& EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-452/R-93-014
October 1993
Air
Effects of the 1990 Clean Air Act
Amendments on Visibility in Class I
Areas: An EPA Report to Congress
-------
ACKNOWLEDGEMENTS
This report is the product of the U.S. Environmental Protection
Agency's (EPA) Office of Air Quality Planning and Standards
(OAQPS). Bruce Polkowsky was project coordinator. Marc
Pitchford of the EPA Office of Research and Development (ORD)
drafted Chapter 2. The analysis of visibility changes for the
eastern U.S. was performed by Robin Dennis of the EPA ORD. The
preliminary national analysis and the southwestern U.S. analysis
were performed by.Systems Application International, Inc., under
contract with the OAQPS. The report incorporates comments from
other staff at the OAQPS, the ORD, the EPA Office of General
Counsel, and the EPA Regional Offices.
The work plan for the analyses which support the conclusions of
the report was reviewed by representatives from the Department of
Energy, the National Park Service, the U.S. Fish and Wildlife
Service, the U.S. Forest Service, and by environmental and
industrial groups including the Environmental Defense Fund and
the Utility Air Regulatory Group in December of 1991. The
preliminary results of the western modeling analysis were
presented at a public meeting of the Grand Canyon Visibility
Transport Commission in January 1993. An informal public review
of a draft of the entire report took place in August 1993.
Comments on this draft were received from the Utility Air
Regulatory Group, the Tennessee Valley Authority, Arizona Public
Service Company, and the U.S. Forest Service.
U.S. Environr:-'''' " ^r-tion Agency
Region 5, Li:,-: '•":
77 West Jp, .--.-d, 12th Floor
Chicago, IL 6uow
ii
-------
-------
Executive Summary
In this report the EPA has assessed changes in regional
visibility expected from implementation of the many provisions of
the Clean Air Act Amendments of 1990. 'This report provides an
estimate of future regional visibility conditions for the
contiguous 48 United States.
This report used a layered approach to assessing visibility
changes. A preliminary assessment was made choosing key
locations and a simple emissions-driven air quality assessment to
ascertain areas likely to see changes in the distribution of man-
made visibility-impairment related pollutants. This preliminary
analysis pointed towards a more focused approach for the Eastern
U.S., concentrating mostly on changes in sulfur dioxide
emissions. For the Southwestern U.S., the mixture of pollutants
is more varied and less dominated by sulfur particles. Thus the
analysis incorporated changes in emissions of sulfur, nitrogen,
organic and primary particulate matter.
Figure ES-1 is a map showing the locations of the Federal
mandatory class I areas. Figure ES-2 shows the regions used in
preliminary analysis. Figure ES-3 highlights which of these
regions were expected to see perceptible changes in visibility
conditions. Figure ES-4 shows the general domains of the more
advanced regional air quality modeling used in this report.
Figure ES-5 shows the current annual average visibility
conditions expressed in standard visual range (kilometers). For
the class I areas of the rural southwestern U.S., the annual
average visibility conditions generally result from 20 to 40
percent natural causes (gases of the atmosphere, estimated
natural fine particles, and coarse particles) and 60 to 80
percent man-made concentrations of aerosols and gases. For the
rural eastern U.S., the annual average visibility conditions
generally result from 10 to 30 percent natural and 70 to 90
percent man-made concentrations. These relative levels do not
apply for class I areas near (within 100-200 km) large urban
areas, nor for class I areas directly on the sea coasts.
Estimates of emissions changes resulting from implementation of
the 1990 Amendments were used to model changes in the man-made
portion of the visibility impairment aerosol concentrations for
two regions of the country. Seasonal changes were estimated
since atmospheric processes which form visibility impairing
particles and resulting levels of extinction of those particles
(due to humidity) vary by season.
Figure ES-6 shows the estimated change (1988-2005 in the
southwest, 1985-2010 in the east) in annual average standard
visual range along with a perception index (deciview) for all
iii
-------
I
i
-------
CO
CO
CO
JS
O
2
d>
•o
0)
u_
I
(0
•o
m
o
a.
(0
co
LU
§
O)
-------
CO
'&
(0
c
J
s>
Q_
0)
CO
CO
o
'5
0)
CVI
CO
111
§
O)
iZ
-------
0)
£
c
•*-»
0)
(0
Q.
(0
o
CO
-------
O)
T3
O
03
3
a
^
<
•e
E
o
O
.2
Q.
2
O)
o
0)
O
CO
UJ
§
O)
U-
-------
i
.8
(0
Q
~O
tn
s
Q)
a
8.
viii
T3
-------
&
sr
S.
t
I
ix
co
0)
O)
E
CD
(0
0)
I
2
o.
E
Q.
C3)
CD £
9- <
UJ |:
CO &'
co
CO
J22
O
CD
CO
LU
I
D)
U.
-------
selected class I areas in the modeling domains that are estimated
to be perceptible. Nearby class I areas should see similar
levels of change. An estimate of 1 deciview change is considered
to be just perceptible. All areas outside of the southwestern
and eastern modeling areas were estimated, based on the
preliminary analysis, to have imperceptible changes in regional
visibility due to the Clean Air Act Amendment programs. The
class I areas inside of the southwestern and eastern modeling
areas not indicated in Figure ES-6 are estimated to have
imperceptible changes in annual average or seasonal regional
visibility conditions. Figures ES-7 and ES-8 show the estimated
visibility change in winter and summer seasons, respectively.
In summary, class I areas from Maine to Georgia are estimated to
see improvements in regional visibility conditions. It is very
important to note that, under the expected implementation of the
Clean Air Act Amendments, no class I areas are estimated to have
perceptible decreases in regional visibility. Winter and summer
seasons show improvements ranging from just perceptible to six
times perceptible. The major improvements expected for class I
areas are for those along the central and southern portions of
the Appalachian Mountains. A change in the average regional
visibility means that some individual days will likely see
dramatic improvements in visibility, perhaps nearly eliminating
man-made impairment due to sulfate episodes. However, the
current models are not reliable enough to predict the exact
magnitude of specific single day events. To better ascertain
events of very good visibility and very bad visibility the entire
distribution of very good to very bad visibility conditions of
the modeling analysis were reviewed. The estimated improvement
in Eastern regional visibility are directly related to provisions
of the Clean Air Act Amendments addressing control of sulfur
dioxide emissions in the East.
Uncertainties in the analysis are mainly from emissions estimates
of the species of pollutants that result in particles that impair
regional visibility. This report only estimated emissions
reductions mandated by the Clean Air Act Amendments to the extent
that those emissions could be quantified. In many urban areas of
the East and West, air pollution management measures may be put
into effect to meet national ambient air quality standards for
ozone and particulate matter that will reduce, pollutants or
prevent the level of growth estimated here. The effect of those
programs would be to increase the regional visual air quality.
Efforts in the Los Angeles area could result in perceptible
improvement in regional visibility levels at class I areas in the
closest class I areas. Overall, over the time horizon of this
study, regional visibility should improve or remain stable across
the continental U.S.
Geographically, the largest uncertainties in predicting visual
air quality, particularly for the sensitive western areas, are
-------
I
I
l!!»/2
r
3
0)
O)
2
(D
CO
(0
0)
CO
2
o
O
(D
,
fit
3:9
O) CO
xi
-------
«
^
o
CO
01
0)
CO
^1
Q- §?
« ff
O) fl
.E li
•G *i
a to
-------
sources of emissions from Mexico. The largest uncertainty for
pollutant type is in estimates of changes in organic particle
concentrations, especially those formed secondarily (in the
atmosphere) from emissions (natural and man-made) of gaseous
volatile organic compounds. Bounding analyses to identify
plausible ranges of man-made contributions to this component of
visibility impairment were preformed for the more sensitive
western areas. These analyses indicated that the range of
plausible changes in man-made secondary organic material would
likely not result in perceptible regional visibility changes.
Although visibility will improve in many eastern class I areas,
based on estimates of the natural annual average visibility
developed for the National Acidic Precipitation Assessment
Program, there will still be perceptible man-made regional - -
visibility impairment in all class I areas nationwide.
xiii
-------
Table of Contents
Chapter 1. Introduction 1
Scope of Report : l
Chapter 2. Defining Visibility and Its Measures 4
What is "Visibility"? 4
Visibility Metrics 7
Chapter 3. Historical Perspective and Current Conditions . . 9
Trends of Visibility 9
Current Conditions at Class I Areas 9
Causes of Current Visibility Conditions 14
Chapter 4. Preliminary Estimation of Future Conditions ... 17
1990 Clean Air Act Amendments 17
Identification of Key 1990 CAAA Provisions 21
Preliminary Analysis of Visibility Changes in the
Federal Class I Areas 21
References for Chapter 4 31
Chapter 5. Assessment of the Eastern U.S 32
The National Acid Precipitation Assessment Program ... 32
Reanalysis Using the RADM Engineering Model 33
Emissions Inventory Revisions 33
Air Quality Model and Visibility Assessment .... 37
Estimating the Effects of Changes in NOX 38
Estimating the Effects of Changes in VOC 47
Summary 47
Comparison with Other Studies 48
References for Chapter 5 50
Chapter 6. Southwestern U.S. 51
Emissions Inventory Development 51
Data Sources 53
Emission Inventory Processing 53
Preparation of 1988 Inventory 54
Preparation of 2005 Base Case Inventory 55
Preparation of 2005 CAAA Case Inventory 58
Summary Comparison of Emission Inventories .... 61
Air Quality Modeling . . . 66
Regional Transport Model 67
RTM-II Application 67
Meteorological and Data Input Preparation 68
Air Quality Model Results 69
Estimation of Visibility 70
Construction of Extinction Budgets ..'...... 70
Visual Range and Deciview Estimates 71
Southwestern Modeling Conclusions 71
Uncertainties 80
xiv
-------
f
f
Emissions 80
Role of Secondary Organic Aerosols 80
References for Chapter 6 84
Chapter 7. Comparison of Modeling Approaches 88
Comparison of Eastern and Southwestern Approaches ... 88
Estimates of Current and Future Emissions 88
Comparison between RTM-II and RADM Visibility
Modeling 89
Sensitivity of Results , , = 92
xv
-------
List of Figures
Figure 2-1. Diagram of Visibility Concepts 5
Figure 3-1. Map of Mandatory Federal Class I Areas .... 10
Figure 3-2. National Weather Service Visibility Trends . . 11
Figure 3-3. Average Visibility (deciview) at Selected
Monitored Class I Areas 12
Figure 4-1. Location of Tabulated Class I Areas and
Regions used in the Preliminary Visibility
Analysis 28
Figure 5-1. Geographical Distribution of Light Extinction
from the 1985 Baseline to NAPAP S4 2010
Emissions Estimate 34
Figure 5-2. Projected S02 Emissions for 1990 CAAA,
.Eastern U.S 35
Figure 5-3. Projected NOx Emissions for 1990 CAAA,
Eastern U.S 36
Figure 5-4. Estimated Baseline Annual Average Visibility . 39
Figure 5-5. Estimated 2010 Annual Average Visibility ... 40
Figure-5-6. Ten Percent of 1985 Days Estimated at Shown
Visibility or Worse 41
Figure 5-7. Ten Percent of 1985 Days Estimated at Shown
Visibility or Better 42
Figure 5-8. Improvement from 1985 to 2010 of the Most
Impaired Days 43
Figure 5-9. Improvement from 1985 to 2010 of the Least
Impaired Days 44
Figure 5-10. Improvement from 1985 to 2010 Average Warm
Season (Apr-Sep) Visibility 45
Figure 5-11. Improvement from 1985 to 2010 Average Cold
Season (Oct-Mar) Visibility 46
Figure 6-1. Southwestern Modeling Domain and Location of
Class I Areas 52
Figure 6-2. Estimated Total Gridded Annual Emissions for
the Southwestern Analysis (1988, 2005 Base,
2005 Control) 62
Figure 6-3. Estimated Total Gridded Annual NOx Emissions
by Source Category for the Southwestern
Analysis (1988, 2005 Base, 2005 Control) ... 63
Figure 6-4. Estimated Total Gridded Annual S02 Emissions
by Source Category for the Southwestern
Analysis (1988, 2005 Base, 2005 Control) ... 64
Figure 6-5. Estimated Total Gridded Annual Fine Particle
Emissions by Source Category for the
Southwestern Analysis (1988, 2005 Base, 2005
Control) 65
Figure 6-6. Estimated Visibility (visual range and -
deciview) at Grand Canyon National Park ... 72
Figure 6-7. Estimated Visibility (visual range and
deciview) at Arches National Park 73
Figure 6-8. Estimated Visibility (visual range and
deciview) at San Gorgonio Wilderness 74
xvi
-------
I Figure 6-9. Estimated Visibility (visual range and
i deciview) at Rocky Mountain National Park . . 75
I Figure 6-10. Estimated Visibility (visual range and
I deciview) at Pinnacles Wilderness 76
- Figure 6-11. Estimated Visibility (visual range and
deciview) at Yosemite National Park ..... 77
i Figure 6-12. Estimated Visibility (visual range and
' deciview) at Chiricahua Wilderness ...... 78
Figure 6-13. Estimated Visibility (visual range and
deciview) at Bandelier Wilderness 79
xvii
-------
Table 3-1,
Table 3-2.
Table 3-3.
Table 4-1.
Table 4-2.
Table 4-3.
Table 5-1.
Table 5-2.
Table 6-1,
Table £-2.
Table 6-3,
Table 6-4,
List of Tables
Comparison of Measured and Reconstructed
Visual Range (IMPROVE March 1988-1991) ... 13
Anthropogenic Visibility (Extinction) Budgets
by Rural Regions of the Country 15
Annual Averages (March 1988-February 1991) of
Reconstructed Light Extinction (Mm"1) for 19
Regions of the IMPROVE Network 16
NO,, SOX, and TSP Emission Control Provisions
Contained in 1990 CAAA 22
List of Mandatory Federal Class I Areas ... 25
Summary of Preliminary visibility Analysis
Results .' 30
Projected Reactive Hydrocarbon Emissions for
•1990 CAAA, Eastern U.S 37
Summary of Visibility Changes for Selected
Class I Areas (Based on RADM Reconstructed
Sulfate plus Non-Sulfate Correction Factor) . 48
Smelter SOX Emissions 56
Eight Largest Power Plants Annual SO,
emissions 57
Ozone Nonattainment Areas 59
Summer and Winter OC/EC Data and Estimated
Summer SOA Concentrations at Selected IMPROVE
Sites 82
xviii
-------
-------
Definitions
Acid precipitation - typically is rain with high concentrations
of acids produced by the interaction of water with oxygenated
compounds of sulfur and nitrogen which are the by-products of
fossil fuel combustion.
Aerosols - gaseous suspension of ultramicroscopic particles of a
liquid or a solid. Atmospheric aerosols govern variations in
light extinction and, therefore, visibility reduction. Aerosol
size distribution and chemistry are key parameters.
Anthropogenic - refers to alteration to the natural environment
caused by human activity, i.e., man-made.
Apportionment - the act of assessing the degree to which specific
components contribute to light extinction or aerosol mass.
Atmospheric clarity - is an optical property related to the
visual quality of the landscape viewed from a distance (see
optical depth and turbidity).
Contrast transmittance - contrast transmittance is the ratio
between apparent and inherent spectral contrast. When the object
is darker than its background, it has a value between 0 and -1.
For objects brighter than their background, the value varies from
0 to infinity. When the contrast transmittance is equal to zero,
the object cannot be seen.
Current conditions - refer to contemporary, or exiting,
atmospheric conditions that are affected by human activity.
Deciview - The scale used in a haziness index designed to be
linear to humanly-perceived changes in visibility caused solely
by air quality changes. The deciview scale is near zero for
pristine atmospheric conditions and increases as visibility
degrades. The perceptibility of a change in haziness of a scene
corresponding to two pollutant levels is proportional to the
difference in the deciview values for those two levels. A 1
deciview (dv) change corresponds to a 10% change in light
extinction, which is thought to be a small but perceptible
visibility change. Larger changes in perceived haziness are
indicated by greater deciview differences.
Deliquescence - the process that occurs when the vapor pressure
of the saturated aqueous solution of a substance is less than the
vapor pressure of water in the ambient air. Water vapor is
collected until the substance is dissolved and in equilibrium
with its environment.
xix
-------
Edge sharpness - describes a characteristic of landscape
features. Landscape features with sharp edges contain scenic
features with abrupt changes in brightness.
Equilibration - a balancing or counter balancing to create
stability, often with a standard measure or constant.
Hydrophobia - lacking affinity for water, or failing to adsorb or
absorb water.
Hygroscopic - an ability or tendency to rapidly accelerate
condensation of water vapor around a nucleus. Also pertains to a
substance (e.g., aerosols) which have an affinity for water and
whose physical characteristics are appreciably altered by the
effects of water.
Kosdmeider constant - the constant in the reciprocal
relationship between standard visual range and the extinction
coefficient (see standard visual range).
Light extinction (Extinction ctJeMYcient) - the attenuation of
light per unit distance due to absorption and scattering by the
gases and particles in the atmosphere.
Mie scattering - the attenuation of light in the atmosphere by
scattering due to particles of a size comparable to the
wavelength of the incident light. This is the phenomenon largely
responsible for the reduction of atmospheric visibility. Visible
solar radiation falls into the range from 0.4 to 0.8 (OR, roughly,
with a maximum intensity around 0.52 fun.
Natural conditions - refer to the prehistoric distribution of
atmospheric states, i.e., atmospheric conditions that are not
affected by human activities.
Nephelometer - an instrument used to measure the light scattering
component of light extinction.
Optical depth - the degree to which a cloud or haze prevents
light from passing through it. It is a function of physical
composition, size distribution, and particle concentration.
Often used interchangeably with "turbidity".
Path radiance - or "airlight", is a radiometric property of the
air resulting from light scattering processes along the sight
line, or path, between a viewer and the object (target).
Primary particles - primary particles are suspended in the
atmosphere as particles from the time of emission (e.g., dust and
soot).
XX
-------
Rayleigh scattering - refers to the scattering of light by air
molecules, also called blue-sky scatter.
Regional haze - the haze that uniformly covers a broad (2000 to
7000 square kilometers) geographic area made up of natural and
man-made components.
Secondary particles - are formed in the atmosphere by a gas-to-
particle conversion process.
Standard visual range - is the reciprocal of the extinction
coefficient. The distance under uniform daylight lighting
conditions at which the apparent^contrast between a specified
target and its background becomes just equal to the threshold
contrast of an observer, assumed to be 0.02.
Threshold contrast - a measure of human eye sensitivity to
contrast. It is the smallest increment of contrast perceptible
by the human eye.
