nearly 500 station-years of National Weather Service (NWS)
visibility and meteorological data.
> About 10 station-years of National Park Service visibility
data.
> Holzworth mixing depth and mixed layer wind speed for all NWS
upper air stations in the United States for 1960 through 1964.
We recommend that further data analysis be coupled with the develop-
ment of a regional visibility model for the Southwest. The objectives of
this analysis would be to determine through an analysis of upper air flow
trajectories typical transport wind fields which could be used for re-
gional calculations. Temporal and spatial variations in visual range
could be studied in conjunction with calculated trajectories to determine
the transport of emissions from source areas to clean nonurban areas (for
example, southern Utah). Also, inferences could be made as to the rate of
sulfate formation and removal by studying trajectories, ground-level sul-
fate measurments, and visual range observations.
In this report, visual range has been shown to be correlated with
many variables. Correlations with meteorological variables as well as
diurnal and seasonal variations have been explored. However, these rela-
tionships have been studied one variable at a time, and no attempts have
been made to elucidate the simultaneous effects of many variables. In
statistical terms, only univariate analyses have been made thus far,
though multivariate analyses are needed.
A question of great interest is what combinations of meteorological
conditions are associated with poor visibility? This question could be
investigated using a multivariate classification technique, such as linear
-------
173
discriminant analysis [e.g., Gnanadesikan (1977), Morrison (1976)], which is
a method that assigns multivariate observations to one of several classes.
This is done by finding planes in the multidimensional space of the inde-
pendent variables (in this case, the meteorological variables) that opti-
mally divide the observations (visual ranges) into different classes (e.g.,
visual range between 80 and 100 km). Discriminant functions for subsets
of the data can be determined, for example, for different years or for the
period of the copper strike. Differences between discriminant functions
would indicate that the conditions associated with various visual ranges
were changing. Misclassification rates should be evaluated for each dis-
criminant function to indicate its utility as a predictor.
Another technique that could be applied to the data set to elucidate
the dependence of visual range on other variables is multiple regression.
However, since many of the independent variables that would be used in the
regression (e.g., the meteorological variables) are likely to be highly
correlated, some dimensionality-reducing method, such as factor analysis,
should be applied to these variables first. In this way, some smaller
number of uncorrelated surrogates could be generated, and the multiple re-
gression could be made more meaningful and useful. Alternatively, a vari-
able selection method such as Cp analysis (Daniel and Wood, 1971) could be
used in conjunction with the regression.
Intervention analysis can be applied to determine whether a sudden
shift in conditions was associated with a corresponding change in visual
range (Box and Tiao, 1965, 1975). This technique would be applicable
to quantifying the changes in visual range that may have occurred during
the copper strikes. This time series method allows for the nonindependence
of successive observations in evaluating a change in level of a series of
observations. (A t-test of the difference in levels before and after the
event would be invalid because of the dependence between successive ob-
servations.) The analysis proceeds by calculating a function of the ob-
servations, with a known statistical distribution, which estimates the
shift in level. Straightforward statistical inference then gives the
-------
174
level of significance of the observed shift. Application of time series
techniques should enable both an analysis of trends in visual range and
sudden shifts due to events such as the copper strike.
E. DEVELOPMENT OF A SOUTHWEST REGIONAL VISIBILITY MODEL
For several reasons, the Southwest is likely to be the first region
in which visibility regulations, which are to be promulgated by August
1979, are implemented. First, there are a large number of mandatory
federal Class I areas in the Southwest, including national parks of such
obvious scenic value as the Grand Canyon, Bryce Canyon, Canyonlands, and
Arches. Second, the existing visual range in clean areas of the Southwest
is probably the best of anywhere in the contiguous United States. Accord-
ing to our analysis of visibility data (see Appendix A), there are several
locations in the Southwest, notably in northern Arizona, Utah, Colorado,
and New Mexico, where visual range is often greater than 160 km (100 miles).
Indeed, based on nephelometer measurements in Bryce Canyon reported by
Charlson (private communication, 1978), visual range at times may approach
the Rayleigh scattering limit of 390 km (240 miles) in the Southwest.
Third, significant energy development is planned for the Southwest, par-
ticularly in Utah and Colorado. Several large coal-fired power plants
are currently being proposed to be located at sites in the Southwest, in-
cluding Harry Allen (2000 Mwe), Intermountain Power Project (3000 Mwe),
Warner Valley (500 Mwe), and Garfield (2000 Mwe). Fourth, several large
coal-fired power plants are currently in operation in the Southwest, some
of whose plumes have been observed from scenic vistas in national parks
such as Bryce Canyon and Mesa Verde. These plants include Four Corners
(2175 Mwe), Mohave (1500 Mwe), Huntington Canyon (800 Mwe), Navajo (2300
Mwe), and San Juan (1500 Mwe). Finally, several very large emissions sources
are located in the Southwest or are in the prevailing upwind direction from
the Southwest. These sources include the copper smelters, whose current
aggregate SO emissions are more than 3000 tons per day, and the metropolitan
/\
areas of Phoenix, Tucson, Las Vegas, Salt Lake City, and Los Angeles, from
which pollution may be transported to Southwest mandatory Class I areas.
-------
175
Since spectacular scenery in the Southwest is enhanced by generally
excellent visibility and since considerable development of coal resources
is planned or has already occurred, a regional visibility model that can
study and answer the following questions is needed:
> Will a proposed power plant have an impact on visual range
in mandatory Class I areas located 100, 200, 500, or even
1000 km away? Will yellow-brown haze be visible? What sit-
ing alternatives exist, and how much pollution abatement is
required?
> Does an existing power plant's plume have a significant im-
pact on visual range and atmospheric color in national
parks? Where? How much? How often? What pollution
abatement equipment is required (i.e., particulate, SOX,
or NOX control)?
> Does the existing copper smelter complex have to control
SO emissions further to reduce visibility impairment?
A
Where does SO get transported, how much is removed by
A.
natural atmospheric processes, and how much is converted
to sulfates?
