United States Office of Air Quality EPA-450/3-78-028
Environmental Protection Planning and Standards June 1978
Agency Research Triangle Park NC 27711
__
Project Da Vinci II: ppnDCDTv
r\ ± n. i -i PROPERTY OF
Data Analysis and DIVISION
Interpretation MnrOF ,....,
-------
EPA-450/3-78-028
Project Da Vinci II:
Data Analysis and Interpretation
by
C.E. Decker, J.E. Sickles, II, W.D. Bach,
P.M. Vukovich, and J.J.B. Worth
Research Triangle Institute
Research Triangle Park, N.C. 27709
Contract No. 68-02-2568
EPA Project Officer: Dr. Edwin L Meyer
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
June 1978
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from the
Library Services Office (MD-35) , U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711; or, for a fee, from the
National Technical Information Service, 5285 Port Royal Road, Springfield,
Virginia 22161.
This report was furnished to the Environmental Protection Agency by
Research Triangle Institute, Research Triangle Park, N.C. 27709, in fulfill-
ment of Contract No. 68-02-2568. The contents of this report are reproduced
herein as received from Research Triangle Institute. The opinions, findings,
and conclusions expressed are those of the author and not necessarily
those of the Environmental Protection Agency. Mention of company or
product names is not to be considered as an endorsement by the Environmental
Protection Agency.
Publication No. EPA-450/3-78-028
11
-------
CONTENTS
Figures vi
Tables x
Acknowledgments xi
1.0 EXECUTIVE SUMMARY 1
1.1 Introduction and Objectives 1
1.2 Data Analysis 2
1.3 Principal Findings 2
1.4 Conclusions 5
2.0 INTRODUCTION 7
2.1 Background 7
2.2 Research Objectives 8
2.3 Report Organization 8
3.0 PROJECT DA VINCI II FIELD EXPERIMENT 9
3.1 Air Quality Measurements 10
3.1.1 Da Vinci II Measurements 10
3.1.2 Supporting Aircraft Measurements 11
3.1.3 Acoustic Sounder Measurement 13
3.1.4 RTI-EML Measurements 13
3.1.4.1 Ground Level Measurements 13
3.1.4.2 RTI Balloon-Borne Measurements 14
3.1.5 State Agency Measurements 16
3.1.6 RAMS Network Measurements 16
3.1.7 Emissions Inventory Data 18
3.2 Flight Description 20
3.3 Meteorological Conditions for the Period Encompassing
Flight Day 25
3.4 Quality Control Program 27
3.5 Analysis Approach 28
4.0 ATMOSPHERIC CHEMISTRY ANALYSIS FOR DA VINCI II 31
4.1 Objectives of Analysis 31
4.2 Approach 31
4.3 Results and Discussion 31
4.3.1 Emissions Inventory 31
iii
-------
Page
4.3.2 Air Contaminant Considerations for the Period
Encompassing Flight Day 37
4.3.3 Air Quality at the Launch Site 42
4.3.4 Air Contaminant Behavior at Ground Level
During the Flight 43
4.3.4.1 Air Quality at RAMS Stations 43
4.3.4.2 Air Quality at the RTI-EML 58
4.3.5 Air Contaminant Behavior Aloft During the Flight 62
4.3.5.1 NOAA Aircraft 62
4.3.5.2 Acoustic Sounder 70
4.3.5.3 Temperature, Relative Humidity, and
Condensation Nuclei on Da Vinci II 70
4.3.5.4 Ozone and Sulfur Dioxide on Da Vinci II 75
4.3.5.5 Hydrocarbons on Da Vinci II 82
4.3.6 Air Quality Aloft and at the Ground 90
4.3.6.1 Sulfur Dioxide 90
4.3.6.2 Ozone 92
4.3.7 Considerations for Future Programs 97
4.4 Findings 98
5.0 MESOMETEOROLOGICAL AND AIR POLLUTION STUDY RELATIVE
TO DA VINCI II 101
5.1 Introduction 101
5.2 Data Sources 102
5.3 Surface Temperature Distribution 102
5.4 Surface Wind Distribution 108
5.5 Temperature and Wind Profiles 113
5.6 Simulation Results 118
5.7 Surface Ozone Distribution 129
5.8 Da Vinci II and RTI-EML Ozone Data 133
5.9 Aircraft Ozone Profiles 138
5.10 Summary and Conclusions 139
6.0 METEOROLOGICAL ANALYSIS FOR DA VINCI II 145
6.1 Plume Identification 145
6.1.1 Introduction 145
6.1.2 Data Sources 146
6.1.3 Method of Computation 146
iv
-------
Page
6.1.4 Results 150
6.1.4.1 Wind Shear 150
6.1.4.2 Low Altitude Plume 153
6.1.4.3 The Plume at Altitude 154
6.1.5 Interpretation 154
6.1.6 Conclusions 157
6.2 Cross Section Analyses 158
6.2.1 Introduction 158
6.2.2 Analysis Approach 159
6.2.2.1 Analysis of Winds Aloft 159
6.2.2.2 Rawinsonde Analyses 162
6.2.2.3 Acoustic Sounder 165
6.2.3 Conclusions 167
6.3 Modeling Atmospheric Chemistry and Physics 168
6.3.1 Need for Modeling 168
6.3.2 Results of Model Application 169
7.0 PRINCIPAL FINDINGS AND CONCLUSIONS 171
7.1 Principal Findings 171
7.2 Conclusions 173
8.0 REFERENCES 175
APPENDIXES
Appendix A. Evaluation of Sample Collection Containers for
Hydrocarbon Sampling 177
Appendix B. Evaluation of Laboratory and Quality Control
Practices for Hydrocarbon Sampling and Analysis
(Da Vinci II Participants) 207
Appendix C. RTI Quality Control Program Relative to Da Vinci II
for Hydrocarbon Sampling 213
Appendix D. Objective Analysis Technique 219
Appendix E. Planetary Boundary Layer Model 223
-------
LIST OF FIGURES
Figure Page
1 Da Vinci II balloon and gondola 9
2 Photograph of gondola and scientific equipment 10
3 RTI Environmental Monitoring Laboratory onsite at
Arrowhead Airport 14
4 Track of Da Vinci II and route of RTI-EML during
first 12 h of flight 17
5 Flight track of Da Vinci II and ground track of the
RTI-EML from launch at Arrowhead Airport to touchdown
in Indiana 21
6 RTI-EML located adjacent to gondola for post-flight
calibration of Da Vinci II analyzers 22
7 Da Vinci II altitude data, 8 and 9 June 1976 24
8 RAMS meteorological parameters for 6-9 June 1976 27
9 Ozone distribution in the northeastern quadrant of
the United States based on daily maximum hourly
concentrations 39
10 Mean profiles for RAMS network on 6-9 June 1976 40
11 Hourly average ozone concentration prior to launch
at Arrowhead Airport 1-8 June 1976 43
12 Mean pollutant concentrations at urban, suburban, and
nonurban sites in St. Louis on 8 June 1976 46
13 Air temperature at urban, suburban, and nonurban
sites in St. Louis on 8 June 1976 48
14 Temperature gradient through the first 30 m (100 ft)
at urban and nonurban sites in St. Louis on 8 June 1976 48
15 Carbon monoxide concentration relative to downtown
St. Louis along west to east and south to north
directions on 8 June 1976 49
16 Nitrogen oxides concentration relative to downtown
St. Louis along west to east and south to north
directions on 8 June 1976 50
17 Ozone concentration relative to downtown St. Louis
along west to east and south to north directions on
8 June 1976 51
18 Nonmethane hydrocarbon concentration at the time of
maximum THC concentration relative to downtown St. Louis
along west to east and south to north directions on
8 June 1976 52
19 Hourly ozone isopleths (computer-generated) for St. Louis
RAPS area on 8 June 1976 55
VI
-------
Figure Page
20 Ground-level ozone concentrations measured by RTI-EML
and nearest RAMS Station during flight of Da Vinci II 60
21 Unadjusted ozone profiles as measured in St. Louis
on 8 June 1976 during vertical flights of the NOAA
aircraft that started at 0915 and 1148 CST 64
22 B profiles as measured in St. Louis on 8 June
scat r
1976 during vertical flights of the NOAA aircraft
that started at 0915 and 1148 CST 65
23 Relative humidity profiles as measured in St. Louis
on 8 June 1976 during vertical flights of the NOAA
aircraft that started at 0915 and 1148 CST 66
24 Nitrogen dioxide profiles as measured in St. Louis
on 8 June 1976 during vertical flights of the NOAA
aircraft that started at 0915 and 1148 CST 67
25 Adjusted ozone profiles as measured in St. Louis
on 8 June 1976 during vertical flights of the NOAA
aircraft that started at 0915 and 1148 CST 68
26 ANL acoustic sounder data along the flight track
of Da Vinci II, 8 and 9 June 1976 72
27 Air temperature on Da Vinci II, 8 and 9 June 1976 73
28 Relative humidity on Da Vinci II, 8 and 9 June 1976 74
29 Condensation nuclei on Da Vinci II, 8 and 9 June 1976 75
30 Ozone on Da Vinci II as measured by Sandia Laboratories
and ASL, 8 and 9 June 1976 77
31 Ozone on Da Vinci II as measured by ASL and represented
by connected dots, 8 and 9 June 1976 78
32 Sulfur dioxide on Da Vinci II as measured by Sandia
Laboratories and represented by connected dots, 8 and 9
June 1976 79
33 Pollutant concentrations as determined by RTI from grab
samples collected on Da Vinci II, 8 and 9 June 1976 84
34 Hourly average sulfur dioxide aloft and at ground level,
8 and 9 June 1976 92
35 Hourly average ozone aloft and at ground level, 8 and 9
June 1976 94
36 Ozone upwind and downwind of Indianapolis at an altitude
of 457 m (1,500 ft) above ground on 9 June 1976 as meas-
ured on the ICFAR aircraft 96
37 The area of immediate interest around St. Louis showing
the locations of the RAPS stations and the two upper air
sounding stations 103
VII
-------
Figure Page
38 The ground track of the Da Vinci II gondola and the
track of the RTI-EML from lift-off at Arrowhead Air-
port through 2000 CST 104
39 The hourly temperature difference between the St. Louis
center city and the rural/suburban area some 15 mi away 105
40 Spatial distribution of the St. Louis urban heat
island [°C] at 0500, 0900, and 1400 CST on 8 June 107
41 Spatial distribution of the St. Louis urban heat
island [°C] at 1800 and 2000 CST on 8 June 109
42 Interpolated surface wind field [m/s] at 0500,
0900, 1000, and 1400 CST on 8 June 110
43 Interpolated surface wind field [m/s] at 1500,
1800, and 2000 CST on 8 June Ill
44 Diurnal variability of the potential temperature [°K]
profile on 8 June over the St. Louis center city (sta-
tion 141) and over the rural area southwest of St. Louis
(station 142) 114
45 Diurnal variability of the wind speed [m/s] and wind
direction profiles over the St. Louis center city on
8 June 116
46 Diurnal variability of the wind speed [m/s] and wind
direction profiles over the rural area southwest of
St. Louis on 8 June 117
47 Horizontal flow and horizontal distribution of the
vertical velocity along with a vertical cross section
of the vertical velocity profile are shown 119
48 Horizontal distribution of the potential temperature
[°K] perturbation and vertical cross section of the
potential temperature field constructed along the
section line are shown 121
49 Simulated horizontal flow and horizontal distribution
of the vertical velocity along with a vertical cross
section of the vertical velocity profile are shown 122
50 Horizontal distribution of the potential temperature
perturbation; vertical cross section of the potential
temperature field; and the horizontal distribution of
the potential temperature perturbation are shown 123
51 Simulated horizontal flow and horizontal distribution
of the vertical velocity along with a vertical cross
section of the vertical velocity profile are shown 125
52 Horizontal distribution of the potential temperature
perturbation and vertical cross section of the poten-
tial temperature field are shown 126
Viii
-------
Figure Page
53 Simulated horizontal flow and horizontal distribution
of the vertical velocity with a cross section of the
vertical velocity profile are shown 127
54 Horizontal distribution of the potential temperature
perturbation and vertical cross section of the poten-
tial temperature field are shown 128
55 Simulated horizontal flow and horizontal distribution
of the vertical velocity along with a vertical cross
section of the vertical velocity profile are shown 130
56 Horizontal distribution of the potential temperature
perturbation and vertical cross section of the poten-
tial temperature field are shown 131
57 Surface ozone distribution in the study area at 0500,
0900, and 1300 CST on 8 June 132
58 Surface ozone distribution in the study area at 1600,
1900, and 2100 CST on 8 June 134
59 Variation of altitude with time for the flight of
Da Vinci II 135
60 Hourly average ozone concentrations measured along
the Da Vinci II flight track 137
61 Vertical ozone profiles flown by a NOAA aircraft
in the vicinity of the Da Vinci II gondola on the
morning of 8 June 1976 140
62 Pilot balloon launch sites 148
63 Typical analysis of air leaving St. Louis during
Da Vinci II flight 148
64 Dispersion of air, initially, over St. Louis at
indicated altitudes (km) and times, resulting
from wind shear 151
65 Movement of air at 50 m from St. Louis between
0800 and 2000 CST, 8 June 1976 and for 18 h of
transport 152
66 Motion of air initially over St. Louis at the
indicated altitudes and times and near the Da
Vinci II altitude 155
67 Horizontal wind vectors aloft along Da Vinci II
track, from pilot balloon observations 161
68 Cross section of potential temperature along
Da Vinci II path (from rawinsonde data) 164
69 Cross section of wind speed along Da Vinci II
path (from rawinsonde data) 166
70 Vertical profiles of u, v, 6, and c at indicated
times from PEL model simulation of 8 June 170
IX
-------
LIST OF TABLES
Table Page
1 Identification of selected data collected on
Da Vinci II 12
2 RTI-EML air quality instrumentation and calibra-
tion procedures 15
3 RAMS instrumentation 19
4 Counties for which emissions inventory data were
retrieved 20
5 Major altitude excursions of Da Vinci II 25
6 Major S02 point sources along the flight track
of Da Vinci II 32
7 Major NO point sources along the flight track
of Da Vinci II 33
8 Major HC point sources along the flight track
of Da Vinci II 34
9 Major CO point sources along the flight track
of Da Vinci II 35
10 Emissions summaries for counties along the
flight track of Da Vinci II 37
11 Summary of air quality parameters at Arrowhead
Airport for the 16-day period 44
12 Nocturnal increases in ozone concentration at
the surface during the flight of Da Vinci II 62
13 Mean HC concentrations at Da Vinci II and at
selected RAMS stations 86
14 Mean hydrocarbon concentrations and excursions
from the mean on Da Vinci II 88
15 Pilot balloon observations 147
16 Winds at 650 m MSL from RAMS pilot balloon
stations 158
17 Rawinsonde data used in analyses 163
-------
ACKNOWLEDGMENTS
This project was conducted by Research Triangle Institute (RTI), Research
Triangle Park, North Carolina, under Contract 68-02-2568 for the United States
Environmental Protection Agency (EPA). The support of this agency is grate-
fully acknowledged as is the advice and guidance of the Project Officer, Mr.
