Comprehensive Report
FIELD STUDY FOR INITIAL EVALUATION
OF AN URBAN DIFFUSION MODEL
FOR CARBON MONOXIDE
Prepared for:
COORDINATING RESEARCH COUNCIL
30 ROCKEFELLER PLAZA
NEW YORK, NEW YORK 10020
ENVIRONMENTAL PROTECTION AGENCY
DIVISION OF METEOROLOGY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
CONTRACT CAPA-3-68 (1-69)
STANFORD RESEARCH INSTITUTE
Menlo Park, California 94025 • U.S.A.
-------
STANFORD RESEARCH INSTITUTE
Menlo Park, California 94025 • U.S.A.
Comprehensive Report
June 1971
FIELD STUDY FOR INITIAL EVALUATION
OF AN URBAN DIFFUSION MODEL
FOR CARBON MONOXIDE
By: W. B. JOHNSON
W. F. DABBERDT
F. L. LUDWIG
R. J. ALLEN
Prepared for:
COORDINATING RESEARCH COUNCIL
30 ROCKEFELLER PLAZA
NEW YORK, NEW YORK 10020
ENVIRONMENTAL PROTECTION AGENCY
DIVISION OF METEOROLOGY
RESEARCH TRIANGLE PARK, NORTH CAROLINA
27711
CONTRACT CAPA-3-68 (1-69)
SRI Project 8563
Approved by:
R. T. H. COLLIS, Director
Atmospheric Sciences Laboratory
RAY L. LEADABRAND, Executive Director
Electronics and Radio Sciences Division
Copy A/o,
-------
ABSTRACT
A measurement program in San Jose, California, during November and
December 1970, provided data to evaluate and improve our existing
receptor-oriented Gaussian diffusion model for calculating urban carbon
monoxide (CO) concentrations. Seven stations were operated in a two-
block downtown area to measure CO at five heights, winds, and temperature
gradients. CO concentrations and temperatures were also measured by
helicopter and two vans. San Jose's automated downtown network provided
traffic data.
A helical air circulation in the street canyon was observed when
roof level winds were within 45 of the cross-street direction. In
these cases, CO concentrations were proportional to vehicle emissions
in the canyon and to the reciprocal of wind speed. In front of buildings
that face the roof-level wind, observed concentrations from street
emissions are inversely proportional to the street width and nearly
constant with height; on the other side of the street they are inversely
proportional to the slant distance from the nearest traffic lane and
hence decrease with height. These relationships are incorporated in a
new street effects submodel. For winds parallel to the street, the
expressions for the two cross-wind cases are averaged, giving concen-
trations that are the same on both sides of the street.
Transport of CO into and out of the downtown area was determined
from vertical profiles of wind and horizontal profiles of CO concen-
trations at various heights up to 300 m along the perimeter of the central
area. The computed CO input at the surface deduced on this basis was
compared with our emissions submodel calculations. On the average, the
two methods gave values within a factor of about 1.5. Mixing depths from
helicopter measurements of temperature and CO profiles were compared with
those from our mixing depth submodel; six of the nine cases agreed within
15 percent, and eight within 41 percent. On the basis of helicopter
observations, the stability submodel was modified to account for more
stable conditions during immediate post-sunrise and pre-sunset hours.
The revised model also uses an increased diffusion rate appropriate to
urban areas. The changes were based on reevaluation of earlier urban
diffusion studies. The Gaussian nature of the model is retained.
Evaluation of the revised model has shown that significant improve-
ments have been made. The model reproduces the observed frequency dis-
tributions very well for street-canyon sites. At these sites, hourly
predictions are well correlated (correlation efficient of about 0.6 to
0.7) with observations, and about 80 percent of the calculated values
are within 3 ppm of the observed, which ranged as high as 16 ppm. This
level of uncertainty is about half that found in earlier work before the
model was revised.
iii
-------
CONTENTS
ABSTRACT 1:L1
LIST OF ILLUSTRATIONS ix
LIST OF TABLES xvli
SUMMARY AND CONCLUSIONS xix
I INTRODUCTION 1
A. General Program Objectives 1
B. Review of the First-Stage Model Development 2
1. Description of the Basic Model 2
2. Capabilities of the Basic Model 8
C. Scope of the Current Project 10
II DESCRIPTION OF THE SAN JOSE FIELD PROGRAM 19
A. Background 19
B. Experimental Area 19
1. Selection of San Jose 19
2. Downtown San Jose 20
C. Instrumentation and Operations 24
1. Fixed-Station Measurements 24
2. Mobile Measurements 33
D. Supplementary Data Available from Other Agencies .. 39
1. Traffic Data 39
2. Meteorological Data 41
3. Air Pollutant Monitoring Data 42
v
-------
CONTENTS (Continued)
E. Preparation of the Data for Analysis 42
1. Streetside Data 42
2. Mobile Data 43
3. Data Available for Analysis 44
III STREETSIDE DATA ANALYSIS AND RESULTS 49
A. Background 49
B. Data Analysis 53
C. Results 54
1. Case Study of Hourly Averaged Data for 11-12
December 1970 54
2. Data Stratified by Wind Direction Classes 62
D. Street-Effects Submodel 72
IV ANALYSIS OF HELICOPTER AND SUBMODEL MOBILE VAN DATA 79
A. Introduction 79
B. Data Reduction Techniques 79
C. Results 82
1. Determination of Vehicular Emissions and
Vertical Diffusion of Carbon Monoxide 82
2. Mixing Depth Estimates 91
3. Stability Estimates 94
V INCORPORATION OF THE RESULTS INTO THE URBAN DIFFUSION
MODEL 97
A. Introduction 97
B. Emissions Submodel 97
C. Estimation of Atmospheric Stability 98
vi
-------
CONTENTS (Concluded)
D. Vertical Diffusion Rates 99
E. Mixing Depth 107
F. Local (Street) Effects 107
VI EVALUATION OF THE PERFORMANCE OF THE REVISED MODEL 109
A. Introduction 109
B. Tests of the Subcomponents 109
1. Emissions Submodel 109
2. Mixing Depth Submodel 114
3. Stability Class Submodel 115
4. Street Effects Submodel 116
C. Evaluation of the Composite Model 116
D. Frequency Distribution of Concentrations 137
VII RECOMMENDATIONS 141
ACKNOWLEDGMENTS . „ 143
Appendix A—FIXED-STATION INSTRUMENTATION SYSTEM
Appendix B—VAN INSTRUMENTATION SYSTEM
Appendix C—HELICOPTER INSTRUMENTATION SYSTEM
Appendix D—DATA PROCESSING
Appendix E—PILOT-BALLOON DATA SUMMARY
REFERENCES
vii
-------
ILLUSTRATIONS
Figure 1 Computer Display of Traffic Links for Chicago 3
Figure 2 Hourly Distribution of Traffic for Two Facility
Types in St. Louis ., 4
Figure 3 Vertical Diffusion According to Gaussian
Formulation . .. „ 5
Figure 4 Diagram of Segments Used for Spatial Partitioning
of Emissions 6
Figure 5 Meteorological Input Parameters Required for the
Diffusion Model—Chicago Data (19-25 October 1964) .. 8
Figure 6 Calculated Carbon Monoxide Concentrations (ppm)
for Chi cago 9
Figure 7 Calculated St. Louis Concentration Patterns
for Two Grid Sizes 11
Figure 8 Calculated Concentration Patterns Based on Forecast
of 1990 St. Louis Traffic 12
Figure 9 Calculated St. Louis CAMP Station CO Concentration
Frequency Distribution for 1965 Traffic Conditions;
0800, 1200, and 1800 Hours 13
Figure 10 Calculated St. Louis CAMP Station CO Concentration
Frequency Distribution for 1965 Traffic Conditions;
1-Hour, 8-Hour, and 24-Hour Averages 14
Figure 11 Comparison of Calculated and Observed Hourly CO
Concentrations at the Denver CAMP Station for a
One-Week Period 15
Figure 12 Chicago CAMP Station 17
ix
-------
ILLUSTRATIONS (Continued)
Figure 13
Figure 14
Figure 15
Figure 16
Figure 17
Figure 18
Figure 19
Figure 20
Figure 21
Figure 22
Figure 23
Figure 24
Figure 25
Figure 26
Figure 27
Figure 28
Aerial Photograph of San Jose Showing Intersection
Studied and Helicopter and Van Routes
Map of Area around Intersection of First and
San Antonio Streets
Looking Toward First and San Antonio from First and
San Fernando
21
22
23
Installation of Instrumentation and CO Inlet Tubing 26
Three-Component Wind Sensor 28
Radiation Shield and Ventilation System
for Temperature Sensor
Instrumented Van
29
34
Helicopter and Van Routes (with check points)
around the Central Business District,
San Jose, California 37
Instrumented Helicopter 38
Area Coverage of San Jose Traffic Sensing System ... 40
Indicated Typical Helical Air Flow over a Street
(adapted from Georgii, 1967) 50
Specification for Leeward and Windward Cases on the
Basis of Receptor Location, Street Orientation,
and Wind Direction 52
CO Patterns in San Jose for Three Heights for Early
Morning on 11 December 1970 55
CO Patterns in San Jose for Three Heights at Noon on
11 December 1970 56
CO Patterns in San Jose for Three Heights during
Late Afternoon on 11 December 1970 57
CO Patterns in San Jose for Three Heights during
Late Evening on 11 December 1970 53
x
-------
ILLUSTRATIONS (Continued)
Figure 29 CO Patterns in San Jose for Three Heights During
the Night of 12 December 1970 59
Figure 30 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 045° (±22.5° ) 63
Figure 31 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 090° (±22.5° ) 64
Figure 32 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 135° (±22.5°) 65
Figure 33 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 180° (±22.5°) 66
Figure 34 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 225° (±22. 5° ) 67
Figure 35 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 270° (±22.5°) 68
Figure 36 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 315° (±22.5° ) 69
Figure 37 Average of Horizontal CO Distribution at 3 m
(above) and Vertical CO Profiles (below), for Mean
Rooftop Wind from 360° (±22.5° ) 70
Figure 38 Schematic of Cross-Street Circulation between
Buildings 73
Figure 39 Vertical Profiles of Carbon Monoxide and
Temperature at Spartan Stadium, San Jose,
California 80
Figure 40 Horizontal Traverses of Carbon Monoxide at
Indicated Heights for Box Pattern over Downtown
San Jose, California 83
' xi
-------
ILLUSTRATIONS (Continued)
Figure 41 Lidar-Observed Time-Height Cross Sections of the
Urban Haze Layer over San Jose, California, on
11 December 1970 92
Figure 42 Dependence of the Vertical Diffusion Parameter a
upon Travel Distance for Selected St. Louis
Tracer Tests Conducted by Leighton and Dittmar
(1953) 102
Figure 43 Comparison of Results from Tracer Tests Conducted
in St. Louis over Short Ranges (Leighton and Dittmar,
1953—"L&D*) with those for Intermediate Ranges
(McElroy and Pooler, 1968--"M&P") 103
Figure 44 Comparison of Urban Vertical Dispersion Data with the
Pasquill-Gifford Curves (Adapted from McElroy and
Pooler, 1968) 104
Figure 45 Vertical Diffusion as a Function of Travel Distance
and Stability Category, as Revised for Urban
Conditions 106
Figure 46 Diurnal Emission Patterns for St. Louis Ill
Figure 47 Total Number of Traffic Counts for All Detectors
in Downtown San Jose 112
Figure 48 Calculated and Observed CO Concentrations for
Station 4 at Two Heights for 19 and 20
November and 7 December 1970 117
Figure 49 Calculated and Observed CO Concentrations for
Station 4 at Two Heights for 9, 10, and 11
December 1970 118
Figure 50 Calculated and Observed CO Concentrations for
Station 4 at Two Heights for 14 and 15
December 1970 119
Figure 51 Calculated and Observed CO Concentrations for
Station 5 at Two Heights for 19 and 20
November and 7 December 1970 120
xii
-------
ILLUSTRATIONS (Continued)
Figure 52 Calculated and Observed CO Concentrations for
Station 5 at Two Heights for 9, 10, and 11
December 1970 121
Figure 53 Calculated and Observed CO Concentrations for
Station 5 at Two Heights for 14 and 15
December 1970 122
Figure 54 Calculated and Observed CO Concentrations for
Station 6 at Two Heights for 19 and 20
November and 7 December 1970 123
Figure 55 Calculated and Observed CO Concentrations for
Station 6 at Two Heights for 9, 10, and 11
December 1970 124
Figure 56 Calculated and Observed CO Concentrations for
Station 6 at Two Heights for 14 and 15
December 1970 125
Figure 57 Calculated and Observed CO Concentrations for
Station 7 at Two Heights for 19 and 20
November and 7 December 1970 126
Figure 58 Calculated and Observed CO Concentrations for
Station 7 at Two Heights for 9, 10, and 11
December 1970 127
Figure 59 Calculated and Observed CO Concentrations for
Station 7 at Two Heights for 14 and 15
December 1970 128
Figure 60 Calculated and Observed CO Concentrations for
Station 8 at Two Heights for 19 and 20
November and 7 December 1970 129
Figure 61 Calculated and Observed CO Concentrations for
Station 8 at Two Heights for 9, 10, and 11
December 1970 130
xiii
-------
ILLUSTRATIONS (Continued)
Figure 62 Calculated and Observed CO Concentrations for
Station 8 at Two Heights for 14 and 15
December 1970
Figure 63 Scatter Diagram of Calculated Versus Observed
CO Concentrations for all Five Levels at
Stations 7 and 8 138
Figure 64 Calculated and Observed Frequencies of One-
Hour Average CO Concentration 139
Figure A-l Block Diagram of Fixed-Station Instrumentation
System A~5
Figure A-2 Type A Terminal Sensors and Support System A-8
Figure A-3 Dual Roof Boom—Part of Sensor Support System A-9
Figure A-4 Beckman Carbon Monoxide Analyzer with Remote
Line Coupler A-ll
Figure A-5 Block Diagram of CO Measuring System Using
Beckman Analyzer A-13
Figure A-6 Block Diagram of Temperature System A-17
Figure A-7 Temperature and UVW Electronics «. A-22
Figure A-8 Block Diagram of Remote Line Couplers A-29
Figure A-9 Mini-Computer and Peripheral Devices at
Central Station A-34
Figure A-10 Block Diagram of NOVA Computer A-36
Figure B-l Electronic Console in Van B-4
Figure B-2 Block Diagram of Van Instrumentation System B-5
Figure B-3 Functional Diagram, Van and Helicopter Carbon
Monoxide Measuring System B-7
xiv
-------
ILLUSTRATIONS (Concluded)
Figure C-l Helicopter Instrumentation C-4
Figure D-l Example of Real-Time Summary Printout D-5
Figure D-2
Example of Partially Corrected Data Contained
in One Record of the Tape Generated by Initial
CDC 6400 Processing
D-7
Figure D-3 Example of Information Contained in One Record
of the Basic Data Summary Tape D-9
Figure D-4 Simplified Flow Chart of Data-Averaging Program D-10
Figure D-5 Example of Averaged Data D-ll
Figure D-6 Computer Output Format for Helicopter Temperature
and Carbon Monoxide Profile Data D-16
Figure D-7 Computer Output Format for Helicopter Temperature
and Carbon Monoxide Traverse Data D-17
Figure D-8 Example of Raw Traffic Data Summary D-18
Figure D-9 Magnetic Tape Format for Traffic Data D-20
Figure D-10 Average Link Volumes, 1100-1130 D-25
xv
-------
TABLES
Table 1 Resolution Limitations Imposed by the Analog-to-
Digital Converter 30
Table 2 Master Data Summary—San Jose Field Program,
5 November-15 December 1970 45
Table 3 Wind Direction Sectors for San Jose Street
Stations 77
Table 4 Corresponding Helicopter and Van Legs, Indicating
Weighting Factors for Determination of Average CO
along Van Route Segments 82
Table 5 Values of p in Eq. 22 After Frost (1947) 86
Table 6 Transport Rates of Carbon Monoxide through the
Sides, Top, and Bottom of the Sublayers of the
San Jose Budget Box 88
Table 7 Average Vehicle Speeds in the Downtown Sector of
San Jose for Specified Times During the Period
9-11 December 1970 89
Table 8 Carbon Monoxide Emission Rates (Q) for the Budget
Area Determined from the Mass Budget Analysis and
Traffic Data [with Eq. 27] 90
Table 9 Comparison of Mixing Depth Estimates Obtained
from the Mixing Depth Submodel and the Subjective
Analysis of the Helicopter Profile Data, with the
Lidar-Observed Haze-Layer Structure at San Jose,
California 93
Table 10 Modified Pasquill-Turner Stability Categories
Used with the Diffusion Model (Ludwig et al., 1970)
as a Function of Insolation, Wind Speed, and
Cloud Cover 95
xvii
-------
TABLES (Concluded)
Table 11
Table 12
Table 13
Table 14
Table 15
Table 16
Table A-l
Table A-2
Table A-3
Table A-4
Table B-l
Table D-l
Table D-2
Table E-l
Comparison of Bulk Stability Coefficients
Computed from Eq. (28) with Stability Categories
Determined from the Diffusion Model for the Period
20 November to 11 December 1970, at San Jose,
California
Revised Stability Categories
96
99
Test Conditions for Analyzed Leighton and Dittmar
(1953c) St. Louis Data
Values of Constants in Eq. (34) as a Function of
Atmospheric Stability Category
105
Ambient Carbon Monoxide Background Concentrations ... 134
Correlation Coefficients (r) between Observed and
Predicted Carbon Monoxide Concentrations 136
Manufacturer's Stated Performance Specifications for
the Nondispersing Infrared CO Analyzer A-12
Mode Codes for Indicated Remote Coupling Units A-27
Remote Line Coupler Assigned Scanning Sequence A-31
Characteristics of NOVA Computer A-37
Manufacturer's Stated Performance Specifications
for the Mercuric Oxide Reduction CO Analyzer ...
Traffic at Intersection of First and San Antonio
Streets for Monday, 2 November 1970
Total Traffic Counts from 291 Sensors in Downtown
San Jose for Tuesday, 24 November 1970
Pilot Balloon Data Summary—San Jose State
College (1970) ,
B-9
D-22
D-23
E-4
XVlll
-------
SUMMARY AND CONCLUSIONS
Background—The measurements and original research described in
this report have been used to test and improve the diffusion model
developed during earlier phases of the program (Johnson et al., 1969;
Ludwig et al., 1970). That model, based on existing experimental data
and previous research results, was designed to calculate carbon monoxide
(CO) concentrations from available traffic and meteorological data. It
calculates CO concentrations from the emissions in the city, vertical
diffusion rate, mixing depth, and the wind.
Predictions based on this model in its initial form were compared
with measured data from Continuous Air Monitoring Program (CAMP) stations;
the calculated and observed values often differed significantly in
magnitude, although they tended to have similar trends. The studies
described in this report, show that there were several reasons for this.
Foremost is the fact that local effects in street canyons and around
buildings can sometimes cause CO concentrations to vary by as much as a
factor of 3, or 10 ppm from one side of the street to the other. It is
obvious that any model that did not account for these effects could be
expected to have large errors. One of the principal accomplishments of
the research reported here is a new submodel that does describe these
street-canyon effects and substantially improves the model.
The program has also uncovered and corrected some other short-
comings of the earlier model. These corrections further improve the
model's performance. The nature of the new street effects submodel and
the other changes are discussed below.
xix
-------
Except for wind, none of the inputs to the model are regularly
measured, so submodels have been developed to derive the emissions,
atmospheric stabilities, and mixing depths from measured quantities.
The program reported here has checked the performance of these sub-
models using special measurements made in San Jose, California, during
November and December 1970.
Emissions Submodel Evaluation—San Jose has an extensive computer-
based traffic-monitoring system that provides detailed information on
the traffic in the central business district. This detailed traffic
information allowed the emissions submodel to be applied in this area
with good confidence. The emissions submodel describes the amount of
CO emitted per vehicle-mile as a function of the average vehicle speed.
The traffic flow is known from the monitoring network; the average
vehicle speed was determined from the movements of a project van around
the downtown perimeter. Emissions calculated from the submodel were
compared to independent estimates made from a CO mass budget analysis
that was based on upper level wind measurements and CO concentrations
measured around the central business district with helicopter- and van-
borne instruments. The difficulties with this method include uncertain-
ties in the wind field and possible significant changes of CO emission
rate during the measurement periods, but the results are sufficiently
reliable to uncover serious inaccuracies in the emissions submodel.
The averages of the four cases studied show that the two types of CO
emission estimates agree within a factor of 1.5. There seems to be no
justification for changing the submodel at this time.
Mixing Depth Submodel Evaluation—Vertical profile measurements of
CO concentration and temperature obtained during helicopter flights up
to 1000 m were used to determine mixing depths, and these values were
compared with submodel calculations. Abrupt decreases of CO or increases
xx
-------
in temperature with height usually marked the top of the mixing layer.
On several days the mixing depth was also estimated from measurements
made with a laser radar (lidar) that can detect a sharp reduction in
aerosol concentration at the top of the mixing layer. The submodel
*
mixing depth was within ±15 percent of the mixing depth obtained from
helicopter CO and temperature soundings in six of nine cases studied,'
and within 41 percent in eight. The submodel used to determine the
mixing depth is adequate if representative low-level morning temperature
soundings are available.
Vertical Diffusion Revisions—The basic model is receptor-oriented
and treats the vertical distribution of CO concentration from a con-
tinuous source as Gaussian, where the standard deviation, a , changes
z
with distance downwind of the source and with atmospheric stability.
The dependence of a on downwind distance and stability has been revised
z
to reflect the vertical diffusion observed during urban fluorescent
particle tracer tests in St. Louis, Minneapolis, and Winnipeg (Leighton
and Dittmar, 1952-1953; Pooler, 1966; McElroy and Pooler, 1968). The
dependence of a on downwind distance, x, is of the form
z
b
a = a x ,
z
where a ranges from 1.35 for slightly stable to 0.07 for very unstable
conditions and b from 0.51 to 1.28. The a and b values are such that
a equals 10 m at x equal to 50 m, regardless of stability category.
z
This value of a , 10 m, represents initial mechanical mixing caused by
z
roughness elements near the source.
*
Sometimes there was evidence of two upper bounds to the mixing layer.
In our comparisons we have used the values that are most consistent
with the calculations.
xxi
-------
Stability Submodel Revisions—The methods used to determine
stability category have also been revised to reflect the more stable
conditions observed during the immediate post-sunrise and pre-sunset
hours. The revisions were based on estimates of stability derived from
surface wind speeds and the helicopter-measured vertical temperature
gradient in the lower layers.
Development of the Street-Canyon Effects Submodel—The diffusion
model as constituted at the beginning of the research described in this
report had no provision for effectively describing the behavior of
pollutants in street canyons and around buildings near sources. Street
effects strongly influence concentrations to which pedestrians are ex-
posed. If the model is to be tested properly against observed concentra-
tions, the calculated values must account for the small-scale effects
that can cause large variations in the concentrations around streets and
busy intersections, where most air monitors are located.
An intersection in downtown San Jose was instrumented to obtain the
data necessary to describe and model the street effects. We measured CO
concentrations (at five levels between 3 m and rooftop) and 3-m winds at
seven sites; roof-level winds and vertical temperature gradients were
measured at two of the locations. Data were automatically collected and
recorded on magnetic tape. Fifteen days' data were used to determine the
air circulations and the distributions of CO in the street canyon. The
data analysis showed that rooftop wind direction is the most important
meteorological factor in determining the distribution of CO in the street.
When the roof-level wind blows within about ±45° of the cross-street
direction, a helical circulation develops in the street. At street level
the cross-street component is opposite the roof-level wind direction,
causing a downflow of relatively clean air in front of the "downwind"
buildings that face the roof-level wind, and an upflow across the street.
xxii
-------
The resulting low-level CO concentrations in front of the downwind
buildings are often only half those observed across the street. Cross-
street gradients of CO are quite small for winds approximately parallel
to the street.
A simple model based on physical principles has been developed that
describes the observed distribution of CO in the street canyon. First,
the model presumes that the emissions from the local street traffic are
added to the CO already present in the air that enters from roof level.
These additive concentrations are proportional to the local street
emissions, Q (g m-1 s"1), and inversely proportional to the roof-level
wind speed (u), which is augmented by a small amount to account for the
mechanically induced air movement caused by traffic; 0.5 m s gives
good results.
In front of the downwind buildings the downward air flow produces
relatively little vertical variation of concentration. "Box model"
reasoning indicates that the CO concentration is inversely proportional
to street width, w, and uniform with height. Thus, on the side of the
street where the buildings face the roof-level wind, the CO concentration
added by street sources, AC, is given by:
Q
AC + k
w(u + 0.5)
On the upwind side of the street the model is also based on box
model reasoning, but the volume into which the emissions can be mixed Is
limited by the helical circulation that transports street emissions
toward the buildings and upward. As the air moves from the source, the
volume into which the pollutants are mixed increases, so the concen-
tration is taken to be inversely proportional to the slant distance, r,
between the receptor and the nearest traffic lane. For concentrations
in front of the upwind buildings, the equation is height dependent:
xxiii
-------
AC = k Q
r(u + 0.5)
The constant, k, is the same for both equations. When the wind
blows nearly parallel to the street, the additive concentration, AC,
is described by the average of the values from the two equations and is
the same on both sides of the street.
Overall Performance of the Model—The performance of the total
model, including all its subunits, has been checked against the data
obtained at the fixed CO-measuring sites in downtown San Jose. Since
traffic data are available only for a limited area, the concentrations
arising from emissions at the greater distances, beyond about 2 km, have
been simulated using concentrations measured during helicopter flights or
estimated on the basis of values measured at a height of 34 m at one of
the fixed stations. Data suitable for the evaluations were available
for 70 hours during eight separate days.
The total model was evaluated in two ways. First, the values
predicted by the model were compared with observations on an hour-by-
hour basis; second, the histograms of observed concentration were compared
with those for calculated concentration.
For the hour-by-hour comparisons, we found that the calculated
values (at heights from 3 m to rooftop) were within ±3 ppm of the
observed for nearly 80 percent of the hours at two midblock street-
canyon stations. Observed values ranged from about 1 to 16 ppm. For
two stations near the intersection, the results were not quite so good;
about 75 percent of the calculated cases were within 3 ppm of the observed.
In general, hour-by-hour calculated values followed the observed
trends of CO concentration. This was particularly true of the midblock
street-canyon stations as evidenced by their correlation. For concen-
xxiv
-------
trations at a height of 3 m the correlation coefficients were 0.55 and
0.68. Somewhat poorer results were obtained .near the intersection, where
correlation coefficients ranged from 0.47 to 0.58. Qualitatively, there
appears to be some lag between the calculated and observed values.
Observed concentration changes lag behind changes of wind or stability,
but calculated values, based on current observations, respond immediately.
As noted earlier, the model was used to calculate frequency dis-
tributions of CO concentrations. It is quite important that the model
perform well in this application, because this type of output is important
in planning applications. The frequency of occurrence of CO concentrations
was determined for five geometrically spaced class intervals (1 to 2 ppm,
2 to 4 ppm, etc.) The frequency distribution of 70 observed and 70 cal-
culated 3-m CO concentrations were compared for each station. For the
two midblock sites the two frequency distributions differed by no more
than four cases (i.e., less than 6 percent of the sample) in any class
interval. On the other hand, there was considerable difference (as many
as 21, i.e., 31 percent) between the frequency distributions for stations
near the intersection.
In conclusion, the results have shown that the combined model is
capable of estimating CO concentrations within about ±3 ppm of those
experienced in the downtown streets of a moderate-size city. This level
of uncertainty is about half that found in our earlier work (Ludwig et al.,
1970), indicating that this research has resulted in significant improve-
ments in the model.
xxv
-------
I INTRODUCTION
A. General Program Objectives
The long-term objective of this research program is the development
of a methodology for predicting the concentrations of motor-vehicle-
generated air pollutants throughout an urban area as a function of local
meteorology and the distribution of traffic. The results of this work
are intended primarily to be used as a tool in planning activities to
predict the pollution patterns in any urban region resulting from
planned traffic changes or predicted growth. In addition, the model
can be used in an operational mode for short-term predictions.
Our first goal has been the development of such a methodology for
a quasi-inert pollutant such as carbon monoxide (CO). For the initial
*
development we used available data from five cities having CAMP air
monitoring stations: Chicago, St. Louis, Denver, Cincinnati, and
Washington, D.C. The current status of this program is as follows:
• A working model for CO has been developed.
• A field program, designed to fill gaps in the available
data, has been carried out in San Jose, California.
• On the basis of these data, the initial refinement and
evaluation of the model have been completed.
*
Continuous Air Monitoring Program stations, run by the Federal Air
Pollution Control Office, Environmental Protection Agency.
-------
In this and following sections, the structure and capabilities of the
basic model are reviewed, the results of the field program are reported,
and the progress in the refinement and evaluation of the model is dis-
cussed in detail.
B. Review of the First-Stage Model Development
1. Description of the Basic Model
Instead of an emissions inventory, which is the basic informa-
tion on sources used by diffusion models for other pollutants, the CO
model uses a traffic inventory. An example of a network of traffic
"links," in this case for Chicago, is shown in Figure 1. Each straight
road segment, or link, is assigned an average daily traffic volume, based
upon historical or forecast data furnished by traffic agencies. Each
link is also identified in the computer memory by its length and the
geographical locations of its endpoints and is designated as a particular
road type.
To calculate emissions at a given time, a temporal adjustment
factor such as is illustrated in Figure 2 is first applied to the daily
traffic volumes to obtain traffic volumes for that particular hour. Then
the emission rate, E (g-CO vehicle-mi ), is estimated from the mean
vehicle speed, S (mi h ), by an empirical equation of the form
E = a s~P . (1)
Here a and p are constants that depend on the characteristics of the
emission control devices installed. These can be parameterized reasonably
well by vehicle model year. Estimated values of S are used that depend
upon the type of road and the time of day (whether during peak or off-
peak traffic conditions). The total CO emission for a given traffic
-------
TA-7874-78
FIGURE 1 COMPUTER DISPLAY OF TRAFFIC LINKS FOR CHICAGO
-------
RADIAL EXPRESSWAY
CIRCUMFERENTIAL ARTERIAL
00
HOUR OF DAY — LST
TA-7874-15
FIGURE 2 HOURLY DISTRIBUTION OF TRAFFIC FOR
TWO FACILITY TYPES IN ST. LOUIS
link is found by multiplying E by the length of the link and by the
hourly traffic volume.
Given the network of emissions, the diffusion model is applied
on an area-source basis. The model uses a "Gaussian plume" diffusion
formulation, in which the vertical concentration profile from a cross-
wind line source (such as a road) is assumed to be Gaussian in shape,
as shown schematically in Figure 3. The spread of this vertical con-
centration distribution is characterized by its standard deviation, a .
z
On the basis of experimental data, 0 , which represents the extent of
z
the vertical diffusion, is taken to have the form
CT = ax
z
(2)
-------
HEIGHT
DEPENDS UPON
• TRAVEL DISTANCE
• ATMOSPHERE STABILITY
GAUSSIAN
VERTICAL
CONCENTRATION
PROFILE
LINE
SOURCE
DISTANCE
TA-8563-49
FIGURE 3 VERTICAL DIFFUSION ACCORDING TO GAUSSIAN FORMULATION
where x is the downwind distance and the parameters a and b depend upon
atmospheric stability.
So that nearby sources can be more precisely located, the
model uses a number of area segments spaced at logarithmic upwind range
intervals, as shown in Figure 4. These area segments are oriented in
the direction of the transport wind, and they overlay the emissions
(traffic) network. The traffic links and portions of links falling
within each area segment are identified, the emissions from the indivi-
dual links accumulated, and the total emission then assumed to be re-
*
leased uniformly over the area segment. The contributions from each of
To save computer time, the emissions within the four segments farthest
from the receptor are calculated by a different technique than that
used for the closer segments. The outermost segments are larger than
those nearby, and it was felt that the spatial resolution achieved by
the link assignment technique was not necessary when the emissions were
to be averaged over the entire large area of each outer segment.
-------
16km
RECEPTOR
POINT
1000m
500
250
EXPANDED VIEW OF
ANNULAR SEGMENTS
WITHIN 1 km OF
RECEPTOR
125
62
RECEPTOR
POINT
TA-7874-1R
FIGURE 4 DIAGRAM OF SEGMENTS USED FOR SPATIAL PARTITIONING OF EMISSIONS
-------
the ten area sources to the CO concentration at the receptor are com-
puted individually using the simple formulation given below, and then
added to find the total concentration (C):
1-b
\ / ii
1 -
-2 -1
where Q is the average area emission rate (g m s ) from a particular
A -
segment, u is the transport wind speed, and the subscripts denote dif-
ferent segments (i) and stability classes (j).
""
A simple box model,
_li (4)
uh '
is applied for distant segments when there is a limited mixing depth (h)
determined by the vertical temperature stratification. Under these
conditions, pollutants tend to become uniformly distributed in the ver-
tical after sufficient travel has taken place. We change from the
Gaussian model to the box model at the distance where the two (in their
respective line source formulations) give equal concentration values.
Besides the traffic data, the input variables required for
the model are (1) transport wind direction, (2) transport wind speed,
(3) mixing depth, and (4) atmospheric stability type, as exemplified in
Figure 5 by a week's data for Chicago. The model is designed to be
generally applicable to any city, where conventional (airport) weather
observations might be the only observations available. None of the re-
quired input variables are directly observed. The airport surface wind
speed and direction have been applied as a first approximation to the
transport wind, but suitable adjustments need to be made to account for
urban effects. Separate submodels had to be developed to estimate
-------
(m)
MIXING DEPTH
WIND DIRECTION
(m/s)
WIND SPEED
STABILITY INDEX (solid)
CLOUD COVER (dash)
4000
2000
0
400
200
0
20
10
0
10
i .-''•>- .-•-I
i
J_
0 20 40 60
MON TUES WED
80 100 120 140
THURS FRI SAT
160 HOURS
SUN
TA-8563-50
FIGURE 5 METEOROLOGICAL INPUT PARAMETERS REQUIRED BY THE DIFFUSION
MODEL — CHICAGO DATA (19-25 October 1964)
mixing depths and stability categories from the available airport
observations.
2. Capabilities of the Basic Model
There are two configurations of the basic model: (1) synoptic
and (2) climatological. The former is useful in an operational sense;
the latter is designed as a tool for planning activities. The synoptic
model uses hour-by-hour values of meteorological and traffic data and
calculates hour-average CO concentrations at a point or for a grid of
points. An example of the latter is presented in Figure 6, where the
-------
20
15
10
o
u. 0
o
-
ui
O
t-
-15
-20
, . .,"•-
-20 -15 -10 -50 5 10
DISTANCE EAST OF CITY CENTER — miles
15
20
TA-7874-62
FIGURE 6 CALCULATED CARBON MONOXIDE CONCENTRATIONS (PPM) FOR CHICAGO.
