xvEPA
United States
Environmental Protection
Agency
Environmental Monitoring
and Support Laboratory
PO Box 15027
Las Vegas NV 89114
EPA-600/4-78-053
September 1978
Research and Development
Environmental
Monitoring Series
Carbon Monoxide
Monitoring Network
Design Methodology
Application in the
Las Vegas Valley
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad categories
were established to facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously planned to foster
technology transfer and a maximum interface in related fields. The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.This series
describes research conducted to develop new or improved methods and instrumentation
for the identification and quantification of environmental pollutants at the lowest
conceivably significant concentrations. It also includes studies to determine the ambient
concentrations of pollutants in the environment and/or the variance of pollutants as a
function of time or meteorological factors.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161
-------
EPA-600/4-78-053
September 1978
CARBON MONOXIDE MONITORING NETWORK DESIGN METHODOLOGY
Application in the Las Vegas Valley
James L. McElroy, Joseph V. Behar, Leslie M. Dunn,
Pong N. Lem, Ann M. Pitchford, Nancy T. Fisher
Environmental Monitoring and Support Laboratory
Las Vegas, Nevada 89114
and
Mei-Kao Liu, Terry N. Jerskey, James P. Meyer,
Jody Ames, Gary Lundberg
Systems Applications, Incorporated
950 Northgate Drive
San Rafael, California 94903
Contract No. 68-03-2399
Project Officer
Edward A. Schuck
Monitoring Systems Research and Development Division
Environmental Monitoring and Support Laboratory
Las Vegas, Nevada 89114
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and
Support Laboratory-Las Vegas, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
ii
-------
FOREWORD
Protection of the environment requires effective regulatory actions
which are based on sound technical and scientific information. This informa-
tion must include the quantitative description and linking of pollutant
sources, transport mechanisms, interactions, and resulting effects on man and
his environment. Because of the complexities involved, assessment of specific
pollutants in the environment requires a total systems approach which tran-
scends the media of air, water, and land. The Environmental Monitoring and
Support Laboratory-Las Vegas contributes to the formation and enhancement of
a sound monitoring data base for exposure assessment through programs de-
signed to:
develop and optimize systems and strategies for monitoring
pollutants and their impact on the environment
demonstrate new monitoring systems and technologies by
applying them to fulfill special monitoring needs of the
Agency's operating programs.
This report discusses the application of a method for the design of a
carbon monoxide monitoring network in the Las Vegas Valley. The procedures
for applying the design methodology are given in sufficient detail to guide
regional or local pollution control agencies who may have a need to plan new,
or modify existing, air quality monitoring networks. The Monitoring Systems
Design and Analysis Staff at the EMSL-LV may be contacted for further infor-
mation on this subject.
George B. Morgan
Director
Environmental Monitoring and Support Laboratory
Las Vegas
iii
-------
ABSTRACT
An objective methodology that uses aerometric data and a physically
based air quality simulation model was proposed in a previous report for the
optimal siting of air pollutant monitoring stations in urban areas. This
report describes the continuation of that workthe application of the pro-
posed methodology to the urban Las Vegas area.
The first part of this report contains an examination of the validity
of the Atmospheric Pollution Simulation Model, a key component of the pro-
posed methodology. It also describes an intensive field measurement program
conducted to provide the necessary data base. The second part describes the
selection of meteorological scenarios associated with high pollution poten-
tial in the Las Vegas Valley and presents the results of the application of
the siting methodology.
One of the principal features of this methodology is the concept of a
Figure of Merit for general air quality monitoring. The Figure of Merit
represents an average pollutant concentration at each grid point as weighted
by the frequency of occurrence of meteorological scenarios.
iv
-------
CONTENTS
Foreword ill
Abstract iv
Figures vi
Tables vii
Abbreviations viii
Acknowledgement ix
I. Introduction 1
II. Summary 2
III. Outline of Siting Methodology 4
IV. Validation of the APSM in the Las Vegas Valley 7
Objective
Summary of the APSM
Field Measurement Program
Las Vegas Valley
Historical Information
Sampling Rationale and Plan
Simulation Model Input Data
Modeling Region
Emissions Inventory for CO
Aerometric and Other Data
Comparisons of Predictions and Measurements
Model Validation Summary and Conclusions
V. Application of Siting Methodology to the Las Vegas Valley 38
Overview
Meteorological Scenarios Pertaining to High CO
Concentrations
Exercise of Siting Methodology
Results and Discussion
VI. Concluding Remarks 50
References 51
Appendices 52
A. Field Program Instrumentation 54
B. Emissions Inventory for the Las Vegas Valley 60
C. Isopleth Maps of Simulation Days 73
-------
FIGURES
Number Page
1 Plan for demonstrating the siting methodology 5
2 Map of the Las Vegas Valley area 11
3 Carbon monoxide measurement sites 15
4 Meteorological measurement sites 16
5 Weekly variation of the average daily maximum CO at
selected stations 18
6 Location of virtual wind stations relative to the actual
stations in the modeling grid 21
7 Diagram of predicted vs. measured CO concentration in
the Las Vegas Valley on January 21, 1976 25
8 Diagram of predicted vs. measured CO concentration in
the Las Vegas Valley on December 3, 1975 26
9 Predicted and measured CO concentrations at selected
sites near Las Vegas on December 3, 1975 28
10 Predicted and measured CO concentrations at selected
sites near Las Vegas on December 4, 1975 29
11 Predicted and measured CO concentrations at selected
sites near Las Vegas on January 14, 1976 30
12 Predicted and measured CO concentrations at selected
sites near Las Vegas on January 16, 1976 31
13 Predicted and measured CO concentrations at selected
sites near Las Vegas on January 21, 1976 32
14 Predicted and measured CO concentrations at selected
sites near Las Vegas on January 22, 1976 33
15 Wind data for 1600 PST, January 16, 1976 34
16 Isopleths of Figures of Merit based on maximum 1-hour
average CO concentrations in the Las Vegas Valley 45
17 Locations of CO measurement sites and those proposed on
the basis of maximum 1-hour average CO concentrations 46
18 Isopleths of Figures of Merit based on morning 8-hour
(0500 to 1200 LSI) average CO concentration in the
Las Vegas Valley 48
19 Isopleths of Figures of Merit based on evening 8-hour
(1200 to 1900 LSI) average CO concentration in the
Las Vegas Valley 49
A-l Helicopter spiral sites in the Las Vegas Valley cc
B-l Emissions inventory grid definition ,-,
B-2 Relationship between grid systems for data used in the
present study go
C-l through C-24 74-97
vi
-------
TABLES
Number Page
1 Data Required as Input to the Model for the Las Vegas
Valley Validation Study 12
2 Roughness Parameter for Land-Use Categories in the Las
Vegas Valley 23
3 Statistical Comparison of Predicted and Measured CO
Concentrations 24
4 Classification of Meteorological Scenarios 41
5 Meteorological Scenarios Selected for the Las Vegas Valley 42
A-l Data Collecting Sites 57
A-2 Land-Use Characteristics in the Vicinity of the CO Monitoring
Sites 58
B-l Total Carbon Monoxide Emissions (kg/h) for the Las Vegas
Valley by Source Type for Constant Reference Temperature
of 10ฐC 70
vii
-------
ABBREVIATIONS
APSM
CCHD
CO
DIF
EPA
h
HDV
kg
km
LDV
LST
LTO
m
tub
m/s
MRI
MSL
NAAQS
NBS
NDH
NLV
NOAA
NSO
NTS
NWS
ppm
s
SAI
UNLV
UTM
atmospheric pollution simulation model
Clark County Health District
carbon monoxide
dual isotope fluorescence
U.S. Environmental Protection Agency
hour
heavy duty vehicle(s)
kilogram(s)
kilometer(s)
light duty vehicle(s)
local standard time
landing and takeoff
meter(s)
millibar
meters per second
Meteorology Research, Incorporated
mean sea level
National Ambient Air Quality Standards
National Bureau of Standards
Nevada Department of Highways
North Las Vegas
National Oceanographic and Atmospheric Administration
Nuclear Support Office
Nevada Test Site
National Weather Service
parts per million
second(s)
Systems Applications, Incorporated
University of Nevada, Las Vegas
Universal Transverse Mercator
NOTE: Certain symbols used in this report bear different connotations or
descriptions. These particular symbols, however, are part of an
equation, table, or figure and are defined specifically to that usage.
viii
-------
ACKNOWLEDGEMENT
A number of Federal, State, and local governmental agencies provided
invaluable assistance in the initiation and execution of this project. The
Nuclear Support Office of the National Weather Service in Las Vegas provided
meteorological equipment and training; Nellis Air Force Base provided meteo-
rological data and emissions data for the Base; current meteorological data
and climatological summaries of the Las Vegas Valley were collected and
assembled by the National Weather Service; the General Services Administra-
tion provided space for a field measurement station; and the U.S. Environ-
mental Protection Agency's Region IX assembled the initial emissions inven-
tory for the Las Vegas Valley. The Nevada State Department of Highways par-
ticipated in the field measurement program by providing several instrumented
sampling sites and furnished updated data on emissions in the form of traffic
summaries. The Air Pollution Control Division of the Clark County Health
District (CCHD) provided field data and operated several sampling sites. The
Clark County Electrical Division supported the field maintenance of meteoro-
logical stations through loan of equipment. The City of Las Vegas and the
City of North Las Vegas permitted use of existing streetlight poles as
structures for locating meteorological equipment.
The cooperation of a number of private organizations is acknowledged for
allowing equipment to be located on their property. These include the:
Central Telephone Company, Desert Inn Country Club, KBMI Radio Station, Hughes
North Las Vegas Terminal, Rob's Motel, and Sky Harbor Airport.
Private corporations that supplied information pertinent to emissions
include Nevada Power Company, Southwest Gas Corporation, Hughes Executive
Terminal, Hughes North Las Vegas Terminal, and Union Pacific Railroad.
Key personnel in all the aforementioned organizations were most helpful
in providing the very necessary support to the field study conducted in the
Las Vegas Valley.
Drs. C. Shepard Burton and Phillip M. Roth of Systems Applications, In-
corporated (SAI), provided many helpful comments on this work.
ix
-------
I INTRODUCTION
The monitoring of ambient air quality is an indispensable element of air
pollution studies. Field measurements, if gathered in an adequate manner,
probably provide the most valuable information on a number of environmental
issues, such as the identification of air pollutants of concern, the deter-
mination of their concentrations, and the search for their concentration pat-
terns and trends (Morgan et al., 1970). It is, therefore, not surprising
that the Clean Air Act, as amended in 1970, required ambient air monitoring
programs as part of the State Implementation Plans.
The establishment of a monitoring network is costly. Since air pollu-
tion in an urban area is the end product of many complicated circumstances
and events with noticeable spatial and temporal variability, pollutant con-
centrations generally vary significantly in time and space. Thus, the selec-
tion of monitoring sites is one of the most crucial problems in the design
of an air quality monitoring network. Designing a network that fulfills its
intended goal at the lowest cost (i.e., with the fewest monitors) is a com-
plex and difficult task. Until very recently, however, this problem had not
been fully addressed. The selection of monitoring sites was usually based
on either subjective judgment or nontechnical considerations such as conveni-
ence and accessibility.
The first in this series of reports (Liu et al., 1977) discussed various
aspects of concern in the siting of air quality monitors, and proposed a
methodology for a rational design of a network for carbon monoxide (CO). Con-
ceptually, the methodology presented is generally valid for any airborne
pollutant. Carbon monoxide was chosen because it is a relatively inert pol-
lutant for which the methodology should be presentable in its basic, simplest
form.
This report describes the application of the siting methodology to the
Las Vegas Valley area of Nevada. A brief review of the methodology is in-
cluded from the theory detailed in the previous report to make this document
complete in itself. The validity of a mesoscale air quality simulation model,
a key component of the proposed methodology, is tested using field data col-
lected in the Las Vegas Valley. The actual application of the siting method-
ology is then described. Proposed changes or additions to the methodology
to increase its versatility are also discussed.
-------
II SUMMARY
This report details the application of the specific siting methodology
for CO to the Las Vegas Valley area of Nevada, Demonstration of the proposed
approach consists of two major tasks. First, the establishment of the valid-
ity of the air quality simulation model selected for this study. The second
major task consists of several subtasks:
Examination of climatological information in order to characterize
meteorological patterns potentially leading to high pollutant concentrations.
This includes calculation of the frequency of occurrence of such patterns.
Exercise of the air quality simulation model for each of the meteoro-
logical patterns identified to generate the corresponding pollutant distri-
bution pattern.
Selection of monitoring sites based on these pollutant distribution
patterns and predetermined selection criteria (in the present case, those
sites most representative of local maximum concentrations).
Utilizing data gathered in the field measurement program, the air pollu-
tion simulation model was exercised for 6 days during the winter of 1975-1976.
The predictions for these 6 days were compared with field measurements in
order to assess the validity of the model. Spatial and temporal distributions
of pollutant concentrations as well as point-by-point comparisons of predic-
tions and measurements were considered in judging the validity of model pre-
dictions.
The model predicted diurnal trends well but often failed to predict the
absolute values of peak concentrations especially in the downtown and Las
Vegas Wash areas. These locations are where the highest value afternoon
traffic peak is experienced (downtown) and at the lowest point (topographi-
cally) in the Valley (Las Vegas Wash). The weakness in the predictions for
these areas implies unresolved microscale effects. Diagrams of comparison
data show a tendency to underpredict at high CO values but most of the corre-
lation coefficients were acceptable for hourly comparisons (in the range of
0.7 to 0.9).
One further observation on the model's behavior was in showing CO stations
to be located where a strong gradient of predicted CO concentrations often
occurred. This suggests that slight uncertainties in the wind field specifi-
cation could greatly affect the comparisons.
Limited sensitivity analysis in this study and those conducted by others
-------
including parametric analysis stress the influence of wind field and emissions
uncertainties and unresolved microscale effects on the results. These diffi-
culties can be resolved and are part of ongoing research.
In this study, the desirability of placing an air quality monitor at a
given location is measured by a Figure of Merit. This measure is defined as
the product of an air quality index at a particular location and the associ-
ated frequency or probability of occurrence of prevailing (or specified)
meteorological conditions at that location. Development of the meteorological
scenarios can be effected using different objective analysis techniques, de-
pending on the availability of data. In this exercise, the scenarios were
developed directly from atmospheric mixing depth and wind speed data for the
dates experiencing the lowest 20 percent of atmospheric ventilation rates.
For this purpose, a two-way contingency table involving atmospheric mixing
depth and wind speed was devised using this subset of the 5-year data set for
the local CO season. These data were further grouped according to wind speed
alone. Existing wind field data collected during an intensive measurement
date provided the balance of the information for the meteorological scenarios
as employed in the air quality simulation model. The predicted hour-by-hour
CO concentrations for each grid point were then used as indices to compute
Figures of Merit. Isolated high values of the Figure of Merit were ranked to
select potential monitoring locations.
A different set of Figures of Merit was calculated based on 8-hour aver-
ages of CO concentrations for the morning and evening periods. These periods
were 8-hour averages centered around the morning and afternoon peak traffic
hours, respectively. The locations of the highest nine values for each set
of 8-hour averages and the worst 1-hour average indicated that the expected
high values of CO concentrations do vary in location during the day. The
selection of the particular locations for monitoring suggested by the Figures
of Merit depends, therefore, on which emission patterns are most desirable to
be measured.
Current research is centered on incorporating a procedure for selecting
the number of sites in the design methodology in terms of either cost/benefit
for a unit of monitoring information or acceptable error concepts.
-------
Ill OUTLINE OF SITING METHODOLOGY
The design of an air quality monitoring network involves a number of
multidiscipline considerations, ranging from policy decisions on the need and
cost of the network to technical decisions on the selection of sites, the
choice of measuring instruments, and the operation and maintenance of the
network. The present project is concerned with the most crucial step in the
design of an air quality monitoring networkthe actual siting of the measure-
ment stations.
Rational design procedures for selecting sampling sites cannot be formu-
lated without a set of well-defined goals or objectives for the monitoring
program. After they are determined, a number of criteria for evaluating the
relative merits of different network configurations can then be clearly
stated. For example, if the objective of the proposed monitoring program is
to detect the long-term trends of air pollution, the proper criterion for
site selection is that the pollutant concentrations at the chosen sites be
most sensitive to changes in source emission strengths. If the objective as
described herein is to assess whether existing concentrations exceed an air
quality standard, the sites should be chosen so that the measurements are
most representative of the local maximum concentrations. These monitoring
goals not only lead to different selection criteria but, also, affect the
nature of siting methodology to be used. In some cases, a rigorous mathemat-
ical optimization technique can be implemented to obtain the best results
(e.g., Seinfeld, 1972; Darby et al., 1974); in other cases, a set of simple
heuristic decision rules may be the appropriate tool.
It is well known that pollutant concentrations are highly variable in
both time and space. The concentrations at a given receptor site are a func-
tion of both emissions (locations, strengths, temporal variations, etc.) and
atmospheric conditions (wind distribution, turbulent diffusion, vertical tem-
perature structure, etc.). The optimal placement of monitoring sites requires
a priori knowledge about the pollutant distributions under a variety of con-
ditions. Therefore, the key component in any logical siting methodology is
a predictive scheme, such as an air quality model, that links emissions from
sources with the observed air quality (Behar et al., 1976).
Through active research and development in the last decade, many air
quality models have been developed that are applicable to both inert and
photochemical air pollutants in urban areas. They range from simple approaches
such as the nonlinear rollback model (Schuck and Papetti, 1973) to complex
numerical models such as the Atmospheric Pollution Simulation Model (APSM)*
*Also known as Atmospheric Pollution Simulation Program (APSP).
-------
(Reynolds et al., 1973, 1974; Roth et al., 1974). These models can provide
valuable assistance in the design of air quality monitoring networks.
An earlier report (Liu et al., 1977) described the approach for the
selection of optimal monitoring sites. As shown in Figure 1, a demonstration
of the approach proposed consists of two tasks. The first is an examination
of the validity of the air quality model. The purpose of this task is to
assess the ability of the air quality model employed to reproduce the pollu-
tant concentration distributions.
J 1
I
FIELD
MEASUREMENTS
!
k
f
1
AIR QUALITY
MODEL
_k
^
ASSESSMENT OF
1 MODEL
CAPABILITY
rmm
CLIMATOGICAL
INFORMATION
AIR QUALITY
MODEL
MONITORING
OBJECTIVES
POLLUTANT
CONCENTRATION
DISTRIBUTION
PATTERNS
SITING
CRITERIA
H
SELECTION OF
MONITORING SITES
Figure 1. Plan for demonstrating the siting methodology
The APSM mentioned earlier was selected for the present project primarily
because it can provide spatial distributions for reactive pollutants. After
the validity of the model is established, the second task can then be carried
out in the following steps:
Characterization of meteorological patterns that potentially lead to
high air pollutant concentrations. This involves the use of historical
records including climatological information for the region of interest.
Utilization of the air quality model for each of the meteorological
patterns to generate the corresponding pollutant distribution patterns.
-------
Selection of monitoring sites based on these pollutant distribution
patterns and predetermined selection criteria.
The above tasks were carried out for the Las Vegas Valley area of Nevada.
The Las Vegas Valley was selected because it is an isolated urban area where
it is expected that very little pollutant of interest would enter the model-
ing region from other areas, and thus, boundary conditions could easily be
determined. Furthermore, complex and costly logistical support was not
required because the field program was directed from the EPA's Environmental
Monitoring and Support Laboratory located in the valley.
The results presented are limited to the chemically inert species, CO,
and analysis is restricted to the mesoscale; i.e., the determination of the
general areas where the monitoring sites should be located. In the present
application, this area is a 1-kilometer (km) square. The extension of the
present work to photochemical air pollutants and the inclusion of a micro-
scale analysis for pinpointing the optimal location for a monitor within each
general area are subjects of current research.
-------
IV VALIDATION OF THE APSM IN THE LAS VEGAS VALLEY
One of the key components of the siting methodology is an air quality
simulation model that relates local emissions to ambient air quality. In the
present project, a 3-dimensional photochemical model (the APSM) was selected
for mesoscale analysis. The APSM was originally developed for use in the
Los Angeles, California, metropolitan area. As such, it contained some com-
ponents applicable only to that area. Consequently, as a first step in the
project, the APSM was modified by removing these site-specific components and
replacing them with ones considered more general in nature. The modifications
are described in Liu et al. (1977), Ames et al. (1978), and later in this
chapter.
The usefulness of the siting methodology depends upon the relative
accuracy with which the APSM predicts ground-level concentration distributions.
Therefore, a model validation study was included as part of this project. This
undertaking included a comprehensive field measurement program conducted in the
Las Vegas Valley as the urban area chosen for development of the siting method-
ology. Using the data collected in the field measurement program, the APSM
was exercised for 6 days and its predictions were compared with measurements.
A detailed description of this effort follows.
OBJECTIVE
The objective of a model validation study is to assess the ability of
the model to achieve the best possible comparison between model predictions
and measurements, given the limitations of the input aerometric and measured
air quality data. The accuracy of the model predictions depends on
the adequacy of the model formulation and solution procedures
the availability and accuracy of data required as input.
Because the quantity of input data which the APSM requires is not always
available, the modeler may have to make some assumptions or estimates about
the input data that are not available. Some uncertainty is introduced by
this procedure, and the modeler has the prerogative of adjusting his estimates
within the limits of physical reality. These limits are determined by the
range of values normally observed or the values of the input data that are
available. Methods of obtaining the required input from limited data range
from simple interpolation to complex solutions of the momentum and energy
equations for the atmosphere. The latter often require other data to obtain
a solution. The method chosen depends on the resources at the disposal of
the modeler and the data available.