E. L. Meyer, Jr., and other staff members of the Office of Air Quality Planning
and Standards.
Work on this project was performed by staff members of the Systems and
Measurements Division and the Energy, Engineering, and Environmental Sciences
Division of RTI, under the direction of Mr. J. J. B. Worth, Group III Vice
President. Mr. Worth was Laboratory Supervisor for the project. Mr. C. E.
Decker served as Project Leader and was responsible for the coordination of
the program. Staff members of RTI who contributed to the preparation of this
report are recognized and listed in alphabetical order: Dr. W. D. Bach, Mr.
C. E. Decker, Mr. R. B. Denyszyn, Dr. J. E. Sickles II, Dr. F. M. Vukovich,
Dr. D. E. Wagoner, and Mr. J. J. B. Worth. Special recognition is given to
Dr. L. A. Ripperton for technical guidance and advice in the interpretation of
the data.
Special recognition is given to Dr. B. Zak of Sandia Laboratories who
directed the Da Vinci II project and contributed valuable information for use
in this report.
xi
-------
1.0 EXECUTIVE SUMMARY
1.1 Introduction and Objectives
During 8-9 June 1976, Research Triangle Institute (RTI) under sponsorship
of the U.S. Environmental Protection Agency (EPA) participated in Project Da
Vinci II, an Energy Research and Development (ERDA) experiment. Project Da
Vinci II was a manned, balloon-borne scientific experiment conducted by Sandia
Laboratories for ERDA in the St. Louis, Missouri, area to study the behavior
of air pollutants in the lower atmosphere. The primary emphasis of the experi-
ment was focused upon the chemical and meteorological processes by which
gaseous effluents from urban areas are reacted or transformed while being
transported away from an urban area. Measurements of sulfur dioxide, sul-
fates, ozone, ozone precursors, and meteorological parameters were made to
determine changes in their relative concentrations in space and time as the
balloon drifted with the wind.
The underlying concept of the Da Vinci II balloon flight takes advantage
of a quasi-Lagrangian frame of reference for an instrumented platform to
monitor atmospheric processes within a moving air parcel. In this study, both
chemical and meteorological processes could be monitored continuously as they
were occurring. This approach afforded the opportunity to follow the transport
of various pollutants and examine their behavior during transport. In addi-
tion, pollutant behavior could be investigated under varying meteorological
conditions, such as occur in the daytime mixed layer or within the layer
bounded by the subsidence inversion and the radiation inversion at night.
Da Vinci II was launched from Arrowhead Airport, 24 km west of St. Louis,
Missouri, on 8 June 1976, at 0756 CST. The craft carried a four-person crew
and a payload of scientific equipment designed to quantify selected chemical
and meteorological parameters associated with the atmosphere. Supporting data
were also collected by an aircraft, the RTI Environmental Monitoring Labora-
tory (RTI-EML), and at ground stations. Da Vinci II landed at 0805 CST on
9 June 1976, in southwest Indiana after having traveled some 340 km in a 24-h
period. A report of the Da Vinci II field program is presented in EPA Report
No. 450/3-77-009.l
The objective of this study was to consolidate, analyze, and interpret
selected data collected during the flight of Da Vinci II in order to expand
the present understanding of the transport of air pollution from urban to
-------
nonurban areas, with primary emphasis on the transport of ozone. Emphasis was
also focused on relating emissions upwind, within, and downwind of the urban
area to observed levels of primary and secondary pollutants upwind, within,
and downwind of the urban area. An additional consideration was the inter-
pretation of the data so that EPA could further develop and apply methodol-
ogies for evaluating the effectiveness of existing or proposed control strate-
gies for meeting ambient air quality standards.
1.2 Data Analysis
The analyses of both the chemical and the atmospheric processes for Da
Vinci II consider data taken from two frameworks. Ground station chemical and
meteorological and upper air data were taken from a Eulerian system (i.e., the
observer is fixed in space and observes changes as the medium moves past),
while Da Vinci II measurements (aboard the balloon) were taken in a nearly
Lagrangian system (i.e., the observer is moving (approximately) with the
medium and measures the changes that occur from processes within the moving
volume). Meteorological characteristics of the balloon environment are re-
lated to the chemical and oxidant concentrations measured aboard the Da Vinci
II balloon and to measurements from other platforms—aircraft, the RTI-EML,
and ground stations.
Information was assembled from many sources to provide a comprehensive
description of the atmospheric and pollutant behavior during the study period
both aloft and at ground level. Aircraft data, when available, were used to
supplement information from the balloon to describe processes that were occur-
ring aloft. Data from the Regional Air Monitoring Study (RAMS) network, the
RTI-EML, and Indiana, Illinois, and Kentucky State pollution monitoring sta-
tions were employed to describe pollutant behavior at the ground. Emissions
inventory data were also examined to locate major pollutant sources within the
study area.
Independent analyses of the data were performed to provide chemical and
meteorological interpretations for Da Vinci II. These analyses are documented
in detail in the body of the report.
1.3 Principal Findings
Selected data collected by various participants during the flight of Da
Vinci II were consolidated, summarized, and segmented into general subject
areas for analysis and interpretation. These subject areas include atmos-
pheric chemistry and meteorological analyses. Data were analyzed and inter-
-------
preted according to the objectives for each analysis and have been incorpor-
ated to combine both a chemical and meteorological interpretation of results.
Principal findings from these studies are summarized below.
1. Conditions of high ozone concentrations persisted aloft for the
study period. These conditions were widespread and extended for
several hundred kilometers. Air containing high concentrations of
ozone, approximately 0.11 ppm, was being transported into the study
area on the flight day.
2. Da Vinci II was launched into air that had an immediate fetch over a
nonurban area, although the total experiment was conducted within a
stagnant, polluted, high-pressure system. Air parcel trajectories
suggest that Da Vinci II was launched into air that 72 h earlier had
been in eastern Kentucky and the Ohio Valley.
3. Da Vinci II did not travel in the heart of the St. Louis urban plume
as defined by ozone concentrations. Examination of the hydrocarbon
concentrations sampled on Da Vinci II suggests that the flight oc-
curred in air characteristic of suburban-to-nonurban areas.
4. The ozone concentration at the flight level (700 m) of Da Vinci II
was greater than 0.08 ppm when Da Vinci II reached that altitude.
Since it reached that level prior to 1030 CST when deep convective
mixing was experienced in the urban boundary layer, a major portion
of the upper level ozone concentration must have come from sources
other than St. Louis on the morning of 8 June.
5. During the morning, ground-level ozone concentrations were less than
concentrations aloft. The observed increase of ozone concentrations
aloft, therefore, could not have resulted from upward mixing of
ozone from the ground, but was probably due to tropospheric photo-
chemical synthesis.
6. Vertical profiles show substantial ozone concentrations aloft be-
tween 762 and 1,981 m (2,500-6,500 ft) on the morning of 8 June with
a sharp decline in ozone concentration above 2,133 m (7,000 ft).
Mean ozone concentrations between 2,133 and 2,743 m (7,000-9,000 ft)
were less than those measured between 762 and 1,981 m (2,500-6,500
ft) by a factor of 2.4. These observations suggest that stratospher-
ic intrusion was not responsible for the elevated ozone concentra-
tions that existed aloft on 8 June.
7. The impact of anthropogenic emissions on ambient concentrations
within the urban area is variable and depends strongly on the time
of day. Ground-level CO and NO were diluted by factors of 5 to 7
between the morning and afternoon. This behavior probably reflects
the significant increase in the mixed volume that occurs with the
dissipation of the surface-based radiation inversion and the estab-
lishment of well mixed conditions.
8. As vertically well mixed conditions were established between 0500
and 1100 CST, both downward mixing and photochemical synthesis
contributed to the observed increase in ground level ozone concen-
tration. This ground level increase exceeded that aloft by a factor
of approximately two.
-------
9. Between 1100 and 1700 CST on flight day, a near zero ozone con-
centration gradient existed from the ground into the mixed layer.
This is based, to a large extent, on the close comparison of ozone
concentrations determined on Da Vinci II with those beneath Da Vinci
II at ground level.
10. After well-mixed conditions were established and until the nocturnal
inversion began to form, the increases in ozone concentrations aloft
and at ground level were approximately equal and probably resulted
from photochemical synthesis.
11. The airflow in the immediate vicinity of St. Louis on 8 June was
governed by the synoptic scale flow and an intense circulation
produced by the urban heat island. Da Vinci II responded to both of
these scales of motion.
12. A major air pollution plume (defined by the ozone distribution) was
found at the surface. The maximum ozone concentration in this plume
was immediately downstream (within 5 km) of the center of downtown
St. Louis between 1000 and 1500 CST. Simulation results (meteorolog-
ical model) suggest that if a plume existed in the upper portions of
the boundary layer, it would have moved in a different direction
than the surface plume.
13. Between 1000 and 1500 CST, the maximum ozone concentration in the
major air pollution plume at the surface (as defined by ozone distri-
bution) was found in a region where there was a zone of horizontal
convergence in the wind field associated with the urban heat island
circulation.
14. After 1500 CST, the maximum ozone concentration in the air pollution
plume (as defined by the ozone distribution) moved farther downstream
(greater than 15 km) when the heat island circulation and its accom-
panying convergence zone was dissipating.
15. Within the layer that had been well mixed during the daytime, strat-
ification, which occurred at night with the establishment of the
nocturnal radiation inversion, resulted in the formation of two
regimes of ozone concentrations. Ozone concentrations within the
radiation inversion were much reduced compared to levels above it.
This is presumably due to destruction by surface deposition and
reaction with ozone-destructive agents emitted into and trapped
within the inversion.
16. Ozone data obtained aboard Da Vinci II were used to calculate a
nighttime ozone half-life of 116 h. Examination of available hydro-
carbon precursor data obtained aboard Da Vinci II suggests that
concentrations of ozone destructive and other HC species aloft were
low during the flight. The dark phase half-life of ozone under the
conditions of presumed low levels of precursors is sufficient to
allow transport of ozone from an urban area overnight to another
area without significant diminishment due to decay.
17. Penetration through the nocturnal radiation inversion resulted in
the mixing of high levels of ozone to the ground and precursors from
the ground to aloft. This was suggested by the sharp nighttime
ozone peaks and the associated declines of NO and CO observed at
-------
selected ground stations. This phenomenon occurred frequently in
May and June and was observed at ground stations covering an area of
several hundred kilometers. It may also provide mechanisms for in-
creasing nighttime ozone destruction aloft and for enhancing early
morning ozone synthesis by distributing ozone precursors aloft above
the inversion before sunrise.
18. The balloon may have entered an air parcel that was enriched in
hydrocarbons after 2100 CST on 8 June and traveled within this par-
cel for the remainder of the flight.
19. Sharp, short-term reductions in ozone concentrations occurred aloft
during the nighttime portion of the flight and were coincident with
increases in S(>2 concentration. Both SC>2 and NO are emitted by
power plants. Although NO was not measured in this study, the
observed ozone behavior is probably the result of destruction by
reaction with NO .
x
20. The influence of ozone precursors emitted in the urban area was
manifested at ground level by ozone concentration enhancement over
nonurban concentrations. The magnitude of this enhancement was 0.06
to 0.11 ppm. Both the buildup and movement of the region of enhanced
ozone were documented, although detailed definition of the extent
and magnitude were severely hampered by the lack of a comprehensive
supporting aircraft measurement program.
21. High ozone concentrations measured aloft aboard Da Vinci II on the
morning of 9 June in Southwestern Indiana are attributed to long
distance transport of ozone. High ozone at the surface is influenced
by synthesis and mixing downward of the ozone aloft in the morning.
After a well-mixed layer is established, changes resulting from
mixing are minimized. Further increases in ozone concentration at
the surface can be attributed to synthesis.
22. The use of a Lagrangian system to document ozone transport and
address atmospheric chemistry problems has been shown to be feasible.
Interpretation of these data are severely limited, however, unless
supporting data that define vertical and horizontal pollutant distri-
butions are also available.
1.4 Conclusions
The conclusions listed below are derived from the analysis of the Da
Vinci II data relative to the objectives of this project.
1. The synoptic wind flow pattern as modified locally by the urban heat
island circulation prevented Da Vinci II from passing over the
St. Louis urban area and taking up a position in the urban plume.
2. During the nocturnal phase of the flight, changes in ozone con-
centrations observed aboard Da Vinci II could not be directly attri-
buted to the downtown St. Louis urban plume.
3. General meteorological conditions—a subsidence inversion aloft and
a strong radiative inversion based at the ground—were nearly ideal
for long-distance transport aloft of ozone at night, while keeping
ozone separated from ground level emissions and destruction and
limiting vertical mixing.
-------
4. The dark-phase stability of ozone above the nocturnal surface-based
radiation inversion suggests that transport of ozone can occur aloft
over long distances at night without significant diminishment. It
is possible for the ozone to be transported overnight several hundred
kilometers and be mixed to the ground on the next day with a signifi-
cant impact on ground-level air quality.
5. The occurrence of high ozone concentrations aloft on the morning of
9 June, in a rural area in southwestern Indiana, is attributed to
long-distance transport of ozone.
6. Balloon-borne experiments are a feasible approach for addressing
atmospheric chemistry problems, provided supporting data from ground
stations, a ground-level chase vehicle, and an instrumented aircraft
are available to complement the interpretation. The omission of any
one of these measurement platforms severely limits the interpreta-
tion effort.
-------
2.0 INTRODUCTION
2.1 Background
Previous studies of nonurban ozone concentrations have led to the conclu-
sion that nonurban ozone is significantly affected by long-distance transport
2 3
of ozone. ' These studies have suggested that the unique precursor, synthe-
sis, destruction, and transport conditions within a high-pressure system are
conducive to the transport of ozone in large concentrations for long distances.
At night, the radiative temperature inversion develops, stabilizes the air,
and inhibits vertical transfer processes. The inversion effectively insulates
the ozone aloft from destruction by contact with the surface or by reaction
with ozone-destructive agents (NO, NC^) that may be emitted near the ground.
Overnight, the air is transported farther downwind and diffused horizontally.