(0700-0800 LST; wind 4 ms"1, 270°; mixing depth 200 m; neutral stability)
-------
concentration calculations, objective contouring, and graphical display
were all controlled by a CDC 6400 computer. The road network is shown
as an underlay. Figure 7 illustrates the telescoping grid or "zoom
capability of the model. In the bottom section of the figure, the grid
spacing was reduced by a factor of ten to depict the detailed concen-
tration pattern in downtown St. Louis. The synoptic model can use
either historical, current, or forecast input data. An example of a
projection of CO concentration patterns in St. Louis, for specified
meteorological conditions, is presented in Figure 8; here forecast
traffic data for the year 1990 were used.
The climatological model is really a trimmed-down version of
the synoptic model, which has been streamlined to reduce the computing
time required for each concentration calculation. This was necessary
because the climatological model is designed to furnish probability or
frequency distributions of CO concentrations at a point, rather than
only a single value, as with the synoptic model. The concentration
frequency distributions are built up from hour-by-hour calculations
using a long series (five years) of climatological data for a given
city. Using this large number of concentration values, one can obtain
various types of frequency distributions, such as for different times
of the day (Figure 9), or for various averaging times (Figure 10).
Thus, for example, the 9Q-percentile concentrations to be expected for
a variety of situations can readily be found.
C. Scope of the Current Project
In evaluating the performance of the model after the first-stage
development, we made extensive comparisons (including regression analyses)
*
Typical computer costs for a 625-point calculation, analysis, and dis-
play, as shown in Figure 6, are about $40.
10
-------
1500-1600 CDT
15 OCTOBER 1964
WIND 310°/1.5m s'1
MIXING DEPTH 1670m
UNSTABLE
-12
-12 -10 -8-6-4-2 0 2 4 6 8 10 12
DISTANCE EAST OF CAMP STATION — miles
TA-7874-26
(a) 1-MILE (1.6 km) GRID SPACING
-1.2
-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0 1.2
DISTANCE EAST OF CAMP STATION — miles
TA-7874-24
(b) 0.1-MILE (0.16 km) GRID SPACING
FIGURE 7 CALCULATED ST. LOUIS CONCENTRATION PATTERNS FOR TWO GRID SIZES
11
-------
o
<
Q_
<
o
oc
o
liJ
o
CO
0
-12 C
-12 -10 -8 -6 -4 -2 0 2468 10 12
DISTANCE EAST OF CAMP STATION — miles
(a) WITHOUT EXHAUST EMISSION CONTROLS
12
10
z 6 h
o
co
a. 2
LL
I *
I -4
-6
co
O
-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12
DISTANCE EAST OF CAMP STATION — miles
TA-7874-61
(b) WITH EXHAUST EMISSION CONTROLS
FIGURE 8 CALCULATED CONCENTRATION PATTERNS BASED ON FORECAST OF 1990
ST. LOUIS TRAFFIC
12
-------
bU
.ASS INTERVAL
w * w
Oo°
u
\ 20
t-
z
u
u 10
cc
a.
0
6O
_> 50
>
ac
£ 40
z
OT 30
V)
<
_i
" 20
i—
z
o 10
UJ
a.
0
60
INTERVAL
4> <"
0 0
en
S 30
_i
u
PERCENT /
— ro
Q0 0 0
| i i | • i i • | i . | . . i i| |
•
••
1 ||
(a) 0800 HOURS
:
i 1 1 1 1 I i . 1 . i 1 .1 l.i
•
-
•
-
1 . ,
(b) 1200 HOURS
-
-
-
-
i 1 i
| , . | . ...| , .,,...., | , .
—
•
-
.1 0.2 0.5
(c) 1800 HOURS
-
-
—
, . .,...7~U-..-
2 5 10 20 5(
CO CONCENTRATION pptn
TA-7874-54
FIGURE 9 CALCULATED ST. LOUIS CAMP STATION CO CONCENTRATION FREQUENCY
DISTRIBUTION FOR 1965 TRAFFIC CONDITIONS; 0800, 1200, AND 1800 HOURS
13
-------
70
5
IT
" 50
z
40
<
o 30
S 20
o
UJ 10
Q-
0
70
^ 60
tr
£ 50
z
-------
of calculated concentrations and those observed at the CAMP stations in
Chicago, St. Louis, Denver, Cincinnati, and Washington, D.C. An example
of such a comparison for hourly concentrations is presented in Figure
11. Generally, the agreement is fair to good, at least in terms of
trends, although there are some instances of poor agreement.
25
20
I 15
oc
I-
UJ
u
1 10
O
O
T [ I I
I I
1 I
DENVER
DEC. 13-19, 1965
'
OBSERVED
CALCULATED
20
MON
40
TUES
60
WED
80
THURS
100 120
FRI
140
SAT
160 HOURS
SUN
TA-8563-51
FIGURE 11 COMPARISON OF CALCULATED AND OBSERVED HOURLY CO CONCENTRATIONS
AT THE DENVER CAMP STATION FOR A ONE-WEEK PERIOD
The weaknesses in the model performance were ascribed to three main
factors:
(1) Deficiencies in the input data, particularly the traffic
data.
15
-------
(2) Weaknesses in the basic model formulation itself and in
the submodels used to estimate some of the input
parameters from conventional meteorological data.
(3) The lack of a suitable street-effects submodel to
handle the problem of local diffusion within street
canyons and to compensate for the influences of
buildings on concentrations measured at nearby receptor
sites.
The last factor is perhaps the most important in terms of achieving
better agreement between calculated concentrations and CAMP-measured
values. This is apparent when one considers that the CAMP stations
were deliberately located to make measurements in areas where concentra-
tions were expected to be high, as in the case of the Chicago station
shown in Figure 12. This station is near a heavily traveled street
and immediately adjacent to a tall building.
The second phase of this research effort, then, has been concerned
with improving and evaluating the model in the areas discussed above.
To do this, we have conducted a field study in San Jose, California,
which has excellent traffic data. In this comprehensive measurement
program, we have collected the special data needed for the following
tasks: (1) improve the submodels for estimation of input parameters,
(2) develop a street-effects submodel, and (3) evaluate the performance
of the basic model and its subunits. The structure of this experimental
program, and its results, will be described in subsequent sections.
16
-------
FIGURE 12 CHICAGO CAMP STATION
-------
II DESCRIPTION OF THE SAN JOSE FIELD PROGRAM
A. Background
As has already been noted, the field program described in this re-
port had been designed to check the operation of the diffusion model
developed earlier (Ludwig et al., 1970). In particular, we were con-
cerned about the ability of the model to cope with small-scale effects
such as occur around buildings and in street canyons. We also wanted
to check some of the fundamental aspects of the model design as they
related to describing the behavior of the emission and dispersion of
pollutants on a scale of a few kilometers. To study all these factors,
we heavily instrumented an intersection in a downtown area and organized
a program for making airborne and street-level measurements around the
perimeter of that downtown area. This section of the report describes
the experimental program only in sufficient detail so that the reader
can understand the results that follow in later sections. Instrumenta-
tion, data processing, and other aspects of the program are described
in greater detail in the appendices to this report.
B. Experimental Area
1. Seclection of San Jose
Two factors entered most heavily into the selection of San
Jose as the location for our initial experiments. San Jose has some
advantages that make it a valuable experimental site. Most important
of these is the City's traffic monitoring network. This network pro-
vides computer-compatible information about traffic in the downtown area.
This information is available with high resolution in both space and
19
-------
time. We considered the availability of these traffic data to be very
important. The nature of these traffic data and of other corollary in-
formation available in San Jose will be discussed further on subsequent
pages. San Jose is also a desirable location with regard to the prac-
tical matters of program management. The experiment we planned was
quite complicated; much of the equipment was newly designed or had been
assembled into unique combinations. The siting arrangements were some-
what unusual and required considerable contacts with building owners
and public officials. All these facts dictated that the experiments be
conducted nearby, so that the resources of the Institute would be avail-
able when we encountered the inevitable difficulties of complex field
experiments.
2. Downtown San Jose
The city of San Jose has a population of about 435,000 (Rand
McNally, 1970). Because of its proximity to San Francisco and Oakland
and because of the plethora of large neighborhood shopping centers, down-
town San Jose is not as developed as many cities of its size. However,
it does have a number of multistory buildings, as can be seen in the
aerial view shown in Figure 13. The location of the streetside ex-
periment at the intersection of First and San Antonio Streets is marked
by the circle on this photograph.
This particular location was selected because it is an area of
relatively uniform building height, compared to most other downtown in-
tersections. When originally selected, the buildings at all four corners
of the intersection were available to be used for the experiment. Sub-
sequently, the building at the south corner of the intersection was ex-
tensively remodeled, which prevented us from making measurements there.
Figure 14 shows a map of the area surrounding the intersection, the
20
-------
to
H
TA-8563-4O
FIGURE 13 AERIAL PHOTOGRAPH OF SAN JOSE SHOWING INTERSECTION STUDIED AND HELICOPTER AND VAN ROUTES
-------
SAN FERNANDO ST.
(BUILDINGS ARE
SHADED AND
NUMBERS OF
FLOORS ARE
INDICATED BY
SMALL NUMERALS)
\
Jl
2
0
? (
1
2
2
2
2
3
3
2
2(1
00
PARK
CENTER
PROJECT
1
1
2
2 ,
2
2
2
2
2
Sffi
2
00
a
2
O
cj
LLJ
00
r 5
0
2
0 5
3
3
3
1
0
2
SAN CARLOS ST.
TA-7874-84R
FIGURE 14 MAP OF AREA AROUND INTERSECTION OF FIRST AND SAN ANTONIO STREETS
locations of our experimental sites, the building heights, and traffic
flow directions. Figure 15 is a picture taken toward the intersection
of First an San Antonio Streets from the top of the seven-story building
at First and San Fernando Streets, Site 9. As can be seen in this
photograph, the area is typical of the downtown regions of many middle-
sized cities.
In addition to the streetside experiment conducted around the
one downtown intersection, large-scale experiments were undertaken to
define the effects from the entire downtown area. These involved
helicopter- and truck-borne instrumentation traveling about the down-
town area. The routes taken by these vehicles around the central down-
town area are marked in Figure 13.
22
-------
ii 111 111 III 111
TA-8563-41
FIGURE 15 LOOKING TOWARD FIRST AND SAN ANTONIO FROM FIRST AND SAN FERNANDO
23
-------
C. Instrumentation and Operations
1. Fixed-Station Measurements
Three types of measurements were considered useful for deter-
mining the street-scale processes operating to disperse the CO emitted
from traffic. Most important, of course, were the CO concentrations at
various locations around the intersection. Next were the winds in the
vicinity. The winds could also be used to determine turbulence inten-
sities, if measured with sufficient temporal resolution. Finally, the
vertical temperature gradients were of some importance in determining
the thermal stability of the air in the street canyons.
In addition to data that we would obtain from measurements,
there were other data that we got from other sources. These included
• the traffic data already mentioned, meteorological observations in the
area, and pollutant concentrations at a location outside downtown San
Jose.
a. Instrumentation
The instrumentation is described in detail in Appendices
A through C. In this section we only outline the basic nature of the
equipment and its principles of operation. Two types of measuring
stations or terminals were employed (see Figure A-l in Appendix A).
At the Type A terminals (Remote Units 1-2 and 3-4), CO was measured at
five different levels; the three wind components were measured at roof
level and at a height of 3 m, and the temperature difference in the
vertical was determined from sensors at five different heights. The
second type of station (Type B terminals) was installed at the other
five locations. At these, CO was also measured, but not temperature
gradients, and only two-component winds at the 3-m level were determined.
24
-------
1) Carbon Monoxide
Nondispersive infrared analyzers were used to measure
the CO concentrations. These devices were manufactured by Beckman
Instruments. They have a 40-inch-long absorption cell and use an optical
filter to remove the effects of water vapor interference. They have a
sensitivity of about 0.5 ppm, and during our operations we found that
they generally maintained their calibration within about this same limit.
Inasmuch as it would have been prohibitively expen-
sive to have provided a CO analyzer for each of the levels sampled, we
used a single analyzer at each site and a manifold system with five in-
lets. This manifold system used solenoid-actuated valves to switch the
CO analyzer from one inlet to another. The four levels that were not
being sampled at any given time were continually purged with an auxiliary
pump. Thus, the CO analyzer pump did not need to exhaust dead air from
the tubing when air from a new level was switched to the instrument.
The 1/4-inch-ID polyethylene tubes used to bring air from the five levels
to the instrumentation were about 200 feet long; the CO analyzer samples
at a rate of 2 liters min , which exhausts the tube volume in about one
minute. The purge-pump flow rate was about three times the CO sampler's
flow rate. Checks of the polyethylene tubing indicate that there was
no interference from this relatively inert material.
Figure 16 shows the CO inlet tubing as it appeared
when installed on the building at the two types of stations. At the end
of each tube was a filter to remove particulates that would interfere
with the proper operation of the CO analyzer. Over these filters,
facing downward, we placed polyethylene bottles to prevent rain from
entering the system.
25
-------
(a) SITE MEASURING CO, TEMPERATURE GRADIENT, AND THREE-COMPONENT WINDS
TA-8563-42
(b) SITE MEASURING CO AND TWO-COMPONENT WINDS
FIGURE 16 INSTALLATION OF INSTRUMENTATION AND CO INLET TUBING
26
-------
2) Winds
Two basic types of wind sensor were used on this
project. At the stations where the winds were measured at only the 3-m
level, we used conventional low-inertia cup and vane sensors. These
have a starting speed of less than 1 mi h"1. They were located on booms
extending about 3 ra from the building, as shown in Figure 16.
Three-dimensional winds were measured at the sites
of Remote Units 1 and 3, at both rooftop and 3-m levels. The units
chosen for these measurements use three low-inertia propeller sensors
whose axes of rotation are orthogonal. The starting speed of these
instruments is about 0.5 mi h . They were placed about 3 m from the
building. The roof-level sensors were above the parapet top; one is
shown in Figure 17. All the wind direction measurements were made
relative to the street direction rather than to north. For those sites
on First Street, a wind along the street, from approximately southeast,
was taken to be a 180° wind.
3) Vertical Temperature Gradients
Platinum wire resistance elements were used to
measure vertical temperature gradients. The actual elements were mounted
in stainless steel tubes 0.125 inch in diameter, which were placed in
silvered, double-walled glass cylinders similar to a Dewar flask. The
inside diameter of these radiation shields is 3.2 cm. Ventilation was
provided at a rate of about 15 ft s by a blower located in a housing
1.1 m from the sensor. The time constant of the aspirated, steel-housed
sensor is about 40 seconds. The whole assembly is shown in Figure 18.
The sensors were suspended at five equally spaced heights
from 3 m to rooftop. The temperature differences between adjacent levels
27
-------
TA-8563-43
FIGURE 17 THREE-COMPONENT WIND SENSOR
28
-------
FIGURE 18 RADIATION SHIELD AND VENTILATION SYSTEM FOR TEMPERATURE SENSOR
were sensed and the voltage signal was electronically amplified. In
this system the temperature difference could be detected to about
±0.01°C. Three ranges of AT were available: 0° to 1°C, 0° to 2°C,
and 0° to 5°C.
b. Control and Data Acquisition System
The block diagram shown in Figure A-l in Appendix A indi-
cates the nine remote units used at the seven different sites. Inputs
1 and 2, and Inputs 3 and 4 were paired at two different sites. The
first-named input of each pair transmitted the three-component winds,
and the second the temperature and CO concentration information. Each
29
-------
remote unit had several sensors connected to it. The signal voltages
from these sensors are selected by a reed relay multiplexer that connects
one signal at a time to the input of an analog-to-digital (AD) converter.
The AD converter has a resolution of 1/128 volt from -1 volt to +1 volt.
Table 1 shows how this relates to the resolution available for each
of the measured parameters. In this table the most commonly used range
settings are marked by asterisks.
Table 1
RESOLUTION LIMITATIONS IMPOSED BY THE ANALOG-TO-DIGITAL CONVERTER
Parameter
Wind speed, components
Wind speed, cup and vane
Wind direction
Temperature difference
Carbon monoxide concentration
Range
-20.3 to +20.3 m s
0 to 6.7 m s
* -1
0 to 13.4 m s
0 to 26.8 m s
0 to 360°
*-l to +1° C
-2 to +2° C
-5 to +5° C
0 to 50 ppm
-1
Resolution
-1
0.16 m s
0.05 m s
-1
-1
O.lms
0.2 m s
2.8°
0.008° C
0.016° C
0.039° C
0.4 ppm
Most commonly used range settings.
A small general-purpose computer with a magnetic tape
recorder was located at one of the stations, marked "l, 2" in Figure
14. Each remote unit was connected to the computer and to the other
remote units through two dc (20-mA) teletype circuits installed by the
telephone company. The computer's transmitter contacts and the
30
-------
remote-unit receiver terminals were connected in series in one circuit,
the command line. The other circuit, the data line, connected all the
remote units' transmitter contacts and the computer's receiver terminals
in series. Thus, the computer transmitted to all remote units simul-
taneously, and a single remote unit transmitted to the computer at any
one time. The computer terminals were in a line-coupler unit that con-
verted the low-voltage (~20 V) operation of the computer's second tele-
type interface to the higher voltage (~100 V) needed to drive the lines.
The sequence of operations was as follows. The computer
sent a command message, consisting of two characters, to all the remote
units. The first character was an address code, which would be recog-
nized by one remote unit. On recognition of the address code, the acti-
vated remote unit began to send its data message to the computer. The
second character of the command message was a sampling-height (level)
code that the remote unit stored temporarily and then decoded to set the
appropriate intake level for the CO analyzer.
The data message returned to the computer from the remote
unit consisted of the level code stored at the time of the preceding
command, followed by the CO measurement and the other measurements that
the unit was programmed to make. The data message ended with its address
code. Each measurement was transmitted as a 7-bit binary number plus
even parity.
After the computer received the complete message, it
checked the address code to verify that it corresponded to the address
that was transmitted. Then another unit was commanded to report. The
timing of command messages was derived from the computer's real-time
clock in a programmed sequence. The sequence was as follows. It started
by interrogating the seven remote units that had CO concentration inputs.
At the time of interrogation, these units were switched to sample another
31
-------
level. This process took about one second per station. While the CO
instruments were equilibrating to the inputs from the new levels, the
three-component wind stations were interrogated. For 52 seconds the
interrogation alternated between the two stations at which these measure-
ments were made. Then the CO measuring stations were again interrogated
in sequence and the complete cycle started once more.
In addition to interrogating the remote units and re-
ceiving their data, the computer had two other functions. It arranged
the data for recording, and it did some averaging and preprocessing.
When sufficient data had been accumulated, it wrote them on the magnetic
tape. After each of the CO levels had been sampled and the data recorded
on magnetic tape, the computer prepared and printed a summary for the 5-
minute period necessary to accumulate those data. This summary was
printed out on a teletype attached to the computer, taking about one
minute to type the summarized information. No data were recorded during
the printout period. An example of this material is shown in Figure D-l
of Appendix D. It proved quite useful for monitoring system performance
in the field.
c. Operations
In general, the equipment was operated only during the
daytime for this test, usually from about 0700 to 1800 PST. The elec-
tronic systems of the carbon monoxide analyzers were left running con-
tinually to maintain their stability. The computer was turned on and
its control and recording functions were started at the beginning of the
day. Each of the sites was then visited and the instruments put on line.
This usually took about 15 minutes. During this time, data were only
being recorded from those instruments that had been turned on. At the
32
-------
end of the operating period a similar situation arises as the instruments
are taken off line and put on "standby" one by one.
After the instruments were all operating in the morning,
the CO analyzers were calibrated. We used one source of zero gas and
one of span gas to calibrate all the instruments. A high-purity helium
was used to zero the instruments. A mixture of 19 ppm (certified
analysis) CO in nitrogen was used to set the span on the equipment. The
calibration results were recorded so that they could be used later in the
data processing to correct the CO readings when the instruments had
drifted significantly. It took about 5 to 10 minutes to calibrate an
individual analyzer. It usually took one to two hours to make the rounds
of the sites and complete the calibration of all the instruments.
The summarized data were monitored periodically, and if
the operator detected any possible equipment malfunctions, they were
checked and corrected where necessary. Of course, if there were equip-
ment problems, they were noted in the station log and accounted for in
subsequent data processing.
The station log was also used to record other significant
events, e.g., street blockage by maintenance crews, idling vehicles near
sampling sites (if noticed), and unusual events such as parades.
2. Mobile Measurements
a. Van
The streetside measurements were supplemented by data
collected using instrumentation in a compact van. Two identically in-
strumented vans were used; one of these is shown in Figure 19. Tem-
perature, wind, and CO concentrations were measured when the vans were
stationary, and temperature and CO concentrations during traverses around
the downtown area.
33
-------
1. Aspirated Radiation Shield for Temperature
Sensor
2. Inlet for CO Analyzer with Rain Shield
3. Telescoping CO Intake Mast (10 m Max. Height)
4. Wind Sensor
5. Telescoping Wind Mast with Plumb Adjustment
(5 m Max. Height)
FIGURE 19 INSTRUMENTED VAN
34
-------
The CO analyzers used in the vans employed a different
principle of measurement than those used at the streetside sites. In
this instrument, heated mercuric oxide is chemically reduced by carbon
monoxide in the atmosphere. This reaction releases mercury vapor, which
is quantitatively detected by its absorption of ultraviolet radiation
from a mercury vapor lamp. The details of this instrument are described
in Appendix B.
The inlet to the CO analyzer was through a length of
polyethylene tubing supported inside a telescoping antenna mounted on
the right side of the front bumper. The inlet could be quickly located
at heights ranging from about 2 to 10 m by extending the supporting
antenna to the desired length.
The temperature sensor was a radiation-shielded, venti-
lated thermistor. It was mounted on the left side at the roof level of
the van, about 2 m high. The readings of both the CO analyzer and the
temperature sensor were recorded on a strip chart recorder. They were
also recorded as analog signals on magnetic tape.
The wind measurements on the van were made at a height
of about 4 m with a propeller and vane device. The wind direction
sensor was oriented consistently with the streetside wind direction
sensors. The low-inertia anemometer had a starting speed of about 0.25
m s"^-, Its signals were recorded in analog form on the same magnetic
tape as the temperature and CO data.
All the equipment was powered from a large bank of storage
batteries. The battery voltage was converted by an inverter to the 117-
volt ac required by the instrumentation. The batteries could power the
instrumentation for about 8 hours without recharging, and could be com-
pletely recharged overnight for the next day's operation.
35
-------
During stationary operation the van could be left un-
attended, but usually the opertor raised and lowered the inlet to dif-
ferent levels at intervals of 10 to 15 minutes. Each change of inlet
height was noted in the log. The chart record was also marked and the
change was noted on the voice channel of the magnetic tape recorder.
The CO analyzer was turned on about an hour before sampling
began. This allowed the temperature of the mercuric oxide cell to sta-
bilize and also ensured stable operation of the electronics. The CO
instrument was calibrated at the beginning of each day's operations.
The CO inlet was kept at a height of about 3.5 m during
mobile operation when the van traversed the perimeter of the downtown
area. The route shown in Figures 13 and 20 was used and the chart records
were marked whenever the van passed one of the numbered points. A voice
record of the traverses was kept on the magnetic tape.
b. Helicopter
The instrumentation used on the helicopter was similar
to that used on the van (except of course that there was no wind sensor)
with the addition of a pressure transducer. The CO analyzer on the
helicopter was identical to those in the vans. The inlet to the instru-
ment was located on the port skid ahead of the helicopter cab.
Temperature was monitored with a thermistor element
mounted forward on the starboard skid. Having the sensors located
ahead of the cab and maintaining a forward speed with the helicopter
made it possible to avoid the effects of the main rotor downwash. The
pressure transducer, of the aneroid potentiometer type, was located in
the cockpit. The outputs of the CO, temperature, and pressure instru-
ments were recorded on a chart recorder.
36
-------
FIGURE 20 HELICOPTER AND VAN ROUTES (with check points) AROUND THE CENTRAL
BUSINESS DISTRICT, SAN JOSE, CALIFORNIA
37
-------
All the equipment was designed so that it could be in-
stalled or removed quickly from the Hughes 300 helicopter that was
chartered for this program. The helicopter with the equipment installed
is shown in Figures 21 and C-l (Appendix C).
FIGURE 21 INSTRUMENTED HELICOPTER
Operations of the mercuric oxide CO analyzer in the heli-
copter impose some special problems, because this instrument is sensi-
tive to the mass flow rate of the sampled air through the device. The
flow rate was sensitive to changes in altitude, so it was necessary for
the operator to make frequent flow rate adjustments. As with the van;
it was necessary to prewarm the CO analyzer for about an hour before use..
The CO analyzer was calibrated at the beginning of each flight. The
calibration technique was much the same as that used with the analyzers
in the vans.
Two basic kinds of measurements were made with the instru-
mented helicopter: vertical profiles, and horizontal traverses around
38
-------
the downtown area. The vertical profiles were usually obtained at a
site near Spartan Stadium, about 3 km southeast of the downtown center,
from heights of about 15 to 1000 m, although on several occasions pro-
files were obtained over the intersection of First and San Antonio
*
Streets. Restrictions imposed by the presence of the city buildings
and the approach pattern for the San Jose airport limited the downtown
observations to the height range between 60 and 300 m.
The horizontal traverses around the downtown perimeter
were made along the track shown in Figures 13 and 20. The first traverse was
at a height of about 60 m. Subsequent traverses were made at heights of
90 m, 150 m, and also at 300 m, if the top of the mixing layer had not
been penetrated below 150 m. The top of the mixing layer was usually
identified by an abrupt decline in CO concentration. During all the
flights the operator annotated the chart records and kept a position log.
D. Supplementary Data Available from Other Agencies
1. Traffic Data
As noted earlier, the detailed traffic data available for San
Jose played a very important part in the selection of San Jose as the
location for our experiments. The area covered by the computer-
controlled signal network in San Jose has two main parts: (1) a four-
by-six-block rectangular grid, covering the 'heart" of the downtown
area, and (2) a 3.5-mile "panhandle, ' along one of the major streets
which connects to the grid. Figure 22 is a schematic representation
of the area covered by the network. Within the grid area, 100-percent
coverage of the traffic movement is provided by approximately 225 mag-
netic vehicle detectors, located in every lane of all street links.
39
-------
JULIAN
Partial Coverage Only
SAN CARLOS
SAN
SALVADOR
TA-8563-45
FIGURE 22 AREA COVERAGE OF SAN JOSE TRAFFIC SENSING SYSTEM
-------
Each detector is connected through an interface to a computer
(IBM 1800) operated by the city of San Jose. Pulses generated by a
vehicle passing over any detector are monitored by the computer and
stored in terms of a sensor identification number and pulse time tag.
Program options permit summarizing these raw data as total vehicle counts
obtained for each sensor during selected time intervals. For this
project, 5-minute intervals were selected. Although longer periods might
have been sufficient, the 5-minute raw summaries were possible at very
little extra cost (all data were generated and processed by machine),
and the more conservative interval was chosen. The 5-minute volume
histories were generated by the traffic control computer as both printed
listing and punched card output.
The punched cards became the input for further processing on an
SRI computer (CDC 6400). The description of this processing appears in
Appendix D (Data Processing) of this report.
;
Traffic data were collected in three phases. During the first,
from 2 to 23 November 1970, traffic volumes were obtained only from the
links adjacent to the intersection of San Antonio and First Streets.
Ten sensors were monitored in all. The second phase, from 23 November
to 11 December 1970, was directed toward monitoring all operational de-
tectors in the system (291). In both phases, data were collected on
weekdays during rush hours (0645-0830, 1100-1300, 1600-1800 PST) only.
The final phase was a two-day effort, 14 and 15 December.
Data were collected from all sensors during the entire period of 0645
to 1800 PST.
2. Meteorological Data
Conventional meteorological surface observations are avail-
able from San Jose Municipal Airport, about 5 km northwest of downtown
41
-------
San Jose. The Navy also makes regular meteorological observations at
Moffett Field, 17 km to the northwest. Radiosondes are released twice
daily from the Oakland Airport, about 50 km northwest of San Jose. In
general we used only the observations from the San Jose Airport.
The San Jose State College Meteorology Department has several
recording instruments, including wind speed and direction, located on
top of a seven-story building about 0.7 km east of the intersection of
First and San Antonio. On several occasions, College personnel made
pilot-balloon measurements of the upper level winds (see Appendix E).
3. Air Pollutant Monitoring Data
The Bay Area Air Pollution Control District maintains a pollu-
tion monitoring station about 2.3 km south-southeast of the downtown
area. A variety of pollutants, among them carbon monoxide, are moni-
tored. The carbon monoxide analyzer is of the nondispersive infrared
type, as are those used at our streetside sites.
E. Preparation of the Data for Analysis
The details of the data processing for analysis are given in
Appendix D, and later sections of the body of the report deal with some
of the analyses that have been prepared from the data. In this section,
some of the corrections that were made on the data will be enumerated
and the preprocessing briefly described.
1. Streetside Data
The streetside data were quite extensive, and it was necessary
to perform some editing and condensation. The data handling began by
translating the binary data as originally recorded to the binary-coded
42
-------
decimal used in subsequent processing. At the time of translation,
those measurements made during periods of calibration (determined from
equipment status codes) were eliminated from the records. The recorded
voltages were converted to engineering units at the same time. The con-
version factors used were also based on recorded equipment status codes.
The next step involved further corrections based on records
kept by operators during the experiment. These corrections included
those made on the basis of CO analyzer calibrations, or to convert wind
components to a common frame of reference.
The corrections were followed by a consolidation of the data
onto a new magnetic tape. Each record contained information for a 5-
minute period. All the CO concentration and temperature data appeared
in this record, as did all the cup-and-vane wind data. The three-
component wind measurements were also consolidated by summarizing these
data for each period. The reduced wind information was recorded as:
(1) the number of observations of each component during the period, (2)
the algebraic sums of each observed component magnitude during the
period, and (3) the sums of the squares of the observed component mag-
nitudes during the period. This information can be used to calculate
mean values and their standard deviations for the individual periods
and for combinations of individual periods.
2. Mobile Data
The mobile data were generally recorded on charts and trans-
ferred to computer-compatible form by hand. Much of the editing and
correcting was done by hand during this part of the processing.
The CO data from the stationary vans were recorded on magnetic
tape. The tape contains CO concentrations at 1-minute intervals, van
43
-------
locations, inlet height, date and time. The data collected during the
mobile traverses were punched on cards that give locations on .the city
perimeter, average CO concentrations over the specified route segments,
date, and time. Similar records of the helicopter data are available;
these include heights in addition to the other information.
3. Data Available for Analysis
All of the times that data were collected are shown in Table 2.
Only those data collected after 18 November 1970 were processed and
analyzed for this report. Prior to 19 November, the magnetic tape
recorder was not working, so the data are available only as paper copy
summaries or punched paper tape. Because data in these forms are not
readily suitable for electronic data-processing techniques and because
we had sufficient data for analysis recorded on magnetic tape, we chose
to exclude the earlier period from our studies.
44
-------
Table 2
MASTER DATA SUMMARY—SAN JOSE FIELD PROGRAM, 5 NOVEMBER-15 DECEMBER 1970
Date
(1970)
5 November
(Thursday)
6 November
(Friday)
9 November
(Monday)
10 November
(Tuesday)
11 November
(Wednesday)
12 November
(Thursday)
Traffic
0645-0830
1600-1800
0645-0830
1100-1300
1600- 1800
0645-0830
1100-1300
1600-1800
0645-0830
1100-1300
1600-1800
0645-0830
1600-1800
Street
Stations
1220-1300
0901-1035
0805-1208
1430- ?
—
0745-0937
0948-1124
1240-1650
0625-1030
Van A
__
—
—
—
—
Van B
__
—
—
—
—
Helicopter
1408-1530 (V)
1435-1445 (V)
1448-1507 (H)
1000-1025 (H)
—
1713-1720 (V)
1720-1750 (H)
1750-1800 (V)
Lidar
8
1337-1552 (F)
—
—
—
—
Time
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
Clouds
23®
25®
25®
40®
40® 100®
150) 30(0)
5® 20®
5® 20®
5® 40®
23® 80®
60® 90®
600) 100®
40®
35®
35®
6® 14®
8® 19®
25®
20®
O
30®
60® 120®
30® 60®
20® 32®
28®
20®
O
O
0
0
Visibility*
(mi)
8R--
15
15
20
35
12
5R-
5RW
10R—
15
10
5HK
6HK
5R-
7
5FK
5FK
7
7
7
3HK
6HK
10
15
12
15
15
15
50
50
Temp/dewpoint
(°F)
63/62
68/60
72/62
67/59
65/57
59/57
60/58
61/60
62/60
63/59
59/59
65/60
69/61
68/64
67/63
60/60
61/59
64/57
67/57
68/58
58/58
65/57
70/58
69/59
67/60
54/52
60/52
64/51
68/48
63/40
Wind
(deg/kt)
190/11
170/11
190/11
290/13
340/04
160/10
140/12
140/10
150/04
140/03
030/04
180/03
010/08
320/09
340/08
320/07
310/06
040/04
090/04
350/07
120/04
130/15
160/13
200/10
220/10
300/10
320/12
320/14
320/15
320/11
en
The alphabetical symbols indicate weather conditions as follows: R—rain, RW—rain shower, H—haze, K—smoke, F—fog. The minus and double minus signs
following the symbols indicate light and very light weather conditions, respectively.
Periods of data collection are given in PST throughout.
Suffixes (V) and (H) denote vertical and horizontal helicopter profiles.
§
Suffixes (F) and (M) denote fixed and mobile operational modes.