-------
Differences between model predictions and measurements depend not only
on model accuracy but also on the accuracy and representativeness of the
measurements that are compared with the predictions. Representativeness in
this context refers to how well the data represent the conditions to which
the model predictions apply. Representativeness is of particular importance
in this project because the concentrations predicted by the APSM are spati-
ally averaged over an area of 1 square kilometer.
This horizontal grid size was considered to be the smallest one consis-
tent with the computational requirements and capabilities of the particular
computer utilized, the areal extent of the modeling region, and the resolu-
tion of the pollutant emissions data.
It is important that monitoring sites not be located in the immediate
vicinity of sources in order that the measured concentrations be representa-
tive of the chosen spatial area. If there is reason to believe that data
from a particular site are not representative of the spatial average in the
area or, if there is some reason to suspect their accuracy, the data should
not be used for model validation.
Models need to be validated in a manner consistent with their intended
use (Simmons, 1974). Herein, the concern is with concentrations which exceed
ambient air quality standards. Thus, meteorological conditions conducive to
high pollutant concentrations were desired. The validation dates were chosen
with this general goal in mind.
There is no simple or single proven measure that can be used to judge
the overall goodness of an air quality simulation model (Olsson and Ring,
1974). Usually, a combination of graphical and statistical measures is used
for this purpose. In the present project the ability of the APSM to locate
"hot spots" or maxima in concentrations in reproducing a data base for the
site-selection process requires assessment.
Thus both the spatial and temporal distributions of concentrations are
considered in judging the validity of model predictions. Point-by-point
comparison of predictions and measurements is made. Also, the predictions
and measurements are examined collectively and comparison of concentration
distributions are made in addition to those on an individual basis.
SUMMARY OF THE APSM
The theoretical framework of the APSM is described in detail in the
previous report (Liu et al., 1977) and in the previously cited references.
The formulation of the APSM for the present application is given by equation
1. A transformation of coordinates was performed on the usual mass continuity
equation with the assumption that the slope of the land and the inversion
layer are negligible over the modeling region. This transformation aids in
the computational stability of the numerical solution to this equation.
-------
IT (AHci>+1? (uAHci)
c = concentration of the i-th species,
T = time coordinate,
,n = horizontal cartesian spatial coordinates,
p = vertical spatial coordinate = ... , , , N
H(x,y,t) - h(x,y)
z = height above mean sea level,
AH = H(x,y,t) - h(x,y),
H(x,y,t) = height of the elevated stable layer above mean sea level,
h(x,y) = height of the ground above mean sea level,
u,v,w = wind velocity components in the ฃ,nป and p directions,
/9h + 9AH\ _ /8h + 9AH \ 9AH
I 9ฃ 8E / \ 8ri 8f| / 8t
\ / \ '
K = horizontal eddy diffusivity,
IL = vertical eddy diffusivity.
Sj = rate of emission of species i from sources located be-
tween H and h. Terms in the generalized equation involving chemical reac-
tions and removal mechanisms have been eliminated for the present applica-
tion to CO. The solution of equation (1) requires as input, the wind field
(from measurements or a predictive scheme), height of the elevated stable
layer, and the vertical and horizontal eddy diffusivities over the entire
region and time to be simulated. Additional input data are required as a
result of the boundary conditions (Reynolds et al., 1973):
*V 8ฐi
P = 0 Q1(e,n,t) = - ^jj-gT- >
P-I
"v a0
-------
*ci
or ^ uct - Kg -jjฃ- = ugi if
<*>
i -ปป
T=- = ฐ if U-n > 0
n = f) or TI vc ~" ^Si "<\ = vg, if U*T) < 0
S N in OT\ i ~
8c
- K.. -r-^ =0 if tf-fi > 0
where
Q. = the mass flux of species i at the surface,
U = ui + vj,
i = unit vector in the e direction,
j = unit vector in the r\ direction,
n = the outwardly directed unit vector normal to the horizontal
boundary,
g. = the mean concentration of species i just outside the modeling
region.
Thus, the data requirements as a result of the boundary conditions are the
pollutant concentrations at both the sides and the top of the modeling region
and the flux of emissions into the bottom of the modeling region. The data
required as input to the APSM are summarized in Table 1.
FIELD MEASUREMENT PROGRAM
Data for the model validation effort were acquired through a field
measurement program. In the Las Vegas Valley, existing data collection
activities were reviewed and historical information examined to determine the
need for supplemental measurement sites. The availability of historical in-
formation and the extent of current monitoring activities influenced the
selection of meteorological and air quality monitoring sites.
Las Vegas Valley
The Las Vegas Valley represents a relatively isolated but mostly
urbanized area. It contains a desert community with a population of over
300,000. The valley, located in Clark County in southern Nevada, is bounded
by the Sheep Range and Las Vegas Range to the north, the Spring Mountains to
the west, Frenchman and Sunrise Mountains to the east, and the McCullough
Range to the south (Figure 2). These mountains, averaging about a kilometer
above the valley floor, impart a bowl shape to the valley with relief passes
10
-------
I I I I I I I I
KILOMETERS ,-,,
Figure 2. Map of the Las Vegas Valley area
11
-------
TABLE 1.
DATA REQUIRED AS INPUT TO THE MODEL FOR THE
LAS VEGAS VALLEY VALIDATION STUDY
Data Required
Spatial Resolution
and Characteristics
Temporal
Resolution
Emissions
Surface
Elevated
Air quality
Initial conditions
Boundary conditions
Winds
Elevated stable
layer heights
Horizontal eddy
diffusivity
Surface roughnesst
Atmospheric stabilityt
Surface cells*
Specified cells*
All cells (assumed vertically
invariant)
Top and upward side
boundary cells
All cells
Assumed spatially
invariant
Assumed constant
Surface cells
Assumed spatially invariant
Hourly
Hourly
Hourly
Hourly
Hourly
Assumed constant
Hourly
* "Cells" in Table 1 are all 1 km by 1 km.
t Required for vertical eddy diffusivity.
to the northwest, southwest and southeast. The floor of the valley slopes
gently from west to east about 980 meters (m) mean sea level (MSL) on the
west to about 550 m MSL on the east-southeast side. To the east of the Las
Vegas Wash, the terrain gently rises again.
The main economic base of tourism is supplemented by a substantial
trucking and warehousing industry, a large mineral and metal mining and
refining industry, and the operations of several large governmental organi-
zations. Because the urban area of Las Vegas consists of vacant desert
scattered among residential developments, the population is distributed over
a larger area than many other urban communities of equivalent population. To
serve this large area, a limited access interstate highway traverses the city
and major highways crisscross the valley providing access from Arizona to
northwestern Nevada, and from California to Utah. A large grid of four- and
six-lane arterial streets and an intermeshed network of secondary roadways
accommodate local traffic in the valley.
Historical Information
To plan the initial measurement program, a survey of available informa-
tion was conducted. Several Federal and State agencies have ongoing data
collection efforts in the Las Vegas Valley, in both meteorology and air
12
-------
quality.
Meteorological data are collected locally by the National Weather Service
(NWS) at McCarran International Airport and at the Nevada Test Site (NTS) and
by the U.S. Air Force at Nellis Air Force Base (see Figure 2). The NTS is
about 100 km north of Las Vegas. Near-surface weather information including
that on cloud cover, visibility, barometric pressure, air temperature and
moisture, precipitation, and wind speed and direction are provided at 1-hour
intervals at the airport location. Upper air data on temperature, moisture,
and wind speed and direction are collected at the NTS. Vertical soundings of
these parameters are made at 0400 and 1600 local standard time (LST). Prior
to 1966 the soundings were taken at McCarran International Airport. Data
from these and other sources are compiled at the National Oceanic and Atmos-
pheric Administration's (NOAA) National Climatic Center in Asheville, North
Carolina. In addition, the NWS operates a local Nuclear Support Office (NSO),
including coverage of the NTS, for the Department of Energy with weather fore-
casting capabilities and a library of publications pertinent to southern
Nevada. Much of the meteorological equipment used in the field program was on
on loan from the NSO under a cooperative agreement.
Aerometric data beginning in the fall of 1974 are available from the
State of Nevada Department of Highways (NDH), which operates two trailers in
the Las Vegas area equipped to monitor air quality and wind speed and direc-
tion. The Clark County District Health Department (CCHD) operates four air
quality measurement sites locally (data archived beginning in 1974). Arrange-
ments to share current and historical data were made with both agencies.
Sampling Rationale and Plan
Historical data indicated that CO, total suspended particulates, and
oxidants were the major criteria pollutants of concern in the valley. Be-
cause CO is relatively inert, this pollutant was chosen as the simpler case
to test the methodology for the design of an optimum monitoring network.
The period for full scale aerometric sampling for both the routine and
intensive programs was chosen as November 1975 through February 1976, inclu-
sive. Limited historical CO data collected by the CCHD indicate that the
highest daily and shorter term concentrations and frequency of exceeding the
National Ambient Air Quality Standards (NAAQS) for CO all occur during this
period of the year. Historical weather data for the Las Vegas Valley show
that the meteorological conditions conducive to high concentrations of CO
also occur most frequently during this period of the year.
Climatologically, the period of minimum dilution in the Las Vegas Valley
for pollutants such as CO, which are emitted primarily from near-ground-level
sources, is that of late fall and early winter. Both the atmospheric mixing
depth (Holzworth, 1964) and the average wind speed through the depth
(Holzworth, 1962) are generally smallest during this time of the year. Thus,
the dilution or ventilation rate, which is comprised of the product of these
parameters, follows a similar pattern. The meteorological potential for air
pollution episodes is also highest then. Holzworth (1974), for example,
observed that the episode of slowest dilution and the episode of
13
-------
longest duration for a 5-year data sample both occurred locally in December.
An episode in this analysis was defined as a period of 5 or more consecutive
days of high meteorological air pollution potential, with the potential being
determined in terms of mixing depth, wind speed through the depth, and the
details of the dominant synoptic scale weather features.
As was discussed in the previous report (Liu et al., 1977), the Las
Vegas Valley is situated in semi-arid terrain and experiences a desert-type
climate. This is typified by clear skies, a large diurnal near-surface tem-
perature change, and a strong nocturnal surface-based temperature inversion,
especially during times experiencing light wind speeds. Periods of minimal
pollutant dispersal are associated with the occurrence of the nocturnal in-
version which dissipates most slowly during the first few hours after sunrise
and forms around sunset. As shown elsewhere in this report, motor vehicles
generate nearly all the emissions of CO. The times of peak activity are
0630-0830 and 1600-1800 LST. Thus, the times of peak emission coincide
closely with diurnal segments of minimum pollutant dispersal during the
selected sampling period, particularly since sunrise occurs later and sunset
earlier than during the rest of the year (see Liu et al., 1977). Consequently,
highest short-term concentrations (on the order of a few hours) can also be
expected to occur more frequently during this period.
Near-surface sites for routine continuous monitoring of aerometric vari-
ables were chosen subjectively on the basis of predominant synoptic scale and
locally induced wind patterns (e.g., upslope and downslope winds), topography,
land use, and road traffic information. Because of cooperative arrangements
with local agencies monitoring in the valley, only new sites required to sup-
plement those existing were maintained by the U.S. Environmental Protection
Agency (EPA). Measurements of CO were made at nine stations (Figure 3) .
Those of wind speed and direction and air temperature were made at 13 sites
and of air temperature alone at another four sites (Figure 4). A special
effort was made to locate CO monitoring sites away from the immediate vicinity
of major emission sources such as roadways.
Special additional sampling of aerometric parameters was accomplished
in intensive periods. During these periods, air temperature, moisture, and
pollutant data in the vertical were collected with an instrumented helicopter,
and measurements of winds aloft were made. The periods began an hour before
sunrise and normally continued until 2 to 3 hours after sunset. This time
frame was chosen in order to cover both the morning and evening peak emission
periods of CO.
Periods for short-term intensive sampling were chosen on the basis of
forecasts of synoptic scale weather patterns over the Las Vegas Valley. Days
with particularly limited atmospheric dilution were desired. Such situations
generally occur when the valley is under the direct influence of a stagnant
or slowly moving high pressure area. Weekends were not considered because an
adequate CO emissions inventory could not be established for this period from
available data.
To insure that this constraint did not invalidate the monitoring systems
design, average daily maximum CO concentrations for several years of record
14
-------
60
50
12345678
I I II 1111
Figure 3. Carbon monoxide measurement sites
15
-------
Charleston Blvd.
-24
^L McCarran
Surface Wind Sites
Pibal Sites
Hygrothermograph
Sites
12345678
1 1 I I 1 I I
Figure 4. Meteorological measurement sites
16
-------
were plotted. Figure 5 shows these data as a function of day of the week.
The plots at these three central city locations generally show slightly higher
values during weekdays than during weekends. This fact indicates that the
monitoring systems design is not unduly constrained by the absence of weekend
emissions data.
Weather forecasts for intensive sampling were made using facilities
available at NWS's NSO in Las Vegas where observed surface and upper air data
for the contiguous United States are routinely collected by teletype. Fac-
simile charts containing observed or forecast weather data for ground-level
and various constant-pressure surfaces in the atmosphere are received which
cover the United States plus portions of adjacent oceans. In addition,
special forecast information is available, generally resulting from simula-
tions with synoptic scale meteorological models.
Specific forecasts of weather events for the local area were normally
made for periods up to only 24 hours in advance as they could usually be
expected to provide reasonably accurate information on the details and timing
of such events. As a result, only 1 day of notice was usually provided for
intensive sampling periods. Extended weather outlooks during favorable sam-
pling situations for periods up to 48 and 72 hours in advance were also pre-
pared when appropriate. On the basis of such outlooks, field personnel were
placed on standby status.
Details of instrumentation used during routine and intensive sampling,
including data collection and reduction and quality assurance procedures
established for the field measurement program, are contained in Appendix A.
Numeration of the sites shown in Figures 3 and 4 is established in Table A-l
of this Appendix.
SIMULATION MODEL INPUT DATA
Modeling Region
As discussed earlier, the Las Vegas Valley represents a relatively iso-
lated urbanized area. It was assumed that boundary conditions could be deter-
mined more easily because of this relative isolation. That is, one could
reasonably expect that very little pollutant of interest would enter the
modeling region from other areas. Consequently, the modeling region was
defined as a 48-km by 70-km area extending to the ridgelines surrounding the
Valley. The region was divided into 1-km by 1-km squares commensurate with
the resolution of pollutant emission data and of the APSM.
It is recognized that the scale of natural variability of CO in the
immediate vicinity of major emission sources like roadways is of the order
of tens of meters. Microscale considerations should be applied to ensure
compatibility between this scale and the resolution of the APSM. However, it
was decided to use the results of model simulations unmodified by microscale
effects for the purpose of demonstrating the basic utility of the design
methodology. It should be noted, as discussed earlier, that detailed CO
emission data on a scale*less than a 1-km by 1-km square were not available
locally and that an effort was made to avoid locating CO monitoring sites in
17
-------
CXI
SUN
MON
TUE WED THU
DAY OF THE WEEK
FRI
SAT
Figure 5. Weekly variation of the average daily maximum CO at selected stations
-------
close proximity to major sources.
Emissions Inventory for CO
Emissions data were inventoried within the modeling region as divided
into 1-km by 1-km squares. Emissions from point sources (electric power
generation plants and industrial plants), area sources (space heating), and
mobile sources (railroads, aircraft, and road vehicles) were determined for
each hour of the day within each such grid square. Light duty vehicle (LDV)
emissions have been shown to depend on ambient temperatures. These emissions
account for 80 percent of the total inventory on a daily basis and must be
adjusted to reflect different ambient temperature conditions. Thus, an emis-
sions inventory was developed specifically for each day simulated.
The emissions inventory was developed from information provided by the
CCHD, NDH, local industries and utilities, the National Environmental Data
System (NEDS) data bank using AP-42 (EPA, 1976) as a guide. Details of
methodology including procedures and equations are described in Appendix B.
Aeroroetric and Other Data
Preparation and use of available aerometric data and other data required
as input to the APSM for the Las Vegas validation study are discussed later.
Air Quality Data
Carbon monoxide measurements made at nine sites were previously shown
with respect to the modeling grid in Figure 3. This figure illustrates that
most of the monitoring sites are located near the center of the modeling
region. As a result, there is a high density of data available for validat-
ing the model near the center of the region and very little or no data avail-
able on the periphery. This situation would ordinarily make it difficult to
specify the initial and boundary conditions. However, because of the almost
total absence of sources outside the region containing the majority of the
stations, it can reasonably be assumed that the initial concentrations in the
unmonitored region approach the background levels at large distances from the
monitoring sites. Furthermore, the concentrations on the boundaries of the
modeling grid can also be safety assumed to be at the background levels. No
measurements aloft were available but, based on our knowledge of the vertical
distribution of CO at other locales, we assumed that concentrations entrained
as the nocturnal inversion rises are also at background levels.
The measurements collected at the monitoring sites are needed to specify
the initial pollutant distribution. In order to provide concentrations at
each grid point, interpolation is necessary. A two-step procedure was used
in this study. First, an automated scheme based on the reciprocal-of-the-
distance rule (Liu et al., 1973) was applied using the measured CO data. The
resultant CO concentration fields were subsequently smoothed manually. This
latter procedure was used because unrealistically high fluctuations in CO
concentrations were obtained during early hours of simulations with only the
automated technique.
19
-------
Surface Wind Speed and Direction
Information on the 3-dimensional wind field for each of the grid cells
in the modeling region is required at hourly intervals. The computer algo-
rithm described in Chapter III-A of the previous report (Liu et al., 1977)
provides these data from the values of horizontal wind speed and direction
reported for the grid locations of each of the 13 near-surface wind stations
in the Las Vegas field network.
The objective interpolation scheme to develop the near-surface horizontal
wind field was modified slightly in order to facilitate application of the
algorithm for the local area. The existing wind stations shown with respect
to the modeling grid in Figure 6 are grouped together in the central portion
of the modeling region, i.e., away from severe terrain. As a result, unreal-
istic irregular features in the near-surface wind field were often noted near
the boundaries when the objective analysis scheme was applied to the Las Vegas
data. Through experimentation, it was found that a wind field that retained
the basic features of the depicted wind patterns and that eliminated these
irregularities could be obtained when nine virtual stations on the periphery
of the modeling region were added to the existing data set. A virtual sta-
tion is not a real measuring site,but a location near the periphery of the
modeling region where a value of the wind velocity vector is assigned by
vector addition of the velocity vectors of its two nearest actual measuring
stations. Specifically, virtual stations were first added to the corners and
then to the central portions of the grid edges until the continued addition
of stations provided no apparent improvement in smoothness to the wind field
near the periphery. These stations are also shown in Figure 6 relative to
the existing stations in the modeling grid.
Mixing Depth
The thickness of the atmospheric mixing layer, or mixing depth, at hourly
intervals was obtained through analysis of vertical temperature profiles
obtained during the helicopter spirals. Wherever possible, data collected by
temperature sensors located at the fixed near-surface stations were used to
provide ground truth for the helicopter temperature profiles. Additionally,
shear information from winds^aloft measurements were used where applicable to
assist in establishing the thickness of the mixing layer; often a large shear
of horizontal wind speed or direction occurred at the top of this layer.
Spatial homogeneity in the thickness of the mixing layer was assumed to
exist across the modeling region in the Las Vegas Valley. On occasion, par-
ticularly at certain times of the day, slight spatial differences in mixing
depth were indicated from the helicopter temperature profiles. The indicated
differences were generally close to the resolution limit of the helicopter
data.
Two general modifications to each of the reported data sets were made in
order to facilitate operation of the APSM. First, a minimum constant value
of the mixing depth of 5 m was specified for the nocturnal hours. Actually,
a surface-based radiation temperature inversion was present in the valley
during this interval for each of the reported data sets. The minimum
20
-------
Figure 6. Location of virtual wind stations relative to the actual stations
in the modeling grid. Station numbers preceded by the letter F
are virtual stations.
21
-------
reference or mixing depth roughly corresponds to the thickness of the mixing
zone in the wake of a moving road vehicle (Ranzieri and Ward, 1975). Road
vehicles produce more than 90 percent of CO emissions on a daily basis in the
local area.
The second modification was concerned with the handling of the transition
period from the occurrence of the maximum mixing depth to the reestablishment
of the surface-based inversion near sunset. The field measurements indicated
that this transition is accomplished in about an hour. However, each of the
data sets was altered by allowing this transition to occur gradually, begin-
ning 2 to 3 hours prior to sunrise. This was done to guarantee computational
stability of the numerical formulation of the APSM.
Vertical Eddy Diffusivity
Profiles of vertical eddy diffusivity for each grid volume in the Las
Vegas Valley are computed using a computer algorithm (see Chapter III, Liu
et al., 1977) from values of near-surface horizontal wind speed, roughness
parameter, height of reference level (i.e., mixing depth or elevated stable
layer), exposure class and cloud cover. Information on cloud cover is
obtained from hourly weather observations made at the NWS station located
at McCarran International Airport. Exposure class at night is specified in
terms of cloud cover. Daytime exposure classes require sun angle determina-
tions which are computed using local latitude, longitude, time of day, and
Julian date as input parameters.
Values of the surface roughness parameter (z ) for inclusion into the
diffusivity scheme are estimated using the bulk aerodynamic technique devised
by Lettau (1969). The technique considers the height, surface area, and dis-
tribution of surface roughness elements exposed to the wind and is formulated
as:
, h* x , s*
z = (x) (-rr) ,,,.
o 2 S* (6)
where h* is the average height of the roughness elements within a horizontal
grid area, s* is the silhouette area of the average element exposed to the
wind, and S* is the lot area. The lot area is the quotient of the total sur-
face area over which the elements are being considered to the number of
elements within the area.