> Are sulfate, nitrate, and organic aerosol*, emitted from
urban areas, such as Phoenix, Salt Lake City, or even Los
Angeles, transported to mandatory Class I areas in the
Southwest? Do they have a significant impact on visibility?
These questions have potentially significant technical, socioeconomic, and
political implications. For example, if anthropogenic pollution is found
to be the cause of significant visibility impairment in the Grand Canyon,
the question to be answered is which combinations of sources contribute--
urban areas, the copper smelters, or a nearby coal-fired power plant?
A Southwest regional visibility model should be developed to point toward
the answers to these questions and to provide guidance for critical decisions.
-------
176
REFERENCES
Alkezweeny, A. J., and D. C. Powell (1977), "Estimation of Transformation
Rate of S02 to $04 from Atmospheric Concentration Data," Atmos.
Environ.. Vol. 11, pp. 179-182.
American Society for Testing and Materials (1974), "Standard Method of
Specifying Color by the Munsell System," Dl535-68, Philadelphia,
Pennsylvania.
Arizona Department of Health Services (1977), "First Annual Report on
Arizona Copper Smelter Pollution Control Technology, Phoenix,
Arizona.
Atkinson, R., R. A. Perry, and J. N. Pitts, Jr. (1976), "Rate Constants
for the Reactions of the OH Radical with N02 (M = Ar and Ne) and
S02 (M = Ar)," J. Chem. Phys., Vol. 65, pp. 306-310.
Barrie, L., and H. W. Georgii (1976), "An Experimental Investigation of
the Absorption of Sulfur Dioxide by Water Drops Containing Heavy
Metal Ions," Atmos. Environ.. Vol. 10, pp. 743-749.
Bassett, H., and W. G. Parker (1951), "The Oxidation of Sulfurous Acid,"
J. Chem. Soc.. Pt. 2, pp. 1540-1560.
Bauer, E. (1973), "Dispersion of Tracers in the Atmosphere: Survey of
Meteorological Data," Paper P-925, Institute of Defense Analysis,
Arlington, Virginia.
Baulch, D. L., D. D. Drysdale, and D. G. Home (1973), Evaluated Kinetic
Data for High Temperature Reactions. Volume 2--Homogeneous Gas
Phase Reactions of the Hg-^-O;? System (CRC Press. Cleveland, Ohio).
Beilke, S., and G. Gravenhorst (1977), "Heterogeneous S02 Oxidation in
the Droplet Phase," International Symposium on Sulfur in the
Atmosphere, Dubrovnik, Yugoslavia.
Beilke, S., D. Lamb, and J. Mueller (1975), "On the Uncatalyzed Oxida-
tion of Atmospheric S02 by Oxygen in Aqueous Systems," Atmos.
Environ.. Vol. 9, pp. 1083-1090.
Benarie, M., A. Nonat, and T. Menard (1973), "Etude de la Transformation
de L'Anhydride Sulfureaux en Acide Sulfurique en Relation avec les
Donnees Climatologiques, dans un Ensembly Urbain in a Caractere
Industriel, Rouen," Atmos. Environ.. Vol. 7, pp. 403-421.
-------
177
Benson, S. W. (1968), Thermgcheniical Kinetics (John Wiley & Sons, New
York, New York).
Bergstrom, R. W. (1973), "Extinction and Absorption Coefficients of the
Atmospheric Aerosol as a Function of Particle Size," Beitra'ge Zur
Physik der Atmosphare, Vol. 46, pp. 223-234.
Bergstrom, R. W., and J. T. Peterson (1977), "Comparison of Predicted
and Observed Solar Radiation in an Urban Area," J. Appl. Meteor.,
Vol. 16, pp. 1107-1116.
Bergstrom, R. W., and R. Viskanta (1974), "Spherical Harmonics Approxima-
tion for Radiative Transfer in Polluted Atmospheres," Progress in
Astronautics and Aeronautics, Vol. 35, pp. 23-40 (MIT Press,
Cambridge, Massachusetts).
Boris, J. P., and D. L. Book (1973), "Flux-Corrected Transport--!: SHASTA
A fluid Transport Algorithm That Works," J Computational Phys.,
Vol. 11, pp. 38-69.
Box, G.E.P., and G. C. Tiao (1975), "Intervention Analysis with Applica-
tions to Economic and Environmental Problems, J. Amer. Stat. Assoc.,
Vol. 70, p. 70.
(1965), "A Change in Level of a Non-stationary Time Series,"
Biometrika, Vol. 52, Nos. 1 and 2, p. 181.
Briggs, G. A. (1969), "Plume Rise," Office of Information Services, U.S.
Atomic Energy Commission, Washington, D.C. (National Technical
Information Service TID-25075).
Brimblecomb, P., and D. J. Spedding (1974), "The Catalytic Oxidation of
Micromolar Aqueous SOg--!. Oxidation in Dilute Solutions Containing
Iron (III)," Atmos. Environ., Vol. 8, pp. 937-945.
Businger, J. A., et al. (1971), "Flux-Profile Relationships in the
Atmospheric Surface Layer," J. Atmos. Sci., Vol. 28, pp. 181-189.
Calvert, J. G., et al. (1978), "Mechanism of the Homogeneous Oxidation
of Sulfur Dioxide in the Troposphere," Atmos. Environ., Vol. 12
(in press).
(1977), "Mechanism of the Homogeneous Oxidation of Sulfur
Dioxide in the Troposphere," International Symposium on Sulfur
in the Atmosphere, 7-14 September 1977, Dubrovnik, Yugoslavia.
Cass, G. R. (1976), "The Relationship Between Sulfate Air Quality and
Visibility at Los Angelas," EQL Memorandum No. 18, California
Institute of Technology, Pasadena, California.
Chamberlain, A. C. (1966), "Transport of Gases to and from Grass and
Grass-like Surfaces," Proc. Roy. Soc.. Vol. A290, pp. 236-260.
-------
178
Charlson, R. J., and N. C. Ahlquist (1969), "Brown Haze: N02 or
Aerosol?," Atmos. Environ., Vol. 3, pp. 653-656.