Transmissometer - an instrument that measures atmospheric
transmittance. From transmittance, the atmospheric extinction
coefficient can be derived.
Transmittance - the fraction of initial light from a light source
that is transmitted through the atmosphere. Light is attenuated
by scattering and absorption from gases and particles.
Turbidity - a condition that reduces atmospheric transparency to
radiation, especially light. The degree of cloudiness, or
haziness, caused by the presence of aerosols, gases, and dust.
Visibility - refers to the visual quality of the view, or scene,
in daylight with respect to color rendition and contrast
definition. The ability to perceive form, color, and texture.
Visibility indexes - have been formalized for aerosol, optical,
and scenic attributes. Aerosol indexes include mass
concentrations, particle compositions, physical characteristics,
and size distributions. The optical indexes include coefficients
for scattering, extinction, and absorption. Scenic indexes
comprise visual range, contrast, radiance, color, and deciview.
Visibility reduction - is the impairment or degradation of
atmospheric clarity. Becomes significant when the color and
contrast values of a scene to the horizon are altered or
distorted by airborne impurities.
Visual air quality - Refers to the influence of air quality on
atmospheric visibility. For a specific scene, changes in
lighting conditions (e.g., position of the sun) and the inherent
appearance of the scene (e.g., adding clouds or snow cover) are
xxi
-------
important factors that influence perceived visibility. However,
they are purposely held constant when evaluating visual air
quality effects. Visual air quality is the only aspect of
visibility addressed by the Clean Air Act.
XXIX
-------
Chapter 1. Introduction
Scope of Report
This report is in response to the requirement of Section l69B(b)
of the Clean Air Act (Act). Section 169B was added to the Clean
Air Act in 19901 as a supplement to Section 169A which
established the general provisions for special protection of
class I Federal areas2 from man-made visibility impairment.
Section 169B(b) requires an assessment and report to Congress on
the progress and improvements expected in visibility in class I
Federal areas from implementation of the provisions of the 1990
Clean Air Act Amendments (CAAA) (other than the provisions of
Section 169B itself). Other parts of Section 169B focus on
conducting additional visibility research and monitoring, and the
formation of visibility transport regions to address "regional
haze" impairment of certain class I Federal areas through
commissions comprised of affected States and Federal Agencies.
In Section- 169A of the Act, Congress established as a national
goal "the prevention of any future, and the remedying of any
existing impairment of visibility in mandatory class I Federal
areas which impairment results from man-made air pollution."
(95th Congress, 1977). Section 169A(a)(3) directed the EPA to
write a report to Congress on methods for implementing"the
national goal. In 1979, the EPA issued its report (1979 Report)3
which, among other things, reviewed the mechanisms of human
perception and fundamentals of atmospheric visibility impairment
and identified various types of visibility impairment. The basic
science and mechanisms of visibility impairment formation are not
reviewed in detail in this report. However, a brief review of
the historical trends and current conditions of regional
visibility impairment are provided for perspective in reviewing
the estimated future conditions.
1 On November 15,, 1990, significant amendments to the Clean
Air Act were signed into law. Pub L. No. 101-549, 104 Stat. 2399.
Section 816 of the 1990 Clean Air Act Amendments added section 169B
to the Act. 104 Stat. at 2695-97.
2 International Parks, National wilderness and National
Memorial Parks exceeding 5,000 acres in size, and National Parks
exceeding 6,000 acres as provided in section 162(a) of the Act.
3 "Protecting Visibility, An EPA Report to Congress", U.S.
EPA, Office of Air Quality Planing and Standards, October 1979,
EPA-450/5-79-008
-------
This report will focus only on regionally homogeneous haze or,
more commonly, "regional haze." In this report, regional haze is
defined to be uniform haze extending over a geographic area of
2000 to 7000 square kilometers. This report does not focus on
t layered haze events or other site specific impairments at class I
I areas. It is possible that reductions'in air pollution resulting
| from the Clean Air Act Amendments of 1990 will result in
I improvements of visibility impairment associated with local
? source impacts on specific class I areas. However, without exact
| knowledge of specific source emissions changes cind detailed
r reviews of the local meteorology no meaningful estimate of
r changes could be produced. In addition, as part of implementing
existing visibility regulations, the EPA has recently concluded a
; regulatory review process to address visibility impairment in
class I areas that is "reasonably attributable" to nearby sources
of pollution. /This review found few sources of such impairment
based on currently available information. While new information
may reveal additional instances of existing "reasonably
attributable" impairment in class I areas, these will be
infrequent cases. Most impairment at class I areas results from
long-range transport of pollution that manifests in a regional,
largely homogeneous haze.4
Given limited resources, this report follows the recent guidance
of the National Academy of Sciences report on regioa^. visibility
by applying a simple approach to estimating visibility conditions
and changes across the U.S. Then the report uses more complex
estimation techniques for regions of the country that are likely
to have perceptible changes in visibility resulting from changes
in pollutant emissions.3
The Clean Air Act Amendments of 1990 and the technical analyses
accompanying their development indicate that visibility
improvements are expected as a corollary benefit from certain
programs, notably the provisions for addressing acidic
deposition. As specified in section 169B(B), this report
summarizes the changes in visibility expected from implementation
of the Clean Air Act Amendments of 1990 without projecting
improvements resulting from actions that may be taken under
section 169B.
Chapter 2 reviews the measures of visibility impairment used in
this report. The measures are estimated spatially (grid cells in
a modeling domain) and temporally (warm and cold seasons) to
4 Protecting Visibility in National Parks and Wilderness
Areas. Committee on Haze in National Parks and Wilderness Areas,
National Research Council, National Academy of Sciences, January,
1993, pg. 6.
5 Ibid. pg. 7.
-------
develop a basic understanding of the distribution of visibility
changes. Chapter 3 reviews the pollutants related to visibility
impairment and the trends of those pollutants and visibility
impairment to lend perspective on future estimates. Chapter 4
covers a preliminary estimate of future national visibility
trends. This preliminary estimate guided the allotment of
resources for the air quality modeling discussed in Chapters 5
and 6. Chapter 5 reviews changes in visibility expected for the
eastern U.S. This review is based on a regional model developed
by the EPA for the National Acidic Precipitation Assessment
Program (NAPAP). Chapter 6 reviews a modeling effort to
characterize the expected change in visibility in the
southwestern U.S. Chapter 7 explores the differences between the
analyses for all regions of the country and explains the
limitations on ..analyses and conclusions.
-------
-------
Chapter 2. Defining Visibility and Its Measures
What is "Visibility"?
Historically, visibility has been defined in a number of ways
depending on the specific application of the term. For the
purpose of visibility protection in class I Areas, visibility can
be defined as the effects of atmospheric constituents on the
appearance of a scene viewed from a distance. However, since
non-air quality factors can influence the appearance of a scene
viewed from a distance, some have chosen to use the term visual
air quality to more clearly specify which aspects of visibility
are of interest.
For an object to be seen against a background, there must be
sufficient contrast between the object and its background. That
is to say that the light from the object and the background must
be sufficiently different in apparent brightness or color to make
the object stand out against the background. Light from objects
and their background viewed through the atmosphere from a
distance are modified by the constituents of the atmosphere.
Figure 2-1 is a schematic of the processes involved. Light as it
traverses the atmosphere is scattered (i.e., redirected in a
random direction) and absorbed (i.e., converted from light to
heat) by the particles and gases in the atmosphere. This affects
the appearance of scenes in two ways. The image-forming light
(also called image radiance) from scenic features is diminished
since a fraction of the light is scattered or absorbed; non-
image-forming light (also called path radiance) is scattered into
the sight path. Both of these effects lower the contrast between
object and background, and cause the scene to be more obscured.
This contrast reduction increases with distance to the scenic
feature being viewed, and with increased concentration of
constituents responsible for scattering and absorbing light.
All atmospheric particles and gases scatter light and some of
them absorb light. Scattering of light by gas molecules
(including the nitrogen and oxygen in pure air), referred to as
Rayleigh scatter, is responsible for a limit on visibility, even
for an atmosphere, that has no air pollution. The only gas that
absorbs visible light to any appreciable extent at concentrations
expected in the atmosphere is nitrogen dioxide which absorbs
preferentially in the blue resulting in a yellowish discoloration
when looking through sufficient quantities of the gas. Regional
concentrations of nitrogen dioxide are too low to be of concern
to visibility, though in urban and combustion plumes they may be
important.
Light scattering, and to a lesser degree, light absorption by
suspended particles are the most important contributors to
visibility degradation. The influence of particles depends on
-------
Figure 2-1.
Diagram of Visibility Concepts
(Source: NAPAP, Sute of Science/Technology Report «24 (1990)
OBSERVED
IMAGE
• RADIANCE
Schematic diagram showing the interaction of direct and diffuse radiance with landscape features and
atmospheric-panicles, to produce, image forming and path radiance.
-------
the concentration, composition, and the size of the particles.
Particles in a size range from about 0.5 to 1.0 microns diameter
scatter more light for the same mass concentration than smaller
or larger particle. Black carbon, primarily from incomplete
combustion such as in diesel exhaust or wood smoke, is the
principal cause of light absorption in the atmosphere. Some
particles are composed of materials such as sulfates and nitrates
that cause water vapor from the air to form solution drops under
high relative humidity condition. Since the solution drops are
larger than the dry particles, visibility impairment by these
particles increases during high humidity conditions.
Non-visual air quality factors such as lighting conditions and
inherent appearance of the scene, also have an influence on the
appearance of the scene viewed at a distance. For example, the
position of the sun with respect to the viewing angle of the
scene and characteristics of the clouds, if any, determines the
nature of shadows in the scene and influences the amount of non-
image- forming light. Also, visibility is usually dramatically
reduced during periods of precipitation and fog. Since factors
of this type are continually changing and are generally
unpredictable, they can not be made a part of a predictive
analysis of visibility effects. However, these non-air quality
factors are assumed to have the same influence when averaged over
sufficient time and space so that they do not need to be
separately evaluated for this report.
Unlike the nearly constant Rayleigh contribution to visibility
impairment, the contribution by particles from both natural and
man-made processes are highly variable. Natural particle sources
such as wildfires, windblown dust, salts from ocean spray, fog
and precipitation, etc. are highly variable across time and space
with the result that natural background levels of visibility are
highly variable. Concentrations of man-made and natural
particles also vary because of the influence of variable
meteorology responsible for atmospheric transport and dispersion.
Estimates of the eastern and western regional annual averaged
natural background visibility levels (Rayleigh scattering plus
contributions by natural particle sources) have been made using
particulate composition data and emission source inventories.
The difference between annual averaged current conditions and the
estimated natural background contributions to visibility
impairment is an estimate of the averaged man-made contribution
to visibility impairment.
For further information on visibility, see the 1979 Report and
the 1990 NAPAP Report "Acidic Deposition: State of Science and
Technology, Volume III, Report Number 24, entitled Visibility:
Existing and Historical Conditions - Causes and Effects.
-------
Visibility Metrics
Visual range, which is defined as the greatest distance that a
large dark object can be seen against the background sky, is the
oldest and most commonly used visibility metric. Visual range
was developed and continues to function well as an aid in
military operations and transportation safety. Airport
observations of visual range have been made since 1919, and have
been computer archived since the late 1940's. Daylight
observations involve viewing preselected visibility markers
(large dark objects) at known distances from the observation
point to determine the most distant marker-that is visible. At
night, lights are used instead of markers. One of the more
serious shortcomings of airport observations of visual range is
the availability of suitable targets at reasonable distances. It
is not unusual for the actual visual range to be greater than the
most distant target.
In spite of such difficulty, visual range is likely to remain a
popular measure of visibility because of its familiar distance
units (kilometers or miles), simple definition, and the fact that
any sighted person can use it to characterize visual conditions
without instrumentation. These very attributes which make visual
range popular also result in its common misinterpretation. For
example, some people mistakenly believe that all objects out to a
distance of the visual range are clearly seen and therefore
visibility is unimpaired as long as the visual range is greater
than the distance to the furthest scenic feature of interest. In
fact, noticeable degradation of scenic appearance including the
disappearance of some features is likelyjEor pbjects as near as
10% of the visual range.
Another traditional visibility metric is extinction coefficient,
which is the attenuation of light per unit distance due to
scattering and absorption by gases and particle in the
atmosphere. Extinction coefficient is expressed in inverse
length units (e.g., km"1) and is used primarily by scientists
studying the causes of reduced visibility. Direct relationships
exist between concentrations of atmospheric constituents and
their contribution to extinction coefficient. Apportioning
extinction coefficient to atmospheric constituents provides a
method to estimate change in visibility caused by change in
constituent concentrations. Calculation of the extinction
coefficients corresponding to air quality model predicted
pollutant concentrations is the approach used in the analyses
done for this report in order to estimate the visibility changes
expected to result from Clean Air Act mandated emission changes.
An estimate of visual range, commonly termed standard visual
range, can be made from extinction coefficient using a simple
transformation known as Koschmieder's relationship. The visual
-------
range values in this report are calculated from predicted or
measured extinction coefficient values.
Neither visual range nor extinction coefficient is linear to
perceived visual changes caused by uniform haze. For example, a
5km change in visual range or a 0.01km'1 change in extinction
coefficient can result in a scene change that is either
imperceptible or very obvious, depending on the baseline
visibility conditions. Presentation of visibility data or model
results in terms of visual range or extinction coefficient
creates the potential for misinterpretation by those who are not
aware of the non-linear relationship, and requires the
inconveniences of further analysis for those who are aware.
To avoid these difficulties, a new visibility index related to
perception of atmospheric haze was recently developed.1 The
scale of this visibility index, expressed in deciview (dv), was
designed to be linear with respect to perceived visibility
changes over its entire range, analogous to the decibel scale for
sound. A one dv change is approximately a 10% change in the
extinction coefficient, which is a small but usually perceptible
scenic change. The deciview scale is near zero for a pristine
atmosphere (Rayleigh conditions) and increases as visibility
impairment increases. Because the index increases as haze
increases, it is characterized as a haziness index. The deciview
scale is defined in terms of the logarithm of the extinction
coefficient, allowing simple transformations between all of the
common visibility metrics. Since the deciview scale is
perceptually linear, a change of any specific number of dv should
appear to have approximately the same magnitude of visual change
on any scene regardless of baseline visibility conditions.
1 Pitchford, M.L. and Malm, W.C. (1992) Development and
Applications of a Standard Visual Index. Presented at the
Conference on Visibility and Fine Particles, Vienna, Austria.
Accepted for publication by Atmospheric Environment, in 1993.
8
-------
Chapter 3. Historical Perspective and Current Conditions
Trends of Visibility
Figure 3-1 shows the location of the 158 mandatory class I
Federal areas. Prior to routine instrument-based monitoring of
visibility conditions begun in 1978, the National Weather Service
(NWS) observations are the only data available to track
visibility changes. Figure 3-2 indioj»tes the trend of the NWS
observations across the country. (NAPAP 24-30) From this
analysis, visibility is noted to improve in the northeast during
the winter, and decline in the east in general and the southeast
in particular in the summer. Western rural values show little
change according to the NWS data likely due to higher visibility
levels and a less dense network of data collection.
Monitoring of visibility conditions specifically in class I areas
was begun in 1978, with emphasis on the western areas. Starting
in 1987, to fulfill requirements for monitoring under the EPA's
visibility regulations, a multi-agency monitoring system began
operation in approximately 30 class I areas. Monitoring of other
class I areas, using similar techniques, takes place in
approximately 13 more class I areas. Limited analysis of these
data have taken place. One study of data from the southwestern
monitoring sites indicates improved summer visibility, but winter
visibility has remained constant between 1988 and 1991.l
Current Conditions at Class I Areas
Monitoring of class I areas through the Inter-agency Monitoring
of Protected Visual Environments (IMPROVE) network provides
information to develop current visibility conditions at class I
areas across the U.S. Figure 3-3 illustrates current visibility
annual average conditions at selected class I areas based on
IMPROVE data. The IMPROVE estimates of current visibility are
based on reconstructing the components of visibility impairment,
including Rayleigh and natural aerosols. Results of two
reconstruction techniques are shown in Table 3-1 along with the
measured visibility impairment. Techniques to estimate
visibility conditions for an annual or seasonal time period
generally compare well (within 10 percent) with measured
atmospheric conditions in the East, Central Rockies and Colorado
Plateau. Reconstructed extinction is typically 70-80 percent of
the measured extinction. The worst agreement is in the Sierra
Nevada (Yosemite National Park monitor) where the reconstructed
1 Visual Air Quality in the Grand Canvon and Golden Circle;
An Assessment 9£ Measurements. Source Contributions, and Trends.
D.'A. Latimer, AWMA 86th Annual Meeting, Paper 93-MP-4.09
-------
to
to
-------
Figure 3-2.
National Weather Service Visibility Trends
(Source: NAPAP, State of Science/Technology, Report K4, 1990)
1948-1954
1955-1964
1965-1974
1975-1983
Q1
United Stales trend maps for the 75th.percentile extinction coefficient (derived from air visual range data)
for the calendrical quarters: winter (QJ). spring (Q2), summer (Q3), and fall (Q4).
11
-------
Figure 3-3.
Average Visibility (deciview) at Selected
Monitored Class I Areas (1988-1991)
(Source: IMPROVE, CIRA Report, Feb. 1993)
15
Average visibility impairment in deciviews calculated from total (Rayleigh included) reconstructed light extinction foi
the first three years of IMPROVE, March 1988 through February 1991.
12
-------
Table 3-1.
Comparison -of Measured and Reconstructed Visual
Range (IMPROVE March 1988-1991)
(Source: IMPROVE, CKA Report, Feb. 1993)
IMPROVE Region
Appalachian
Colorado Plateau
Central Rockies
Pacific Coast
Northeast
Northern Great Plains
Northern Rockies
Southern California
Sonoran Desert
Sierra Nevada
West Texas
Measured Annual
Average Visual
Range (km)
32
145
165
78
96
120
84
49
105
66
89
Reconstructed
Annual Average
Visual Range (km)
(Dry Organics)
36
151
167
98
94
127
103
64
135
133
124
Reconstructed
Annual Average
Visual Range (km)
(50% of Organics
Assumed
Hygroscopic)
34
149
164
97
91
123
95
63
134
131
123
13
-------
extinction is only 50% of the measured value. This may be due to
the fact that the aerosol monitor is located above the mixed
layer much of the time. For this report, reconstructed
extinction will be used to estimate changes in future conditions.
The baseline extinction of the models was adjusted for measured
extinction levels.