Average values of z were computed for each of the 1-km x 1-km horizontal
grids of the model area in the valley. For this purpose, representative land-
use categories were established and percentages of them within each horizontal
grid element determined using aerial photographs, results of visual ground
surveys, street maps, and real estate maps and charts. Values of z calcu-
lated using the above formulation for the specific land-use categor?es estab-
lished are shown in Table 2. A composite value of z for each of the hori-
zontal grid elements was then derived by linear weighting of the percent
coverage of each land-use category existing within the element.
22
-------
TABLE 2. ROUGHNESS PARAMETER FOR LAND-USE CATEGORIES
IN THE LAS VEGAS VALLEY
Land-Use Category z (m
6
1. Undisturbed desert: shrubs 2 m high and 1 m wide,
spaced about 5 m apart 0.05
2. Disturbed desert: shrubs 1 m high and 0.5 m wide, spaced
about 3 m apart 0.01
3. Residential: new subdivisions; one-story structures
and few or small trees 0.40
4. Residential: older subdivisions; one-story structures
and mature trees 0.50
5. Commercial: one to three-story structures with small
to medium parking areas 0.30
6. Industrial: one-story structures with large parking
and open areas 0.10
7. Downtown: high rise structures 1.00
Variations in silhouette area and hence of z as a function of both wind
direction and season of the year were assumed to be negligible in the valley.
These are not considered to be severe restrictions since tree cover, parti-
cularly deciduous species, is quite sparse, agricultural activity is minimal,
and land-use tracts are relatively homogeneous.
COMPARISONS OF PREDICTIONS AND MEASUREMENTS
The APSM was exercised, utilizing the data base described above, for 6
days in the Las Vegas Valley. In this section, model predictions and measure-
ments are compared.
The relative magnitudes of predictions and measurements can be compared
by making plots of the two sets of data for each validation day independent
of the location and the time. Ideally, the regression line calculated for
such a plot would have a slope of one. When the measurements are plotted on
the ordinate and the predictions on the abscissa, if the slope is less than
one but greater than zero, then the predictions tend to be high relative to
the measurements. If the slope is greater than one, the predictions tend to
be low relative to the measurements.
Table 3 presents linear correlation coefficients and regression para-
meters for the 6 days of the validation study. Also listed are the number of
23
-------
data points used in the statistical analyses. It is apparent from the slopes
of the regression lines that, in general, the model tends to underpredict.
TABLE 3. STATISTICAL COMPARISON OF PREDICTED AND
MEASURED CO CONCENTRATIONS
Number of Correlation
Day Data Points Coefficient Slope Intercept
Dec.
Dec.
Jan.
Jan.
Jan.
Jan.
3,
4,
14
16
21
22
1975
1975
, 1976
, 1976
, 1976
, 1976
92
91
79
67
83
92
0.
0.
0.
0.
0.
0.
76
73
74
42
91
77
1.
0.
1.
0.
1.
0.
62
91
09
48
22
99
- 1.
0.
0.
0.
- 0.
- 0.
23
26
02
78
87
41
Figures 7 and 8 are diagrams of predictions and measurements for
January 21, 1976, and December 3, 1975, respectively. As shown in Table 3,
January 21, 1976, yields the best correlation between predictions and measure-
ments. The correlation for December 3, 1975, is reasonably high, but the
measurements tend to be much higher than the predictions. One interesting
feature of Figures 7 and 8 is the large number of points below the unit slope
and zero intercept line for measured concentrations below 5 parts per million
(ppm). In fact, the portion of the plot for January 21, 1976, below measured
concentrations of 10-ppm is fitted quite closely by the line of unit slope and
zero intercept.
For most of the validation days, and at most of the measurement sites,
the trends of the predicted concentrations follow closely the trends of the
measurements. As expected from the results in Table 3, the magnitudes of the
predicted minima are close to the minima of the measurements, and the pre-
dicted and measured maxima do not compare well. The temporal variations of
the predicted and measured CO concentrations are compared in Figures 9 through
14 for four of the measurement sites. The Arden and Nellis sites were chosen
to represent model predictions at the remote monitoring locations and the re-
maining sites reflect predictions and measurements at the sites under the
influence of the urban CO emissions.
The late afternoon peak in the measured CO concentration that occurred
on January 16, 1976 (see Figure 12), seems to reflect microscale effects on
measured concentrations which cannot be reproduced by the mesoscale model.
The Shadow Lane, Casino Center, and East Charleston monitoring sites are
separated by a maximum of 5 km yet on that day East Charleston had virtually
no afternoon concentration peak while concentrations at Shadow Lane and
24
-------
IB.OOr-
o.
a.
09
/
/
/
CM CO
Predicted Cone, (ppm)
Figure 7. Diagram of predicted vs. measured CO concentration in the
Las Vegas Valley on January 21, 1976. The solid line is the
regression line. The dashed line is of unit slope, zero
intercept.
25
-------
Predicted Cone, (ppm)
Figure 8. Diagram of predicted vs. measured CO concentration in the Las
Vegas Valley on December 3, 1975. The solid line is the regres-
sion line. The dashed line is of unit slope and zero intercept.
26
-------
Casino Center reached approximately 8 ppm and 14 ppm, respectively. The
model predicts an afternoon peak at all three stations. The reason for the
discrepancy between the measurements and predictions cannot be specified with
certainty. However, an explanation based on the wind direction and the loca-
tions of the monitoring stations relative to sources appears to be reasonable.
The stations at Shadow Lane and at Casino Center are both surrounded by areas'
of heavy traffic. The region to the east of the East Charleston site has a
lower source density, and the region to the west has a source density similar
to the areas surrounding the other two sites. During the afternoon of Janu-
ary 16, 1976, the wind at the two meteorological stations nearest to the East
Charleston site was out of the northeast. On the other 5 days studied, the
wind was blowing from the west at this time of day. Hence, it would appear
that the change of the wind direction may have had a more pronounced effect
on the CO concentration at East Charleston than at the other two sites.
Changing the mixing depth or the atmospheric stability inputs to the
model can change the magnitude of the concentrations as well as the temporal
dependence of the pollutant concentrations. However, because we assume that
the stability and the mixing depth are spatially uniform, any change in these
parameters will lead to a change in concentration of the same size at all
locations. It is clear in Figures 10 and 11 that the low concentrations pre-
dicted at the Shadow Lane station could not be improved by lowering the in-
version heights or increasing the stability without making the correspondence
between measured and predicted concentrations worse at the other stations.
This fact suggests that either the winds have not been adequately estimated
or there is some microscale phenomenon affecting the Shadow Lane CO concen-
trations in these two instances. The initial conditions are unlikely to be
the cause of the low prediction because the poor correspondence between mea-
surements and predictions also occurs in the afternoon. It should be noted
that a high pollution concentration gradient frequently occurs in the vicinity
of the Shadow Lane station. This situation could magnify the discrepancies
between prediction and measurements with small changes (or errors) in wind
direction.
Failure of a model like the APSM to simulate correctly very high mea-
sured values of CO often reflects either a microscale phenomenon not resolved
by the model or uncertainties in the emissions inventory or in the interpo-
lated wind field. An explicit example of uncertainties in the interpolated
wind field significantly affecting the location and magnitude of high CO
values is discussed below. No attempt in the current study was made to adjust
the emissions inventory or to treat microscale effects although such adjust-
ments would be expected to improve the validation results.
At 1600 LST on January 16, 1976, a large horizontal shear existed in the
surface wind field between sites L and Z and sites H and K as shown in Figure
15. As expected, this resulted in large spatial variations in the automated,
interpolated wind field in the vicinity of CO monitoring sites D, S, C, and E
and generally poor agreement between predicted and measured CO values at these
sites. The agreement was improved substantially when virtual stations in the
manner previously discussed were added between the affected wind measurement
sites prior to running the automated interpolation scheme.
27
-------
a
5-
Measured
Predicted
I
T T
12 14
Measured
Predicted
16 18 20
10-
10
12 14 16 18 20
40
20-
8 10 12 14 16 **I8 20
20^
Measured
Predicted
10
12
14
1*6
18
20
TIME
Figure 9. Predicted and measured CO concentrations at selected sites near
Las Vegas on December 3, 1975. a. Arden b. Shadow Lane c. East
Charleston d. Winterwood.
28
-------
8-1
6-
a 4.
Measured
Predicted
6 8 10 12 14 16 18 ' 2*0
16-1
Measured
Predicted
6 8 10 12 14 16 18 20
Measured
Predicted
6 8 10 12 14 16 18 20
d 8
Figure 10. Predicted and measured CO concentrations at selected sites near
Las Vegas on December 4, 1975. a. Arden b. Shadow Lane
c. East Charleston d. Winterwood.
29
-------
8-1
6-
a 4-
> Measured
Predicted
8 10 12 14 16 18
20
12-
4-
.ซ Measured
__ Predicted
Q. o.
Q. 6 ' B ' 10 " 12 14 ' 16 " 18 20
8
12
C 8-
4-
6 8 10 12 14 16 18 20
10
16 18
20
12 14
TIME
Figure 11. Predicted and measured CO concentrations at selected sites near
Las Vegas on January 14, 1976. a. Arden b. Shadow Lane
c. East Charleston d. Casino Center.
30
-------
a 4.
2ซ
> Measured
Predicted
6 8 10- 12 14 16 18 20
p**!iirii
ฃL 6 8 10 12 14 16 18 20
8 10 12 14 16 18 20
TIME
Figure 12. Predicted and measured CO concentrations at selected sites near
Las Vegas on January 16, 1976. a. Nellis b. Shadow Lane
c. East Charleston d. Casino Center.
31
-------
a
CL
Q.
O
o
C e
d 4
2-
Measured
Predicted
10 12 14 16 18 20
8 10 12 14 16 18 20
\
10
16 18 20
12 14
TIME
Figure 13. Predicted and measured CO concentrations at selected sites near
Las Vegas on January 21, 1976. a. Arden b. Shadow Lane
c. East Charleston d. Casino Center.
32
-------
8-
6-
12-
b 8
18-
C 12'
6.
16
12.
4-
> Measured
Predicted
i
> 8 10 ' 1*2
m
._.
14 16
Measured
Predicted
18 20
2
8: ฐ-
ฐ- <
7T 24^
o
'^"^ '^.
i 8 10
^- ^ /
12 14 16 18 20
. Measured
Predicted
6 8 10 12 14 16 18 20
6 8 10 12 14 16 18 20
TIME
Figure 14. Predicted and measured CO concentrations at selected sites near
Las Vegas on January 22, 1976. a. Arden b. Shadow Lane
c. East Charleston d. Winterwood.
33
-------
60
Figure 15. Wind data for 1600 PST, January 16, 1976.
34
-------
A detailed study of the APSM to determine the effect of uncertainties in
input parameters was not conducted as part of the present model validation
effort. Such uncertainties may be in the data themselves or in the interpo-
lation routines utilized to prepare the data for use with the model However
a study on this subject was conducted by Liu et al. (1976a), using an earlier '
version of the APSM with data appropriate for the Los Angeles metropolitan
area. In the study, wind speed, vertical eddy diffusivity, mixing depth, or
pollutant emission rate was varied within its expected range of uncertainty
with the other parameters kept at specified base values. The relative change
in CO concentrations associated with these variations were than determined
through model simulations for a specific data base in the area. The results
were evaluated in terms of deviations from areawide averages of CO concentra-
tions. Qualitatively and in a relative sense, the simulated CO values were
most sensitive to changes in wind speed, less sensitive to changes in mixing
depth and pollutant emission rates, and least sensitive to changes in the
vertical eddy diffusivity. For further details on the study, the paper by
Liu et al. (I976a) should be consulted.
The temporal variations of the CO concentrations shown in Figures 9
through 14, the diagrams in Figures 7 and 8, and the results of the statis-
tical analysis shown in Table 3 illustrate how well the predictions compare
with the measurements at the monitoring sites. However, to expect a highly
favorable comparison of the predictions with the measurements by these cri-
teria alone is overly optimistic, considering the uncertainties in the input
data, for the model. The criteria presented thus far for comparing predic-
tions with measurements do not take into consideration the possibility of
slight shifts in the spatial distribution of the predictions relative to the
measurements. Shifts might arise from errors in the input data (e.g., the
emissions distributions or the windfield). Hence, if an error of 0.5 km/h
in a wind vector (or perhaps an error in the direction) arose from the in-
terpolation procedure, it would not be unreasonable to expect that after
several hours of simulation a measurement might be in better agreement with
the prediction at a location one or more grids from the monitoring site than
at the site itself. Such shifts of the predictions can be detected only by
inspection of the predicted ground-level concentrations on the entire grid
system.
This inspection would be a formidable task because of the quantity of
output from the model, but a simplification of the format of the model output
made it less difficult. The predictions were converted into isopleth diagrams,
with lines of constant CO concentration printed out on the grid at intervals
of 2 ppm. Isopleths for predictions and the measurements at each of the
monitoring sites for each hour of each validation day are presented in
Appendix C.
A fair appraisal of model performance can be obtained by comparing the
temporal distributions of the predictions and the measurements (in Figures 9
to 14) with the corresponding spatial distributions for the hours of interest
(in Appendix C). From Figure 10, it appears that the model does poorly at
reproducing the magnitudes of the peaks in CO concentrations in the morning
and afternoon at Shadow Lane. The isopleth diagrams for December 4, 1975,
35
-------
show that the predicted concentration gradients in the vicinity of Shadow
Lane at the times of these two peaks are very sharp. At any hour, a slight
shift of the predicted concentration distribution to the west would result
in a very close match between predictions and measurements. The isopleth dia-
grams support similar observations in many other cases of substantial devia-
tions between predictions and measurements but generally show that the model
does well in predicting the general location of hot spots or local maxima.
The highest values for CO occur in the vicinity of the Las Vegas Wash
area. Since this is an area of convergence for nighttime drainage flow,
pollutants are expected to collect here, resulting in high values following
the evening traffic peak.
The measured CO concentrations on all days for the hours from 1700 to
2000 indicate very sharp concentration gradients between East Charleston and
Casino Center. These features of the concentration distributions cannot be
predicted by the model without more highly resolved meteorological data or
some knowledge of the subgrid-scale processes important in the vicinity of
East Charleston.
In general, when the predicted concentrations exceed 10 ppm there are
very sharp concentration gradients in the central city area around the
Shadow Lane, Casino Center, and East Charleston measurement stations. The
spatial scale over which concentrations change is much less than the distance
between the wind measurement stations. Hence, we expect the uncertainty in
interpolated wind fields to have a significant effect on the comparison be-
tween measured and predicted concentrations at these central city measurement
stations.
Finally, the validation results can be assessed in terms of the indivi-
dual monitoring stations with emphasis on the specific locations regarding
land use and emission sources. For example, the Winterwood and Henderson
stations lie in the Wash area where the model appears to perform well in pre-
dicting both diurnal trends and peaks (except for microscale influences).
The same is true for the East Charleston station which also lies in the
Wash but is near the downtown area. The model does well in predicting
diurnal trends but misses the magnitude of the peaks more so than in the
Wash area. Many of these peaks are definitely associated with microscale
effects due to location of station (as evidenced by observed wind directions).
Arden and Nellis are background stations where little diurnal variability
is observed in the measurements. The model depicts this well at Arden. At
Nellis it indicates some diurnal variability associated with apparent emis-
sions. It is likely that these are the result of the assumption of apportion-
ment of Nellis emissions from aircraft to several squares. These emissions
are probably not reflected in surface concentrations at outlying squares as
aircraft operations here are aloft. Also, emissions in the square for the
Nellis site are probably downwind of it but this model cannot account for
such microscale emissions effects.
The Shadow Lane and northwest residential stations are represented by
36
-------
residential/commercial or resldental land use. At these sites the model simu-
lates the diurnal trends but not the magnitude of the peaks. Both sites are
in an area of some heterogeneity in emissions which cannot be resolved well
by the model on the mesoscale.
MODEL VALIDATION SUMMARY AND CONCLUSIONS
The APSM was evaluated for CO for the Las Vegas Valley using data for
six dates which were basically chosen for their meteorological potential for
high pollutant concentrations. The results show that the model generally
predicted trends at stations and distributions, including the relative loca-
tion of mesoscale (1-km by 1-km resolution) of hot spots or local maxima well.
It did not perform as well in predicting the absolute values of these peaks,
especially in the downtown area. These discrepancies, though, can usually be
accounted for based on either uncertainties in input data or on microscale
phenomena. The hourly averaged point-by-point comparison between predictions
and measurements of the stations taken collectively yielded linear correlations
between 0.7 and 0.9 for five of the six validation dates.
Results of past model validation studies indicate that point-by-point
comparisons between pollutant predictions and measurements on an annual or
seasonal basis generally result in linear correlation coefficients between
about 0.6 and 0.9 (e.g., Koch and Thayer, 1972; Slater and Tikvart, 1974).
Limited results are available on hourly comparisons. Those for sulfur dioxide
have reported as yield coefficients between 0.2 and 0.6 (e.g., Koch and Thayer,
1972; Shirr and Shieh, 1974). Johnson et al. (1973) reported coefficients for
CO between 0.4 and 0.7 but utilized monitoring sites in the immediate vicinity
of roadways and included a microscale submodule for specific microscale effects.
Liu et al. (1976b) reported coefficients for CO in the range 0.6 and 0.8 for
models finely tuned to a specific metropolitan area.
Thus, the comparisons between predictions and measurements on a point-
by-point basis for the present study are at least as good as those presented
in the literature. In addition, as required by the project objective, it
generally predicted distributions well,including relative locations of peaks.
As a result, the APSM is considered at least adequate for its proposed task
in this projectthat of providing a data base for exercising the network
siting procedure in the Las Vegas Valley as a demonstration case.
37
-------
V APPLICATION OF SITING METHODOLOGY TO THE
LAS VEGAS VALLEY
As illustrated in Figure 1, the demonstration of the siting methodology
consists of two parts. The first part, the model validity study, was discussed
in Chapter IV. This chapter addresses the second partthe application of the
methodology for the selection of CO monitoring sites in the Las Vegas Valley.
OVERVIEW
A review of various considerations pertaining to the selection of air
quality monitoring sites was presented earlier (Liu et al., 1977). It was
emphasized that the formulation of an objective scheme for selecting monitor-
ing sites requires a clear statement of the goals or objectives of monitoring.
For the sampling of air contaminants, several possible objectives were listed:
Compliance with air quality standards.
Determination of long-term air pollution trends,
Enforcement of regulations on pollutant emissions.
Estimation of regional air pollutant fluxes.
Procurement of data for air quality model development or validation.
For Federal or regional air pollution control agencies, compliance with regu-
lations and standards is probably the most important objective for air quality
monitoring, and it is the principal objective in the design of the siting
methodology.
In the methodology, the desirability of placing an air quality monitor
at a given location is measured by a Figure of Merit. Quantitatively, the
Figure of Merit for a particular location is defined as the product of an air
quality index (either observed or expected) and the assbciated frequency or
probability of occurrence:
=2_, (Air Quality Index) (Probability of Occurrence).
(7)
The summation is to be performed over all meteorological scenarios that lead
to high air pollution concentrations. The Figure of Merit is weighted by the
frequencies of occurrence of scenarios because the pollutant concentration at
any location is a function of the prevailing meteorological conditions. Thus,
it varies significantly with time. Consequently, air quality information
related to a single event or period would not necessarily be the best index
38
-------
for the determination of a permanent or semipermanent site for a monitoring
station. By careful selection and characterization of the meteorological
scenarios, it may be possible to include the effects of variable meteorological
conditions in the selection of the optimum site by means of equation (7).
The air quality index in equation (7) may be chosen either as the concen-
tration of a specific pollutant for general air quality monitoring or, for the
case of detecting measurements which exceed an ambient air quality standard
(AAQS), as a delta function defined by
jl if the observed or expected concentration exceeds the AAQS,
if not. (8)
It is possible that in the first case, high values of the Figure of Merit
may be calculated for locations which never exceed the AAQS or some large
fraction of it. Such locations may be excluded from consideration as poten-
tial monitoring sites by truncation of the calculation for them. In the case
of locating a site for the measurement of many pollutants, the air quality
index can apparently be generalized using a composite index c*,
N
c* = ฃ w*c , (9)
1=1
where c is the observed or expected concentration of species i, and w* is the
weighting factor reflecting the importance of pollutant species i.
METEOROLOGICAL SCENARIOS PERTAINING TO HIGH CO CONCENTRATIONS
Selection of optimal locations for CO monitoring stations by the Figure
of Merit technique was accomplished using the results of simulations with the
APSM as a data base. The simulations were run on meteorological scenarios
developed from historical weather data supplemented by current aerometric in-
formation collected during the intensive and routine field sampling programs.
Analysis of historical CO data, especially that with concentrations from
slightly less than to greater than the NAAQS, in relation to the prevailing
meteorological situation or pattern,should provide definitive information
useful for scenario selection. Unfortunately, the historical aerometric data
base in the Las Vegas Valley is insufficient for this purpose. Consequently,
the meteorological pollution potential expressed as the ventilation rate dis-
cussed previously was substituted for the air quality data in the scenario
selection process. This was not considered to be a severe restriction since
the concentration of a relatively inert pollutant such as CO released primarily
at near ground-level should be proportional to the ventilation rate (McCormick,
1968).
Ventilation rates were calculated using upper air data collected at
McCarran International Airport. Specifically, 5 years of mixing depth and
wind speed data were used for the period between 1959 and 1964. These data
were available on magnetic tape from NOAA's National Climatic Center. Ven-
tilation rates were only determined for 1600 LST data since a nocturnal
surface-based temperature inversion usually existed locally at the time of
39
-------
the 0400 LST upper air sounding.