-------
Table 2 (Continued)
Date
(1970)
13 November
(Friday)
16 November
(Monday)
17 November
(Tuesday)
18 November
(Wednesday)
19 November
(Thursday)
20 November
(Friday)
23 November
(Monday)
Traffic
1100-1300
1600-1800
0645-0830
1100-1300
1600-1800
0645-0830
1100-1300
1600-1800
1100-1300
1600-1800
0645-0830
1100-1300
1600-1800
0645-0830
1600-1800
0645-0830
1100-1300
1600-1800
Street
Stations
0730-1113
1128-1410
0730-1439
0850-1434
0740-1211
*
0725-1208
0735-0755
0850-1045
1105-1800
0735-0800
0831-1245
Van A
—
0725-1453 (F)
0815-1223 (F)
1230-1315 (M)
—
0822-1240 (F)
0718-1530 (F)
Van B
—
—
—
—
—
Helicopter
__
—
0827-0837 (V)
—
0854-0903 (V)
0903-0934 (H)
0938-0948 (V)
1654-1713 (V)
1719-1744 (H)
1751-1801 (V)
0807-0825 (V)
0825-0901 (H)
0903-0921 (V)
Lidar
1625-1632 (F)
1540-1627 (F)
—
—
1540-1815 (M)
—
Time
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
100O
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
Clouds
O
0
O
O
0
16®
0
O
O
O
O
4®
0
0
O
wox
WIX
O
O
O
O
0
O
0
0
0
0
O
0
1500)
40-®
2500
2500
250CD
100®
Visibility
(mi)
50
50
50
50
50
10
10
6HK
6HK
10
4HK
IF
2F
5HK
6HK
1/4F
3/8F
3F
6HK
15
20
30
50
30
30
10
5HK
5HK
5HK
5HK
7
5HK
5HK
5HK
5HK
Temp/dewpoint
(°F)
52/40
64/36
68/37
70/36
63/39
47/42
60/46
64/55
66/56
63/57
46/46
55/55
59/55
66/55
63/55
52/52
54/54
57/56
65/51
63/55
47/46
64/43
68/42
69/43
62/47
44/43
57/46
66/46
—
61/45
47/43
58/45
64/45
68/48
63/47
Wind
(deg/kt)
320/06
120/04
240/07
300/12
270/06
120/04
360/06
340/07
330/09
330/10
120/03
340/04
100/04
350/06
340/08
350/04
310/06
320/06
340/09
350/08
140/04
230/07
320/08
300/11
330/08
180/04
150/05
130/04
330/08
320/08
000/00
090/04
320/04
340/04
320/08
Start of data recording on digital magnetic tape.
Start of data collection from all traffic detectors in downtown area; previous data are for First and San Antonio Streets only.
-------
Table 2 (Continued)
Date
( 1970)
24 November
(Tuesday)
25 November
(Wednesday)
30 November
(Monday)
1 December
(Tuesday)
2 December
(Wednesday)
3 December
(Thursday)
4 December
(Friday)
7 December
(Monday)
Traffic
0645-0845
1100-1300
1600-1800
0645-0845
1600-1800
1100-1300
1600- 1800
0645-0845
1100-1300
1600-1800
1100-1300
1600-1800
0645-0845
1100-1300
1600-1800
0645-0845
1100-1300
1600-1800
0645-0845
1100-1300
1600-1800
Street
Stations
0850-1325
0745-0955
1020-1226
0725-1045
1058-1329
—
-_
0753-0830
0942-1332
—
0741-1200
1230-1800
Van A
0819-1142 (F)
1142-1310 (M)
0819-1630 (F)
0735-1620 (F)
—
—
0731-0824 (M)
0827-1514 (F)
—
0803-1245 (F)
1254-1410 (M)
1449-1630 (F)
1636-1740 (M)
Van B
—
—
—
—
~
0849-1426 (F)
1448-1703 (F)
Helicopter
1217-1233 (V)
1233-1243 (H)
0824-0849 (H)
0855-0905 (V)
—
—
0753-0813 (V)
0817-0856 (H)
—
1226-1246 (V)
1247-1308 (H)
1652-1652 (V)
1701-1728 (H)
Lidar
1430-1507 (M)
—
—
—
—
—
1240-1410 (F)*
Time
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
Clouds
25®
17® 1500
400 150®
45®
50®
23®
25®
25®
25®
25®
230 38®
300 45®
300 45®
30®
10®
300 100®
300 45®
200 45®
200 45®
25®
40®
40®
38®
40®
400
45®
60®
40®
40®
28®
30®
35®
35® 50®
35®
25®
6®
7®
200®
100®
450 55®
•Visibility
(mi)
4HK
5HK
12
15
15
7R-
15
10RW-
15
10
12RW-
12
12
7
7
10
10RW-
20
15
7RW-
10RW
15
15RW
15
30
20
20
15
15
10
7R-
10
12
12RWN
15
3F
3HK
5HK
15
15
Start of 90° lidar observations for height-time displays; previous data are RHI elevation scans.
Temp/dewpoint
(°F)
48/44
62/50
70/51
70/51
69/52
61/57
62/56
64/57
65/57
62/56
51/46
54/47
59/47
58/49
51/50
53/47
56/47
57/50
59/48
57/49
51/44
55/41
54/39
55/42
52/37
47/37
51/36
55/36
57/39
58/43
56/51
58/51
61/53
64/54
63/52
57/52
59/52
65/52
68/46
62/50
Wind
(deg/kt)
080/04
130/05
030/05
260/05
320/05
130/10
140/07
140/06
210/08
320/08
150/13
160/12
160/16
190/16
150/05
140/09
160/13
170/12
180/13
160/09
180/04
250/10
290/18
240/15
210/04
140/13
160/14
150/12
140/15
150/13
130/10
160/10
140/11
160/10
160/08
320/07
180/03
140/13
160/10
040/06
-------
Table 2 (Concluded)
Date
(1970)
8 December
(Tuesday)
9 December
(Wednesday)
10 December
(Thursday)
11 December
(Friday)
12 December
(Saturday)
14 December
(Monday)
15 December
(Tuesday)
Traffic
0645-0845
1100-1300
1600-1800
0645-0845
1100-1300
1600-1800
0645-0845
1100-1300
1600-1800
0645-0845
1100-1300
1600-1800
0645-1800
0645-1800
Street
Stations
0721-1320
0715-1059
1105-1800
0745-0840
0845-1800
0800-2400
0000- 1000
1120-1225
1231-1800
0740-1800
Van A
0750-0906 (M)
0701-0740 (F)
0740-0853 (M)
1005-1115 (F)
1132-1252 (M)
1313-1341 (F)
1627-1805 (M)
0746-0858 (M)
0900-1140 (F)
1145-1247 (M)
1317-1420 (F)
1610-1749 (M)
0814-1400
—
1255-1320 (F)
1400-1800 (F)
0742-1800 (F)
Van B
__
0700-1357 (F)
1700-1748 (F)
0720-1422 (F)
1640-1802 (F)
0742-0851 (M)
0909-1130 (F)
1137-1255 (M)
1409-1628 (F)
1631-1851 (M)
—
—
—
Helicopter
0758-0817 (V)
0819-0841 (H)
0747-0806 (V)
08O6-0832 (H)
1145-1206 (V)
1206-1223 (H)
1642-1702 (V)
1707-1731 (H)
0758-0808 (V)
0811-0843 (H)
1150-1210 (V)
1211-1239 (H)
1650-1710 (V)
1711-1740 (H)
0802-0820 (V)
0826-0837 (H)
1156-1215 (V)
1222-1239 (H)
1641-1702 (V)
1702-1725 (H)
—
—
—
Lidar
_.
0746-1733 (F)
0730-0825 (F)
1433-1539 (F)
1632-1746 (F)
0725-1800 (F)
—
—
—
Time
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
0700
1000
1200
1400
1700
Clouds
20® 50®
20CD 45®
10® 23®
32®
25®
40®
400)
40®
40®
200®
®
®
250®
®
®
8
59/53
61/54
62/55
62/52
56/50
45/39
55/45
57/43
58/41
53/42
40/36
50/41
56/40
59/41
56/41
44/39
52/46
56/45
60/42
56/38
36/31
49/37
53/39
56/42
51/42
42/38
48/40
54/42
51/45
57/45
43/39
52/44
59/44
63/46
58/45
Wind
(deg/kt)
160/07
140/13
140/15
190/12
340/09
120/03
320/07
330/11
350/10
340/08
200/03
330/03
320/05
350/07
300/08
140/04
020/03
330/10
340/09
280/09
310/05
020/03
320/04
320/08
320/07
160/04
160/08
150/06
060/04
110/04
210/04
130/08
140/14
150/15
140/16
00
-------
Ill STREETSIDE DATA ANALYSIS AND RESULTS
A. Background
Because of the finite spacing of sources within a city, area-source
simulation such as is used in the intraurban model is best applied at
scales above a certain lower limit. This minimum spatial scale can be
considered to be on the order of a city block, hence the choice of 62 m
as the finest resolution in the intraurban model. For shorter source-
receptor distances, an alternative technique is needed.
Additional complications arise because, contrary to the usual situa-
tion in nonurban diffusion studies, the scale of the largest urban
roughness elements (buildings and so on) is very large compared to the
local scales of emission and reception. This means that the aerodynamic
effects of structures become important.
Models that do not include the effects of microscale diffusion will
normally undercalculate concentrations in comparison with those measured
at CAMP Stations, which are often located near streets. For example,
the model used by Ott et al. (1967) gave average concentrations that
amounted to 36 percent of the CAMP average.
The street effects have great importance for two reasons. First,
they must be considered if we are to use existing data to verify the
performance of our model. Most available observations are taken near
streets in downtown areas where local effects are likely to be signifi-
cant. The second reason for the importance of the street effects is
that they contribute substantially to those concentrations to which
large parts of the population are exposed.
49
-------
Current knowledge about street effects on CO concentrations is
based largely upon the extensive measurements of Georgii et al. (1967)
in Frankfurt/Main, Germany; McCormick and Xintaras (1962), Schnelle
et al. (1969), and Rouse (1951) have also made experimental contribu-
tions in this area.
Georgii's experiment involved extensive measurements of CO concen-
trations and wind speeds at different levels above three different
streets in built-up areas, along with occasional traffic counts. A
major finding was that the CO concentrations on the leeward sides of
buildings were considerably higher than those on the windward sides,
implying a helical cross-street circulation component near the surface
in the opposite direction from the roof-level wind (see Figure 23).
In addition, the averaged data showed that (1) the vertical concentration
profiles on either side of the street assume an exponential form, (2) the
TA-7874-20
FIGURE 23 INDICATED TYPICAL HELICAL AIR FLOW OVER A STREET (adapted from
Georgii, 1967)
50
-------
mode of air circulation above the street apparently changes when the roof-
_ i
top wind speed exceeds about 2ms , and (3) the concentrations are ex-
ponentially related to traffic density. Examination of the measurements
reported by Schnelle et al. (1969) also indicates general agreement with
(1) and (2) above; their data are insufficient for verifying (3).
During the first year's effort (Johnson et al., 1969), an attempt
was made to develop a street submodel, based largely upon Georgia's
findings. The final equations, applicable to a street canyon, were
C = C exp 29(Q)I1 - z/z M (5)
CL = Cb expK45.6 + 4.68 U)(Q)(l - z/zjl , (6)
where
C = streetside concentration on the windward side of the
W
building
C = streetside concentration on the leeward side of the
L
building
C = background urban concentration
b
Q = line source emission rate (g m s )
U = rooftop wind speed (m s )
z/z = ratio of receptor height to building height.
The specification of leeward and windward cases was carried out
according to the quadrant including the observed wind direction, as
illustrated in Figure 24. For intermediate wind directions, the
concentration profile was taken to be the average of that given by Eqs.
(5) and (6).
51
-------
10
LEEWARD
8 \
210— — : * "*"* ~
f*
V
WINDWARD
+ 315°
\
\
/ \
WINDWARD
V-
/
/
+ 225°
I
!;;$ . •
IS3
II
X
l
i
i
i
|3 + 45
/
/
A
|g) LEEWARD
\y
\
\
|3 + 135
(3 + 180
INTERSECTION
SINGLE STREET
TA-7874-22
FIGURE 24 SPECIFICATION FOR LEEWARD AND WINDWARD CASES ON THE BASIS
OF RECEPTOR LOCATION, STREET ORIENTATION, AND WIND DIRECTION
Although these formulas give CO profiles that fit Georgii's averaged
data reasonably well, the results were not satisfactory when the submodel
was applied to the CAMP data. There are several shortcomings to this
methodology that may be to blame: (1) the equations are strictly em-
pirical, rather than physical; (2) the concentration component due to
locally generated emissions should be additive to the background concen-
tration, rather than proportional to it; (3) there is no provision for
a variable distance from the roadway; and (4) the concentration at the
receptor level z = z may not be equal to the background concentration.
r
Because of a number of problems encountered in applying the empirical
model on a general basis, it has not been incorporated into the diffu-
sion model. A new technique has been developed with the aid of the data
from the San Jose street experiment. The data analysis, the results,
and the development of the new methodology are described in the next
sections.
52
-------
B. Data Analysis
The basic processing of the data has been discussed in a preceding
section of this report, and is described in detail in Appendix D. Most
of this processing was directed toward simplifying the stratification
and class-averaging of the data. The first step was to calculate hour
averages and standard deviations of all measured and derived variables
(CO concentrations, wind speeds, wind directions, turbulence intensities,
and vertical temperature differences) for all time periods when the data
were available on magnetic tape. These averages were reduced to a format
similar to that shown in Figure D-5 of Appendix D. CO concentrations
for the 7.5-m and 12-m levels were obtained by interpolation.
Contour analyses were prepared for these, hour-averaged data for the
levels of 3, 7.5, and 12 m for each hour for which there were a sufficient
number of cases to constitute a representative average. This basic set
of analyses has been helpful in determining the general types of concen-
tration and circulation patterns, in establishing the criteria for strati-
fying the data into classes, and in judging the representativeness of
patterns subsequently shown by the analyses of the stratified data.
In selecting bases for stratification of the data, we were also
guided by the work of Georgii et al. (1967). Their studies, as well as
our hour-averaged data, suggested that wind direction, wind speed, and
CO emission rates in the street were the most significant variables in-
fluencing the distribution of CO within a street canyon. Accordingly,
the data were stratified in three forms:
(1) By wind direction (±22.5° sectors) only
(2) By wind direction and hour of the day
(3) By wind direction, wind speed, and traffic volume in the
downtown area.
53
-------
As with the hour-averaged data, these three sets of stratified data
were plotted and analyzed for three different heights. In addition,
CO profiles for all seven stations were graphed for the first two sets
of data.
C. Results
To illustrate the basic nature of the wind circulations and CO
patterns observed within the street canyon during the San Jose experi-
ment, we have selected two sets of data to be presented as examples from
among the many analyses that have been prepared and studied. The first
set is a case study of hour-averaged data for selected hours throughout
a 24-hour period on 11-12 December 1970. The second set consists of the
total body of data stratified into eight wind-direction classes. When
viewing these analyses, one should keep in mind that they are based on
only a few widely separated data points, and that in some cases there is
little justification for drawing the concentration contours in the region
of the intersection itself. Nevertheless, this was done to aid in the
interpretation.
1. Case Study of Hourly Averaged Data for 11-12 December 1970
The weather during the period of 11-12 December (see Table 2)
was characterized by scattered low clouds during the morning hours of
11 December, becoming clear after noon and for the rest of the period.
Winds were relatively light throughout the period; the wind direction
was quite variable.
Horizontal CO distributions for three heights (3, 7.5, and 12 m)
for the periods 0800-0900, 1200-1300, 1700-1800, 2300-2400, and 0300-0400
PST are presented in Figures 25 to 29. The first three periods
were selected to coincide with the peak traffic periods. Vehicle
54
-------
0800-0900 PST
01
en
CO Concentration
(ppm)
6.3
FIRST
STREET
7.5m
9
9.3
12m
TA-8563-55
FIGURE 25 CO PATTERNS IN SAN JOSE FOR THREE HEIGHTS FOR EARLY MORNING ON 11 DECEMBER 1970
-------
1200-1300 PST
05
CO Concentration
i \
ippuif f\ e
^"**"1*-»^w ^'^ '
Wind Speed 9.1
(cm/s) *-(50)
980
veh/hr
*^~ ^-««
*^^ --^
/
SAN 315 BV-X'
ANTONIO ;"b
STREET veh/nr
7.8 g
9
10
9.7"
(20)
'
\
\
9.1
9 (60)
8
765 4 3.5 (60)
/III 302 veh/nr
J 5 4
1
> s
8
0 6
1015
veh/hr
6.3
(100)
_ 6.0
(100)
FIRST
STREET
3m
FIRST
STREET
7.5m
FIRST
STREET
12m
TA-8563-56
FIGURE 26 CO PATTERNS IN SAN JOSE FOR THREE HEIGHTS AT NOON ON 11 DECEMBER 1970
-------
1700-1800 PST
CO Concentration
Wind Speed •
(cm/s)
-10.3
•(90)
10
SAN
ANTONIO 294 veh/hr)
STREET
3m
7.5m
12m
TA-8563-57
FIGURE 27 CO PATTERNS IN SAN JOSE FOR THREE HEIGHTS DURING LATE AFTERNOON ON 11 DECEMBER 1970
-------
2300-2400 PST
CO Concentration
(ppm)
oo
3m
7.5m
12m
TA-8563-58
FIGURE 28 CO PATTERNS IN SAN JOSE FOR THREE HEIGHTS DURING LATE EVENING ON 11 DECEMBER 1970
-------
0300-0400 PST
CO Concentration
(ppm) '\w
11.0
Wind Speed _^(30)"
(cm/s)
11
J- -,
SAN
ANTONIO
STREET
11.9^
(20)
/
11.2
"(10)
10 9 8.6 (30)
fir
L
\
V
V
\
vr
FIRST
STREET
E9
9.3
(0)
to
11
11.3
(50)
10.6
10.8
10 „
3m
FIRST
STREET
7.5m
rr
9 8
9.7
10
10.8
11
12m
TA-8563-59
FIGURE 29 CO PATTERNS IN SAN JOSE FOR THREE HEIGHTS DURING THE NIGHT OF 12 DECEMBER 1970
-------
counts as obtained from the San Jose traffic monitoring network are
indicated for First Street and San Antonio Street on the first three
analyses. Unfortunately, no traffic data were available after 1800 PST,
since the monitoring network was not in operation after that time.
*
Wind directions and speeds are plotted for six stations for
the 3-m levels, and for Stations 2 and 4 (the only stations having upper-
level winds) on the 12-m analyses. These winds were actually observed
at heights of 22 and 15 m, respectively. It turns out that the rooftop
wind, as characterized by that at Station 2, furnishes the best indicator
of the type of street-level CO pattern observed at any given time.
t
For the first period, there is a light easterly rooftop wind,
which sets up a leeward-windward circulation for Stations 7 and 8. The
observed CO concentration at Station 7 at 3 m is almost double that at
Station 8. However, the 3-m wind directions at these two stations are
opposite to what would be expected from the primary vortex circulation.
This occurred frequently; the most likely explanation for this apparent
anomaly is that the wind sensors, which are located 3 m away from the
buildings as well as 3 m above the street, are influenced by the secon-
dary corner vortices, which are depicted in Figure 38 (presented later
in this section).
•x-
The data from Station 9 were not included in these analyses because
of its separated location, at the intersection of First and San
Fernando (see Figure 14). This station was mainly used as a control
to ensure that the First and San Antonio data were representative of
other intersections, and to furnish data up to higher levels (35 m)
for comparison with the helicopter data.
t
First Street is oriented at an azimuth direction of 330 -150 . All
observed winds and discussions of these winds are relative to the
First Street direction, which for convenience is taken to be N-S.
60
-------
Another feature worth noting is the difference between the
First Street concentrations and that at Station 4 on San Antonio Street,
which has only about one-third the traffic present on First Street.
Also, the concentrations at most stations decrease with height, as would
be expected, with much less horizontal variation at the 12-m level.
The traffic volumes for the second period, 1200-1300 PST
(Figure 26), are about 30 percent higher than for the morning period,
but the observed concentrations are about the same in magnitude. This
probably reflects the increase in observed wind speed rather than any
changes in atmospheric stability. The rooftop wind direction has shifted
to NNW, and the concentration gradient at First Street between Stations
7 and 8 has reversed from the earlier case.
This CO pattern stays approximately the same for the late
afternoon case, 1700-1800 PST (Figure 27). The rooftop wind has
shifted slightly to a westerly direction, and the leeward situation at
Station 8 is still apparent. The traffic volumes are about the same as
for the early morning case, but concentrations are slightly higher.
Figure 28 depicts an unusual situation of high CO concen-
trations during what is normally an off-peak traffic period, at 2300-
2400 PST. However, at this time on Friday nights, a local custom of
"cruising" of the downtown business district by the young people of the
area produces traffic volumes estimated to be roughly double that of the
late afternoon period, or about 1500 vehicles/hour on First Street.
(Unfortunately, no traffic data are available for this period.) This
heavy traffic, combined with light winds, caused the observed high con-
centrations. There is an indication of some leeward-windward effects at
street level, as well as a general down-street gradient at 12 m, in
accordance with the northwesterly rooftop wind direction.
61
-------
The last period, 0300-0400 PST (Figure 29), is interesting
in that it shows the result when the traffic is essentially "turned off.
The traffic volume at this time was probably only about 10 percent of
the daytime values, or roughly 100 vehicles/hour on First Street. How-
ever, the winds are very light over the city, and the general urban
background concentration, much of which is probably due to traffic during
earlier hours, remains high. The analysis shows that there is almost no
vertical variation in the observed CO concentrations, as would be ex-
pected for a weak street-level emissions source, and very little hori-
zontal variation.
2. Data Stratified by Wind Direction Classes
As previously discussed and as noted in the analyses just pre-
sented, the roof-level wind direction is apparently a controlling factor
of the street-level CO patterns. Accordingly, we used the rooftop wind
direction at Station 2 as the primary classification parameter. Other
multiparameter stratifications were carried out, using in one case wind
direction and time of day, and in the other case, wind direction, wind
speed, and traffic volume. However, the simple classification by wind
direction alone turned out to be the most revealing and the easiest to
use. The CO patterns varied only slightly with time of day, except as
influenced by traffic changes; the consistency of the concentration
distributions from hour to hour was quite striking. In general, the
stratification that included wind speed showed that there was a form of
inverse dependence of concentration upon wind speed, as was to be
expected.
Figures 30 to 37 depict the total CO and wind data
classed and averaged by eight 45° wind direction sectors. Horizontal
3-m concentration distributions as well as vertical concentration
62
-------
8.3 ^
(28)
SAN
ANTONIO
STREET
Solid arrows: 3-m wind direction
Dashed arrows: rooftop wind direction
CO concentrations in ppm
Wind speeds (cm/s) in parentheses
Vertical wind component (positive up)
in brackets, roof level
!
I
g
7
1
above 3-m level 1
7.3
(m)"
25
20
It
\ 15
HEIGHT -
0
5
0
(
'
•«^
12
v*
^J
8^S
8
(81)
y
8
FIRST
STREET
I ' I ' I
2
A
\ \
• 1A
*\ N \
> \» \
••' \V\
t V "r
— \
\
\
O 4
I , I
) 2 4 6
11.9
'(46) pie]
y / / ^]v 4.7
7/6/5/^14)<60'
'8.2 r+18i
(28) |+12J
^
,8.4
(13)
i ' i • i
6
\
^
\ \
I , I , I
8 10 12
Wind Direction
at 24 m
Station 2
e
*
I
6 '
U-
~ .
7 ~
—
—
I
14
CO CONCENTRATION — ppm TA-8563-6O
FIGURE 30 AVERAGE OF HORIZONTAL CO DISTRIBUTION AT 3 m
(ABOVE) AND VERTICAL CO PROFILES (BELOW), FOR
MEAN ROOFTOP WIND FROM 045° (±22.5°)
63
-------
SAN
ANTONIO
STREET
Solid arrows:
Dashed arrows:
s: 3-m wind direction
ws: rooftop wind direction
ration in ppm
s (cm/s) in parentheses
id component (positive up)
ets, roof level
-m level
7.0
(m)
7
8
^
\l (46)
/I8.9 f+22l
/
/
.
1 1
11
/10
m
9 11
FIRST
STREET
fi) b-11j
0
\M',f\f
Wine
090° -^^B at 2
Stati
10.8
(76)
20
15
I-
1 10
LU
I
I
4 6 8 10
CO CONCENTRATION — ppm
12
7 _
14
TA-8563-61
FIGURE 31 AVERAGE OF HORIZONTAL CO DISTRIBUTION AT 3 m
(ABOVE) AND VERTICAL CO PROFILES (BELOW), FOR
MEAN ROOFTOP WIND FROM 090° (±22.5°)
64
-------
7.3 >
(52)
9
n-
^ *-
SAN ^— ^
.8.7 r ^
(76) -8
L-36J
/ S***"(W
ANTONIO ^t^r / .4.4
STREET ^ 6 / / (63)
iinov .X^ X
Solid arrows: 3-m wind direction
Dashed arrows: rooftop wind direction
CO concentration in ppm
Wind speeds (cm/s) in parentheses
Vertical wind component (positive up)
in brackets, roof level
above 3-m level
4.7
(m)~*
\ •"«•/ J
<_5JXT
5 r
/
/
/ j
/ /
f
I f
4 /-y
5 6 7
FIRST
STREET
4.8 4
r r^i
b«J
5 o
r&>
***
6 J\ Wind
at 24
Statior
7
- 6.9
(177)
25
20
15
I
<2 10
u
I
1
1
4 6 8 10
CO CONCENTRATION — ppm
12
14
TA-8563-62
FIGURE 32 AVERAGE OF HORIZONTAL CO DISTRIBUTION AT 3 m
(ABOVE) AND VERTICAL CO PROFILES (BELOW), FOR
MEAN ROOFTOP WIND FROM 135° (±22.5°)
65
-------
SAN
ANTONIO
STREET
(36)
Solid arrows: 3-m wind direction
Dashed arrows: rooftop wind direction
CO concentration in ppm
Wind speeds (cm/s) in parentheses
Vertical wind component (positive up)
in brackets, roof level
above 3-m level
5.8
(m)
FIRST
STREET
3'7 [-161
(48) _J
L J
180°
Wind Direction
at 24 m
Station 2
5.8
(148)
25
20
15
10
LU
I
J.
_L
J.
4 6 8 10
CO CONCENTRATION — ppm
12
14
TA-8563-63
FIGURE 33 AVERAGE OF HORIZONTAL CO DISTRIBUTION AT 3 m
(ABOVE) AND VERTICAL CO PROFILES (BELOW), FOR
MEAN ROOFTOP WIND FROM 180° (±22.5°)
66
-------
8.0
(90)
SAN
ANTONIO
STREET
Solid arrows: 3-m wind direction
Dashed arrows: rooftop wind direction
CO concentration in ppm
Wind speeds (cm/s) in parentheses
Vertical wind component (positive up)
in brackets, roof level
above 3-m level
Wind Direction
at 24 m
Station 2
25
20
15
10
I
I
8* •
_L
1
4 6 8 10
CO CONCENTRATION — ppm
12
14
TA-8563-64
FIGURE 34 AVERAGE OF HORIZONTAL CO DISTRIBUTION AT 3 m
(ABOVE) AND VERTICAL CO PROFILES (BELOW), FOR
MEAN ROOFTOP WIND FROM 225° (±22.5°)
67
-------
9.5
(69)
10
SAN
ANTONIO
STREET
Solid arrows: 3-m wind direction
Dashed arrows: rooftop wind direction
CO concentration in ppm
Wind speeds (cm/s) in parentheses
Vertical wind component (positive up)
in brackets, roof level
above 3-m level
12.8
•»• 270°
Wind Direction
at 24 m
Station 2
25
20
15
H
v 10
UJ
X
I
I
4 6 8 10
CO CONCENTRATION — ppm
12
I
14
TA-8563-65
FIGURE 35 AVERAGE OF HORIZONTAL CO DISTRIBUTION AT 3 m
(ABOVE) AND VERTICAL CO PROFILES (BELOW), FOR
MEAN ROOFTOP WIND FROM 270° (±22.5°)
68
-------
profiles for each station are shown. In addition, the vertical wind
components at Stations 2 and 4 are included. These analyses are fairly
self-descriptive, but a few comments are in order.
Generally, for rooftop winds from the eastern sectors, Stations
6 and 7 show a significant leeward effect, while those across First
Street, Stations 5 and 8, indicate a windward case. This situation re-
verses for rooftop winds from the western sectors. Station 4 shows
little variation for north and south winds, possibly because of the much
lighter traffic on San Antonio Street.
\
For rooftop winds up First Street from the south (Figure 33),
there is no cross-street gradient, and the concentrations increase north
of the intersection. For the opposite case of north winds (Figure 37),
the concentrations at Stations 5 and 6 are uniform, but Stations 7 and 8
still indicate a cross-street gradient. For slightly more easterly
winds, from 045° (Figure 30), the cross-street gradient for these
stations reverses, with Station 7 becoming slightly higher. This indi-
cates that the concentrations at the two stations become uniform (analo-
gous to the 180° case) for a wind direction of about 020° to 030°. The
reason for this asymmetry in the CO patterns is unknown, but the sub-
stantial difference in building heights at Stations 7 and 8 may be in-
volved. In addition, the proximity of Stations 2, 5, and 6 to the
intersection complicates the circulation patterns compared to that
characteristic of a typical street canyon, which is basically a two-
dimensional situation. The latter is best represented in our data by
that from Stations 7 and 8.
As expected, the CO profiles generally indicate a decrease
with height. However, it is worthy of note that this slope is larger
for stations in the lee of buildings, while more uniform concentrations
71
-------
with height are apparent for those stations on the windward side. A
good example of this is found in the profiles for Stations 7 and 8 in
Figure 31.
Little relation was found between the observed CO concentra-
tions and the atmospheric thermal stratification, as characterized by
the observed temperature profiles between 3 and 20 m. This probably
reflects the dominance of mechanical mixing effected by the air flow
around the buildings and the motion of the vehicles, as compared with
mixing caused by convective processes.
D. Street Effects Submodel
In the development of this submodel, we have sought to find a
simple technique that has a sound physical basis and, of course, that
gives good results. In view of the importance of the upper-level wind
direction relative to the street orientation, as found by Georgii et al.
(1967) and confirmed by the San Jose data, we have retained the prin-
ciple of separation into leeward and windward cases, as used before.
Figure 38 illustrates the basic rationale behind the street
submodel. Given a general helical air circulation of the type shown,
the receptors on the leeward side of the building (on the right in the
figure) are exposed to substantially higher concentrations than those
on the windward (left) side, because of the reverse flow component
across the street near the surface. It is assumed that the concentra-
tion (C) at a receptor consists of two components that are superimposed.
One component is the background concentration (C ) in the air entering
the street canyon from above, and the other concentration component (AC)
arises from the locally generated CO emissions in the street,
72
-------
; BUILDING
S TRAFFIC;
LANE
-W-
MEAN
WIND
(U)
BACKGROUND
CO CONCENTRATION
-------
Y =
(9)
Also U may be taken to be linearly related to the roof-level wind,
s
U (m s"1):
U = k (U + 0.5)
S &
(10)
where the additive 0.5 m s"-1- is an estimate that accounts for the
mechanical air movement caused by traffic. The motions of the cars
also mix the CO into an initial volume of dimensions (L ) comparable
o
to the vehicle size, about 2m; Eq. (9) becomes
Y = k L + L I = k (L + 2)
IV o / 1
(11)
Now L is the diagonal distance from the center of the nearest traffic
lane to the receptor. From Figure 38, we have,
/ 2 2
L = x + z
(12)
where x and z are the horizontal distance and the height of the receptor
relative to the center of the traffic lane. Combining Eqs. (8), (10),
(11), and (12), we have
Q
*!*
2 (U +
0.5)
7 2 2\1/2
Ix + z 1 +
2
(13)
We can represent a 50/50 mix of pre- and post-1965 model cars by the
emission formula (Ludwig et al., 1970):
74
-------
E = 0.5 I 245 S --" + 1120 S °-85J[ & 1 (14)
\veh-mile/
or approximately,
E ~ 700 s"°-75( g ) (15)
\veh-mile / '
where S is the average vehicle speed (miles per hour). If E is multiplied
by the traffic flow, N (vehicles per hour), we obtain the line source
strength, Q in units of g mi h . If we change the units of Q to
-1 -1
mg m s , we get
-0. 75
Q = 0.12 N S . (16)
Combining Eqs. (13) and (16), and introducing a factor of 0.87 to
-3
convert units of mg m to ppm, we have
Ap (0.1) K N S"°-75
L = - r/~2 — —
(U + 0.5) (x + ' z
i '
1 +2
where K = 1/k k . A reasonable value of the dimensionless constant k
X £
was found from our data to be about 7. In downtown San Jose, we measured
S as approximately 14 mi h"1. Substituting these values into Eq. (17) gives
0.07 N (18)
2 2 \i/2
(U + 0.5) MX + z I +2
i.
This equation and Eq. (7) were used to calculate leeward concentra-
tion profiles for comparison with the San Jose data, using x = 8 m for
First Street and x = 7 m for San Antonio Street.
75
-------
On the windward side, the air flow is mostly downward. Hence the
air should be fairly well mixed, since it has traveled a considerable
distance from the source, and there should be little vertical concentra-
tion variation. We have again used the box model, assumed the mixing
volume to be constrained only by the width of the street (W), and have
considered that the vertical concentration is uniform, giving
-0.75
0.1 K N S
*CW = W (U+0.5) '
or
W W (U + 0.5)
when the previously used values of K and S are incorporated.
When the wind direction is such that neither a leeward nor a wind-
ward case is appropriate, an intermediate concentration (AC ) is found
by averaging the results of Eqs. (18) and (20):
AC = 1/2 (AC + AC 1
I \ L W/
\
1 T
(21)
0.'035 N ) 1
(U + 0.5) \
2 2
x + z
1
1/2 "I + W
) +2J )
There remains only the specification of wind direction sectors for
the three different cases. This is carried out basically as set out in
Figure 24. However, for a few of the stations in San Jose, these
sectors had to be shifted by 15° to 45°, presumably because of the
complexities in the flow caused by the differing building heights and
the proximity to the intersection of several stations. The wind direc-
tion sectors used for the various stations are given in Table 3.
76
-------
Table 3
WIND DIRECTION SECTORS FOR SAN JOSE STREET STATIONS
Station
No.
2 (B)
4 (D)
5 (E)
6 (F)
7 (G)
8 (H)
9 (I)
Wind Direction Ranges for Various Cases
Leeward
060°-150°
315°-045°
225°-315°
045°-135°
060°-150°
210°-360°
225°-315°
Windward
210°-360°
135°-225°
045°-135°
225°-315°
210°-360°
060°-150°
045°-135°
Intermediate
000°-060°, 150°-210°
045°-135°,, 225°-315°
315°-045% 135°-225°
315°-045% 135°-225°
000°-060°, 150°-210°
000°-060°, 150°-210°
315°-045°, 135°-225°
Equations (18), (20), and (21), along with Table 3 constitute the
methodology developed to handle street effects for San Jose. The
results of the verification tests using this simple street submodel
are described in Section VI.