Causes of Current Visibility Conditions
As noted before, regional visibility conditions are dominated by
a continuously varying mix of natural and man-made fine particles
in the. troposphere. Many studies have been published on
visibility conditions and related aerosol concentrations. The
NAPAP lists 33 aerosol and visibility databases. From these
studies, the major contributors to visibility impairment from
natural and man-made sources are sulfate particles, organic
particles, elemental carbon, suspended dust, nitrate particles,
and nitrogen dioxide. The National Academy of Sciences (NAS)
developed an estimate of regional man-made contributions of these
categories for three regions of the country that contain 90
percent of the class I areas. These regions are the East (states
east of the Mississippi River), the Southwest (California,
Nevada, Arizona, New Mexico, Utah, and Colorado), and the Pacific
Northwest (Oregon, Washington, and Idaho). Table 3-2 summarizes
the NAS model results.
The EPA's estimate of changes in visibility in the class I areas
focuses on estimating changes in these identified categories of
particles. The contribution of nitrogen dioxide to light is very
small, particularly in and near the remote class I areas, and,
therefore, is not considered in development of future cases as a
direct cause of regional impairment. Regional nitrogen dioxide
emissions and their resulting conversion to nitrate particles is
considered. The primary pollutant emissions that lead to
formation of the identified components of regional man-made
visibility impairment reviewed in this report are sulfur dioxide,
oxides of nitrogen, primary organics and primary particulate
matter.
IMPROVE monitoring data were used to establish annual and
seasonal apportionment of current aerosol components to the total
visibility impairment for specific class I areas. An example of
the IMPROVE data for annual averages is given in Table 3-4.
14
-------
Table 3-2. Anthropogenic Visibility (Extinction) Budgets by
Rural Regions of the Country*
(Source: Protecting Visibility in National Park* and WildmaM Ami, National Itimarch Council,
National Aead«ny Pnu, 1993)
* '-•••- E«stb Southwest Northwest11
Sulfates
• Organic*
Elemental carbon
: Suspended dust
Nitrates
"Nitrogen dioxide
65
14
11
2
5
3
39
18
14
15
9
5
33
28
15
7
13
_4
: 'Percentage contribution by specific pollutant to anthropogenic light extinc-
tion in three regions of the United States.
bBased on Table 9, Table 18, Figure 45, Appendix A, and Appendix E of
NAPAP Visibility SOS/T Report (Trijonis et al.t 1990). It is assumed that
sulfates (3% natural) account for 60% of non-Rayleigh extinction, organics
(33% natural) account for 1896, elemental carbon (3% natural) accounts for
10%, suspended dust (50% natural) accounts for 4%, nitrates (10% natural)
account for 5%, and nitrogen dioxide (10% natural) accounts for 3%.
''Based on Table 9, Table 18, Figure 45, Appendix A, and Appendix E of
the NAPAP Visibility SOS/T Report (Trijonis et al., 1990). It is assumed
that sulfates (10% natural) account for 33% of non-Rayleigh extinction, or-
ganics (33% natural) account for 20%, elemental carbon (10% natural) ac-
counts for 12%, suspended dust (50% natural) accounts for 23%, nitrates
(10% natural) account for 8%, and nitrogen dioxide (10% natural) accounts
for4%.
Extinction efficiencies (relative to organics) are chosen as 1.5 for sul-
fates, 2.5 for elemental carbon, 0.3 for fine crustal materials, and 1.5 for
nitrates (Trijonis et al., 1988, 1990). Coarse dust extinction is assumed to be
three times fine dust extinction (Trijonis et al., 1988, 1990). Natural aerosol
particle fractions are assumed to be one-tenth for sulfates, one-third for
organics, one-tenth for elemental carbon, one-half for crustal materials, and
one-tenth for nitrates. These assumptions are applied using the fine mass
concentrations in Trijonis et al., (1990). The percentage contribution for
nitrogen dioxide is assumed to be 4%.
15
-------
Table 3-3.
Annual Averages (March 1988-February 1991) of
Reconstructed Light Extinction (Mm'1) for 19 Regions
of the IMPROVE Network.
(Source: IMPROVE, CIRA Report, Feb. 1993)
REGION .
ALASKA
APPALACHIAN
BOUNDARY
WATERS
CASCADES
CENTRAL
ROCKIES
CENTRAL
CALIFORNIA
COAST
COLORADO
PLATEAU
FLORIDA
GREAT BASIN
HAWAII
NORTHEAST
NORTHERN
GREAT
PLAINS
NORTHERN
ROCKIES
SIERRA
NEVADA
SIERRA
HUMBOLDT
SONORAN
DESERT
SOUTHERN
CALIFORNIA
WASHINGTON,
D.C.
WEST TEXAS
Total
Extinction
25.4
112.2
68.2
58.8
28.1
56.3
27.1
87.5
23.4
53.2
71.3
39.7
54.3
33.4
28.0
31.3
63.5
164.3
36.7
Aerosol
Extinction
15.4
~ 102.2
58.2 .
48.8
18.8
46.3
17.1
77.5
13.4
43.2
61.3
29.7
44.3
24.4
18-0
21.3
53.5
154.3
26.7
Sulfate
6.7
.. 69.7
29.8
19.0
5.8
15.4
6.0
42.4
3.4
31.5
38.3
13.1
12.4
5.7
4.4
8.1
7.7
75.6
12.2
Nitrate
0.7
6.9
8.4
3.3
1.3
12.1
1.4
9.5
0.9
1.0
5.1
3.3
4.0
3.6
1.4
1.3
23.8
24.6
1.4
Organic*
4.6
16.7
14.1
19.2
6.1
10.6
4.7
15.4
4.6
5.0
11.0
73
19.6
8.1
7.7
5.5
9.7
25.0
5.7
Element*!
carbon
0.5
4.6
2.2
4.9
13
2.7
1.5
3.6
0.6
0.7
4.0
1.4
4.3
2.5
1.8
1.8
4.8
18.4
1.5
Soil and
Coane
2.6
4.3
3.8
2.4
3.6
5.6
3.5
6.7
4.0
5.1
2.9
4.7
3.9
3.4
2.7
4.5
7.5
10.6
5.9
16
-------
-------
Chapter 4. Preliminary Estimation of Future Conditions
In order to estimate the effects on visibility due to emissions
changes expected to occur as a result of implementation of the
CAAA, it was necessary to predict the changes in pollutant
concentrations for those species that are most important to
visibility degradation. In the East, sulfates predominate the
man-made visibility degradation. The estimates of visibility
change address changes in sulfur dioxide emissions and nitrogen
dioxide emissions. Changes in man-made volatile organic
compounds in the East are driven by air quality control measures
to attain the ozone health standard. These occur primarily in
urban areas, particularly in the northeast. In the southwest, it
was determined that the modeling exercise should include
sulfates, nitrates, primary carbonaceous particles (organic and
elemental), and remaining fine particle mass. Although secondary
organics are sometimes a large portion of the ambient organic
particulate matter in rural atmospheres, it is not clear whether
these rural concentrations are anthropogenic or natural in
origin. For analysis purposes, the secondary organic particulate
matter effects on visibility are reviewed as a sensitivity
analysis on the eastern and southwestern modeling analyses.
Modeling only primary anthropogenic organic material eliminated
the need to prepare base and CAAA-controlled volatile organic
compound (VOC) emissions inventories, and this resulted in a
considerable savings in labor and computing resources.
1990 Clean Air Act Amendments
With the passage of the CAAA on November 15, 1990, the U.S. EPA
was given a mandate to control air pollution using a variety of
innovative new approaches. The breadth of this new mandate poses
significant new regulatory and analytical requirements for states
and air pollution sources located within them.
Each of the Titles of the CAAA has been reviewed to identify
provisions that may affect visibility in class I areas. The
focus is on provisions which will directly affect emissions of
visibility-related pollutant species. A discussion of important
provisions in each Title follows. Subsequently, the
methodologies for quantifying emission reductions and determining
the effects on visibility are presented.
The Clean Air Act as amended by the CAAA will continue to require
attainment and maintenance of the national ambient air quality
standards through state implementation plans (SIPs). The CAAA,
however, outline a number of changes in the SIP process and new
requirements in SIP submittals. Title I of the CAAA addresses
areas that are not in attainment of the national ambient air
quality standards for criteria pollutants, including ozone.
17
-------
I
I Title I expands the boundaries of nonattainment areas to include
I the entire Consolidated Metropolitan Statistical Area (CMSA) .
f This means previously unregulated sources will have to comply
I with the new requirements of the CAAA. The CAAA include
; 'reasonable further progress' requirements which call for 15
: percent VOC reductions by 1996 and additional reductions of three
? percent per year after that until attainment is achieved. In
certain cases, NO, reductions may be substituted for VOC
reductions in order to meet the three percent per year reduction
;; requirement. The reductions must be taken mainly by stationary
sources; specific program mandated emissions reductions for motor
vehicle sources are covered in Title II. NO, emissions controls
are also required for some sources. This is the first time NO,
controls have been required nation-wide for stationary sources.
Previously VOC control was considered sufficient to bring areas
into attainment with ozone standards. It is also the first time
that quantified emissions reductions targets have been specified
for such areas.
New planning requirements will have a direct impact on emissions.
For example, SIPs must include enforceable emission limitations
and other control measures or techniques (including economic
incentives such as fees, marketable permits, and auctions of
emission rights). Also, SIPs must include provisions that
prohibit any activity from emitting pollutants that will
contribute significantly to nonattainment in another state or
that will interfere with measures required to prevent significant
deterioration of ambient conditions.
Under the particulate matter provisions within Title I,
Reasonably Available Control Measures (RACMs) are to be
implemented by December 10, 1993 in those moderate nonattainment
areas (as defined in Title I); Best Available Control Measures
(BACMs) must be implemented in current serious nonattainment
areas by February 8, 1997 (within 4 years of reclassification).
In those serious nonattainment areas failing to reach attainment
by December 31, 2001, 5 percent annual PM10 or PM10 precursor
emission reductions will be required.
Title II of the CAAA addresses mobile source-related emissions.
Among its provisions are requirements for: (1) tighter emission
standards on motor vehicles; (2) longer warranty periods on
certain motor vehicle emission control equipment; (3) fuel
producers to produce and sell reformulated/oxygenated gasolines
which meet specific physical and compositional requirements
including reduced vehicle emissions of VOCs; (4) gasoline
content, such as reduced volatility, inclusion of detergents, and
no lead; (5) fleet operators to purchase clean-fuel vehicles, and
(6) a pilot program in the State of California to demonstrate the
effectiveness of clean-fuel vehicles in controlling air pollution
in ozcfcre nonattainment areas.
18
-------
The air toxics provisions in Title III of the CAAA will require
reductions in hazardous air pollutants, including carcinogenic
chemicals. Many of these hazardous air pollutants are volatile
organic compounds (VOCs), and some are particles. Reduction in
VOCs that are both toxic compounds and.photochemically reactive
is likely to result in some changes in visibility. Similarly,
reductions in particulate matter that come about as a result of
air toxics emissions rules may also affect visibility. Although
reactive VOCs contribute to ozone formation and thus affect
visibility, it is expected that the reductions resulting from
Title III regulations will not be significant in comparison to
controls implemented via Title I of the CAAA. Similarly, the
majority of particulate emissions will be controlled .under Tittle
I and other provisions of the CAAA.
Provisions in Title IV of the CAAA call for large reductions in
NO, and SO, emissions from utility power plants. Nationally, SOX
emissions are expected to be reduced by 10 million tons per year
(TPY) and NO, emissions by approximately 2 million TPY as a
result of Title IV and other CAAA provisions. Since NO, and SO,
emissions play an important role in visibility, these emission
reductions have the potential for producing significant
visibility improvements in at least some class I areas.
Under Title IV, power plant emissions are to be reduced in a two
phase process. The second phase is scheduled to begin in the
year 2000 at which time annual S02 emissions from all units with
capacities in excess of 25 megawatts (MW) are to be reduced to a
rate equal to 1.2 Ib S02 / MMBtu times their 1985 - '87 average
annual fuel consumption. In addition, units operating at less
than 1.2 Ib S02 / MMBtu before 2000 may not increase emissions by
more than 20 percent. To achieve these emission reductions at
the lowest possible cost, a system of marketable emission credits
will be established. Thus, some sources may choose to buy
credits rather than reduce emissions. A series of bonus and
incentive allowances will be provided to assist midwestern
sources in achieving the required emission reductions, to allow
for growth in states with average emissions below 0.8 Ib S02 /
MMBtu and at sources experiencing increased utilization, and as
an incentive for installing flue gas desulfurization (FGD)
systems, implementing conservation programs and increasing use of
renewable energy sources.
Title IV also calls for the implementation of more stringent NO,
controls. By 2000, most boilers affected by the S02 reduction
requirements of Title IV must meet EPA-established performance
standards. In addition, a revised New Source Performance
Standard (NSPS) for NO, is to be issued.
Title V of the CAAA requires states to issue operating permits
for major stationary sources, including major sources subject to
19
-------
I
t
air toxics regulations. The purpose of the operating permit is
to allow a single document to list all the operating requirements
and emission limitations that apply to a particular source. This
means that conditions from prevention of significant
deterioration (PSD) and other emission'limits imposed by the SIP
must be contained in one central document. It is unlikely that
implementation of the Title V permit program will result in
significant changes in emissions, since best available retrofit
technology (BART)l and other substantive requirements are already
imposed through various regulatory mechanisms. The operating
permit may increase the efficiency of enforcement, but this is
difficult to measure in terms of emission reduction and
visibility improvement.
Provisions in Title VI of the CAAA are designed to protect the
stratospheric ozone layer which has been damaged by the emissions
and migration into the stratosphere of very slowly reactive
clorofluorocarbon (CFC) species. The presence of CFCs in the
lower atmosphere does not affect visibility. If the use of CFCs
are replaced by other, more reactive, hydrocarbon species, then
it is possible that, while reducing stratospheric ozone
depletion, there is a potential for increased econe formation
near the surface. However, it is expected that the provisions of
Title VI will add no additional controls relevant to visibility
in class I areas that have not been addressed by other titles
(e.g., Title I).
Title VII of the CAAA addresses issues concerning provisions for
enforcement of regulatory measures. Although enforcement
provisions are necessary for the realization o< emission
reductions, these reductions have been specified in other Titles
of the CAAA. This Title does not directly stipulate emission
reductions so there is no measurable effect on visibility in
class I areas.
Title VIII of the CAAA addresses miscellaneous issues such as
provisions for limiting emissions from outer continental shelf
(OCS) development, future visibility studies to be undertaken,
and authority for discounting emissions emanating from outside
the U.S. for certain SIP purposes in international border areas,
and establishment of a program to monitor and improve air quality
in regions along the border between the U.S. and Mexico.
Although the outcome of some of these provisions may lead to
development and implementation of control measures, it was not
possible to estimate emission reductions so no attempt was made
1 BART applies to certain existing major stationary sources
that emit any air pollutant which may reasonably be anticipated to
cause or contribute to any visibility impairment in a mandatory
class I Federal area for which visibility is an iragprtant value.
(See section 169A(b)(2)(A) of the Clean Air Act.
20
-------
to assess the effect that Title VIII provisions will have on
class I area visibility. The requirements for this report
specifically exclude from consideration any programs enacted
under Section 169B.
Identification of Key 1990 CAAA Provisions
Each title of the 1990 CAAA was reviewed to identify provisions
that may effect visibility in class I areas. Results of this
review process are summarized in Table 4-1. Provisions which do
not contain specific, quantifiable emission limits were dropped
from consideration, even though it could be argued that some of
these provisions may result in some emission reductions.- This
includes the permitting provisions in Title V, the enforcement
provisions in Title VII, and the miscellaneous provisions in
Title VIII (including the DCS source provisions and the
Mexico/Canada border provisions). Since it was determined at the
outset that secondary anthropogenic organics are not as major a
component of the extinction budget as sulfur and elemental carbon
in most of the class I areas considered, provisions relating to
VOC emissions reductions were not directly addressed in air
quality modelling, although the effect of these provisions on
ambient ozone levels were included in the modeling analysis for
the southwestern U.S. A bounding exercise was also undertaken to
establish an upper limit on visibility improvements expected to
result from reductions in secondary anthropogenic organic
aerosols.
Of the provisions addressed in this study, those in Title I are
responsible for most of the controls on stationary source NOX and
particulate matter (PM) emissions in ozone, CO, and PM10 non-
attainment areas. Utility NOX and SO2 emissions are controlled by
Title IV provisions. Mobile source emission reductions arise
from both Title I and Title II requirements. Although the air
toxics provisions in Title III may result in some reductions of
primary particulate matter emissions, these reductions are
extremely difficult to quantify at this point and were therefore
not included in the analysis.
Preliminary Analysis of Visibility Changes in the Federal Class I
Areas
A preliminary assessment was conducted to establish the
geographic region selection and classification of class I areas
and to provide a qualitative estimate of the expected changes in
visibility due to the CAAA. Statewide emission reduction
estimates corresponding to the CAAA provisions were used to scale
ambient concentrations (i.e., source category rollback) and
approximate the changes in visibility-related pollutant
concentrations and visual range at all class I areas in the U.S.
21
-------
Table 4-1.
NO,, SO., and TSP Emission Control Provisions Contained in 1990 CAAA.
Title
I
I
I
I
I
I
I
i&n
i
Attainment
Status of
Affected
Areas
Ozone:
Marginal and
above
Ozone:
Moderate and
above
Ozone:
Serious and
above
Ozone: All
PM-10
Ozone
Transport
Regions
Ozone:
Serious and
above
COr Moderate
(>12.7ppm)
and above
Ozone:
Serious and
above
Ozone: Severe
and above
CO: Serious
Provision
Stationary sources must perform new source review (Nippon all
new and modified major sources.
Stationary sources must apply reasonably available control
technology (RACT) to all major sources of VOC and NOx.
Stationary sources must demonstrate 3% per year reduction in VOC
emissions averaged over 3 year period, beginning Nov. 1996.
Alternatively, provide for reductions in VOC and NOx that result in
ozone reductions at least equivalent to those that would result from
3% per year reduction in VOC alone.
New CTGs requiring RACT for VOC emissions from application of
aerospace coatings and solvents. VOC emissions from removal or
applicatiwtff paints, coatings, and solvents used in shipbuilding ant
repair
Reasonably available control measures (RACM) and best available
control measures (BACM— including BACT) on fugitive dust,
residential wood burning, prescribed burning, and other sources of
PM-10
Enhanced vehicle inspection and maintenance programs for mobile
sources, existing and n0w CTGs, RACT for SO tpy VOC sources
and 100 tpy NO, sources
Mobile
-------
Table 4-1. Concluded.