Initially, classification of meteorological situations into categories
was accomplished through statistical analysis of data available on 850-millibar
(mb) constant pressure charts. For this analysis, heights on the charts were
available at intervals of 1 degree latitude and longitude for the contiguous
United States and large portions of adjacent bodies of water. Six years of
the data encompassing the Western United States and portions of the Eastern
Pacific Ocean were utilized from a magnetic tape furnished by the NWS's NSO
in Las Vegas. Objective classification was accomplished in the manner out-
lined by Lund (1963) and Roach and McDonald (1975). Basically, this involves
establishing the linear correlation of heights on the constant pressure sur-
face for all pairs of grid points. The charts or maps correlating highest
with each other are grouped together to form classes. The percentage frequency
of each class determines its probability of occurrence.
The results of the study for the period of the local CO season were not
satisfactory. The classes chosen for the 850-mb constant pressure surface
charts were not usually relatable to distinct features present on ground-level
charts. In addition, a large percentage of charts from the data set could not
be placed into discernible classes. It is likely that better results might
be attained if ground-level data were utilized in the analysis. This expecta-
tion is supported by the fact that the transport and diffusion of ground-based
emissions such as CO on a scale of the Las Vegas Valley will be largely deter-
mined by processes in the planetary boundary layer; the layer rarely extends
to the 850-mb level except in mid-to-late afternoon. However, such ground-level
data were not readily available.
Consequently, meteorological situations were grouped into classes through
visual examination of historical ground-level weather charts. This was accom-
plished using data from 1959-1964 available on microfilm from the NSO. The
classification was done using the entire data set and a selected portion of
the data set for the local CO season. The selected data set consisted of the
dates experiencing the lowest 20 percent of the ventilation rates. This sub-
set of the data was assumed to encompass most of the situations for which the
NAAQS would actually, or nearly, be exceeded. In each case, the frequency of
occurrence of the chosen classes established the requisite probability of
occurrence.
Comparisons between the resulting synoptic meteorological patterns or
classes and the ventilation rates were subsequently made using discriminant
analysis. A computer algorithm of the discriminant analysis procedure applic-
able to this problem has been devised by Meyers (1971). A detailed discussion
of discriminant analysis is found in Hoel (1962) . As in the previous case,
the results of the comparisons were poor. That is, very few of them were
statistically significant at the standard 5 percent level. The results are
likely colored by the fact that the synoptic classes that evolved spanned a
very narrow range of conditions, i.e., largely consisting of high pressure
areas in various locations in relation to the Las Vegas Valley. In addition,
pollutant transport and diffusion are determined by the details of atmospheric
circulations which may not be readily divisible by a classification scheme
based fundamentally upon synoptic scale weather features.
40
-------
Because the above techniques yielded limited results, the scenarios were
developed directly from the mixing depth and wind speed data for the dates
experiencing the lowest 20 percent of the ventilation rates. For this purpose,
a two-way contingency table involving mixing depth and wind speed was devised
using this subset of the 5-year data set for the local CO season. The result-
ing table, also known as a test block, is presented as Table 4. Maximum mix-
ing depth (1600 LST value generally corresponds to the daily maximum value)
and average wind speed through this depth are, respectively, the ordinate and
abscissa in the Table. Statistics on the mean and standard deviation of data
values are provided in each block. The number of data sets (n) in each class
is also shown.
TABLE 4. CLASSIFICATION OF METEOROLOGICAL SCENARIOS
Wind Speed
(u = m/s) <200
calm
0.3-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
>9
*
*
u = 2
H = 185
n - 1
u = 2.1
H = 56
n = 1
u = 3.5
H = 163
n = 1
*
*
u = 6.4
H - 107
n = 1
*
*
*
200
u -v
H =
n =
u =
H =
n =
u =
H =
n ป
u =
H =
n =
u =
H =
n =
u =
H =
n =
u =
H =
n =
u =
H =
n =
- 400
0
394
1
0.84 + 0.23
288 + 44
8
1.60 + 0.31
342 + 55
11
2.7
320
1
3.5
266
1
4.7
355
1
5.4 + 0.1
309 + 31
2
*
*
9
220
1
*
Mixing Depth (H = m)
400 - 600 600 - 800 800 - 1000 >1000
u i/ 0 * * *
H = 590
n = 1
u = 0.82 + 0.21 u = 0.82 + 0.21 u = 0.88 + 0.25 u = 0.87 + 0.19
H = 507 + 61 H = 702 + 53 H = 848 + 55 H = 1237 + 182
n=26 n=25 n=8 n = 7
u = 1.68 + 0.52 * * *
H = 518 + 54
n = 27
* * * *
* * * *
* * * *
* * * *
* * * *
*
*
*
indicates n = 0
41
-------
The data in the test block were further grouped according to wind speed
alone: calm -3, 3-5, and 5 m/s. Since the second and third classes together
consist of less than 6 percent of the subset and, hence, only about 1 percent
of the total data set, they were excluded from further consideration. Data
in class one were then classified by mixing depth to form six scenarios, as
follows: 0-300, 301-450, 451-600, 601-800, 801-1,000 and >1,000 m. Representa-
tive example dates for each such scenario were then chosen from the data. This
information, along with that yielding the frequency of occurrence of each such
scenario for the data subset, is provided in Table 5.
TABLE 5. METEOROLOGICAL SCENARIOS SELECTED FOR THE LAS VEGAS
VALLEY
Scenario
1-1
1-2
1-3
1-4
1-5
1-6
II
III
Wind Speed
Range
(m/a)
0 < U <
0 < U <
0 < U <
0 < U <
0 < U <
0 < U <
3 < U <
5 < U
: 3
: 3
: 3
: 3
: 3
: 3
c 5
Mixing Depth Frequency of
Range Occurrence
300
450
600
800
1000
H
< H
< H
< H
< H
< H
< 300
< 450
< 600
< 800
< 1000
0.081
0.106
0.431
0.203
0.065
0.057
0.024
0.033
Representative
Data
December 1, 1964
November 24, 1964
November 5, 1962
December 13, 1960
November 3, 1961
November 28, 1961
January 27, 1961
January 9, 1962
EXERCISE OF SITING METHODOLOGY
For each given meteorological scenario, the APSM was exercised to provide
a corresponding set of air quality patterns or scenarios. These air quality
patterns were used to compute Figures of Merit which form the basis for the
selection of monitoring sites.
Information on the following was required for operation of the APSM for
each of the chosen scenarios: near-surface temperature, initial and boundary
CO concentrations, atmospheric stability, mixing depth, and near-surface wind
speed and direction fields.
Mixing depths for daylight hours of each scenario were computed using
hourly near-surface temperatures and the 0400 LST upper air sounding taken
by NWS at McCarran International Airport for the dates shown in Table 5. The
technique utilized was developed by Holzworth (1964) and consists of extending
a dry, adiabatic lapse-rate line from the surface temperature to the 0400 tem-
perature sounding on a thermodynamic chart. The height of the intersection
is the mixing depth. The resulting diurnal curves were modified for the noc-
turnal and late afternoon periods as discussed in Section IV.
Near surface wind data for the example scenario dates existed for two
locations in the valley: Nellis Air Force Base and McCarran International
42
-------
Airport. It was not feasible to develop wind fields for the modeling region
using the objective interpolation in the APSM from only two data points at
the low wind speeds of the scenarios. For such wind speeds, it is likely that
local topographically and thermally induced circulations predominate with a
lesser influence provided by the existing synoptic scale pressure gradient.
These local circulations are not readily quantifiable without extensive addi-
tional field measurements. Thus, an objective wind field model or meteoro-
logical simulation model using the topographical information and wind field
data from the routine sampling program will be required to develop an ade-
quate wind field from the few data points in the historical data base. Since
such models were not available to the project, it was necessary to use the
existing wind field from the validation date for which the local topographi-
cally and thermally induced circulations appeared to be the most pronounced
and apply this for all scenario runs. The date chosen was December 3, 1975.
With this restriction, the credibility is diminished for any network design
developed using the chosen scenarios. The present work at least represents
a quantitative example of the application of the network design methodology
for a realistic situation for a location such as the Las Vegas Valley.
Near-surface temperature data for quantifying the CO emissions from cer-
tain motor vehicle sources and exposure classes for determination of the ver-
tical eddy diffusivities were both taken from the December 3, 1975, data base.
Neither was considered to represent a severe restriction in relation to that
imposed by the near-surface wind field utilized. For the nearly clear sky
and light wind situations that typically exist for the scenarios, a very simi-
lar diurnal temperature curve occurs. The range of temperatures reasonably
expected under such conditions for the local CO season is too small to exhibit
a substantial effect on the calculated emissions inventory. The range of ex-
pected exposure classes for such conditions is likewise too small to signifi-
cantly alter the calculated values of the diffusivities. It should be noted
that the above data could have been obtained from hourly weather observations
made by the NWS at McCarran Airport for each of the example scenario dates.
Background and initial CO concentration data used were those developed
for the December 3 intensive date. The results of the intensive field program
and the model validation studies indicated that background concentration would
not have a major impact on the local CO concentration field. These same
studies show that the initial CO fields have a significant impact only for the
first 2 or 3 hours of simulation. Another approach for model initialization
would have consisted of using average CO fields for intensive days or for
representative routine sampling dates. The potential impact of the initial
CO data field could, of course, be diminished by choosing an earlier diurnal
starting time for initiating the simulations.
RESULTS AND DISCUSSION
The APSM was exercised for each of the above described cases. The pre-
dicted hour-by-hour CO concentrations for each grid point were then used to
compute Figures of Merit. Based on maximum 1-hour CO concentrations, the
43
-------
following was computed:
/ \ / S,1
6 / Frequency of Occurrence \ / Maximum 1-hour surface \
F (i,j) = V" I ฐf Meteorological I . I CO concentration at Grid I
1 o_ \ Pattern ฃ / \ Point i, j under Pattern U
Isopleths of these Figures of Merit are plotted in Figure 16.
It should be noted that the Figure of Merit calculation is a mathematical
expression of the concept of overlapping isopleth maps in order to locate high
frequency-high concentration coordinates as described earlier (Liu et al.,
1977). However, use of equation (10) in lieu of actual overlapping of iso-
pleth maps may give rise to the situation where the maximum concentration
location is not always selected as the prime location for a station since
there is a frequency factor which must be considered. The network designer
must be aware of this possibility and make adjustments as appear appropriate.
Such adjustments should be subject to a set of rational criteria in order to
maintain the universality of application built into the methodology.
As a part of the siting methodology, a computer program was written that
searches for the highest values of the Figure of Merit. The number of loca-
tions is arbitrary. The program then eliminates locations with high Figures
of Merit that are adjacent to locations with higher Figures of Merit without
an intervening trough. Such locations are considered to be adequately repre-
sented by the adjacent location with the highest value of the Figure of Merit.
The isolated peaks of the Figure of Merit thus selected are chosen as poten-
tial candidates for monitoring stations. The locations are ranked alphabetic-
ally according to the order of importance based upon the computed Figure of
Merit. The nine locations that rank highest are plotted in Figure 17 along
with the locations of the nine existing CO monitors in the Las Vegas Valley.
A different set of Figures of Merit can be calculated based on the 8-hour
average CO concentrations. This was done for both the morning period and the
evening period (Figures 18 and 19). A comparison of the locations selected
based on maximum 1-hour average CO concentrations with those based on 8-hour
evening averages (hours 1200 to 1900 LST) shows that the first three locations
are identical and the remainder are shifted only slightly. The locations
selected based on the 8-hour evening averages are quite similar to those based
on the 8-hour morning averages (hours 0500 to 1200 LST). The location of the
first station is the same for both cases, but the second one is located at
the southern end of the Las Vegas Strip based on the evening averages, and in
the vicinity of Henderson based on the morning averages. It seems that the
siting methodology developed under the present project can detect subtle di-
urnal variations in the emissions pattern which is unique for the Las Vegas
area.
The calculation of the Figure of Merit has also been considered for run-
ning 8-hour averages rather than the 8-hour periods centered around the morn-
ing and evening traffic peaks. Although this was not done, the result can be
explained qualitatively on the basis of the concentration distributions
44
-------
ฎ
CD
U)
<&- _
CO
LU
cn
CM
10
m
NORTH
20 30 40
I i i i i I I I I Tl I I I I I I | | i | |
..E
ID
PERKS PflNKED flLPHflBETICRLLY
CONTOUR IfTERVflL -- 1 PPM
m:::
(a
cc
"la-:
i i i i i i i i
i i
10
fJ
20 30
SOUTH
CO
cr
UJ
on
(S)
OJ
esi
Figure 16. Isopleths of Figures of Merit based on maximum 1-hour average
CO concentrations in the Las Vegas Valley
45
-------
T
'60
50
40
20
'10
Figure 17
Code
*
N
S
C
D
E
A
J
M
U
in
C
0
^5
ซ
C!
C
"vt
X
HI
Type ot Stations
Proposed Stations
Nellis Air Force Base
CCHD (Shadow Lane)
CCHD (Casino Center)
Desert Inn Goll Course
East Charleston
Arden
Northwest
Henderson
Winterwood
30
i
A-U Present Locations
* Proposed Locations
12345678
I I I I i I I I
Locations of CO measurement sites and those proposed on the
basis of maximum 1-hour average CO concentrations
46
-------
produced in the validation and scenario runs. During the late morning and
midday hoursfrom approximately 1000 to 1500 LSTCO concentrations are very
low and the spatial distribution is nearly uniform. On the other hand, as
seen in the concentration isopleth data in Appendix C, at the periods of peak
traffic the concentrations become high and the distribution of pollutants be-
comes very inhomogeneous. Hence, a running average which included the period
from approximately 1000 to 1500 would tend to distribute the Figures of Merit
more uniformly than an average taken around the period of peak traffic. How-
ever, since the concentrations are much higher during the peak traffic periods,
it is expected that locations of stations selected on the basis of the Figure
of Merit would not change significantly provided the period of peak concen-
trations is included in the running average.
It should be noted that the Figure of Merit does not yield an optimum
solution in a rigorous mathematical sense. That is, the derivative of an ob-
jective function subject to specified constraints is not maximized or minimized.
However, the procedure is similar to this strict mathematical one in that it
searches out maximum values of a well-defined function. The locations of these
values are then prioritized in descending order as potential monitoring sites
for the stated purpose of detection of violations of the NAAQS.
The exercise described here optimizes the existing network with respect
to siting the stations. The question still arises about the optimum number
of stations: is it nine, eight, or ten? In order to approach this question
rationally, the cost and the associated benefit of a unit of monitoring infor-
mation- must be evaluated. A comparison must then be made of the cost/benefit
ratios resulting from the addition or deletion of a station. The optimum
number of stationsoptimum with respect to cost/benefit ratiowould be that
number of stations which gives the smallest non-zero value of the cost/benefit
ratio.
A second constraint might be to optimize the network such that the data
retrieved reproduces pollutant isopleths to within some predetermined level
of error at a specified confidence interval. Then, by application of advanced
statistical procedures, one may determine the number of data points (monitor-
ing sites) required to generate a surface (pollution isopleths) of known
accuracy with known or predetermined error limits.
The important point is that a system may be optimized with respect to
many potential constraints and that it is incumbent upon the network designer
to carefully define those constraints and to assure that they are compatible
with the network objectives prior to any network design effort.
47
-------
NORTH
Si.
fv
03
LO
CO
LU
CM
10 20 30 l|0
1 1 1 1 1 1 1
-
-
:
-
-
-
-
-
-
_
-
-
PEfiKS R
CONTOUR IN
1 1 1 1 1 1 | 1 1
1
1 1 1 1 1 1 1 I'
_,.
flNKED PLPHfiB
TEPVflL -- I PPM
U_I_1_1_1J_J_LJU
3 2
I 1 1 1 1 1 1 1
E:> r/'"
f,*' .' ~" '""'
-"'"..' ':" ' f ..'^T--*.
.... ; i ! \ f \3L~-*.
;...:: // i U
\ \ "" \t _ "''
:. \
1 fi:--'
m
ET 1C ALLY
Mill I 11
3 3
i i i i i i i i i
H
/ ( '" -, \
''': / \
---.. '- '"..
S '-.. ": \
\ / ,.''' \
"^..-"" ...,-.. ,.j
"".. ..-"'
\ f
''"'; '"':"' ...';
\ ^
y ....--
1 i 1 I 1 1 1 t
3 i|-
f~TTTTTTn
':
\
\ \
; :.
/
\ i
1 1 1 1 1 I 1
3
SOUTH
..&>
.(Si
to
.G!
U5
co
cc
LU
ro
OJ
Figure 18. Isopleths of Figures of Merit based on morning (0500 to 1200 LST)
8-hour average CO concentration in the Las Vegas Valley
48
-------
CO
LU
IS.
0
C1-.
uT
tn~
C\J~
%
It
"| I 1 i i i i i
-------
VI CONCLUDING REMARKS
An objective method that uses aerometric data and a mesoscale air quality
simulation model was proposed for selecting sites for pollutant monitoring
networks in urban areas (Liu et al., 1976). This report discusses that method
and its applicability and potential as a planning tool, using the Las Vegas
Valley as a test region.
An advantage of this method is not only that it avoids subjectivity in
the choice of monitoring locations but also that it offers an optimum network
configuration for a given set of criteria. Furthermore, objective methods are
very flexible. For example, if a weather forecast is available, an objective
method such as the one described herein can be used to locate monitoring sites
that accommodate future anticipated emission distribution patterns. The utility
of the method is by no means limited to the design of a new monitoring network.
It should also be useful for the modification (through addition or relocation
of stations) of an existing network that has known deficiencies (e.g., Gold-
stein, 1976).
The application of the method is not limited to the test area (Las Vegas
Valley). The use of local parameters as model inputs and locally measured
data for model verification is possible in order to apply the method to a
different city or locale. Depending on available resources and the specific
situation, the model may be more or less fine-tuned to the specific area for
more accurate predictions. Costs involved in the application of the method
will vary widely depending on the amount and quality of data available. For
example, if long-term aerometric data from a dense network are available, one
might even consider eliminating the model validation steps by generating pol-
lutant distributions and their frequencies strictly from historical data.
The validity of the results obtained by the proposed method necessarily
depends on the reasonableness of many hypotheses or assumptions that are in-
voked. The most critical assumptions are that
the air quality model used simulates pollutant concentration
distributions in a reasonably accurate manner;
the chosen scenarios are representative of the meteorological
conditions during which high pollutant concentrations occur;
the criteria for locating the monitoring sites are appropriate.
These assumptions are believed to be valid for the application presented
in this report.
50
-------
REFERENCES
Ames, J., J. D. Reynolds, D. C. Whitney, and N. T. Fisher. 1978. User's
Guide to the SAI Photochemical Air Pollution Simulation Program. Final
report on Contract 68-03-2399, U.S. Environmental Protection Agency,
Las Vegas, Nevada.
Behar, J. V., L. M. Dunn, J. L. McElroy, R. R. Kinnison, and P. N. Lem. 1976.
Development of Criteria for Establishing Guidelines for Optimization of
Environmental Monitoring Networks: Air Monitoring Networks. In: Pro-
ceedings of the International Conference on Environmental Sensing and
Assessment, 20-1.
Darby, W. P., P. J. Ossenbruggen, and C. J. Gregory. 1974. Optimization of
Urban Air Monitoring Networks. Journal Environmental Engineering Division
(American Society Civil Engineering), EE3, pp. 577-591.
Goldstein, I. F. 1976. Use of Aerometric Network Data to Monitor Acute Health
Effects. Paper No. 76-32.6, 69th Annual Meeting of the Air Pollution
Control Association, Portland, Oregon.
Hoel, P. G. 1962. Introduction of Mathematical Statistics (Third Edition).
John Wiley and Sons, Inc., New York, 427.
Holzworth, G. C. 1962. A Study of Air Pollution Potential for the Western
United States. Journal of Applied Meteorology, Vol. 1, No. 2, pp. 366-382.
Holzworth, G. C. 1964. Estimates of Mean Maximum Mixing Depths in Contiguous
United States. Monthly Weather Review, Vol. 92, No. 5, pp. 235-242.
Holzworth, G. C. 1974. Meteorological Episodes of Slowest Dilution in Con-
tiguous United States. EPA-650/4-74-002, U.S. Environmental Protection
Agency .
Johnson, W. B. , F. L. Ludwig, N. F. Dabberdt, and R. J. Allen. 1973. An
Urban Diffusion Model for Carbon Monoxide. Journal of Air Pollution
Control Association, Vol. 23, pp. 490-498.
Koch, R. C., and S. D. Thayer. 1972. Validity of The Multiple Source Gaus-
sian Plume Urban Diffusion Model Using Hourly Inputs of Data. In: Pro-
ceedings of Conference on Urban Environment and Second Conference on
Biometeorology, Philadelphia, Pennsylvania, October 31-November 2;
p. 64-68.
51
-------
Lettau, H. H. 1969, Note on Aero-Dynamic Roughness - Parameter Estimation on
the Basis of Roughness - Element Description, Journal of Applied Meteor-
ology, Vol. 8, pp. 828-832.
Liu, M. K. 1973. Further Development and Evaluation of a Simulation Model
for Estimating Ground-Level Concentrations of Photochemical Pollutants.
Vol. III. Automation of Meteorological and Air Quality Data for the SAI
Urban Airshed Model. Report R73-SAI-32, Systems Applications, Inc.,
San Rafael, California 94903.
Liu, M. K. , D. C. Whitney, and P. M. Roth. 1976a. Effects of Atmospheric
Parameters on the Concentration of Photochemical Air Pollutants.
Journal of Applied Meteorology, Vol. 15, pp. 829-835.