After sunrise, heat is gradually added to the surface layers, while concurrent-
ly ozone is being generated by photosynthesis in the precursor-rich air. When
enough heat has been added to eliminate the "inverted" temperature structure,
vertical mixing is no longer inhibited and ozone-rich air from aloft may be
brought to ground.
Understanding the role of urban ozone and its interactions with rural air
quality background levels through transport, diffusion, and chemical reactions
is important for strategy decisions relating to regional air quality control.
One of the specific questions yet to be answered is, "How do urban ozone and
associated precursors affect the observed high levels of rural oxidant?" The
basic strategy question is whether urban hydrocarbon control will reduce
oxidant concentrations equally in the city and in the downwind rural areas.
To examine the question of urban transport, it is necessary to trace or
move with an air parcel from a highly urban area into a "relatively clean"
rural environment. An opportunity for such a definitive field experiment
existed during June 1976 in association with a scheduled Energy Research and
Development Administration (ERDA) experiment, Project Da Vinci II.
Project Da Vinci II was a manned, balloon-borne scientific experiment
conducted by ERDA Sandia Laboratories in early June in St. Louis, Missouri, to
study the behavior of air pollutants in the lower atmosphere. The primary
emphasis of the experiment was focused upon the chemical and meteorological
processes by which gaseous effluents from urban areas are reacted or trans-
formed while being transported away from an urban area. Measurements of
-------
sulfur dioxide, sulfates, ozone, ozone precursors, and meteorological param-
eters were made to determine changes in their relative concentrations or
values in space and time as the balloon drifted nearly with the wind. The Da
Vinci II field program is described in EPA Report No. 450/3-77-009.
The balloon-borne measurement program offered these distinct advantages:
1. Air quality measurements could be made in a quasi-Lagrangian frame
of reference (i.e., within a moving air parcel).
2. The results of chemical processes occurring in an urban air parcel
could be continuously monitored over the time interval of transport.
3. Continuous airborne sampling could be conducted within a layer of
air bounded aloft by the subsidence inversion and below by the
ground-based radiation inversion.
4. This experiment provides the needed data to document uniquely the
contribution of a single city's effluvia to the background pollution
levels within a given air parcel.
2.2 Research Objective
The objective of this study was to consolidate, analyze, and interpret
selected data collected during the flight of Da Vinci II in order to expand
the present understanding of the transport of air pollution from urban to
nonurban areas. Emphasis was focused on relating emissions upwind, within,
and downwind of an urban area to observed levels of primary and secondary
pollutants upwind, within, and downwind of the urban area. An additional
consideration was the interpretation of the data, so that EPA could further
develop and apply methodologies for evaluating the effectiveness of existing
or proposed control strategies for meeting ambient air quality standards.
2.3 Report Organization
This report includes four major sections that describe the field ex-
*
periment and present an in-depth analysis and interpretation of the Da Vinci
II data. Section 3.0 contains a general description of the Da Vinci II field
experiment and includes mobile and airborne measurements, flight description
and meteorological considerations during the flight, quality control program,
and the analysis approach. In section 4.0, atmospheric chemistry analyses are
described and results are discussed. In section 5.0, an in-depth evaluation
of the mesometeorological conditions during the flight and the effect of the
heat island circulation on Da Vinci II are addressed. Meteorological analyses,
such as plume identification, cross section analysis, and modeling of atmos-
pheric chemistry and physics are described in section 6.0.
-------
3.0 PROJECT DA VINCI II FIELD EXPERIMENT
The Da Vinci II system was assembled under the direction of Sandia Labora-
tories and consists of a double-decked gondola approximately 3 m square; a
22-m suspension harness that connects the gondola with the balloon; data-re-
ceiving systems for the chase vehicles; and navigational and life support
equipment for a crew of four. The gondola and balloon system was designed to
permit several days of flight for a crew of four with approximately 909 kg of
scientific and power supply equipment on board.
The balloon is constructed of 2-mil, double layer, polyethylene (fig-
ure 1) and has a nominal volume of about 186,000 ft3. It has a diameter of
about 22 m, a gore length of 36 m, and 49 load-bearing tapes of 450 kg test.
Helium was used as the lifting gas. Two valves at the balloon apex are opera-
ted on separate electrical systems to vent helium as needed to assist in
Figure 1. Da Vinci II balloon and gondola.
-------
controlling the altitude of the balloon. Figure 1 is a picture of the inflated
balloon and gondola prior to launch.
The gondola was constructed by the Grumman Aircraft Corporation and is
3 m square and double-tiered. It is especially designed to protect crew and
equipment in event of rough landing and yet to be used again. The lower deck
has about 1 m of head room and is used for batteries, supplies, ballast, life
support equipment, and crew sleeping quarters. The upper deck is fiberglass
and is used for flight operations and for conducting the scientific experi-
ments. The bottom of the gondola is equipped with a layer of impact material
or "crush pads" about 0.5 m thick to provide shock absorption in landings.
Scientific equipment is located throughout the flight train. A closeup photo-
graph of the gondola and supporting scientific instrumentation is shown in
figure 2.
3.1 Air Quality Measurements
3.1.1 Da Vinci II Measurements
During the course of the Da Vinci II experiment, both air quality and
meteorological data were collected at many locations by various organizations.
These included data collected aloft on the balloon and by supporting aircraft;
and data collected at the ground by the RTI-EML chase vehicle, Argonne National
Figure 2. Photograph of gondola and scientific equipment.
10
-------
Laboratory (ANL) acoustic sounders, selected State agencies, and the Regional
Air Monitoring System (RAMS) network.
The types of data that were collected on the balloon are identified and
described in table 1. Sandia Laboratories provided altitude data and concen-
tration data for ozone, sulfur dioxide, carbon monoxide, methane, hydrogen,
and neon. The last four species were determined by analysis from grab samples.
Altitude data are 1-min averages obtained from an encoding altimeter. Ozone
and S(>2 concentration data are also 1-min averages. The 63 and 863 instruments
were operated only during selected intervals over the flight in an effort to
conserve electrical power. The Dasibi ozone monitor and the TECO pulsed
fluorescence sulfur dioxide monitor were calibrated by RTI both before and
after the flight. Personnel from the National Oceanic and Atmospheric Agency
(NOAA) aboard Da Vinci II collected temperature, relative humidity, and conden-
sation nuclei data manually during the flight. Temperature and relative
humidity were measured with a sling psychrometer. Condensation nuclei were
determined with a manually operated Gardner instrument.
®
Grab samples were collected in light-shielded Tedlar bags for subsequent
determination of selected hydrocarbons and halocarbons by the Research Triangle
Institute. Grab samples were also collected in evacuated stainless steel cans
for analysis by Washington State University. Samples were collected sequen-
tially; this prevented the simultaneous collection of samples of the same air
parcel.
The U.S. Army Atmospheric Sciences Laboratory (ASL) flew instruments
aboard Da Vinci II for determining temperature, relative humidity, and ozone.
Ozone was determined continuously by the solid-phase chemiluminescent instru-
ment located on the balloon. Data from this device were recorded at 2-s
intervals. The solid-phase instrument was calibrated against the Dasibi.
3.1.2 Supporting Aircraft Measurements
Two vertical flights were conducted within 2 km of the balloon on 8 June
by a NOAA aircraft. The first vertical was flown from 0915 to 0933 CST and
the second was flown from 1148 to 1211 CST. During these flights 03, NO, N02,
NO , Bscat, and relative humidity were measured. Both 03 and NO were deter-
X X
mined by Monitor Labs instruments. Cross calibrations were not conducted,
thus preventing a comparison with other instruments employed in the Da Vinci
II experiment.
11
-------
(—1
• H
U
d
•H
>
(0
Q
d
o
73
01
U
01
r-l
r-l
O
u
03
4-1
(0
•o
*o
01
4J
u
01
r-l
01
M
*4-f
O
d
o
•H
4J
(0
U
•H
•H
d
01
73
•
rH
01
r-l
f*i
H
4^
d
4)
O
CJ
r-l
0)
4J
4)
a
03
PM
^J
d
(0
•H
U
•rl
4-)
(0
PL,
i M id
1 >>O 00 >* 01
2d (0 W /— , 0) 7© M di 734-1(0 d W 4->
Od 4IM U 43 r-l 4) M 5>> OIM -HOO)
CJ r-l 4J CJ 73 i— 1 U 4-)(Or-IOI 4-> CJ 8 <•» iH
W UMOrH 3 -H O r-H tO r-l U CO O • 3 D ft
U H 4J OIM 3 CQ V-l 4-1 OI73dOO OIM < r— 1 (-1 Q
. 4-1 « CJ r-l H «3 rH rH 0) 8 OJ U
o) oi d u o d 1-1 u (-1 t—l f>> W T3 73 W 4-1 4J S >
r-l 4-1 U O OIMOIPJ M 43 « H 0) r-H 0) 43 OIMO)r-l 01 ••« O
<1 W 0) 3 rH 73< 4)Ua rHOIiHU r-l 73r-l M4->
dr-lrH ft73 CJ 4J>>CO~ Q4-H-HM 0,73 3 (OdM
ao w H PI., a D ^ iz; O>M t-i a 43 « m SUMOH 43 oi
d 1 fl) 4-1 O a P-l H O (OM4-IOI fl) 4-1 O ft B CM •
•rl-HO73 W n) m >, O OI4-) MIOIM M03MH4J 13 4-1
O 'r4 W M 4-i 43 U M M d 73 0) 43 <-] p^ jQ o M "H 4J W M
UM4)i-!d(Q(OdM -l 4-1 (OOOM • (0(0d«> i— 1 W 0) -H
da34434)M>(Q-H 43 rH (0 D )>l-rlO4-)CJ )-l>(0>i-l O d > r-H
MQHPUaOOIUM HCOUQ O rH m nj Z CD
01 -rt 3 O »f«H W -H
oiiz; 73!z; ~iW «~ T3
73 -rl CM -H ^ >- CJ W -H
•H- sdWOrH o< a
X« OI3O M *CJ O « CJ CJ (0 >
3!l« U-HdWHO U-H
4JOI3CJ 0)4^01 CJ'-SCi'H «s^ 4IOI4-)
•rid1*-! 0(0373 CD VP3 CM -*CM d ft «
4->OiH~ Hr-td •> W CJ U» StjOjrH QBiH
rHN3O 4IO>O OniCJ 0
73 /*-> d -H -H
M d < M d w >>
0) (fl < O 01 M
•rl O 01 4S O
(H U % rH 01 ft 4->
O -H v— ' 00 4-1 M (0
4J d d oj o u
(0 (0 r*> 03 4-1 a O
!-l 01 U -rl CO 4-1 43
o u d )-i < (0
43 O 41 H /-N d'-v 1-3
OJ 00 M O P >%
ri rH«( 43H 4JCO BM
(0 c> • 0) t-H
d 4-101 W 4-1 W 4-1 CO -ft CO
(0 0343 0)3 OJ-H 'O !3BM ?CO v— '
12
-------
Between 1500 and 1743 CST on the afternoon of 9 June after the balloon
had landed, two regional-scale horizontal flights were conducted by the Indian-
apolis Center for Advanced Research (ICFAR). Ozone was measured at 458 m
(1,500 ft) above ground level (AGL) and 671 m (2,200 ft) AGL over a distance
extending from 225 km northeast to 290 km southwest of Indianapolis. Cross
calibrations were not conducted with other instruments employed in the Da
Vinci II experiment. Both the NOAA and ICFAR data were used as reported by
the respective investigators.
3.1.3 Acoustic Sounder Measurement
Sequential acoustic soundings of the lower atmosphere were conducted by
Argonne National Laboratory (ANL). Two mobile acoustic sounders were employed
at ground level to measure the thermal structure of the planetary boundary
layer. These devices were employed in a "leap frog" fashion to follow the
track of the balloon.
3.1.4 RTI-EML Measurements
3.1.4.1 Ground Level Measurements
Continuous ozone, nitrogen oxide (NO, NO ), and sulfur dioxide measure-
A
ments and continual (i.e., once every 5 rain) measurements for THC, €#4, and CO
were made aboard the RTI-EML for a 16-day period prior to the launch of Da
Vinci II and during the actual flight. During that period, the RTI-EML was
parked at Arrowhead Airport, located approximately 24 km west of St. Louis.
®
Sample air was aspirated through a Teflon -glass manifold system from a height
of 10 m. Figure 3 shows the RTI-EML onsite at Arrowhead Airport prior to the
launch of Da Vinci II. During the in-transit measurement period, sample air
®
was provided to the instruments through a Teflon -glass manifold. Its inlet
was located approximately 3.5 m from the ground and extended 1 m in front of
the RTI-EML. Theoretical calculations and extensive road tests were conducted
to insure that effects due to aerodynamic characteristics of the RTI-EML and
vehicular exhausts on the ambient air sample were minimized. A minimum flow
of 0.1 m3/min was maintained through the manifold at all times. To demonstrate
the feasibility of making mobile air quality measurements using this system,
drive-by comparisons were conducted on the RTI campus using a stationary ozone
monitor (i.e., ground station) and multiple passes of the RTI-EML. Ozone
measurements were selected for comparison purposes. Results of multiple
13
-------
Figure 3. RTI Environmental Monitoring Laboratory
onsite at Arrowhead Airport.
comparisons showed agreement, in the mean, between stationary and mobile ozone
measurements to within 0.003 ppm.
Instrumentation and calibration procedures used to obtain ambient air
measurements prior to the flight and during the flight for the above-mentioned
pollutants are summarized in table 2. Appropriate quality control procedures
and a sufficient number of instrument calibrations were performed to obtain
high quality data. Quality control procedures included verifying calibration
procedures, standards, and operating procedures; performing dynamic calibra-
tions and checks; maintaining adequate records to describe instrument per-
formance; and thorough training of the instrument technicians.
3.1.4.2 RTI Balloon-Borne Measurements
Instrumentation and equipment were installed on board the Da Vinci gondola
®
by RTI for the collection of time-integrated samples in Tedlar bags at hourly
intervals for subsequent selected hydrocarbon and halocarbon analyses at RTI.