Title
i&n
i&n
n
i&n
m
IV
rv
V
VI
vn
vm
Area
Attainment
Status
Ozone: Severe
and above
CO: Serious
All
All
CO: Moderate
and above
All Areas
All Areas
All Areas
All Areas
All Areas'
All Areas
All Areas
Description of Requirement
Mobile sources: By Nov. 1992, require employers of 100+ to
increase passenger occupancy per vehicle during peak hours by
25%. Reformulated gasoline for 9 worst ozone areas with
population greater than 250,000 and ozone areas reclassified severe.
Onboard Vehicle Controls, Reid Vapor Pressure programs.
1995 reduced tailpipe NO, emission standards
Mobile sources: Require at least 2.7% oxygen in any gasoline sold
in nonattainment area during season of high CO concentrations,
beginning Oct. 1993. Fuels with higher oxygen content can be usec
to offset those with lower content.
Provisions related to the emissions of air toxics.
By 2000, most utilities must meet Phase n total SO, emission limits
and may participate in the SO, allowance trading system
Reduced NO, emission factor limits for utility boilers subject to
Phase I and Phase n SO, controls; similar implementation schedule
as Phase I and n SO, allowances.
Increased permitting for a variety of sources. Permit enforcement
provisions.
Stratospheric ozone protection.
Federal enforcement provisions.
Miscellaneous provisions: off-shore emissions
Methodology for
Addressing Requirement
Not addressed - effects on total
emissions expected to be
generally small
Not addressed - effect on total
emissions expected to be small
Modeled with MOBILE 5.0
Modeled with MOBILE 5.0,
for all episode months
Not addressed
Based on CEUM results from
ICF Resources for "EPA RIA
low trading case"
Application of Phase II
emission factor limits by boiler
type
Not addressed
Not addressed
Not addressed
Not addressed
23
-------
I Table 4-2 lists the 158 mandatory Federal class I areas. Figure
| 4-1 shows the location of each class I area as well as the
| regional grouping used for the preliminary analysis. For each
; geographical group, one or two class I areas were selected to
I characterize the current aerosol/extinction conditions for all
5 other class I areas in the group. In most cases, a single class
i I area was considered sufficient to represent the group; however,
! two class I areas were used to represent geographically larger
groups, and those groups containing class I areas with larger
: variations in aerosol mixtures.
Annual average concentrations of the major.fine particle (less
than 2.5 pan) constituents (sulfate, nitrate, organic and
elemental carbon, and soil dust) for the selected class I areas
were extracted from the National Park Service (NFS) air
monitoring network data collected during 1983-1986.
Average extinction resulting from each of five fine mass
components was determined by multiplying each average species
concentration by an individual extinction efficiency based on
literature review. A simple functional form for sulfate and
nitrate humidity-dependent extinction efficiency was used to
approximate the deliquescent nature of sulfate and nitrate
particles. Annual mean daytime relative humidity from the
nearest major city was used for each representative class I area.
Average extinction for N02 and coarse mass were also taken from
the literature.
Extinction for the anthropogenic portion of the five fine
particle species, coarse mass, and N02 was determined by
subtracting an estimated natural background extinction from the
total extinction estimate. Background values for SO4, N03, and
elemental carbon were based on the NAPAP. Natural extinction for
organic carbon was assumed to be 75 percent of the total organic
carbon in rural areas, and 25 percent near urban areas. All
. extinction from coarse mass and fine mass soil dust was assumed '
to be natural sources. All extinction from N02 was assumed
anthropogenic.
The contribution of each source category to atmospheric fine mass
component concentrations was approximated using the relative
emission strength for the source category. Ambient
concentrations (anthropogenic portion) of sulfate, nitrate (and
N02), elemental carbon, primary organic carbon, and secondary
organic carbon were assumed to be directly related to emission
rates of sulfur dioxide, nitrogen oxides, soot, organic carbon
and volatile organic carbon, respectively. Relative percent
contributions of each source category from upwind states were
combined via a weighted sum of emission loadings. The estimated
source attributions were tabulated as a matrix displaying the
24
-------
Table 4-2. List of Mandatory Federal Class I Areas
Class I
Area
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
as
16
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Group
Number
5
5
4
5
5
5
5
5
5
5
5
4
4
5
4
6
6
6
6
6
6
Class I Are* Name/location
Pasayien W, WA
North Cascade* NP, WA
Glacier Peak W, WA
Alpine Like. W. WA
Olympic NP, WA
Mount Kaiaer NP, WA
Goat Rocks W, WA
Mount Adaou W, WA
Mount Hood W, OK
Eagle Cap W, OR
Belli Canyon W, OR.ID
Strawberry Mountain W, OR
Mount Jefferson W. OR
Mount Waihiagton W, OR
Three Sisters W, OR
Diamond Peak W. OR
Crater Lake NP. OR
Gearhan Mountain W, OR
Mountain Lakes W, OR
Kalmiopsis W, OR
Redwood NP, CA
Marble Mountain W, CA
Lava Beds W, CA
South Warner W, CA
Thousand Lake* W, CA
Yolla-Bolly-Middle-Eel W, CA
Lanen Volcanic NP, CA
Caribou W.CA
Deviation W, CA
Mokelumme W, CA
Point Reyet W. CA
Emigrant W, CA
Yowmite NP. CA
Hoover Vr, CA
Minarets W, CA
John Muir W. CA
Kaiser W.CA
Kings Canyon NP, CA
Sequoia NP, CA
Pinnacles W, CA
Ventana W, CA
Dome Land We CA
San Rafael W, CA
San Gabriel W, CA
Cucamonga W, CA
Saa Gorgonio W, CA
Agu« Tibia W, CA
San Jacimo W, CA
Joshua Tree W. CA
KEY
« » • 'Visibility unimportant*
7p m lateroaoonai Park
NP - National Park
NM — National Monument
W - Wilderness Area
IfP . - Proposed area has been
accepted a* mandatory
25
-------
Table 4-2.
Continued
IIP
50
51
52
S3
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
7S
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
7
7
7
7
7
7
^
V
7
7
7
7
ChiricahuaNM, AZ
OuricanunW, AZ
SaguaroW, AZ
GalinroW.AZ
SoptntitioB W, AZ
Mount BaldyW.AZ
Sssne Anehi W, AZ
MazatzalW.AZ
KM Mountain W. AZ
Petrified Forest NP, AZ
Sycenofc Canyon lw v AZ
Omni Canyon NP, AZ
ZknNP. UT
Bryce CaayoD NP, UT
Capitol Reef NP.UT
Cnyonlands NP, UT
Arches NP.UT
larbidge W. NV
CMen of the Moon W. ID
Sawtooth W, ID
SclwwyBinenwM W. ID.MT
Cabiaet Moonuinf W. MT
GlaeierNP, MT
Munon MouatiiD W. MT
BoftM*nfa*HW. MT
Secpcgou W, MT
GiiMofteMtaW.MT
U. L. Bend W, MT
Medicine lake W, MT
Anacoodfr-PintUr W. MT
Red Reek Like. W, MT
YeUowsune NP, MT.WYJD
North AbaroluW.WY
WuhdcieW.WY
TetonW.WY
Grand Teton NP, WY
Fttzpxtnck Pedro Parks W, NM
BudelierW.NM
PecosW.NM
Botque del Apache W. NM
26
-------
Table 4-2. Continued
104 7 Gilt W, NM
105 7 While Mountain W, NM
106 7 Salt Creek W, NM
107 7 Cariebad Caverai NP, NM
108 7 Guadalupe Mounuini NP. TX
109 7 Bi( Bead HP, TX
110 10 Wiebiu Mountain* W, OK
111 3 Badlands W, SD
112 3 Wind Cave NP, SD
113 3 Theodore Rooceveh NMP, NO
114- 3 LoBwoodW, ND
115 11 Voyageur* NP, MN
116 11 Boundary Water* Canoe Area W, MN
117 10 MingoW, MO
118 10 Herculet-Glade* W, MO
119 10 Upper Buflalo W, AR
120 10 Canty Creek W.AR
121 10 Breton W, LA
122 13 Sipsey W, AL
123 13 Mammoth Cave NP, KY
124 11 « Rainbow Lake W, WI »
125 11 We Royale NP. MI
126 11 Seoey W, MI
127 14 Roowvelt Campobello IP, NB, Canada
128 14 . Mooeeborn W, ME
129 14 Acadia NP, ME
130 14 Great Gulf W,NH
131 14 Preridential Range-Dry River W. NH
132 14 Lyie Brook W, VT
133 15 Brigantine W, NJ
134 15 Dolly So
-------
I
Figure 4-1.
Location of Tabulated Class I Areas and Regions
used in the Preliminary Visibility Analysis
IZ7U2I
[NE U.S.
28
-------
relative contribution of each source category to each of the four
major anthropogenic aerosol constituents.
Estimates of the 1990 Clean Air Act Amendments-mandated
reductions in annual emissions of precursor species from
industrial, vehicular, and commercial sources were applied to the
1985 NAPAP inventory at the county level, then aggregated to the
state level. Changes in statewide S02 and NOX emissions from
electric utilities were taken from projects developed for the EPA
by ICF, incorporated assuming high economic growth and national
trading of sulfur emissions. This resulted in some states having
estimated future S02 and NOX emissions from.utilities increasing
over the estimate period due to economic growth factors. These
increases are over and above any reductions indicated for acid
rain control. This situation was addressed later, in more
detailed estimates.
Source attribution tables were reconstructed for the Clean Air
Act Amendment control scenario; the current (1985) relative
contribution of each source category to the four fine particle
constituents were scaled by the percent change of annual emission
rates in the future case estimate. Extinction budget tables were
produced that combined the current estimated anthropogenic
concentrations of sulfate, nitrate, organic carbon and elemental
carbon with the future emissions case scaling factors. Natural
background concentrations of these species (held constant at 1985
estimate levels) were added to these concentrations, and each sum
was multiplied be each species' corresponding extinction
efficiency to obtain new estimates of extinction. Table 4-3
lists the qualitative current and magnitude of change visibility
conditions based on this process.
From this preliminary analysis, the EPA decided to focus
resources on more detailed modeling of visibility changes in the
East and the Southwest. For the eastern U.S. analysis, the EPA
used the available Regional Acidic Deposition Model (RADM)
developed under the NAPAP with an additional processor to examine
visibility changes due to changes in sulfate concentrations. For
the southwestern U.S. analysis the EPA conducted a new modeling
analysis. The remaining areas of the country, notably the
Pacific Northwest and region just west of the Mississippi River
are not likely to see perceptible changes in regional-visibility
due to Clean Air Act Amendment programs.
29
-------
Table 4-3. Summary of Preliminary Visibility Analysis Results
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Geographical
Group
Cascades
Northwest Coast
Northern Tier
Central Coast
Sierra
Southern California
Desert Southwest
Golden Circle
Rockies
Central U.S.
Northern Lakes
Florida
Southeast U.S.
Northeast U.S.
Appalachia
Current
Extinction
Moderate
Small
Moderate
Moderate
Moderate
Large
Moderate
Small
Small
Large
Moderate
Large
Large
Large
Large
Estimated Change
in Extinction
Small
Small
Small
Moderate
Small
Large
Small
Small
Small
Moderate
Small
Moderate
Large
Moderate
Large
30
-------
References for Chapter 4
1. Trijonis, J., "Existing and natural background levels of
visibility and fine particles in the rural east," Atmos Environ..
16, 2431 (1982).
2. Trijonis, J., R. Charlson, R. Husar, W. Malm, M. Pitchford,
and W. White, "Visibility: Existing and historical conditions --
causes and effects," in NAPAP State of Science/Technology Report.
National Acid Precipitation Program, Washington, B.C., 1990.
3. Stoeckenius, T.E., H.A. Gray, and C.A. Emery, Report of
Results for Work Assignment 2-7. Contract. No. 68D00091:
Classification of CAAA Implementation on Parks and Wildernesses.
letter report to B. Polkowsky, dated February 19, 1992.
4. Gray, H.A., T.E. Stoeckenius, R.E. Morris, L.R. Chinkin, B.
Garelick, D.S. Eisinger, L.A. Gardner, and L.W. Richards,
Workplan for the Assessment of the Effects of the 1990 Clean Air
Act Amendments on Visibility in class I Areas. SYSAPP-91/-124,
prepared for the EPA by Systems Applications International, San
Rafael, California, 1991.
31
-------
-------
Chapter 5. Assessment of the Eastern U.S.
The National Acid Precipitation Assessment Program
As part of its comprehensive work, the National Acid
Precipitation Assessment Program (NAPAP) estimated changes in
visibility for the eastern U.S. that would result from various
sulfur emission reduction scenarios. The NAPAP State of the
Science Report #24, cites the major cause of man-made visibility
impairment for the eastern U.S. as fine particle scattering,
accounting for 75 to 95 percent. Of that fine particle
scattering, sulfates and organic fine particles account for
three-fourths of the dry fine particle mass with sulfuric acid
and its ammonium salts accounting for at least half of the fine
mass. For the .eastern U.S., natural background levels of
visibility are considered to be 150 +-45 km of standard visual
range on an annual average.1 This corresponds to an average
deciview range of 7.0 to 13.1 dv.
As part of the assessment of changes in effects related to
changes in pollutants associated with acid deposition, the NAPAP
reviewed the visibility conditions' sensitivity to change for
certain reductions in ambient sulfate concentrations. Then NAPAP
reviewed several specific emission reduction strategies based on
policy options. These visibility changes were presented in the
1990 Integrated Assessment Report. In Chapter 4 of the NAPAP
Integrated Assessment, the visibility changes resulting in set
levels of reductions and additions (20% increase, 40%, 60%, 80%,
100% decrease) of sulfate were estimated for three locations in
the eastern U.S. No changes were made in levels of organic or
other fine particle constituents. The integrated assessment
found:
• On an annual ave.rage basis, calculations show that sulfate
contributes 57 percent of light extinction in the rural
East.
• That reductions in ambient sulfate concentrations of 20
percent, 40 percent, 60 percent, 80 percent, or 100 percent
yield approximately 11 percent, 23 percent, 34 percent, 46
percent, or 57 percent decreases in light extinction.
Chapter 5 of the NAPAP Integrated Assessment focused on several
emission reductions scenarios that were being considered for
legislation in 1989. The major differences among these scenarios
were timing of sulfur emissions reductions. For the purposes of
assessing changes in visibility resulting from these emission
1 State of Science and Technology Reports, U.S. National
Acid Precipitation Assessment Program, Volume III, Report 24,
Visibility: Existing and Historical Conditions, Causes and
Effects, December 1990.
32
-------
scenarios, only a 36 percent sulfur dioxide emissions reduction
was considered. For that reduction in sulfur dioxide emissions,
holding other pollutant fine particle concentrations steady, the
Regional Acid Deposition Model outputs were used to calculate
sulfate and corresponding visibility changes. A 36 percent
reduction in sulfur dioxide emissions, resulting from application
of a 1.2 Ib of S02 per million Btu of heat input emission
limitation on all major sources in the Eastern U.S. results in an
annual average decrease in light extinction of 21 percent. This
improvement in visibility is a slightly greater 26 percent in
summer and slightly less in winter.
The geographic pattern of the visibility improvement was
investigated by using information from the RADM output on the
geographical pattern of sulfate concentration reductions. This
was combined with the approximate spatial distribution of
baseline light extinction as provided by an analysis of airport
data from the State of the Science Report. According to this
analysis, the greater visibility improvements are expected to
occur in a wide band along the Appalachian mountains from
northern Georgia to southern New England, with greater than
average improvements also occurring throughout the Ohio River
Valley. Figure 5-1, taken from the Integrated Assessment,
displays that pattern.
This NAPAP emission scenario is very close to the legislative
requirements ultimately adopted in the 1990 Clean Air Act
Amendments. The effects on visibility of trading S02 emissions
across the region for implementation cost reduction purposes as
allowed under the Amendments program was not estimated in the
NAPAP report.
Reanalysis Using the RADM Engineering Model
Since the NAPAP report, improvements in the RADM results for air
concentrations of sulfur and development of emissions databases
for regulatory impact assessments have taken place. (See
References 3,4,5) In addition, the EPA has developed a
visibility post-processor to the RADM air concentration outputs
which enables direct calculation of extinction changes, due to
changes in fine particle sulfur concentrations. With these
changes, a more refined estimate of changes in visibility for the
eastern U.S. was possible without extensive new runs of the RADM
model which would have been beyond the scope of this report.
Emissions Inventory Revisions
The emissions projections for the revised RADM engineering model
runs used in this report were provided by Pechan for the year
2010. Figures 5-2 and 5-3 show the estimated emissions for the
1985 base year and 2010 case for sulfur dioxide and nitrogen
33
-------
> 34% > 23% > 12%
Figure 5-1.
Geographical Distribution of Light Extinction from
the 1985 Baseline to NAPAP S4 2010 Emissions
Estimate.
(Source: NAPAP SOS/T Report 124, 1990)
34
-------
Figure 5-2.
Projected S02 Emissions for 1990 GAAA, Eastern
U.S.
(Source: EPA Compilation of Acid Rain Emission* Inventory work by Pechan, 1991)
Projected SO2 Emissions For 1990 CAAA
Eastern United States
S20000000
7/5
S 15000000
c
•2 10000000
v>
UJ 5000000
o
CO
Utility Industrial Minor Area
Emissions Sources
Total
1985 • 2010
35
-------
Figure 5-3.
Projected NOx Emissions for 1990 CAAA, Eastern
U.S.
(Source: EPA Compilation of Acid Rain Emissions Inventory work by Pechan, 1991)
Projected NOx Emissions for 1990 CAAA
Eastern United States
20000000
15000000 -
10000000
£ 5000000 —
x
O
Utility Industrial Minor Area Mobile Total
Emissions Sources
1985112005
36
-------
dioxide. Estimates of the changes in reactive hydrocarbon
emissions are given in Table 5-1.
Table 5-1.
Projected Reactive Hydrocarbon Emissions for 1990
CAAA, Eastern U.S.
Source Group
Major/Industrial Sources
Area/Non Traditional Stationary
Sources
(e.g., dry cleaning)
Mobile Sources
Total
Projected 1985 Emissions
(Million tons)
1.63
9.49
3.81
14.93
Projected 2010 Emissions
(Million tans)
2.36
8.85
1.08
12.29
(Source: EPA Compilation of ROM emissions inventory data base on Pechan, 1992)
2010 was the year selected to represent full implementation of
Title IV of the 1990 CAAA for many analyses the EPA is
conducting. The emissions projections assume no trading of S02
allocations. In addition, the emissions reductions for S02 for
the Canadian program, as predicted by Environment Canada for the
1990 NAPAP Integrated Assessment are included in these
projections.
Air Quality Model and Visibility Assessment
In this assessment, a RADM Engineering Model post-processor, (EM-
VIS) has been created to calculate hourly estimates of visual
range and light extinction for sulfate plus Rayleigh extinction.