Liu, M. K., D. C. Whitney, J. H. Seinfeld, and P. M. Roth. 1976b. Continued
Research in Mesoscale Air Pollution Simulation Modeling, Vol. I. Assess-
ment of Prior Model Evaluation Studies and Analysis of Model Validity
and Sensitivity. EPA-600/4-76-016A.
Liu, M. K., J. P. Meyer, R. I. Pollack, P. M. Roth, J. H. Seinfield, J. V.
Behar, L. M. Dunn, J. L. McElroy, P. N. Lem, A. M. Pitchford, and N. T.
Fisher. 1977. Development of a Methodology for the Design of a Carbon
Monoxide Monitoring Network. EPA-600/4-77-019. U.S. Environmental
Protection Agency.
Lund, I. A. 1963. Map-Pattern Classification by Statistical Techniques.
Journal of Applied Meteorology, Vol. 2, pp. 56-65.
McCormick, R. A. 1968. Air Pollution Climatology. Air Pollution; A. C.
Stern, editor, Academic Press, New York, New York.
Meyers, J. P. 1971. Discriminant Analysis in Laterite and Lateritic Soils
and Other Problem Soils of Africa. An engineering study for Agency for
International Development. AID/csd-2164. June.
Morgan, G. B., G. Ozolins, and E. C. Tabor. 1970. Air Pollution Surveillance
Systems. Science, Vol. 170, p. 289.
Olsson, L. E., and S. Ring. 1974. Validation of Urban Air Pollution Models.
In: Proceedings of 5th Meeting NATO/CCMS Expert Panel on Air Pollution
Modeling, Roskilde, Denmark, June 4-6; Chapter 25.
Ranzieri, A. J. , and C. E. Ward. 1975. Caline Z - An Improved Microscale
Model for the Diffusion of Pollutants From a Line Source. Air Quality
Workshop, Washington, D.C.
Reynolds, S. D. , P. M. Roth, and J. H. Seinfeld. 1973. Mathematical Model-
ing of Photochemical Air PollutionI: Formulation of the Model.
Atmospheric Environment, Vol. 7, pp. 1033-1061.
52
-------
Reynolds, S. D., P. M. Roth, and J, H, Seinfeld, 1974. Mathematical Modeling
of Photochemical Air PollutionIII: Evaluation of the Model.
Atmospheric Environment, Vol. 8, pp. 563-596.
Roach, G. E., and A. E. MacDonald. 1975. Map-Type Precipitation Probabil-
ities for the Western Region. U.S. Department of Commerce. NOAA NWS
Com-75-10428.
Roth, P. M., P. J. W. Roberts, M. K. Liu, S. D. Reynolds, and J. H. Seinfeld.
1974. Mathematical Modeling of Photochemical Air PollutionII. A
Model and Inventory of Pollutant Emissions. Atmospheric Environment,
Vol. 8, No. 2, pp. 97-130. ~
Schuck, E. A., and R. A. Papetti. 1973. Examination of the Photochemical
Air Pollution Problem in the Southern California Area. Appendix D of
Technical Support Document for the Metropolitan Los Angeles Intrastate
Air Quality Control Region Transportation Control Plan Final Pro-
mulgation, Region IX, U.S. Environmental Protection Agency, San
Francisco, California.
Seinfeld, J. H. 1972. Optimal Location of Pollutant Monitoring Stations in
an Airshed. Atmospheric Environment, Vol. 6, pp. 847-858.
Shirr, C. C., and L. J. Shieh. 1974. A Generalized Urban Air Pollution
Model and Its Application to the Study of SO-Distributions in the
St. Louis Metropolitan Area. Journal of Applied Meteorology, Vol. 13,
pp. 185-203.
Simmons, W. 1974. Comments on Modeler User Conference. In: Proceedings of
5th Meeting NATO/CCMS Expert Panel on Air Pollution Modeling, Roskilde,
Denmark, June 4-6; Chapter 40.
Slater, H. H., and J. A. Tikvart. 1974. Application of a Multiple-Source
Urban Model. In: Proceedings of 5th Meeting NATO/CCMS Expert Panel
on Air Pollution Modeling. Roskilde, Denmark, June 4-6; Chapter 14.
U.S. Environmental Protection Agency. 1976. Compilation of Air Pollutant
Emissions Factors, AP-42, and Supplements 1 through 5, Second Edition.
53
-------
APPENDIX A
FIELD PROGRAM INSTRUMENTATION
Routine and special sampling of aerometric parameters was accomplished in
the Las Vegas Valley to provide data for model verification and supplementary
data used in the design of a CO-monitoring network. Details of equipment
utilized, data collection and reduction, and quality assurance established for
the field program are presented in this appendix.
EQUIPMENT
The total monitoring network in the Las Vegas Valley was composed of 25
stations operated cooperatively by the CCHD, the NDH, and the EPA. In addi-
tion, wind speed, wind direction, and temperature data from NWS at McCarran
International Airport were utilized. Continuous measurements of CO were made
at nine stations (Figure 3). The instruments used in making these measurements
are
Beckman Model 6800 gas chromatograph (GC)
Bendix Model 8501-5BA non-dispersive infrared (NDIR) analyzer
Beckman Model 7000 dual isotope fluorescence (DIF) analyzer
Energetic Science "Ecolyzer"
Continuous wind speed, wind direction, and air temperature measurements
were made at 13 sites in the network using Meteorology Research Inc. (MRI)
Models 1072 and 1022 weather stations. In addition, near-surface air tempera-
tures were monitored at 4 sites using a Belfort Instrument Company Hygrothermo-
graph, Model 594 (Figure 4).
Measurements of winds aloft were made at two sites (Figure 4) during in-
tensive measurement periods with single theodolite observations of standard
20-gram helium-filled pilot balloons (pibal). Elevation and azimuth angles
at nominal 30-second intervals are read to the nearest 0.1ฐ.
Low altitude aerometric monitoring (spirals) over the Las Vegas Valley
was performed during intensive periods from a Sikorsky S-58 helicopter operated
by the EPA. Air temperature and dew point were measured using a Cambridge
System Model CS-137 (CO data were collected but unusable due to instrument
malfunction). Four flights were conducted on each intensive sampling day:
sunrise to 1 to 2 hours after sunrise, mid-morning, mid-afternoon, and 1 hour
before sunset to 1 to 2 hours after sunset. Locations of helicopter spirals
are shown in Figure A-l.
All data with the exception of those for pibals and the helicopter were
recorded as analog signals on strip charts. Pibal data were recorded manually
54
-------
Helicopter Spiral Sites
1234567
I I I I I I I
Figure A-l. Helicopter spiral sites in the Las Vegas Valley
55
-------
on computer coding forms. Aerometric data obtained using the helicopter were
recorded digitally on magnetic tape with a Cipher Model 70M-7 interfaced to a
Monitor Laboratories Model 7200 data acquisition system.
The distribution of instruments among the 25 sites in the Las Vegas net-
work is given in Table A-l. The table gives the site number and description,
the organization responsible for instrument operation, and the parameters
measured at each site. Table A-2 describes the general land use character-
istics in the vicinity of the CO monitoring sites. Wind sensors were mounted
on towers or utility poles about 10 m above ground where they were well-exposed
and away from obstructions. Hygrothermographs were housed in standard, lou-
vered Stevenson shelters.
FIELD DATA COLLECTION AND REDUCTION
Hygrothermographs measured near-surface air temperature continuously.
Data were retrieved once a week. At site 6, one instrument was placed on the
roof of an air monitoring trailer, and the other on the roof of a nearby four-
story building in order that differences in temperature could be estimated
over the 12-meter change in height.
Strip chart data from the MRI weather stations recording wind run, wind
direction, and temperature were collected at biweekly intervals.
The CO instruments made continuous measurements during the entire field
sampling period. Data from the Ecolyzer instruments (sites 19, 20 and 21)
were collected at either weekly or biweekly intervals. That from the NDH gas
chromatographs (sites 2 and 10) were obtained at 1-month intervals. At CCHD
sites (sites 5, 6 and 9) the strip chart data were collected daily and usually
reduced the same day by CCHD personnel.
At one EPA trailer (site 18), CO data were collected weekly.
Reduction of field data charts from aerometric stations for preparation
of data bases was accomplished on a Hewlett-Packard 9830 calculator and digi-
tizer. Preprocessing procedures for CO data included digitizing initial and
final zero and span values for the determination of drift corrections to be
applied to each of the sample points and corrections to span gas changes over
time. For the MRI charts, wind speed was determined through measurement of
the slope of the wind run.
Pibal data were processed on a CDC 6400 using an algorithm based on stan-
dard trigonometric relationships and a standard rate of balloon rise from NWS
tables. Helicopter data were also processed on the CDC 6400. Data were first
screened, then converted from voltages to engineering units, edited, and
finally plotted as a single parameter as a function of height.
QUALITY ASSURANCE
A quality assurance program was established to ensure that measurements
within the individual programs were comparable and that the data handling was
as error-free as possible. Methods to achieve this are described below.
56
-------
TABLE A-l. DATA COLLECTING SITES
WINTER 1975 - 1976
- LAS VEGAS VALLEY
Ul
LOCATION
OPERATOR PARAMETERS
HELI-
EQUIPMENT COPTER
SPIRALS
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
WS
X
Tule - Wild Animal Reserve
(extreme N.W. )
Nellis Air Force Base
Fire Station #3 - NLV
Pump Station 2, Henderson
CCHD - Shadow Lane
CCHD - Casino Center
ii it ii
McCarran Airport
Rob's - S.E. of Henderson
CCHD - E. Charleston
Arden (extreme S.W.)
Regency - Nellis & E.Charleston
Leisure World-Desert Inn and
Eastern
Cemetery - Paradise Valley
W. Charleston
Jade Park
Sky Harbor Airport
N. 5th & Regina
D.I. Golf Course
Private residence (N.W.)
Private residence (Henderson)
Private residence
(Winterwood golf course)
Apex (extreme N.E.)
R.R. Pass (extreme S.E.)
UNLV
NLV Air Terminal
- wind speed; WD - wind direction
EPA
NDH
EPA
CCHD
CCHD
CCHD
ii
EPA,
EPA
CCHD
NDH
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
; T -
- parameters measured with helicopter s;
WS,
CO,
WS,
WS,
, EPA CO
, EPA CO
n tri
NWS T,
WS,
CO
CO,
WS,
WS,
WS,
WS,
WS,
WS,
WS,
CO
CO
CO
CO
-
-
WD
WS
WD
WD
WS,
WD
WS
WD
WD
WD
WD
WD
WD
WD
, T
, WD
, T,
, T
T
, T
, WD
, T
> T
, T
, T
, T
> T
, T
Pibal
T
temperature ;
pirals are
T,
CO -
dew :
MRI #1072
Beckman 6800 GC, MRI #1072
Pibal Theodolite MRI #1072
MRI #1072
Bendix NDIR #8501-5BA
Bendix NDIR
Belford #594 (2)
Belford #594
MRI #1072
Beckman 7000 DIF
Beckman 6800 GC, MRI #1072
MRI #1072
MRI #1072
MRI #1072
MRI #1072
MRI #1072
MRI #1072
MRI #1072
Beckman 7000 DIF
Ecolyzer
Ecolyzer
Ecolyzer
Theodolite
Belford #594
carbon monoxide
point and altitude
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
TABLE A-2. LAND USE CHARACTERISTICS IN THE VICINITY OF THE
CO MONITORING SITES
Site
No.
Ln
00
Site
2 Nellis AFB
5 CCHD Shadow Lane
6 Casino Center
9 East Charleston
10 Arden
18 D. I. Golf Course
19 Private Residence
20 Private Residence
21 Private Residence
Trailer in vacant lot 2 km from major streets/highways. Scattered storage and
training facilities 0.4 km to N and S, fuel storage facility 1 km to NE, and
runways 2.5 km to E.
CCHD office building surrounded by large parking lot in residential/commercial
area with nearest major streets 0.75 km to W, 1 km to S and 0.5 km to E.
Trailer in edge of parking lot adjoining alley in downtown commercial area.
Nearby structures 1 to 2 stories except for 4-story building 0.1 km to E.
Nearest street 0.05 km to S.
Trailer in construction yard bordering alley adjacent to a small shopping
center and parking lot bordering on a major street 0.1 km S. Area is mixed
commercial/residential with large shopping center 0.5 km to S and other major
streets 0.5 km to W and SW.
Trailer in fenced area adjacent to office building for local civil defense
agency. Surroundings open desert except for office building and scattered
structures along railroad 0.25 km to E. Nearest major road 1 km to N.
Trailer adjacent to tennis courts in center of a large golf course. Nearest
major street is 0.25 km to S with the Las Vegas Strip 0.5 km to W, other major
streets 0.5 km to N and E and major hotel 0.25 km to NW.
Storage shed in the backyard of a home in a residential area. Nearest major
street is 0.2 km to S with other major streets 0.5 km1 and 1 km from the site.
House in a residential area of Henderson. Nearest major street 0.4 km to N.
Open desert beyond 0.5 km to E.
Home bordering a golf course in residential area. Nearest major streets are
0.3 km to the N and 0.8 km to W.
-------
Uniformity in air quality measurements was attained through calibration
of span gases against National Bureau of Standards (NBS) standard reference
gas. EPA and NDH span gases were calibrated against the same NBS standard by
the EPA while the CCHD used its own NBS standard reference. Cross references
were acquired through a single EPA cylinder of known concentration passed as
a blind sample to each agency.
Routine and preventive maintenance and frequent instrument calibrations
were also carried out by each agency. The EPA used the following procedures.
Standardized checklists and log books were kept for all instru-
mentation. (Instrument recorder charts were labeled at each
calibration with site name, span value, time, and date.)
For pilot balloon measurements, a standard alignment reference
point was determined for each site, and used by all technicians.
Mechanical weather stations were serviced routinely every 2
weeks. At this time, the station's alignment with true north
was verified using predetermined reference points.
Ecolyzer CO analyzers were calibrated (i.e., set to zero and
spanned) and inspected for maintenance purposes every 12 or 24
hours depending on the particular instrument's drift character-
istics.
The Beckman 7000 Dual Infrared Fluorescence instrument (site 18)
was calibrated every week.
Hygrothermographs were inspected weekly for maintenance purposes.
In addition, each instrument was calibrated in an environmental
chamber at various known temperatures.
Helicopter instruments were calibrated prior to and following
each flight.
The NDH followed maintenance and check procedures as follows.
At each of two trailers, checks were made twice weekly to detect
possible instrument malfunction. Additional checks were carried
out during intensive study periods.
The gas chromatograph was set to zero and spanned automatically
every midnight.
The CCHD maintained three sites with procedures similar to those used by
the NDH. Instruments were set to zero, spanned, and calibrated every second
or third day.
59
-------
APPENDIX B
EMISSIONS INVENTORY FOR THE LAS VEGAS VALLEY
INTRODUCTION
An inventory of emissions from point sources, area sources, and mobile
sources, expressed as emissions per hour per grid square, was developed for
the Las Vegas Valley for carbon monoxide (CO) for the winter of 1975-1976.
These data were adjusted for temperatures specific to the simulation days.
Information for assembling the inventory was provided by the local electric
power and gas companies, the Nevada Department of Highways (NDH), the Clark
County District Health Department (CCHD), and industries in the area of con-
cern. The National Emissions Data System (NEDS) was also interrogated for
basic emission sources. Emission factors from the U.S. Environmental Protec-
tion Agency (EPA) publication, Compilation of Air Pollutant Emission Factors,
2nd Edition, and its Supplements 1 through 5 (AP-42), were used except as
noted.
The inventory encompassed a modeling region of 48 km by 70 km delineated
by the ridgelines of the mountains surrounding the Las Vegas Valley. Since
much of the Valley is undeveloped, the inventory involved surveys within the
occupied areas and the specification of zero emissions in the bordering areas.
GRID
The final non-zero grid for the emissions data is a network of 1-km
squares, 48 km by 47 km in size, which is skewed by 1.45ฐ from the Universal
Transverse Mercator (UTM) Coordinate System. The UTM zone 7 coordinates of
the corners and the definition of the squares are shown in Figure B-l. Point
(XY) = (0,0) is at the lower left corner of the grid. The grid squares are
numbered 1 to 48 increasing to the east (in the X direction), and numbered
1 to 48 increasing to the north (in the Y direction). A grid square is iden-
tified by the distance of its upper right corner in the X and Y directions
from the origin. Point locations are assigned to squares by using the integer
part of the distance from the point to the origin, and then adding 1 kilometer
to the X and Y values. Thus, the lower left square is the square (1,1). This
grid aligns exactly with that used by the NDH for 1976 traffic data and for
its traffic flow model.
The emissions inventory coordinates (X,Y) and UTM coordinates are related
by the following expressions:
X = (X + 0.66)cosa - (Y + 11.67)sina + 642.0
utm
Yutm = (X + 0-66)sina + (Y + 11.67)cosa + 3964.0
Where: a = 0.025237 radians or 1.445984 degrees.
60
-------
641180 mE X= 1 2 3 4 5 6 7 8 9 10 11 12
4022670 mN 47
46
45
44
43
42
41
40
39
38
37
36
689160 rnE, 4023880 mN
43 44 45 46 47 48 /
642370 mE, 3975680 mN
690350 mE, 3976890 mN
Figure B-l. Emissions inventory grid definition.
61
-------
Emissions data distributed according to several other grids systems with
different origin references were used to develop the inventory. The relation-
ships among these grids are illustrated in Figure B-2. These data were first
located using the EPA coordinate system which is parallel to the UTM coordinate
system.
To convert EPA coordinates to emission inventory coordinates:
X = (X - 0.37)cosa + (Y - 11.68)sina
epa epa
Y = -(x - 0.37)sina + (Y - 11.68)cosa
epa epa
Where:
a = 0.025237 radians or 1.445984 degrees.
To convert EPA coordinates to UTM coordinates:
X = X - 642.0
epa utm
Y = Y - 3964.0
epa utm
Traffic data were supplied by the NDH on a 50"km x 50km grid parallel-
ing the UTM grid.
X = X
traffic grid epa
Y ., = Y + 12
traffic grid epa
SOURCES
The inventory involved the summing of emissions from point, area, and
mobile sources, resulting in emissions per hour per grid square. Point
sources considered were power plants and industrial manufacturing. Space
heating was handled as an area source. Mobile sources considered were air-
craft, railroads, and automobile traffic.
Area (Space Heating)
Estimates of CO emissions from space heating were made using fuel flow
data obtained from Southwest Gas Company records. These data were hourly
averages for the same consecutive 6 days used to estimate power plant emissions.
Space heating included all residential and commercial use and was derived from
the total gas flow data taking into account electric power plant and industrial
use. The consumer distribution figures used were those reported in the South-
west Gas Annual Operating Report for 1975. The distribution figures were
applied uniformly through each day throughout the year. Total space-heating
emissions were calculated by the following equation:
E = (0.79 x C + 0.21 C ) x G,
space h r c h
62
-------
EPA Coordinate System
Traffic Grid
3964000 mN -
y
642000 mE
Hydrocarbon Grid (CCHD)
4034000
mN
690000 mE
Figure B-2
Relationship between grid systems for data used in the
present study
63
-------
Where:
E = emissions from space heating for hour (h)
space h
C = residential space heating factor
r
C = commercial space heating factor
c
G = average gas flow excluding power plant and industrial use
h for hour (h)
The emissions were apportioned to each grid square according to a population
distribution in the Valley linearly interpolated from 1970 census data and
1980 projections found in the NDH (1970) Las Vegas Valley Transportation Study.
Point Sources
Point sources are divided into two groups: power plants, which have
variable hourly emissions, and the other industrial sources with relatively
uniform hourly emissions. Data for the latter were developed from the NEDS
annual totals, using the assumption that hourly emissions were uniform over
the full year.
Calculations for the Sunrise and Clark Power Plants were made using
6-day averages of fuel flows during January 1975. This period was assumed
to be representative of a winter season. Power plant emissions per hour are
given by the following equation:
E. = 0, x FO + G,x FG
h h h
Where:
E = power plant emissions in kilograms for hour (h)
0, = power plant oil flow in pounds for hour (h)
FO = fuel oil emission factor
G = average gas flow for hour (h)
h
FG = natural gas emission factor
Aircraft
Aircraft emissions were calculated for the three main airports in the
Valley: McCarran International Airport, North Las Vegas Air Terminal, and
Nellis Air Force Base. The majority of data for the former were developed
from the Official Airline Guide, North American Edition. This book contains
up-to-date airline schedules including flight times, flight numbers, and
aircraft used. The number of nonscheduled flights and aircraft types (6)
were estimated by Hughes Executive Terminal personnel. Using this information,
hourly engine landing and take-off (LTO) cycles were obtained using the
64
-------
following equations:
Engine LTD cycles = //operations
x # engines per aircraft
An operation is an aircraft arrival or departure.
The calculation of total aircraft emissions per hour at McCarran Inter-
national Airport is given by the equation:
\ -ฃ.! Vi
Parameters for the above equation are defined as follows:
E = total aircraft emissions for hour (h)
Lhi = the number of engine LTD cycles for hour (h) and
aircraft class (i)
C^ = the emission factor in kg/engine LTD cycle for
aircraft class (i).
The North Las Vegas Air Terminal is used principally by private aircraft
and during daylight hours. Information on the number of flight operations
conducted each day were provided by the air traffic controller. These opera-
tions were estimated to occur uniformly each day over the year and equally
over all daylight hours on any given date.
Total hourly aircraft emissions from McCarran International Airport and
North Las Vegas Air Terminal were distributed across grid squares which were
covered by the runways of the two airports. The calculation of aircraft emis-
sions per grid square is given by:
EhXY = \FXY
Where:
EhXY - aircraft emissions in kilograms for hour (h) and grid
square (XY)
E = aircraft emissions from the airport for hour (h)
h
F = fraction of airport runway contained in grid square (XY)
A.X
Emissions data, supplied by the U.S. Air Force Environmental Section, were
similarly distributed over the grid square encompassing Nellis Air Force Base.