Charlson, R. J., N. C. Ahlquist, and H. Horvath (1968), Atmos. Environ.,
Vol. 2, p. 455.
Chen, T., and C. H. Barren (1972), "Some Aspects of the Homogenous
Kinetics of Sulfite Oxidation," Ind. Eng. Chem. Fund., Vol. 11,
pp. 446-470.
Cheng, R. T., M. Corn, and J. I. Frohliger (1971), "Contribution to the
Reaction Kinetics of Water Soluble Aerosols and SO? in Air at ppm
Concentrations," Atmos. Environ., Vol. 5, pp. 987-1008.
Committee on Colorimetry, Optical Society of America (1963), "The Science
of Color," Washington, D.C.
Conner, W. D., and J. R. Hodkinson (1972), "Optical Properties and Visual
Effects of Smoke-stack Plumes," AP-30, Office of Air Programs,
Environmental Protection Agency, Research Triangle Park, North
Carolina.
Covert, D. S., R. J. Charlson, and N. C. Ahlquist (1972), "A Study of
the Relationship of Chemical Composition and Humidity to Light
Scattering by Aerosols," J. Appl. Meteor., Vol. 11, pp. 968-976.
Cox, R. A., and S. A. Penkett (1972), "Aersol Formation from Sulphur
Dioxide in the Presence of Ozone and Olefinic Hydrocarbons," J. Chem.
Soc., Vol. 68, p. 1735.
(1971a), "Photo-Oxidation of Atmospheric S02>" Nature. Vol. 229,
pp. 486-487.
(1971b), "Oxidation of Atmospheric S02 by Products of the Ozone-
Olefin Reaction," Nature, Vol. 230, pp. 321-322.
(1970), "The Photo-Oxidation of Sulphur Dioxide in Sunlight,"
Atmos. Environ., Vol. 4, p. 425.
Crutzen, P. J., and J. Fishman (1977), "Average Concentration of OH in
the Northern Hemisphere Troposphere, and the Budgets of CH., CO,
and HC," Geophys. Res. Letters, in press.
Daniel, C., and F. S. Wood (1971), Fitting Equations to Data (John Wiley
& Sons, New York, New York).
Dave, J. V. (1970), "Subroutines for Computing the Parameters of the
Electromagnetic Radiation Scattered by a Sphere," IBM System 360
Program 3600-17.4.002
-------
179
Davis, D. D., W. Heaps, and T. McGee (1976), "Direct Measurements of
National Tropospheric Levels of OH Via an Aircraft Borne Tunable
Dye Laser," Geophys. Res. Letters, Vol. 3, pp. 331-333.
Davis, D. D., G. Smith, and J. Klauber (1974), "Trace Gas Analysis of
Power Plant Plumes Via Aircraft Measurements: (h, NOX and SO?
Chemistry," Science. Vol. 186, pp. 733-736.
Durran, D. D., et al. (1978), "A Study of Long Range Air Pollution Problems
Related to Coal Development in the Northern Great Plains," Systems
Applications, Incorporated, San Rafael, California.
Egan, B. A., and J. R. Mahoney (1972), "Numerical Modeling of Advection
and Diffusion of Urban Area Source Pollutants," J. Appl. Meteor.,
Vol. 11, pp. 312-322.
Ensor, D. S., L. E. Sparks, and M. J. Pilat (1973), "Light Transmittance
Across Smoke Plumes Downwind from Point Sources of Aerosol Emissions,"
Atmos. Environ., Vol. 7, pp. 1267-1277.
EPA (1976a), "1973 National Emissions Report," EPA-450/2-76-007, National
Air Data Branch, Office of Air Quality Planning and Standards,
Environmental Protection Agency, Research Triangle Park, North
Carolina.
(1976b), "Monitoring and Air Quality Trends Report, 1974,"
EPA-450/1-76-001, Environmental Protection Agency, Washington, D.C.
(1976c), "Existing and Proposed Fuel Conversion Facilities
Summary," TS-5, EPA Region VIII, Denver, Colorado.
ER&T (1977), "Development of an Atmospheric Model for Sulfate Formation,"
Document No. P-1534, Environmental Research & technology, Incorporated,
Santa Barbara, California.
Ettinger, H. J., and G. W. Royer (1972), "Visibility and Mass Concentra-
tion in a Nonurban Environment," J. Air Poll. Contr. Assoc.. Vol. 22,
PP. 108-m.
Federal Power Commission (1976), "FPC Form 67: Steam Electric Plant Air
and Water Quality Control Data for the Year Ending December 31, 1975,"
Washington, D.C.
Forrest, J., and L. Newman (1977a), "Further Studies on the Oxidation of
Sulfur Dioxide in Coal-Fired Power Plant Plumes," Atmos. Environ..
Vol. 11, pp. 465-474.
(1977b), "Oxidation of Sulfur Dioxide in the Sudbury Smelter
Plume," Atmos. Environ., Vol. 11, pp. 517-520.
-------
180
Forsythe, G. E., and W. R. Wasow (1960), Finite-Difference Methods for
Partial Differential Equations (John Wiley & Sons, New York,
New York).
Foster, P. M. (1969), "The Oxidation of S02 in Power Station Plumes,"
Atmos. Environ., Vol. 3, pp. 157-175.
Freiberg, J. (1975), "The Mechanism of Iron Catalyzed Oxidation of S02
in Oxygenated Solutions," Atmos. Environ., Vol. 9, pp. 661-672.
(1974), "Effects of Relative Humidity and Temperature on Iron-
Catalyzed Oxidation of S02 in Atmospheric Aerosols," Environ. Sci.
Techno! . , Vol. 8, pp. 731-734.
Friedlander, S. K. (1977), Smoke, Dust and Haze: Fundamentals of Aerosol
Behavior (John Wiley & Sons, New York, New
Fuller, E. C., and R. H. Crist (1941), "The Rate of Oxidation of Sulfite
Ions by Oxygen," J. Am. Chem. Soc., Vol. 63, pp. 1644-1650.