These estimates were adjusted for total extinction based on NAPAP
estimates of relative contribution of components to eastern
extinction. Distributions of these estimates were developed for
all for mid-day averages (10:00 a.m. to 4:00 p.m.) following the
NAPAP methodology. The distributions were rank-ordered by RADM
cell for the base emissions and for the new 2010 projections,
representing implementation of the 1990 Clean Air Act Amendments
for acid deposition. Given the changes in emissions, the
meteorology producing the 90th percentile (highest visibility
impairment) in 1985 (the base year) is not necessarily the same
as that producing the 90th percentile for the year 2010 since
RADM accounts for chemistry by pollutant loadings.
Absolute and percentage changes were computed for the deciles for
each RADM cell relative to the 1985 base case. Since EM-VIS only
models changes in sulfate, the modeled changes in those species
37
-------
contribution were applied to 1985 baseline levels of the total
extinction budget. In general, the 1985 baseline sulfate was
well duplicated by the RADM model when compared with monitoring
data. The major differences occurred due to underprediction of
sulfate by RADM in the upper Midwest. In addition, the
monitoring data showed higher levels of sulfate than those
predicted by RADM for southern Florida and coastal New England.
This could be due to maritime climate contributions to the
measured sulfate levels that would not be predicted by RADM's
handling of major stationary source emissions.
The results of the analysis of RADM sulfate data indicate that
regional visibility will improve in class I areas located from
New Hampshire to northern Florida. Figure 5-4 shows the
estimated annual median visual range and deciview calculated by
RADM with adjustments for non-sulfur extinction for the period
1982-1985 which is the baseline period for this analysis. Figure
5-5 shows the estimated levels for 2010.
The annual distribution has a range of highly impaired days to
very clean days. For 1985, ten percent of the days in the year
are expected to have visibility levels equal to or less than
those indicated in Figure 5-6. Ten percent of the days in the
year are expected to have visibility levels equal to or greater
than those indicated in Figure 5-7. The amount of improvement
varies along this distribution as shown by the expected
improvements in the ten percent clearest days and the ten percent
most impaired days as modeled by RADM, shown in Figures 5-8 and
5-9, respectively.
Most of the high impairment days occur in summer due to the
higher atmospheric temperatures and higher daytime humidity
levels. The improvement in median visibility level for the warm
season (April through September) between 1985 and 2010 is similar
to the level of visibility improvement for the most impaired days
of annual distribution, as shown in Figure 5-10. This indicates
that the annual distribution very impaired days are dominated by
events in the warm season. An examination of the change in the
cold season (October through March) average reflects a change
more like that of the cleaner days in the annual distribution.
See Figure 5-11. The cold season has much better visibility for
the region as a whole.
Estimating the Effects of Changes in NO,
Nitrate aerosols contribute between 6 and 12 percent of the total
(Rayleigh included) extinction budget in the rural east as an
annual average. By comparison, sulfates contribute between 44
and 62 percent of the annual average extinction budget. Given
the magnitude of change expected by 2010 in nitrogen dioxide of
approximately 10 to 15 percent reduction, it
38
-------
Figure 5-4,
Estimated Baseline Annual Average Visibility
Visual Range (km) and Haziness (dv) based on reconstructed extinction from RADM
with adjustment for non-sulfate extinction
LEGEND:
Haziness (dv)
16-20 • 79-53
20-24 • 52-35
24-28 • 34-23
>28 • <23
Visual Range (km)
39
-------
Figure 5-5.
Estimated 2010 Annual Average Visibility
Visual Range (ton) and Haziness (dv) based on reconstructed extinction from KADM
with adjustment for mon-sulfate extinction
LEGEND:
Haziness (dv)
16-20 •
20-24 •
24-28 •
>28 •
79-53
52-35
34-23
<23
Visual Range (km)
40
-------
Figure 5-6.
Ten Percent of 1985 Days Estimatcsd at Shown
Visibility or Worse
Visual Range (km) and Haziness (dv) bated on reconstructed extinction from RADM
with adjustment for mon-tulfate extinction
LEGEND:
Haziness (dv)
20-24 • 52-35
24-28 • 34-23
>28 • <23
Visual Range (km)
41
-------
Figure 5-7.
Ten Percent of 1985 Days Estimated at Shown
Visibility or Better
Visual Range (km) and Haziness (dv) based on reconstructed extinction from RADM
with adjustment for mon-sulfate extinction
LEGEND:
Haziness (dv)
0-16 £
16-20 I
20-24 •
>79
79-53
52-35
Visual Range (kr
42
-------
Figure 5-8.
Improvement from 1985 to 2010 of the Most Impaired
Days
Haziness (dv) based on reconstructed extinction from RADM with adjustment for non-
fulfate extinction. A one dedview change is considered perceptible.
LEGEND:
is 0
IE 1
• 2
• 3
- 1
- 2
- 3
- 4
43
-------
Figure 5-9.
Improvement from 1985 to 2010 of the Least
Impaired Days
Haziness (dv) based on reconstructed extinction from RADM with adjustment for non-
sulfate extinction. A one deaview change is considered perceptible.
LEGEND
i 0 - 1
• 1-2
44
-------
Figure 5-10.
Improvement from 1985 to 2010 Average Warm Season
(Apr-Sep) Visibility
Haziness (dv) based on reconstructed extinction from RADM with adjustment for non-
gitffate extinction. A one dedview change it considered perceptible.
LEGEND
E 0 - 1
11-2
• 2-3
45
-------
Figure 5-11.
Improvement from 1985 to 2010 Average Cold Season
(Oct-Mar) Visibility
Haziness (dv) based on reconstructed extinction from RADM with adjustment for non-
sulfate extinction. A one deciview change is considered perceptible
LEGEND:
E 0 - 1
B 1 - 2
46
-------
is unlikely that rural regional visibility levels will change
perceptibly due changes in nitrate levels.
Estimating the Effects of Changes in VOC
There is little information available to distinguish between the
man-made organic aerosols which impair visibility from those that
occur naturally from vegetation and other causes. The currently
monitored levels of organic aerosol are also subject to more
uncertainty than those of sulfate and even nitrate. In general,
organic aerosols are estimated to be between 15 and 20 percent of
the regional fine particle loadings based on IMPROVE monitoring.
Since RADM cannot account for the organic compounds, a simple
rollback estimation based on estimates of emissions changes is
the only means to determine the possible effects of these changes
on visibility.: From Table 5-1, the eastern region should
experience a reduction of man-made VOC of approximately 17.7
percent. Assuming as an upper bound that the entire organic
fraction measured by IMPROVE is related to man-made emissions,
then a 17.7 percent reduction of that organic fraction would be
expected at the class I areas. Reducing the organic aerosol
fraction of the extinction budgets developed through the IMPROVE
monitoring by 17.7 percent would result in approximately a 3 to 4
percent increase in the total 1985 extinction and a 4 to 5
percent reduction based on the projected 2010 median extinction.
This is well below a 1 dv threshold for perceptibility. Highly
impaired days that have a larger concentration of organic
aerosols would likely see a greater improvement, but it is highly
unlikely that all these organic aerosols are man-made. So at
these levels of regional reductions of VOC, it is unlikely that
the rural class I areas in the eastern U.S. will see regional
improvement in visibility related to VOC emissions reductions.
Summary
This assessment indicated that a noticeable improvement in
visibility should occur across the eastern U.S. for the entire
year, with most of the change occurring in the warm seasons.
Winter improvements outside of the central Appalachian area may
not be noticeable, depending on the scene. With respect to class
I areas, those in the East, particularly along the Appalachian
chain should see substantial improvements in regional visibility,
particularly in the summer. Class I area improvements were
obtained by selecting the RADM grid cell that covered geographic
location of each class I area. Table 5-2 lists the changes in
visual range and deciview for selected class I areas in the East.
Improvement beyond those estimated here would have to come from
further reductions in sulfur dioxide emissions coupled:with
47
-------
Table 5-2.
Summary of Visibility Changes for Selected Class I
Areas (Based on RADM Reconstructed Sulfate plus
Non-Sulfate Correction Factor)
Location
AcadiaNP
Everglades NP
Great Smoky Mtns
NP
Lye Brook W
MingoNWR
Okefeookee NWR
Presidential W
Shenandoah NP
Sipsey W
Voyager NP
NP= National Park, N
Annual Average
Visibility
km(dv)
1985
50
(21)
58
(19)
30
(26)
45
(22)
41
(23)
28
(26)
51
(20)
40
(23)
31
(25)
69
(17)
2010
53 •
(20)
63
(18)
39
(23)
50
(21)
44
(22)
31
(25)
55
(19)
52
(20)
34
(24)
72
(17)
10 Percent of
Annual Days
Have Visibility
Less than:
km (dv level
greater man)
1985
17
(31)
32
(25)
15
(32)
14
(33)
16
(32)
12
(34)
22
(29)
15
(33)
13
(34)
43
(22)
WR= National Wildlife Ref
2010
21
(29)
37
(24)
20
(30)
17
(31)
20
(30)
14
(33)
27
(27)
21
(29)
17
(31)
46
(21)
10 Percent of
Annual Days
Have Visibility
Greater man:
km (dv level less
than)
1985
76
(16)
73
(17)
52
(20)
79
(16)
62
(18)
56
(19)
83
(16)
60
(19)
66
(18)
86
(15)
2010
78
(16)
78
(16)
56
(19)
83
(16)
66
(18)
62
(18)
85
(15)
67
(18)
70
(17)
88
(15)
Warm Season
(Apr-Sep)
Average Visibility
km(dv)
1985
50
(21)
63
(18)
22
(29)
39
(23)
36
(24)
23
(28)
49
(21)
25
(28)
28
(27)
83
(16)
2010
54
(20)
69
(17)
28
(26)
44
(22)
42
(22)
28
(26)
54
(20)
31
(25)
31
(25)
86
(15)
uge, W= Wilderness
Cold Season
(Oct-Mar)
Average
Visibility
km(dv)
1985
55
(20)
58
(19)
51
(20)
59
(19)
59
(19)
43
(22)
65
(18)
54
(20)
49
(21)
61
(19)
2010
59
(19)
63
(18)
57
(19)
63
(18)
62
(18)
49
(21)
67
(18)
61
(19)
54
(20)
67
(18)
reduction in the man-made fraction of organic aerosols, likely to
be emissions located outside of urban ozone NAAQS nonattainment
areas.
Comparison with Other Studies
t
The results of this analysis compares reasonably with recent work
done by Zannetti et. al, which indicates that sulfur emissions
reductions of approximately 12 million tons per year would result
in an average improvement in median visual range of 8 percent to
48
-------
11 percent. This work used a completely different technical
approach which separated three years of daily meteorological data
into classes of regional transport and regional atmospheric
conditions. Their work also focused on the effects on visibility
in 4 impact regions rather than much smaller irfHividual grid
cells. For each grouping of weather classifications and each
impact area, an estimate of visibility improvement was made on a
daily basis. The visibility estimates of change were also based
on modifying national weather service readings at airports. The
smaller change in visibility reported in that paper than in the
RADM modeling presented in this report may be due to the
following factors: 1) the use of airport data which might
underestimate the regional visibility thereby reducing the
1 relative effect of fine particle scattering, 2) differences-in--•
? emissions changes used in computing reduction of sulfate aerosol,
; and 3) different treatment of fog and precipitation conditions.
1 Overall, both approaches conclude there will be broad geographic
? areas of improvement in average visibility, which indicate large
• changes for certain days when sulfur aerosol dominates the
-- extinction.
Previous work for the EPA conducted in 1984, alipo assessed
changes in eastern visibility using a different air quality
modeling system (SAI). That work estimated that a 50 percent
reduction in S02 emissions across 31 eastern states would improve
the regional annual average visibility by 22 percent. This work
underestimated visibility improvement, because it failed to
account for changes in the boundary conditions that would result
from emissions controls Therefore, within the realm of
uncertainties for air quality modeling and the translation of air
quality data to visibility estimates, that estimate is
essentially in agreement with work performed* for this report.
49
-------
References for Chapter 5
1. State of Science and Technology Reports of the U.S. National
Acid Precipitation Assessment Program, Volume III, Report 24,
Visibility: Existing and Historical Conditions, Causes and
Effects, December, 1990.
2. 1990 Integrated Assessment Report of the U.S. National Acid
Precipitation Assessment Program, November, 1991.
3. Correcting RADM's Sulfate Underproduction: Discovery and
Correction of Model Errors and Testing the Corrections Through
Comparisons Against Field Data, Dennis et al., Atmospheric
Environment, (manuscript date August 10, 1992).
4. RIA, EPA Acid Rain Program regulations, September, 1991.
5. Documentation of S02 and NO, Forecasts, Assignment Report, EPA
Contract 68-D00120, Work Assignment 33, Prepared for Joan Novak,
USEPA, Res. Tri. Pk, Prepared by E.H. Pechan & Associates, Inc.,
Springfield, VA, September, 1991
6. Regional Oxidant Modeling Emissions Inventory Development and
Emissions Control Scenarios, EPA Contract 68-D9-0168, WA 43,
Prepared by E.H. Pechan & Associates and E R/C, June 1992
7. Notes on Revised EM-VIS Postprocessor, Memorandum from John
McHenry to Robin Dennis, U.S. EPA, Research Triangle Park, NC,
March 23, 1992.
8. Memorandum RADM Data for OAQPS Report to Congress on Effects
of 1990 CAAA on Visibility in class I areas. From Robin L.
Dennis, Atmospheric Characterization and Modeling Division, to
Bruce Polkowsky, Ambient Standards Branch, January 14, 1993.
9. Zannetti, P., Tombach, I., Cvencek, S., and Balson, W.
"Calculation of Visual Range Improvements from S02 Emission
Controls -- II. An Application to the Eastern United States"
Atmospheric Environment, Vol 27A, No. 9.
10. Visibility and Other Air Quality Benefits of Sulfur Dioxide
Emissions Controls in the Eastern United States, Systems
Applications, Inc. Prepared for the U.S. Environmental
Protection Agency, Draft Final Report, January, 1984.
50
-------
-------
Chapter 6. Southwestern U.S.
For class I areas in the southwestern U.S., a more comprehensive
analysis was carried out to estimate the changes in visual range
due to implementation of the CAAA on a finer spacial and temporal
scale than in the preliminary analysis. Figure 6-1 shows the
geographic domain of the southwestern modeling analysis. This
chapter summarizes the steps taken in preparing emissions
inventory for the modeling, the modeling approach, and the
results.
Emissions Inventory Development
Three annual emission inventories were prepared for the
southwestern U.'S. modeling domain:
a 1988 base case inventory,
a 2005 base case inventory designed to reflect emission
levels which would be expected in the absence of the
1990 CAAA, and
a 2005 control case inventory which reflects
anticipated emission reductions resulting from
implementation of the 1990 CAAA (excluding provisions
of Section 169B).
For purposes of this modeling, all provisions of the amendments
were assumed to be fully implemented as of 2005. Thus, control
measures designed to bring the Los Angeles area into attainment
with the national ambient air quality standard for ozone by 2020
were assumed to be fully implemented by 2005.
Each annual inventory was temporally disaggregated in the manner
described below, resulting in one inventory for each month
modeled (April, July, October, and December).
Anthropogenic sources of NOX, SOX, and PM were included in the
inventories -- biogenic sources of these species were not
included. For non-mobile source categories, emissions of PM10 and
PMjj (particles equal to or smaller than 10 /an and 2.5 /xm
aerodynamic diameter, respectively) were estimated from total
suspended particulate (TSP) emissions by assuming a
representative size distribution for each source category. PM^
emissions were further speciated into primary organic and
elemental carbon emissions for each source category. For mobile
sources, estimates of PM,0 emissions were obtained directly from
the PARTS model (Shepard, Gray, and Heiken, 1992). Speciation
factors were applied to these estimates to estimate the primary
organic and elemental carbon fractions.
51
-------
Figure 6-1. Southwestern Modeling Domain and Location of Class I Areas
-650
10
-150
350
650
- 4350
oJH-
10
20
3850
j ill i t i_ V i i_
30
'3350
Modeling domain for the southwest visibility modeling study; numbers indicate locations of Class I areas.
52
-------
Data Sources
The 1985 National Acid Precipitation Assessment Program (NAPAP)
emission inventory, Version 2 (EPA, 1989) was used as the basis .
for the annual inventories. This inventory includes emission
estimates, stack parameters, seasonal, weekly, and daily
operating schedules and location information for all major point
sources (defined as sources with annual emissions of any one
species of 100 TPY or more), area source emissions by county
(area sources include total emissions from .individual point
sources emitting less than 100 TPY), and-mobile source emissions
by county. Extensive data quality checking was performed on the
NAPAP inventory in connection with a previous modeling study of
the southwestern U.S. (Gray et al., 199lb)i. Missing,
inconsistent, or unrealistic source identification, location,
operating schedule, and stack parameter data were replaced with
best available estimates.
Electric utility emission estimates in the 1985 NAPAP inventory
are largely out of date and were not used for this study.
Instead, the National Allowance Data Base, Version 2.1 (NADB
V2.1; Rothschild, 1992) was used as the basis for the 1988
utility emission estimates together with updated information for
specific plants obtained from utility industry sources (Teague,
1992, 1993). Future year (2005 base and control case) utility
emissions were obtained from the regulatory impact analyses for
the acid rain and NO, regulations implementing Title IV of the
1990 CAAA (ICF, 1992a,b) as described below.
Few data are available regarding emissions from northern Mexico
which may effect visibility in U.S. class I areas. SO, emission
estimates were obtained for two large smelters in the border area
(Nacozari and Cananea) using data obtained in a previous study
(Gray et al., 1991b) . NO, and PM emissions for these facilities
were estimated assuming species emission ratios similar to those
at large U.S. smelters. Emissions from other sources in Mexico
were not included in the inventory.
U.S. smelter stack parameters were obtained primarily from the
1985 NAPAP inventory, with corrections incorporated as in the
previous study mentioned above. Emissions estimates for smelters
located in Arizona were obtained from the Arizona Department of
Environmental Quality (Costello, 1992). Emissions estimates for
the Kennecott Utah copper smelter were obtained from the plant
operator (Salmon, 1993) as were estimates for the two New Mexico
smelters (Kendall, 1993).
Emission Inventory Processing
Modeling inventories were prepared using the Emission
Preprocessing System (EPS 2.0; EPA, 1992a) in accordance with EPA
53
-------
procedures (EPA, 1991b). County-level area and mobile source
emissions in the NAPAP inventory were allocated to grid cells in
the modeling domain based on population density and county area.