Railroads
An estimate of annual emissions from railroads was taken from Transporta-
tion Control Plan Development for Clark County. Nevada (1975).
65
-------
According to the Union Pacific Trainmaster in Las Vegas, approximately
13 trains pass through Las Vegas per day. Since service is unscheduled, the
arrivals and departures of trains are distributed in an unknown pattern. To
assign a distribution of these sources in the Valley, the total daily railroad
emissions were divided into 13 equal parts and, using a random number generator,
they were distributed among the 24 hours of the day. An equal portion of emis-
sions for each train was assigned to each of the 70 grid squares through which
the trains pass. The time lag of a train moving from one end of the valley
to the other was assumed to be less than 1 hour.
Traffic
Emissions from traffic were developed from actual data for 1974 supplied
by the NDH. For each grid square, data are given for roadway types, average
speeds, the number of vehicle miles traveled (VMT) and hourly-averaged traffic
flow. Cold/hot start information and the percentages of light-duty (LDV) and
heavy-duty (HDV) vehicles traveling on the various roadway types are also
included. Data for diesel trucks and motorcycles were not available. However,
these were shown to contribute less than 3 percent of all CO (TRW, Inc., 1975).
The calculation of HDV emissions is given by:
(%TRAF,
100
where:
= HDV emissions in kilograms for hour (h) and grid square (XY)
%TRAF = percentage of daily traffic for hour (h) as developed from
hourly traffic count data
%HDV = percentage of HDV for roadway type (q)
VMT = the vehicle miles traveled per day for roadway type (q) and grid
square (XY)
cin = the 1975 Federal Test Procedure (FTP) mean emission factor for
the i-th model year HDV in calendar year (n)
min = the fraction of annual travel by the i-th model year HDV during
calendar year (n)
vig = the speed correction factor for the i-th model year vehicles for
average speed (s)
For c. and m. , n = 1975.
in in
66
-------
LDV emissions require correction for temperature dependence. This cor-
rection is included in the following equation:
=75 Cin min Vis
Where:
EhXY = LDV emisslons in kilograms for hour (h) and grid square (XY)
Zit = the temPerature correction factor for non-catalyst vehicles
(pre-1975 model years) for ambient temperature (t)
itw = the hot/cold vehicle operation correction factor for non-catalyst
LDV at ambient temperature (t) and percentage cold operation (w),
defined below
%TRAF = percentage of daily traffic for hour (h) as developed from hourly
traffic count data
%LDV = percentage of vehicles that are LDVs for roadway type (q) and grid
square (XY), given in Table 16.
VMT = the vehicle miles traveled for roadway type (q) and grid square
(XY)
c. = the FTP mean emission factor for the i-th model year LDV during
calendar year (n)
m. = the fraction of annual travel by the i-th model year LDV during
calendar year (n)
v. = the speed correction factor for the i-th model year LDV and aver-
age speed (s)
z'. = the temperature correction factor for catalyst equipped light-duty
lfc vehicles (post-1974 model years) and ambient temperature (t)
ritwx = the hot/cold vehicle operation correction factor for catalyst-
equipped LDV ambient temperature (t), percentage cold operation
(w) and percentage hot operation (x), defined below
For variables c. and m. , n = 1975.
in in
The hot/cold vehicle operation correction factors rฑtw and r'itwx
are
67
-------
determined by the following equations:
'itw
itwx 20 + 27f (t)
(for post-1974 mode! years)
Where:
f(t), g(t) = functions of ambient temperature (t)
w = the percentage of vehicle operation in the cold
start condition, defined to be the first 500 seconds
of engine operation after the engine has not been
used for a period of at least 5 hours
x = the percentage of hot start operation (vehicle start-
up after a short, i.e., less than 1-hour engine-off
period). Hot start operation is assumed to occur at
the national average of 27 percent.
Cold start weighting was determined using results of the NDH gravity model
for traffic flow prediction in the Las Vegas Valley. This model computes the
number of trips from one zone to another for different purposes (e.g., work,
recreation, commercial, shopping, etc.). With the percent of starts which
are cold starts and the average time per trip for each purpose, the cold
starts for each zone are determined on the basis of the following:
w ,., =V ' COLD
(TRIP )
P
p TOT TRIP
8.333
MIN
P
or 1 whichever is less
Where:
w ,.. = the percent of vehicle operation in the cold start mode
J for zone (j)
p = purpose for trip
COLD = percent cold starts for a purpose
TRIP = number of trips in zone (j) for purpose (p)
TOT TRIP = total number of trips zone (j) square
MIN = average number of minutes for a trip of purpose (p)
The zones used in the gravity model overlap the grids used in the emissions
inventory. To obtain the percent of cold operation for each grid square,
the fractions of zones in each grid square were estimated using a grid
68
-------
overlay on a map of the zones. Based on this overlay, the cold start percent-
ages within a given grid square were calculated using the following formula:
m
/
(ZJW) (TRIP.)
Where:
w = the percent of vehicle operation in the cold start mode in grid
square XY
Z.XY = the percent of the j-th zone contained in grid square X, Y
J
TRIP. = total number of trips for zone (j).
This method produced values for cold starts for grid squares ranging from
less than 1 percent to 26 percent, with a mean of 14 and a standard deviation
of 4.7, compared with the national average of 20 percent. The values were
assumed to apply equally during all hours of the day.
Ambient temperature data for use with the LDV equations were obtained
from the NWS station at McCarran International Airport assuming spatial homo-
geneity across the Valley.
Illustrative Example
An illustrative example of total emissions by hour of the day for the
various categories described in previous sections is presented in Table B-l.
This example is for a constant reference ambient temperature of 10 C.
69
-------
TABLE B-l. TOTAL CARBON MONOXIDE EMISSIONS (kg/h) FOR LAS VEGAS VALLEY BY SOURCE
TYPE FOR CONSTANT REFERENCE TEMPERATURE OF 10ฐ c
Category
(kg/h)
% of Total
Source Type
2 34567
ower Other Space
lant Industrial Heating Aircraft Railroad LDV HDV
TOTAL
(continued)
Hour
0
1
2
6.2
0.22
6.3
0.28
6.5
0.38
3
4
5
6.6
0.40
6.5
0.47
6.4
0.35
6 6.8
0.15
7 7.6
0.09
I
8
8.0
0.09
9 8.3
1 0.10
314
11.
314
14.
314
18.
314
19.
314
22.
314
17.
314
7.
314
3.
314
3.
314
3.
.2
34
.2
10
.2
56
.2
05
.2
74
.2
14
.2
00
.2
79
.2
58
.2
84
13.2
0.48
14.0
0.63
13.6
0.80
11.6
0.70
13.0
0.94
12.7
0.69
13.2
0.29
13.6
0.16
14.0
0.16
14.2
0.17
43.0
1.55
97.6
4.38
11.6
0.69
0.0
0.0
0.0
0.0
2.8
0.15
111.1
2.48
324.0
3.91
548.4
6.24
449.1
5.49
0
0
0
0
0
0
119
7.
0
0
0
0
0
0
0
0
119
1.
56
0.
.0
.0
.0
.0
.0
.0
.0
22
.0
.0
.0
.0
.0
.0
.0
.0
.0
35
.0
68
2124.8
76.66
1593.6
71.51
1195.2
70.59
1062.4
64.43
929.6
67.28
1328.0
72.45
3585.6
79.91
6772.7
81.66
6905.5
78.58
6507.1
79.58
270.4
9.76
202.8
9.10
152.1
8.98
135.2
8.20
118.3
8.56
169.0
9.22
456.2
10.17
861.8
10.39
878.7
10.00
828.0
10.13
2771.
2228.
1693.
1649.
1381.
1833.
4487.
8293.
8787.
8176.
8
5
2
0
6
1
1
9
8
9
-------
for the determination of a permanent or semipermanent site for a monitoring
station. By careful selection and characterization of the meteorological
scenarios, it may be possible to include the effects of variable meteorological
conditions in the selection of the optimum site by means of equation (7).
The air quality index in equation (7) may be chosen either as the concen-
tration of a specific pollutant for general air quality monitoring or, for the
case of detecting measurements which exceed an ambient air quality standard
(AAQS), as a delta function defined by
jl if the observed or expected concentration exceeds the AAQS,
6 = \0 if not. (8)
It is possible that in the first case, high values of the Figure of Merit
may be calculated for locations which never exceed the AAQS or some large
fraction of it. Such locations may be excluded from consideration as poten-
tial monitoring sites by truncation of the calculation for them. In the case
of locating a site for the measurement of many pollutants, the air quality
index can apparently be generalized using a composite index c*,
N
c* = ฃ w*c. , (9)
i=l X 1
where c is the observed or expected concentration of species i, and w* is the
weighting factor reflecting the importance of pollutant species i.
METEOROLOGICAL SCENARIOS PERTAINING TO HIGH CO CONCENTRATIONS
Selection of optimal locations for CO monitoring stations by the Figure
of Merit technique was accomplished using the results of simulations with the
APSM as a data base. The simulations were run on meteorological scenarios
developed from historical weather data supplemented by current aerometric in-
formation collected during the intensive and routine field sampling programs.
Analysis of historical CO data, especially that with concentrations from
slightly less than to greater than the NAAQS, in relation to the prevailing
meteorological situation or pattern,should provide definitive information
useful for scenario selection. Unfortunately, the historical aerometric data
base in the Las Vegas Valley is insufficient for this purpose. Consequently,
the meteorological pollution potential expressed as the ventilation rate dis-
cussed previously was substituted for the air quality data in the scenario
selection process. This was not considered to be a severe restriction since
the concentration of a relatively inert pollutant such as CO released primarily
at near ground-level should be proportional to the ventilation rate (McCormick,
1968).
Ventilation rates were calculated using upper air data collected at
McCarran International Airport. Specifically, 5 years of mixing depth and
wind speed data were used for the period between 1959 and 1964. These data
were available on magnetic tape from NOAA's National Climatic Center. Ven-
tilation rates were only determined for 1600 LST data since a nocturnal
surface-based temperature inversion usually existed locally at the txme of
39
-------
the 0400 LSI upper air sounding.
Initially, classification of meteorological situations into categories
was accomplished through statistical analysis of data available on 850-millibar
(mb) constant pressure charts. For this analysis, heights on the charts were
available at intervals of 1 degree latitude and longitude for the contiguous
United States and large portions of adjacent bodies of water. Six years of
the data encompassing the Western United States and portions of the Eastern
Pacific Ocean were utilized from a magnetic tape furnished by the NWS's NSO
in Las Vegas. Objective classification was accomplished in the manner out-
lined by Lund (1963) and Roach and McDonald (1975). Basically, this involves
establishing the linear correlation of heights on the constant pressure sur-
face for all pairs of grid points. The charts or maps correlating highest
with each other are grouped together to form classes. The percentage frequency
of each class determines its probability of occurrence.
The results of the study for the period of the local CO season were not
satisfactory. The classes chosen for the 850-mb constant pressure surface
charts were not usually relatable to distinct features present on ground-level
charts. In addition, a large percentage of charts from the data set could not
be placed into discernible classes. It is likely that better results might
be attained if ground-level data were utilized in the analysis. This expecta-
tion is supported by the fact that the transport and diffusion of ground-based
emissions such as CO on a scale of the Las Vegas Valley will be largely deter-
mined by processes in the planetary boundary layer; the layer rarely extends
to the 850-mb level except in mid-to-late afternoon. However, such ground-level
data were not readily available.
Consequently, meteorological situations were grouped into classes through
visual examination of historical ground-level weather charts. This was accom-
plished using data from 1959-1964 available on microfilm from the NSO. The
classification was done using the entire data set and a selected portion of
the data set for the local CO season. The selected data set consisted of the
dates experiencing the lowest 20 percent of the ventilation rates. This sub-
set of the data was assumed to encompass most of the situations for which the
NAAQS would actually, or nearly, be exceeded. In each case, the frequency of
occurrence of the chosen classes established the requisite probability of
occurrence.
Comparisons between the resulting synoptic meteorological patterns or
classes and the ventilation rates were subsequently made using discriminant
analysis. A computer algorithm of the discriminant analysis procedure applic-
able to this problem has been devised by Meyers (1971). A detailed discussion
of discriminant analysis is found in Hoel (1962). As in the previous case,
the results of the comparisons were poor. That is, very few of them were
statistically significant at the standard 5 percent level. The results are
likely colored by the fact that the synoptic classes that evolved spanned a
very narrow range of conditions, i.e., largely consisting of high pressure
areas in various locations in relation to the Las Vegas Valley. In addition,
pollutant transport and diffusion are determined by the details of atmospheric
circulations which may not be readily divisible by a classification scheme
based fundamentally upon synoptic scale weather features.
40
-------
Because the above techniques yielded limited results, the scenarios were
developed directly from the mixing depth and wind speed data for the dates
experiencing the lowest 20 percent of the ventilation rates. For this purpose,
a two-way contingency table involving mixing depth and wind speed was devised
using this subset of the 5-year data set for the local CO season. The result-
ing table, also known as a test block, is presented as Table 4. Maximum mix-
ing depth (1600 LST value generally corresponds to the daily maximum value)
and average wind speed through this depth are, respectively, the ordinate and
abscissa in the Table. Statistics on the mean and standard deviation of data
values are provided in each block. The number of data sets (n) in each class
is also shown.
TABLE 4. CLASSIFICATION OF METEOROLOGICAL SCENARIOS
Wind Speed
(u = m/s) <200
calm
0.3-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
>9
*
*
u = 2
H = 185
n = 1
u = 2.1
H = 56
n = 1
u = 3.5
H = 163
n - 1
*
*
u = 6.4
H = 107
n = 1
*
*
*
*indicates n = 0
200
u \-
H -
n =
u =
H =
n =
u =
H =
n =
u =
H =
n =
u =
H -
n =
u =
H =
n =
u =
H =
n =
u =
H =
n =
- 400
0
394
1
0.84 + 0.23
288 + 44
8
1.60 + 0.31
342 + 55
11
2.7
320
1
3.5
266
1
4.7
355
1
5.4 + 0.1
309 + 31
2
*
*
9
220
1
*
Mixing Depth (H = m)
400 - 600 600 - 800 800" - 1000 >1000
u <\, o * * *
H = 590
n = 1
u = 0.82 + 0.21 u = 0.82 + 0.21 u = 0.88 + 0.25 u = 0.87 + 0.19
H = 507 + 61 H - 702 + 53 H = 848 + 55 H = 1237 + 182
n = 26 n=25 n = 8 n=7
u = 1.68 + 0.52 * * *
H = 518 + 54
n = 27
* * * *
* * * *
* * * *
* * * *
*
*
*
*
41
-------
The data in the test block were further grouped according to wind speed
alone: calm -3, 3-5, and 5 m/s. Since the second and third classes together
consist of less than 6 percent of the subset and, hence, only about 1 percent
of the total data set, they were excluded from further consideration. Data
in class one were then classified by mixing depth to form six scenarios, as
follows: 0-300, 301-450, 451-600, 601-800, 801-1,000 and XL,000 m. Representa-
tive example dates for each such scenario were then chosen from the data. This
information, along with that yielding the frequency of occurrence of each such
scenario for the data subset, is provided in Table 5.
TABLE 5. METEOROLOGICAL SCENARIOS SELECTED FOR THE LAS VEGAS
VALLEY
Scenario
1-1
1-2
1-3
1-4
1-5
1-6
II
III
Wind Speed
Range
(H/B)
0 <
0 <
0 ซ
0
0
0 <
3 <
5 <
: U <
c U <
c U <
' U <
' U <
( U <
' U <
c U
: 3
: 3
: 3
: 3
: 3
: 3
: 5
Mixing Depth Frequency of
Range Occurrence
300
450
600
800
1000
H <
< H <
< H <
< H <
< H <
< H
300
450
600
800
1000
0.
0.
0.
0.
0.
0.
0.
0.
081
106
431
203
065
057
024
033
Representative
Data
December
November
November
December
November
November
January
January
1,
24
5,
13
3,
28
27,
9,
1964
, 1964
1962
, 1960
1961
, 1961
1961
1962
EXERCISE OF SITING METHODOLOGY
For each given meteorological scenario, the APSM was exercised to provide
a corresponding set of air quality patterns or scenarios. These air quality
patterns were used to compute Figures of Merit which form the basis for the
selection of monitoring sites.
Information on the following was required for operation of the APSM for
each of the chosen scenarios: near-surface temperature, initial and boundary
CO concentrations, atmospheric stability, mixing depth, and near-surface wind
speed and direction fields.
Mixing depths for daylight hours of each scenario were computed using
hourly near-surface temperatures and the 0400 LST upper air sounding taken
by NWS at McCarran International Airport for the dates shown in Table 5. The
technique utilized was developed by Holzworth (1964) and consists of extending
a dry, adiabatic lapse-rate line from the surface temperature to the 0400 tem-
perature sounding on a thermodynamic chart. The height of the intersection
is the mixing depth. The resulting diurnal curves were modified for the noc-
turnal and late afternoon periods as discussed in Section IV.
Near surface wind data for the example scenario dates existed for two
locations in the valley: Nellis Air Force Base and McCarran International
42
-------
Airport. It was not feasible to develop wind fields for the modeling region
using the objective interpolation in the APSM from only two data points at
the low wind speeds of the scenarios. For such wind speeds, it is likely that
local topographically and thermally induced circulations predominate with a
lesser influence provided by the existing synoptic scale pressure gradient.
These local circulations are not readily quantifiable without extensive addi-
tional field measurements. Thus, an objective wind field model or meteoro-
logical simulation model using the topographical information and wind field
data from the routine sampling program will be required to develop an ade-
quate wind field from the few data points in the historical data base. Since
such models were not available to the project, it was necessary to use the
existing wind field from the validation date for which the local topographi-
cally and thermally induced circulations appeared to be the most pronounced
and apply this for all scenario runs. The date chosen was December 3, 1975.
With this restriction, the credibility is diminished for any network design
developed using the chosen scenarios. The present work at least represents
a quantitative example of the application of the network design methodology
for a realistic situation for a location such as the Las Vegas Valley.
Near-surface temperature data for quantifying the CO emissions from cer-
tain motor vehicle sources and exposure classes for determination of the ver-
tical eddy diffusivities were both taken from the December 3, 1975, data base.
Neither was considered to represent a severe restriction in relation to that
imposed by the near-surface wind field utilized. For the nearly clear sky
and light wind situations that typically exist for the scenarios, a very simi-
lar diurnal temperature curve occurs. The range of temperatures reasonably
expected under such conditions for the local CO season is too small to exhibit
a substantial effect on the calculated emissions inventory. The range of ex-
pected exposure classes for such conditions is likewise too small to signifi-
cantly alter the calculated values of the diffusivities. It should be noted
that the above data could have been obtained from hourly weather observations
made by the NWS at McCarran Airport for each of the example scenario dates.
Background and initial CO concentration data used were those developed
for the December 3 intensive date. The results of the intensive field program
and the model validation studies indicated that background concentration would
not have a major impact on the local CO concentration field. These same
studies show that the initial CO fields have a significant impact only for the
first 2 or 3 hours of simulation. Another approach for model initialization
would have consisted of using average CO fields for intensive days or for
representative routine sampling dates. The potential impact of the initial
CO data field could, of course, be diminished by choosing an earlier diurnal
starting time for initiating the simulations.
RESULTS AND DISCUSSION
The APSM was exercised for each of the above described cases. The pre-
dicted hour-by-hour CO concentrations for each grid point were then used to
compute Figures of Merit. Based on maximum 1-hour CO concentrations, the
43
-------
following was computed:
110)
6 / Frequency of Occurrence \ / Maximum 1-hour surface
F (i,j) = V^ I of Meteorological I . I CO concentration at Grid
1 e=i \Pattern ฃ / \ Point i, j under Pattern
*- 1 \ / \
Isopleths of these Figures of Merit are plotted in Figure 16.
It should be noted that the Figure of Merit calculation is a mathematical
expression of the concept of overlapping isopleth maps in order to locate high
frequency-high concentration coordinates as described earlier (Liu et al.,
1977). However, use of equation (10) in lieu of actual overlapping of iso-
pleth maps may give rise to the situation where the maximum concentration
location is not always selected as the prime location for a station since
there is a frequency factor which must be considered. The network designer
must be aware of this possibility and make adjustments as appear appropriate.
Such adjustments should be subject to a set of rational criteria in order to
maintain the universality of application built into the methodology.
As a part of the siting methodology, a computer program was written that
searches for the highest values of the Figure of Merit. The number of loca-
tions is arbitrary. The program then eliminates locations with high Figures
of Merit that are adjacent to locations with higher Figures of Merit without
an intervening trough. Such locations are considered to be adequately repre-
sented by the adjacent location with the highest value of the Figure of Merit.
The isolated peaks of the Figure of Merit thus selected are chosen as poten-
tial candidates for monitoring stations. The locations are ranked alphabetic-
ally according to the order of importance based upon the computed Figure of
Merit. The nine locations that rank highest are plotted in Figure 17 along
with the locations of the nine existing CO monitors in the Las Vegas Valley.
A different set of Figures of Merit can be calculated based on the 8-hour
average CO concentrations. This was done for both the morning period and the
evening period (Figures 18 and 19). A comparison of the locations selected
based on maximum 1-hour average CO concentrations with those based on 8-hour
evening averages (hours 1200 to 1900 LST) shows that the first three locations
are identical and the remainder are shifted only slightly. The locations
selected based on the 8-hour evening averages are quite similar to those based
on the 8-hour morning averages (hours 0500 to 1200 LST). The location of the
first station is the same for both cases, but the second one is located at
the southern end of the Las Vegas Strip based on the evening averages, and in
the vicinity of Henderson based on the morning averages. It seems that the
siting methodology developed under the present project can detect subtle di-
urnal variations in the emissions pattern which is unique for the Las Vegas
area.