The detailed hydrocarbon analyses were performed using a modified Per-
kins-Elmer Model 900 gas chromatograph coupled to a Hewlett-Packard Model
2100A computer. The following ten hydrocarbons were selected for routine
analysis:
14
-------
CO
OJ
^1
3
TJ
OJ
U
O
n
ft
d
o
• H
4_)
CO
f-t
n
•H
rH
CO
U
d
«0
d
o
•H
4_)
CO
4J
d
0)
M
4_)
CO
d
• H
^*
4_)
•H
rH
CO
3
0<
M
•H
CO
(--I
W
I
«
^"—^ *H
EH
^3 oj
00 CO
CO
co .d
^jr
CO
3
O
3
d
•H
^j
d
0
CJ
rO
ft
U")
OJ
u
d
OJ
CJ
CO
OJ
d
•H
O
rH
•H
S
OJ
f{
rj
X
•H m
d o
OJ rH
PQ 00
X
0
%
OJ
fl
3
H
d
o
•H
^j
CO
OJ
c
OJ
CM
CO
hyj
CO
3
O
3
d
•rl
4-1
d
O
CJ
rQ
ft
ft
C**I
OJ
CJ
d
OJ
M
OJ
O
3
rH
TJ
OJ
ca
rH
3
O4
oo
vj
o
u
OJ
H
M
O
CO
my
OJ
•r-l W
•rl OJ
t-l d
OJ -H
CJ rH
d cj
•H
0 1
cj d
•H
"O S
d 3
CO rH
^
W
CJ
CO
.. aj
CJ 4-1
•H 3
rH d
CJ -rl
^t S
LO
^Q
ft
ft
o
CM
d
o
•H
4-1
CO
N
•H
d
o
M
OJ
a
CO
rH
d
CO O
J§
CJ VO
OJ
PQ
u *<
EH CJ CJ
.
M
d
o
ft OJ
M T3
CO -rl
0 X
O O
>-i d
T3 O •
>> • a d
Jd QJ QJ
d d «>
rH CO O O
CO XI ft H
4J 4J V4 4->
O OJ CO -H
4J a cj d
1 1 1 1
1 1 1 1
CJ * O CM
g«CJ*
.
d
OJ
00
o •
M OJ
• -P TJ
OJ -H -H
T> d X
•H O
X >« -H
O O TJ
• U W M
OJ -H OJ 3
d rJ TJ "4-1
O 4-> -H rH
N -rl X 3
o d o M
lilt
III!
«O XCN
O SB O o
•i< Jz; co
15
-------
ethylene/ethane n-butane
acetylene 1-butene
propane trans-2-butene
propylene isopentane
isobutane cyclopentane
Separation of the C2~Cs hydrocarbons was made on a 1.8-m x 0.15-cm i.d.
Durapak n-octane (100-120 mesh) column that was operated at 23° C. The sum of
the above 10 nonmethane hydrocarbons was computed for each grab sample collec-
ted on board the Da Vinci II system and is hereafter referred to as I NMHC.
Methane and carbon monoxide concentrations were measured on each grab sample
using a Beckman 6800 Air Quality Chromatograph.
Freon 11 and Freon 12 analyses were performed on grab samples collected
during flight using a Perkins-Elmer Model 3920 gas chromatograph with electron-
capture detector. These compounds were separated on a 2-m glass column packed
with Chromosorb W-H.P. and coated with 10 percent DC-200. A quality control
program was conducted by RTI in conjunction with hydrocarbon sampling and
analysis and is described in appendix C.
3.1.5 State Agency Measurements
Hourly average ozone data were acquired for the period 6-10 June from 35
air monitoring sites operated by various State air pollution control agencies.
The data were collected at nine Kentucky stations, six Indiana stations, and
twenty Illinois stations.
3.1.6 RAMS Network Measurements
The U.S. Environmental Protection Agency has sponsored a multiyear air
pollution measurement program in St. Louis, Missouri. This program is known
as the Regional Air Pollution Study or RAPS. A network comprised of 25 ground
stations was established to provide selected air quality and meteorological
data for the St. Louis area on a continual basis. This network is known as
the Regional Air Monitoring System (RAMS).
RAMS stations are located on concentric circles having radii of 4, 9, 20,
and 40 km from a center in downtown St. Louis. The stations are numbered from
101 to 125 with the station numbers generally increasing with distance from
the downtown area. The locations of stations 101 through 121 are shown in
figure 4. The remaining four stations, the most remote in the network, are
located just off the map; station 122 is located north, 123 is east, 124 is
south, and 125 is west.
16
-------
I
.inn!
jjj D D 0 ® | 4
HI
iiiiilihii
ttr
is*
ill
ill il
17
-------
A summary of selected instrumentation for the RAMS network is shown in
table 3. With few exceptions, each station is fully equipped. Monitor Labs
instruments were employed to determine 63, NO, N02, and NO . Beckman 6800
A
environmental chromatographs were used to monitor CO, THC, CH4, and NMHC. Two
types of sulfur monitors using flame photometric detection were employed. In
13 stations the Tracer sulfur chromatograph was used to determine total sulfur
(TS), S02, and H^S; the Meloy instrument was used to monitor TS in the remain-
ing 12 stations. Integrating nephelometers (MRI) were employed to measure
B . Wind speed and wind direction were monitored by MRI instruments at 10
S C81-
or 30 m, depending on the site. Ambient temperature was monitored with MRI
temperature probes at 5 m at all sites. Dew point was determined by EG & G
thermoelectric dew point hygrometers. At selected sites, vertical temperature
gradients of up to ±5° C were determined using MRI temperature probes located
on towers at altitudes of 5 and 30 m. Barometric pressure and total solar
radiation were also monitored at selected sites.
Validated RAMS data for June 5 through 9 were retrieved on tape from the
EPA data bank as hourly averages. The data were then transferred from tape to
disc files by RTI for subsequent examination and analysis.
The time designations employed by RAMS have been adopted for all hourly
data in this study. Hour designations progress from 0000 to 2300 and are
reported in Central Standard Time (CST). Averages for a particular hour
represent the mean of data taken during that hour, i.e., the 0700 value is the
average of data collected between 0700 and 0759 CST.
3.1.7 Emissions Inventory Data
Emissions inventory data were retrieved from EPA's National Emissions
Data System (NEDS). The retrievals were made in May and June of 1977. The
counties for which data were requested are listed in table 4. Two types of
data were retrieved: raw data reports and county emissions summaries. Raw
data reports were employed to identify and locate the major pollution sources
of SO , NO , THC, and CO. County emissions summaries describe the contribu-
a A
tions of area sources such as automobiles and of point sources to the total
countywide emissions. It should be noted that different practices of report-
ing and updating records by regulatory agencies introduce uncertainties into
these data. For this reason, officials with various State and local agencies
having jurisdiction within the study area were interviewed to obtain current
18
-------
(3
0
• H
4->
03
4-1
d
U
§
i-i
4->
CO
d
•H
CO
Ja
PS
.
CO
0)
i— 1
£>
B)
H
SZT
m
ezT
ZZT
T3T
OZT
6TT
8TT
LU
9TT
STT
m
o en
d 3TT
o TTT
% °TT
03
^ 60T
CO
SOT
£OT
90T
SOT
^OT
eoT
ZOT
TOT
CO
4->
•H
d
£3
4->
>
O> 4-1
M -H
3 4->
W d
03 03
*
4J CO
•H
d
0
se
X
o
z
1
o
55
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CJ
e
ft
ft
o
o
00
vrt
>w
d
J
u
o>
m
u
e
1
•*
w
u
1
o
u
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
e s
ft ft
ft ft
•< m
W co
O rH
f""**
>
u o
03 rH
M 0)
H S
CO
IN
w
IN CO
O H
CO
CO
H
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
1— 1
1
s
1— 1
vrt
^u
in
rH
n
K
s
00
d
•H *•
M 4->
CU
4-> -rl
a) U
O -H
co m
<+-!
4-1 1>
ft O
oou
•H
r^
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
rH
1
CO
s
CO
fxl
tN
CM
o
I— 1
M
^H
T3
cu
4-) CU
(0 ft
U CO
CO
39 TJ
d
•H
2e
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CO
cu
cu
M
00
(U
Q
Q
fNJ
V^I
CM
o
1— 1
rH
K
s
d
o
•H
4-»
u
0)
M
•H
Q
"O
d
•H
»
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
CJ O
O 0
o
co
oo
CM
1 flj
1 VU
O 00
^* T3
co •!-!
M
M ,0
K a
S 03
u
CU
M 4-1
3 d
4-1 -H
03 O
M Pn
CU
i* *
S cu
0) Q
H
X
X
X
X
X
X
X
X
X
X
X
X
CJ
o
CM
<£
l~H
o
^a-
oo
1— 1
OA
4->
d
OJ
•H
•0
03 v— /
03 p-i OK
CQ H
19
-------
Table 4. Counties for which emissions inventory data were retrieved
Missouri
Illinois
Indiana
Lincoln*
Warren*
Franklin*
Jersey*
Macoupin*
Madison*
St. Clair*
Monroe*
Montgomery
Bond
Clinton
Washington
Sullivan
Knox
Gibson
Posey
St. Charles*
St. Louis*
Jefferson*
Shelby
Fayette
Marion
Jefferson
Effinghans
Clay
Wayne
Cumberland
Jasper
Richland
Edwards
White
Clark
Crawford
Lawrence
Wabash
*Counties included in the greater St. Louis area.
information on major sources. In addition, the detailed emissions inventory
established in support of RAPS was used to supplement NEDS data. The RAPS
emissions inventory can resolve emissions data within-the study area to time
increments as fine as 1 h. The RAPS data bank was incomplete at the time of
the current study and validated data were not available. Preliminary retriev-
als of annual emissions rates were therefore used in the present study; the
counties for which RAPS emissions inventories were used are noted in table 4.
3.2 Flight Description
On 8 June 1976, at 0756 CST, the Da Vinci II system was launched from
Arrowhead Airport, 24 km west of St. Louis, Missouri. During the next 24 h
Da Vinci II drifted south, then turned north, and finally east across the
southern Illinois plains before landing in a wheatfield in southwestern Indi-
ana (see figure 4). Figure 4 shows the flight track of Da Vinci II for the
first 12 h in and about the St. Louis area, the RTI-EML ground track, and
20
-------
Figure 5. Flight track of Da Vinci II (above) and ground track of the RTI-EML (below) from
launch at Arrowhead Airport to touchdown in Indiana.
21
-------
major pollution sources in the area. Figure 5 shows the entire flight track
of Da Vinci II and the ground track of the RTI-EML from launch at Arrowhead
Airport to touchdown in Indiana. The flight ended at 0805 CST on 9 June 1976.
During the 24-h flight, the RTI-EML was used to obtain measurements of
ozone, nitrogen oxides, sulfur dioxides, total hydrocarbons, methane, and
carbon monoxide at ground level approximately along and underneath Da Vinci
II's track (see figure 5). With the exception of a 2-h period of time early
in the morning («0300 to 0500 CST) on 9 June 1976, the RTI-EML was in visual
and radio contact with Da Vinci II. At the conclusion of the flight (0805 CST,
9 June 1976), the RTI-EML was within 0.8 km of the landing site. The RTI-EML
was then brought adjacent to Da Vinci II for postcalibration of the Dasibi
ozone monitor that was flown aboard the balloon (figure 6).
Da Vinci II spent the first 12 h of the flight drifting in the vicinity
of St. Louis. (See figure 4.) After launch at 0756 CST, Da Vinci II drifted
to the southeast crossing U.S. Highway 40 at 0830 and Interstate 244 at 0838
CST. At 0940 it crossed 1-44 passing within 5 km of a Chrysler assembly
plant. At approximately 1027 CST Da Vinci II turned and began moving to the
northwest recrossing 1-44 at 1140 and U.S. 40 at 1215. It then moved directly
Figure 6. RTI-EML located adjacent to gondola for post-flight
calibration of Da Vinci II analyzers.
22
-------
north for the next hour and 20 minutes. Moving to the northwest during the
period between 1336 and 1356 CST, Da Vinci II passed over 1-70 and 1-270.
This location is within 4 km of Lambert Field (St. Louis International Airport)
and a Ford assembly plant. The direction shifted to slightly east of north
and Da Vinci II crossed the Missouri River for the first time at 1410 CST. At
1500 the direction began to shift more toward the east. The airspeed was
reduced significantly, and from 1615 to 1830 CST the craft covered only about
6 km. During this period Da Vinci II passed within about 3 km of the Union
Electric Sioux Power Plant and crossed the Missouri River and Pelican Island.
At 1830 the airspeed increased with the craft moving directly to the east. Da
Vinci II crossed the Mississippi River at 1910 and had passed over the south-
eastern tip of Alton, Illinois, by 1945 CST. This location is within 1 km of
the Illinois Power Wood River Power Plant in East Alton. Between 1945 and
2015 CST it traveled 3 to 5 km north of a heavily industrialized area. Three
major oil refineries are located in Wood River and Roxana. Drifting to the
east, Da Vinci II moved outside the St. Louis study area at approximately
2140 CST when it crossed 1-55 near Hamel at the intersection with Illinois
140.
The entire flight track of Da Vinci II from launch to touchdown is illus-
trated in figure 5. The portion of the flight track traveled by Da Vinci II
after it left the greater St. Louis area until touchdown is referred to as the
"downwind flight track." Da Vinci II moved eastward from St. Louis on the
evening of 8 June crossing 1-55 at 2140 CST, and passed 8 km north of Green-
ville and 14 km south of Coffeen at just after midnight. An 870-MW coal-fired
power plant located at Coffeen was the only major point source in close prox-
imity to the downwind flight track. Da Vinci II crossed 1-70 near the inter-
section with Illinois 140 at 0100 on 9 June. The craft continued its eastward
path and passed 10 km to the south of Vandalia at 0200. Between 0300 and
0330 CST it crossed 1-57, and its direction began to shift toward the south-
east. From this point until touchdown Da Vinci II was passing over oil fields.
During the interval between 0505 and 0635 CST, the flight track crossed U.S.
50 and moved toward the south. Da Vinci II maintained this direction as it
drifted over the Wabash River into Indiana, crossed 1-64, and touched down in
a wheat field. The craft landed at 0805 CST on the morning of 9 June after
having flown approximately 340 km in 24 h.
The altitude-time profile presented in figure 7 shows the vertical move-
23
-------
CO
-** —* •*
«**• IDf*
sf
-^•»'-
OJ
fr2
.
y
'
-°.»
,.*'j"-,'
CM
vO
CO
0
UJ
oo
OJ
CO
-------
Table 5. Major altitude excursions of Da Vinci II
No.
la
2a
3a
4a
5a
6a
7a
8a
9a
lOa
lla
12a
13a
14
15a
I6a
17a
I8a
19
20a
Time
CST
0756
0937
1025
1056
1110
1247
1306
1355
1543
1625
1719
1820
1941
-
2318
0100
0151
0509
-
0710
Minimum
altitude
feet, MSL
450
1483
1473
1550
4370
3677
4650
5177
1850
1550
1350
1650
2150
-
2550
2250
2650
2550
-
1763
No.