Mobile source NO, emissions were estimated by applying emission
factors obtained from the MOBILE 5.0 model (EPA, 1992b) to the
appropriate county vehicle miles traveled (VMT) estimates
(obtained as described below) and spatially allocated to each
grid cell based on population density. Mobile source PM and SO,
| emission factors .were estimated using the PARTS model recently
I developed by Systems Applications International (Shepard, Gray,
I and Heiken, 1992). The annual inventory was temporally
* disaggregated using seasonal and diurnal adjustment factors.
I Point sources were disaggregated on a point-by-point basis using
\ the operating schedule information in the NAPAP inventory
| together with diurnal profiles provided in the EPS 2.0. Area and
• mobile sources were disaggregated by source category using EPA
: default temporal profiles.
Preparation of 1988 Inventory
Non-utility stationary source emission estimates for 1988 were
• obtained by applying appropriate 1985 - 1988 activity indicator
1 projection factors to the 1985 NAPAP estimates. State-level
projection factors for 2-digit SIC (standard industrial code)
designations and NAPAP area source category codes were obtained
from projected demographic and industrial economic activity
? indicators developed by the Department of Commerce's Bureau of
Economic Analysis. Data for smelters in Arizona and New Mexico,
1988 emission estimates were obtained from other sources as noted
above.
Mobile source emissions were estimated by first calculating the
1985 VMT levels from total emissions and associated emission
factors listed in the 1985 NAPAP inventory. These VMT estimates
were then grown to 1988 levels by applying national average VMT
growth factors developed by Argonne National Labs. The national
VMT growth factors were prorated to each state based on the
change in state population between 1985 and 1988. Finally, the
1988 VMT estimates were combined with emission factors obtained
from MOBILE 5.0 and PARTS to compute 1988 mobile source
emissions.
Utility SO, emissions for 1988 were assumed to be approximately
equal to the 1985 - 1987 average fuel use times the 1985 emission
factor for each source as listed in the NADB V2.1 file
(Rothschild, 1992). Although not specific to 1988, emission
estimates obtained in this way were judged to be more accurate
than those obtained by applying projection factors to the 1985
NAPAP utility emissions. Utility 1988 NOX and PM emissions were
estimated by scaling the 1985 NAPAP NO, and PM emissions for each
source by the ratio of NADB SO, emissions to 1985 NAPAP SO,
54
-------
emissions. For a significant number of the larger power plants,
specific 1988 emission estimates for SO,, NO,, and PM were
obtained from operating companies as compiled by the Utility Air
Regulatory Group (Teague, 1992) .
Preparation of 2005 Base Case Inventory
Non-utility stationary and mobile source emission estimates for
the 2005 base case inventory were obtained by applying 1988 -
2005 projection factors to the 1985 NAPAP inventory in a manner
analogous to that used to obtain the 1988 inventory. It should
be noted that the use of demographic and economic indicators
projected to 2005 results in significantly more uncertain,
emission estimates than is the case for 1988. Projection factors
were not used to estimate future U.S. smelter emissions which
were obtained from other sources as described above. Emissions
from the Mexico smelters were projected based on the applicable
industry projection factors for New Mexico. A summary of smelter
SO, emissions is presented in Table 6-1. Motor vehicle fleet
turnover effects are incorporated via the MOBILE 5.0 emission
factor model.
Utility NO, and SO, emission estimates for 2005 were calculated by
ICF Resources using results from the Coal Electric Utility Model
(CEUM) for the so-called "EPA Low" base case originally developed
in 1989 (ICF, 1989) and subsequently used in the acid rain
regulation regulatory impact analysis (ICF, 1992). The low base
case assumes electricity sales growth in the states included in
the modeling domain ranging from two to three percent per year
through 2000 and one to two percent per year thereafter.
Although the projected oil and gas prices assumed in this base
case are higher than most current projections, distortions in
emission estimates generated by this assumption were minimized by
updating the EPA forecasts using updated power plant build
decision information in the NADB V2.1 file, and recent data from
the National Electric Reliability Council (NERC) and the Energy
Information Agency (EIA). To account for the much lower gas
prices currently projected, all new coal-fired capacity included
in the EPA base case that is not confirmed by NADB or NERC/EIA
information were assumed to be built as gas-fired units.
Furthermore, the 1000 MW Allen coal-fired generating station
projected in NADB but not included in the EPA Low base case was
assumed to be indefinitely deferred. These changes resulted in
the projection of essentially no new coal-fired and very little
new oil-fired generating capacity. All new capacity which was
included in the calculations for which location information was
unavailable was assumed to be distributed over existing sites.
Table 6-2 summarizes SO, emissions from the eight layout utility
sources in the modeling domain.
55
-------
Table 6-1. Smelter SO,, Emissions (10* TPY).
2005
Arizona
Magma
Miami-Inspiration
Asarco
New Mexico
Phelps Dodge
Chino Mines
Utah
Kennecott Minerals
Mexico
Nacozari
Cananea
1988
113
11
25
66
35
27
29
64
Base
73
12
21
38
23
19
32
70
CAAA
73
12
21
38
23
1
32
70
56
-------
Table 6-2. Eight Largest Power Plants Annual SO, emissions (10* TPY).
2005
1988 Base CAAA
Navajo 68.4 8.1 8.1
Mohave 41 41 41
Four Corners 33 30 25
. San Juan 28 25 25
Comanchee 13 23 16
Cholla 17 22 22
Pawnee 12 18 18
Huntington 11 18 13
57
-------
SO, emissions at the Navajo Generating Station power plant in
Arizona were adjusted to reflect 90 percent scrubbing as is
currently required under provisions of the Clean Air Act
independent of the 1990 CAAA. For sources in California, utility
NO, emissions were restricted to the lesser of the CEUM
predictions and NO, emission limits imposed by state and local
authorities before 1990. Utility PM emissions were estimated
from projected fuel consumption by using uncontrolled PM emission
factors from EPA AP-42 and applying a typical control efficiency
of 98.5 percent.
Preparation of 2005 CAAA Case Inventory
A 2005 CAAA or "control" case inventory was obtained by adjusting
the 2005 base case inventory to account for the effects of
emission controls expected to result from implementation of the
1990 CAAA. A summary of the emission control provisions of the
1990 CAAA by title and affected non-attainment areas together
with a brief description of the methods used to account for the
provisions in the inventory is presented in Table 4-1.
Provisions calling for VOC emission reductions were not addressed
since VOCs were not included in the inventory. For stationary
sources, Title I NO, controls [e.g., reasonably available control
technology, (RACT)) were applied using the control efficiency
(CE), rule effectiveness (RE) and rule penetration (RP) factors.
These factors conform to those used in the EPA's emission
projections for the Northeast Ozone Transport Region (EPA,
1991a). Affected nonattainment areas are identified in Table 6-
3. It should be recognized that, as in any inventory development
process, the selection of appropriate CE, RE, and RP factors
represents a major source of uncertainty.
Title I of the 1990 CAAA call for implementation of certain
control measures (RACM, including RACT and BACH, including BACT)
for sources of'particulates in moderate and serious PM10
nonattainment areas respectively. At the time this report was
prepared, however, insufficient information was available with
which to estimate specific emission reductions which might be
expected to result from these control measures. For this study,
therefore, a linear rollback technique was used in which PM10
emissions in nonattainment areas were assumed to be reduced
across the board by an amount proportional to the ratio of the
1990 design value to the 24-hour average national ambient air
quality standard of 150 fig/m3; the 24-hour standard was assumed
to be the controlling standard in all cases. For certain areas
which were nearly in attainment as of 1990, PM emissions were
assumed not to grow beyond 1988 levels.
58
-------
Table 6-3. Ozone Nonattainment Areas Within the Modeling
Domain, with Minimum Source Size for RACT
Applicability.
Ozone
Nonattainment
Status
Minimum
Source Size
(tons
N0,/year)
Nonattainment Areas
Marginal
100
Reno, NV
Moderate :
100
Salt Lake City, UT
Monterey Bay, CA
, Phoenix, AZ
San Francisco Bay
Area, CA
Santa Barbara, CA
Serious
50
El Paso, TX
Sacramento, CA
San Joaquin Valley, CA
Severe
25
SE Desert AQMA, CA
San Diego, CA$
Ventura Co., CA
Extreme
10
Los Angeles SCAB, CA|
* Reformulated gasoline areas.
59
-------
For on-road motor vehicles, emission reductions anticipated as a
result of the following provisions of the 1990 CAAA were included
in addition to the projected VMT and fleet turnover effects
included in the 2005 base case inventory:
enhanced inspection and maintenance (I/M) programs in
serious, severe, and extreme ozone nonattainment areas
(areas with design values equal to or higher than 160
ppb and CO nonattainment areas with design values of
12.7ppb and above);
• basic I/M programs in moderate ozone nonattainment
areas (areas with design values equal to or higher than
138 ppb but less than 160 ppb and CO moderate
nonattainment areas with design values 12.7ppb and
below);
• reduced fuel Reid vapor pressure (RVP) specifications;
• new emission standards for light duty gasoline vehicles
and trucks;
reformulated gasoline in the Los Angeles and San D£ego
nonattainment areas.
Clean fuels provisions were not addressed since only hydrocarbon
emission reductions are predicted for these fuels by MOBILE 5.0.
Mobile source PM and SOX emissions were estimated using emission
factors from the PARTS model.
Utility SOX and NO, emissions for the 2005 CAAA case were
estimated using results from CEUM runs based on the "RIA low
trading case" as described in the acid rain regulatory impact
analysis (RIA) prepared by ICF Resources. These results take
into account the Title IV Phase II SO2 emission allowances, which
are for the most part based on an emission rate of 1.2 Ib/MMBtu
or less. The effects of trading of allowances between sources
and allowance banking are accounted for by the CEUM. New units
are allocated allowances if construction commenced by the end of
1990 and they are expected to be on-line by the end of 1995. All
other new units were assumed to be required to purchase
allowances to cover their emissions. As in the 2005 base case,
scrubbers with 90 percent S02 average removal efficiency were
assumed to be installed at the Navajo Generating Station power
plant.
Utility NO, emissions were calculated assuming an emission limit
of 0.45 Ib/MMBtu for existing tangentially-fired coal units and
0.5 Ib/MMBtu for existing wall-fired coal units and for any new
coal-fired units, as called for in Title IV. Since rules
allowing for averaging of NO, emissions across units for purposes
60
-------
of meeting these limits have not been finalized, the effects of
such averaging were not considered. In addition to these limits,
the 0.2 Ib/MMBtu and 0.3 Ib/MMBtu NOX RACT limits for tangential
and wall-fired gas and oil units, respectively, were imposed for
sources located in ozone ncnattainment areas.
Utility PM emissions were calculated as in the 2005 base case
inventory, but using the fuel consumption estimates calculated
from the CKUM for the "RIA low trading case."
Summary Comparison of Emission Inventories
Figures 6-2 through 6-5 provide a comparative summary of the
three annual emission inventories generated for this project
(1988, 2005 base case, 2005 CAAA case). Only the fine particle
(PMjj) fraction of the PM emissions is shown in these figures.
Figure 6-2 shows total emissions by species. The largest
reductions in the 2005 CAAA case are for NO, emissions which drop
below 1988 levels. SO, emissions decrease slightly while fine
particle emissions decreases in the CAAA case only partially
offset the expected growth from 1988 levels.
Motor vehicle and area sources dominate the NOX emissions (Figure
6-3), with the bulk of the total reduction expected to come about
as a result of mobile source controls.
For the modeling region, total utility SO, emissions are
comparable to those from other point sources (e.g., smelters,
refineries, pulp mills) and area sources (Figure 6-4). Only
utility sources are expected to contribute to decreased SO,
emissions as a result of the 1990 CAAA.
Fine particle emissions are dominated by motor vehicles and area
sources (Figure 6-5). A significant amount of growth in the
motor vehicle component is expected due to projected VMT growth
and the lack of any specific PM controls for this source category
in the 1990 CAAA. This growth accounts for the bulk of the
increase in total PM^ emissions. Area source PM emissions are
expected to decrease due to the requirement to bring certain
areas into attainment with the PM10 NAAQS.
Seasonal variations in NO, and SO, emissions were estimated to be
generally small, while the seasonal variations in PM^ emissions
exceed the expected increase from 1988 to the 2005 base case.
61
-------
Figure 6-2,
Estimated Total Gridded Annual Emissions for the
Southwestern Analysis (1988, 2005 Base, 2005
Control)
3,500
U1988
Base Case
B2005 GAAA
SOx
NOx
PMF
62
-------
Figure 6-3.
Estimated Total Gridded Annual NOx Emissions by
Source Category for the Southwestern Analysis
(1988, 2005 Base, 2005 Control)
3,500
3,000
2,500
rt
-2,000
Q.
W
.0 1,500
n
O
1,000
500
D1988
02005 Base Case
B2005 CAAA
Motor Vehicle Utility
Area/Low-Level Point
All
Other Point
Source Category
63
-------
Figure 6-4. Estimated Total Gridded Annual S02 Emissions by
Source Category for the Southwestern Analysis
(1988, 2005 Base, 2005 Control)
1,400
1,200
1,000
w,
«
$
800
ft
o
600
400
200
D1988
H2005 Base Case
S2005 CAAA
Motor Vehicle Utility
Area/Low-Level Point
All
Other Point
Source Category
64
-------
Figure 6-5.
Estimated Total Gridded Annual Fine Particle
Emissions by Source Category for the Southwestern
Analysis (1988, 2005 Base, 2005 Control)
700
600
500
(
-400
CL
(0
Lp 300
n
O
200
100
D1988
02005 Base Case
§2005 CAAA
M I
E-asa i
Motor Vehicle Utility
Area/Low-Level Point
All
Other Point
Source Category
65
------- OCR error (C:\Conversion\JobRoot\00000554\tiff\2000HCLQ.tif): Saving image to "C:\Conversion\JobRoot\00000554\tiff\2000HCLQ.T$F.T$F" failed.
-------
Air Quality Modeling
The emissions inventories prepared for the 1988, 2005 base, and
2005 CAAA control cases were used in an air quality modeling
application to estimate the effects of the projected emission
changes on long-term regional air quality concentrations in the
southwestern U.S. Modeled concentrations of visibility-impairing
species were then converted to visibility estimates in each of
the class I areas in the modeling domain. To select the most
appropriate modeling methodology for this project, several
important air quality modeling issues were identified:
The model must carry the important anthropogenic gaseous and
particulate components responsible for visibility
impairment, and their precursors;
The model should adequately address the important
environmental/meteorological influences on the non-linear
chemical interactions between primary precursor species and
secondary particulate and gaseous species;
Operating on monthly-seasonal-annual time scales, the model
must remain time and cost efficient;
The model should maintain an appropriate level of complexity
to adequately represent emission impacts on visibility-
related metrics at regional scales (on the order of 100-1000
km) .
The use of a computer-based dispersion model was preferred to
simulate the physical processes occurring in the atmosphere,
including emissions by time and location, their subsequent
transport, chemical transformation and deposition. A variety of
modeling options, however, were explored for use in this study.
The issues listed above were carefully considered for the
following modeling options:
Statistical models (linear chemistry);
Transfer matrices (linear chemistry);
Simple grid models (linear or pseudo-first order 'chemistry);
Complex grid models (detailed non-linear chemistry).
Statistical models and transfer matrices were determined to be
inappropriate because they over-simplify treatment of transport
and transformation processes, and because they are most
applicable for analyses focusing on particular time scales
(rather than a range of time scales). The complexity associated
with non-linear grid models prohibited their application to the
long integration periods required for this project, particularly
for a large regional domain such as the Southwest U.S.
66
-------
The National Park Service's Air Quality Modeling System (NPSAQMS;
Morris and Chang, 1992) was selected for this project. The
NPSAQMS contains version II of the Regional Transport Model (RTM-
II; a linear chemistry grid model) and a regional version of the
SAI Diagnostic Wind Model (METDWM). RTM-II was easily modified
to treat all of the major species found to be most abundant in
anthropogenic aerosols (PM,C and PM^); its modular structure
| allowed for the insertion of a pseudo-first order chemical
t mechanism; and its simplicity allowed integration over long time
I scales, providing gridded estimates of both short-term (hourly to
I daily) and long-term (monthly, seasonally, annually) averaged
I concentrations. The model was found to provide a reasonable
\ characterization of regional-scale transport, transformation, and
' deposition.
i Regional Transport Model
i
i The RTM-II (Morris and Chang, 1992) is a simple three-dimensional
• Eulerian air quality model that estimates the emission,
dispersion, deposition, and chemical transformation of several
anthropogenic gaseous and particulate species considered to be
most important in visibility degradation and acid deposition.
• The model originally carried primary anthropogenic emissions of
' sulfur dioxide (S02), sulfate (S04"), and nitrogen oxides (NO and
N02 combined into the single species NO,), along with secondary
linear chemical formation of nitric acid (carried as total
nitrate, N03") and sulfate (combined with primary sulfate) . A
number of species were added to the model in order to upgrade the
sulfate/nitrate chemical mechanism, to allow for the splitting of
total nitrate into nitric acid and particulate nitrate, and to
account for other primary anthropogenic particulate species;
these improvements are discussed in greater detail below. The
model now carries primary anthropogenic emissions of SO2, S04~,
NO, N02, total ammonia (NHj), fine particulate organic carbon,
fine particulate elemental carbon, "other" fine particulate mass
(metals, etc.), and coarse particulate mass, along with the
original secondary species of total nitrate and sulfate. Modeled
species concentrations and deposited mass from all grid cells are
output as 3-hour averages for the duration of the model run.
RTM-II Application
An RTM-II modeling domain consisting of a 38 by 26 horizontal
grid with 50 km grid cell size was selected for the current study
(Figure 6-1). The grid covers most of the significant source
areas expected to substantially contribute to visibility
degradation in class I areas throughout the southwestern U.S.
This rather coarse resolution was not expected to degrade the
quality of model predictions when considering the large temporal
and spatial.scales involved in the analysis.
67
-------
The three emission scenarios described above were simulated: 1988
base case, 2005 base case (without 1990 CAAA), and 2005 control
case (with 1990 CAAA). For each emission scenario, the RTM-II
was run for four 30-day periods using meteorological inputs
developed for April, July, October, and December 1988. These
months were determined to satisfactorily represent each of the
four seasons climatologically, except perhaps in terms of
rainfall. On one hand, 1988 was one of the worst of the last six
drought years in the Southwest, particularly in California. On
the other hand, these particular months were selected based on
their relative lack of rainfall for two reasons: (1) the mapping
of rainfall necessary for 120 days of model application was
rather labor intensive; and (2) uncertainties in model results
relating to the rainfall analyses and theoretical limitations of
the wet deposition routine were minimized.