The calculation of the Figure of Merit has also been considered for run-
ning 8-hour averages rather than the 8-hour periods centered around the morn-
ing and evening traffic peaks. Although this was not done, the result can be
explained qualitatively on the basis of the concentration distributions
44
-------
ซ>
to
in
CO
UJ
m
CM
10
NORTH
30
m
m
i r
PEflKS RPNKED PLPHflBETICftLLY
CONTOUR IfTERVRL -- 1 PPM
i i i i i j
20 30
SOUTH
40
I
CO
d
UJ
Figure 16. Isopleths of Figures of Merit based on maximum 1-hour average
CO concentrations in the Las Vegas Valley
45
-------
60
'50
40
20
'10
T
Code
*
N
S
C
D
E
A
J
M
U
tations
c
V)
X
UJ
Type of Stations
Proposed Stations
Nellis Air Force Base
CCHD (Shadow Lane)
CCHD (Casino Center)
Desert Inn Golf Course
East Charleston
Arden
Northwest
Henderson
Winterwood
30
N
T
60
\
10
I
A-U Present Locations
* Proposed Locations
10
40
Figure 17. Locations of CO measurement sites and those proposed on the
basis of maximum 1-hour average CO concentrations
46
-------
produced in the validation and scenario runs. During the late morning and
midday hoursfrom approximately 1000 to 1500 LSICO concentrations are very
low and the spatial distribution is nearly uniform. On the other hand, as
seen in the concentration isopleth data in Appendix C, at the periods of peak
traffic the concentrations become high and the distribution of pollutants be-
comes very inhomogeneous. Hence, a running average which included the period
from approximately 1000 to 1500 would tend to distribute the Figures of Merit
more uniformly than an average taken around the period of peak traffic. How-
ever, since the concentrations are much higher during the peak traffic periods,
it is expected that locations of stations selected on the basis of the Figure
of Merit would not change significantly provided the period of peak concen-
trations is included in the running average.
It should be noted that the Figure of Merit does not yield an optimum
solution in a rigorous mathematical sense. That is, the derivative of an ob-
jective function subject to specified constraints is not maximized or minimized.
However, the procedure is similar to this strict mathematical one in that it
searches out maximum values of a well-defined function. The locations of these
values are then prioritized in descending order as potential monitoring sites
for the stated purpose of detection of violations of the NAAQS.
The exercise described here optimizes the existing network with respect
to siting the stations. The question still arises about the optimum number
of stations: is it nine, eight, or ten? In order to approach this question
rationally, the cost and the associated benefit of a unit of monitoring infor-
mation- must be evaluated. A comparison must then be made of the cost/benefit
ratios resulting from the addition or deletion of a station. The optimum
number of stationsoptimum with respect to cost/benefit ratiowould be that
number of stations which gives the smallest non-zero value of the cost/benefit
ratio.
A second constraint might be to optimize the network such that the data
retrieved reproduces pollutant isopleths to within some predetermined level
of error at a specified confidence interval. Then, by application of advanced
statistical procedures, one may determine the number of data points (monitor-
ing sites) required to generate a surface (pollution isopleths) of known
accuracy with known or predetermined error limits.
The important point is that a system may be optimized with respect to
many potential constraints and that it is incumbent upon the network designer
to carefully define those constraints and to assure that they are compatible
with the network objectives prior to any network design effort.
47
-------
10
NORTH
20 30
U5
=*
1
LJ
2
G>
00
CM
* I
1 1 1 1 1 1 1 1 1
-
_
_
_
_
-
-
-
-
~
-
-
_
-
-
PEflKS R
CONTQUP IN
1 1 1 1 1 P 1 1 1
1 1 1 ! 1 < 1 1 1
PNKED flLPHHB
TEF'VflL -- 1 PPM
1. 1- .1 1 1 1 1 ! 1
1 1 1 1 1 1 1 1 1
g-} f " .......,,.
y'" .'
'::!... :" f .'/' ^
\ ; ',- .-Yini--.
: / LHJ ...
: .<'(' i I !
> / / f ) "--,.
<. i ffl.-,.i-~^.
t I. ,., ,.., -...
:.
'. ''v ...-'
i /"-.. ''
tj''""'---:- j
m
ITICflLLY
i i i i i i i
III III!)
B
;' i""'... ''-\
/ I ! /
(EM
)' < '\
"*-,
r--\-\-\-\
/ / ,> \
^L// \
"IJT -" -". '
":
'"--''
''
t
*-. IDJ .."'-. i1
"-...t "'""'.... :-!'.
/.i.
\ '%
iii 1 1 1 1 1
ni 1 1 1 1 1 j r-.
J Cfi
I U5
-1
j
J
j
iฎ
;, _j ^~
\ ~ i
* '* i ง
~~] (-f-\
-. ': _1 \J}
I \ H cr
/ .! -1 LLJ
/ / J
;' J ^
-! J ^
| J
, 1 ]
' / j
_J
3
J~01
1 ri
i i i i i i i.JL
10 20 30
SOUTH
Figure 18. Isopleths of Figures of Merit based on morning (0500 to 1200 LSI)
8-hour average CO concentration in the Las Vegas Valley
48
-------
CO
LU
0
IN
CQ~
L/T
ง-
C0~
cu~
%
H
v^
-
-
-
-
-
-
PEfiKS R
CONTOUR IN
1 1 1 1 1 1 1 1_1
1
NORTH
ป 20 30 40
:
fa
i.
fiNKED filFHHB
TEfr'.'flL -- i PFT1
j LJ_LJ.J 1 1 JLJ-
0 2
< I II II
I '--""".,
% ;''(/' i:'r"^'
";. / !
Ufa, J
'""'* '': i ' .'-
' ''-. f'"''
m
mCfil.LY
Q 3
i i i i i i rrn
m
m "}
'*
'.'' / '-.
"*-..
k
._1_1_1_J_L-LJ_J-J-
9 M-
i i i i i i r~
-
-
-
;
-.,
-
-
L J- L L-L J_i I
Ol
CO
CT
LU
(T)
(M
i*^. '""V 1 1 T" I 1
bUu i h
Figure 19. Isopleths of Figures of Merit based on evening (1200 to 1900 LST)
8-hour average CO concentration in the Las Vegas Valley
49
-------
VI CONCLUDING REMARKS
An objective method that uses aerometric data and a mesoscale air quality
simulation model was proposed for selecting sites for pollutant monitoring
networks in urban areas (Liu et al., 1976). This report discusses that method
and its applicability and potential as a planning tool, using the Las Vegas
Valley as a test region.
An advantage of this method is not only that it avoids subjectivity in
the choice of monitoring locations but also that it offers an optimum network
configuration for a given set of criteria. Furthermore, objective methods are
very flexible. For example, if a weather forecast is available, an objective
method such as the one described herein can be used to locate monitoring sites
that accommodate future anticipated emission distribution patterns. The utility
of the method is by no means limited to the design of a new monitoring network.
It should also be useful for the modification (through addition or relocation
of stations) of an existing network that has known deficiencies (e.g., Gold-
stein, 1976).
The application of the method is not limited to the test area (Las Vegas
Valley). The use of local parameters as model inputs and locally measured
data for model verification is possible in order to apply the method to a
different city or locale. Depending on available resources and the specific
situation, the model may be more or less fine-tuned to the specific area for
more accurate predictions. Costs involved in the application of the method
will vary widely depending on the amount and quality of data available. For
example, if long-term aerometric data from a dense network are available, one
might even consider eliminating the model validation steps by generating pol-
lutant distributions and their frequencies strictly from historical data.
The validity of the results obtained by the proposed method necessarily
depends on the reasonableness of many hypotheses or assumptions that are in-
voked. The most critical assumptions are that
the air quality model used simulates pollutant concentration
distributions in a reasonably accurate manner;
the chosen scenarios are representative of the meteorological
conditions during which high pollutant concentrations occur;
the criteria for locating the monitoring sites are appropriate.
These assumptions are believed to be valid for the application presented
in this report.
50
-------
REFERENCES
Ames, J., J. D. Reynolds, D. C. Whitney, and N. T. Fisher. 1978. User's
Guide to the SAI Photochemical Air Pollution Simulation Program. Final
report on Contract 68-03-2399, U.S. Environmental Protection Agency,
Las Vegas, Nevada.
Behar, J. V., L. M. Dunn, J. L. McElroy, R. R. Kinnison, and P. N. Lem. 1976.
Development of Criteria for Establishing Guidelines for Optimization of
Environmental Monitoring Networks: Air Monitoring Networks. In: Pro-
ceedings of the International Conference on Environmental Sensing and
Assessment, 20-1.
Darby, W. P., P. J. Ossenbruggen, and C. J. Gregory. 1974. Optimization of
Urban Air Monitoring Networks. Journal Environmental Engineering Division
(American Society Civil Engineering), EE3, pp. 577-591.
Goldstein, I. F. 1976. Use of Aerometric Network Data to Monitor Acute Health
Effects. Paper No. 76-32.6, 69th Annual Meeting of the Air Pollution
Control Association, Portland, Oregon.
Hoel, P. G. 1962. Introduction of Mathematical Statistics (Third Edition).
John Wiley and Sons, Inc., New York, 427.
Holzworth, G. C. 1962. A Study of Air Pollution Potential for the Western
United States. Journal of Applied Meteorology, Vol. 1, No. 2, pp.366-382.
Holzworth, G. C. 1964. Estimates of Mean Maximum Mixing Depths in Contiguous
United States. Monthly Weather Review, Vol. 92, No. 5, pp. 235-242.
Holzworth, G. C. 1974. Meteorological Episodes of Slowest Dilution in Con-
tiguous United States. EPA-650/4-74-002, U.S. Environmental Protection
Agency.
Johnson, W. B., F. L. Ludwig, N. F. Dabberdt, and R. J. Allen. 1973. An
Urban Diffusion Model for Carbon Monoxide. Journal of Air Pollution
Control Association, Vol. 23, pp. 490-498.
Koch, R. C., and S. D. Thayer. 1972. Validity of The Multiple Source Gaus-
sian Plume Urban Diffusion Model Using Hourly Inputs of Data. In: Pro-
ceedings of Conference on Urban Environment and Second Conference on
Biometeorology, Philadelphia, Pennsylvania, October 31-November 2;
p. 64-68.
51
-------
Lettau, H. H. 1969, Note on Aero-Dynamic Roughness - Parameter Estimation on
the Basis of Roughness - Element Description, Journal of Applied Meteor-
ology, Vol. 8, pp. 828-832.
Liu, M. K. 1973. Further Development and Evaluation of a Simulation Model
for Estimating Ground-Level Concentrations of Photochemical Pollutants.
Vol. III. Automation of Meteorological and Air Quality Data for the SAI
Urban Airshed Model. Report R73-SAI-32, Systems Applications, Inc.,
San Rafael, California 94903.
Liu, M. K. , D. C. Whitney, and P. M. Roth. 1976a. Effects of Atmospheric
Parameters on the Concentration of Photochemical Air Pollutants.
Journal of Applied Meteorology, Vol. 15, pp. 829-835.
Liu, M. K., D. C. Whitney, J. H. Seinfeld, and P. M. Roth. 1976b. Continued
Research in Mesoscale Air Pollution Simulation Modeling, Vol. I. Assess-
ment of Prior Model Evaluation Studies and Analysis of Model Validity
and Sensitivity. EPA-600/4-76-016A.
Liu, M. K., J. P. Meyer, R. I. Pollack, P. M. Roth, J. H. Seinfield, J. V.
Behar, L. M. Dunn, J. L. McElroy, P. N. Lem, A. M. Pitchford, and N. T.
Fisher. 1977. Development of a Methodology for the Design of a Carbon
Monoxide Monitoring Network. EPA-600/4-77-019. U.S. Environmental
Protection Agency.
Lund, I. A. 1963. Map-Pattern Classification by Statistical Techniques.
Journal of Applied Meteorology, Vol. 2, pp. 56-65.
McCormick, R. A. 1968. Air Pollution Climatology. Air Pollution; A. C.
Stern, editor, Academic Press, New York, New York.
Meyers, J. P. 1971. Discriminant Analysis in Laterite and Lateritic Soils
and Other Problem Soils of Africa. An engineering study for Agency for
International Development. AID/csd-2164. June.
Morgan, G. B., G. Ozolins, and E. C. Tabor. 1970. Air Pollution Surveillance
Systems. Science, Vol. 170, p. 289.
Olsson, L. E., and S. Ring. 1974. Validation of Urban Air Pollution Models.
In: Proceedings of 5th Meeting NATO/CCMS Expert Panel on Air Pollution
Modeling, Roskilde, Denmark, June 4-6; Chapter 25.
Ranzieri, A. J., and C. E. Ward. 1975. Caline Z - An Improved Microscale
Model for the Diffusion of Pollutants From a Line Source. Air Quality
Workshop, Washington, D.C.
Reynolds, S. D., P. M. Roth, and J. H. Seinfeld. 1973. Mathematical Model-
ing of Photochemical Air PollutionI: Formulation of the Model.
Atmospheric Environment, Vol. 7, pp. 1033-1061.
52
-------
Reynolds, S. D., P. M. Roth, and J, H, Seinfeld, 1974, Mathematical Modeling
of Photochemical Air PollutionIII; Evaluation of the Model.
Atmospheric Environment, Vol. 8, pp. 563-596.
Roach, G. E., and A, E. MacDonald. 1975. Map-Type Precipitation Probabil-
ities for the Western Region. U.S. Department of Commerce, NOAA, NWS,
Com-75-10428.
Roth, P. M., P. J. W. Roberts, M. K. Liu, S. D. Reynolds, and J. H. Seinfeld.
1974. Mathematical Modeling of Photochemical Air PollutionII. A
Model and Inventory of Pollutant Emissions. Atmospheric Environment,
Vol. 8, No. 2, pp. 97-130.
Schuck, E. A., and R. A. Papetti. 1973. Examination of the Photochemical
Air Pollution Problem in the Southern California Area. Appendix D of
Technical Support Document for the Metropolitan Los Angeles Intrastate
Air Quality Control Region Transportation Control Plan Final Pro-
mulgation, Region IX, U.S. Environmental Protection Agency, San
Francisco, California.
Seinfeld, J. H. 1972. Optimal Location of Pollutant Monitoring Stations in
an Airshed. Atmospheric Environment, Vol. 6, pp. 847-858.
Shirr, C. C., and L. J. Shieh. 1974. A Generalized Urban Air Pollution
Model and Its Application to the Study of SO-Distributions in the
St. Louis Metropolitan Area. Journal of Applied Meteorology, Vol. 13,
pp. 185-203.
Simmons, W. 1974. Comments on Modeler User Conference. In: Proceedings of
5th Meeting NATO/CCMS Expert Panel on Air Pollution Modeling, Roskilde,
Denmark, June 4-6; Chapter 40.
Slater, H. H., and J. A. Tikvart. 1974. Application of a Multiple-Source
Urban Model. In: Proceedings of 5th Meeting NATO/CCMS Expert Panel
on Air Pollution Modeling. Roskilde, Denmark, June 4-6; Chapter 14.
U.S. Environmental Protection Agency. 1976. Compilation of Air Pollutant
Emissions Factors, AP-42, and Supplements 1 through 5, Second Edition.
53
-------
APPENDIX A
FIELD PROGRAM INSTRUMENTATION
Routine and special sampling of aerometric parameters was accomplished in
the Las Vegas Valley to provide data for model verification and supplementary
data used in the design of a CO-monitoring network. Details of equipment
utilized, data collection and reduction, and quality assurance established for
the field program are presented in this appendix.
EQUIPMENT
The total monitoring network in the Las Vegas Valley was composed of 25
stations operated cooperatively by the CCHD, the NDH, and the EPA. In addi-
tion, wind speed, wind direction, and temperature data from NWS at McCarran
International Airport were utilized. Continuous measurements of CO were made
at nine stations (Figure 3). The instruments used in making these measurements
are
Beckman Model 6800 gas chromatograph (GC)
Bendix Model 8501-5BA non-dispersive infrared (NDIR) analyzer
Beckman Model 7000 dual isotope fluorescence (DIF) analyzer
Energetic Science "Ecolyzer"
Continuous wind speed, wind direction, and air temperature measurements
were made at 13 sites in the network using Meteorology Research Inc. (MRI)
Models 1072 and 1022 weather stations. In addition, near-surface air tempera-
tures were monitored at A sites using a Belfort Instrument Company Hygrothermo-
graph, Model 594 (Figure 4).
Measurements of winds aloft were made at two sites (Figure 4) during in-
tensive measurement periods with single theodolite observations of standard
20-gram helium-filled pilot balloons (pibal). Elevation and azimuth angles
at nominal 30-second intervals are read to the nearest 0.1ฐ.
Low altitude aerometric monitoring (spirals) over the Las Vegas Valley
was performed during intensive periods from a Sikorsky S-58 helicopter operated
by the EPA. Air temperature and dew point were measured using a Cambridge
System Model CS-137 (CO data were collected but unusable due to instrument
malfunction). Four flights were conducted on each intensive sampling day:
sunrise to 1 to 2 hours after sunrise, mid-morning, mid-afternoon, and 1 hour
before sunset to 1 to 2 hours after sunset. Locations of helicopter spirals
are shown in Figure A-l.
All data with the exception of those for pibals and the helicopter were
recorded as analog signals on strip charts. Pibal data were recorded manually
54
-------
MEASURED CO CONCENTRATION
CODE STATION
HELLIS AIR FORCE BASE
CCHD (SHADOW LANE)
CCHD (CASINO cram)
EAST CHARLESTON
NORTHWEST
DESERT INN GOLF COURSE
ARDEN
WINTERWOOD
PPM
0
MEASURED CO CONCENTRATION
STATION PPM
NEU.IS AIR FORCE BASE
CCBD (SHADOW LAHE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTBUEST
DESERT INN GOLF COURSE
LBS VEGRS VBLIDBT10N RUN 16 JBN 76
CO BETWEEN THE HOURS OF 900. BND 1000. PST
CO
LBS VEGBS VBL1DBTION RUN -- 16 JBN 76
BETWEEN THE HOURS OF 1000. BUD 1100. PST
MEASURED CO COHCENTRA1
COPE STATION
N HELLIS AIR FORCE BASE
S CCHD (SHADOW LANE)
C CCHD (CASINO CENTER)
E EAST CHARLESTON
J NORTHWEST
M HENDERSON
D DESERT INN GOLF COURSi
A ARDEN
U WINTERHOOD
"I"?': \ ': : ' '"':-l ': '- "!*: 'J ' "
.;.;;.;:.; ; :.:;..; . :': ;.=; ..
.-,..i..;.j. ,,,;.. ,...;. :. .;.-...-. ..:.;,.;..: -
: ' ' '. : '. '.'.'.':,: ' '. . .
.;..:.,.::. i.; ,n \\\',^( ;..:.:
.x:r.' I.;.',.; ;j '' rr.H \ ' :
El:Hli.n:h :.:!i-I:
.:;.; ; i: ;..;.:; ' \ \ -'^ ,
', ;.;',- ..;:;..:.
.:!:.: ': \ : 1 rrt.i : : : ; ^. ::: :
: ; .-:'.; . : .
Bfi;;:;1;
ION
fra
0
i
2
0
-
0
J^
,' ; f
VVS;C
: : '0 " .
'-..;.".' '''"'
t
:
K
* - -':
ti -
.;.,!:,,, .-
1 ^
-
N - -
KEASURED CO CONCENTRATION
CODE STATION; PPM
N HELLIS AIR FORCE BASE 0
S CCHD (SHADOW LANE) 0
C CCHD (CASINO CENTER) 1
E EAST CHARLESTON 0
J NORTHWEST
M HENDERSON 0
D DESERT INN GOLF COURSE 1
A ARDEN
U UINTERWOOD
;. i ; - i- -. ; ' - . ' i
, . . ,;,:. ;; -.-:.. ;
,;.-. :.}:.'.;;- = -
....... .,. . ,.,.,..... ...^
iL^iii/i
ii:ฐ !;;:
. . ': . : .
v ..;- : -
- N
, y - , , .
M ; : :
;: 'J; '
LOS VEGB? VBLIDBTION PUN -- If- JBH 7f
CO BETWEEN THE HOUPS OF 1100. ftND 1200. P'T
LBS VEGBS VBL1DBT10N PUN -- IS JBN
CO BETWEEN THE HOURS OF J2J0. flND 13W. P3T
Figure C-14.
87
-------
CO CONCENTRATION
CODE STATION
NELLIS AIR FORCE BASE
CCHD (SHADOW LANE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHVEST
HENDERSON
DESERT INN GOLF COURSE
ARDEV
WINTERUOOD
D :
MEASURED CO CONCENTRATION
CODE STATION PPM
N NELLIS AIR FORCE BASE 0
S CCHD (SHADOW LAKE) 0
C CCHD (CASINO CENTER) 1
E EAST CHARLESTON 0
J NORTBWEST
M HENDERSON 1
D DESERT INN GOLF COURSE 1
A ARDEH
D WINTERWOOD
Ji '''::' '
. I . ;.. . '
-: '; i -I r- ;
-> ''' :' , ' -
..;~ :... <:; ; : :.
-: --: : - -; . r
-V : ' :: -:
..... ... . ^.. ,,.-. ...
".'.'. I".' '.
..:': \\.\4. . .:
..'.:.:.. ..,.'::. . :
;: ; .....:;.:. O . ; .
-. -Jl'-;- '
:!::'-:: :::J'::'
i j ' :-Hi i-'t:'?? i ':': ''
: \ . ; ' s .