77
-------
IV ANALYSIS OF HELICOPTER AND MOBILE VAN DATA
A. Introduction
The helicopter and mobile van measurements of carbon monoxide con-
centration and temperature provide a unique set of data for use in the
refinement and validation of the urban diffusion model. The vertical
profiles provide detailed information on the structure of the lower at-
mosphere and have been used for the determination of background CO con-
centrations, mixing depth, and stability. Additionally, they provide a
qualitative 'feel ' for the experimental conditions in terms of the
meteorology and traffic. The aerial and surface traverses were made
about the perimeter of the central business district (see Figures 13 and 20);
analysis of these data provides an independent estimate of the rate of
vehicular emission of CO over the area, while the data also permit us
to compute the vertical diffusion of CO. Vertical profiles of wind
structure over the city during selected periods were obtained from pilot
balloon (pibal) ascents at nearby San Jose State College; the pibal data
are summarized in Appendix E.
B. Data Reduction Techniques
The vertical profiles of temperature and CO were determined at
intervals of about 15 m to a height of 152 m and then at 30.5-m intervals
to the top of the profile (nominally 1000 m). As examples, four sets of
CO and temperature profiles are given in Figure 39, corresponding to
periods for which pibal data are available.
The traverse data obtained by the helicopter comprise 19 point
values around the circuit shown in Figure 20 for each level. Values
79
-------
(a)
EVENING DATA RUN
9 DECEMBER 1970
ASCENDING
DESCENDING
SPECIAL
RUN
900
800
700
V)
% 600
E
I 500
I-
1*00
I
300
200
100
0
T
T
(b)
NOON DATA RUN
10 DECEMBER 1970
I
I
I
2.5 5.0 7.5
CARBON MONOXIDE — ppm
5 10
TEMPERATURE
I
15 20
°C
TA-8563-27a
FIGURE 39 VERTICAL PROFILES OF CARBON MONOXIDE
AND TEMPERATURE AT SPARTAN STADIUM,
SAN JOSE, CALIFORNIA
80
-------
900
800
700
600
500
400
300
200
100
0
900
800
700
600
500
400
300
200
100
0
(c)
EVENING DATA RUN
10 DECEMBER 1970
ASCENDING
DESCENDING
(d)
EVENING DATA RUN
11 DECEMBER 1970
I
0 2.5 5.0 7.5 0
CARBON MONOXIDE — ppm
I
I
5 10 15 20
TEMPERATURE — °C
TA-8563-27b
FIGURE 39 VERTICAL PROFILES OF CARBON MONOXIDE
AND TEMPERATURE AT SPARTAN STADIUM,
SAN JOSE, CALIFORNIA (Concluded)
81
-------
are given at each corner of the pattern (the first corner is repeated
at the end of the flight) and for three intermediate points on the two
short legs and four on each of the two long legs. Examples of the CO
traverse data for four periods are given in Figure 40. For the mobile
van measurements, averages over the route segments shown in Figure 20
were used rather than point values, owing to the highly variable nature
of the street level concentrations. An average of five van traverses
were made during each data period. Because of street patterns, the route
of the van differs slightly from that of the helicopter, particularly
along the southern leg (see Figure 20). The effect of these differences
is minimized when leg averages are determined corresponding to the heli-
copter pattern. The weighting procedure is straightforward and is sum-
marized in Table 4.
Table 4
CORRESPONDING HELICOPTER AND VAN LEGS, INDICATING WEIGHTING FACTORS
FOR DETERMINATION OF AVERAGE CO ALONG VAN ROUTE SEGMENTS
Helicopter Leg
I-II
II-III
III-IV
IV- 1
Corresponding Van Leg
Segment
V5-V4
V4-V3
V3-V2
VI- V6
Weight
1.00
0.75
0.33
0.40
Segment
V3-V2
V2-V1
V6-V5
Weight
0.25
0.67
0.60
C. Results
1. Determination of Vehicular Emissions and Vertical Diffusion
of Carbon Monoxide
The vehicular emissionsand the vertical diffusion of carbon
monoxide within the downtown area have been determined through mass
82
-------
budget analysis. The horizontal perimeter of the budget box is described
by the. near-coincident routes of the helicopter and van traverses; the
top is defined by the vertical extent of the aerial measurements. The
volume is divided into various sublayers by the heights of the heli-
copter traverses (61, 92, 152, 213, and 305 m).
The mean transport of CO into or out of the four sides of the
sublayers is given by the area integral of the product of the component
of the wind speed normal to the sides and the mean CO concentration along
the sides. In the absence of detailed wind measurements in the sublayer
nearest the surface, it was assumed that the wind profile could reasonably
be approximated by a simple power law,
S = SJz/zJ" , (22)
where S is the wind speed normal to the side at height z, and S is the
•f
normal component of the measured wind speed at z , the lowest level re-
'P
solved from the pibal ascents (nominally 72 m). Values of the exponent
p are given in Table 5 as A function of atmospheric stability (repre-
sented by the difference between the 122-m and 2-m temperatures). Car-
bon monoxide concentration was assumed to change linearly with height
within each sublayer. Therefore, the mean transport through each of
the sides of the lowest layer is given as
h
S*L
S X dz dL =
'0 "L Z*
x h"*1 .«•
o ah
p + 2 p + 2
(23)
where X is mean CO concentration along the side, h is the layer thick-
ness, L is the length of the side, the subscript zero denotes a surface
value, and the parameter a is the bulk CO gradient over the layer,
X - XQ
a = —^ . (24)
h
83
-------
14 -
12
I
10
w
Q
X 8
O
o
O
m
cc
< 4
O
•t
r...
I
(a)
EVENING DATA RUN
9 DECEMBER 1970
Sfc Wind: 305°/5kts
200'
500'
700'
1
-
- - 200'
r >^x^ 300'
1 1
(b)
NOON DATA RUN
10 DECEMBER 1970 —
Sfc Wind: 310°/8kts
—
.— ¥^~" " * r^=^r ^ ,
"*""- -~lr~~ ----^N
500' ~~-B" 700. ^ef^:ioo~'~~~ — — —
14
12
I
10
LU
Q
X
O
o
CO
DC
<
O
III
ROUTE POINT
IV
TA-8563-26a
FIGURE 40 HORIZONTAL TRAVERSES OF CARBON MONOXIDE
AT INDICATED HEIGHTS FOR BOX PATTERN OVER
DOWNTOWN SAN JOSE, CALIFORNIA
84
-------
14 -
12
10
LU
Q
X 8
O
O
I 6
O
m
DC
< 4
O
1
—
-
-
-
fc -- -'"'£-—
r — 1"
i i
(c)
EVENING DATA RUN
10 DECEMBER 1970 —
Sfc Wind: 310°/8kts
200' —
"^/ ' *%x- j. --''" Vv-
* 500' "-^----
1 1 7°°'
14 -
12
a
a
10
LU
Q
X 8
O
O
O
ca
DC
<
O
6
(d)
EVENING DATA RUN
11 DECEMBER 1970
Sfc Wind: 250°/8kts
r:
in
ROUTE POINT
IV
TA-8563-26b
FIGURE 40 HORIZONTAL TRAVERSES OF CARBON MONOXIDE
AT INDICATED HEIGHTS FOR BOX PATTERN OVER
DOWNTOWN SAN JOSE, CALIFORNIA (Concluded)
85
-------
Table 5
VALUES OF p IN EQ. (22) AFTER FROST (1947)
T122nfT2m
(°C)
-2.2 to -1. 1
-1.1 to 0
0 to 1.1
1.1 to 2.2
2.2 to 3.3
3.3 to 4.4
P
0.145
0.25
0.32
0.44
0.59
0.63
Computation of the mean CO transport through the upper layers
incorporated the additional assumption of linear wind changes with height
between pibal data levels, where
dz dL = S
Z~hAz
(25)
Implicit in the budget analysis is the assumption that the net
horizontal turbulent flux of CO for each layer is negligibly small com-
pared to the net transport by the mean wind. This condition is satisfied
by a horizontally homogeneous emission source and wind field over the
area. Furthermore, when the mean vertical component of the wind is taken
as zero, the net horizontal transport for any given layer is then balanced
by the net vertical (turbulent) flux of CO through the bottom and top.
The uppermost traverse was made sufficiently high such that the vertical
diffusion through that level is essentially zero. Working down toward
the surface, the vertical CO fluxes through the various levels can be
determined as the residue in the budget, where the near-surface (3-m)
flux corresponds to the vehicular CO emission rate. The vertical separa-
tion of approximately 3 meters between the emission source (automobile
86
-------
exhausts) and the receptor height of the van may be considered negligible
for the majority of atmospheric conditions although it may lead to slight
underestimates of the emission rate during near-stagnation episodes.
Results of the budget analysis are given in Table 6. The
analysis for the 1211-1238 PST period on 10 December 1970 was done twice
with two different assumed shapes for the wind profile. The only pibal
sounding available on that date was made approximately 4 hours later.
However, because of the constancy of the surface winds during the after-
noon (obtained from the hourly observations at the nearby San Jose
Municipal Airport), it was felt that the sounding would be representative
of the midday vertical wind field, and the first analysis for the noted
period follows from this assumption. The second analysis uses an assumed
power profile [Eq. (22)] for the wind based on the airport surface wind
observation and the vertical temperature gradient obtained from the heli-
copter measurements. This second approach is essentially one which a
researcher might be forced to employ in the absence of low-level wind
profile data. The two methods differ by a factor of 2.4. This relatively
large difference emphasizes the importance of the low-level wind field
for pollution transport computations in general, and for mass budget
analysis in particular.
Vehicular carbon monoxide emissions were also determined with
the empirical emissions model presented by Ludwig et al. (1970), and
slightly modified for this study, where
E = 0.5 [245 S~°'48 + 1120 S~ ' ) . (26)
E is the emission rate in g-CO per vehicle-mile, and S is the average
vehicle speed in miles per hour. The first term in the parentheses
represents emissions from post-1968 vehicles; the second term is for
87
-------
Table 6
TRANSPORT RATES OF CARBON MONOXIDE THROUGH THE SIDES, TOP, AND BOTTOM
OF THE SUBLAYERS OF THE SAN JOSE BUDGET BOX
Date/Time
9 December 1970
1707-1729 PST
10 December 1970
1211-1238 PST*
10 December 1970
1211-1238 PST"f
10 December 1970
1711-1743 PST
11 December 1970
1702-1724 PST
Height
of Layer
(m)
Top
213
152
92
61
305
213
152
92
61
305
213
152
92
61
213
152
92
61
213
152
92
61
Bottom
152
92
61
3
213
152
92
61
3
213
152
92
61
3
152
92
61
3
152
92
61
3
Horizontal
Transport
(g-CO s"1)
Total
In
678
624
474
2875
139
280
478
274
1270
722
1105
1590
846
2324
287
442
278
1428
301
562
446
3626
Total
Out
779
870
642
3184
167
306
490
278
1507
870
1211
1628
859
2759
275
464
338
1632
276
821
749
4253
Vertical Flux
(g-CO S'1)
In through
Bottom
101
347
515
823
29
55
67
72
310
148
253
292
305
739
-12
23
83
287
-25
259
562
1189
Out through
Top
0
101
347
515
0
29
55
67
72
0
148
253
292
305
0
~-0
23
83
0
^X)
259
562
With 1700 PST pibal data.
t.
With assumed "power law" wind profile.
88
-------
earlier-model vehicles. The average vehicle speed was taken as 13.7
-1
mi h for all periods as determined from an analysis of the mobile van
movement through the downtown sector. Table 7 gives a synopsis of
the van speeds for the period 9-11 December 1970.
Table 7
AVERAGE VEHICLE SPEEDS IN THE DOWNTOWN SECTOR OF SAN JOSE
FOR SPECIFIED TIMES DURING THE PERIOD 9-11 DECEMBER 1970
Time Period
0740-0900
1137-1245
1616-1757
Number of
Van Circuits
15
13
17
Average Speed
(mi IT1)
14.3
14.3
12.7
Standard Deviation
(mi h'1)
1.5
1.1
3.4
The traffic monitoring network comprises approximately 44
percent of the surface area of the budget box. Therefore, it was neces-
sary to estimate the traffic volume outside the network, but within the
box. Since daily, routine data are not available for this outer area,
it was assumed that the traffic volume for this region is proportional
to that within the monitoring network. To determine the proportionality
constant, we used a selective traffic count conducted by the City of
San Jose (Turturici, 1970) where traffic volumes obtained during August
1969 are presented by street segments for a region encompassing all of
the budget area. The ratio of traffic in the budget area to that within
the monitoring network was found to be 1.64. Furthermore, it was assumed
that the mean vehicular speed was constant over the entire area. Using
the results of the emissions model [Eq. (26)], the total emission rate
(Q, gm-CO s ) over the budget area is obtained from the equation
89
-------
Q = 1.64 E L N
(27)
where L is the mean link length (0.1 mile) and N is the total number of
vehicle counts registered in the monitoring network per second (averaged
over a 90-minute period). The results of this computation are given in
Table 8, together with those obtained from the mass budget analysis.
Table 8
CARBON MONOXIDE EMISSION RATES (Q) FOR THE BUDGET AREA DETERMINED
FROM THE MASS BUDGET ANALYSIS AND TRAFFIC DATA [WITH EQ. (27)]
Date
(1970)
9 December
10 December
10 December
11 December
Average
Time
(PST)
1707-1729
1211-1238
1711-1743
1702-1724
N*
(cts s'1)
25.76
22.43
26.24
28.15
25.64
Q, Traffic Data
(g-CO s-1)
408
355
416
446
406
Q, Budget Analysis
(g-CO s"1)
823
310
287
1189
652
N is the number of traffic counts per second within the traffic moni-
toring network.
The mean area emission rate determined from the traffic data
differs from the mean of the budget values by 38 percent. The dif-
ferences in the individual cases probably arise because of the following
factors: (1) inhomogeneities in the horizontal wind field, (2) nonzero
vertical wind components, (3) changes of the CO field during the obser-
vation period, and (4) inaccurate representation of the total traffic
flow by assuming it to be a constant factor of that within the monitoring
network. It is encouraging that these two totally independent methods
90
-------
agree to the extent indicated. At this time there does not appear to
be any justification for changing the emissions model.
2. Mixing Depth Estimates
The helicopter profiles of temperature and carbon monoxide
concentration have been used to estimate the depth of the San Jose urban
mixing layer for the period 7, 9, 10, and 11 December 1970. These esti-
mates are compared with values from the mixing depth submodel (Ludwig
et al., 1970) and, additionally, with values obtained from the SRI/APCO
Mark VIII lidar system, which was operated coincident with this program
under another project.
The Mark VIII lidar is composed basically of a laser trans-
mitter, which emits a very brief, high-intensity pulse of coherent mono-
chromatic light, and a receiver, which detects the energy at that
wavelength backscattered from atmospheric aerosols, as a function of
range. Some of the features of this ruby-lidar system are: (1) coaxial
transmitter-receiver alignment, (2) high pulse rate (20/minute), (3)
range compensation, and (4) automatic programmed elevation scanning and
firing. The data are recorded on a magnetic disk in a format that per-
mits an intensity-modulated range-height indicator (RHI) display on ah
oscilloscope. The resulting vertical cross sections through the haze
can be analyzed to determine time-average composites through multiple
displays and photographic exposures.
The lidar was used to monitor the mixing layer depth as repre-
sented by the lower haze layer(s). The displays in Figure 41 represent
a series of time-height cross sections obtained by vertical observations
12 seconds apart on 11 December at the Spartan Stadium area. The time
variation in the haze layer heights is striking. The mean heights, how-
ever, correspond reasonably well with significant levels on the
91
-------
12 18 24
ELAPSED TIME — mm
30 36
f;;?O'
'-•••:-
n 1 r
1206 PSJ
-— 1651 PST
0 2.5 5.0
CO — ppm
5 10 15 20
TEMPERATURE —°C
FIGURE 41 LIDAR-OBSERVED TIME-HEIGHT CROSS SECTIONS OF THE URBAN HAZE LAYER
OVER SAN JOSE, CALIFORNIA, ON 11 DECEMBER 1970. Concurrent helicopter
profiles of carbon monoxide (CO) and air temperature are shown at the bottom right
92
-------
corresponding helicopter profiles; these values are summarized in
Table 9. Mixing depths computed for the same time periods during the
morning helicopter temperature soundings and the submodel used with
the diffusion model are also given in the table. The agreement among
the three methods is generally good.
Table 9
COMPARISON OF MIXING DEPTH ESTIMATES OBTAINED FROM THE MIXING DEPTH
SUBMODEL AND THE SUBJECTIVE ANALYSIS OF THE HELICOPTER PROFILE DATA,
WITH THE LIDAR-OBSERVED HAZE-LAYER STRUCTURE AT SAN JOSE, CALIFORNIA
Date
(1970)
7 December
9 December
10 December
11 December
Time
(PST)
1240
0800
1200
1700
0800
1200
1700
0800
1200
1700
Mixing Depth
(m)
Model
—
115
950
672
143
567
567
148
539
539
Helicopter
215
100
125, 675^
200, 725
125
425
500, 750
60, 550
500, 725
600
Lidar-Observed Haze Layers
(m)
450, 850^
300-500 (variable)
450-600 (variable)
750
300
450*
250
Low Clouds
350-700 (variable)
600-750 (variable)
t.
Indicating the height variation over a one-half hour period.
Multiple values for the helicopter and lidar data represent the
tops of the surface and elevated mixing layers.
Lidar observation made at 1430 PST.
93
-------
The helicopter and lidar data often indicate the presence of
multiple stable layers in the lowest 1000 m above the surface. The CO
profiles on these occasions show that there is some penetration of CO
through the near-surface mixing layer. This occurs when the lid is not
particularly strong (as indicated by the temperature profile) while the
concurrent lidar observations frequently indicate intermittancy in the
occurrence and/or height of the lid. This intermittancy may be the
result of local (convective) or advective effects and seems to represent
a transitional stage in the lower atmospheric structure.
3. Stability Estimates
The helicopter temperature profiles were used in conjunction
with the airport wind speeds to compute a bulk stability coefficient (B),
- T
2 , (28)
°3
where T is temperature (°C), U is the wind speed (knots), and the sub-
*
scripts are the heights (m) of the various measurements. The bulk
stability coefficients were compared with modified Pasquill-Turner
stability categories determined from the diffusion model (Ludwig et al.,
1970) as a function of insolation strength, wind speed, and cloud cover
as summarized in Table 10.
T-
The coefficient B provides an estimate of the ratio of the production
of energy by buoyant forces to the dissipation of mechanical energy by
turbulence. Neutral atmospheric conditions are indicated by values of
B near zero; unstable conditions result in B < 0 and stable conditions
in B > 0.
94
-------
Table 10
MODIFIED PASQUILL-TURNER STABILITY CATEGORIES USED WITH THE
DIFFUSION MODEL (Ludwig et al., 1970)
AS A FUNCTION OF INSOLATION, WIND SPEED, AND CLOUD COVER
Surface Winds
(knots)
£3
3-6
6-10
10-12
;>13
Daytime Insolation
Strong
1
1
2
3
3
Moderate
2
2
3
3
4
Slight
2
3
3
4
4
Night Clouds
2:5/10
5
4
4
4
4
<4/10
5
5
4
4
4
1 = extremely unstable, 2 = moderately unstable,
3 - slightly unstable, 4 - neutral, 5 = slightly stable.
The stability estimates from the diffusion model agree very
well with the observed thermal stability (given by B) for neutral and
slightly unstable conditions as shown in Table 11. However, the model
often appears to break down by predicting moderately unstable conditions
during the morning hours when, in fact, stable conditions are observed.
The presence of a morning, surface-based inversion with corresponding
low wind speeds is, in essence, indicative of "night" conditions, despite
the slight insolation. With reference to Table 10, employing the
nighttime hypothesis leads to the prediction of neutral or slightly
stable conditions in agreement with observed conditions.
In summary, the model predicts atmospheric stability reasonably
well except for the few morning hours shortly after sunrise during light
wind conditions. On these occasions, reasonable stability estimates are
obtained by considering the situation to be better simulated by the
nighttime case in the stability estimation methodology.
95
-------
Table 11
COMPARISON OF BULK STABILITY COEFFICIENTS COMPUTED FROM EQ. (28)
WITH STABILITY CATEGORIES DETERMINED FROM THE DIFFUSION MODEL
FOR THE PERIOD 20 NOVEMBER TO 11 DECEMBER 1970, AT SAN JOSE, CALIFORNIA
Time
(PST)
0800-0900
1200-1300
1700-1800
Average
Stability Categories
Moderately Unstable
(Category 2)
0.0440
0.1300
0.0035
-0.0140
-0.0074
0.0312
Slightly Unstable
(Category 3)
-0.0468
-0.0040
-0.0035
-0.0531
-0.0269
Neutral
(Category 4)
0.0041
-0.0023
-0.0062
-0.0039
-0.0030
-0.0054
-0.0097
-0.0039
96
-------
V INCORPORATION OF THE RESULTS INTO THE URBAN DIFFUSION MODEL
A. Introduction
In this section we summarize the actual refinements that have been
made to the basic urban diffusion model as a result of the San Jose
field program and of further analyses of data from experiments by other
groups. Our rationale in determining the necessary revisions has been
to measure the input variables and CO concentrations as completely and
as accurately as possible. In this way, (1) the performance of the
submodels that estimate certain of the input parameters for the model
can be assessed, and (2) the accuracy of the basic diffusion model in
predicting CO concentrations may be determined by using accurately
measured input parameters, rather than indirect estimates. However,
the fundamental design specification that, in practice, the model will
use available conventional meteorological data has remained as a guiding
principle.
B. Emissions Submodel
In past work we have assumed that CO emissions (E, g veh mi )
can be estimated from the mean vehicle speed (S, mi h ) on the basis
of empirical relationships of the type proposed by Rose et al. (1964):
E = cS , (29)
where c and g are constants. We used the formulas
-0.85
E = 1120 S (30)
97
-------
for pre-1965 vehicles without exhaust control systems, and
E = 245 S °' (31)
for post-1968 model year automobiles.
As shown in the preceding section, equations of the form of Eq.
(29) can give results that agree well with independent measurements
of emissions if some accounting is made of the mix of exhaust-controlled
and uncontrolled vehicles. On this basis, there appears to be no reason
to change the emissions model substantively. Of course, some account
must be taken of the vehicle mix. For instance, if p is the fraction
of the cars newer than 1965 (and 1 - p, the fraction older) then an
equation of the following form would be used in the model:
-0.85 -0.45
E = 1120 (1 - p)S + 245 p S . (32)
When the mix is 50/50, i.e., p = 0.5, Eq. (32) reduces to
1 / -0.85 -0.48 \
E = - (1120 S + 245 S ) . (33)
2 \ I
Over a reasonable range of speeds the two terms of Eq. (33) can be well
approximated by a single exponential, as was done in Eq. (15).
C. Estimation of Atmospheric Stability
The helicopter temperature profiles were used to drive a bulk
stability parameter with which to check the stability category estima-
tion procedure, as described in Section IV. The results of this analysis
98
-------
showed that the original model underestimated atmospheric stability when
it was applied to early-morning, light wind situations. Accordingly,
the table based upon modified Pasquill-Turner categories has been
further revised to the form shown in Table 12. The changes
Table 12
REVISED STABILITY CATEGORIES
Surface
Winds
(knots)
<. 3
3-6
6-10
10-12
> 13
Daytime
(SR + 4 hours to SS - 3 hours)
Strong
Insolation
1
1
2
3
3
Moderate
Insolation
2
2
3
3
4
Slight
Insolation
2
3
3
4
4
Early AM
and Late PM
(SR + 1 to SR + 3
and
SS - 2 to SS - 1)
4
4
4
4
4
Nighttime
(SS to SR)
:> 5/10 < 4/10
Clouds Clouds
5 5
4 5
4 4
4 4
4 4
SR = Sunrise, SS = Sunset.
consist of adding an additional time classification for early morning
l
and late afternoon cases, and of adjusting the times for the daytime
classification accordingly.
D. Vertical Diffusion Rates
(.
The Pasquill-Gifford curves that have been used to find values of
the vertical dispersion parameter (a ) as a function of travel distance
z
and stability have been a subject of discussion in the meteorological
community for some time. These curves are based upon measurements taken
in England over rolling, wooded countryside containing small towns, and
hence the applicability of the data to urban areas has been questionable.
99
-------
In an effort to develop more representative a curves for urban
z
areas, we have examined available data from the few urban experiments
carried out by other groups. There are two comprehensive field programs
that are the most help in this regard. The data that are best known
are those of Pooler (1966) and McElroy and Pooler (1968) ("M & P"), who
conducted tracer tests with florescent particles (FP) in St. Louis during
the period 1963-1965. An additional data source that has been largely
neglected because of its originally classified nature is the extensive
series of tracer (FP) experiments carried out in St. Louis, Minneapolis,
and Winnipeg by Leighton and Dittmar (1952, 1953a-e) ("L and D") during
the period 1952-1953.
We have carried out further analyses of certain of the test results
from the latter study. The line-source releases are of special interest
because the FP was released from a dispenser mounted on a moving auto-
mobile, which closely simulates the typical emission conditions for
automobile exhaust. The automobile was driven approximately cross-
wind for a route of about 2 miles across the city, and a network of
samplers downwind gave ground-level dosage values. If the release rate,
release time, transport wind speed, and the cross-wind integrated dosage
L
as a function of distance downwind are known, the variation of cr with
z
travel distance can be computed from the requirement for mass balance,
if the assumption of a Gaussian-shaped vertical concentration distribu-
tion is made. A similar procedure was used by McElroy and Pooler (1968).
Table 13 summarizes the test conditions for the five cases analyzed,
which included three line-source tests (all that were carried out), and
two point-source releases that were conducted close in time to the line-
source tests. The objectives of the analysis were to determine a versus
z
distance for the five tests and to see whether this depended signifi-
cantly upon the type of release.
100
-------
Table 13
TEST CONDITIONS FOR ANALYZED LEIGHTON AND DITTMER (1953c) ST. LOUIS DATA
Test
No.
1012 A
1012 B
1006 A
1006 B
1006 C
Release
Type
Point
Line
Line
Point
Line
Date
(1953)
6/15
6/15
5/29
5/30
5/30
Time
(CST)
2045
2227
2335
0126
0335
Wind
(m s'1)
135° /2. 3
135° /2.0
200° /I. 8
220° A. 6
220° /I. 5
Kf Lapse
(SFC-lOOm)
1.7°C
1.2°C
1.0°C
0.9°C
0.8°C
Mixing
Depth
(m)
*
125
*
125
130
120
120
Sky
HI (D)
MID©
CLEAR
CLEAR
CLEAR
Weak inversion at top of mixing layer.
As indicated in Table 13, these tests were all conducted in the
early summer at night, with light winds and a neutral to slightly un-
stable lapse rate. The wiresonde temperature data for the 15 June tests
showed a weak inversion at 125 m, whereas those for the 29-30 May experi-
ments revealed a strong inversion at about the same height.
The results of the analysis are presented in Figure 42. Test
numbers 1006 A and 1006 C gave essentially identical results, so they
are plotted as the same line to reduce clutter on the graph. No consis-
tent differences between the diffusion from the point and line sources
are indicated. It is interesting that the a values for the 29-30 May
z
tests show a leveling-off between about 130 and 200 m in magnitude. This
is probably a reflection of the vertical trapping due to the strong in-
version lid present on that occasion.
Since the L and D data are mostly for short ranges between 0.1 km
and 1.5 km, while the M and P data generally cover the intermediate ranges
from 0.7 km to 10 km, it is appropriate to compare the results from both
101
-------
400 —
200 —
8
-------
10"
103
10'
IT
L&D DATA
ST. LOUIS
(NIGHT) x
-------
10°
10"
10'
10U
I 1 I MINI /I
A
10"
PASQUILL - GIFFORD
1 ST. LOUIS (M&P, 1968)
JOHNSTOWN, PA.
FORT WAYNE, INDIANA (E-F)
FORT WAYNE, INDIANA (D)
10° 10"*
DOWNWIND DISTANCE — m
10°
TA-8563-73
FIGURE 44 COMPARISON OF URBAN VERTICAL DISPERSION
DATA WITH THE PASQU ILL-GIF FORD CURVES
(ADAPTED FROM McELROY AND POOLER, 1968)
104
-------
curves (Pasquill, 1961; Gifford, 1961). It is apparent that the latter
give underestimates of vertical diffusion in urban areas. We have
revised the model to incorporate better vertical diffusion estimates
based on the M and P data and the limited L and D results. As shown in
Figure 45, we have approximated and extrapolated the a curves of
z
Figure 43 by expressions of the form
a = a x
z
(34)
where x is downwind distance. The values of the constants a and b for
the various stability categories are given in Table 14. These new
values replace those previously used with the model, which were taken
from the Pasquill-Gifford curves. Figure 45 shows a unique feature of
the new curves; they all intersect at a = 10 m for x = 50 m. This
z
represents a reasonable value for the initial mechanical mixing due to
roughness elements near the source. Accordingly, for the model we have
assumed that a = 10 m for x ^ 50 m.
z
Table 14
VALUES OF CONSTANTS IN EQ. (34)
AS A FUNCTION OF ATMOSPHERIC STABILITY CATEGORY
Stability
Category
1 (A)
2 (B)
3 (C)
4 (D)
5 (E)
Stability Type
Very unstable
Unstable
Slightly unstable
Neutral
Slightly stable
a
0.07
0.12
0.23
0.50
1.35
b
*
1.28
1.14
0.97
0.77
0.51
Estimated from extrapolation of a's and b's
for other categories (no data available).
105
-------
10
10" 10° 10"
DOWNWIND DISTANCE — meters
TA-8563-98
FIGURE 45 VERTICAL DIFFUSION AS A FUNCTION OF TRAVEL DISTANCE
AND STABILITY CATEGORY, AS REVISED FOR URBAN CONDITIONS
106
-------
E. Mixing Depth
Table 9 in the preceding section shows that the methods used in
the model to define the mixing depth give values that are comparable to
those determined from the helicopter profiles of CO and temperature.
There do not seem to be any systematic differences that would warrant
changing the mixing depth submodel at this time.
F. Local (Street) Effects
This new submodel, used to compensate for effects of the street
canyon on CO concentration at streetside receptors, is thoroughly
described in Section III by Equations (7), (15), (17), and (19),
plus Table 3. In effect, a concentration increment (AC) is computed
from this procedure and added to the urban concentration (C ) calculated
b
from the basic urban diffusion model. Since the traffic on the street
passing directly by the receptor is included in the calculation of faC,
this same traffic is neglected in calculating C . The concentration
b
increment, AC, depends upon rooftop wind direction and wind speed, local
traffic volume and average vehicle speed, width of the street, height of
the receptor above the street, and horizontal distance of the receptor
from the center of the nearest traffic lane. The results of verification
tests using this new procedure and the other model revisions are de-
scribed in the next section.
A final word is in order regarding the appropriate wind speed to
use as an input to the street effects submodel. Rooftop winds above the
street will not generally be available, and airport wind speeds will have
to be used. In this case an appropriate relation to apply, derived from
the San Jose data, is
U = 0.47 U - 0.60 m s~ , (35)
a
107
-------
where
U = rooftop wind speed and
U = airport wind speed.
a
This does not differ substantially from the relationship that
Schnelle et al. (1969) found for Nashville, Tennessee.
108
-------
VI EVALUATION OF THE PERFORMANCE OF THE REVISED MODEL
A. Introduction
From the beginning, the diffusion model has been developed as a
composite of separate modules (Johnson et al., 1969; Ludwig et al.,
1970). The reasons for this are simple: it allows changes in various
parts of the model without a complete disruption of the entire struc-
ture. It also allows us to check the performance of each of the modules
and thereby diagnose the model's weaknesses. We have exploited this
feature in designing different parts of the San Jose field project to
test different subunits of the model.
In this section, the performance of the various modules will be
reviewed. We will also present the results of our attempts to check
the overall performance of the composite model. It is this overall per-
formance that is of greatest interest to the potential user. Finally,
the results of the tests are discussed in terms of those conditions
under which the model performs well and those under which it performs
poorly. This aspect of the validation study was undertaken partly to
help the designers of the model decide where to concentrate the future
efforts at improvement and partly to help users avoid conditions for
which the model may give unreliable results.
B. Tests of the Subcomponents
I. Emissions Submodel
a. General
The emissions submodel requires traffic data as inputs.
Hourly traffic volumes and average vehicular speeds are needed for the
109
-------
different highway segments in the area. In practice the model uses
measurements of total 24-hour volumes and distributes the total traffic
among the hours of the day according to a temporal function based on a
limited number of observations of diurnal traffic patterns. To obtain
the speeds to be used in the model, each road segment is assigned to
one of several classifications (e.g., suburban freeway, local or feeder
street). Each street classification has two "typical" speeds, which are
used for emission calculations. The slower speeds are used for the peak
traffic hour calculations. As with the diurnal traffic cycle, these
typical speeds are based on very limited data.
b. Traffic Data
It is clear from the preceding discussion that there are
at least three potential sources of error in calculating the emissions:
the nature of the equation relating emissions to traffic parameters
(discussed in the next section), traffic volumes, and average speed.
Traffic volumes can be misestimated because we have inaccurate total
24-hour counts or because we have an. inaccurate diurnal assignment
function.
The detailed traffic data from San Jose indicate that
the traffic on any given street link, for a given hour of the day, does
not vary much from one weekday to another. The standard deviations of
day-to-day values of half-hour traffic volumes in downtown San Jose at
a given time of day are generally in the range of 5 to 15 percent of
the average volume. This implies that the standard deviation of the
24-hour traffic volume totals is probably less than 15 percent. Since
emissions are directly proportional to traffic volumes, the emission
errors would be similar in relative magnitude to the traffic volume
errors that cause them.
110
-------
It appears that uncertainties in the diurnal traffic
patterns may be more serious than uncertainties in 24-hour total traffic
volumes. Some evidence of this was presented in an earlier report
(Ludwig et al., 1970). The solid line in Figure 46, taken from that
report, shows the daily emission cycle used for the model. The dashed
line shows a daily emission cycle derived from a statistical analysis
of five years of St. Louis CO observations. This derived emission cycle
0.10
•
ui 0.06
C3
<
o:
w
O 0.04
O
0.02
I
I
ORIGINAL
STATISTICALLY
DETERMINED
10 15
HOUR OF DAY
20
25
TA-7874-70
FIGURE 46 DIURNAL EMISSION PATTERNS FOR ST. LOUIS
111
-------
gives the best agreement between calculated and observed CO concentra-
tion values. The emission cycle is very closely related to the traffic
cycle, differing only in that it includes the effects of reduced rush-
hour speeds. It is evident from Figure 46 that there are considerable
differences between the statistically determined values and those
originally hypothesized for the model. The dashed curve is less peaked
at the morning and evening rush hours than hypothesized. Figure 47
shows the total traffic in the downtown San Jose area on two successive
days, for the hours between 0630 and 1730. These curves are different
from either of those in Figure 46 but are closer to the statistically
determined case in that they show little decrease in traffic at midday.