Ib
2b
3b
4b
5b
6b
7b
8b
9b
lOb
lib
12b
13b
I4b
-
I6b
-
18b
19b
~
Time
CST
0813
0956
1047
1105
1130
1254
1343
1409
1603
1651
1741
1901
2009
2304
-
0123
-
0538
0637
—
Maximum
altitude
feet, MSL
2150
3350
2150
4850
6350
5450
6450
7093
4350
3050
1950
2950
2550
3050
-
3150
-
2950
2950
"•
ment of Da Vinci II during the flight. The numbered altitude excursions are
identified in table 5. Shortly after launch, Da Vinci II achieved an altitude
of approximately 579 m (1,900 ft) MSL and maintained this altitude until it
began to experience thermal turbulence. The craft descended to 452 m (1,483
ft) at 0937 CST, rose to 1,021 m (3,350 ft) at 0956 CST and descended to 449 m
(1,473 ft) at 1025 CST.
It was at this time that the balloon turned from a southeastward to a
northwestward horizontal direction. The balloon continued to experience
vertical perturbations, was allowed to rise at 1056 CST to conserve helium,
and at 1130 CST achieved an altitude of 1,935 m (6,350 ft). Most of the north-
ward leg of the flight track was traversed at altitudes between 1,219 m and
2,133 m (4,000 and 7,000 ft). Shortly after crossing the Missouri River
during the interval from 1425 to 1543 CST, the craft experienced a sharp
descent from 2,149 m (7,050 ft) to 564 m (1,850 ft). As the balloon was
moving slowly eastward across the Missouri and Mississippi Rivers, four less
25
-------
severe altitude excursions occurred at 1603, 1651, 1741, and 1901 CST. From
1941 until the final descent for landing at 0637 on 9 June, altitude was
relatively stable between 655 m and 930 m (2,150 and 3,050 ft).
3•3 Meteorological Conditions for the Period Encompassing Flight Day
The general meteorological conditions were examined for 6-9 June 1976.
This 4-day period encompasses the flight day. The conditions that were exam-
ined include synoptic-scale air mass behavior, wind speed, wind direction,
insolation, and temperature. These factors exert a significant influence on
the extent to which pollutants may interact in the atmosphere. Physical
factors determine the extent of both vertical mixing and horizontal transport.
Atmospheric mixing dictates the extent to which pollutants such as hydrocarbons
and nitrogen oxides accumulate and therefore determines their ambient concen-
tration levels. Insolation and temperature are influential in determining the
progression of photochemical reactions such as those that result in the forma-
tion of ozone.
The study area came under the influence of a high pressure system during
the first week of June. This system followed an unusual track moving south-
westward from Ontario on 2 June toward the mid-Mississippi Valley where it
stagnated and gave rise to high levels of air pollution. As the center of the
high pressure system moved to the south on 3-6 June, winds from the east
brought polluted air into the midwest. As the system continued to move to the
south, return flow caused a shift in wind direction. By the morning of 8 June,
the surface winds had shifted 180° and were from the west. The air mass then
began to progress slowly to the east. By 12 June, this system had been pushed
off the east coast by snor-her air mass.
Mean profiles of selected meteorological parameters monitored in St. Louis
by RAMS describe nourly behavior of these parameters averaged across the RAMS
network. These profiles are presented in figure 8 for the 4-day period,
6-9 June,
The 0700 CST surface weather maps prepared by the National Weather Service
were examined ;."o.- 6-9 June, A large high pressure system was centered over
Wisconsin on 6 Juue a^d covered most of the midwestern and eastern States. By
7 June the high pressure system was centered over the St. Louis area and
remained there for the flight day, 8 June. The system began to weaken on
9 June, although it continued to dominate the weather map. The behavior of
26
-------
WIND DIRECTION, DEGREES
03S/W 'Q33dS ONIM
'3yniyy3dW3i
I
-9
9
CD
a>
CO
i
3
in
o
I—
Ol
- o
cc
ac
'85
ac
. ct
. oo
o
00
o
o
V)
<
CC
00
|
O)
OH 'Ni '3ynss3Md
3inNiw/sx3ioNui 'NOiiuiauy yyios 10101
27
-------
the high pressure system during this period is reflected by the RAMS network
mean barometric pressure profile, figure 8a.
Near the center of the high pressure system, calm conditions and clear
skies were recorded at many reporting stations. These conditions prevailed in
the St. Louis area. Mean hourly surface wind speeds from the RAMS network as
indicated in figure 8b were generally less than 3ms . Mean RAMS solar
radiation data shown in figure 8c also indicate clear skies. The stagnant
conditions associated with this type of weather system allow the accumulation
of both primary and secondary air pollutants. The regional scale problem of
elevated summertime ambient ozone levels is associated with atmospheric condi-
tions of this type. In addition, calm winds and clear skies provide favorable
conditions for the development of strong nocturnal radiation inversions.
Although the winds were generally light and disorganized, the RAMS mean
wind direction profile shown in figure 8b indicates distinct directional
changes in surface winds over the 6-9 June period. On 6 June winds were
generally from the east; however, they began to shift to a more southerly
component shortly after sundown. On 7 June, winds were basically from the
south, shifting from southwesterly to southeasterly and back. On flight day,
mean winds were from the west: for the first 5 or 6 h they were from slightly
south of west; during the 0700 CST hour they had begun shifting to a northwest-
erly component; before noon they had shifted from a northwesterly to a more
southerly component; by sunset the winds had once more shifted and were from
the west. Although it may be somewhat fortuitous, these RAMS surface wind
patterns were qualitatively reflected by the daytime movement of Da Vinci II.
On the morning of 9 June the winds were from the southwest.
A slight warming trend was evident from 6 to 9 June. The mean temperature
profile from the RAMS network for this period is presented in figure 8d. The
maximum and minimum ambient temperatures on flight day were 30.2° C (86° F)
and 13.8° C (57° F).
3.4 Quality Control Program
To achieve and maintain a high level of confidence in air quality data,
it is essential to have a well-designed, operational quality control program.
Quality control for air quality measurements on board the Da Vinci II gontfpla
and RTI-EML included (1) the use of reference or equivalent analyzers, whenever
possible, reference or accepted calibration techniques, NBS-SRMs as primary
standards for calibration, and standard operation, maintenance, and calibration
28
-------
procedures; (2) reference of calibration systems and personnel to systems and
standards maintained by the Quality Assurance Branch, Environmental Monitoring
and Support Laboratory, Environmental Protection Agency, Research Triangle
Park, North Carolina; (3) performance of dynamic calibration of air quality
analyzers prior to and after the flight of Da Vinci II; (4) maintenance of
adequate records that describe instrument performance; (5) design and checkout
of intake manifold of RTI-EML for mobile sampling; and (6) system comparison
(RTI-EML) to stationary monitor (i.e., drive-by comparison) to assess compara-
bility of mobile measurements to ground station measurements. In addition,
logs and records were maintained during the flight that documented the position
of the RTI-EML relative to the Da Vinci II gondola, identified times when the
RTI-EML was stationary and generator exhaust might influence ambient measure-
ments, and described the Da Vinci II flight and RTI-EML tracks. Data were ex-
amined and validated immediately after the flight using accepted techniques.
Primary standards used for calibration purposes were verified prior to and
immediately after the Da Vinci II flight.
3.5 Analysis Approach
The atmosphere is the active medium where all chemical processes occur.
The investigation of the role of the atmospheric processes upon the air chemi-
cal processes, as revealed through measurements, is an essential part of the
total analysis effort. If the atmosphere were purely passive, laboratory
models would be sufficient to explain the measurements.
The analyses of both the chemical and the atmospheric processes for Da
Vinci II consider data taken from two frameworks. Ground station chemical and
meteorological and upper air data were taken from a Eulerian system; i.e., the
observer is fixed in space and observes changes as the medium moves past,
while Da Vinci II measurements (aboard the balloon) were taken in a nearly
Lagrangian system; i.e., the observer is moving (approximately) with the
medium and measures the changes that occur from processes within the flow.
Meteorological characteristics of the balloon environment are related to the
chemical and oxidant concentrations measured aboard Da Vinci II and to measure-
ments from other platforms—aircraft, RTI-EML, and ground stations.
Information was assembled from many sources to provide a comprehensive
description of the atmospheric and pollutant behavior during the study period
both aloft and at ground level. Aircraft data, when available, were used to
supplement information from the balloon to describe processes that were occur-
29
-------
ring aloft. Data from the Regional Air Monitoring Study (RAMS) network, the
RTI-EML, and Indiana, Illinois, and Kentucky State pollution monitoring sta-
tions were employed to describe pollutant behavior at the ground. Emissions
inventory data were also examined to locate major pollutant sources within the
study area.
Independent analyses of the data were performed to provide chemical and
meteorological interpretations for Da Vinci II. These analyses are discussed
in the following sections.
30
-------
4.0 ATMOSPHERIC CHEMISTRY ANALYSIS FOR DA VINCI II
4.1 Objectives of Analysis
The purpose of this analysis of Da Vinci II data was to examine the
atmospheric chemistry processes that occurred and to investigate relationships
between primary and secondary air pollutants upwind, within, and downwind of a
city. The behavior of ozone and its precursors was a major concern. Specific
objectives were to describe and examine the following:
1. The levels of ozone and ozone precursors that occurred upwind of the
urban area;
2. The impact of urban sources of ozone precursors on the ambient con-
centrations of both ozone and its precursors;
3. The impact of precursor and ozone transport on urban ozone concen-
tration levels;
4. The transformations of ozone and precursors during transport;
5. Differences between daytime and nighttime pollutant behavior;
6. The scale or extent of urban transport; and
7. The feasibility of balloon-borne experiments for obtaining the data
needed to address these points.
4.2 Approach
Information from many sources was assembled to provide a comprehensive
description of the atmospheric and pollutant behavior during the study period
both aloft and at ground level. Aircraft data, where available, were used to
supplement information from Da Vinci II to describe the processes that were
occurring aloft. Data from the RAMS network and the RTI-EML were employed to
describe pollutant behavior at the ground. Emissions inventory data were also
examined to locate major pollutant sources within the study area.
4.3 Results and Discussion
4.3.1 Emissions Inventory
Before an assessment of atmospheric pollutant behavior was conducted, the
major sources in the study area were identified. Major point sources of SCL,
NO , HC, and CO were ranked and are listed in tables 6-9. In these tables the
sources within the St. Louis area covered by the first 12 h of the flight are
segregated from the sources within the area covered by the last 12 h of the
flight (downwind flight track). The positions of the major point sources and
highways are located along the flight track of Da Vinci II as illustrated in
figure 4.
31
------- An error occurred while trying to OCR this image.
------- An error occurred while trying to OCR this image.
------- An error occurred while trying to OCR this image.
-------
M
M
O
C
•rt
£>
CO
Q
M-l
O
03
r-l
4-J
4-»
GO
-H
^
r<
"5
^
%
o
rH
cd
W
QJ
f,
M
g
CO
4-1
C
•H
O
ex
o
^
>-i
0
'c?
S
CTi
QJ
, — j
•3
H
V
0
*"
M
0
CD
g
H
*
01
<3
«
•rt
00
i
Cd
01
^i
rH
•rt
0
O
u
•*•*,
•J*
C rH
PS
"cJ|
O
vO 1
° !
•O- 1
§o
oo
r» vo
m
oo co
•rt Q>
C U
•rt O
KH IH
rH
I 5
01 rl
o co
Vi 9
Q) Ci
PH H
§ g
CO CO
•rt *rt
T3 TJ
.5 5
•O O
M 4-1
0 rH
KH 4-1
rH CO
O 01
•o
& 01
cd y
rH Cd
CJ >-4
CM en
"Jj|
I
j
I
o
in
r*»
rH
CO
CO
01
u
o
£
rH
•rt
CO
9
•o
6
M
O
CO
•rt
•o
rH
01
0)
4J
CO
4-1
•rt
CJ
01
4J
•rt
G
cd
O
•*
"Jjj
Q
t en
1
O O
O VO
rH O
,rH rH
C
0) O
CO -rt
01 4J
u cd
O rl
£ S
CU
rH O
cd
•rt 0
CO 4J
9 U
•O 0)
O rH
M H
IH CO
•rt -rt
cd 9
rH 0
0 r3
* *
4J 4J
CO CO
CO
•rt CJ
9 CU
c? S
* ^
CO
4-1 U
CO iH
fd rl
u
. 01
0 rH
^j Pd
§ c
0] O
C iH
£ fi
in vo
o o
o o\
in en
CM
O O
m oo
o rv
rH
C C
4J 4J
cd cd
rl rl
01 CU
g £
o c5
u u
rl rl
4J 4J
U U
01 CD
rH rH
M M
rH
rt &
CO
0)
rH
fa
§cd
,c
to u
•rt
•a
Cd 4J
S co
c
0
4-1
i—t
^
X
CO O
Cd iH
Cd CO
A «
VJ U
01 -H
g a
P-. U
CU
CO rH
C C
•rt O
rH -rt
rH a
r~ oo
O
oo
00
CM
^
c^ 1
1
1
1
rH
01
CO
O
P.
CO
•rt
Q
0)
4-1
CO
•o
•rt
H
O
CO
s
CO
•rt
1
•
4-1
co
o
4-1
cd
cu
.g
g
H ^B
u
rH Cd
Cd rl
•H _£!
C 00
9 iH
S rH
Fx
•O
•rt
g
cr> y
(3
"ol
i
i
i
o
0
vO
rH
rH
00
•rt
•fj
<4H
CU
PA
1
rH
O
4J
01
•0
rS
0)
co
9
cd
cu
rl
S
e
cd
i
•rt
TJ
e
M
rH
•
CO
S
§
4J
CO
cd
&
§
•rt
co
CO
rH
•rt
4-1
cd
2
0)
£
i
•4-1
•o
cu
>
0)
•rt
t-l
4J
01
rl
§
•rt
4-1
o
pj
•rt
C
O
•8
CO
M
CU
•a
rl
O
60
CO
cd
01
rl
u
0)
TJ
C
•rt
•O
0)
Ai
C
"«9|
b^
fa
o
4J
§
a
M
CO
a
o
•rt
00
co
g
w
CO
2
PH
e
o
rl
>4H
-------
Within the St. Louis area, electric power plants and petroleum refineries
are the major point sources of SO- and NO (see tables 6 and 7). Petroleum
£* A
refineries and automobile assembly plants (surface coating) are the major
point sources of HC (see table 8). Petroleum refineries are the primary point
sources of CO (see table 9).