Galbally, I. F., (1974), "Gas Transfer Near the Earth's Surface," Advances
in Geophysics. Vol. 18B, pp. 329-340.
Gartrell, F. E., F. W. Thomas, and S. B. Carpenter (1963), "Atmospheric
Oxidation of SO? in Coal -Burning Power Plant Plumes," Am. Ind. Hyg.
J_., Vol. 24, pp. 113-120.
Gerhard, E. R., and H. F. Johnstone (1955), "Photochemical Oxidation of
Sulfur Dioxide in Air," Ind. Eng. Chem., Vol. 47, p. 972.
Gnanadesikan, R. (1977), Methods for Statistical Data Analysis of Multi-
yariate Observations (John Wiley & Sons, New York, New York).
Grosjean, P., and S. K. Friedlander (1975), "Gas-Particle Distribution
Factors for Organic and Other Pollutants in the Los Angeles
Atmosphere," J. Air Poll. Contr. Assoc., Vol. 10, pp. 1038-1044.
Hagen, L. J., and N. P. Woodruff (1973), "Air Pollution from Duststorms
in the Great Plains," Atmos. Environ.. Vol. 7, pp. 323-332.
Halow, J. S., and S. J. Zeek (1973), "Predicting Ringelmann Number and
Optical Characteristics of Plumes," J. Air Poll. Contr. Assoc..
Vol. 23, pp. 676-684.
Hanson, J. E., and L. D. Travis (1974), "Light Scattering in Planetary
Atmospherics," Space Science Reviews, Vol. 16, pp. 527-610.
Heisenberg, W. (1948), "Zur Statistischen Theorie der Turbulenz," Z_.
Phys . , Vol. 124, p. 614.
-------
181
Hidy, G. M., et al. (1975), "Characterization of Aerosols in California
(ACHEX)," Science Center, Rockwell International, Thousands Oaks,
California.
Hodkinson, J. R. (1966), "Calculations of Colour and Visibility in
Urban Atmospheres Polluted by Gaseous N02," Air Water Poll. Int. J.,
Vol. 10, pp. 137-144.
Holzworth, G. C. (1972), "Mixing Heights, Wind Speeds, and Potential for
Urban Air Pollution Throughout the Contiguous United States," AP-101,
Office of Air Programs, Environmental Protection Agency, Research
Triangle Park, North Carolina.
Horvath, H. (1972), "Author's Reply: On the Brown Colour of Atmospheric
Haze," Atmos. Environ. , Vol. 6, PP- 143-148.
(1971), "On the Brown Colour of Atmospheric Haze," Atmos.
Environ., Vol. 5, pp. 333-344.
Husar, R. B., and W. H. White (1976), "On the Color of the Los Angeles
Smog," Atmos. Environ., Vol. 10, pp. 199-204.
Irvine, W. M. (1975), "Multiple Scattering in Planetary Atmospheres,"
Icarus, Vol. 25, pp. 175-204.
Jarman, R. T., and C. M. DeTurville (1969), "The Visibility and Length
of Chimney Plumes," Atmos. Environ., Vol. 3, pp. 257-280.
Jerskey, T. N., and C. S. Burton (1977), "The Prediction of Visibility
on a Regional Scale: The Status of Visibility Models and Measure-
ments," El77-23, Systems Applications, Incorporated, San Rafael,
California.
Johnstone, H. F., and D. R. Coughanowr (1958), "Absorption of Sulfur
Dioxide from Air: Oxidation in Drops Containing Dissolved Catalysts,"
Ind. Enq. Chem., Vol. 50, p. 1169.
Johnstone, H. F., and A. J. Moll (1960), "Formation of Sulfuric Acid in
Fogs," Ind. Eng. Chem., Vol. 52, p. 861.
Judd, D. B., and G. Wyszecki (1975), Color in Business Science and
Industry (John Wiley & Sons, New York, New York).
Junge, C. E. (1963), Air Chemistry and Radioactivity (Academic Press,
New York, New York]"!
Junge, C. E., and T. G. Ryan (1958), "Study of the S02 Oxidation in
Solution and Its Role in Atmospheric Chemistry," Quart.- J. Roy.
Meteor. Soc.. Vol. 84, pp. 46-55.
Katz
, M. (1950), "Photoelectric Determination of Atmospheric S02 Employ-
ing Dilute Starch-Iodine Solutions," Anal. Chem.. Vol. 22, p. 1040.
-------
182
LACAPCD (1971), "Profile of Air Pollution Control," Air Pollution Control
District, County of Los Angeles, Los Angeles, California.
Larson, T. V., N. R. Horike, and H. Harrison (1977), "Oxidation of Sulfur
Dioxide by Oxygen and Ozone in Aqueous Solution: A Kinetic Study
with Significance to Atmospheric Rate Processes," Atmos. Environ.,
in press.
Latimer, D. A., and G. S. Samuel sen (1978), "Visual Impact of Plumes
from Power Plants," Atmos. Environ., Vol. 12, pp. 1455-1465.
(1975), "Visual Impact of Plumes from Power Plants," UCI-ARTR-
75-3, UCI Air Quality Laboratory, School of Engineering, University of
California, Irvine, California.
Leighton, P. A. (1961), Photochemistry of Air Pollution (Academic Press,
New York, New York).
Levy, A. H., D. R. Drewes, and J. M. Hales (1976), "S02 Oxidation in
Plumes: A Review and Assessment of Relevant Mechanistic and Rate
Studies," EPA-450/3-76-022, Environmental Protection Agency,
Research Triangle Park, North Carolina.
Liu, M. K., and D. Durran (1977), "The Development of a Regional Air
Pollution Model and Its Application to the Northern Great Plains.,"
EPA-908/1-77-001, Systems Applications, Incorporated, San Rafael,
California.
(1976), "On the Modeling of Transport and Diffusion of Air
Pollutants over Long Distances," ER76-55, Systems Applications,
Incorporated, San Rafael, California.
Liu, M. K., M. A. Wojcik, and D. Henderson (1978), "Numerical Experiments
of the Effect of Lateral Diffusion on Mesoscale Pollutant Dispersion,"
Summer Computing Conference, 24-26 July 1978, Newport Beach,
California.