The fact that the two curves in Figure 47 are nearly superimposed
Z
i
oc
LLJ
Q-
8
I
o
I
I
o
tr
i-
30
25
20
15
10
14 DECEMBER 1970
15 DECEMBER 1970
I
I
I
0600
0800
1000 1200 1400
TIME OF DAY — PST
1600 1800
TA-8563-77
FIGURE 47 TOTAL NUMBER OF TRAFFIC COUNTS FOR ALL DETECTORS
IN DOWNTOWN SAN JOSE
112
-------
suggests that there is very little day-to-day variability and that it
would not take very many observations in a given area of a city to ob-
tain reliable diurnal traffic cycles for that area.
Not enough data were collected on this program to deter-
mine the accuracy of our estimates of speed. Our three-day sample of
95 total circuits around the downtown area indicates an average speed
of about 14 mi h , but there is considerable variation in speed among
route segments and from circuit to circuit, particularly at certain
times, such as 1700. The variability of speed is substantially greater
in the morning and evening than at midday.
Spot speeds can be estimated from the time required for
a vehicle to pass over a magnetic sensor in the traffic network. Limited
data of this type were collected on First Street during morning and
evening rush hours on 6 May 1971. These data indicate that the after-
noon traffic at this location tended to travel more slowly, about 70
percent of the morning speeds, for all comparable traffic volumes. Data
from all sources suggest that there may be considerable variability in
average speeds, even among streets that appear, in other respects, to
be similar. The speed variability is evident, also, on the same street
for equal volumes, when measured at different times of the day.
The model assumes that traffic on downtown arterials
-1 -1
travels at a rate of 24 mi h during peak traffic and 30 mi h at
-1
other times. This is substantially different from the 14 mi h that
we measured in the downtown area. Speed uncertainties of this magnitude
could cause significant changes in calculated emissions; for instance,
according to Eq. (26), the emissions per vehicle-mile at 14 mi hri are
1.7 times those at 30 mi h . Thus, better estimates of vehicle speed
on the various road types and at the different times of day would sub-
stantially improve the performance of the emissions submodel.
113
-------
c. The CO Emission Equation
Ihe special traffic data in San Jose allowed us to check
the validity of Eq. (26) in a situation where the input parameters,
traffic volume, and vehicle speed were accurately known. The results
of the comparison of emission rates determined from the equation with
those determined independently have been presented in Section IV. Since
the independent measurements have substantial uncertainty, they cannot
serve as a precise check on the emissions submodel, but they do lead .us
to believe that the equation is probably at least as accurate as the
traffic data that are usually available. We hope that future work can
improve our assessment of the performance of the emissions submodel.
2. Mixing Depth Submodel
The mixing depth submodel proved to be reasonably accurate,
as shown by the comparison presented in Table 9. The calculated
mixing depths, with one exception, were all within 150 m of observed
values based on helicopter profiles of temperature and carbon monoxide.
The model makes use of the early morning vertical temperature gradient
in the lowest few hundred meters to determine the early morning mixing
depth. It is important that these soundings come from an open area that
is geographically similar to that for which we wish to calculate the
mixing depth. This was the case for the soundings used to calculate
the model values shown in Table 9, and as a result the morning values
were quite accurate. However, when the values were recalculated using
radiosonde data from Oakland, 50 km distant, the calculated depths were
found to be considerably greater, which might be expected.
Afternoon mixing depths, as determined by the submodel, are
also based on the morning sounding, but in this case, the higher levels
of the sounding are generally more important than the lower layers. The
114
-------
Oakland sounding yields overestimates of afternoon mixing depths as well
as of morning mixing depths, but the afternoon values are more nearly
correct, probably because of the smaller effect of the Bay at higher
levels.
The above discussion indicates that some care should be exer-
cised in selecting a source of upper air data for use with the model.
The closest radiosonde station may not give the best results if it is
characterized by substantially different geographical surroundings than
the city of interest. If, however, a representative sounding is avail-
able, the mixing depth submodel appears to perform well.
3. Stability Class Submodel
The performance of the submodel that was originally used to
determine stability was discussed in Section IV. As noted there, and
in Section V, that submodel gives good results except in the hours just
after sunrise and just before sunset. The changes described in Section
V make the submodel more consistent with the observed values of the
bulk stability parameter. However, we have only a limited number of
measurements of the bulk stability parameter, so a full evaluation of
the revised stability index submodel will have to await the availability
of more data. Such data should be available from the field program that
we have recommended be undertaken in St. Louis.
Although we have not collected data suitable for checking the
variations of a with distance from the source, other studies (Leighton
z
and Dittmar, 1952, 1953; McElroy and Pooler, 1968) indicate that the
changes described in Section V will give results more appropriate to
urban conditions.
115
-------
4. Street Effects Submodel
The street effects submodel presumes that the emissions from
traffic on the local street are added to the concentrations from emis-
sions outside the immediate area that enter the street canyon from above.
We assume that the basic diffusion model calculates the concentrations
of CO entering the street canyon and adds to those the concentration
component from the street effects submodel. To evaluate the performance
of the street effects submodel separately requires some measure of CO
concentrations free of the effects of emissions in the immediate area.
Unfortunately, such measurements are not available, since even the top
sampling levels at our streetside stations are exposed to considerable
CO from the adjacent street. Thus, it has been necessary to evaluate
the street effects submodel indirectly, and by inference from the re-
sults of the composite model evaluation.
The street effects submodel is designed for application in a
street canyon, i.e., away from an intersection, near the center of the
block. We would therefore expect it to perform best at sites 7 and 8,
not quite so well at site 4, and poorly at the other intersection sites.
The data collected during the experiments confirm this. It is apparent
that the modeling of intersection effects will require further study,
perhaps including wind tunnel work. Nevertheless, the data collected
in San Jose can be valuable in confirming the applicability of any future
theoretical or laboratory work.
C. Evaluation of the Composite Model
The composite model has been evaluated through comparison of
observed and predicted hourly concentrations (Figures 48 to 62) of CO
for eight days at two levels at each of five stations (Nos. 4, 5, 6,
7, and 8). The predicted concentrations may be resolved into the
116
-------
^!>
20
1
i 15
:
•
J 10
>
5
5
;
5
0
25
20
1
. 15
3™
C
C
j 10
5
>
' 5
o
I I
- 3-m HEIGHT
•••
-
19 NOV
\ TRAFFIC
1 BLOCKED
*""* '""'
_^_
-
1 1
l i i ii i i
\r+ i f*
l*1~ + Ub
cb
Observed
20 NOV 7 DEC
-
V^/ <4 J-
>/y \ / / -A — •-///,-'
^x^ // _
ii i ii i i
1 1
" 15-m HEIGHT
-
19 NOV
I TRAFFIC
\ BLOCKED
•X\
xJir.^
i i
ii i ii i i
AC + Cb
cb
Observed —
20 NOV 7 DEC
'^-7:?rv
J 1 1 II 1 1
08 13 18 08 13 18 08
TIME — PST
13 18
TA-8563-80
FIGURE 48 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 4 AT TWO HEIGHTS FOR 19 AND 20
NOVEMBER AND 7 DECEMBER 1970
117
-------
25
20
15
10
o
o
o
o
0
25
20
n i i r
3-m HEIGHT
Ac + Ch
9 DEC
10 DEC
Observed ~
11 DEC
JS N :
I ° i
I D i
J A;.
I
i
i i
a
a
Z 15
O
cr
UJ 10
o
O
o
o
CJ
i i i r
15-m HEIGHT
Ac -i- C_
Observed —
9 DEC
10 DEC
11 DEC
08
13
18 08 13
TIME — PST
18 08
13
TA-8563-81
FIGURE 49 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 4 AT TWO HEIGHTS FOR 9, 10, AND
11 DECEMBER 1970
118
-------
25
20
DC
I-
§ 10
o
o
o
0
0
25
I
a
GC
O
o
o
(J
20
15
T
3-m HEIGHT
I I
Ac + ch
— c
Observed
14 DEC
15 DEC
I I 1 T
15-m HEIGHT
1 I
---- c
Observed
14 DEC
15 DEC
\ \
08 13 18 08 13
TIME — PST
18 08
13
18
TA-8563-82
FIGURE 50 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 4 AT TWO HEIGHTS FOR 14 AND 15
DECEMBER 1970
119
-------
25
20
I
<
DC
g 10
o
o
o
o
0
25
20
O
r-
cc
H
ui 10
O
z
o
o
I I I I
3-m HEIGHT
19 NOV
TRAFFIC
BLOCKED
II
H I | T
11-m HEIGHT
19 NOV
TRAFFIC
BLOCKED
II
20 NOV
20 NOV
___ AC + c
----- c
- Observed
7 DEC
I T
Ac + CK
---- c
Observed —
7 DEC
1
08 13 18 08 13
TIME — PST
18 08
13 18
TA-8563-83
FIGURE 51 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 5 AT TWO HEIGHTS FOR 19 AND 20
NOVEMBER AND 7 DECEMBER 1970
120
-------
25
a
a
O
LU
O
O
O
O
O
20
15
10
-------
25
20
a
<
DC
r-
O
O
o
O
(J
10
o
25
20
2 15
O
<
cc
I-
| 10
z
O
o
1 I I
3-m HEIGHT
14 DEC
1^ I I
11-m HEIGHT
14 DEC
'X
I I
Ac
Observed ~
15 DEC
I
15 DEC
II
08 13 18 08 13
TIME — PST
18 08
II I I
AC H- Cu
Observed _
13 18
TA-8563-85
FIGURE 53 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 5 AT TWO HEIGHTS FOR 14 AND 15
DECEMBER 1970
122
-------
2b
20
L
a.
- 15
(
C
j 10
z
3
3
5
0
25
20
i
i
»• 15
3
t
C
a 10
P
3
3 5
0
1 1 1 1 1
- 3-m HEIGHT
-
19 NOV 20 NOV
\ TRAFFIC / N
X\ BLOCKED v / °
- V ^ °
\ ',- i
I i ii i
I I I
AC + Cb
cb
Observed ~
7 DEC
\/\ \ i .
VI \ / *
\ 11
^'\ ', -
II 1 1
i i 1 \ \
~ 15-m HEIGHT
-
19 NOV 20 NOV
\
~ \ TRAFFIC
\ BLOCKED \ / g
>^X ^ D
- I V s' A
I III I
08 13 18 08 13
II 1 1
AC + Cb -
cb -
7 DEC
- - "" \ * / *
V J >*
II 1
18 08 13 18
TIME — PST
TA-8563-86
FIGURE 54 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 6 AT TWO HEIGHTS FOR 19 AND 20
NOVEMBER AND 7 DECEMBER 1970
123
-------
25
20
E
a
a
15
<
oc
H
X 10
i
o
(J
0
25
20
15
cc
H
5 10
o
o
o
3-m HEIGHT
9 DEC
N
O
1 1 II
15-m HEIGHT
9 DEC
\I
D:
A:
T !
A;
10 DEC
10 DEC
08 13 18 08 13
TIME — PST
AC + Ch
Observed ~
\ 11 DEC
AC +
Observed _
11 DEC
18 08
13 18
TA-8563-87
FIGURE 55 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 6 AT TWO HEIGHTS FOR 9, 10, AND
11 DECEMBER 1970
124
-------
20
a
a
§ 15
<
DC
I-
| 10
z
o
o
o
o
1
3-m HEIGHT
14 DEC
I I
15 DEC
I
Ac + cb
Cb
Observed
£b
20
1
Z 15
O
<
oc
S 10
2
o
o
0
0 5
0
1 1 1 1 1 1
• 15-m HEIGHT
—
-
—
14 DEC 15 DEC
— /*\
- /Q ^^
*-,
i i it i i
08 13 18 08 13
TIME — PST
AC + Ch
18 08
Observed _
13 18
TA-8563-88
FIGURE 56 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 6 AT TWO HEIGHTS FOR 14 AND 15
DECEMBER 1970
125
-------
25
20
a.
a
EC
H
I 10
o
o
o
o
0
25
I
O
oc
I-
LU
O
o
o
8
20
15
10
I \
3-m HEIGHT
I I
I
19 NOV
TRAFFIC
BLOCKED
20-m HEIGHT
19 NOV
TRAFFIC
BLOCKED
20 NOV
20 NOV
Ac +
c
b
- Observed
7 DEC
1 P
AC + Cb
'-- cb
Observed
7 DEC
08 13 18 08 13
TIME — PST
FIGURE 57 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 7 AT TWO HEIGHTS FOR 19 AND 20
NOVEMBER AND 7 DECEMBER 1970
126
-------
I
25
20
8 10
o
o
o
o
0
25
20
a
a
15
Ju 10
U
O
o
o
o
I I
3-m HEIGHT
9 DEC
I I I I
AC + Cb
Cb ~~
Observed ~
I I
I I
20-m HEIGHT
cb
Observed _
9 DEC
10 DEC
11 DEC
08 13 18 08 13
TIME — PST
18 08
13 18
TA-8563-90
FIGURE 58 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 7 AT TWO HEIGHTS FOR 9, 10, AND
11 DECEMBER 1970
127
-------
25
20
o
<
cc
K
I 10
o
o
o
0
25
20
I
Z 15
O
K
cc
5 10
o
o
o
I I I
3-m HEIGHT
14 DEC
1
1 I I I
AC + Cb
cb
Observed
15 DEC
~TII IT
20-m HEIGHT
15 DEC
AC + Ch
Observed —
08 13 18 08 13
TIME — PST
18 08
13 18
TA-8663-91
FIGURE 59 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 7 AT TWO HEIGHTS FOR 14 AND 15
DECEMBER 1970
128
-------
25
20
I
<
DC
H
o
o
o
o
10
Q.
D.
E
o
o
o
o
O
0
25
20
15
10
I I I
3-m HEIGHT
I I I
Ac +
Observed ~
19 NOV
20 NOV
7 DEC
TRAFFIC
BLOCKED
I I I I
13-m HEIGHT
1 I
Ac
Observed _
19 NOV
20 NOV
' V'\
7 DEC
\ TRAFFIC
r*~" BLOCKED
I I
I I
08 13 18 08 13
TIME — PST
18 08
13 18
TA-8 563-92
FIGURE 60 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 8 AT TWO HEIGHTS FOR 19 AND 20
NOVEMBER AND 7 DECEMBER 1970
129
-------
25
20
a
a
§ 15
cc
I-
o
o
o
o
10
0
25
20
I I
3-m HEIGHT
1 I
--- AC + CK
9 DEC
10 DEC
— Observed
11 DEC
V \
I I
0.
a
2 15
cc
o
o
o
o
10
I I I I
13-m HEIGHT
I I
Ac + a.
9 DEC
I
I I
I
I I
08 13 18 08 13 18 08
TIME — PST
13 18
TA-8563-93
FIGURE 61 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 8 AT TWO HEIGHTS FOR 9, 10, AND
11 DECEMBER 1970
130
-------
25
20
a
a
cc
I-
1 10
z
o
o
o
o
0
25
20
Z
O
o
o
o
o
o
10
T
3-m HEIGHT
14 DEC
I
I
i i i r
13-m HEIGHT
14 DEC
I
Ac + cb
Cb
Observed
15 DEC
I I
1 I
Ac + a.
ch
Observed —
15 DEC
J L
08 13 18 08 13
TIME — PST
18 08
13
18
TA-8563-94
FIGURE 62 CALCULATED AND OBSERVED CO CONCENTRATIONS
FOR STATION 8 AT TWO HEIGHTS FOR 14 AND 15
DECEMBER 1970
131
-------
sum of (1) "street background" values determined from the basic, receptor-
oriented Gaussian diffusion model, and (2) values computed with the
street effects submodel. Evaluation of the composite model indicates
that the predicted values generally agree quite well with the observa-
tions. The figures depict the observed concentrations as well as both
the predicted values from the composite and street background (basic)
models.
Street background concentrations at the five stations were com-
puted from the vehicular emissions in the local downtown area immediately
adjacent to the monitoring stations and added to the mean, ambient urban
CO background representative of conditions farther upwind. The basic
model was used to compute the contribution from the sources within
about 2 km of the stations. Traffic data from the downtown monitoring
network were used for these computations. Because the traffic network is
oriented asymmetrically about the monitoring stations, it is augmented
with historical traffic data (Turturici, 1970) to avoid an unrealistic
dependence of the computed concentrations on wind direction. The enlarged
grid is a circle of 1.9 km radius centered on the monitoring stations.
The historical traffic data give the fraction that each street contributes
to the total mean daily downtown traffic volume. We have assumed that this
fraction is constant throughout the day; the hourly traffic volumes for
these streets were computed from the hourly total traffic volume measured
in the traffic monitoring network. The diurnal patterns for three
different street types (major arterial, high volume collector, and col-
lector) given by Turturici indicate only minor differences in the patterns.
It was not possible to compute the ambient urban background concentra-
tion entering the augmented grid area because we lacked detailed traffic
data for the greater San Jose area; the magnitude of CO sources farther
upwind are also uncertain. Background concentrations at the upwind
edge of the circular grid were derived from the helicopter observations
132
-------
when available. It was assumed that the mean concentration along the
upwind leg at the lowest level (60 m) was a representative value. During
those days when helicopter measurements were not available, subjective
estimates were based on the existing meteorological conditions and the
CO concentration at the 30-m level of Station 9. A majority of the
helicopter flights were made with northwesterly winds. The high correla-
tion found to exist during these periods between the helicopter back-
ground values and the concentration at the uppermost level of Station 9
allowed us to "compute"the background in the absence of helicopter
measurements for days with northwesterly winds. Ambient urban back-
ground concentrations of CO are given in Table 15. Because only
three or fewer helicopter flights were made on any given day, background
values were taken to remain constant during the following periods:
0700-1100, 1100-1600, and 1600-1800 PST.
On 19 November, data were collected during the morning hours only
(see Figures 48, 51, 54, 57 and 60); traffic was blocked on two lanes
of First Street south of San Antonio Street commencing at approximately
0830 PST and continuing throughout the rest of the data collection
period on that day. The effect of this disruption in the traffic flow
is reflected in the generally poor agreement between model and observa-
tions.
The performance of the model on the other seven days (20 November,
and 7, 9, 10, 11, 14, and 15 December) is quite good. In general, both
the trend and the magnitude of the predicted concentrations agree very
well with the observations. During those cases when the composite model
exhibited poor agreement with the observations, we assessed the relative
performance of the basic model and the street effects submodel in order to
identify those areas where additional refinements might be necessary.
We found that both components contributed to poor results approximately
133
-------
Table 15
AMBIENT URBAN CARBON MONOXIDE BACKGROUND CONCENTRATIONS (ppm)
Date
(1970)
19 November
20 November
7 December
9 December
10 December
11 December
14 December
15 December
Time
Period*
I
II
I
II
III
I
II
III
I
II
III
I
II
III
I
II
III
I
II
III
I
II
III
Background
(ppm)
2.7
2.7
2.0
2.0
2.5
2.0
1.0
1.9
0.9
0.7
2.6
5.8
2.4
2.4
5.5
3.0
1.5
2.0
2.0
2.0
1.0
1.0
1.0
Source
>• Measured
Estimated
Estimated
Computed
Estimated
Measured
Measured
v Measured
"|
> Measured
J
>• Measured
> Estimated
1
> Estimated
J
I - 0700 to 1100 PST
II - 1100 to 1600 PST
III - 1600 to 1800 PST
134
-------
the same number of times, indicating that refinements in both may be
necessary. We have identified some of the problem areas and the changes
necessary to rectify them.
Determination of the CO concentration in the air entering the cir-
cular grid is one of the most difficult problems in the computation of
the street background at the receptor. As noted above, we have assumed
that the upwind, 60-m helicopter measurement is a representative value;
this seems to provide a reasonable estimate in the absence of reliable
near-surface observations or detailed traffic data for the greater
metropolitan area. In this manner we have obtained good qualitative
estimates of the changes in the upwind background. The absolute value
of the changes is subject to some uncertainty. The basic limitation
in this procedure, however, is the small number (three or less) of such
measurements that are available during the day. Because of this we have
had to assume that the upwind background remains constant for a period
of several hours and then changes abruptly to the value appropriate to
the subsequent period. This can introduce serious errors in the computed
street background, especially when meteorological conditions are
changing rapidly in the interim between background observations.
A limitation in the street effects submodel is associated with the
specification of the windward, leeward, and intermediate wind direction
sectors (Table 3). Wind tunnel research (Hoydysh, 1971) indicates
that the arcs of the various sectors may not be constant, but rather
are dependent on the intensity of atmospheric turbulence. Hence, they
will be a function of atmospheric stability, wind speed, and the
aerodynamic roughness of the site. This effect will obviously be
most significant at the outer limits of the wind direction sectors
where the flow regime changes abruptly (i.e., from a lateral, cross-
street circulation to longitudinal flow, or vice versa).
135
-------
The street effects submodel is designed for application in the
street canyon away from the influences of intersections. Stations 7
and 8 are most representative of this situation, while Stations 5 and
6 are less representative because of their proximity to San Antonio
Street. Station 4, on the other hand, is quite poor because it is in
a short block with irregular building heights. The effect of location
is illustrated in Table 16, which lists the correlation between
observed and predicted concentrations at each station for all eight
days. As one would expect, the correlation coefficient (r) is highest
for Station 7 (r = 0.68) and lowest for Station 4 (r = 0.47). Because
of the highly uncertain conditions that existed on 19 November, the
correlation coefficient for Station 7, as an example, was recomputed
for the other seven days and increased to 0.71. Figure 63 is included
to illustrate the degree of scatter represented by these correlations.
It has the calculated CO values plotted versus the observed values at
Stations 7 and 8 for all five levels and the same days illustrated in
Figures 57 through 62.
Table 16
CORRELATION COEFFICIENTS (r) BETWEEN OBSERVED
AND PREDICTED CARBON MONOXIDE CONCENTRATIONS
Station
4
5
6
7
8
r
0.47
0.51
0.58
0.68
0.55
Although encouraging, the correlations are not as large as we
would desire. Aside from the problems already cited, one additional
point should be mentioned. The model, in essence, presumes that
changes in the concentration occur simultaneously with changes in
136
-------
the input parameters, e.g., wind speed, stability, and traffic. In
reality, the response of the concentration function will lag changes in
the inputs. This may especially be important during stable conditions
with low wind speeds.
In summary, the composite model predicts CO concentrations that
agree quite well with values measured in downtown San Jose. Additional
refinements to the model have been proposed which we hope to evaluate
during the next phase of the program and incorporate in the final version
of the composite model.
D. Frequency Distribution of Concentrations
The model can also be used to calculate the frequency distributions
of concentration from the hour-by-hour values. This is a somewhat less
demanding use of the model than hour-by-hour prediction, and therefore
better performance might be expected.
For those sites for which it is most applicable, we get good agree-
ment between the calculated and observed frequency distributions of 3-m
CO concentration. Figure 64 shows these results for Stations 4, 7,
and 8. The concentration values represented by these figures are for
the same days used for the calculations described in the preceding
section. As before there is good agreement between calculations and
observations at the midblock Stations, 7 and 8. Station 4 gives poorer
results, as do the stations not shown, 5 and 6. As noted above, the
model's results, when street effects are included, are best for those
stations that best fit the assumptions of the street effects submodel.
137
-------
20
15
I
Q
111
cc
HI
10
0
20
15
O
O
Q
III
cc
in
10
STATION 7
1:1.
STATION 8
5 10 15
CALCULATED CO — ppm
20
TA-8563-100
FIGURE 63 SCATTER DIAGRAM OF CALCULATED VERSUS
OBSERVED CO CONCENTRATIONS FOR ALL
FIVE LEVELS AT STATIONS 7 AND 8
138
-------
oc
ffi
^
z
OT
UJ
0
2
UJ
OC
D
O
0
LL
O
ou
40
30
10
n
r- ,
STATION 4
—
I
_ _____ f
I
I
OBSERVED —
«« — ~ ^ C A LCU LATE D
—
u I
1
oc
HI
m
i
z
OC
UJ
ffi
5
3
Z
10
in
^
UJ
rr
tc
fj
0
O
u_
O
f/i
u
UJ
oc
oc
0
8
LL
O
50
40
30
10
40
30
20
10
0
_
STATION 7
—
STATION 8
—
—
|
—
—
—
4 8
CO CONCENTRATION — ppm
16 32
TA-8563-78
FIGURE 64 CALCULATED AND OBSERVED FREQUENCIES OF ONE-HOUR
AVERAGE CO CONCENTRATIONS
139
-------
VII RECOMMENDATIONS
Our results indicate that the revised model gives very good results
for midblock, street-canyon locations. However, some additional research
is desirable, and the San Jose program has provided valuable information
to define the best directions for future research. It would be very
worthwhile to extend our evaluations of the stability, mixing depth, and
traffic emissions submodels. We would like to examine the wind fields
in urban areas to determine the representativeness of airport wind data.
Finally, it is imperative that the model be tested further under
substantially different "canyon" conditions, i.e., with a ratio of
building height to street width larger than those in the San Jose
experiment.
Analysis of the San Jose data has shown that the instrumentation
could be more profitably used in the future for the detailed study of
the canyon situation alone. The next phase of the program should be
confined to the study of these effects. We can make good use of the
background from our own San Jose studies and from several laboratory
wind tunnel studies (Roshko, 1955; Maull and East, 1963; Fox, 1964 and
1965; Burggraf, 1966; and Hoydysh, 1971).
Study of the extremely complex flow patterns at an intersection
should be deferred and should be first studied in a wind tunnel to provide
qualitative background information for field efforts. It should be
emphasized that wind tunnel simulations are often limited by difficulties
in scaling atmospheric turbulence and stability, and the results need
to be confirmed by field measurements.
141
-------
A user's manual per se is not included in this report; a separate
manual should be prepared upon completion of the recommended additional
research, incorporating further refinements that may result. The manual
should provide potential users with the appropriate computer programs,
and input and output formats and specifications. Additionally, it should
specify the types of problems for which it will serve as a useful diag-
nostic and/or prognostic tool, as well as an account of conditions which
may limit its applicability in certain problem areas.
142
-------
ACKNOWLEDGMENTS
We are appreciative of the support provided by the Coordinating
Research Council and the Air Pollution Control Office (EPA), and of the
assistance furnished by their representatives on the CAPA-3 monitoring
committee: J. F. Black (Chairman), A. P. Altshuller, J. M. Colucci,
R. P. Doelling, C. R. Hosier, R. G. Larson, J. J. Mitchell, J. S. Seward,
and A. E. Zengel.
We are grateful to a number of SRI personnel for their assistance.
E. L. Younker, F. H. Burch, and W. B. Guthoerl were responsible for the
design, building, and installation of the remote coupling units; B.
Wheeler did the programming for the mini-computer. Equipment was in-
stalled in San Jose by A. H. Smith, L. Salas, W. Ward, W. Crossen,
W. Fulcher, and D. Mouton. G. L. Williams, R. Mancuso, H. Shigeishi,
R. Trudeau, J. Kealoha, and B. Ripple assisted in various aspects of
dats reduction and analysis. V. Klein, M. Kucinski, E. Cox, S. Hanson,
D. Orr, M. Taylor, P. Monti, and T. Davis all aided in the preparation
of this and other project reports.
The City of San Jose was most cooperative; particular thanks are
deserved by G. Mahoney and B. Todd of the Traffic Engineering Research
Group. The San Jose Redevelopment Agency was very helpful in arranging
for suitable equipment sites.
Professors A. Miller, K. MacKay, and D. Mage of San Jose State
College provided the project with access to data collected at the college,
and arranged pilot balloon ascents. Professor Miller also made the
arrangements necessary for the lidar van to be parked and operated on
College property.
143
-------
During some of our operations in San Jose, W. Ott of Stanford
University was making special measurements of CO concentration in con-
junction with the Bay Area Air Pollution Control District (BAAPCD).
Much of their data was made available to us, for which we are grateful.
The BAAPCD also provided us with their routinely collected data.
G. Young piloted our chartered helicopter in the heavy air traffic
over downtown San Jose. He also made helpful suggestions concerning
the installation of equipment in the aircraft.
144
-------
Appendix A
FIXED-STATION INSTRUMENTATION SYSTEM
A-l
-------
Appendix A
FIXED-STATION INSTRUMENTATION SYSTEM
1. Brief System Description
The San Jose fixed-station instrumentation system consisted of a
central station and seven satellite remote data-gathering terminals
located in a two city block area. Figure 14 of the text shows the
location of these sites. Figure A-l of this appendix is a schematic
representation of the system. There are two different types of ter-
minals designed to gather data as follows:
Type A Terminal—Carbon monoxide (CO) concentrations are
(two total)
measured at five equally spaced heights
to produce a vertical profile from 3
meters to the top of the building where
installed. Temperature differences are
measured between adjacent levels to ob-
tain a vertical profile corresponding to
the CO profile. Absolute temperature is
monitored at either the top or bottom
level (manually selected). Three-axis
orthogonal (UVW) wind speed is sensed
at the top of the building and near the
3-m level.
Type B Terminal—CO concentrations are measured at five
(five total)
equally spaced heights as at the Type A
terminals. Horizontal wind speed and
directions are measured at the 3-m level.
A-3
-------
Both the Type A and Type B terminals contain Model 1 remote line couplers;
in addition, the A terminals contain a second remote line coupler called
the Model 2. Both model couplers contain an address decoder, analog and
digital multiplexers, an analog-to-digital converter, and associated
modules. The Model 2 couplers also contain circuitry for controlling a
low-level multiplexer used for selecting the various temperature sensor
inputs. All remote couplers received commands and transmitted data via
teletype lines leased from the phone company. Spare channels are avail-
able, so additional sensors can be added if needed.
"Mode code" generators were included so that "fixed" data such as
the selected "range" for wind measurements could be supplied to the
central station; this allowed the correct scale factor to be selected by
the computer during data processing.
The two types of terminals also had differences in the support
systems used to suspend the sensors from the buildings. These dif-
ferences arose from the differences in the set of parameters measured
at the two types of terminals. The Type A terminals required a more
complicated suspension system than the Type B, because of the greater
numbers of parameters to be measured.
The central station contained a small minicomputer/controller, a
teletype unit, and magnetic tape storage. The purposes of this central
station are (1) to interrogate the sensors at the various terminals;
(2) to select the heights at which CO was to be monitored; (3) to store
the gathered data returned by the seven satellite terminals; (4) to per-
form certain computations with part of the data; (5) to generate and
place a "time elapsed" tag with each group of data collected; (6) to
log the information digitally on magnetic tape; and (7) to print sum-
maries of processed data periodically for monitoring purposes.
A-4
-------
r
STATION ( S J
I TYPE B L-
TERMINAL P~
INPUTS /
(FIRST STREET!
i kr
1 TERMINAL f"
I __J
r 1
I TYPE B L
| TERMINAL I
L 1
TERMINAL Is
MAGNETIC TAPE ADAPTER
*
NOVA COMPUTER
CENTRAL PROCESSOR
AND CORE MEMORIES
TELETYPE ' TELETYPE
CONTROL | CONTROL
NO. 2 i NO. 1
MAGNETIC
TAPE
RECORDER
TELETYPE
CENTRAL STATION
n
TYPE A TERMINAL
@
REMOTE
LINE
COUPLER
Model 2
t
CO
UNITS
TEMPERATURE
UNITS
DUAL-
MODE CODE
GENERATOR
REMOTE
LINE
COUPLER
Model 1
THREE- AXIS
WIND
SPEED
UNITS
VJLX
CO
UNITS
HORIZONTAL
WIND
VELOCITY
UNITS
/
REMOTE
LINE
COUPLER
Model 1
t
SINGLE-
MODE CODE
GENERATOR
TYPE B TERMINAL
r
TYPE A TERMINAL
' AIR
INLET
TUBING
TEMPERATURE
SENSOR
INPUTS
(FIRST
STREET)
WIND
SENSORS
INPUTS
®
REMOTE
LINE
COUPLER
Model 2
t
CO
UNITS
TEMPERATURE
UNITS
* •
*
•
-rJ
~H
DUAL-
MODE CODE
GENERATOR
i
• (£>
REMOTE
LINE
COUPLER
Model 1
THREE-AXIS
WIND
SPEED
UNITS
*
AIR
INLET
TUBING
TEMPERATURE
SENSOR
INPUTS
ISAN ANTONIO
STREET)
WEND
SENSORS
INPUTS
T6-8S63-23
FIGURE A-1 BLOCK DIAGRAM OF FIXED-STATION INSTRUMENTATION SYSTEM
-------
In the following section the two types of terminals are described,
starting with the sensor supports and proceeding through the instrumen-
tation system in the direction of the data flow. The final section in
this appendix contains a description of the central station.
2. Description of Terminals
a. Sensor Support System
The prime requirements in the design of the two different types
of sensor support systems were that they: (1) provide for the sensors to
be suspended above the sidewalk and clear of the building; (2) be capable
of supporting the weights involved without hazard to pedestrians; (3)
not damage the roof or building; (4) be transportable and easy to in-
stall even on sloped roofs; and (5) not allow the suspended sensors to
move about substantially. The sensor support systems shown in Figures
A-2 and A-3 accomplished these goals with only two minor problems
*
evolving from incorrect installations. As can be noted, the complete
sensor support system consists of a roof-located boom and counter-weighted
support, rope(s) holding the air (CO) inlets and temperature aspirators,
and the lower support boom installed 3 meters above the sidewalk.
Standard commercial 1-inch galvanized water pipes and fittings
were used to support the single upper booms used at Type B terminals.
At the Type A terminals, 1-1/4-inch pipes were used to support the dual
upper boom. A 1.5- by 4-ft plywood pallet was used for mounting four
front pipe floor flanges, and a 2.5- by 4-ft pallet for mounting four
*
A truck hit one lower boom after it was moved from over a wide sidewalk
on First Street to a narrower one on San Antonio Street. High winds
bent a single roof boom that was improperly secured to an existing roof
structure.
A-7
-------
>
oo
COUNTER
WEIGHT
PALLET^ 1 A
BUILDING
FELT
PADDING~
_i
D-
D~
D-
D-
i
-
-q
-q
-z-
^-
•q
q-
^
i
^
3
THREE-AXIS WIND SPEED SENSOR
(UVW ANEMOMETER)
SIDE
WALK
ASPIRATED TEMPERATURE-
RADIATION SHIELD AND
AIR (CO) INLET FILTER
TA-8563-24R
FIGURE A-2 TYPE A TERMINAL SENSORS AND SUPPORT SYSTEM
-------
FIGURE A-3 DUAL ROOF BOOM—PART
OF SENSOR SUPPORT SYSTEM
rear flanges plus providing an area for the concrete block and/or sand-
bag counter weights.