The electric power plant at Coffeen, Illinois, is the major point source
associated with the downwind flight track. Da Vinci II was in close proximity
to this source at midnight when it passed to within 14 km of Coffeen. Da
Vinci II passed over the eastern edge of Gibson County just prior to its land-
ing in Posey County, Indiana. Although a power plant is located in Gibson
County and a petroleum refinery is in Posey County, the effects of these
sources were probably not monitored on Da Vinci II because the instruments
were shut down in preparation for landing prior to passage over this area.
Total emissions are summarized in table 10 for the greater St. Louis and
downwind areas. Point sources such as those listed in table 6 are the major
sources of S02, contributing 98 and 96 percent of the totals for the two study
areas.
Automobiles are the major sources of HC and CO, comprising 60 and 77
percent of the totals for the St. Louis area. Although it is not reflected in
table 10, the petrochemical industry in Madison County, Illinois (Wood River),
is the dominant source of HC and CO, comprising 67 and 62 percent of Madison
County's HC and CO emissions. Automobiles are the major HC and CO sources in
the downwind flight track area and contribute 73 and 88 percent of the total.
In the two study areas, power plant NO emissions exceed vehicular emis-
sions by 65 to 110 percent. Power plants contribute 52 percent of the NO
X
emissions in the St. Louis area and 61 percent downwind. In contrast, vehicu-
lar sources contribute only 31 and 28 percent in these areas.
The point of release of air pollutants is important in determining air
quality within a study area. This may be particularly significant at night
when a radiation inversion traps emissions near the ground and insulates the
ground from pollutants released through stacks into the air above the radia-
tion inversion.
Various ratios of pollutants may be calculated for the two regimes based
on emissions data. The S02/N0 mass ratio calculated across all power plants
is 4.5. This corresponds to a molar concentration ratio of 3.3. The HC/NO
and CO/NO mass ratios calculated from St. Louis area vehicular emissions are
A
36
-------
Table 10. Emissions summaries for counties along the flight track of Da Vinci II
Tons Per Year Of Emissions
a/
Greater St. Louis Area—
HC N0y CO S02
Total 192,209 232,056 771,857 823,084
Point 65,125 (34)-/ 140,266 (60) 156,839 (20) 802,588 (98)
Power Plant 992 (<1) 120,691 (52) 3,496 (<1) 662,970 (81)
Area 127,084 (66) 91,790 (40) 615,018 (80) 20,496 (2)
Auto 114,921 (60) 72,325 (31) 592,075 (77) 5,593 (<1)
c/
Downwind Flight Track—
Total 65,582 115,971 304,581 238,647
Point 9,058 (14) 78,258 (67) 13,499 (4) 228,359 (96)
Power Plant 594 (<1) 70,815 (61) 1,573 (<1) 203,602 (85)
Area 56,524 (86) 37,713 (33) 291,082 (96) 10,288 (4)
Auto 47,790 (73) 32,349 (28) 267,272 (88) 1,457 (<1)
a/
— The following counties were included: Lincoln, Warren, Franklin, St. Charles,
St. Louis, and Jefferson (Missouri); Jersey, Macoupin, Madison, St. Clair, and
Monroe (Illinois).
— Numbers enclosed with parentheses represent the contribution to the total in
percent.
c/
— The following counties were included: Montgomery, Bond, Clinton, Washington,
Shelby, Fayette, Marion, Jefferson, Effingham, Clay, Wayne, Cumberland,
Jasper, Richland, Edwards, White, Clark, Crawford, Lawrence, and Wabash
(Illinois); Sullivan, Knox, Gibson, and Posey (Indiana).
1.6 and 8.2. These values correspond to molar ratios of 4.3 and 13.4. The
overall HC/NO and CO/NO mass ratios based on emissions from all sources in
X X
the St. Louis area are 0.8 and 3.3. The discrepancies between these overall
ratios and those for vehicular sources is contributed primarily by NO emis-
X
sions from power plants.
4.3.2 Air Contaminant Considerations for the Period Encompassing
Flight Day
Air parcel trajectories performed by Bujnoch of the Institute of Storp
37
-------
Research suggest that Da Vinci II was launched into relatively dirty air that
had been in eastern Kentucky and the Ohio Valley 72 h earlier. This is gen-
erally consistent with the southwestward movement of the high pressure system
noted in section 3.3.
Maps of ozone distribution for the northeastern quadrant of the United
4
States are presented in figure 9 through the courtesy of Wolff. These dia-
grams are based on daily maximum ozone concentrations recorded at approximately
81 locations within the area. On 6 June high ozone and high pressure dominated
much of the midwest. Regions of high ozone emanating from St. Louis and other
areas were evident on 7 and 8 June. As the high pressure system began to weak-
en in St. Louis on 9 June, the region of high ozone moved to the east covering
a portion of middle Ohio and much of the east coast. Thus, for the period from
6 to 9 June, much of the country east of the Mississippi River was experiencing
high levels of ozone.
A polluted condition also developed in the St. Louis area during this time
period. Mean profiles of selected chemical parameters monitored by the RAMS
network are presented in figure 10. These illustrations depict hourly pollut-
ant behavior averaged across the RAMS network for this 4-day period. A dete-
rioration of air quality is illustrated in the 0,, and B . profiles shown in
figures lOa and b. The daily maximum 0_ levels increased from 6 June through
8 June but were reduced on 9 June. Ozone levels exceeded 0.08 ppm on each
day. In general, B also increased during the period. Nonmethane hydro-
SCcl L
carbon and NO profiles presented in figures lOc and d also show a slight
tendency toward increased levels from 6 through 8 June with a small reduction
on 9 June. Nitrogen oxide levels ranged between 0.005 and 0.11 ppm. Nonmeth-
ane hydrocarbon concentrations ranged between 0.00 and 1.00 ppmC and also ex-
ceeded 0.24 ppmC on each day.
The CO profile shown in figure lOe behaves similarly to the NMHC and NO
profiles. In addition, maximum NMHC, NO , and CO concentrations occurred at
X
night. Frequently, two peaks occurred--the first shortly before midnight and
the second in the early morning hours. Absolute maximum NMHC and NO levels
A
may occur before midnight rather than in the early morning between 0600 and
0900 CST. In the absence of mixing, CO, NO , THC, and CH,, having ground-
A T"
based sources, are expected to accumulate beneath the nocturnal radiation
inversion. An injection of cleaner air such as that from aloft would bp mani-
fest on a concentration profile as a negative inflection surrounded by two
38
-------
ULJ
2
CO
O
oc
O
u.
>•
UJ
Ills
s ii 3 $
o S
8 S ^ *
d o d d
Oi
UJ
z
D
UJ
Z
UJ
.Z
39
-------
O CO U3
C\J ^ —i
CM O
O O
w/i '
Wdd
-------
o
CO
o
r»
o
lO
O
U1
O
CO
o o
OJ »H
UJ
z
vO
r-
C7)
UJ
z
x
o
UJ
z
I ! I
O
CO
i
£
§>
'yruins
Wdd '03
41
-------
positive peaks. This "dual peak" behavior observed in figure lOc through e
may be due to increased vertical turbulence mixing air from aloft to the ground
at around midnight.
The NMHC/NOx ratio (R) was calculated for the study period. The R profile
shown in figure lOf exhibits a 4-day mean value of 7.5. Automobiles are major
ground-based sources of HC and NO emissions. The mean R value is 75 percent
higher than the value calculated earlier from vehicular emissions estimates.
Sulfur dioxide concentrations measured as total sulfur (TS) were generally
below 0.03 ppm. The TS profile is displayed in figure lOg. Sulfur dioxide is
primarily emitted aloft from power plant stacks. This may contribute to the
erratic behavior of the profile. The RAMS stations located at ground level
were frequently insulated from any SOg-laden air aloft by a nocturnal radia-
tion inversion. This insured low nighttime SC>2 levels. During the day, when
vertical mixing was enhanced, the opportunity existed for power plant plumes
to become distributed through the mixed layer. Thus, as observed, the daytime
ground-level S02 profile could exceed the nighttime concentration.
4.3.3 Air Quality at the Launch Site
The hourly network mean data from RAMS have provided a general descrip-
tion of air quality behavior in the St. Louis area. The RTI-EML collected
air quality data at the launch site, Arrowhead Airport, for 16 days prior to
launch. This site is located approximately 24 km west of downtown St. Louis,
7 km west of the nearest RAMS station, number 120, and 18 km east of RAMS sta-
tion 125.
Much of the RTI-EML data is summarized in table 11. The ozone data are
particularly interesting. Over the 16-day period (384 h), ozone levels ex-
ceeded 0.08 ppm for 59 hours or 15.5 percent of the time. Ozone levels ex-
ceeding 0.08 ppm occurred daily from 1 June to launch. These high-ozone con-
ditions prevailed for 8, 3, 5, 1, 3, 10, and 11 h for the first 7 days of June.
The hourly average ozone profile at Arrowhead for the period 1 June until
launch is presented in figure 11. This profile is in good agreement with data
from RAMS stations 120 and 125. "In addition to exceeding 0.08 ppm, ozone lev-
els exhibited elevated minimum concentrations from 2 to 4 June. During this
period, winds were from the west and north indicating air flow from rural and
suburban areas. On 6 and 7 June, winds were from the east to the south with
flow from St. Louis. Ozone concentrations that exceeded 0.08 ppm prevailed
for 10 h on 6 June and 11 hours on 7 June during the day, and near-zero mini-
42
-------
in
CM
in
CM
o
CM
o
CM
in
«—I
•
o
o
v-4
•
o
in
o
in
o
o o
Wdd '3NOZO
o
o
2s
o
cc
o
43
-------
Table 11. Summary of air quality parameters at Arrowhead Airport for the
16-day period prior to launch
Parameter
03
NO
NO
X
S02
THC
CH4
NMHC
CO
Range of Hourly Average
Concentrations, ppm
0*
0
0
0
1.43
1.37
0
0.15
- 0.149
- 0.057
- 0.106
- 0.084
- 4.16
- 3.44
- 0.58
- 1.880
Mean, ppm
0.050
0.004
0.016
0.020
1.825
1.694
0.132
0.535
*A reported zero corresponds to the MDC for the instrument employed.
mum values occurred at night. On the afternoon of 7 June, the winds shifted
to a more westerly component, and increased minimum nighttime ozone levels were
measured at Arrowhead and RAMS station 125. Winds were from the west on the
morning of the launch.
Urban and nonurban environments exhibit characteristically different diur-
nal ozone profiles. The amplitude is greater at urban sites than at nonurban
sites. Urban stations exhibit elevated maximum and low minimum ozone concen-
trations. The near-zero nighttime ozone minima are attributed to scavenging
by anthropogenic ozone-destructive agents such as NO and HC, which are emitted
by automobiles. The 6 and 7 June profiles typify this behavior. Ozone pro-
files at nonurban sites exhibit a reduced amplitude, reduced daytime maxima,
and increased nighttime minima. These increased minima presumably result from
a relative deficiency of ozone-destructive agents in nonurban air. The 2 to
4 June profiles are typical of this behavior. Westerly winds and the elevated
nighttime ozone minimum on 7 June suggest that Da Vinci II was launched into
air that was coming from nonurban environs.
4.3.4 Air Contaminant Behavior at Ground Level During The Flight
4.3.4.1 Air Quality at RAMS Stations
4.3.4.1.1 Mean Profiles
Mean profiles of selected parameters averaged across the RAMS network for
44
-------
the 4-day period, 6-9 June, were presented earlier in figures 8 and 10. Al-
though attention can be focused on flight day, 8 June, these illustrations pro-
vide resolution of temporal behavior only.
To achieve spatial resolution, RAMS stations were stratified into three
groups according to distance from the central urban area (CUA). Stations 101
through 113 comprise the "inner" group and are located in the St. Louis urban
area. Stations 114 through 121 make up the "middle" group and are largely
suburban sites. Stations 122 through 125 comprise the "outer" group and rep-
resent nonurban sites. A profile for each group of stations was then drawn
on the same figure, and figures were drawn for selected parameters. Such
illustrations allow an assessment of both temporal and spatial behavior of
selected parameters within the RAPS area.
Figure 12a illustrates the characteristic shapes of urban and nonurban
ozone profiles discussed earlier. The nonurban ozone profile has a reduced
amplitude, with a higher minimum and a lower maximum than the urban profile.
The morning increase and afternoon decline of the nonurban profile are the
slowest of the three profiles. Although the suburban ozone maximum occurred
3 h after the urban maximum, both concentrations were of similar magnitude.
The suburban ozone minima fall in between those for the urban and nonurban
sites.
The profiles of CO, CH4, NMHC, and NO in figures 12b through 12e are
similar in shape. The profiles are characteristic of pollutants monitored at
the ground that are emitted by ground-based sources. These pollutants accumu-
late beneath nocturnal radiation inversions both in the early morning and the
late evening to concentrations roughly proportional to the magnitude of their
emission rates. As daytime solar heating occurs and enhances mixing through a
much larger volume, the concentration of these species declines to lower lev-
els. Although the relative rank ordering of the three profiles is generally
maintained throughout the day, the profiles seem to approach a common level
during the well-mixed portion of the day.
Profiles of the NMHC/NO ratio, R, presented in figure 12f, are more er-
A
ratic than previous species' profiles. The variation increases with distance
from the CUA. This behavior may be associated both with instrument variability
as ambient hydrocarbon concentrations approach the noise range of the in^tru-
ment's performance and with real concentrations in areas somewhat removed from
high emissions density areas. The urban stations are in the closest proximity
45
-------
0.25
0.20
I ^ I ^ I ' I ' I '
12 0 RflMS URBflN. SUBURBflN. 4 NONURBflN OZONE
0.05
0.00
0000 0100
0800 1200 1600
TIME. CST
0.25 3.0
2.5
2.0
0 20
0 15
0 10 o
0.05
0.00
1.5
1.0
0.5
0.0
, 1 1 1 1 1 1 1 r-
12b RflMS URBflN. SUBURBflN. 4 NONURBflN;
-»-r H,
3.0
2.5
2.0
1.5
1.0
0.5
2000 0000 "'0000 0100 0800 1200 1600 200 1 7
— ,' ' * 1 \ —
•*' ' ' ', ' 1 f
.' ', • '' ,* ' ' ', ' * 1 ' \ "~
t=~^*~v^/ii^
*"•"*" -"t T^*rt~7v « ^ \ Tv^-V'-*1-*.-* "
00 0100 0800 1200 1600 2000 "oO
TIME. CST
av
70
60
50
10
30
20
10
0
00
Figure 12. Mean pollutant concentrations at urban (A), suburban (+ ), and nonurban ( o ) sites in
St. Louis on 8 June 1976. RAMS stations 101-113 comprise the urban group. Stations
114-121 make up the suburban category and stations 122-125 constitute the nonurban
group.