Matteson, M. J. (1978), "Capture of Atmospheric Gases by Water Vapor Con-
densation on Carbonaceous Particles," Conference on Carbonaceous
Particles in the Atmosphere, 20-22 March 1978, Lawrence Berkeley
Laboratory, Berkeley, California.
Matteson, J. J., W. Stober, and H. Luther (1969), "Kinetics of the
Oxidation of SOo by Aerosols of Manganese Sulfate," Ind. Eng. Chem.
Fund., Vol. 8, pp. 677-687.
McKay, H.A.C. (1971), "The Atmospheric Oxidation of Sulphur Dioxide in
Water Droplets in Presence of Ammonia," Atmos. Envi ron.. Vol. 5,
pp. 7-14.
Megaw, W. J. (1977), "Thin Layer Brown Haze," J. Aerosol Sci., Vol. 8,
pp. 21-29.
-------
183
Middleton, W.E.K. (1952), Vision Through the Atmosphere (University of
Toronto Press, Toronto, Canada).
Miller, J. M., and R. G. DePena, (1972), "Contribution of Scavenged
S02 to the Sulfate Content of Rain Water," J. Geophys. Res.,
Vol. 77, pp. 5905-5916.
Morrison, D. F. (1976), Multivariate Statistical Methods (McGraw-Hill,
New York, New York).
Munsell Color Company (1976), "Munsell Book of Color," Baltimore,
Maryland.
Newman, L., J. Forrest, and B. Manowitz (1975), "The Application of an
Isotopic Ratio Technique to a Study of the Atmospheric Oxidation
of Sulfur Dioxide in the Plume from an Oil-Fired Power Plant,"
Atmos. Environ., Vol. 9, pp. 959-968.
NGPRP (1974), "Atmospheric Aspects Work Group Report," Northern Great
Plains Resource Program, Denver, Colorado.
Niki, H. (1974), "Reaction Kinetics Involving 0 and N Compounds," Can.
J. Chem., Vol. 52, pp. 1397-1404.
Nixon, J. K. (1940), "Absorption Coefficient of Nitrogen Dioxide in the
Visible Spectrum," J. Chem. Phys.. Vol. 8, pp. 157-160.
Noll, K. E., P. K. Mueller, and M. Imada (1968), "Visibility and
Aerosol Concentration in Urban Atmospheres," Atmos. Environ.,
Vol. 2, pp. 465-475.
Nordlund, G. G. (1975), "A Quasi-Lagrangian Cell Method for Calculating
Long-Distance Transport of Airborne Pollutants," J. Appl. Meteor.,
Vol. 14, pp. 1095-1104.
Nordo, J. (1973), "Meso-Scale and Large-Scale Transport of Air Pollutants,"
Proc. Third International Clean Air Congress, B105-B108, Dusseldorf,
Federal Republic of Germany, VDI-Verlag.
Nordo, J., A. Eliassen, and J. Saltbones (1974), "Large-Scale Transport of
Air Pollutants," Advances in Geophysics. Vol. 18B, pp. 137-150.
Novakoy, T., S. G. Chang, and A. B. Marker (1974), "Sulfates in Pollu-
tion Particulates: Catalytic Oxidation of S02 on Carbon Particles,"
Science, Vol. 186, pp. 259-261.
Orel, A. E., and J. H. Seinfeld (1977), "Nitrate Formation in Atmospheric
Aerosols," Environ. Sci. Techno!., Vol. 11, pp. 1000-1007.
Payne, W. A., L. 0. Stief, and D. D. Davis (1973), "A Kinetics Study of
the Reaction of H02 with $03 and NO," J. Am. Chem. Soc.. Vol. 95,
pp. 7614-7619.
-------
184
Penkett, S. A. (1972), "Oxidation of SC>2 and Other Atmospheric Gases by
Ozone in Aqueous Solution," Nature, Phys. Sci., Vol. 240, pp. 105-106.
Penkett, S. A., and J. A. Garland (1974), "Oxidation of SO? in Artificial
Fogs by Ozone," Tell us, Vol. 26, pp. 284-290.
Pilat, M. J., and D. S. Ensor (1971), "Comparison Between the Light
Extinction Aerosol Mass Concentration Relationship of Amtospheric
and Air Pollutant Emission Aerosol," Atmos. Environ., Vol. 5,
pp. 209-215.
Roberts, E. M,, et al. (1975), "Visibility Measurements in the Painted
Desert Through Photographic Photometry," Engineering Bulletin 47,
Dames & Moore, Santa Barbara, California.
Roberts, P. T., and S. K. Friedlander (1975), "Conversion of $03 to
Sulfur Particulate in the Los Angeles Atmosphere," Environ. Health
Perspectives, Vol. 10, pp. 103-108.
Rodhe, H. (1972), "A Study of the Sulphur Budget for the Atmosphere over
Northern Europe," Tell us. Vol. 24, pp. 128-138.
Samuels, H. J., S. Twiss, and E. W. Wong (1973), "Visibility, Light
Scattering and Mass Concentration of Particulate Matter,"
California Air Resources Board, Sacramento, California.
Sander, S. P., and J. H. Seinfeld (1976), "Chemical Kinetics of Homogeneous
Atmospheric Oxidation of Sulfur Dioxide," Environ. Sci. Techno!.,
Vol. 10, pp. 1114-1123.
Schroeter, L. C. (1963), "Kinetics of Air Oxidation of Sulfurous Acid
Salts," J. Pharm. Sci.. Vol. 52, pp. 559-563.
Schwartz, S. E., and L. Newman (1978), "Processes Limiting Oxidation of
Sulfur Dioxide in Stack Plumes," Environ. Sci. Techno!., Vol. 12,
No. 1, pp. 67-73.
Scott, W. D., and P. V. Hobbs (1967), "The Formation of Sulfate in Water
Droplets," J. Atmos. Sci.. Vol. 24, pp. 54-57.