The RTM-II was configured to estimate the emissions,
transformation, dispersion and deposition of anthropogenic
aerosol mass, as discussed above, since visibility impacts from
future year emissions growth and the 1990 CAAA mandated controls
are manifested in changes to only the anthropogenic fraction of
particulate mass (and N02). Seasonally dependent natural and
anthropogenic background components of aerosol mass generated
within the domain and entering through the domain boundaries
(including sulfate, nitrate, natural/boundary/secondary organic
carbon, elemental carbon, and fine and coarse dust) were added to
the predicted species for the 1988 base case performance
evaluation and for all visibility calculations. Therefore, in
the RTM-II model applications, initial and boundary conditions
for all aerosol mass were set to zero.
Secondary species such as sulfate, nitric acid, and ammonium
nitrate, are formed from both anthropogenic and
natural/background precursor gasses. Due to the non-linearity of
sulfate and nitrate chemical transformations, it was necessary to
include the contributions of background concentrations of NO,
N02, S02, and NH3 to the formation of the secondary particulate
species. In principle, then, modeled "anthropogenic" sulfate and
nitrate do include a very small fraction of mass generated from
background precursor species. Initial and boundary conditions
for the gaseous precursor species were set to global background
values (Table 3-2) obtained from previous modeling studies and
the literature. The RTM-II was allowed to initialize from the
initial state (zero aerosol mass, non-zero gaseous precursor
concentrations) over the course of one or two days at the
beginning of each month modeled.
i
Meteorological and Data Input Preparation
The RTM-II requires several environmental input fields to drive
the advective, diffusive, chemical, and deposition processes
68
-------
simulated in the model. Along with two-dimensional static fields
containing gridded terrain heights and land use-dependent gridded
dry deposition parameters, the model requires time-dependent
three-dimensional gridded inputs of wind components, temperature,
humidity, pressure, and two dimensional gridded inputs of mixing
heights, rainfall rate, and cloud cover. All of the gridded
meteorological fields (except rainfall rate) for April, July,
October, and December 1988 were prepared using the NPSAQMS METDWM
meteorological preprocessor. Data storage considerations and the
long-term modeling timescales prompted the development of all
I time-varying model input data to three-hourly, rather than
$ hourly, formats.
i
I The RTM-II requires gridded ozone'concentrations that are used as
I an estimate of the oxidation potential of the atmosphere for its
* pseudo-first order chemical mechanism. Observed hourly ozone
: concentration data for 1988 from the EPA AIRS network were
'": averaged to 3-hourly values then interpolated to each grid cell
: in the model's mixed layer using an inverse distance weighting
•: within a particular radius of influence (200 km). It was assumed
that interpolated surface ozone measurements were representative
for the entire mixed depth for a given grid column. For remote
areas lacking ozone monitors (i.e. beyond the radius of influence
of any monitor) and known to likely experience tropospheric
background ozone concentrations (e.g., over ocean, in remote
rural areas), ozone concentrations were set to 40 ppb. It was
also assumed that ozone at the top of the model (3000 m) was at a
tropospheric background value of 40 ppb; ozone concentrations
within the upper two layers were linearly interpolated between
the mixed layer value and the top of the modeling coluagft.
Reduced ozone input fields were developed for the 2005 control
case, following the assumption that full implementation of the
1990 CAAA within all nonattainment areas covered by the modeling
domain will be effective in reducing ozone design values to the
federal ozone standard. For each 3-hour period, the ozone
concentration over a background of 40 ppb, in each grid cell
contained within a nonattainment area, was scaled back by the
percent difference of the design value for the nonattainment area
and the 120 ppb standard.
Air Quality Model Results
After each month was modeled by the RTM-II, season-dependent
natural background and anthropogenic boundary contributions were
added to the predicted aerosol species at eight representative
class I areas containing IMPROVE particulate samplers. These
eight sites are Grand Canyon (AZ), Arches (UT), San Gorgonio
(CA), Pinnacles (CA), Yosemite (CA), Chiricahua (AZ), Handelier
(NM), and Rocky Mountain (CO). All subsequent performance
evaluations and visibility calculations were carried out for
these eight class I areas.
69
-------
Model-predicted 3-hourly concentrations of visibility-related
species from the 1988 Base Case scenario were averaged to
seasonal values for grid cells covering the eight representative
class I areas. Background contributions were then added on a
site- and season-specific basis as described above. Resulting
concentrations were then compared to 1988 season average IMPROVE
data at each site for those species for which direct comparison
could be made; these include ammonium sulfate, ammonium nitrate,
organics, elemental carbon, total fine mass, and coarse mass.
In terms of absolute differences between modeled and.measured
concentrations, model performance was quite good at all sites
(with minor exceptions); absolute error ranged from 0.0 to about
1.0 /zg/m3 for most; fine mass species. Coarse mass performance
tended to be worse in terms of absolute differences, but since
the model predicted very little coarse mass for class I areas
removed from urban areas, this appeared to be a result of a
positive bias between the 1988-1991 average coarse mass used for
background conditions and the 1988 average coarse mass. In terms
of relative differences, the model performance for most species
and most sites ranged from 0% to 50%, averaging 10-20%.
Estimation of Visibility
The air quality modeling results were used to estimate the
changes in seasonal and annual average visibility expected in
each of eight representative class I areas in the southwestern
U.S. due to implementation of the CAAA. First, a light
extinction budget was developed for each representative class I
area in order to apportion total extinction by atmospheric
constituent.
Construction of Extinction Budgets
The 3-hour average concentrations of visibility-related species
plus background values at each site were multiplied by their
respective extinction efficiencies, and their products were
averaged over each season (i.e., modeling month) and for the
year. Adding all species extinction coefficients to a Rayleigh
scattering coefficient of 10 Mm*1 yields seasonal and annual
average total light extinction coefficients.
For the 1988 Base case, the relative distribution of extinction
agrees rather well with observations and measurements for all
sites: extinction in class I areas in the desert Southwest is
dominated by sulfate, organics, and dust; extinction at the San
Gorgonio site in southern California is dominated by nitrate and
organics; while extinction for sites in heavily forested
mountainous areas is dominated by organics most likely from
biogenic sources.. All sites indicate a large fraction of
extinction due to elemental carbon; even at typically low
70
-------
concentrations below 1 jig/m3, the very high extinction efficiency
(mostly absorption) causes elemental carbon to contribute
substantially to overall iight extinction. Overpredictions of
ammonium nitrate at Yosemite and Rocky Mountain are manifested in
high nitrate contributions to overall extinction at those sites.
Visual Range and Deciview Estimates
Visual range and deciview were calculated from the 3-hour average
extinction budgets and averaged to seasonal and annual average
values. Figures 6-6 through 6-13 display visual range and
deciview at all eight class I areas by season (and annual
average) for all three emission scenarios. The growth in
emissions between 1988 and 2005 base cases is reflected in a
slight reduction of visual range predicted at all sites during
r the entire year. Controls required by the CAAA are predicted to
| offset this emissions growth so that visual ranges for the CAAA
I control case are predicted to be similar to 1988 base case
levels. Changes are quite small (less than one DV) in all cases
'• except San Gorgonio. In the case of San Gorgonio wilderness a
i special review of the influence of nitrate on the extinction
t budget and the resulting change in nitrate concentrations that
i might occur with attaninnent of the ozone ambient standard in the
• Los Angeles area, indicates that this area could see perceptible
* improvements in visibility of 1-2 deciviews due almost entirely
• to reductions in aerosol nitrate.
Southwestern Modeling Conclusions
The RTM-II dispersion model was used to estimate the expected
changes in air quality and visibility at class I areas in the
southwestern U.S. due to implementation of the 1990 CAAA.
Provisions of the CAAA were examined to determine the effects on
future emissions and then these emission projections were used in
the model to develop estimates of changes in ambient
concentration of visibility-related species. The effect on
extinction budgets, visual range and deciview were then estimated
from the change in concentrations.
Significant emission reductions mandated by the CAAA are targeted
primarily at nonattainment areas. However, those provisions of
the CAAA, for which quantifiable emission reductions could be
determined, were estimated to result in only modest emission
reductions throughout the southwest modeling domain. Not
surprisingly, therefore, modeling results indicate that little.
change is to be expected in the distribution of visual range at
class I areas. For example, implementation of the CAAA by 2005
is estimated to increase visual range at Grand Canyon National
Park from 122.5 km to 123.2 km, an increase of only 0.6 percent.
71
-------
Figure 6-6.
Estimated Visibility (visual range and deciview)
at Grand Canyon National Park
140
120
£100
v
to
§
K
80
I 40
10
8
6
5 4
"o
v
o
Seasonal Average Visual Range
Grand Canyon
Spring Summer Autumn Winter
Seasonal Average Deciview
Grand Canyon
Annual
-------
Figure 6-7.
Estimated Visibility (visual range and deciview)
at Arches National Park
Seasonal Average Visual Range
Arches
140
120
5100
g1 80
0
20|
0
14
12
I
10
8
t
T 6
"5
v
0 4
2
0
Spring
Summer
Autumn
Winter
Seasonal Average Deciview
Arches
Annual
Spring
Summer
Autumn
Winter
Annual
73
-------
Figure 6-8. Estimated Visibility (visual range and deciview)
at San Gorgonio Wilderness
100
80
60
BC
0
K
40
Seasonal Average Visual Range
San Gorgonio
w
>
20
Spring Summer Autumn Winter
Seasonal Average Deciview
San Gorgonio
Annual
\V
Spring Summer Autumn
Winter
Annual
74
-------
Figure 6-9.
Estimated Visibility (visual range and deciview)
at Rocky Mountain National Park
Seasonal Average Visual Range
Rocky Mtns
140
120
v
c 80
e
K
•Z 60
20
0
14
12
10
B
6
13
v
0 4
2
0
v
Spring Summer Autumn Winter
Seasonal Average Deciview
Rocky Mtns
Annual
Spring
Summer Autumn Winter
Annual
75
-------
Figure 6-10. Estimated Visibility (visual range and deciview)
at Pinnacles Wilderness
Seasonal Average Visual Range
Pinnacles
140
120
£.100
v
s? 80
c
- 60
40
20
0
Spring
Summer
Autumn
Winter
Annual
Seasonal Average Deciview
Pinnacles
14
12
10
B
I 6
Spring
Summer
Autumn
Winter
Annual
76
-------
.*•
*
*
3
Figure 6-11. Estimated Visibility (visual range and deciview)
at Yosemite National Park
Seasonal Average Visual Range
Yosemite
140
120
£.100
C 80
0
K
, eo;
20
0
Spring
Summer Autumn
Winter
Seasonal Average Deciview
Yosemite
Annual
Spring
Summer
Autumn
Winter
Annual
77
-------
Figure 6-12. Estimated Visibility (visual range and deciview)
at Chiricahua Wilderness
Seasonal Average Visual Range
Chiricahua
140
120
<-»
5100
v
00
an
BO
, 60
so
0
Spring Summer Autumn Winter
Seasonal Average Deciview
Chiricahua
14
12
10
8
v
«
I
Annual
\\\
Spring
Summer • Autumn
Winter
Annual
78
-------
Figure 6-13. Estimated Visibility (visual range and deciview)
at Bandelier Wilderness
140
120
£100
60 on
C BO
0
K
1 6°
20
0
14
12
10
6
5
T 6
I*
2
0
Seasonal Average Visual Range
Bandelier
Spring
Summer Autumn
Winter
Annual
Seasonal Average Deciview
Bandelier
Spring
Summer Autumn Winter Annual
79
-------
This shift in the distribution of visibility conditions is not
significant. The only exception to this limited change was the
review of nitrate changes at San Gorgonio wilderness showing a 1-
2 deciview improvement in annual average visibility levels.
However, the predicted 1-2 deciview improvement is dependent on
the Los Angeles area meeting attainment of the ozone national
ambient air quality standard by 2005.
Uncertainties
Emissions
The results presented in this report are based on a number of
assumptions. The air quality modeling results, and hence the
estimates of future visibility, are sensitive to a number of
inputs, some of which are fairly uncertain. For example,
emission projections indicated that, without the CAAA, there
would be sufficient growth in emissions generating activities
between 1988 and 2005 to largely offset the expected NO, emission
reductions and more than offset the modest reductions expected in
fine particle emissions. A significant amount of growth in
particulate emissions was due to predicted growth in mobile
source VMT estimates which are fairly uncertain. If actual VMT
growth was less than estimated, larger improvements in visibility
would have been demonstrated. However, the predicted visibility
difference between the two future year scenarios (with and
without CAAA) would not be significantly changed.
As in any study of this type, uncertainties in emissions are
likely to be large. The VOC emissions were not examined due to
their relatively small role in visibility in the remote Southwest
and the high cost of including them in the study. It should be
noted that the 2005 base case is a highly artificial case, but it
is useful as a reference case for assessing the magnitude of
effects from uncontrolled growth in emission-generating
activities. The emission increases due to increases in VMT are
particularly striking. It should also be noted that the 2005
CAAA case includes only quantifiable emission reductions directly
required by the 1990 CAAA. The effects of other CAAA provisions
and state and local programs are not included, so actual 2005
emissions may possibly be lower than assumed in this study,
resulting in improvements in visibility not estimated here.
Role of Secondary Organic Aerosols
Secondary organic aerosols (SOA) arise from the photochemical
oxidation of certain VOC emissions, which produce low vapor
pressure products that can condense into the aerosol phase. SOA
can form from either anthropogenic or biogenic VOC emissions; in
fact, the aerosol-forming potential of biogenic VOC is much
greater than that of most anthropogenic VOC (e.g., Grosjean and
80
-------
Seinfeld, 1989). Estimates of the secondary contribution to
organic aerosol concentrations have been made for the Los Angeles
area (Gray et al., 1986; Turpin and Huntzicker, 1991; Pandis et
al., 1992), and tend to be in the range of 20 to 40 percent
during the summer. In the winter, little SOA would be expected •
because of reduced photochemical activity. Since organic
aerosols often account for 30 to 50 percent of fine mass,
jr secondary organic aerosols potentially contribute up to 20
* percent of fine mass during the summer, a fraction comparable to
i the sulfate contribution in many locations and greater than the
* nitrate contribution in all but the most polluted locations.
Methods for identifying the secondary component of organic
aerosols often focus on the ratio of organic to elemental carbon
(OC/EC). If OC/EC for the primary aerosol is known, then SOA can
be identified by increases in OC/EC. The difficulty with this
approach is that primary OC/EC is not constant for all source
types. If variations in primary OC/EC are large, then
uncertainties in estimating the SOA will also be large.
The RTM-II does not currently include SOA formation, as discussed
above. Thus, the estimates derived in the present study of the
impact of the CAAA on fine particle concentrations and visibility
do not include any possible reductions in SOA that might arise
from the CAAA. Yet, reductions in precursor VOCs are required by
the CAAA in ozone nonattainment areas, and could potentially
impact SOA concentrations in downwind class I areas.
To estimate the contribution of SOA at class I sites, and the
potential reductions in SOft due to the CAAA, the OC and EC data -
from selected IMPROVE sites within the modeling domain were
examined. In the Los Angeles studies, ambient OC/EC ratios were
available either with high temporal resolution (Turpin and
Huntzicker, 1991), or for several sites at various downwind
distances from the primary source region (Gray et al., 1986). In
both cases, primary OC/EC ratios could be estimated readily,
either from the early morning ambient OC/EC ratio or from the
ambient OC/EC ratio in near-source regions. However, neither of
these approaches could be used with the IMPROVE data. Instead,
winter minimum OC/EC ratios were used as a surrogate for the
primary OC/EC ratios, and the SOA contribution in summer was
estimated from the differences between the summer and winter
OC/EC.
Table 6-4 shows the measured winter minimum OC/EC and the summer
mean OC/EC for selected IMPROVE sites. For each site, the winter
minimum was calculated based on an average of the lowest 6 to 10
values. The estimated summertime contribution of SOA ranges from
13 to 55 percent, amounting to 0.5 to 1.2 fig/m3. The maximum
estimated SQA Concentration occurs at San Gorgonio, located
downwind Of the Los Angeles* area.
81
-------
Table 6-4.
Summer and Winter OC/EC Data and Estimated Summer
SOA Concentrations at Selected IMPROVE Sites.
SITE
Grand
Canyon
Arches
Pinnacles
Bandelier
Rocky Mtn
Yosemite
San
Gorgonio
Chiricahua
Winter
minimum
OC/EC
2.02
2.80
.2.42
•: 3.09
2.90
3.65
2.13
2.82
Summer
mean
OC/EC
4.60
5.99
4.38
6.41
5.94
4.18
3.37
6.29
Summer
SOA
percent
56
53
.45
52
51
13
37
55
Summer
OC
1.26
1.40
1.15
1.18
2.20
4.13
3.22
1.35
Summer
SOA
(/ig/m*
0.71
0.74
0.52
0.61
1.12
0.54
1.19
0.74
82
-------
The uncertainties in these values are high, mainly because of the
uncertainty in the validity of the assumption that the OC/EC
ratio is constant for summer and winter. There are numerous
reasons why this may not be the case. Wood burning emissions,
for example, are characterized by very high OC/EC ratios and have
strong seasonal patterns. Areas impacted by residential wood
combustion in winter, or forest fires, prescribed burns, or even
campfires in summer, may exhibit high OC/EC ratios that are not
due to secondary formation. In addition, primary biogenic
organic aerosols may have a significant impact upon total organic
aerosol concentrations in class I areas (Simoneit and Mazurek,
1982).
To estimate an upper bound to the possible reduction in SOA due
to the CAAA, the San Gorgonio site was used for further analysis.
First, the estimate of Pandis et al. (1992) of 0.5 fig/m3 for
biogenic SOA at Claremont was assumed to represent a lower bound
for the biogenic SOA concentration at San Gorgonio (San Gorgonio
is much more heavily wooded than the Los Angeles basin) .
The contribution of primary biogenic organic aerosol was 0.2
/ig/m3 (Simoneit and Mazurek, 1982) . Of the 1.2 estimated at
estimated for total SOA concentration in summer,, subtraction of
0.7 jig/m3 due to biogenic primary and secondary organic aerosol
leaves an estimated anthropogenic SOA concentration of 0.5 /xg/m3.
In order to meet the CAAA deadline for ozone attainment, the Air
Quality Management Plan for the Los Angeles area calls for VOC
reductions of 83 percent (SCAQMD, 1991). Therefore, the
reduction in anthropogenic SOA at San Gorgonio due to the CAAA is
estimated at 83 percent of 0.5 /zg/m3, or 0.42 pig/m3. Applying the
scattering efficiency for organic aerosol of 6 m2/g gives an
estimated reduction in summertime extinction of 2.5 Mm'1 for SOA.