46
-------
to sources of HC and NO and exhibit the least variability. The R profiles are
*l
less variable under stable nighttime conditions than during the well-mixed day-
time period. Nighttime urban R values are also slightly higher than those at
suburban and nonurban stations although no trend is apparent for the daytime
period.
Profiles of two other parameters, temperature and temperature gradient,
reveal interesting behavior. The temperature profiles shown in figure 13
illustrate the existence of an urban heat island. At night, urban air was the
warmest, and both urban and suburban air temperature were much greater than
that at nonurban sites. During the day the temperature differences across
these three groups of stations were much reduced.
The temperature gradient profiles in figure 14 are useful in resolving
the timing of the establishment and breakup of the nocturnal radiation inver-
sion. Such inversions usually begin to break up in the urban areas prior to
nonurban areas. Likewise, they are established in nonurban areas prior to ur-
ban areas. On the morning of 8 June the inversion began to break up during the
0600 CST hour. Mixing had occurred through the first 30 m (100 ft) at the ur-
ban sites a full hour before it occurred at the nonurban stations. At the
1800 CST hour in the afternoon, the inversion began to reform at the nonurban
sites. A 1-h delay occurred between the onset at the nonurban and urban sta-
tions. Thus, during the day, well-mixed conditions prevailed through the
lowest 30 ra (100 ft) for approximately 2 h longer at urban sites than at nonur-
ban sites.
4.3.4.1.2 Concentration-Distance Plots
To gain another perspective on the spatial and temporal distribution of
atmospheric contaminants within the RAPS area, concentration data have been
plotted with respect to distance from station 101 in both the west-to-east and
south-to-north directions. Maximum, 0500-0700 CST mean*, and 1100-1500 CST
mean concentration data are presented for CO, NO , and 03 in figures 15, 16,
A
and 17. In addition, NMHC data are presented at the time of the maximum THC
concentrations in figure 18. The figures clearly indicate that anthropogenic
emissions are concentrated in the heart of the urban area.
The substantial difference between the maximum and the 0500-0700 CST mean
concentration data for CO and NO indicate that these compounds do not always
A
*Note that the mean of the 0500, 0600, and 0700 CST hourly average con-
centrations encompasses the period between 6 and 9 a.m., local time.
47
-------
•a*
35.
30.0 -
o
,;25.o -
-------
z:
o_
o_
o
o
6.0
5.0
t.O
3.0
2.0
1.0
0.0
1 I ' I ' I ' I ' I ' I '
15 a WEST-EflST RflMS CO 8 JUNE 1976 -
I
'50 tO 30 20 10 0 10 20 30 HO
WEST KM FROM RflMS 101 EflST
50
6.0
5.0
t.O
3.0
2.0
1.0
0.0
Q_
Q_
O
O
6.0
5.0
t.O
3.0
2.0
1.0
0.0
1 I ' I ' I '~1~^T r F"^ I T--|—r f—i
15b SOUTH-NORTH RflMS CO 8 JUNE 1976 -\
I
'50 tO 30 20 10 0 10 20 30 '
SOUTH KM FROM RflMS 101 NORTH
50
6.0
5.0
t.O
3.0
2.0
1.0
0.0
Figure 15. Carbon monoxide concentration relative to downtown St. Louis along west to east and
south to north directions on 8 June 1976: (A) maximum hourly concentration; ( +)
0500-0700 CST mean concentration; and ( o ) 1100-1500 CST mean concentration.
49
-------
0.
a.
x
o
0.30
0.25
0.20
. 0.15
0.10
0.05
0.00
Q_
Q_
X
O
0.30
0.25
0.20
. 0.15
0.10
0.05
0.00
I6a WEST-EflST RflMS NOX 8 JUNE 1976
50 fO 30
WEST
20 10 0 10 20 30 fO 50
KM FROM RflMS 101 EflST
0.30
0.25
0.20
0.15
0.10
0.05
0.00
1 r I T] ' I ' I ' I T 1^
16 b SOUTH-NORTH RflMS NOX 8 JUNE 1976
50 tO 30
SOUTH
20 10 0 10 20 30
KM FROM RflMS 101 NORTH
50
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Figure 16. Nitrogen oxides concentration relative to downtown St. Louis along west to east and
south to north directions on 8 June 1976: (A) maximum hourly concentration; ( + )
0500-0700 CST mean concentration; and ( o ) 1100-1500 CST mean concentration.
50
-------
0.25
0.20
0.15
0.10
0.05
0.00
I ' I ' I ' I ' I ' I ' I ' I '
17 a WEST-ERST RflMS OZONE 8 JUNE 1976
I i I i"*l i I /V+.-T*. I
* - +
I . I
0.25
0.20
0.15
0.10
0.05
50 10 30 20 10 0 10 20 30 10
WEST KM FROM RflMS 101 EflST
50
0.00
0.25
0.20
t 0.15
a.
0.10
0.05
0.00
I ' I ' I ' I ' I ' I ' I ' I ' I '
I7b SOUTH-NORTH RflMS OZONE 8 JUNE 1976
e--
1,1,1 >'i ,*T .1.
0.25
0.20
0.15
0.10
0.05
0.00
50 10 30 20 10 0 10 20 30 10 50
SOUTH KM FROM RflMS 101 NORTH
Figure 17. Ozone concentration relative to downtown St. Louis along west to east and south to
north directions on 8 June 1976: ( A) maximum hourly concentration; ( + ) 0500-
0700 CST mean concentration; and ( o ) 1100-1500 CST mean concentration.
51
-------
1.0
3.0
o
2:
0.
Q_
2.0
1.0
1 I ' I ' I ' I ' I ' I ' I ' I ' I '
18 a WEST-EflST RflMS NMHC 8 JUNE 1976
1.0
3.0
2.0
1.0
0 pi «r I . I i I . I . I . I . I i I i I i IQ
50 10 30 20 10 0 10 20 30 10 50
WEST
KM FROM RflMS 101
EflST
1.0
a.
O_
o
3.0
2.0
1.0
0.0
1 I ' I ' I ' I ' I ' I ' I ' I ' I '
18b SOUTH-NORTH RflMS NMHC 8 JUNE 1976
3.0
2.0
1.0
50 10 30 20 10 0 10 20 30 10 50
SOUTH KM FROM RflMS 101 NORTH
0.0
Figure 18. Nonmethane hydrocarbon concentration at the time of maximum THC concentration
relative to downtown St. Louis along west to east and south to north directions on
8 June 1976.
52
-------
achieve their daily maximum between 6 and 9 a.m. local time. As noted earlier,
maxima for CO and NO occur in the evening.
The 0500-0700 CST mean CO and NO data exceed the corresponding 1100-1500
A
CST data. This difference may be attributed to temporal variation of both
emission rates and atmospheric stability. Early morning automotive emissions
accumulate in a relatively stable layer of air. Between 1100 and 1500 CST, ve-
hicular emission rates are reduced and atmospheric mixing dilutes these emis-
sions through a large volume. The mean ratio of the 0500-0700 CST CO data to
those for the 1100-1500 CST period is 5.1 ± 2.4 for the inner RAMS sites. The
corresponding value for NO is 9.0 ± 4.2. This suggests that mixing reduced
early morning anthropogenic ozone precursor concentrations in the urban area
by factors of 5 to 9. In spite of this dilution effect, the 1100-1500 CST data
continue to show slightly elevated CO and NO levels in the urban area in com-
A
parison to the less urban areas.
The behavior of the secondary pollutant, ozone, shown in figure 17 is
quite different from that of CO or NO . As noted in previous discussions,
X
early-morning ozone levels are reduced in the urban area by reaction with
ozone-destructive agents such as NO and HC, which accumulate in the relative-
ly stable air.
Figure 17 also provides additional insight toward defining the location
of the area affected most strongly by species that act as ozone percursors or
ozone-destructive agents. The location of sharp inflections in both the morn-
ing and afternoon ozone levels near the center of the city correspond to the
sharp increases in the early morning NO concentrations.
A
The close agreement of the maximum and the 1100-1500 CST Og data results
from the fact that maxima were observed between the 1100 and 1500 CST hours at
18 of 22 RAMS stations.
The significant difference in absolute magnitude between the 0500-0700 CST
and 1100-1500 CST ozone data reflects several factors. Beneath the nocturnal
inversion, ozone levels are reduced to the early morning levels by destructive
processes. Between the 1100 and 1500 CST hours, synthesis and mixing contrib-
ute significantly to the observed increase in ozone levels.
As shown earlier, elevated ozone levels prevailed on a regional scale from
St. Louis to the east coast on 8 June. Ozone levels during the well-mixed part
of the day, therefore, reflect the effects of synthesis superimposed on the
existing high-ozone air that had been mixed to the ground. The relative con-
53
-------
tribution of synthesis increases with increased precursor input. The differ-
ence between urban and nonurban ozone maxima may be considered to be a lower
limit estimate of the net contribution of urban precursors to ozone behavior.
Based on the data in figure 17, a net increase of 0.060 ± 0.015 ppm of 0$ was
associated with urban precursors.
On 8 June, winds were from the west. A more realistic estimate of the
urban influence on ozone behavior may be the difference between the 1100-1500
CST mean ozone level at the westmost site, station 125, and the maximum ozone
level observed within the network. This difference corresponds to 0.112 ppm.
Thus on 8 June, a net increase of between 0.06 and 0.11 ppm of ozone was asso-
ciated with precursors released in the St. Louis urban area.
4.3.4.1.3 Ozone Isopleths
Ozone isopleths, computer-drawn constant concentration lines (see appendix
D), have been prepared for selected hours of 8 June. These isopleths, shown in
figure 19, have been superimposed on maps of the RAPS area. Each illustration
provides hourly resolution of the ground-level spatial distribution of ozone
within the RAPS area. The hourly flight track of Da Vinci II is also shown on
isopleths for the hours between the time of launch (the 0800 CST hour) and the
time when Da Vinci II passed out of the RAPS area (the 2000 CST hour). The
following observations have been drawn from these figures.
1. Before dawn (0436 CST), Os concentrations were generally low (<0.01
ppm) at points in the downtown area and eastward, whereas higher
levels (0.02-0.06 ppm) existed outside the city. This presumably
reflects a relatively high level of 03 in the air mass over this
region with the reduced urban levels due to anthropogenic Os-destruc-
tive agents (NO and HC) emitted locally within the city and extend-
ing eastward by westerly winds.
2. The above situation prevailed from the 0000 to 0600 CST hours. From
the 0600 to 0700 CST hour a noticeable increase (~0.01 ppm) occurred
within the study area.
3. By the 0800 CST hour, several 63 regimes existed in the study area:
areas of elevated Og were separated by areas of reduced 63 content.
Ozone concentrations both outside and within the city were greater
than 0.06 ppm. These regions were separated by areas of reduced 63
content (0.04 to 0.06 ppm). Ozone levels in the less-urban outlying
areas had increased by approximately 0.01 ppm over the 0700 value,
whereas ozone levels in the CUA had increased by 0.02 to 0.04 ppm.
The increase in ozone concentration is due to mixing from aloft and
photochemical synthesis. The greater increase in the CUA is presuma-
bly due to the earlier breakup of the nocturnal radiation inversion
with downwind mixing of 03-rich air and generation by increased lev-
els of ozone precursors (HC and NO ).
A
54
-------
005 0040D3 OQg 001
HOUR -0600 CST
0052
0009
0055
0025
HOUR-0700 CST
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
M
0025
00)5
0018
0011
0013
0055
0035
0034
M
0026
0032
M
0066
0050
0.051
0038
0033
0027
M
0069
0075
0051
124 - 0070
125 - 0038
HOUR-0800 CST
HOUR-0900 CST
006 009 DO
OD oil di2
HOUR-1000 CST
M
00/7
0096
M
0120
0107
0096
0100
0 108
0138
0102
0103
M
0093
0099
0132
0107
0081
Figure 19. Hourly ozone isopleths (computer-generated) for St. Louis RAPS area on 8 June 1976.
Ozone concentrations (ppm) are listed for each RAMS site; missing data are indicated by
M; dashed lines represent county boundaries; heavy line indicates position of Da Vinci II
during indicated hour.
55
-------
HOUR-1100CST
HOUfl-1200 CST
017 018
HOUR-1300 CST
HOUR-1400 CST
M
0180
0192
0168
0.165
0157
0144
012 QI3QI4 OB
012 OBQrt
oi4 ais as
HOUR-1500 CST
HOUR-1600 CST
Figure 19. (con.)
56
-------
014 060)60.17 OB
HOUR-1700 CST
0,13 OK OI5
HOUR-1800 CST
oc QOOM'OB
oo op an 012
008 008 006 008
HOUR-1900 CST
HOUR-2000 CST
0061
0086
HOUR-2100 CST
0.05 OD4 003
HOUR-2300 CST
101 •
102 •
103 •
104 •
105 •
106 •
107 •
108 •
109 •
110 •
111 •
112 •
113 •
114 •
115 •
116 •
117 •
118 •
119 •
120 •
121 •
122 •
123 •
124 •
125 •
0003
0003
0003
0003
0003
0003
0003
0003
0003
0003
0003
0006
M
0004
0024
0003
0031
0063
0003
M
0057
0033
0016
M
OO7 006 O05 004
Figure 19. (con.)
57
-------
4. By the 0900 CSX hour, 63 levels at many of the stations had increased
by 0.01 to 0.02 ppm and exceed 0.08 ppm. The first indications of a
pocket of high ozone became evident in the southeastern quadrant of
the study area; this is generally consistent with the northwesterly
winds that existed during and just prior to this time.
5. During the 1000 CST hour, 03 levels at the outlying (N, W, and S)
stations were approximately 0.09 ppm, and a distinct area of high
ozone (0.11-0.14 ppm) was evident in the southeastern quadrant. An
area of slightly reduced ozone content was apparent in the north
central quadrant. In general, higher 03 concentrations occurred
in the CUA, and lower 0$ concentrations occurred at the less urban
outlying stations. This is the first indication of the formation
of an urban plume for St. Louis and is a complete reversal of the
urban-nonurban ozone gradient that existed between the 0000 and
0700 CST hours.
6. By the 1200 CST hour, the mean 03 level at the outlying (N, W, and S)
stations was approximately 0.12 ppm. The area of high ozone (0.16-
0.18 ppm) had extended to the northern part of the southeastern quad-
rant; this movement was probably due to the change in wind direction
from northwesterly to southwesterly that occurred between the 0900
and 1100 CST hours.
7. For the next 5 h (1300-1700 CST) the mean ozone levels at the outly-
ing stations (N, W, and S) remained nearly constant at 0.12-0.13 ppm.