Scriven, R. A., and B.E.A. Fisher (1975a), "The Long-Range Transport of
Airborne Material and Its Removal by Deposition and Washout--
I. General Considerations," Atmos. Environ., Vol. 9, pp. 49-58.
(1975b), "The Long-Range Transport of Airborne Material and
Its Removal by Deposition and Washout--!!. The Effect of Turbulent
Diffusion," Atmos. Environ., Vol. 9, pp. 59-68.
Sellers, W. D. (1965), Physical Climatology (University of Chicago Press,
Chicago, Illinois).
-------
185
Smagorinsky, J. (1963), "General Circulation Experiments with the Primitive
Equations: I. The Basic Experiment," Mon. Wea. Rev.. Vol. 91,
pp. 99-164.
Spicer, C. W., and P. M. Schumacher (1977), "Interferences in Sampling
Atmospheric Particulate Nitrate," Atmos. Environ., Vol. 11, pp. 873-876.
Thorn, A. S. (1972), "Momentum, Mass and Heat Exchange of Vegetation,"
Quart. J. Roy. Met. Soc., Vol. 98, pp. 124-134.
Trijonis, J., and K. Yuan (1978), "Visibility in the Northeast: Visibility
Trends and Visibility/Pollutant Relationships," Technology Service
Corporation, Santa Monica, California.
(1977), "Visibility in the Southwest: An Exploration of the
Historical Data Base," Technology Service Corporation, Santa Monica,
California.
Urone, P., and W. H. Schroeder (1969), "S02 in the Atmosphere: A Wealth
of Monitoring Data but Few Reaction Rate Studies," Environ. Sci.
Techno!.. Vol. 3, pp. 436-445.
Urone, P., et al. (1968), "Static Studies of Sulfur Dioxide Reactions in
Air," Environ. Sci. Technol., Vol. 2, pp. 611-618.
van der Heuvel, A. P., and B. J. Mason (1963), "The Formation of Ammonium
Sulfate in Water Droplets Exposed to Gaseous S02 and NH3," Quart. J.
Roy. Meteor. Soc.. Vol. 89, pp. 271-275.
Van de Hulst, H. C. (1957), Light Scattering by Small Particles (John
Wiley & Sons, New York, New York).
Waggoner, A. P., and R. J. Charlson (1971), "Simulating the Color of
Polluted Air," Applied Optics. Vol. 10, pp. 957-958.
Waggoner, A. P., R. J. Charlson, and N. C. Ahlquist (1972), "Discussion:
On the Brown Color of Atmospheric Haze," Atmos. Environ., Vol. 6,
pp. 143-148.
Waggoner, A. P., et al. (1976), "Sulfate-Light Scattering Ratio as an
Index of the Role of Sulfur in Tropospheric Optics," Nature.
Vol. 261, pp. 120-122.
Weir, A., et al. (1976), "Factors Influencing Plume Opacity," Environ.
Sci. Techno!., Vol. 10, pp. 539-544.
(1975), "Measurement of Particle Size and Other Factors Influ-
encing Plume Opacity," International Conference on Environmental
Sensing and Assessment, 14-19 September 1975, Las Vegas, Nevada.
-------
186
Whitby, K. T., and G. M. Sverdrup (1978), "California Aerosols: Their
Physical and Chemical Characteristics," to be published in the
ACHEX Hutchinson Memorial Volume, Particle Technology Laboratory
Publication Number 347, University of Minnesota, Minneapolis,
Minnesota.
White, VI. H. (1977), "NOX-03 Photochemistry in Power Plant Plumes:
Comparison of Theory with Observation," Environ. Sci. Techno!.,
Vol. 11, No. 10, pp. 995-1000.
White, W. H., and P. T. Roberts (1975), "The Nature and Origins of
Visibility Reducing Aerosols in Los Angeles," 68th Annual Meeting,
Air Pollution Control Association, 15-20 June 1975, Boston,
Massachusetts.
Whitten, G. Z., and H. Hogo (1977), "Mathematical Modeling of Simulated
Photochemical Smog," EPA-600/3-77-001, Systems Applications,
Incorporated, San Rafael, California.
Williams, M. D., and R. Cudney (1976), "Predictions and Measurements of
Power Plant Plume Visibility Reductions and Terrain Interactions,"
Proc. of the Third Symposium on Atmospheric Turbulence Diffusion
and Air Quality, American Meteorological Society, pp. 415-420.
Williams, M. D., M. J. Wecksung, and E. M. Leonard (1978), "Computer
Simulation of the Visual Effects of Smoke Plumes," Society of Photo-
Optical Instrumentation Engineers Technical Symposium, 30 March 1978,
Washington, D.C.
Winkelmann, D. (1955), "Die Electrochemische Messung du Oxidations-
geschwindigkeit von Na?S03 Durch Gelosten Sauerstoff," Z. Elektrochemie,
Vol. 59, pp. 891-895.
Winkler, P. (1973), "The Growth of Atmospheric Aerosol Particles as a
Function of the Relative Humidity--!I. An Improved Concept of
Mixed Nuclei," Aerosol Sci., Vol. 4, pp. 373-387.
Yanenko, N. N. (1971), The Method of Fractional Steps (Springer-Verlag,
Berlin, Germany).
-------
187
GLOSSARY
accumulation mode--Submicron particles in the range 0.1 to 1.0 ym, which
are formed from smaller nuclei by coagulation or by direct conden-
sation of gases. This particle mode is most effective per unit mass
in light scattering. Secondary particles such as sulfates, nitrates,
and organics are found predominantly in this size range. Since coag-
ulation is relatively slow for particles larger than 1 ym, particles
in the accumulation mode are generally removed from the atmosphere by
precipitation and surface deposition before they grow larger.
aerosol—A suspension of fine solid or liquid particles in air or a gas.