Although this is a significant fraction of the extinction
reduction for San Gorgonio simulated by the RTM-II, it is still
small in an absolute sense. Keeping in mind that this estimate
has a high degree of uncertainty, it suggests that future
modeling analyses of visibility impacts in Southern California
should address SOA.
The impact of the CAAA on total extinction at all other class I
areas will be considerably smaller than the estimate derived for
San Gorgonio, for two reasons. First, the estimated total SOA
concentration (including primary biogenic organic aerosol) in
Table 6-4 is lower at all other sites, ranging from 0.5 to l.l
jig/m3. Assuming a minimum total biogenic contribution of 0.7
/ig/m3 leaves only Rocky Mountain National Park with any
significant estimated anthropogenic SOA contribution. Second,
the VOC reductions required by the CAAA will be lower for air
basins located upwind of all other sites than the 83 percent
assumed for the Los Angeles area.
83
-------
References for Chapter 6
1. Taback, H.J., A.R. Brienza, J. Macko, and N. Brunetz, Fine
Particle Emissions from Stationary and Miscellaneous Sources in
the South Coast Air Basin. KVB Inc. document no. KVB 5806-783,
TUBtin, California, 1979 (includes an important appendix volume).
Applications to Los Angeles. 2 vols., Ph.D. thesis, California
Institute of Technology, Pasadena, EQL report no. 16-2, 1977.
3. Gray, H.A., Control of Atmospheric Fine Primary Carbon
Particle Concentrations. Ph.D.. thesis, California Institute of
Technology, Pasadena, EQL report no. 23, 1986.
4. Gray, H.A.,' Reid, S.B., and Chinkin, L.R., Carbon Particle
Emissions Inventory for Denver Brown Cloud II; Development and
Assessment, prepared for Colorado Department of Health by
Systems Applications International, San Rafael, California, 1992.
5. Gray, H.A., M.P. Ligocki, G.E. Moore, C.A. Emery, R.C.
Kessler, J.P. Cohen, C.C. Chang, S.I. Balestrini, S.G. Douglas,
R.R. Schulhof, J.P. Killus, and C.S. Burton, Deterministic
Modeling for Navajo Generating Station Visibility Study (Volume 1
and Appendix D). SYSAPP-91/-45a, prepared for Salt River Project
by Systems Applications International, San Rafael, California,
1991.
6. Rothschild, S.S., The National Allowance Data Base Version
2.1 -- Technical Support Document. Pechan & Associates, Inc.,
Springfield, Virginia, 1992.
7. EPA, User*s Guide for the Urban Airshed Model. Volume IV;
User's Manual for the Emissions Preprocessor System 2.0; Part A:
Core FORTRAN System. EPA-450/4-90-007D(R), U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards,
1992.
8. EPA, Procedures for Preparing Emissions Projections. EPA-
450/4-91-019, U.S. Environmental Protection Agency, 1991.
9. EPA, MOBILE 5.0. draft version. U. S. Environmental
Protection Agency, San Francisco, California, 1992.
10. Shepard, S.B., H.A. Gray, and J.G. Heiken, User's Guide to
PARTS: A Program for Calculating Particulate Emissions from Motor
Vehicles. SYSAPP-92/101, prepared for U.S. Environmental
Protection Agency, Office of Mobile Sources, by Systems
Applications International, San Rafael, California, 1992.
84
-------
11. BEA, BEA Regional Projections to 2040. Volume 1; State
Projections. Bureau of Economic Analysis, United States
Department of Commerce, 1990.
12. Castello, M., personal communication, letter to B. Polkowsky
dated April 14, 1992.
13. Teague, M., personal communication, 1992.
14. ICF, 1989 EPA Base Case Forecasts, prepared for EPA by ICF
Resources Inc., May 1989.
15. ERC, "Generating Unit Additions and Retirements," in
Electricity Supply and Demand 1992-2001. 1992.
16. DOE/EIA, "Projected Generating Unit Additions," in
Inventory of Powerplants 1991. Table 21, 1991.
t
[ 17. EPA, Regional Ozone Modeling for Northeast Transport
; (ROMNET). EPA-450/4-91/002, U.S. Environmental Protection Agency,
1991.
18. ICF, Regulatory Impact Analysis of the Proposed Acid Rain
Implementation Regulations. prepared for EPA by ICF Inc.,
September 16, 1991.
19. Morris, R.E, and C.C. Chang, Users guide to the National
Park Service's Air Quality Modeling Svs-tem. SYSAPP-92/009,
prepared for the National Park Service by Systems Applications
International, San Rafael, California, 1992.
20. Morris, R.E., T.C. Myers, B.L. Carr, M.C. Causley, and S.G.
Douglas, Users Guide for the Urban Airshed Model. Volume II:
Users Manual for the UAMfCB-IV) Modeling System. (EPA-450/4-90-
007b, prepared for U.S. Environmental Protection Agency by
Systems Applications International, San Rafael, California, 1990.
21. Smolarkiewicz, P.K., "A simple positive definite advection
scheme with small implicit diffusion," Monthly Weather Review.
Ill, 479-486 (1983).
22. Smagorinsky, J., "General circulation experiments with the
primitive equations: I. The basic experiment," Monthly Weather
Review. 91, 99-164 (1963).
23. Morris, R.E., R.C. Kessler, S.G. Douglas, and K.R. Styles,
The Rockv Mountain Acid Deposition Model Assessment Project.
SYSAPP-87/218, prepared for U.S. Environmental Protection Agency
by Systems Applications International, San Rafael, CA, 1987.
24. Douglas, S.G., R.C. Kessler, and E.L. Carr, Users Guide for
the Urban Airshed Model. Volume III; Users Manual for the
85
-------
Diagnostic Wind Model. EPA-450/4-90-007C, prepared for U.S.
Environmental Protection Agency by Systems Applications
International, 1990.
25. Benkley, C.W. and L.L. Schulman, "Estimating hourly mixing
depths from historical meteorological data, " J. Appl. Meteor.,
18, 772-780 (1979).
26. Davidson, J.A., C.A. Cantrell, A.H. McDaniel, R.E. Shetter,
S. Madronich, and J.G. Calvert, "Visible-ultraviolet absorption
cross sections for N02 as a function of temperature," J. Geophys.
Res.. 93, 7105-7112 (1988).
27. Szkarlat, A.C. and S.M. Japar, "Light absorption by
airborne aerosols: comparison of integrating plate and
spectrophone techniques," Appl. Opt.. 20, 1151-1155 (1981).
28. Japar, S.M., A.C. Szkarlat, and W.R. Pierson, "The
determination of the optical properties of airborne particle
emissions from diesel vehicles," Sci. of the Total Environ.. 36,
121-130 (1984).
29. Adams, K.M., L.I. Davis, S.M. Japar, D.R. Finley, and R.A.
Gary, "Measurement of atmospheric elemental carbon: real-time
data for Los Angeles during summer 1987," Atmos. Environ.. 24A,
597-604 (1990).
30. Turpin, B.J., J.J. Huntzicker, and K.M. Adams,
"Intercomparison of photoacoustic and thermal-optical methods for
the measurement of atmospheric elemental carbon," Atmos.
Environ.. 24A, 1831-1835 (1990).
31. Watson, J.G., J.C. Chow, L.W. Richards, D.L. Haase, C.M.
McDade, D.L. Dietrich, D. Moon, and C. Sloane, The 1989-90
Phoenix Urban Haze Study. Volume III; The Apportionment of Light
Extinction to Sources. DRI Document No. 8931.5D1, prepared by
Desert Research Institute, Reno, Nevada, 1991.
32. Gray, H. A., G. R. Cass, J. J. Huntzicker, E. K. Heyerdahl
and J. A. Rau. 1986. Characteristics of atmospheric organic and
elemental carbon particle concentrations in Los Angeles.
Environ. Sci. Technol., 20:580-589.
33. Grosjean, D. and J. H. Seinfeld. 1989. Parameterization of
the formation potential of secondary organic aerosols. Atmos.
Environ., 23:1733-1747.
34. Pandis, S. N., R. A. Harley, G. R. Cass, and J. H. Seinfeld.
1992. Secondary organic aerosol formation and transport. Atmos.
Environ., 26A:2269-2282.
86
-------
35. SCAQMD. 1991. Air Quality Management Plan. Final 1991
Revision. South Coast Air Quality Management District. July
1991.
36. Simoneit, B. R. T. and M. A. Mazurek. 1982. Organic matter
of the Troposphere. II. Natural Background of Biogenic Lipid
Matter in Aerosols over the Rural Western United States. Atznos.
JSnviron., 16:2139-21f,9.
37. Turpin, B. J. and J. J. Huntzicker. 1991. Secondary
formation of organic aerosol in the Los Angeles basin: a
descriptive analysis of organic and elemental carbon
concentrations. Atmos. Environ., 25A:207-215.
87
-------
Chapter 7. Comparison of Modeling Approaches
Comparison of Eastern and Southwestern Approaches
Estimates of Current and Future Emissions
Independent procedures for estimating base case and future year
emission levels under the provisions of the 1990 CAAA were used
for the eastern and southwestern U.S. visibility modeling
studies. Differences in the projection procedures arise
primarily as a result of the different requirements of the
eastern and southwestern visibility modeling, approaches: emission
estimates for the Southwest were prepared for S02, NOX, and
particulate matter for 1988 (base case) and 2005 (base and
control cases), while emission estimates for the East were
prepared for sd2 only for 1985 (base case) and 2010 (control
case). In both areas, future year S02 emission reductions were
assumed to result solely from changes in emissions by electric
utilities in response to changes in electric power demand and
full implementation of the provisions of Title IV of the 1990
CAAA.
In the southwestern analysis, future year utility S02 emissions
were estimated using the Coal Electric Utility Model (CEUM) as
described in the S02 regulatory impact analysis (RIA) prepared by
ICF Resources. This model accounts for the S02 emission
allowance trading and banking provisions of Title IV. The CEUM
results were based on the demand forecast and fuel price
assumptions contained in the "RIA low trading case" used in the
S02 RIA. Since these assumptions are currently considered
outdated, modifications were made to the CEUM results to bring
the power production and fuel use predictions in line with
current government and industry forecasts for the Southwest.
These modifications resulted in virtually no increase in coal
consumption, and very little increase in oil consumption between
1988 and the 2005 control case.
In the eastern analysis, projected utility S02 emissions were
estimated based on DOE forecasts of electricity demand ("Annual
Outlook for U.S. Electric Power 1991", DOE/EIAOl), July, 1991).
These estimates were used in conjunction with assumptions about
unit retirement rates, capacity utilization, etc. contained in
the "EPA High Base Case" as input to the AIRCOST/PC model to
allocate generation by existing, planned, and (new) unplanned
units and thus obtain projected S02 emissions for 2010. This
approach did not account for any effects of trading or banking of
S02 emission allowances on the total amount and geographical
distribution of S02 emissions. However, the effect of allowance
banking over the time period of interest (to the year 2010) is
likely to be negligible and any geographic shifts in emissions
within the East due to large scale allowance trading are
88 :
-------
speculative and not likely to effect the conclusions regarding
visibility improvements for the region as a whole.
Although the procedures used to estimate reductions in S02
emissions resulting from implementation of the 1990 CAAA were
different in the Southwest and the East, both approaches are
believed to provide reasonable estimates of future year S02
emissions under the 1990 CAAA.
Comparison between RIM-II and RADM Visibility Modeling
Several significant differences exist between the air
quality/visibility modeling efforts carried out for the eastern
U.S. (using RADM-EM) and the southwestern U.S. (using RTM-II).
First and foremost, these two air quality models differ in the
chemical species they carry, and in their treatments of the
physical processes responsible for dispersion, deposition, and
< chemical transformation. Although RADM-EM is a simplified
j version of RADM2.0, it incorporates many more physical processes,
I and more complex parameterizations, than those included in the
, RTM-II. Second, the methods used to calculate visibility
'; estimates from model output concentrations differ substantially
! between the two regions. While the methods employed for the
\ southwest are rather complete and detailed, the analysis for the
• • east is primarily based on sulfate concentrations, modified to
; include contributions of non-sulfur species to total light
extinction.
The full RADM 2.0 modeling system incorporates a 3-D Eulerian
framework, in which 6 layers of 35x38 80 km grid cells are
specified to cover the eastern U.S. from the surface up to about
15 km. This results in resolving the lowest 300.0 m of atmosphere
into 4 vertical layers. The model is based on solving a
conservation equation that treats emissions, dispersion,
deposition, and chemical transformation of 63 chemical species
within each model cell. RADM employs the Smolarkewicz advection
scheme, along with explicit first-order vertical diffusion based
on bulk parameterizations of turbulent exchange coefficients in
various atmospheric layers. The model is supplied with emissions
of S02, S04, NO, N02, 16 species of VOC, NH3, and CO. The gas-
phase chemical mechanism is a highly detailed non-linear
mechanism that simultaneously solves 157 chemical reactions that
include 42 species of organics and 20 photolysis rates. Dry
deposition is calculated based on the resistance method in three
physical regimes: the turbulent boundary layer, the laminar
sublayer, and the canopy layer. RADM requires the specification
of cloud fields (including location, type and coverage) to
calculate aqueous chemical transformations, wet removal, and the
vertical redistribution of pollutants in convective-type cloud
systems.
89
-------
The RADM-EM was designed to reduce the excessive computational
cost of running RADM2.0. It is internally identical to RADM in
all respects except that the chemical transformation mechanisms
are a mathematical approximation of the chemical
parameterizations in the full non-linear model. RADM-EM directly
treats only those gas- and aqueous-phase reactions related to
sulfur, and therefore only predicts the emissions, dispersion,
transformation, and deposition of S02, S04, and HjOj. Photolysis
and aqueous chemical rates are calculated in a manner identical
to that of the full RADM model, but OH and H02 concentrations are
externally supplied from the RADM output chemical files. This
procedure is based on the inherent assumption that NOI/VOC
chemistry remains unchanged for a particular set of RADM-EM
simulations; therefore, RADM-EM can handle input emissions
inventories in which only sulfur loadings are altered. The model
requires RADM input meteorological files and RADM output chemical
files, and recomputes sulfur air quality using various sulfur
emission inventories.
Whereas RADM-EM explicitly treats the production of sulfate via
both gaseous and aqueous pathways, and parameterizes dry and wet
surface removal (all using highly complex and non-linear
methods), sulfate is the only aerosol species carried by the
model. On the other hand, RTM-II condenses gaseous sulfur and
nitrogen oxidation into pseudo-first order calculations while
parameterizing humidity-dependent oxidation rates and dry surface
deposition rates in a rather simplistic manner. However, RTM-II
carries several species (nitrate, primary organics, etc.) known
to play crucial roles in visual air quality throughout the
western U.S. Besides the chemical aspects, RTM-II also departs
from RADM-EM in its treatment of vertical grid structure in that
the first layer of RTM-II contains the entire mixing layer, while
two upper layers evenly divide the region between the top of the
mixed layer and the top of the model at 3000 m. Many other
aspects of RTM-II, however, are similar to RADM-EM, including wet
deposition and Smolarkewicz advection algorithms. Calculated dry
deposition rates, although simplistic compared to RADM, utilize
RADM land use categories and first-order reaction rates based on
RADM calculations. The 3-D Eulerian RTM-II was configured for
the southwest using 3 layers of 38x26 50 km grid cells.
As a result of the simplifications required in modeling sulfur
controls with RADM-EM, the calculation of visibility-related
parameters in the eastern U.S. was limited to light extinction
from natural temperature- and pressure-dependent Rayleigh
scattering and aerosol scattering due to sulfate. Sulfate
scattering efficiencies were determined using a formulation in
which efficiencies increase logarithmically with relative
humidity. Gridded total base case extinction (Rayleigh plus
sulfate) was uniformly scaled up so that total extinction
averaged over the entire RADM domain equaled the domain-average
90
-------
observed extinction from eastern U.S. IMPROVE sites. This was
done to incorporate the contributions of other non-controlled
aerosol species to total extinction. This same extinction
increment was used for the sulfur control results (again, impacts
to exti»ction resulted from changes to sulfate air quality only).
Seasonal and annual distributions of visual range and deciview
were then calculated from the gridded extinction data for each
| emissions scenario.
i
I For the southwest U.S., visibility-related parameters were
| calculated using RTM-II concentrations augmented with specified
\ background concentrations based on measurements at several
I western IMPROVE sites. Light extinction was calculated using a
I widely accepted constant value for Rayleigh scattering, and
1 impacts from the following constituents: N03; modeled and natural
'-, background sulfate; modeled and natural background aerosol
• nitrate; modeled primary, natural, and anthropogenic secondary
l organics; modeled and natural elemental carbon; modeled "other"
* fine mass (metals, road dust, etc.) and natural fine dust; and
modeled coarse mass and natural coarse dust. Extinction
efficiencies for sulfate, nitrate, and organics were calculated
from humidity-dependent logarithmic functions similar to that
> . used for sulfate in the eastern U.S. Extinction efficiencies for
- the other species were held constant at widely accepted values.
- In the South Coast Air Basin, extinction efficiency for organics
was held constant, assuming that organics in that area do not
tend to be as hygroscopic due to their differing origin and
formation pathways. Seasonal and annual distributions of
deciview and visual range were then calculated from gridded «
extinction data for each emissions scenario.
In summary, RADM-EM models only sulfur species, but the model
contains superior mechanisms over RTM-II for gaseous and aqueous
sulfur oxidation pathways, and for dry and wet deposition
calculations. This is important for estimating baseline and
future visibility-related parameters in the eastern U.S. since
sulfate dominates visual air quality in that region. RTM-II, on
the other hand, simplifies treatments of sulfur oxidation
chemistry and deposition, yet models several other significant
primary and secondary aerosol species. The design of RTM-II
makes it more applicable to modeling visual air quality in the
western U.S.; in comparison to the east, the smaller and more
wide-spread sulfur emission loadings, combined with the drier
environment, lead to less sulfate formation. This results in a
larger fraction of several other aerosol chemical species such as
nitrate, organics, and dust, most of which are in turn influenced
by 1990 CAAA controls.
91
-------
Sensitivity of Results
As mentioned in the discussion of the modeling approaches, the
emissions growth factors for general economic development are key
to regional changes in visibility beyond the major sulfur dioxide
reductions expected under Title TV of the Clean Air Act as
amended. If economic growth is slower than projected or if new
industrial processes are developed that produce less pollution
for the same industrial output, then more improvements in
regional visibility could occur than those predicted in this
report. This may be the case for class I areas near the Los
Angeles Basin which are close to showing perceptible improvement
in this review.
92
-------
------- |