-.';; i j ' '
|rv;...:::
I '.'':'...
i ':'.. ! i ! ! ! ! '
I .
M
a
: : ' . : '
. ..:..'.., ' ' '
: '.:-: ;:;.;. -
v! !H r: : '
:>> :.;
'. l'~. : !' i./
ป'
, : -.
LOS VEGflS VPLlDNTIi^N RiJM -- i ft JON
BETWEEN THE HOUR'S Of I jttfij. flND 1100. P3T
CO
LPS VEGOS VOL 1 DOT ION RUN -- 16 JON 7f,
BETWEEN THE HOURS OF 1U00. fiND 1500. FS7
MEASURED CO CONCENTRATION
CODE STATION
NELLIS AIR FORCE RASE
CCHD (SHADOW LANE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT INN GOLF COURSE
ARDEH
WINTERVOOD
LOS VEOOS VQLlPftTlON PUN -- H? JON
CO BETWEEN THE HGUP"? OF 15ซJO. ftND ie*>*. P3T
MEASURED CO CONCENTRATION
COM STATION PPM
R NELLIS AIR FORCE BASE 0
5 CCHD (SHADOW LANE) 3
C CCHD (CASINO CENTER) 6
E EAST CHARLESTON 0
J NORTHWEST 0
H HENDERSON I
D DESERT INN GOLF COURSE 1
A ARDEN
U WINTERUOOD
.-.: .j .;:!. i., = i : i i i .:.;:!.
:- ''>>':' -\-\ :'}:
: '..::
>< ' ! ' ; ;
ฑ!::: :.r^;
"". Y;I ; . i
' "" ": ' - !' :
i5';v.-l:-:j:
:i.'::n:!.i.i.r:
.:..;;:.-: ; .r;
.) t . 1 : : . -.
... . . . r
.:..i i ',--, :. ~. . V
: i.ri " '.:;':-
'. L;.-;t.Li..i:
1.': ; ;.;', o .
: | : . -. ; j.j. :
, i . i .,.; i .'. :
'.; ::.-::': :
j VT; ; :/
: (*/Vl J
': '.'ffy\-\i
hMJ;:.
- = :.;.= .
: r! . i : ;.ฃ;
: ;..' - ; i ;
::::;.:; |:i,.
' -:-:-:-
.... ,.^..... ,.,......
:-:i:::.;:f:'
: : ~
......
& ' : - : .
- " ' ' V\
;;-;-!!i!g
:.':; '=1 ; :
;.... ;. ' f i . ; . ':'
';.!.;; i .
.:! iii'lr
i ;. .. [.. .
. ,.,. , . .
f:::!;
- -'I '
LPS VEGBS VflLIDPTION RUN -- 16 JON 7f
CO BETWEEN THE HOURS OF 1680. PND 170ซ. P3T
Figure C-15.
88
-------
MEASURED CO COKEHTUTIOI
STAncn
ELLIS Alt FORCE 1ASI
CCHD (SHAMV LAIR)
CCHD (CASIIB CUTTER)
EAST CUtUSTOI
ELLIS AO. ram USE
CCm (SBADOV LUB)
ccm (CASIIB cram)
EAST CBARLESTDI
NOMlBfEST
LOS VEGRS VBLIDPTION RUN -- 16 JBN 76
CO BETWEEN THE HOURS OF 1700. UNO 1800, PST
CO
LflS VEGHS VHL1DHTION RUN -- 16 JBM 76
BETWEEN THE HOURS OF 1680. BND 1900. PST
HEASVRB) CO COKBHTRATIOI
STATIOa
HELL1S AIB FORCE EASE
CCHD (SHADOW LAME)
CCHD (CASIIB CEHTEE)
EAST CHARLESTON
HOETEUEST
HEIIDERSOd
DESERT \m GOLF COUESE
ARDEH
uiRTEnnco
LOS VEOOS VOL 1 DOT I ON PUN -- 16 MN 7S
CO BETWEEN THE HOURS OF 1900. BND 2000. PST
Figure C-16.
89
-------
MEASURED CO CQปCENTRATI01i
CODE STATION
KELLIS AIR FORCE BASE
CCHD (SHADOW LAKE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT IKH GOLF COURSE
ARDEN
W1NTERVOOD
MEASURED CO CONCENTRATION
COTE STATIOH
HELLIS AZR FORCE BASE
CCHD (SHADOW LAMB)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT INN GOLF COURSE
ARDEN
WINTERHOOD
CO
LOS VEGOS VRLIDPTIOH F'MN -- ?! 'Mil
BETWEEN THE HOURS OF 50O. ftNC- fciปfl F.T
LRS VEGOS VflLJDfiTlON RUN -- 2J Jfill
CO BETWEEN THE HOURS OF 600. fiND 70S. PST
MEASURED CO CONCENTRATIOI)
CODE STATION
KELLIS AIR FORCE BASE
CCHD (SHADOW LANE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT INN GOLF COURSE
ARDEN
UINTERHOOD
NELLXS All FORCE BASE
CCHD (SHADOW LANE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
CO
LPS VEGflS VPLIDflTlON PUN -- 21 JfiN
BETWEEN THE HOUR* OF 790. flND 800. PST
LPS VEGfiS VPLIDQTION RUN -- 21 JOH
BETWEEN THE HOURS OF 800. fiND 900. PST
Figure C-17.
90
-------
MEASURED CO COKCmUTIOII
CODE STATIC*
mils Ail FORCE USE
COD (SHADOW UWE)
COD (CASHO CHRBI)
EAST CHAILESTrai
mmwEST
K*
0
3
t
3
DESERT n* coir COURSE
MEASURED CO COKQITJATIOH
COM STATO?
NELLIS AIR FORCE USE
CCHD (SHADOW LAKE)
CCTO (CASIW CERE!)
EAST CHAILESTOI
NORTRUZST
BEMDERSON
DESERT IHR COLr COURSE
AIDBI
UIimSVOOD
Pt;
CO
LOS VECPS VBLIOPTION RUN --21 JdH 7t
BETWEEN THE HOURS OF 909. BND 1008. PST
CO
LOS VEGBS VHL-1DBT10N RUN --21 JBM
BETWEEN THE HOURS OF 1000. BND 1180. PST
MEASURED CO COHCEBTUTION
CODE STATIOป P7H
IELLIS AIR FORCE USE
CCHD (SHADOW LAKE)
CCHD (CASMO CUTTER)
EAST CHAELESTC*
NORTBUEST
BDTDERSOII
DESERT IKH COLT COURSE
AROm
UTXTERWOOD
..'...}-. i--i!-..;. ~:.-g--!-i-- i-
, ,j. ;..;. i9.^ ,..: f.
.. -.?...
MEASURED CO CONCEHTRATlCni
CODE StATIOJI
MELLIS AIR FORCE BASE
CCED (SBADOH IAIE)
CCHD (CASIBO CEITER)
EAST CHARLESTOH
DESERT im GOLF COURSE
ARDEH
VHRERHOOD
CO
LBS VEGBS VBLIDBTION RUN --21 JBN
BETWEEN THE HOURS OF 1100. BND 1200. PST
LBS VEGBS VPL1DBT10N RUN 21 JBN
CO BETWEEN THE HOURS OF 1200. flND 1380. PST
figure C-18.
91
-------
MEASURED co CONCENTRATION
S95? II*US!
NELLIS AIR POftCE BASE
CCND (SHADOW LANE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT INN GOLF COURSE
ARDEN
WIHTERWOOD
MEASURED CO COHCEIttRATION
ฃ951 SHU
NELLIS AIR FORCE RASE
CCBD (SHADOW LAKE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT INN GOLF COURSE
ARDEN
UINTERUOOD
LBS VECflS VOL 1 DOT I ON RUN -- 21 JflM
CO BETWEEN THE HOURS OF 1300. SND 1100. F5T
LOS VEGflS VRL1DRTJON PUN -- 21 JfiN
CO BETWEEN THE HOURS OF 1400. ftNO 1"?CK). F'S T
MEASURED CO CONCENTRATION
CODE STATION
NELLIS AIR FORCE BASE
CCHD (SHADOW LANE)
COD (CASINO LANE)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT INN GOLF COURSE
ARDEN
unnxRwooD
- -Of
-e
MEASURED CO CONCENTRATION
CODE STATION
NELLIS AIR FORCE BASE
CCHD (SHADOW LANE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHWEST
HENDERSON
DESERT INN GOLF COURSE
ARDEN
WINTERWOOD
LOS VE003 VOL1DOTION PUN -- 2J JAN
BETWEEN THE MOUPซ OF 1SO<5. &N& 16CW. PS I
LflS VECHS VBLIDflTlOH RUN -- 21 JflN
CO BETWEEN THE HOUP^J OF 1600. flND 1700. PST
Figure C-19-
92
-------
MEASURED co CONCENTRATION
CODE ซIAT10II
mun AIR roues use
COD (SIAKN LANE)
CCHD (CASINO CU1U)
KMT CHARLESTON
NORTHUE8T
DESERT INN COIT COURSE
AUDI
uiNTERtfooD
H!
5
13
5
3
4
0
MEASURED CO CONCENTRATION
SSS ilปป
N NELLIS AIR FORCE RASE
S CCHD (SHADOW LANE)
C CCHD (CASINO CENTER)
E EAST CHARLESTON
10
3
D DESERT INN COLT COURSE
A ARDEN
U HINTERHOOD
LOS VECPS VflLlDOTION RUN --21 JBN
CO BETWEEN THE HOUPS OF 170ซ. flND 160-3. PdT
LflS VEGPS VOLlOfiTION RUN -- 21 Jfltj
BETWEEN THE HOURS OF 1800. flND 1900. PST
MEASURED CO CONCENTRATION
CODE STATION PPH
NEU.XS AIR FORCE iASE
CCHD (SHADOW LANE)
CCHD (CASINO CENTER)
EAST CHARLESTON
NORTHHEST
5
14
IBS VEGBS VBLIDBTION RUN 2\ JflM "'
CO BETWEEN THE HOURS OF 1900. flND ฃ000. PS1
Figure C-20.
93
-------
MZASUUD CO CORCBTT1LATIOB
COPE STATIOIf PPM
H nLUS All rOtCI BASE 0
S COLD (SHADOW LA*!} 1
C COD (CASIK> CimU &
D DESDT IKK GOLF COOISt 2
E EAST CHAKLESTOH 6
A AftCBI 0
J NOKTWEST 5
H rtEKDEHSOM 0
U tflHTEKUOOD 2
CODE
N
S
C
D
E
A
J
H
0
MEASURED CO COIKWmmoH
SIATIOII
NELLIS AIR FORCE BASE
CCHD (SHADOW LAMB)
com (CASIHO CENTER)
DESERT Iปr COLT COURSE
EAST CHARLESTON
ARDEN
NORTHWEST
HEHDERSOH
WIKOTIWOOI)
PPM
0
2
6
.
9
0
6
2
4
LOS VEOBS VBLIOflTION RUN -- ^^ JBN
CO BETHEEN THE HOURS OF S08. HND 6i30. PST
CO
LBS VEGfiS VBL1DBT10N RUN --22 Jfit!
BETWEEN THE HOURS OF ฃ00. PND 700. P5T
KEASURED CO COHCERTRATIOD
CODE STATION
NELLIS AIR FORCE RASE
CCHD (SHADOW LAKE)
CCHD (CASINO CENTER)
DESERT INN GOLF COURSE
EAST CHARLESTON
ARDEN
NORTHWEST
HENDERSON
WIHTERUOOD
MEASURED CO CONCENTRATION
CODE STATION
NELLIS AIR FORCE BASE
CCHD (SHADOW LAME)
CCHD (CASINO CENTER)
DESERT INN GOLF COURSE
EAST CHARLESTON
ARDEN
NORTHWEST
HENDERSON
WINTERWOOD
LOS VEOflS VBL10STION RUM -- 22 JftN
BETWEEN THE HOURS OF 700. QND 888. PST
CO
LOS VEGOS VRUDOTION RUN --22 JPH
BETWEEN THE HOURS OF 880. OND 900. PST
Fi-gur-e- O21.
94
-------
MEASURED CO COKENTUTION
CODE STATIC*
HILLIS AIR FORCE BASE
CCHD (SHADOW LANE)
CCHD (CASIKO CENTER)
DESERT INN COIF COURSE
EAST CHARLESTON
ARDEN
NORTHWEST
REXDERSON
WINTERWOOD
t.-l.-l. ;,.!. i. ;
,j..L.I.^..!
MEASURED CO CONCENTRATION
CODE STATION
HEU.IS AIR FORCE USE
CCHD (SHADOW LANE)
CCKD (CASINO CENTER)
DESERT im GOLF COURSE
EAST CHARLESTON
AKDEH
NORTHWEST
HENDERSON
UINTERWOOD
CO
LOS VEGRS VBLIDBT10N RUN -- 22 JfiM
BETWEEN THE HOURS OF 900. BND 1080. PST
CO
LOS VEGfiS VflLIDBTION PUN --22 JPM
BETWEEN THE HOURS OF 10OT. BND 1100. PST
MEASURED CO CONCENTRATION
CODE STATIOH FPM
N NZLLIS AIR FORCE EASE 0
S CCBD (SHADOW LAKE) 2
C CCBD (CASINO CENTER) 2
D DESERT INN GOLF COURSE
E EAST CHARLESTON 1
A ARDEN 0
J NORTHWEST 1
H HENDERSON 0
U W1NTERWOOD 2
l:.:;1rLi.i::!J.::\:vf:;.f:lL:!::b:;-nir:::
;.:.:.";:. . .:::; : : :
' :"."'-?'. ?-
"'":"' T I";"":' *"'
:f:li!:Ht:!:
-:.;:;:;;::; :|;i
:fy::i::i:::::rri:
:t:rj-:;4'W4"!
:.:-i.;'i:::T'i:.;Tj,n:.a'::::
: . ;. : . :. , t;..u.;..;^..|* | '
;; HriH+ -H -*;;;
. ;. .;.....;. A.; .;..;... .;..;.. i. ;.;.; ;..
:.:' ;::;::::'.;.:::.(ป:;.:;:' ..,:. :::';..../. :::'
t-M'ttt J-1-T":
-i~?"? 71 - ;"
fSHJHH
tTrrr;"'"!T"
~'g44T:;::rr
i_(..i.i..;.*--i.- -;..;-.
:::::..
;--r'T-T-:"r":--:" - -;
f-j-r-j-r"T-i--i' <:
^ ? ! ! * 1 ' !--<"*--:--
6-rri-'-ป-i-;-- -f^
|-f44-44-|'f- ;--
f"~'~-T -t-^---- : '; '
;.-i.^.i..i.4 .I--. -:- - --
n-T-f-H-'--;-rrv
4--4-44"i-:-l-:- -
i-K*-~ป4-r-!..-
: . ; ;'--;-; f~f t fr '-."" l"
'"' '.- ";.'.''. "XT'tj '.'{.' "\. r.1'*."l'.. /;"'. I'l-i-Ll^p'T-i
... ;.... .... . .: :. ..;.;...-..,......, ;....( -A.i..^.j..: .: :. .
- ": '
K
....
MEASURED CO CONCENTRATION :
CODE STATION
N NELLIS AIR FORCE BASE
S CCHD (SHADOW LANE)
C CCHD (CASINO CENTER)
D DESERT INN GOLF COURSE
E EAST CHARLESTON
A ARDEN
J NORTHWEST
M HENDERSON
U UIHTERWOOD
'-.':. . .'.! ! . i .:
IT H 1 ::':.!'!'
..: [. : ."'.""::;.
ID.:!: : \
|i;: '-\\\i\
T \\\ ..'. '.
PPM
0 ;;
2 ;
i ;
0
0 '.
;...-.. ': ....'.: t '. ' '- ' ': \
: : :
j : i.i-^ :-. : .
. ., :".4 .;
..... --,,_
'. \ :: i ' ' ' :::
;.| H| !};.;:
-' r : ' ' :
. . -,;..:;
:-; ' ; ! ; ; :
' ' ; : :
:.:::.: " : ' i : :
.-.!..> '..'. '..'.....
; .:::.. \ : .
":ซ h :.:.': ! ;
..-,:;": ' i
' : r
LBS VEGBS VBLIDBTION ROW 22 JBM
CO BETWEEN THE HOURS OF 1100. BND tl**. PST
CO
LBS VEGBS VBLIDBT10N RUN -- 2? IWI
BETWEEN THE HOURS OF 1200. ONP 1300. P51
Figure C-22.
95
-------
MUSURE!) CO CONCENTRATION
CODE STATION PPM
N SELLIS AIR FORCE BASE 0
S CCHD (SHADOW LANE)
C CCHD (CASINO CENTER) 2
D DESERT 1N"N GOLF COURSE
E EAST CHARLESTON 1
A ARDEN 0
J NORTHWEST
M HENDERSON 0
U WINTERWOOD 1
MEASURED CO CONCENTRATION.
CODE STATION PPM
N NELLIS AI8 FORCE BASE 0
S CCHD (SHADOW LANE) 1
C CCHD (CASINO CENTER) 1
D DESERT INN GOLF COURSE
E EAST CHARLESTON 1
A ARDEN 0
J NORTHWEST
M HENDERSON 0
U WINTERWOOD 1
. ; .
..,.;.., ..;!.;. -,.:>
' i -i-f . : :T -:-: ~
; Y: rz.rij.o.;: :in. .
'.;'.". 1 ;"!: ' .." '.-,. ; ;ป . '*". .
- ' : ; " ' ':'
.,.;.,, :.~ :
, , ...'. .,. 1 . . , ., -, :.ft '
..,..l........|..;..^.-,.|.....:..,..j..^-..-.j..:.
;:|:: IB Ely .:j.;;/:ii.v; ;
;;!;: f.H.ri : ':;.::;: :
-; : i- ...;..;..^.*..f. -.;. , .{-; ? :-!
::;- , n.:;:i .:. :v;::r: ';:
r '! !! ; >r: ' ' ' : '
,j.~. ,...:. ;....;..
t~:-i";
N :
.:!tf,*:.:.g|.t.. r -:.;.,
:.,,;..:.;: : :.;.; ;ซ; ....:.:..
..... i - .: :..;..,.....,..;.;..:....... ....
: -: --& V:-';":"V;--T --:; : -
;:'H:-.ir:'.:;;.H;r;:n M ; , ;
... ,..r f ;.....: . .; l .; - ; .1 . . : .
-i"i- >-.- ::--
...:.....;.: ;..,.;,:.:..,.. : .=...,.
.;.: .,.:;::::.. ;.iJ.::i:::;::|::' ;..U ;;
. . . 1 ;-.. i ! :....:. . . -
::rii:::-.-:h::i ' :;
' ; ; . " ": :" ': ; ' ' -
1
: '".
r:^t
M ,-; -
- : r:: *
r ;.: ,
CO
LBS VE6BS VBLIDBTION RUN -- 22 JBN
BETWEEN THE HOURS OF I38B. BND 1U08. PiT
LBS VECBS VOLIDBT10N RUN -- 22 JBN 7T
CO BETWEEN THE HOURS OF 1U00. BNO 158a. PST
MEASURED (X) CORCEHTRATION
CODE STATION PPM
N NELLIS AIR FORCE BASE 0
S CCHD (SHADOW LAKE) 1
C CCHD (CASINO CENTER) 2
D DESERT INN GOLF COURSE
1 EAST CHARLESTON 1
A ARDEN 0
J NORTHWEST
M HENDERSON 0
U WINTERUOOD 1
--.; - _:; j - ;-;-i ; i-.:;;;:;::r:'::;l:
..-.;....;. I I ....-.}.
.-: .!......: ;. .. . T^ j-: -
,....;-;,, ..;... ...... ~f ? ;-.-
-''" -.!'' : . '" ' "'.
II: ::.. ;-L... .-"..'
'"'' i - '-.'
.\ .:. . . .. ; ' . . -
t; ;.; :.-;.:; :;];.;:. ..'4 ;:;;:-
I7 ''::!. : : : : : -
. - ~f - : . .
: - -....,;..---...
-i..|..i..y. ;..,-., -', ,
-ซ..(..fl. .;...; i i ., ...
. : , ,
ft
:. ~~'r~::::
.;.:.;::;:;;.;"
.:.:.:.- :y :
,.i i .-,;.....-
~::rn1:;-;:--:-
^ . M : '
::!;::!.::
---;! ^
..:.;.;.:.::.::...:..,.
; ' .;..-
'-;
t!
MEASITRED CO COHCENTRATIOn
HBLLIS All FORCE BASE
COD (StUDW LAHl)
ccra (CASIK CBHTER)
DESERT INN GOLF COURSE
EAST CHARLESTON
ADDEN
1KWTHUEST
BEHDERSOK
WINTERUOOD
LBS VEOBS VBLIDBT10N RUN --22 JQH
CO BETWEEN THE HOURS OF 1588. BND ISaa. PST
LSS VEGOS VRLIDBTION RUN --22 JflN '6
CO BETWEEN THE HOURS OF 1G0B. BND 17ซ8 PST
Flgtare C-23.
96
-------
MEASURED CO COBCranATUm
CODE STATI PPM
II SILLIS All FOUCe USE
S CCKD (SHADOW LAKE)
C CCHD (CASIKO CHITEII)
D DESERT tm GOLF COURSE
E EAST CHARLESTON
A ARDDI
J NORTHWEST
v wiimtwooo
LOS VEGflS VM.1DPTJON RUN -- 22 JPN 7
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA-600/4-78-053
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
CARBON MONOXIDE NETWORK DESIGN METHODOLOGY
Application in the Las Vegas Valley
5. REPORT DATE
September 1978
6. PERFORMING ORGANIZATION CODE
7.AUTHOR
------- |