In the context of visibility impairment, aerosol means a suspension of
nuclei, accumulation mode, and coarse particles (and gaseous pollutants)
in ambient air. Aerosol is also sometimes used to signify the liquid
or solid particles themselves.
anthropogenic—Caused directly by man or indirectly by man's technology
(e.g., anthropogenic pollutant emissions from combustion sources such
as automobiles and boilers).
a'ppearance--The subjective visual impression or aspect of a thing observed
by a human (e.g., the appearance of distant mountains or a plume).
atmospheric discoloration—An imprecise term describing the change in color
of the sky or distant mountains or clouds observed through the atmos-
phere due to natural or man-made pollution. The term implies that
there is an atmospheric color or set of colors that can be defined
as natural and not discolored. Examples of atmospheric discoloration
are white, grey, yellow, brown, or black haze or plumes.
background—An object being viewed by an observer (e.g., sky, cloud, or
mountain). Light reflected (if any) from the object and light scat-
tered and absorbed by the atmosphere along the sight path or line of
sight between the object and the observer determines the color per-
ceived by the observer. The nature of the background affects the
apparent coloration of a plume or haze layer.
back scatter—Situation where the sun is behind the observer (e>90°).
chroma—The numerical index in the Munsell color notation system that
describes the degree of saturation or the departure from grey. A
chroma of 0 is grey; a chroma of 2 is slightly colored; a chroma of
6 is highly colored.
-------
188
chromaticity diagram and coordinates—A mathematical description of the
relative spectral distribution (hue and saturation) of a given color.
Chromaticity coordinates uniquely describe the position of a given
color on a two-dimensional plot called a chromaticity plot. Chroma-
ticity coordinates provide information only on the relative mix of
light of different wavelengths, not the overall light intensity or
luminance. For example, chromaticity coordinates are not sufficient
to describe the difference between yellow and brown (a dark yellow).
C.I.E.--Commission Internationale de 1'Eclairage, or International Commission
on Illumination, a scientific organization responsible for setting stan-
dards for light and color measurement and specification.
Class I areas—Areas such as national parks, wilderness areas, and national
forests that are afforded the most stringent air quality protection
both in terms of the significant deterioration air quality increments
and the visibility protection and restoration mandates of the Clean
Air Act (Section 169). Mandatory Class I areas are those areas desig-
nated as such by the legislature; however, Class II areas may be redes-
ignated as Class I.
coagulation—The process of particle growth resulting from particle collision,
coarse particles—Particles larger than 1 ym, caused principally by grinding
and mechanical process (e.g., soil dust).
color difference parameter (A£)—An index quantifying the difference between
two colors, in terms of light intensity differences as we'll as chroma-
ticity differences, transformed such that equal values of AE correspond
to equally perceived differences. The parameter is useful to character-
ize the overall perceptibility of haze layers resulting from color
differences.
color solid or volume—A three-dimensional representation of color. A color
can be located in a color volume by specifying Munsell hue, value, and
chroma or by specifying overall light intensity or luminance (Y) and
chromaticity coordinates (x, y). The color difference parameter (A£)
may be visualized as a distance between two points in a color volume
which has been transformed such that"equal distances correspond to
equally perceived color differences.
contrast—The fractional difference in light intensities of two colors. The
contrast defined at specific wavelengths of light is useful in charac-
terizing color changes. The contrast between a black object and the
clear, horizon sky is used in defining visual range.
diffuse radiation—See "multiple scattering."
elevation angle (g)--Angle between the horizontal and the line of sight
(sight path).
-------
189
extinction coefficient (bext)"The derivative of transmitted light intensity
with respect to distance along a sight path of light attenuation due to
light scatter and absorption. The extinction coefficient is a function
of wavelength and depends on concentrations and characteristics of
aerosol particles and absorbing gases such as NO;?. The extinction
coefficient is the sum of the scattering coefficient (bscat) anc' the
absorption coefficient (baDS).
fly ash—Primary particulate matter emitted from furnaces and boilers,
usually consisting chiefly of silica with traces of oxides of metals
such as aluminum and iron. Plume opacity at the top of a stack is usu-
ally caused by fly ash.
forward scatter—Situation where the sun is in front of the observer so
that direct solar radiation is scattered less than 90° into the
observer's line of sight (e<90°).
hue—Index in the Munsell color notation system characterizing the dominant
coloration (e.g., red, green, or blue).
integrating nephelometry—A technique that measures the scattering coeffi-
cient of a small volume of air within a chamber.
Koschmieder relationship--The mathematical expression for calculating the
visual range in a homogeneous atmosphere:
r = 3.912
v bext
liminal contrast—The contrast that is barely perceptible. This contrast
threshold will depend on the observer and lighting conditions, but a
liminal contrast of 0.02 is common and is used in the definition of
visual range.
line of sight (or sight path)—The line connecting the observer and the
observed object. Particles and gases in the atmosphere along this
line will affect the perceived color of the object by absorbing light
and by scattering light into and out of the line of sight.
luminance (Y)--The overall light intensity within the visible spectrum,
weighted by the photopic response of the human eye.
Mie scatter!ng--The theory describing scattering of electromagnetic
radiation by spherical particles of diameters of the same order as
the wavelength (A) of the radiation. Rayleigh scattering theory
covers scattering by particles with diameters much shorter than A.
multiple scattering—Radiation that has been scattered more than once.
Single scattering results when direct solar radiation is scattered into
-------
190
the line of sight. Multiple scattering occurs when direct solar
radiation is scattered at least once by gases and particles not in
the line of sight or is reflected from the surface of the earth or
clouds before being scattered by particles and gases along the line of
sight. Diffuse radiation has been scattered or reflected at least
once, while direct solar radiation, as its name implies, is radiation
directly from the sun.