The upper booms were 1-1/2-inch rigid electrical aluminum
conduit and fittings instead of steel as used in the support. The
longest boom was assembled from two 10-ft sections and extended approxi-
mately 12 ft beyond the front steel pipe support. This front support
extended 2 ft above the aluminum boom and supported an aircraft control
type cable used to back-guy the far end of the boom overhang, as indi-
cated in Figure A-3. Turn buckles and thimbles were used to adjust the
tension on the guy wire. Four mercury leveling type switches were in-
stalled parallel to the two horizontal axes of the UVW sensor at the
tip of the boom. Four lights installed in a remote readout box indi-
cated when the wind sensor was properly oriented. One axis was leveled
by rotating the booms^ the other by adjusting guy-wire tension.
A-9
-------
The lower support boom was constructed of a 1-inch rigid con-
duit and installed at a height of about 3 meters. A 1-ft-square, felt-
backed piece of marine plywood was mounted on the end of a 10-ft section
of conduit that fit against the building. Fittings as shown in Figure
16 were used to support the wind sensors and to provide means for
fastening for the guy ropes on this lower boom. Mercury leveling switches
were also used on this boom.
For the Type A terminals, two low-stretch 1650-lb test ropes
were used to support the aspirated temperature sensor assembly, and the
air inlet filter and associated tubing. These units are pictured in
Figures 16 and 18 of the text. For the Type B terminal, a single
low-stretch rope supported the air inlet filters and tubing as shown in
Figure 16.
During installation these ropes were measured and laid along
the sidewalk. The aspirator assemblies, the wiring, the air inlet
filters, and the tubing were attached to the ropes so that the sensors
would be properly located. The wire, tubing, and ropes were then pulled
through the upper boom support until all sensors were suspended in their
correct positions. The lower ends of the rope were secured to the lower
boom to prevent sway. Cabling and tubing were routed from the roof,
over the edge of the building, and through a window into the room con-
taining the instrumentation units.
b. Carbon Monoxide (CO) System
The CO measuring system was the same at both types of ter-
minals. It consisted of five rain-proof inlet filters, each attached
to 200 ft of 1/4-inch-ID low outgassing polyethylene tubing, a five-
inlet selector unit, and a Beckman model 315-AL CO analyzer (see Figure
A-10
-------
A-4). A block diagram of the system is shown in Figure A-5; specifica-
tions are listed in Table A-l.
FIGURE A-4
BECKMAN CARBON
MONOXIDE ANALYZER
WITH REMOTE LINE
COUPLER
Upon receiving a command from the remote coupling unit, one
of five solenoid valves was energized; it opened so that air from a
selected level could be pumped through the CO analyzer. The remaining
four inlet lines were continuously purged by a pump located within the
selector unit. Sampled air passed through a diaphragm pump in the CO
analyzer; then excess air was bled off through a 2-psi relief valve. A
needle valve was used to adjust the flow rate to 2 liters min , as
monitored with a rotameter. This flow flushed the cell every 8.2
seconds.
The analyzer section uses a double-beam optical system to
measure differential absorption of infrared energy. Two infrared sources
A-11
-------
Table A-l
MANUFACTURER'S STATED PERFORMANCE SPECIFICATIONS FOR THE
NONDISPERSIVE, INFRARED CO ANALYZER
Operating Specifications:
Range
Flow rate
Output
Linearity
*
Zero drift (maximum)
*
Span drift (maximum)
Response speed (amplifier)
Response speed (analyzer)
Sensitivity
Repeatability
Interference
Environmental Specifications:
Ambient temperature range
Physical Specifications:
Weight
Power
0 to 50 ppm by volume
2 j&pm
0 to 1.0 V
(curve provided)
±1 percent of full scale per 8 hours
±1 percent of full scale per 24 hours
0.5 second (90 percent)
(not reported)
0.5 percent of full scale
±3.0 percent of full scale
HO (3 percent equivalent to 1 ppm
£t
response)
-20 to 120° F
110 Ib
410 watts, 115 + 15 V ac, 60 ± 0.5 Hz
Span and zero drift specifications are based on ambient temperature
shifts of less than 40° F at a maximum rate of 20° F per hour.
A-12
-------
L
.±o
GasV_y
Upscale (Span)
Calibration
Chopper Motor -_ 1
SoUrce--~ U (
Tobmg Con
1
-, U -"Source
necnons^-- ^
^ix^
-Tvrl—
1 — — •'
Ve
^
Sample | Valve
Referei
Chamber
^Diaphragm
Undistended
Dram
Filter
with
Condensate
Trap
~1
10 MHr
OSCILLATOR
-^
AMPLITUDE
MODULATOR
-*
RF
DEMODULATOR
FILTER
-*•
AC AMPLIFIER
AND PHASE
INVERTER
-*"
10 Hz
SYNCHRONOUS
DEMODULATOR
-^
FILTER
AND DC
AMPLIFIER
BECKMAN CO ANALYZER
(
CHART
RECORDER
'
Non-Absorbing Molecules
Infrared Absorbing CO
Input
to Remote
Coupling
Unit
• A
)
-X
alyzer
ER
Vent-*—
1
'
P
XX
4-Lin
^Pump
T
LEVEL 5
DIRECTIONAL
CONTROL
SOLENOID
VALVE
*
VALVE
1
3
VALVE
*
VALVE
1
VALVE
1
200 ft of I/
Lovu-Outgas
^
-inch ID
Tub,ng I
F
Ra
an
Inlet
FIVE INLET SELECTOR UNIT
FIGURE A-5 BLOCK DIAGRAM OF
CO MEASURING SYSTEM
USING BECKMAN ANALYZER
-------
are used, one for the sample energy beam, the other for the reference
energy beam. The beams are blocked simultaneously ten times per second
by the chopper, a two-segmented blade rotating at 5 revolutions per
second. In the unblocked condition, each beam passes through the asso-
ciated cell and into the detector.
The sample cell is a flow-through tube that receives a continu-
ous stream of sampled air. The reference cell is a sealed tube filled
with a reference gas. This gas is selected for minimal absorption of
infrared energy in the wavelengths absorbed by CO.
The detector consists of two sealed compartments separated by
a flexible metal diaphragm. Each compartment has an infrared-transmitting
window, and is filled to the same sub-atmospheric pressure with CO. This
detector responds only to the net difference in transmitted IR energy
resulting from absorption by CO in the sample cell. Inside the detector,
gas in the reference chamber is heated more than in the other chamber be-
cause less energy has been absorbed from the reference beam. The higher
temperature of the gas in the reference chamber raises the pressure in
this compartment above that in the sample chamber and distends the dia- .
phragm toward the sample chamber. The diaphragm and an adjacent stationary
metal button constitute a two-plate variable capacitor. Distention of the
diaphragm away from the button decreases the capacitance.
When the chopper blocks the beams, pressures in the two chambers
equalize, and the diaphragm returns to the undistended condition. As the
chopper alternately blocks and unblocks the beams, the diaphragm pulses,
thus changing detector capacitance cyclically. The detector is part of
an amplitude modulation circuit that impresses the 10-Hz information
signal on a 10-MHz carrier wave provided by a crystal-controlled radio-
frequency oscillator. Additional electronic circuitry demodulates and
filters the resultant signal, yielding a 10-Hz signal that is amplified,
A-15
-------
phase inverted, and synchronously rectified. The resulting fullwave-
rectified signal is filtered and the dc voltage is amplified to drive
the meter and recorder and to provide a signal to the remote line coupler
unit. The signal is proportional to the CO concentration in the sample
cell.
c. Temperature System
The temperature system was required to provide one absolute
temperature in the range from -25° C to +50° C and four differential
temperatures with a resolution of 0.01° C. Originally it was planned
to have electronics capable of operating on the roof in an environment
that could have ranged from -25° C to +50° C. The sensors had to operate
in the electrical noise environment of the city, and at distances of
200 ft from the electronics unit. To simplify signal processing, we
wanted sensors with a linear signal output over their operating range.
Their response to temperature had to be stable for periods of at least
3 months. The temperature system shown in the block diagram of Figure
*
A-6 generally accomplished these goals.
The complete temperature system consisted of: (1) the
aspirated-radiation shielded sensor; (2) 200 ft of AWG #28 eight-conductor
cable with an overall shield and jacket; (3) four differential and one
absolute temperature bridge networks with a common 10-V dc power supply
installed on one chassis; and (4) on a second chassis, a gold-plated
low-level stepping switch and associated control circuitry, low-noise/
low-drift differential amplifiers with automatic gain changing, meters
c
The absolute temperature measurement was eliminated during the San Jose
experiments because of ground-loop problems among the absolute and dif-
ferential bridges, their common 10-V power supply, and the lead-wire
shielding. This problem has subsequently been corrected.
A-16
-------
I 1
READY TO OPERATE
INTERLOCKS
Light 'on' when:
1 Aspirator on
2. Auto-manual switch m 'auto
3. Oper-calib switch in 'oper'
4 Mode-Code Unit oper-calib
switch in 'oper'
5. Stepping switch in 'home'
position
Front Panel
Yellow Lights
Air temperature IT, or T5)
Amplifier reference voltage*
Amplifier zero*
- T
- T3 2c FULL
T - T3 ( OR SCALE
- T 6°C
4-IAIire Lead Compensatio
(200 feet AWS #28»
Dual Platinum Resistance Wire
Temperature Sensors and
Aspirated Radiation Shield
Bottom Level
13-m Above Sidewalk}
FIGURE A-6
BLOCK DIAGRAM OF
TEMPERATURE SYSTEM
-------
for visually reading the absolute and differential temperatures at in-
dividual levels, and interlocking circuitry to ensure that the various
switches are in the proper position during operation.
*
Dual platinum resistance-wire-type temperature sensors were
installed at each level. The small additional cost was paid for this
type of sensor because of its inherent linearity (as opposed to thermis-
tors), its long-term stability, and its relatively high signal output
(as opposed to thermocouples). The dual resistance wire sensors were
installed within a 1/4-inch-OD stainless steel housing. Their calculated
time constant is about 40 seconds for the 15-ft s ventilation used. A
commercial low thermal-conduction teflon bushing was purchased and the
stainless steel housings were installed within a silvered, double-walled
glass cylinder open at both ends, but otherwise similar to a Dewar flask.
The silvered cylinder provides a radiation shield for the temperature
t
sensor. An aspirator assembly draws air upward through the end of the
cylinder, and past the stainless steel housing containing the dual sen-
sors. A small blower located at the other end of the aspirator-radiation
shield assembly provides the ventilation. The entire assembly is pic-
tured in Figure 18 of the text.
The cabling was connected to the sensor leads with crimped,
uncoated copper connectors to minimize thermocouple errors. This cable
was shielded and supported by a different rope than the 110-V ac aspira-
tor power cable. This minimized 60-Hz pickup. Four-wire lead compensa-
tion was used to minimize the effect of the differences in resistances
of the lead wire and changes in the lead wire resistance due to tempera-
ture effects.
*
Model 104MK-57-BB-CC, Rosemount Engineering.
No. 43404, R. W. Young Company.
A-19
-------
A separate chassis with a 3.5-inch panel for rack mounting
contained four plug-in hermetically sealed differential bridge networks,
one plug-in type absolute bridge network, and a 10-V dc regulated floating
power supply. These bridge networks contained both span and zero ad-
justments for calibration purposes. Variations in the supply voltage,
if any, were monitored each time data were read by supplying an attenu-
ated voltage just under 1 V dc to the remote line coupler. A further
attenuated bridge voltage is also used as one of the inputs to the dif-
ferential amplifiers, thereby providing a means for checking amplifier
gain and determining long-term drift. These networks for attenuating
the bridge voltage where assembled within plug-in modules similar to
those used for the bridge networks and plugged into spare sockets on
this bridge chassis.
Wires carrying the absolute temperature signal, the four dif-
ferential temperature signals, and the calibration signals were routed
to another chassis with a 5.25-inch rack-mounting front panel. The
calibration signals consisted of the previously mentioned attenuated
bridge voltage used as an amplifier reference (span) voltage and a 100-
ohm "short circuit" which is used as a differential amplifier zero.
These two signals, when compared to the 1-V dc attenuated bridge voltage,
allow any amplifier drift and/or bridge voltage fluctuations occurring
while gathering the experimental data to be determined during off-line
data processing.
Upon receiving an "advance pulse," a gold-plated low-level
stepping switch sequentially selects inputs and routes the individual
signals to low-noise/low-drift differential amplifiers. Additional
contacts on this stepping switch are used to select various combinations
A-20
-------
*
of feed-back/input resistors that control the amplifier gains. Since a
separate, center-zero type meter was required for visual readout of the
differential temperatures, one of the stepping switch levels was used to
select the correct meter. The meters are disconnected from the circuit
during automatic stepping to minimize noise during the ten-step-per-second
cycling run.
The fourth level of the stepping switch selects the indicator
lights installed on the front panel of the multiplexer unit. The lights
indicate the channel selected and are particularly useful when the
channels are selected by manual stepping.
To ensure that the stepping switch always starts with the same
selected input, a reset pulse is supplied from the Model 2 remote line
coupler following the completion of the advanced pulse stepping sequence.
Also a manual reset push button is provided to return the stepping switch
to its "home" or "rest" position after completing some manual readout or
calibration processes. To check that the stepping switch is in the
"home" position prior to automatic operation, a "ready to operate" inter-
lock has been included. This interlock as well as the four others indi-
cated in Figure A-6 provide power to a green "unit ready" indicator light
when all switches are in their correct position for system operation under
control by the central station. The rack containing the temperature
system electronics is shown in Figure A-7.
jt
The gains of the differential amplifiers are 100 when measuring the
absolute temperature; they are 500, 250, and 100 when measuring dif-
ferential temperatures on the 1°, 2°, and 5° C ranges, respectively.
A-21
-------
1. Remote Line Coupler
2,3. UVW Wind Component Indicators
4. Temperature Bridge
5. Temperature Multiplexer Amplifier
6. Mode Code Generator
7. Line Coupler Power Supply
FIGURE A-7 TEMPERATURE AND UVW ELECTRONICS
A-22
-------
d. Wind System
1) Three-Axis (UVW) Orthogonal Anemometer
*
The Gill UVW anemometer is a wind instrument designed
for direct measurement of three orthogonal vectors of the wind. Three
helicoid propeller sensors are mounted at right angles to each other on
a common mast with sufficient separation between propellers so that there
is no significant effect of one on the others for normal wind measure-
ments. The two propellers sensing horizontal components of wind are
designated U and V; and the third sensing the vertical component of wind
is designated W. This instrument is shown in Figure 17 of the text.
The foamed polystyrene propellers respond only to that component of the
wind parallel to the axis of rotation. Both forward and reverse air
flow are measured. The propeller stops rotating when the wind is per-
pendicular to the axis. The propeller response very closely approximates
the cosine law, making the instrument especially suited for vertical
wind component measurements.
The propeller rotates 0.96 revolutions per foot of wind
for all wind speeds above 4 ft s (2.7 mi h ). Slippage increases
down to the threshold speed of 0.8 ft s (0.5 mi h ). It has a dis-
tance constant of 94 cm. The propeller drives a miniature dc tachometer
generator providing an analog voltage output that is proportional to
wind speed. The design allows for an optimum dynamic response in winds
ranging from threshold to 50 mi h
Model 27002/27302, R. M. Young Company.
+
The distance constant is the wind passage required for the sensor to
reach 63 percent of a stepwise change in wind speed.
A-23
-------
The two propellers in the horizontal plane were aligned
parallel and perpendicular to the street. The horizontal sensors provide
a positive signal for wind flow from the front of the sensor; this causes
counterclockwise propeller rotation. With wind flow from the rear of the
sensor, propeller rotation and signal polarity reverse. The W or
vertical sensor receptacle is wired for opposite signal polarity, so
wind flow from the front of the sensor (a down draft) provides a nega-
tive signal for the meter and remote line coupler.
The lower housing of the UVW anemometer contains a small
continuous-duty air blower with a polyurethane foam intake filter. This
blower keeps the internal section of the instrument under a slight
positive pressure. Filtered air moves continuously up the mast and out
through each of the three sensors to the propeller hubs, where it is
expelled to the atmosphere, preventing moisture and dust from entering
the precision ball bearings and other internal parts.
The UVW anemometer signals are transmitted through a
multiconductor cable to the "indicator-translator" unit. This unit,
shown in Figure A-7, has three 4.5-inch meters that indicate the speed
of each wind component in miles per hour. Two ranges are available:
-25 to +25 mi h and -50 to +50 mi h . A switch on the front panel
selects the desired range, which can be calibrated independently. A
separate adjustment is provided for the output to the remote coupler,
allowing the scale factor to be preadjusted. The instrument is cali-
brated with a small motor (Young Company model 27230 calibrator) that
is connected to the propeller axle through a flexible shaft. This cali-
brator rotates at a constant 1800 rpm (using 60-Hz power). This corre-
sponds to 21.3 mi h and meter readings and output voltages are adjusted
accordingly.
A-24
-------
2) Horizontal Wind Velocity Sets
Two different types of measuring sets were used to de-
termine horizontal wind speed and direction. Climet model CI-3 units
were used at four sites and a military type AN/GMQ-12 was used at one
site. The Climet wind speed sensor is a three-cup anemometer with a
threshold less than 1 mi h and a 5-foot distance constant. The wind
speed transmitter produces a frequency signal that is proportional to
the wind speed from threshold to full range (60 mi h ).
The Climet wind direction sensor has a threshold less
than 1 mi h . It uses a potentiometer to produce a voltage proportional
to wind direction over a range of 354°. There is a "dead region around
0°/360°.
Adjustable dc output voltages from the Climet were routed
to the remote coupler and to local meters. The accuracy of the voltages
in representing the wind is ±1.3 percent or ±0.15 mi h } whichever is
greater, and ±4° in azimuth. Three ranges are manually selectable: 15,
30, and 60 mi h . The computer was informed of which range had been
selected by also manually adjusting a similar appearing rotary switch
on the mode code unit discussed in the next section.
Wind speed and direction indicators were calibrated by
removing sensor cables from the translators and inserting those from a
special calibrator. Calibration points were provided for 0, 6, 15, and
30 mi h , and for 0°, 180°, and 360°. The indicated wind direction was
also checked by manually holding the wind vane at known angles.
The AN/GMQ-12 wind measuring set is similar to the Climet.
It also uses a three-cup anemometer. It has four wind speed ranges: 6,
12, 30, and 60 mi h . The wind direction covered a range of 354° with
accuracy of ±3 percent at full scale. The wind speed accuracy was ±4
A-25
-------
percent of full scale, with a threshold less than 1 mi h . Provisions
are incorporated within the electronics circuitry for a cursory check of
the full-scale and zero outputs of the wind direction channel, but no
easy method is available for wind speed calibration in the field. Direc-
tion output was checked by manually holding the vane at known angles, as
with the Climet instruments.
e. Mode Code Generator
The mode code units were designed to provide the computer with
a signal that indicates "fixed" data. These fixed data include the
selected ranges of the wind and temperature measuring units, and whether
each unit is being calibrated. Provisions have also been included to
inform the computer which of two possible orientations of the wind
direction sensor is being used. This latter provision was incorporated
since the horizontal wind direction sensors have a gap near their 360°
point and could require a change of orientation dictated by prevailing
winds. The actual voltages for various switch combinations and their
corresponding acceptable octal ranges, as indicated by the remote
coupler unit, are shown in Table A-2.
The mode code generator has a dc power supply, a resistor-
ladder network, and a series of switches similar to a digital-to-analog
converter. The generator used to supply the code for the two remote
couplers located at Type A terminals has a second ladder, but only the
one dc power supply. Since the range of analog voltages for a given set
of switch positions was ±10 mV, an expensive matched-resistor ladder
network was not required.
A-26
-------
Table A-2
MODE CODES FOR INDICATED REMOTE COUPLING UNITS
to
Mode Code
Generator
Output
(mv ± 10)
15.6
46.9
78.1
109.4
140.6
171.9
203.1
234.4
265.6
296.9
328.1
359.4
390.6
421.9
453. 1
484.4
515.6
546.9
578.1
609.4
640.6
671.9
703.1
734.4
765.6
796.9
828.0
859.4
890.6
921.9
953.0
984.4
Acceptable Octal
Range for Wind
Direction
Orientation
Number 1
0-3
4-7
10-13
14-17
20-23
24-27
30-33
34-37
40-43
44-47
50-53
54-57
60-63
64-67
70-73
74-77
100-103
104-107
110-113
.114-117
120-123
124-127
130-133
134-137
140-143
144-147
150-153
154-157
160-163
164-167
170-173
174-177
Number 2
374-377
370-373
364-367
360-363
354-357
350-353
344-347
340-343
334-337
330-333
324-327
320-323
314-317
310-313
304-307
300-303
274-277
270-273
264-267
260-263
254-257
250-253
244-247
240-243
234-237
230-233
224-227
220-223
214-217
210-213
204-207
200-203
Remote Coupler
Units 1 and 3
Calib. (0)
Oper. (1)
low UVW
0
0
1
1
high UVW
0
1
0
1
Remote Coupler Units 2 and 4
Delta T
Multiplier
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
5
1
1
1
1
1
1
1
1
Tl (1)
T, (0)
Ta
0
0
0
0
1
1
1
1
0
0
0
0
1
I
1
1
0
0
0
0
1
1
1
1
Calib. (0)
Oper. (1)
AT,T,,
0
0
i
i
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
CO
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Remote Couplers
5, 6, 7, and 9
Wind Speed
(mi IT1)
15
1
1
1
1
1
1
1
1
30
1
1
1
1
1
1
1
1
60
1
1
1
1
1
1
1
1
Calib. (0)
Oper. (1)
WS
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
CO
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
WD
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Remote Coupler 8
Wind Speed
Range
(mi IT1)
6
1
1
1
1
1
1
1
1
12
1
1
1
1
1
1
1
1
30
1
1
1
1
1
1
1
1
60
1
1
1
1
1
1
1
1
Calib.
Oper.
WS
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
CO
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
(0)
(1)
WD
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
-------
f. Remote Line Couplers
The block diagram (Figure A-8) shows the main functional ele-
ments of the remote line coupler unit. The inputs from the previously
discussed signal-processing units are connected to the high-level analog
multiplexer that acts as a single-pole selector switch. This module
connects one input at a time, at 0.1-second intervals, to the input of
the analog-to-digital converter in the sequence shown in Table A-3. The
multiplexer controls are the "reset/' which returns the switch to a rest
(home) position, and the "advance," which steps it along to the next in-
put. The rest position always corresponds to the mode code signal.
The analog-to-digital (AD) converter, when commanded by a
"convert" signal, determines a binary number corresponding to the value
of the input voltage within a range of 0 to ±1 volt. The conversion
process takes about one millisecond, after which an end-of-conversion
signal is generated that informs the control circuitry that the AD con-
version outputs are ready for use. The end-of-conversion signal also
steps the multiplexer to the next input.
The output of the AD converter, the contents of the register
indicating the CO sampling level, and the output of the address code
switch are selected by the digital multiplexer to make up the output
message. This is done one character at a time by setting up a shift
register in the parallel-to-serial converter, under control of signals
generated by the output control section.
The remote coupler unit is in a quiescent condition most of
the time. The busy signal is off (forcing the analog multiplexer reset),
and the AD converter is set to an internal trigger mode, so that it will
continually measure the mode code output and display this on the unit's
binary indicator lights.
A-28
-------
Command
Line
I I Advance Pulse
I Reset Pulse
Outputs to Temperature
Low-Level Multiplexing Unit
[Model 2 Only!
FIGURE A-8
BLOCK DIAGRAM OF
REMOTE LINE COUPLERS
-------
Table A-3
REMOTE LINE COUPLER ASSIGNED SCANNING SEQUENCE
oo
Model 1 Remote Line Coupler
Input Jack
J 11
(Rest)
J 12
J 13
J 14
J 15
J 16
J 17
J 18
J 19
J 20
J 21
J 22
Measurement
(Type A Terminal)
Mode Code
U. (Lower level
wind component)
V-, (Lower level
wind component)
W, (Lower level
wind component)
U2 (Upper level
wind component)
V2 (Upper level
wind component)
Wo (Upper level
wind component)
(Spare)
(Spare)
(Spare)
(Spare)
(Spare)
Measurement
(Type B Terminal)
Mode Code
CO
V (Wind speed,
lower level)
9 (Wind direction,
lower level)
*
(Spare)
(Spare)
(Spare)
(Spare)
(Spare)
(Spare)
(Spare)
(Spare)
Model 2 Remote Line Coupler
Input Jack
J 11
(Rest)
J 12
J 13
J 14
J 15
J 16
J 17
(Dwell)
Measurement (Type A Terminal)
Mode Code
CO
(Spare)
(Spare)
Ground (Remote unit zero
calibration)
Bridge voltage
Absolute (air) temperature
Amplifier reference voltage
T - T
2 1
T3 -T2
T4 -T3
T - T
5 4
Amplifier zero
f
Subsequent inputs without data were not scanned.
changing a jumper wire.
Last input jack scanned is controlled by
-------
When the computer sends a command message, the first character
is shifted into a register in the serial-to-parallel converter in each
remote unit. This character is compared to the setting of a thumb-
wheel address switch. In one remote unit the two will match; in the
remaining remote units, there will be no match and the message is ig-
nored. In the one remote unit where the match is found, an ACR (Address
Code Recognized) signal is generated. This activates its control module,
which starts a sequence of operations. First, the contents of the level
register are directed to the output circuits, and a character output
cycle is started. While this is going on, the second character of the
command message is being received. This is the level code, and after
it has been received and checked, it is transferred into the level
register. Near the end of the output cycle a convert command is issued
to the analog-to-digital converter, and the converter measures the mode
code output. Next, the AD converter output signals are directed to the
output module and another character cycle started. The end-of-conversion
signal steps the analog multiplexer to the next signal.
The sequence continues, with a new input signal being digitized,
the digital word being sent to the computer, and the next signal being
selected for digitizing, until all signals connected in the particular
remote unit have been sampled. At the end of the last measurement trans-
mission, the output of the address code switch is directed to the output
module and sent as the last character of the data message. The unit
returns to its quiescent condition, with the analog multiplexer reset
and the AD converter operating on internal trigger, continuing to measure
and display the mode code output. The computer then issues a new command
that will activate another unit.
A-32
-------
Two of the nine remote line couplers have been modified
*
slightly for use with the temperature system. These Model 2 couplers
provide an advance pulse at a 10-pps rate for stepping the low-level
temperature multiplex unit. These signals are provided only after the
remote unit has reached its "dwell" input, which is the J 17 input
(Table A-3) of its high-level analog multiplexer. After a sufficient
number of advance pulses have been sent to the temperature multiplexer
unit, the Model 2 remote unit sends a "reset" pulse that returns the
temperature multiplexer to its "rest" position, completing the cycle.
All the other functions of the Model 2 remote unit are the same as that
of the Model 1.
3. Description of Central Station
As indicated in Figure A-l, the fixed-station instrumentation system
consists of seven remote terminals, described in the preceding section,
and the central station, described in this section. The central station
pictured in Figure A-9 was at the location marked "l,2" in Figure 14
of the text. It consisted of the NOVA computer, magnetic tape adapter,
digital magnetic tape recorder, teletype unit (including a paper tape
punch and reader), and a line coupler.
a. Line and Line Coupler
The nine remote line couplers located within the seven ter-
minals were all connected in series with each other and to the computer
through two dc (20 mA) teletype loops installed by the telephone company.
One loop, the command line, was a series connection of the computer
*
See dashed lines in Figure A-8.
A-33
-------
1. Line Coupler
2. Digital Magnetic Tape Recorder
3. Magnetic Tape Adapter
4. NOVA Computer
5. Teletype
FIGURE A-9 MINI-COMPUTER AND PERIPHERAL DEVICES AT CENTRAL STATION
transmitter contacts and the nine remote unit receiver terminals. The
other loop, the data line,, was a series connection of all the remote
transmitter contacts and the computer receiver terminals. Thus,, the
computer transmits to all remote units simultaneously, but only one re-
mote unit at a time transmits to the computer.
The computer terminals were the output of a line coupler unit
that converts the low-voltage (approximately 20 V) operation of the
computer's Teletype Control No. 2 interface to the high voltage (approxi-
mately 100 V) needed to drive the lines. This coupler provides dc iso-
lation, through an optically coupled light-emitting diode isolator,
between the low-voltage computer loops and the high-voltage remote loops.
At the same time, it permits data to flow straight through. The coupler
also supplies the voltage necessary to establish the remote loop currents.
A-34
-------
Separate isolators are used for each of the remote loops. Meters and a
potentiometer allow the loop currents to be adjusted to 20 mA.
b. Computer
*
A general-purpose NOVA mini-computer was used as the control
and real-time data-processing unit for the fixed-station instrumentation
system. As used in the San Jose operations, the computer interrogated
and directed the changing of sampling heights at the seven CO measuring
sites. This took about seven seconds of every minute. During the re-
maining 53 seconds of the minute, the readings of the U, V, W sensors
were recorded. All the data, together with the time from the computer's
real-time clock, were stored in the memory until the available space had
been filled. It was then transferred to magnetic tape through the
adapter described in the next section.
In addition to interrogating the remote units, and receiving
and recording their data, the computer also did some preprocessing.
After each of the five CO intake heights had been sampled at each of
the stations, the computer prepared and printed, on its peripheral tele-
type, a summary of data from the preceding five-minute period. An
example of this printout is given in Appendix D, Figure D-l.
The block diagram in Figure A-10 shows the organization of
this computer and Table A-4 lists its important characteristics. The
central processor is the control unit for the entire system: it governs
all peripheral in-out equipment, performs all arithmetic, logical, and
*
Model 4001, Data General Corporation, Southboro, Massachusetts.
A-35
-------
CORE
MEMORY
4096 16-BIT
WORDS
PROVISIONS FOR |
| ADDITIONAL I
I 4096 WORD I
I CORE MEMORY \
Front
Panel
Switches
CONSOLE SWITCH
REGISTER
T
TA-8S63-31
FIGURE A-10 BLOCK DIAGRAM OF NOVA COMPUTER
A-36
-------
Table A-4
CHARACTERISTICS OF NOVA COMPUTER
Operational:
Word Size
Memory Size
Cycle Time
Number of Accumulators
Arithmetic
Addressing
Physical:
Dimensions
Power
16 bits
4096 words (expandable)
2.6 (is
4
2's complement
8-bit displacement, may be in page
zero, or relative to either of two
accumulators, or program counter
5-1/4-inch high, mounted in standard
19-inch rack
115 V, 60 Hz, 6 A (with teletype)
data handling operations, and sequences the program. It is connected
to the memory by a memory bus and to> the peripheral equipment by an in-
out bus. It is organized around four accumulators (ACO-AC3), two of
which can be used as index registers. The arithmetic instructions
operate on fixed-point binary numbers, either unsigned or equivalent
signed numbers, using two's complement conventions.
Diagnostic programs are provided by the manufacturer to test
various logic areas of the computer and I/O. The manufacturer also
supplies special programs to help the user edit, assemble, load, and
debug his programs. In addition, arithmetic subroutines are available
with the computer.
A-37
-------
c. Magnetic Tape Adapter
The magnetic tape adapter provides the interfacing between
the NOVA computer and the Kennedy digital magnetic tape recorder. An
internal switch provides selection of either 556 or 800 BPI recording
densities with a transport speed of 15 ips.
The "Write Data Lines" are active on the tape bus as "low-
level" (true state) signals. Each of these "Write Data Lines" can be
delayed, as needed. The read data lines from the tape transport carry
high-level (true state) signals amplified by switching transistor
circuitry.
Inter-Record Gap is produced by an internal clock. There is
a 20-ms delay between the command "move" and Tape Unit Ready, which is
triggered at the end of the RUN signal to allow the tape drive to
stabilize.
d. Magnetic Tape Recorder
The tape recorder is a seven-track synchronous digital mag-
netic tape recorder with dual gap read-after-write operation. A tape
speed of 15 ips allows a read or write data transfer rate of 8.34 kHz
at 556 BPI, or 12 kHz at 800 BPI. The lower density recording was used
on this project. Selection of the recording densities can be accomplished
either locally or remotely. The data were recorded on magnetic tape when
the data storage block in the computer was filled, or when a data-
collecting cycle was completed. Magnetic tapes were created to be com-
patible with the Control Data 6400 computer.
*
Model 4035, Data General Corporation.
Model 2812, Kennedy Co., Altadena, California.
A-38
-------
This unit has the electronics necessary to control tape motion
and the reading and writing. Available tape motion commands include the
usual forward/reverse, run/stop, rewind, and so on. The formatting,
parity generation, gap control, and the like, are provided from the
Magnetic Tape Adapter.
Tape is controlled by a single capstan drive. When commanded
to run, the capstan drive velocity servo responds to a linear ramp in-
put, starting the tape with a constant acceleration. Since the capstan
servo conforms almost identically to the ramp, start/stop times and dis-
tances are highly consistent and accurate. They are inversely propor-
tional to rated tape speed (15 ips). In Rewind, start is a ramp of
approximately 1/2-second duration to the rewind speed of 150 ips. Stop
is an equivalent ramp to zero speed.
Write electronics produce NRZ1 (nonreturn to zero) recording.
Data presented to the machine are recorded upon application of the "Write
Data" strobe signal. NRZ1 recording continuously saturates the tape in
either the positive or negative direction. The direction of current
flow is altered to record a logic 1 and allowed to flow unaltered for a
zero. This is the commonly used method of digital recording.
e. Teletype
*
A model 33-YZ teletypewriter with a tape punch and tape reader
provided a means of communicating with the NOVA computer. Entries into
the computer are made either on the keyboard or by prepunched paper tape.
Keyboard entries were usually used to prepare and debug programs or to
start sampling. Prepunched paper tape was used to load computer programs.
*
Teletype Corporation, Skokie, Illinois.
A-39
-------
The five-minute summaries were always listed (see Figure D-l,
Appendix D), and often punched on paper tape. The paper tape output
was used as a backup for the magnetic tape records.
A-40
-------
Appendix B
VAN INSTRUMENTATION SYSTEM
B-l
-------
Appendix B
VAN INSTRUMENTATION SYSTEM
1. Brief System Description
Two vans were used during the San Jose experiments. They were
instrumented (Figures B-l and B-2) for measuring carbon monoxide (CO)
concentrations, temperature, and wind from a parked location, or CO
concentrations and temperature while driving around selected city routes.
All data plus voice comments were recorded on magnetic tape, and CO and
temperature with superimposed event markers were recorded on a two-pen
chart recorder. All equipment could be operated for an 8-hour period
using wet batteries in the vans. The van engine was not run during
stationary operation, so no CO was produced to bias the measurements.
The effect of van-produced CO was minimized during driving by locating
the sampling inlet at a height of about 3.5 m at the front of the van.
Other equipment in the van included a battery charger, 115-V ac extension
cord, a short ladder, and miscellaneous tools needed to be completely
self-suff ic ient.