Munsell color notation—A system of describing a color quantitatively by
reference to three indices (hue, value, and chroma). The Munsell Book
of Color displays color paint chips at specific intervals of hue, value,
and chroma.
nuclei mode—Particles less than 0.1 ym in diameter, which are not effective
in light scattering, but grow by coagulation to the accumulation mode.
opacity—A term characterizing the optical thickness of an aerosol layer,
usually used to characterize smoke plumes in or near the stack. Opacity
is usually expressed in percent and is defined as
Opacity = 1 - Transmittance ,
where transmittance is related to the optical thickness T as follows:
Transmittance = e~T
optical thickness (T)—The integral of the extinction coefficient of an aero-
sol between two points along a given line of sight.
particulate matter—Small solid or liquid particles, consisting of many mole-
cules, that are suspended in air.
perceptibility—As used herein, the characteristic of an object that makes
it visible to a human observer. Perceptibility results from differ-
ences in light intensity and color between two objects. For example,
a distant mountain is perceptible because it is darker than the back-
ground sky. Air pollution is perceptible if color differences exist
between a plume and a background, a haze layer and a capping layer, or
between a haze and a recollection of a clear day. (See also "liminal
contrast").
phase function—See "scattering distribution function."
pollutant flux—The total mass of a pollutant species in a plume passing
through a plane perpendicular to the plume centerline per unit of time.
primary particulate—Particles emitted directly from an emissions source
(e.g., fly ash).
-------
19.1
Rayleigh scattering—The theory describing scattering of radiation by
molecules or particles much smaller than the wavelength of the radi-
ation, resulting from a dipole interaction with the electric field
of the radiation.
saturation—See "chroma."
scattering angle—The angle between the vector describing the unscattered
radiation on an object and the vector along the line of sight between
the object and the observer.
scattering distribution function (or phase function)--The function describing
the direction in which radiation is scattered. For aerosols the scat-
tering distribution function is largest in forward scatter (e<45°),
which explains why haze layers are bright when the sun is in front of
the observer.
2
solar flux--The intensity of solar radiation (watts/m ) incident on a given
plane perpendicular to the solar rays.
spectral light intensity—The light?intensity along a particular line of
sight at wavelength A (watts/m /steradian). The spectral light inten-
sity can be considered the increment of radiant energy of wavelength A
to A + dA passing through an elemental area dA within a solid angle dw
along the given line of sight:
telephotometry--A technique for measuring the light intensity of a distant
object using a photometer coupled to a telescope. By measuring the
differences in light intensity between the clear horizon sky and dis-
tant mountains, one can estimate the visual range.
transmissometry—A technique of measuring the transmission of light through
the atmosphere by which the overall extinction can be determined.
tristimulus values—Indices that describe a given color by indicating the
amount of red, green, and blue light needed to match the color. Tris-
timulus values (X, Y, Z) are keyed to the wavelength responses of the
three color sensors in the human eye. They can also be translated
into chromaticity coordinates (x, y).
value—The index in Munsell color notation related to brightness (luminance
or overall light intensity).
visibility—See "visual range."
visibility impairment—A reduction in visual range, the presence of atmos-
pheric discoloration, or both. The Clean Air Act Amendments of 1977
-------
192
refer to "any" and "significant" visibility impairment, terms which
have yet to be defined in regulations.
visual range (or visibility)--The distance at which a large, black object
is barely visible when viewed against the horizon sky. For calculations,
it is convenient to use a more strict definition of visual range, i.e.,
the distance at which the contrast between a black object and the clear
(cloudless) horizon sky is reduced to 0.02. When calculating contrast,
one should use overall light intensity or luminance (Y) or, as an approx-
imation, the spectral light intensity at X = 0.55 ym, which is the mid-
point of the visible spectrum and the wavelength to which the human eye
is most sensitive.
zenith angle—The angle between the solar beam and the vertical at a given
location on Earth.
-------
193
TECHNICAL fJEPOHT DATA
({'lease read linlruc nuns un il r rcrcnc liclnrc cotni'lvtittfl
1. REPORT NO. 2.
EPA-450/3-78-nOa,b,c
4. TITLE AND SUBTITLE
THE DEVELOPMENT OF MATHEMATICAL MODELS FOR THE
•PREDICTION OF ANTHROPOGENIC VISIBILITY IMPAIRMENT
7. AUTHOR(S)
D. A. Latimer, R. W. Bergstron, S. R. Hayes, M. K. Liu,
J. H. Seinfeld, G. Z. Whitten, M. A. Wojcik, M.d. Hillye
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications, Incorporated
950 Northgate Drive
San Rafael, California 94903
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Waterside Mall
401 M Street, S.W.
Washington, D.C. 20460
3. RECIPIENT'S ACCESSION-NO.
6, REPORT DATE
November 1 978
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
» EF78-68A,B,C
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
EPA 68-01-3947. and 68-02-2593
13. TYPE OF REPORT AND PERIOD COVERED
Rnal Report: 10/77 to 9/78
14. SPONSORING AGENCY CODE
EPA-OPE/OAQPS
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report describes a nine-month study to recommend and develop models that pre-
dict the contribution of man-made air pollution to visibility impairment in federal
Class I areas. Two models were developed. A near-source plume model based on a
Gaussian formulation was designed to compute the impact of a plume on visual range
and atmospheric coloration. A regional model was designed to calculate pollutant
concentrations and visibility impairment resulting from emissions from multiple
sources within a region with a spatial scale of 1000 km and a temporal scale of
several days. The objective of this effort was to develop models that are useful
predictive tools for making policy and regulatory decisions, for evaluating the
impacts of proposed new sources, and for determining the amount of emissions reduc-
tion required from existing sources, as mandated by the Clean Air Act Amendments
of 1977. Volume I of this report contains the main text; Volume II contains the
appendices; Volume III presents case studies of power plant plume visual impact for
a variety of emission, meteorological, and ambient background scenarios.
17. KEY WORDS AND DOCUMENT ANALYSIS
». DESCRIPTORS
Air quality modeling
Visual range
Atmospheric discoloration
Power plants
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b.lDENTIFlERS/OPEN ENDED TERMS
19. SECURITY CLASS (TltisKvpon)
UNCLASSIFIED
20. SECURITY CLASS (Tliit pancj
UNCLASSIFIED
c. COSATI i icid/Group
Vo. 'II--494J
22. PRICE
Vol. III--91
(PA farm 2220-1 (9-73)
-------
3 <£
0) Q)
33OO
CT -5 -£
y) — —
CD o n
Q3 fD ttl
>z
m > TJ m TI ~
§ 3 5.0,
I a
------- |