The CO analyzer air inlet and the wind sensors were mounted on
vertical extendable antenna masts (Figure 19 in the text) modified for
the purpose. It was possible to sample air from any height between 3
and 10 m. The wind sensor (which was stowed in the van during driving)
was adjustable in height from about 3 to 5 m; however, it was usually
operated at its maximum height when the van was stationary.
B-3
-------
1. Magnetic Tape Recorder
2. Temperature/CO Signal Conditioner
3. CO Analyzer
4. Wind Indicator Panel
5. Power Supply Translator for Wind System
6. Dual Channel Chart Recorder
7. Battery Charger
8. 24 V dc to 115 V ac Inverter
9. Storage Batteries
FIGURE B-1 ELECTRONIC CONSOLE IN VAN
B-4
-------
Rain Proof
Air Ir
Filter
UI
Adjustable
Mast
nlet Selectable Ranges:
r 5, 10, 25 and 50 ppm
AIR DRYER CO k
AND PUMP ~~^ ANALYZER *
Aspjratfd
diation Shield ^-^BLOWER pfw
"^ I y-y^ o ^> 1 1 5 v ac SUF
Thermistor .
Temperature . f .
Sensor [\]
TEMPERATURE
BRIDGE ' *"
X
<
f 1
TWO-CHANNEL
INK RECORDER
KER 4
PLY 1
EVENT MARKER
L *
4
V\ Selectable 1. -18 to 2 °C
A U Propeller Vane Ranges. 2 0 to 20°C
^r* Sensor 3 1g to .^^
H L _J
TEMPERATURE/CO SIGNAL CONDITIONER CHASSIS
WIND SPEED.
1 WIND UIHbCI ION WIND DIRECTION
*" TRANSLATOR- *" INDICATORS •
POWER SUPPLY
Wind Speed
1
Wind Direction
MAGNETIC
TAPE
RECORDER
41 \
Voice
Channel
MICROPHONE
115 Vac
Commercial •
Power
TIMER
BATTERY
CHARGER
r
i
Of 1o
T
Load
^
Charge
L J
INVERTER
8-HOUR 24 V dc
BATTERY
BANK
-115 Vac 60 Hz Bus
HIGH CURRENT FUSE
BOX AND SWITCH
TB-8563-38
FIGURE B-2 BLOCK DIAGRAM OF VAN INSTRUMENTATION SYSTEM
-------
2. Carbon Monoxide (CO) System
The Bacharach CO analyzers used in the two vans and the helicopter
are considerably smaller in size than the Beckman analyzers used with
the fixed-station instrumentation system (Appendix A). The Bacharach
analyzer combines a chemical and an optical technique for the automatic
measurement of ambient carbon monoxide. The heterogeneous reaction be-
tween carbon monoxide and mercuric oxide is used to generate mercury
vapor in the sample gas stream at a concentration proportional to the CO
concentration. Optical absorption at the strong mercury absorption band,
2537 A, in the near-ultraviolet is used to measure the resultant mercury
vapor with high sensitivity. This combination of techniques provides a
method for determining very low CO concentrations with a portable instru-
ment, as has been demonstrated in the work of Robbins, Borg, and Robinson
(1968).
The CO measurement system is shown schematically in Figure B-3, and
its performance specifications are listed in Table B-l. Air is sampled
through a filter installed in a rain shield outside the van on an ad-
justable mast. The air is passed through 1/4-inch-ID teflon tubing and
into one of two silica-gel cartridges. Operation of the instrument re-
quires that the air be dried, and silica-gel was used rather than the
standard heatless air dryers in order to conserve power.
After drying, interfering hydrocarbons are removed by absorption
in an activated charcoal filter. A diaphragm pump is used to avoid
contamination from lubricants. Next, the sampled air is passed through
*
Bacharach Instrument Company, 2300 Leghorn Avenue, Mountain View,
California 94040.
B-6
-------
I RAIN PROOF AIR INLET
FILTER ON ADJUSTABLE MAST
\
ADJUST
BLEEDER
VALVE
Tf M
L
400°F ± 0.1°F
PROPORTIONAL
CONTROLLER
>|
r<^
HEATER L
' /////////
/////////^
HEAT EXCHAN
AND REACT!
THERMISTOR
HgO PELLET
I
"MERCURY-FREE
REFERENCE"
SOLENOID
.'CONTROL VALVE
r
CO ANALYZER
MERCURY VAPOR
(IODINE CHARCOAL)
FILTER
I
WME
2 SCFH
FER
MERCURY
VAPOR
[IODINE CHARCOAL)
FILTER
150 F
PROPORTIONAL
CONTROLLER
L
. TO AMPLIFIER
WITH 5, 10, 25
AND 50 PPM
• RANGE SELECT
FIGURE B-3
FUNCTIONAL DIAGRAM,
VAN AND HELICOPTER
CARBON MONOXIDE
MEASURING SYSTEM
-------
Table B-l
MANUFACTURER'S STATED PERFORMANCE SPECIFICATIONS
FOR THE MERCURIC OXIDE REDUCTION CO ANALYZER
Operating Specifications:
Range
Flow rate
Output
Linearity:
0-20 ppm
>20 ppm
Zero drift (average)
Span drift (positive)
Response
lag
rise time
fall time
Minimum detectable CO
concentration
Reproducibility
Warm-up time
Environmental Specifications:
Ambient temperature range
Shock and vibration
Physical Specifications:
Weight
Power
0 to 50 ppm
3 -1
2 ft h
0 to 0.5 V
good
nonlinear (curve provided)
\ can be greater with
[large ambient tem-
3 percent in 24 h )perature fluctuations
1.5 ppm in 12 h
10 s
15 s
18 s
0.1 ppm
±2 percent
30 minutes
0 to 125° F
can operate in normal mobile uses
20 Ib
60 to 80 watts, 115 V ac ±10 percent
60 Hz
B-9
-------
1/4-inch-ID teflon tubing to the CO analyzer chassis. Flow is adjusted
3 -1
to 2 ft h using a bleeder valve and an end-of-line flowmeter. The
purified, dried, regulated air next flows to a heat exchanger/reactor
module, where it is preheated to the reactor temperature. The reactor
is maintained at 400 ±0.1° F by a thermistor-bridge, proportionally
controlled heater. At this operating temperature, the mercuric oxide
produces a mercury vapor background by thermal decomposition. In the
reactor the air contacts a mercuric oxide pellet and the CO produces
mercury vapor by the reaction:
400° F
CO + HgO *• CO + Hg
2 v
The output of interest from this reactor is the sum of the thermally
produced and the CO-generated mercury vapor.
When CO concentrations are monitored, the hot, mercury-laden gas is
routed through a two-way solenoid valve to an "active" absorption cell
in the optical detection compartment. When activated, this solenoid
valve passes the mercury-laden gas through an iodine charcoal filter to
remove mercury vapor. This provides a mercury-free "reference" to the
active absorption cell. This signal can be used to correct for drift
in the optical detection compartment and its associated circuitry. The
solenoid can be switched either manually or automatically at regular
intervals to provide this reference signal.
The mercury-free gas produces an arbitrary reference base line on
the recorder, but the actual zero CO level is obtained by calibrating
with a reference zero gas. Helium was used as the zero gas since it
has been shown to be essentially CO free. The "true" zero signal in-
cludes the contribution from the thermally produced mercury vapor.
B-10
-------
In the automatic switching mode, the mercury-free air reference
signal was produced for 45 seconds, once every 5 minutes. This signal
was summed with an offset voltage whose amplitude was adjusted by means
of a trim pot. The polarity of this offset voltage provided up-scale
recording of mercury-free air reference signals, allowing use of the
full chart width for recording the atmospheric CO concentration.
When the hot, mercury-laden gas passes through the absorption cell
in the optical detection compartment, ultraviolet light from a mercury
lamp is passed through the gas and detected by a photocell that serves
as one leg of a bridge. A second photocell with a separate optical cell
provides a compensating reference in a second leg of the measuring bridge
circuit. Close control of the UV lamp intensity was obtained by a UV
source oscillator/controller tied into the "reference" leg of the photo-
cell bridge circuit. A trim pot was provided for adjustment of the lamp
intensity. Temperature of the compartment was maintained at 150° F by
use of a second proportional controller and associated circuitry.
The active and reference photocell bridge circuitry included a zero
control to provide fine zero adjustment for zero gas calibration. The
bridge output circuit has two "gain" pots to adjust the calibration of
the instrument, using a span gas of known CO concentration. One pot is
used to adjust the meter reading on the front panel to the correct value.
The second control is used to set the up-scale (span) reading of the
chart and magnetic tape recorders. Electronic filtering reduces noise
to the recorders.
An iodine filter at the exhaust of the instrument removes mercury
vapor before venting.
B-ll
-------
The electronic CO concentration signal from the CO analyzer chassis
goes to a separate differential amplifier on the air temperature/CO
analyzer signal conditioner phassis (Figure B-3). This amplifier has
selectable ranges for full-scale recorder readings of 5, 10, 25, and 50
ppm of CO (see Figure B-2). The switch can also short the input of the
recorders electrically for recorder zeroing; a low-impedance short be-
tween the differential amplifier inputs is also available for adjusting
the amplifier zero.
3. Temperature Instrumentation
*
The van temperature sensor was a thermistor installed in an
aspirated radiation shield mounted on top of the van as shown in Figure
19. The thermistor served as one leg of a bridge network. The bridge
output was amplified and input to a chart and a magnetic tape recorder.
The thermistor was suspended in a 6-inch-long, 1-inch-OD polystyrene
tube. This tube was glossy white outside and flat black inside. It was
located concentric with and near the front of a larger outer tube. This
larger tube was 1-5/8 inches ID, aluminum, 5 ft long. The aluminum tube
was also black inside and;white outside. The aspirator motor was at the
opposite end of the tube from the thermistor, toward the rear of the van.
The aspirator power was controlled inside the van from the same chassis
that contains the temperature bridge and amplifier. This unit is pic-
tured in Figure B-l.
The aspirated radiation shield minimizes errors from self-heating
of the sensor. The self-heating error at a ventilation rate of 10 mi h"1
is less than 0.1° C, and was neglected. The time constant of the aspi-
rated thermistor is 8 seconds.
Veco 34D1.
B-12
-------
A Wheatstone bridge was used with the thermistor as one leg. Any
of three ranges could be selected: -18 to 2° C, 0 to 20° C, or 18 to
38° C. Values of bridge resistors were chosen for optimum linearity.
Low-temperature-coefficient trim potentiometers were incorporated for
zero and span adjustments during calibration.
Mercury batteries provided a stable bridge voltage source. Battery
drift was checked with a reference potentiometer (with locking pro-
visions), which was adjusted immediately after calibration to give full-
scale output on the middle range. The reference can be switched into
the circuit to check whether the supply voltage has changed.
The bridge output served as input to a differential amplifier on
the same chassis. Switches were provided for zeroing the recorder and
amplifier during calibration.
4. Wind Instrumentation
*
A Gill propeller vane was used on the van for wind measurements.
It has a 9-inch, four-blade helicoid propeller coupled to a miniature dc
tachometer generator. The voltage output is directly proportional to
rpm from 2.7 mi h to 100 mi h . Below 2.7 mi h , slippage increases
down to the threshold speed (approximately 0.5 mi h ). The voltage to
the magnetic tape recorder is unipolar and adjustable from 0 to approxi-
mately 5.5 V dc for full scale on the 50-mi h range.
A low-density foamed polystyrene vane, which is coupled to a pre-
cision linear conductive plastic type potentiometer, was used for wind
direction measurements. A regulated power supply provides a constant
*
Model 35002/35402/35602, R. M. Young Company.
B-13
-------
voltage to this potentiometer to produce voltage output directly propor-
tional to the azimuth angle of the vane, over 342 degrees of azimuth.
The output voltage is unipolar and is adjustable from 0 to approximately
±8 volts.
The wind speed system was calibrated in the same way as the wind
component (UVW) sensors at the fixed station (see Appendix A). Wind
direction calibration uses circuitry in the translator unit. A calibra-
tion switch provides a signal that corresponds to either the zero-degree
position or the full-scale vane position. The wind vane was oriented
to give the proper wind direction reading when it was aligned with the
street. The mast supporting the wind sensor could be adjusted to the
vertical, as determined from a level mounted on it. Ihe wind sensor was
taken from the mast whenever the van was moved.
5. Recorders
Chart and magnetic-tape analog-signal recorders were used with the
van instrumentation. We used a dual channel Beckman Model 2550 servo-
balanced potentiometer type strip chart recorder.
An event marker can be superimposed on the temperature trace by
pushing a button within reach of the van driver. This back-loaded the
temperature amplifier with a resistor that simulated the recorder input
impedance, presenting a low-impedance (100 ohm) short to the recorder.
This causes the recorder to return to its zero voltage position, providing
an event mark. The event marks were numbered on the chart paper and in
the operator's log.
The magnetic tape recorder was a standard Hewlett-Packard model
3960-A four-channel FM-type recorder. Voice comments could be recorded
on Channel 4 while the van was being driven. When the van was parked,
B-14
-------
this channel was used primarily for recording the temperature signal
with occasional voice comments superimposed.
6. Primary Power System
Primary power for the van was provided by four, 12-volt, 210-
*
ampere-hour batteries connected in series-parallel to provide 420
ampere-hours of primary power at 24 volts dc. These batteries were
diesel truck type that allowed "deep cycling/' i.e., they can be com-
pletely discharged and recharged each day without damage, unlike most
vehicle batteries. The same batteries were used during the entire ex-
perimental period.
A commercial 40-A battery charger, a separate timer, and a heavy-
duty extension cable were installed in the van to be used for recharging
the batteries during the night when the van was not in experimental use.
The battery condition was periodically monitored with a hydrometer. A
60-A house-type double-pole, double-throw pull box was installed to re-
move the electrical load from the batteries and to connect the charger.
A Topaz model 1000GW 24-V dc to 115-V ac inverter was mounted near
the batteries and supplied power to the various instruments. This high-
quality inverter produces a low-harmonic-content sine-wave output. It
is well regulated to provide stabilized 60 ± 1.0 Hz power output.
*
Prestolite 8908X.
t
Exide EM-40.
Topaz Electronics, San Diego, California.
B-15
-------
A wooden frame constructed of 2 X 10-inch lumber coated with acid-
proof paint secured the batteries over the rear axle of the van. Solidly
secured batteries and equipment were important for safety in case of an
accident or sudden stop.
7. Equipment Installation
The electronic units were installed in a standard 19-inch panel
mounting cabinet which was 24 inches deep and 4 ft tall. It was shock
mounted to a 3/4-inch-thick circular plywood base that was center bolted
to the van floor. This "lazy susan" allowed the cabinet to be rotated
for easy access. A safety cable was attached to the top of the cabinet
to prevent forward travel in an accident. The cabinet could be oriented
to provide for operation from either the driver's or passenger's seat.
There is enough space to allow passage from the seats to the rear of the
van.
The pump and air dryer assembly for the CO analyzer was located on
the floor of the cabinet. This allowed easy replacement of the silica-
gel cartridges. The cartridges were removed in the evening and dried by
baking in an oven until the silica-gel crystals turned bright blue. This
usually took about 2 hours.
The two-channel ink recorder was installed in the van on the engine
cover (Figure B-l). A felt-lined wooden box housed the recorder and
allowed the driver to write identification numbers on the chart paper
periodically for later correlation with experimental notes.
B-16
-------
Appendix C
HELICOPTER INSTRUMENTATION SYSTEM
C-l
-------
Appendix C
HELICOPTER INSTRUMENTATION SYSTEM
1. Brief System Description
The helicopter instrumentation recorded CO concentration, air
temperature, and altitude. The air inlet and the temperature sensor
were mounted on the helicopter skids ahead of the cockpit so that the
effects of rotor downwash could be avoided by maintaining the heli-
copter's forward speed. The pressure transducer for the altimeter was
inside the unpressurized cockpit.
The recorder, CO analyzer chassis, and the temperature/pressure
bridge chassis were separated by padding and stacked on the seat between
the pilot and the observer as shown in Figure C-l. This equipment was
secured with a safety strap. The pump and dryer package for the CO
analyzer was placed on the floor beneath the experimenter's legs. A
12-V dc to 115-V, 60-Hz ac inverter was installed with rain protection
on the outside cargo rack where the pilot could turn it off or on. All
the equipment was designed to operate within the limitations of avail-
able power from the helicopter's electrical system.
The CO analyzer used on the helicopter was the Bacharach instru-
ment described in Appendix B. It differed from the version used in the
two vans only in that it was painted white to minimize solar heating in
the "bubble" cockpit.
C-3
-------
1. CO Analyzer
2. Te,
mperature/Pressure Bridge Chassis
3. 4-Channel Chart Recorder
FIGURE C-1 HELICOPTER INSTRUMENTATION
C-4
-------
2. Temperature System
The helicopter temperature instrumentation used a thermistor in a
>
radiation shield on the helicopter skid, and bridge network. The bridge
output was recorded by a chart recorder with a differential amplifier
input.
The radiation shield has two 6-inch-long, concentric cylinders.
The smaller was 1-inch-ID polystyrene, glossy white outside and flat
black inside. Two VECO 33D12 thermistors were mounted at the center
of this cylinder. The 1-11/16-inch-OD brass outer shield has a 1/32-
inch wall that is chrome plated outside and flat black inside. The
inner cylinder is centered and supported by machine screws in the outer
cylinder. Ventilation is provided by the forward motion of the heli-
copter. With this ventilation the time constant is less than 2 seconds,
i
and self-heating is negligible.
The two thermistors, connected in parallel to reduce self-heating
errors, formed one leg of a Wheatstone bridge. A gang switch selects
any of six ranges: -18 to -8, -9 to 1, 0 to 10, 9 to 19, 18 to 28, or
27 to 37° C. The many ranges were used to increase resolution, which
was restricted by the recorder chart width. The bridge resistors were
chosen for optimum recorder scale linearity. The maximum deviation from
linearity on any range is 0.25° C. D-size mercury batteries provide a
stable voltage for the bridge. Battery drift could be checked with a
reference potentiometer as in the van's temperature measuring system
(see Appendix B). The bridge circuitry and its batteries were mounted
in a 3.5 X 17 X 14-inch chassis that also housed the pressure sensing
system described in the next section.
The sensor was calibrated in a temperature control chamber at two
temperatures for each range, one near the low end of the range and one
near the upper end. Trim pots were available for each range for adjusting
C-5
-------
the circuitry to give correct readings. Adjustments were made first for
the low-end temperature using a pot in the leg of the bridge opposite
the thermistors. Adjustments at high-end temperatures were made using
a potentiometer that controlled the voltage applied to the bridge. The
adjustments were repeated to minimize errors that might arise because
the two adjustments are not totally independent.
3. Altitude-Recording System
During the San Jose experiment, altitude was determined from the
helicopter's altimeter and periodically marked on the chart records by
the operator. A system was designed and built that had been planned
to provide a primary record of altitude but was actually used only as
a backup to the altimeter. Its resolution was less than originally
planned because the desired transducer could not be delivered in time
for the experiment.
One major problem in the design of an inexpensive pressure-sensing,
altitude-recording system is that of altitude resolution. The atmospheric
pressure might only go from 14.7 psi at sea level to approximately 12.3
psi at 5000 feet. A pressure transducer with a 3-psi full range con-
taining a bellows with an internal reference pressure of 12 psia appears
to be a logical choice. However, this 12-psia reference will vary as
temperature and cannot be used without costly corrections. This tem-
perature error becomes negligible as the reference pressure is reduced
towards zero, but this also reduces resolution. A bellows with a near
vacuum internally is relatively insensitive to temperature fluctuations,
but has a smaller movement for a given pressure change.
The pressure transducer used in the helicopter system has a bellows-
actuated potentiometer that serves as two legs of a bridge. Output
voltage is measured at the movable contact. Any one of seven ranges
C-6
-------
can be selected: 0 to 1000 ft, 900 to 1900 ft, 1800 to 2800 ft, 2700 to
3700 ft, 3600 to 5600 ft, and 5400 to 7400 ft. A larger part of the
chart width can be used by having a large number of available ranges.
Nevertheless the nature of the transducer limited the resolution to
about 88 feet.
Variations in atmospheric pressure from day to day would produce
errors in recorded altitude if adjustments were not made. A dual
ganged potentiometer in the bridge circuit was used to adjust the reading
to the correct airport altitude before takeoff. The ganged potentiometers
were wired as opposing rheostats in adjacent legs of the bridge and con-
nected to the ends of the pressure transducer element. Adjustment
caused an effect in opposition to that introduced by the atmospheric
pressure variations at the time of adjustment.
The pressure transducer was calibrated in the laboratory in a
partially evacuated bell jar. Trim potentiometers in the circuitry were
adjusted to give the same readings as the pressure transducer at selected
"altitudes." Once set, these potentiometers could be switched into the
circuits later to provide field simulation of seven heights ranging from
50 to 7300 ft.
The output of the bridge was recorded on a chart recorder with a
high-input-impedance differential amplifier. All the circuitry and
the transducer were installed on the same chassis as the temperature
measurement circuitry.
4. Chart Recorder
Power, space, weight, and cost were important factors in the recorder
*
selection. The MFE Model 24C-AHA recorder was used. It has three
*
Mechanics for Electronics, Inc., Wilmington, Massachusetts.
C-7
-------
channels to record CO, temperature, and altitude and a fourth for possible
future measurement of humidity. Amplification is provided internally by
a high-gain, high-input-impedance differential amplifier, as required
for the relatively low bridge output signals. It uses thermal writing,
which is preferable to ink for use in a helicopter.
The recorder has a narrow chart. Each record is confined to a strip
only 5 cm wide with 1-mm divisions. The four records are written side
by side on the chart. There are four chart speeds, ranging from 1 to
50 mm s , also with 1-mm divisions. One event marker and one timer
marker were included. Calibration is provided by an internal 20-mV
signal.
C-8
-------
Appendix D
DATA PROCESSING
D-l
-------
Appendix D
DATA PROCESSING
1. Streetside Data
The San Jose Streetside data were obtained with a data-acquisition
*
system under the control of a Data General NOVA mini-computer and re-
corded on a Kennedy magnetic tape recorder (Model 2812). These data
consisted of a station identifier, CO level number, CO concentration,
instrument status code, wind speed and direction, temperature, date,
time and other measurements. These were stored by groups in the com-
puter memory whenever one of the various sensor units was activated upon
command by the computer. The data were recorded on magnetic tape when
the data storage block in the computer was filled, or when a data-
collecting cycle was completed.
The data collection cycles were generally 5 minutes long. During
a cycle, CO concentration was recorded for each of the five levels at
each of the stations. Also recorded during a cycle were five values
of wind speed and direction for each of the locations that used cup and
vane sensors. Similarly, five sets of readings of temperature gradient
information were recorded for each of the two stations at which these
parameters were measured, as were approximately 125 values of each of
the wind component values. Also, other recorded material included
information on equipment operation, e.g., bridge voltages, range factors,
*
Data General Corp., Southboro, Massachusetts.
Kennedy Co., Altadena, California.
D-3
-------
and so on. At the end of each cycle, the data were averaged for the
period. These averages were summarized on the tape and printed out on
the teletype for real-time monitoring of system operation. Figure D-l
is a sample of this real-time summary printout.
The first word of each magnetic tape record was assigned either a
0 or 1 (0 identified a data record and I, a summary record). Records
were of variable length and each was made up of variable-length sub-
units, which we will refer to as "groups." Each group was preceded by
two words containing the addresses of the first and last words. These
addresses defined the number of words in that group. The last two
words in a group were the date and time of the sample. The end of a
record was signified by 7777 following the last data group. Either
an "end of file" or a blank record signified the end of tape. The
magnetic tape was created to be compatible with the Control Data 6400
computer. Ten NOVA computer-stored data words are contained in three
CDC 6400 computer words (60 binary bits per word). Hence, a record of
data could easily be buffered out, masked, and operated on. A density
of 556 bits per inch, odd parity mode, and seven-track tapes were used.
A new tape was written by the large computer. This tape contained
all the information on the original field data tapes. In creating this
new tape, the data were converted to engineering units (from the originally
recorded voltages). Some of the editing and data correction was also
accomplished at this time. For instance, data from certain instruments
were eliminated during periods when those instruments were known to be
operating improperly. Where connections were incorrect, they were re-
assigned properly at this step. For instance, the inlets from two CO
sampling levels had been reversed; the readings were attributed to the
proper level in the rewritten tape. Similarly, there was some confusion
D-4
-------
12/11/70 1838
G
01
U*V*W 1/100'S
A +00108 -00336
+00108 +00072
C +00072 +00072
+00072 +00072
Parameters
-00072 +0y252 +00072 +00000
+00072 +00144 +30072 +00072
+00108 +00072 -00144 +00000
+00072 +00108 +00144 +00072
Average Wind Components at Lower and Upper Levels
Standard Deviations of Wind Components
Average Wind Components
Standard Deviations
CO 1/100'S PPM
B +00780 +00741 +00507 +00702 +00546\
D +00585 +00507 +00390 +00390 +80429 I
E +00702 +00819 +00897 +01014 +010141
F +01677 +01092 +01131 +31014 +00858 \ gt Leve|s
G +00936 +00819 +00741 +00741 +00663 I
H +01521 +01365 +01053 +01638 +01092 1
I +00663 +00663 +00702 +00624 +00585^
TABS DTI DT2 DT3 DT4
8 -00255 -00005 -00014 +00000 -00017
D -00250 -00002 -00007 -00004 -00013
Observed CO Concentrations
Through 5
Temperature 1/10°C and AT's 1/100°C
(Temperature Sensor not Connected)
WIND VEL
E +00083
+00125
+00200
+00208
+00196
WIND DIR
+00118
+00039
+00000
+00101
+00104
Average Wind Speeds (1/100 mi h~1)
and Wind Directions (degrees,
relative to First Street)
l_
•Letters used as station identifiers in this printout
A-1. B-2...I=9
TA-8563-20
FIGURE D-1 EXAMPLE OF REAL-TIME SUMMARY PRINTOUT
-------
among the various wind components. In translating the tape, the obser-
vations were converted to a single right-handed coordinate system for
all sensors. An example of the information contained on the tape gener-
ated by the CDC 6400 processing is shown in Figure D-2. At this point
in the data processing, the data had not been corrected to account for
drift of calibration in the CO analyzers. In general, the instruments
were calibrated sometime during the first 2 or 3 hours of operation
each day. Usually they read within ± 0.5 ppm of the analyzed zero
or span gas values when these gasses were introduced for calibration.
If not, the error was recorded in a notebook. For those cases where
there were calibration changes greater than 0.5 ppm, the following
correction was applied to the recorded CO values:
/ \ 19
c = 'c - c I • — .
c , R o/ C^
where C is the corrected value. C is the recorded value, and C and
c R o
C are the readings when the zero gas and the 19 ppm span gas were
.L y
introduced. The zero gas was introduced first, and the instrument was
adjusted to read zero; then the span gas was introduced and that adjust-
ment made. The correction was usually applied to those data collected
between the time the instrument was put on line and the time it was
calibrated. In a few special cases where the instrument was recalibrated
during the day, the correction was applied for a period of an hour or
two before the recalibration.
To this point the complete data set was still on tape and there
had been no attempt to digest or simplify the records. We learned that
it was difficult to obtain statistical summaries of the data from the
unconsolidated tapes. For this reason we decided to produce a new tape
where each record corresponds to one of the cyclical 5-minute data
D-6
-------
l.o
3.0
1.0
3.0
1.0
3.0
1.0
3.0
1.0
3.0
1.0
3.0
10.0
2.0
2.0
4.0
4.0
S.n
6.0
7.0
8.0
9.0
1.0
3.0
1.0
3.0
1.0
3.0
l.o
3.0
1.0
3.0
1.0
3.0
1.0
3.0
1.0
3.0
1.0
3.0
1.0
3."
1.0
3.0
1.0
3.0
1.0
3.0
1.0
3.0
If No. in 1st
Column is:
Station No.
1 or 3
Station No.
2 or 4,
1st Line
Station No.
2 or 4,
2nd Line
Station No.
5,6,7,8,
or 9
10, Sum-
mary
Record
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.)
701123.0
701123.0
701123.0
701123.3
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
701123.0
105449.2
105450.4
105451.4
1054S2.4
105453,4
105454. »
105455.4
105456.4
105457.4
105458,4
105459.3
105500.3
1050.0
105621.5
1(15621.5
105623.1
1U5623.1
1'15623.H
105624.5
105625,2
105625.9
105626.5
105627.5
105620.5
105629.5
105630.5
105631.5
105632.5
105633.5
105634.5
105635.4
105636.4
105637.4
105638.4
105639.4
1U5640.4
105641.4
105642.4
105643.4
105644.4
105645.3
1H5646.3
105647.3
10564B.3
105649.3
105650.6
105651.5
105652.5
105653.5
105654.5
30.0
IS.o
30.0
40.0
30.0
35.0
30.0
40.0
30,0
34.0
30.0
40.0
1054.0
5.0
5.0
5.0
5.0
5...
b.O
5.0
5.0
5.0
29.0
11.0
29.0
40.0
30.0
31.0
30.0
29.n
30.0
8.0
30.0
40.0
30.0
35.0
29.0
40.0
30.0
37.0
30.0
40.0
30.0
36.0
30.0
29.0
30.0
21.0
29.0
40.0
-0.0
-1.2
-0.0
i.a
-0,0
1.5
-.2
2.0
-0.0
1.0
-0.0
2.2
.2
2.3
99.9
1.2
99.9
3.S
T.O
4.3
3.5
4.7
-.3
-3.3
-.3
1.5
-.3
.3
-.3
0.0
-.2
-3.7
-.2
1.7
-.2
1.2
-.2
1.5
-.3
1.2
-.3
1.5
-.5
1.0
-.5
0.0
-.5
o.o
-.3
1.5
.3
-.7
.2
1.5
0.0
.2
o.o
1.7
0.0
.3
0.0
1.7
163.6
99.9
-1.0
99.9
0.0
.3
1.1
.5
.0
.7
0.0
1.7
0.0
1.0
.2
.3
.3
-.2
.3
1.7
.2
1.5
.2
0.0
.2
1.5
0.0
0.0
-.2
1.5
-.2
0.0
o.o
0.0
-.2
-2.2
-.2
1.7
-O.o
-1.9
-.2
-.4
-.2
-o.o
-.2
-0.0
-.2
-.2
-.4
-.2
.5
99.9
-1.0
99.9
67.0
188.3
197.4
320.3
39.3
999.9
-.4
-1.9
-.4
-0.0
-.4
-0.0
-0.0
1.4
-0.0
-1.9
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
,2
-0.0
-.2
-0.0
-0.0
-0.0
4.3
-.2
-1.9
-.6
-.4
-.6
1.3
-.5
.3
-.5
-2.0
-.5
.3
-.5
-3.3
-.5
.3
257.9
99.9
-128.0
99.9
-128.0
-.3
-1.1
,4
.0
99.9
-.5
-.2
-.3
-.2
-.3
-3.3
-.3
1.3
-.3
.3
-.3
0.0
-.3
-2.2
-.3
-.2
-.3
-.5
-.3
-.3
-.3
-.5
-.2
1.3
-.2
1.0
-.2
0.0
-.3
o.o
-.2
-.2
-.2
0.0
-0.0
-.2
-0.0
1.2
-0.0
-.2
8.1
99.9
1.0
99.9
69.0
.0
.3
.3
-.0
99.9
-.2
0.0
-.2
-3.5
-.2
1.5
-.2
1.0
-0.0
0.0
-0.0
-.2
-0.0
0.0
-0.0
0.0
-0.0
-1.0
-0.0
0.0
-0.0
-1.3
-0.0
1.5
-.2
0.0
-.2
-.2
-0.0
,2
-.2
4.3
-0.0
-1.7
-0.0
3.7
-0.0
-1.7
-0.0
3.1
4.9
99.9
29.0
99.9
93.0
28.0
61.0
62.0
62.0
-97.0
-0.0
-.4
-.2
-1.7
-.2
-2.5
-.2
-.4
-.4
-.4
-.6
3.1
-.4
-2.3
-.2
3.5
-0.0
-1.9
-0.0
4.1
-.2
-2.1
-.4
-.6
-.2
-.4
-.2
3.9
LEGEND
Date
(11/23/70)
in this
case)
Date
Date
Date
Date
Time
(nearest
1/10 sec)
Time
Time
Time
Time of
1st CO
obs.
(hr. min.)
Instru-
ment
Status
Indicator
Level of
CO obs.
Level of
CO obs.
Level of
CO obs.
Time of
Last CO
obs.
(hr. min.)
U1
V1
W1
U2
V2
•« Wind Components, 3m and Roof Levels (msec"
CO Con-
centration
(pprn)
AT,
AT,
AT,
AT,
Temp, at Level 2-Level 1,
Level 3-Level 2, etc. (°C)
W2
Temp.,
at 3m
__ Information Concerning Instrument
Status and Operational Parameters
CO Con-
centration
(ppm)
Wind
Speed
(msec"1 )
Average Roof Level
Wind
Direction
(deg.)
u
V
Wind Components
(Calculated from
Recorded Speed
and Direction)
Average Roof Level
Winds, Sta. 1 Winds, Sta. 3
Calculated from Components
Speed ! Direction ( Speed , Direction
Average
CO for
all Sta.
at 3m
Instru-
ment
Status
Indicator
Average
CO for
all Sta.
Roof Lev.
99.9 or 999.9 indicate missing data.
TA-8563-22
FIGURE D-2 EXAMPLE OF PARTIALLY CORRECTED DATA CONTAINED IN ONE RECORD
OF THE TAPE GENERATED BY INITIAL CDC 6400 PROCESSING
D-7
-------
collection periods. On this consolidated tape, each record contains
all the wind, temperature, and CO concentration information from the
stations where CO was measured and summaries of the data from the wind-
component stations. A group of derived data was introduced for purposes
of stratifying the data for statistical summaries.
On this basic summary tape, each record has 420 words. An annotated
example of the recorded information is shown in Figure D-3. The first
ten words include the date and times covered by the record, the average
rooftop winds at Stations 1 and 3, and the CO concentration at 3 m and
rooftop, averaged over all the stations.
The next 30 words on the basic summary tape contain date and time
information, and the sums of all the Station 1 wind component observa-
tions during the 5-minute periods and the sums, their squares, and the
numbers of observations. The following 30 words have the same informa-
tion for Station 3. These sums, and sums of squares, are easily com-
bined with the same items from other records to obtain means and standard
deviations of the wind components.
The rest of the record contains the data from Stations 2, 4, 5,
..., 9, in ten-word groups. The data in these groups are the same as on
the earlier unconsolidated tape, except that the recorded information
relating to instrument operation (bridge and amplifier voltages, etc.)
is not retained.
The basic summary tape has been used to obtain averages for several
different data stratifications. Many of these have been discussed in
the body of this report. A simplified flow chart of the data averaging
routine is shown in Figure D-4. A sample of the type of output obtained
from this processing is shown in Figure D-5. In this example, the means
and standard deviations are given for all those cases occurring between
D-8
-------
10. on
1.00
1.00
1.00
3.00
3. OP
3.00
3.00
4.00
S.OO
6.00
7.00
8.00
9.00
2.00
4.00
S.OO
6.00
7.00
8.00
9.00
2.00
4.00
5.00
6.00
7.00
8.00
9.00
2.00
4.00
5.00
6.00
7.00
8.00
9.00
2.00
*.oo
5.00
6.00
7.00
8.00
9.00
If NO. in 1s
Column is:
10
(Sum-
mary)
Station No
1 or 3
Station No
2 or 4
Station No
5,6,7,8,9
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
70113C.OO
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
701130.00
1210.00
1210.00
1210.00
1210.00
1210.00
1210.00
1210.00
121022.50
121024.10
121024.80
121025.41)
121026.10
121026.8Q
121027.50
121103.00
121104.60
121105.30
121106.00
121106.70
12U07.30
121108.00
121201.60
121203.20
121203.90
121204. 6P
121205.30
121205.90
121206.60
121302.20
121303.80
121304.50
121305.20
121305.80
121306.50
121307.20
121402. 80
121404.4Q
121*05.10
121405.70
121406.40
121407.10
121407.80
1214.00
1214.00
1214.00
1214.00
1214.00
1214.00
1214.00
5.00
5.00
5.00
5.00
S.OO
5.00
5.00
2.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
2.00
2.00
2.00
2.00
2.00
2.00
3.00
3.00
3.00
3.00
3.00
3.00
3.00
4.00
4.00
4.00
4.00
4.00
4.00
4.00
.97
45.35
61.39
124.00
164.84
793. 4B
124.00
1.56
.39
8.97
3.90
.78
5.46
99.90
2.73
.78
8.19
4.29
.78
6.24
99.90
2.34
-.39
5.46
3.12
99.90
13.26
99.90
.78
-.39
4.68
4.29
.78
7.41
99.90
1.56
2.34
5.46
7.41
.78
2.34
99.90
337.92
-111.71
245.32
124.00
-86.3?
377.19
124.00
.91
.03
1.56
1.86
.74
.04
2.83
.89
.02
.89
3.7?
.82
.06
2.38
.91
.06
2.6o
1.79
99. 9n
.06
3.2(1
.95
.07
1.12
2.90
1.19
.06
4.91
.9<>
.04
.97
1.26
2.31
.04
3.57
1.50
146.37
398.84
124.00
54.03
409.85
124.00
-.23
.03
140.50
186.17
177.03
87.11
999.90
-.28
0.00
151.74
160.88
171.41
95.54
999.90
-.30
0.00
115.21
28.81
999.90
250.09
999,90
-.31
-.02
148.93
186.17
177.03
140.50
999.90
-.25
-.02
191.08
166.50
151.74
115.21
999.90
297.65
86.50
221.78
124.0(1
72.87
329.82
124.00
-.05
.05
-1.21
-1.85
-.74
.00
99.90
-.02
0.00
-.79
-3.51
-.81
-.01
99.90
-.05
.02
-1.11
1.56
99.90
-.02
99.90
-.02
.09
-.96
-2.88
-1.19
-.05
99.90
-.11
.09
-.95
-1.23
-2.03
-.02
99.90
3.77
342,39
1033.98
124.00
37.85
351.45
124.00
-.11
.06
-.99
.20
-.04
-.04
99.90
-.14
.05
-.42
-1.22
-.12
-.06
99.90
-.08
.01
-2.36
-.86
99.90
.06
99.90
-.14
-.05
-.58
.31
-.06
-.04
99.90
-.14
-.02
.19
-.30
-1.09
-.04
99.90
3.51
34.47
74.86
124.00
-67.90
760.37
124.00
99.90
99.90
62.00
61.00
63.00
62.00
60.00
99.90
99.90
62.00
61.00
63.00
62.00
61.00
99.90
99.90
63.00
61.00
64.00
62.00
62.00
99.90
99.90
62.00
61.00
62.00
62.00
61.00
99. 9fl
99.90
6?. 00
61.00
62.00
62.00
62.00
LEGEND
Date
(11/30/70
in this
case)
\
Date
Date
Date
Start
Time
of Period
(hr. min.)
Start
Time
of Period
Time
(hr., min.,
sec.)
Time
End Time
of Period
(hr. min.)
End Time ,
of Period
Level of
CO
Reading
Level of
CO
Reading
Average
Rooftop
Wind
Speed
Sta. 1
f Su
2u2
No. of
. obs.
CO Con-
centration
CO Con-
centration
Average
Rooftop
Wind
Direction
Sta. 1
2v
2v2
No. of
obs.
Average
Rooftop
Wind
Speed
Sta. 3
2w
Zw2
No. of
obs.
AT
Level 2-
Level 1
3-m
Wind
Speed
AT
Lev/el 3-
Level 2
3-m
Wind
Direction
Average
Rooftop
Wind
Direction
Sta. 3
2u
2u2
No. of
obs.
AT
Level 4-
Level 3
u
Com-
ponent
Average
of all
3-m CO
Concen-
trations
Zv
Zv2
No. of
obs.
Rooftop Leve
AT
Level 5-
Level 4
V
Com-
ponent
Average of
all Rooftop
CO
Concen-
trations
Zw "\
Zw2 I
No. of [
obs. J
3-m
Temp.
No. In-
dicating
Instru-
ment
Status
• 1st Line
2nd Line
3rd Line
99.9 or 999.9 indicate missing data.
TA-8563-21
FIGURE D-3 EXAMPLE OF INFORMATION CONTAINED IN ONE RECORD OF THE BASIC
DATA SUMMARY TAPE
D-9
-------
START
SET ALL SUMS TO ZERO
READ 420 WORD RECORD
IS IT THE LAST RECORD?
YES
NO
DETERMINE CATEGORIES FROM FIRST
TEN WORDS, e.g. TIME OF DAY AND
ROOF LEVEL WIND
DIRECTION CATEGORIES
ADD VALUES OF DATA (AND SQUARES OF
DATA) TO THE APPROPRIATE SUMS FOR THE
CATEGORIES. KEEP COUNT OF NUMBERS
OF DATA
USE SUMS, SUMS OF SQUARES AND
NUMBERS OF DATA TO DETERMINE
MEANS AND STANDARD DEVIATIONS
PRINT RESULTS—CAN ALSO BE RECORDED
ON MAGNETIC TAPE. SEE FIGURE D-5
FOR SAMPLE PRINTOUT
END
TA-8563-17
FIGURE D-4 SIMPLIFIED FLOW CHART OF
DATA-AVERAGING PROGRAM
D-10
-------
ROOFTOP WIND
DIRECTION AT
STA. 2 =180° ±22.5°
INCLUDES DATA
FROM 1100-1200
h.D.
ien
STA
1
3
ST«
7
4
5
6
7
e
9
ST«
7
4
MSI
.57
1.25 1
C01
2.85
3.52
6.58
6.79
4.07
4.98
5.58
DTI
.61*
.09
VIM
1.57
2
-------
1100 and 1200 where the wind at roof level of Station 1 was from
180° ± 22.5° (relative to First Street). Similar summaries were ob-
tained for data stratified by wind direction only and by wind direction,
wind speed, and traffic amounts.
The basic summary tape can also be processed to give the data
categories for each date and time interval. This information can be
combined with the data from other sources, such as the vans, to provide
additional stratified average concentrations.
2. Mobile Data
a. Van
The basic data from the instrumented vans were recorded on
strip charts. The inlet to the carbon monoxide analyzer could be ad-
justed in height. Three different heights were used during the program,
3, 6, and 9 m for stationary measurements. While the van was traveling,
the inlet was kept at about 3.6 m. When the van was stationed at a
location near the intersection of First and San Antonio Streets, the
height was generally changed at periodic intervals. The chart records
were maked each time the inlet height was changed, and a log was kept
of the van location and inlet heights. The inlet was generally kept
at each height for 10 or 15 minutes.
The first step in reducing the stationary van data was to
read the CO concentrations from the chart records at 1-minute intervals.
These CO values, along with time and date information, were transferred
to punched cards. Another deck of punched cards was prepared from the
log information. These two decks were merged to yield a magnetic tape
that contained all the data collected while the van was located in the
vicinity of the intersection of First and San Antonio Streets. Each
record on this tape contained 510 words, arranged in 85 groups of 6
D-12
-------
words. Each group contained the van number, date, time, location, inlet
height, and CO concentration.
In the preceding section, it was mentioned that lists of data
categories versus time could be obtained. For example, it was possible
to obtain, on punched cards, all those dates and time intervals during
which the roof-level wind direction was 180° ± 22.5°. Then all the
van measurements that were taken during these same time periods could
be averaged and analyzed in combination with the stratified streetside
averaged data. Thus, we were able to calculate average CO concentrations
from the van data that were comparable to those obtained from the fixed
station network, and we could thereby extend some of our analyses.
One van usually traveled a circuit around the downtown area
while the helicopter was operating. This ground-level circuit was
similar, but not identical, to the helicopter route, as can be seen in
Figures 13 and 20 of the main text. As the van passed each of the
numbered points shown in Figure 20, the chart record was marked and
the time noted in a log.
In reducing the data, an average CO value was obtained visually
from the chart record for each route segment between numbered points.
The van generally made five or six complete circuits during each heli-
copter observation period. The five or six CO concentration averages
for each route segment were averaged with a desk calculator to give a
single value for each segment. These values were then used in combina-
tion with the corresponding helicopter data for further calculations.
The considerable spatial and temporal averaging represented by the
process described above is necessary to eliminate unrepresentative CO
values arising from small-scale, short-period traffic variations.
D-13
-------
b. Helicopter Data
Aerial CO and temperature measurements were made during the
morning, noon, and evening peak downtown traffic periods with a chartered
Hughes 300 helicopter. The data were recorded on a multichannel strip
chart recorder.
Before each flight, recorder gains were checked with various
known input voltages. The output voltage of the CO analyzer was first
checked with a CO-free calibration ("zero") gas and adjusted to zero
output, if necessary. Subsequently, a span gas (19 ppm CO from the
same source used for all the analyzers) was fed into the instrument and
the gain adjusted to a nominal value of 10 mV/ppm. This procedure
was repeated to adjust for minor interactions between the zero and gain
adjustments. The calibration procedure was occasionally repeated at
the end of a flight.
Before taking data, the electronic reference signal of the
analyzer was checked. This provided a reference for possible drift of
the analyzer output. The electronic reference was frequently checked for
drift in flight. This procedure was especially necessary for morning
flights when ambient cockpit temperature changed as much as 15° C over
a short period.
The helicopter measurements were of two types, vertical pro-
files and horizontal traverses. Two vertical profiles of CO and tempera-
ture to 1000-m altitude were made in the vicinity of Spartan Stadium
(approximately 3 km SE of the San Jose central business district) at
the beginning of each flight; a profile took approximately 6 minutes to
complete. Horizontal traverses were made at various altitudes (62, 92,
152, 213, and 304 m) about a 1.03 X 1.45-km area encompassing the central
business district, depicted in Figures 13 and 20 of the main text.
The minimum traverse altitude was prescribed by the heights of nearby
D-14
-------
buildings and towers; the maximum was a function of several factors:
(1) absolute value of CO concentration, (2) ceiling height, and (3)
traffic control restrictions of the nearby San Jose Municipal Airport—
the flight pattern intercepted the airport's localizer.
The vertical profiles of CO (mV) and temperature (°C) mea-
surements were put onto cards from the strip chart at 15-m intervals
from the surface to 150 m, and thereafter at intervals of 30 m to the
top of the profile. The horizontal traverse data were reduced over
225-m increments, giving 19 values of each parameter for every height.
Additionally, each profile and traverse was assigned a representative,
mean electronic reference value as the result of frequent in-flight
checks. The data were processed on the CDC 6400 computer with both
printed and graphed output; the output format is illustrated in Figures
D-6 and D-7.
3. Traffic Volumes
The raw 5-minute volume histories for selected sensors moni-
tored by the traffic system were available from the traffic control
computer on punched card output. Figure D-8 is an example of a card
image listing for one trial (1109-1304 on 8 December 1970). Each line
represents one card according to the following format:
Card Number Columns Contents
1
32-33
35-36
38-39
43-44
46-47
49-50
Month
Date
Year
Hour
Minute
Second
Trial
date
Time of completion of
initial 5-minute data
collection interval
D-15
-------
o
DATE s 70/12/ 9/1645
ANALYSIS OF SAN JOSE CO DATA OHTAlNFD FROM A HELICOPTER.
35
NO.
1
2
3
4
5
6
7
a
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
HEIGHT
15.24
30.49
45.73
60.9fl
76.22
91.46
106.71
121,95
137.20
152.44
182.93
213.41
243.90
274.39
304.88
335.37
365.85
396.34
426.33
457.32
487. «C
518.29
54f?,78
579.27
609.76
640.24
670.73
701.22
731.71
762.20
792. 68
823.17
fl5l.66
884.15
CC
3.40
3.53
3.63
3.74
3.51
3.11
2.91
2.72
Z.H3
2.64
I.H2
1.44
J.2S
1.14
1.12
1.12
1.12
1.08
.99
.87
.74
.66
.65
.65
.55
.53
.57
.59
.51
.4?
.34
.IS
.27
.42
TEMP
12. 2u
11.90
12.00
12.00
11. *0
11. HO
11.60
11.50
11.30
11.30
11.? '.'
lO.HO
10.70
1 0 . 4 0
10.10
9.a<-
9.60
9.3')
9.20
9.0*
8.6"
H.?o
7.90
7..90
5.30
2
I AT
8.10
Z25
46.50
EO
46.90
EOR
171.50
G*
I AT = Recorder Attenuation (mV/mm)
Z25 = Reference Zero Gas Value (mm)
EO = In-flight Electronic Reference (mm)
EOR = Calibration Electronic Reference (mm)
G* = Gain of CO Analyzer (mV/19ppm)
HEIGHT = Altitude (ml
TEMP = Temperature (°C)
CO = Carbon Monoxide Concentration (ppm)
TA-8563-33
FIGURE D-6 COMPUTER OUTPUT FORMAT FOR HELICOPTER TEMPERATURE AND CARBON MONOXIDE PROFILE DATA
-------
DATE « 70/11/10/1000
ANALYSIS OF SAN JOSE CO DATA OBTAINED FROM A HELICOPTER.
N a 19 HT s 200 IAT * 5 Z25 a 10.90 EO a 19.10 POP a 18,50 6
8.54
KG.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
HEIGHT
200.00
200.00
200.00
200.00
200.00
200.00
200.00
200.00
200.00
200.00
200.00
200*00
200.00
200.00
200.00
200.00
200.00
200.00
200.00
TEMP
15.00
15.00
14.90
14.80
14.90
14.90
14.80
15.00
14.80
14.90
15.10
15.00
15.10
15.00
15.10
15.00
15.00
15.10
15.00
CO
5.44
5.50
5.68
5.56
5,56
5.27
4.80
4.74
5.33
5.68
5.50
5.39
5.33
5.09
4.45
4.16
4.63
4.57
4.22
AV6 CO
5.56
5.15
5.40
4.49
IAT = Recorder Attenuation (mV/mm)
Z25 = Reference Zero Gas Value (mm)
EO = In-flight Electronic References (mm)
EOR = Calibration Electronic Reference (mm)
G = Gain of CO Analyzer (mV/ppm)
HEIGHT = Altitude (ft)
TEMP = Temperature (°C)
CO = Carbon Monoxide Concentration (ppm)
TA-8563-34
FIGURE D-7 COMPUTER OUTPUT FORMAT FOR HELICOPTER TEMPERATURE AND CARBON MONOXIDE TRAVERSE DATA
-------
12/08/70 11 OB 52
TIME DET DET DFT DET DET OET DET DET DET DET DET DET DET DET DET DET DFT OET
113 107 106 105 092 077 075 073 071 067 065 000 000 000 000 000 000
001 021 042 024 006 016 009 Oil 016 012 024 000 007
002 013 042 030 005 012 012 007 015 015 025 000 003
003 014 028 023 007 015 009 010 014 015 035 000 006
004 019 034 019 006 015 009 012 014 015 022 000 005
005 022 041 022 007 013 003 008 013 017 019 000 003
006 022 034 0?7 003 015 006 012 012 018 025 001 001
007 010 035 025 007 014 015 009 018 016 030 000 004
008 020 034 017 006 017 008 016 020 016 025 000 002
009 015 033 023 005 019 003 018 013 015 020 000 002
010 019 041 037 009 018 Oil 012 008 013 023 001 004
Oil 016 040 018 005 018 Oil 008 021 017 036 000 003
012 021 048 028 008 022 Oil 002 025 021 030 000 004
013 016 044 025 007 023 013 014 015 021 029 001 006
014 024 048 021 008 025 016 015 021 023 023 001 006
015 021 048 031 010 022 017 009 023 029 033 001 002
016 023 040 027 010 024 016 018 019 028 031 001 009
017 019 052 028 009 019 006 020 027 025 043 001 007
018 029 050 034 005 029 010 Oil 014 021 029 000 002
019 016 038 019 010 029 Oil 024 019 024 033 000 003
020 017 047 016 Oil 018 010 015 012 017 033 000 004
021 024 034 020 007 028 008 007 020 024 023 000 006
022 020 041 026 Oil 018 008 019 015 020 021 000 005
023 018 049 021 002 Oil 009 Oil 023 015 027 000 003
TA-8563-35
FIGURE D-8 EXAMPLE OF RAW TRAFFIC DATA SUMMARY
D-18
-------
Card Number Columns
2
3
6-9
10-13
Contents
Ignore—Alpha heading material
Detector I.D. number
Detector I.D. number
74-77 Detector I.D. number
4—final 1-5 Sequence number for each 5-minute
interval (Card 4 will contain "00001"-
Card 5, "00002"; etc.)
6-9 Traffic count for detector identified
on Card 3, Columns 6-9
10-13 Traffic count for detector identified
on Card 3, Columns 10-13
74-77 Traffic count for detector identified
on Card 3, Columns 74-77.
Thus, each volume data card represents the total counts during
a particular 5-minute interval for a maximum of 18 detectors. If less
than 18 are of interest, Card 3 will contain only zeros in the final
fields.
For more than 18 detectors, one or more additional decks,
identical in format to the above, are appended, with two blank cards
separating each deck.
Because of the large number of cards generated during the total
data collection period, a preliminary processing step consisted of trans-
ferring the card data to magnetic tapes for more efficient processing.
Three different tapes were generated, according to particular processing
requirements. The first contains only data from the sensors adjacent
to the San Antonio and First Street Test Site from 2 November to 11
December 1970. The format of this tape is depicted in Figure D-9.
D-19
-------
TAPE I ("10 DETECTOR" DATA)
Detector
Table
Trial
I.D.
30 Words of
Trial Count
Data for
Detector 234
30 Words of
Trial Count
Data for
Detector 233
30 Words of
Trial Count
Data for
Detector 179
1st Record
234
233
232
231
230
229
228
181
180
179
2nd Record
1000 M + 10 D + P
Hour
Min
Sec
M = Month
D = Day
1 for Hour <10
P = 2 for 10 14
k Spare Words
The 30 words should be set up in time sequence
FOR P = 1
1st entry should be made in word number:
3600 H + 60 M + S - 23250
INT
> FOR P = 2
INT
FOR P = 3
INT
300
3600 H + 60 M + S - 38550
300
3600 H + 60 M + S - 56550
300
TA-8563-36
FIGURE D-9 MAGNETIC TAPE FORMAT FOR TRAFFIC DATA
D-20
-------
The actual traffic volume data in the second and subsequent
records are located in ten 30-word blocks, arranged chronologically for
each sensor listed in the detector table, respectively. The time interval
identification is implicit; for example, for morning trials, the first
word in the block corresponds to data collected during a 5-minute inter-
val that terminates between 0632.5 and 0637.5 PST, the second word is 5
minutes later, etc. If data collection had not yet begun or had been
terminated during any particular interval, the corresponding word con-
tains a "999."
The second tape contains data collected during the rush hour
periods from 23 November to 15 December 1970 from all 291 sensors. The
format is identical except that the detector table record contains 291
entries and the data records contain 291 blocks of 30 words.
The third tape contains only the all-day trials of 14 and 15
December. The format is the same as for Tape II, except that (1) the 30-
word blocks were extended to 144 words; (2) maximum tape record size
required the 291 sensors to be divided among four data records, the
first three applying to 84 sensors each and the final record to 39 sen-
sors; and (3) "P" in the identification number = 0.
The processing programs permitted flexibility with respect
to the type of output. A basic summary of daily raw data, totaled
for 15-minute intervals, was prepared for traffic at San Antonio and
First Streets. A similar daily record was prepared for the total of
all 291 detectors. Examples of these are shown in Tables D-l and D-2,
respectively.
Finally, mean traffic volumes and their standard deviations
were computed for all streets and for all intersections within the grid.
These means were based on data from the entire period of operation.
Means were calculated as a function of time of day using 30-minute
D-21
-------
Table D-l
TRAFFIC AT INTERSECTION OF FIRST AND SAN ANTONIO STREETS
FOR MONDAY, 2 NOVEMBER 1970
Time Interval
0645-0700
0700-0715
0715-0730
0730-0745
0745-0800
0800-0815
0815-0830
1105-1120
1120-1135
1135-1150
1150-1205
1205-1220
1220-1235
1235-1250
1250-1305
1603-1618
1618-1633
1633-1648
1648-1703
1703-1718
1718-1733
1733-1748
1748-1803
First St.
In
130
177
207
304
287
220
196
214
235
210
221
212
207
251
258*
178
198
190
180
156
159
177
168
San Antonio
In
22
32
45
67
105
87
87
68
65
84
83
86
71
70
*
93
65
70
84
59
76
52
55
*
48
First St.
Out*
138
172
225
307
288
220
198
247
244
184
211
189
195
264
*
274
169
163
192
163
141
142
177
*
171
San Antonio
Out
15
26
27
49
104
65
67
59
68
76
74 )
71
65
72
*
87
57
72
75
57
69
51
48
*
30
Intersection
In
152
209
252
371
392
307
283
282
300
294
304
298
278
321
*
351
243
268
274
239
232
211
232
216
Intersection
Out*
153
198
252
356
392
285
265
306
312
260
285
260
260
336
361*
226
235
267
220
210
193
225
201*
Estimate—only partial data available.
D-22
-------
Table D-2
TOTAL TRAFFIC COUNTS FROM 291 SENSORS
IN DOWNTOWN SAN JOSE FOR TUESDAY,
24 NOVEMBER 1970
Time Interval
0648-0703
0703-0718
0718-0733
0733-0748
0748-0803
0803-0818
0818-0833
1104-1119
1119-1134
1134-1149
1149-1204
1204-1219
1219-1234
1234-1249
1249-1304
1603-1618
1618-1633
1633-1648
1648-1703
1703-1718
1718-1733
1733-1748
1748-1803
Total Traffic
10,154
12,927
16,810
22,172
22,218
18,341
*
18,441
19,293
19,229
19,946
21,409
20,866
20,742
20,359
20,153
22,914
22,542
25,775
24,042
27,183
20,813
18,292
*
16,353
Estimate—only partial data available.
D-2 3
-------
periods. Figure D-10 shows an example of this kind of summary plotted
on a schematic representation of the downtown traffic grid. For any
given time of day the traffic in the downtown area remains quite con-
sistent from day to day. This can be seen from the relatively small
deviations in Figure D-10. The relative variation of total traffic,
on all links, is even less than on the individual links. Figure 47
of the text shows the total downtown traffic volume on two successive
days in December 1970. It can be seen in that figure that the two
curves nearly coincide.
D-24
-------
Julian
St. James
St. John
Santa Clara
317
35
218
35
814
57
+"
a
^
u
fl
1
1
3:
•}
\
4
22
330
23
811 381 380
63 20 17
376
26
467
39
826 484 560
59 38 37
196
23
232
34
794 438
50 28 4'
907
46
1115
53
7*
824 560 501
49 35 47
i
:
J
'
2C
3
6
42
392
24
332
21
265
39
324
25
384
22
388
28
105
10
432
32
740
41
880
46
572
46
j
•
c
u
2!
3
3
53
16
260
35
275
17
270
•
8
298
21
344
22
369
40
87
10
845
48
San Fernando
San Antonio
San Carlos •
San Salvador
I
352
20
/
/
\
\
\
913 \
27
221
\ ^
276
28
252
26
235
25
165
19
\ 431 513 569 336
\
\130
/ 14
28 *
J2 1
121
12
6 4f
9 E
160
18
!5 5:
2 't
29
9
23 3
M
)0
36 28 43 21
/ 771
23
788
29
609
32
305 321 4(
t
2'
'
20 :
'5
26
37
553
29
)5
)3
Upper numbers are averages; lower numbers are standard deviations
based on 13 days' observations.
* Estimate—Insufficient data due to defective sensors.
TA-8563-39
FIGURE D-10 AVERAGE LINK VOLUMES, 1100-1130
D-25
-------
Appendix E
PILOT BALLOON DATA SUMMARY
E-l
-------
Appendix E
PILOT BALLOON DATA SUMMARY
Pilot balloon (pibal) soundings were made by the Meteorology
Department at San Jose State College. The pibals were released at
the College, which is located on the eastern leg of the helicopter
and van traverses (see Figure 20), and were tracked by double theodolite,
The soundings provide wind data for 11 periods with a maximum vertical
resolution of 72 m. The winds are given in geographic coordinates in
units of m s .
E-3
-------
Table E-l
PILOT BALLOON DATA SUMMARY—SAN JOSE STATE COLLEGE (1970)
z
72
144
216
282
348
414
480
546
612
675
738
801
864
927
990
2 November
0830 PST
u v
-1.4 0.0
-1.2 0.2
-1.4 0.6
-0.9 0.2
-0.3 0.0
2 November
1720 PST
u
3. 1
2.6
2.4
2.4
2.8
2.7
1.6
0.7
0.6
0.6
0.4
0.4
1.4
1.1
-0.6
v
-4.1
-4.6
-4.3
-4.6
-3.9
-1.9
-0.8
-0.9
-0.9
-0.7
0.4
2.4
2.7
2.0
2.9
3 November
0815 PST
u
-0.3
-0.5
-1.3
-2.7
-4.1
-4.2
-4.5
-5.0
-2.7
-1.3
-2.1
-1.6
-0.4
0.6
0.8
v
0.7
1.9
3.7
5.6
6.8
5.6
5.4
7.1
6.3
5.6
7.5
7.5
6.4
6.3
6.3
3 November
1645 PST
u
-0.2
-1.2
-1.6
-1.7
-1.8
-1.5
-0.9
-0.5
-0.2
-0.1
-0.2
-0.1
-0.4
-0.9
-1.4
v
4.6
5.0
4.9
4.3
4.0
4.1
4.1
4.1
4.3
6.1
7.8
8.8
10.4
10.8
10.3
4 November
1700 PST
u
-4.7
-5.0
-5.3
-6.1
-7.2
-7.4
-7.1
-7.3
-7.0
-5.8
-5.4
-5.0
-3.7
-2.4
-1.8
v
4.8
6.5
7.8
9.6
11.9
13.1
13.0
13.1
12.4
11.0
11.2
11.7
12.1
12.2
12.3
5 November
0900 PST
u
-1.0
-0.9
-0.9
-0.5
-0.1
0.3
0.2
0.4
0.6
0.6
1.2
2.6
3.9
4.2
5.2
V
5.2
7.1
7.5
8.2
8.7
8.3
8.2
8.7
9.5
9.1
10.3
10.7
10.6
13.0
15.3
5 November
170O PST
u
-0.1
-0.7
-1.0
-0.9
-0.2
0.2
0.8
1.2
0.8
0.7
0.1
3.0
-0.7
0.4
3.7
v
-1.5
-1.6
-1.9
1.7
-1.5
-0.8
0.0
0.9
1.7
1.9
2.4
1.6
1.7
5.3
6.4
10 December
1647 PST
u
1.6
1.9
1.6
1.3
0.9
0.6
0.9
0.7
0.4
0.1
-0.6
-0.3
-0.3
-1.3
-2.3
v
-1.5
-1.5
1.2
-1.0
-0.8
-0.5
0.1
0.0
-0.5
-1.4
-2.2
-1.1
0.0
-0.9
-2.6
11 December
0825 PST
u
-2.0
-1.4
0.2
1.7
1.7
1.2
2.0
1.8
0.0
-1.3
-2.0
v
0.4
1.2
0.5
-0.4
-0.3
0.5
0.0
-2.9
-4.5
-4.8
-5.2
11 December
1600 PST
u
4.8
4.4
1.7
0.4
0.4
-0.3
0.0
-0.1
-0.8
-1.3
-2.0
-2.6
-2.0
-1.2
-0.2
v
-1.2
-2.2
-4.2
-5.0
-4.2
-4.2
-4.6
-5.1
-5.2
-4.8
-4.4
-3.8
-3.3
-2.8
-3.0
15 December
1630 PST
u
-7.0
-6.1
-4.0
-2.2
-2.2
-2.2
-2.2
-3.3
-3.0
-1.9
-2.0
-2.1
-1.2
0.5
2.5
v
7.6
9.2
8.6
8.4
10.8
11.5
10.8
12.5
14.6
17.3
20.0
19.7
21.6
21.4
17.1
a
z = height (m).
u = east-west wind component (m s ).
v = north-south wind component (m s~ ).
-------
REFERENCES
Burggraf, O. R., 1966: Analytical and numerical studies of the
structure of steady separated flow, J. Fluid.Mech., 24 (1),
pp. 113-151. ~
Fox, J., 1964: Surface pressure and turbulent airflow in transverse
rectangular notches, NASA TN D-2501, 18 pp.
Fox, J., 1965: Heat transfer and air flow in a transverse rectangular
notch, Int. J. Heat Mass Trans., 8, pp. 269-279.
Frost, R., 1947: The velocity profile in the lowest 400 ft,
Meteorological Mag., 76, pp. 14-19.
Georgii, H. W., E. Busch, and E. Weber, 1967: Investigation of the
temporal and spatial distribution of the emission concentration
of carbon monoxide in Frankfurt/Main, Report No. 11 of the
Institute for Meteorology and Geophysics of the University of
Frankfurt/Main (Translation No. 0477, NAPCA).
Gifford, F. A., Jr., 1961: Use of routine meteorological observations
for estimating atmospheric dispersion, Nuclear Safety, £, 48.
Hoydysh, W., 1971: Personal communication, Env. Eng. Res. Labs., New
York Univ., Bronx, New York 10453.
Johnson, W. B., Jr., F. L. Ludwig, and A. E. Moon, 1969: Development
of a practical, multipurpose urban diffusion model for carbon
monoxide, Proceedings of Symposium on Multiple-Source Urban
Diffusion Models, Chapel Hill, North Carolina, 28-30 October 1969,
pp. 5-1 to 5-38.
Leighton, P. A., and R. B.Dittmar, 1952: Behavior of aerosol clouds
within cities, Joint Quarterly Report No. 2, October-December 1952,
Contracts DA-18-064-cmc-1856 and DA-18-064-cml-2282, Stanford
Univ. and Ralph M. Parsons Co., 100 pp., DDC No. AD 7261.
and , 1953a: ibid., Joint Quarterly Report No. 4,
January-March 1953, 218 pp., DDC No. AD 31509.
R-l
-------
REFERENCES (Continued)
Leighton, P. A., and R. B. Dittmar, 1953b: ibid., Joint Quarterly
Report No. 4, April-June 1953, 196 pp., DDC No. AD 31508.
and , 1953c: ibid., Joint Quarterly Report No. 5,
July-September, 238 pp., DDC No. AD 31507.
and , 1953d: ibid., Joint Quarterly Report No. 6,
Vol. I, October-December 1953, 246 pp., DDC No. AD 31510.
and , 1953e: ibid., Joint Quarterly Report No. 6,
Vol. II, October-December 1953, 187 pp., DDC No. AD 31511.
Ludwig, F. L., W. B. Johnson, A. E. Moon, and R. L. Mancuso, 1970: A
practical multipurpose urban diffusion model for carbon monoxide,
Final Report, Coordinating Research Council Contract CAPA-3-68,
National Air Pollution Control Administration Contract CPA 22-69-64,
Stanford Research Institute, Menlo Park, California, 184 pp.,
National Technical Information Service No. PB 196003.
Maull, D. J., and L. F. East, 1963: Three-dimensional flow in cavities,
J. Fluid Mech., 16, pp. 620-632.
McCormick, R. A., and C. Xintaras, 1962: Variation of carbon monoxide
concentrations as related to sampling interval, traffic, and
meteorological factors, J. Appl. Meteor., 1, pp. 237-243.
McElroy, J. L., and F. Pooler, Jr., 1968: St. Louis dispersion study,
Vol. II—Analysis, 51 pp., National Air Pollution Control
Administration Publication No. AP-53, 51 pp.
Ott, W., J. F. Clarke, and G. Ozolins, 1967: Calculating future carbon
monoxide emissions and concentrations from urban traffic data,
PHS Publ. No. 999-AP-41, National Center for Air Pollution Control,
Cincinnati, Ohio.
Pasquill, F., 1961: The estimation of the dispersion of windborne
material, Meteorological Mag. (London), 90, pp. 33-49.
Pooler, F., 1966: A tracer study of dispersion over a city, J. Air
Poll. Contr. Assoc., 11, pp. 677-681.
R-2
-------
REFERENCES (Concluded)
Rand McNally, 1970: Commercial Atlas and Marketing Guide, 101 ed.,
Rand McNally Co., Chicago, 657 pp.
Robbins, R. C., K. M. Borg, and E. Robinson, 1968: Carbon monoxide
in the atmosphere, J. Air Poll. Cont. Assoc., 18, pp. 106-110.
Rose, A. H., R. Smith, W. F. McMichael, and R. E. Kouse, 1964:
Comparison of auto exhaust emissions from two major cities,
U.S. Public Health Service, Cincinnati, Ohio.
Roshko, A., 1955: Some measurements of flow in a rectangular cutout,
NACA TN 3488, 21 pp.
Rouse, H., 1951: Air tunnel studies of diffusion in urban areas,
Meteor. Mono., 1, pp. 39-41.
Schnelle, K. B., F. G. Ziegler, and P. A. Krenkel, 1969: A study of
the vertical distribution of carbon monoxide and temperature
above an urban intersection, APCA Paper No. 69-152.
Turturici, A. R., 1970: 1969 Traffic Volumes, Metropolitan San Jose,
California, Dept. Pub. Works, City of San Jose, California, 39 pp.
R-3
------- |