During this time, however, the area of high 03 behaved in a dynamic
fashion. The area of highest 63 concentration, the 63 "hot spot,"
moved from the southeastern quadrant through the northeastern quad-
rant. This movement was to a large extent a result of the south-
westerly winds that prevailed over this time period. The absolute
level of 03 in the "hot spot" increased from 0.178 at the 1200 CST
hour to 0.221 at the 1600 CST hour: a net change of 0.043 ppm.
This increase in 63 over time in comparison to the relatively con-
stant behavior at the outlying stations is probably a direct result
of synthesis from precursors released in the CUA. The maximum ob-
served concentration of 63 (0.22 ppm) occurred at the 1600 CST hour.
For this hour a substantial ©3 gradient of approximately 0.11 ppm
existed between the westernmost station and the "hot spot."
8. Both ozone buildup and plume movement are evident in these isopleths.
The limited area encompassed by the RAPS in comparison to the scale
of air movement that can occur over a single-day period prevented
detailed definition of both the areal extent and the maximum ozone
concentration for the St. Louis urban plume.
9. By the 1800 CST hour, the mean ozone level at the outlying (N, W,
and S) stations had decreased slightly to 0.11 ppm. High QS levels
(0.13-0.15 ppm) remained in the northeastern quadrant, although in
comparison to values for the previous hour, they had been reduced
to a large extent by transport out of the study area.
10. Three QS regimes had been established by 1900 CST. At outlying
stations (N, W, and S), 63 levels were between 0.08 and 0.10 ppm.
The remnants of the St. Louis plume were evident in the north-
eastern quadrant (0.11-0.12 ppm 03). These two areas, however,
were separated by an area of reduced 63 concentration (0.04-0.08
58
-------
ppm) in the CUA. Compounds that act as precursors in sunlight
also act as Os-destructive agents in the dark. The observed re-
duction of 63 in the CUA is presumably due to reaction with an-
thropogenic NO and HC emissions under conditions of reduced light
intensity.
11. The three 03 regimes persisted in the study area for the next 2 h.
The areal extent of the low-Os air increased. The concentration
of 03 in the CUA had dropped to less than 0.005 ppm by the 2100
CST hour in comparison to values of 0.04 to 0.08 ppm at the outly-
ing stations.
12. By the 2300 CST hour, reduced 03 levels (~0.005 ppm) were prevalent
over most of the study area. Somewhat higher levels (0.02 to 0.06
ppm) occurred at the outlying stations, and the predawn situation
had been reestablished.
4.3.4.2 Air Quality at the RTI-EML
Hourly average ozone data collected by the RTI-EML are presented in fig-
ure 20. In addition, ground-level ozone concentrations for the track of Da
Vinci II were interpolated from hourly RAMS isopleths, such as figure 19, and
are also presented in figure 20. The overall agreement between these two pro-
files is good. The 1100-1500 CST mean ozone concentration from the RTI-EML
of 0.123 ± 0.005 ppm compares closely with the 0.134 ± 0.012 ppm from inter-
polated RAMS data.
The agreement is better between the 0800 and 1700 CST hours than for
other times. The flight track of Da Vinci II for the major portion of the day
was on the periphery of the RAMS network. Between the 0900 and 1400 CST hours
the RTI-EML was near, but not actually in, the area having a high density of
RAMS sites. Interpolated RAMS data for other times are less certain due to the
small number of stations in the vicinity of the flight track. In addition,
because the RTI-EML was traveling on roadways, it sampled air contaminated to
an undefined extent with vehicular exhaust that was rich in (^-destructive
agents. The slight discrepancy between the two profiles for the 1800, 1900,
and 2000 CST hours may be due both to nonrepresentative station location rela-
tive to the route and to possible roadway contamination of air sampled by the
RTI-EML. The RTI-EML moved outside the RAPS area at approximately 2140 CST;
therefore, RAMS data are not included for times after the 2000 CST hour.
It is interesting to note the elevated nighttime ozone concentration
observed on the RTI-EML between the 2100 CST hour on 8 June and the 0100 CST
hour on 9 June. Based on the hourly data of figure 20, the magnitude of the
peak is 0.072 ppm, which represents a net increase of 0.038 ppm. Fifteen-
59
-------
CO
cc,
CO
O
C 0)
8-E
O 3
CO
8
I
•6
d
CM
£
§>
Wdd '3NOZO
60
-------
minute average data, not shown, suggest that the elevated ozone condition
lasted for at least 2 h, from approximately 2315 until 0115 CST.
This phenomenon was investigated by visual inspection of plots of both
RTI-EML and RAMS data for the period 1-7 June 1976. Nighttime ozone peaks
were observed at Arrowhead Airport on most of the first 7 days of June (see
figure 11). At RAMS sites during the same period, nighttime 0^ peaks repre-
senting a net increase of greater than 0.010 ppm occurred in approximately
80 cases. Additional RAMS parameters were examined for associations. Sharp
nighttime ozone peaks were accompanied by equally sharp declines of NO and
A
CO. Although somewhat more obscure, similar associations were observed with
NO, N02, THC, and CH4. Nighttime ozone peaks were weakly associated with in-
creased wind speed and temperature, and with reduced temperature gradient.
Species such as CO, NO , THC, and CH4 that have ground-based sources are
A
expected to accumulate beneath the nocturnal radiation inversion. Air aloft,
insulated from ozone-destructive agents emitted at ground level, is expected
to be higher in ozone concentration than air near the ground. Vertical mixing
would act to reduce the temperature gradient near the ground and could be indi-
* .
cated by increased wind speeds. If this turbulence is sufficient to penetrate
the inversion layer, then increased ozone levels and air temperature may be
accompanied by reduced CO, NO , THC, and CH4 levels. These simultaneous phe-
A
nomena were observed in many cases. This suggests that the nighttime ozone
peaks observed on 1-7 June resulted from vertical transport through the noc-
turnal radiation inversion.
Air quality data collected at several ground stations on the night of
8 June were examined for associations with the nighttime ozone peak observed
at the RTI-EML (see figure 11). Ozone peaks were observed at many of the RAMS
sites. Associations similar to those noted above for CO, NO , THC, CH4, wind
speed, temperature, and temperature gradient were also observed in many cases.
A summary of ozone behavior at many of the ground stations is presented in ta-
ble 12. Occurrences of nighttime ozone peaks were widespread: net increases of
at least 0.010 ppm were observed at 10 RAMS, 4 Kentucky, and 9 Illinois ground
stations. A mean maximum net increase of 0.024 ± 0.012 ppm was observed, and
the condition of elevated ozone lasted for 4 ± 2 h.
Based on the RAMS data in table 12, the strong nighttime Og increases
may occur more frequently at suburban and nonurban sites than at urban loca-
tions. This hypothesis should be examined and explored using portions of the
61
------- An error occurred while trying to OCR this image.
-------
RAMS data set. The behavior is probably governed by the ozone concentration
above the nocturnal radiation inversion, the concentration of both 63 and 63-
destructive agents beneath the inversion, and the meteorological processes
that define the severity and the duration of the mixing. The net result of
the interaction of increased downward mixing with such local situations as an
urban heat island circulation system is unclear.
The data suggest that the nighttime ozone peak observed at the RTI-EML
was due to downward mixing. The implications of this apparently widespread
condition are poorly defined. This phenomenon, however, may provide mechanisms
for depleting ozone aloft during nocturnal transport and for distributing noc-
turnal emissions of ozone precursors aloft. A major question, then, regards
the impact that nighttime injections of ozone precursors can have on the ozone-
producing potential within the daytime mixed layer.
4.3.5 Air Contaminant Behavior Aloft During the Flight
4.3.5.1 NOAA Aircraft
Data collected on the NOAA aircraft can be used to describe events that
occurred aloft. Two vertical flights were conducted within 2 km of Da Vinci
II: the first from 0915 until 0933 CST and the second between 1148 and 1211
CST. Ozone, B . , relative humidity, and N02 concentration profiles from
S C3 u
these flights are presented in figures 21 through 24. For comparison, time-
resolved data from the nearest RAMS station (hourly average), the RTI-EML
(15-min average), and Da Vinci II (2-s average) are included, as available,
on these graphs.
The discrepancies between the aircraft and the Da Vinci II ozone data
shown in figure 21 may be real or they may be indicative of a calibration
or other measurement error. Both NOAA profiles were adjusted by the amount
required to align the noontime NOAA ozone level with the concentration meas-
ured on Da Vinci II, a factor of 0.64. These adjusted data are presented in
figure 25. The adjusted ozone data appear to be more consistent with the RTI-
EML and RAMS data for both periods. The 0915 profile coupled with the data
point from Da Vinci II show a slight ozone bulge below 914 m (3,000 ft). This
may be due either to an enhancement beneath the rising mixed layer presumably
due to synthesis aloft or to inherent inhomogeneities that reflect incomplete
mixing of the layered atmosphere. The distinct layering of ozone that existed
at 0915 CST had disappeared by noon as the atmosphere became well mixed. The
63
-------
IOK
8K
UJ
4K
2K
I2IICST
0933 CST
09I5CST
II48CST
Olfc
1
1
+o
I
3.0K
2.4 K
§
1.8 K
s
m
d
I.2K w
0.6K
0.05 0.10 0.15
OZONE,PPM
0.20
Q2§
>.OK
Figure 21. Unadjusted ozone profiles as measured in St. Louis on 8 June 1976 during vertical
flights of the NOAA aircraft that started at 0915 and 1148 CST. Time-resolved
ozone data from Da Vinci II ( A), RTI-EML (+ ), and RAMS stations ( o ) are in-
cluded for comparison.
64
-------
IOK
8K
6K
8
FJ4K
2K
OK
I ' I ' I ' I ' I ' I ' I '
. J2HCST
I . I . I . I . I . I . I
3.0K
2.4K
I.8K
-I
S
L2K
m
3
V)
0.6K
0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.1
BSCAT, I/M x I04
).OK
Figure 22. Bscat profiles as measured in St. Louis on 8 June 1976 during vertical flights of
the NOAA aircraft that started at 0915 and 1148 CST. Time-resolved Bscat data
from RAMS stations ( o ) are included for comparison.
65
-------
IOK
8K
6K
UJ
U
u_
z>
h 4K
2K
OK
I2IICST
0933CST
I
3.0K
2.4K
L8K H
I.2K
H
O
m
m
m
73
CO
0.6K
20 40 60 80
PERCENT RELATIVE HUMIDITY
O.OK
100
Figure 23. Relative humidity profiles as measured in St. Louis on 8 June 1976 during
vertical flights of the NOAA aircraft that started at 0915 and 1148 CST.
Time-resolved humidity data from Da Vinci II ( A), RTI-EML ( + ), and
RAMS stations ( o } are included for comparison.
66
-------
I OK
8K
h-
UJ
LJ
6K
UJ
0
D
t 4K
2K
* .1211 CST
'0933 CST
OK
0.00
3.0K
2.4K
l.8K=j
o
m
m
I.2K3
0>
O)
0.6K
0.01 0.02 0.03
N03 ,PPM
0.04
0.05
O.OK
Figure 24. Nitrogen dioxide profiles as measured in St. Louis on 8 June 1976 during
vertical flights of the NOAA aircraft that started at 0915 and 1148 CST.
Time-resolved NO2 data from RAMS stations ( o ) are included for com-
parison.
67
-------
IOK
8K
CO
* 6K
LJ
a
5 4K
<
2K
OK
1211 CST
0933 CST
,1148
/ CST
3.0K
2.4K
I.8K
I.2K
H
C
o
m
m
m
c/>
— 0.6K
0.00 0.05
0.10
OI5
0.20
0.25
0.0 K
OZONE, PPM
Figure 25. Adjusted ozone profiles as measured in St. Louis on 8 June 1976 during vertical
flights of the NOAA aircraft that started at 0915 and 1148 CST. Time-resolved
ozone data from Da Vinci II (A ), RTI-EML (+ ), and RAMS stations ( o ) are
included for comparison.
68
-------
noontime ozone profile suggests that a near-zero gradient condition prevailed
from the surface to 1,981 m (6,500 ft). The mean adjusted 63 concentration
aloft between 762 and 1,981 m (2,500-6,500 ft) was 0.097 ± 0.010 ppm at 0915
CST and was 0.117 ± 0.010 ppm at noon. This corresponds to an adjusted net
increase of 0.020 ppm and an unadjusted increase of 0.032 ppm.
Portions of the above interpretation that rely on the adjusted data of
figure 25 are somewhat speculative. Several general features are apparent,
however, from either the raw or the adjusted data:
1. Substantial amounts of ozone existed aloft in a layer between
1,219 and 2,133 m (4,000-7,000 ft) at 0915 CST. This is probably
indicative of the elevated ozone levels that prevailed from the mid-
west to the east coast during the time of the study.
2. Furthermore, this ozone could not have been synthesized from pre-
cursors released in St. Louis on 8 June, because by 0915 CST, the
balloon was still upwind of St. Louis and the mixing layer had not
yet reached the 1,219-m (4,000-ft) level.
3. Due to the early hour it is also improbable that the high levels
of ozone prevalent at 0915 CST were photochemically synthesized on
the morning of 8 June.
4. A net ozone increase of approximately 0.020-0.032 ppm occurred
aloft during the 2.5-h interval between the two vertical flights.
This ozone was either synthesized or transported into the area.
Photochemical synthesis is thought to be primarily responsible
for observed increase.
5. Ozone concentrations increased on the ground by 0.056 ppm at the
RTI-EML and by 0.042 ppm at the nearest RAMS station between the
0900 and 1100 CST hours. These increases probably reflect both
mixing and photochemical synthesis.
B data were collected on the aircraft and at the nearest RAMS sites.
s ca L
These data, shown in figure 22, are in fair agreement. Substantial atmospheric
layering is indicated by the 0915 CST profile. In general agreement with the
ozone profiles, well-mixed conditions were established by noon up to 1,676 m
(5,500 ft). Only a slight increase in B from 2.5 to 2.8 x 10" m~ oc-
S Celt
curred aloft at 1,524 m (5,000 ft) over the 2.5 h interval between vertical
flights. These values are somewhat low in comparison to urban levels. How-
ever, it may be too early in the day to use B as an indicator of urban
S Cci L.
influence, because in typical photochemical systems, increased B normally
-£° -1
lags ozone. The highest noontime RAMS B value was 3.0 x 10 m and the
/ - S C3L-
highest daytime value was 4.0 x 10 m
Relative humidity data are presented in figure 23. These data were col-
69
-------
lected on the aircraft and Da Vinci II aloft and at RAMS stations and the RTI-
EML on the ground. The data are in good agreement and indicate layering aloft.
Elevated RH between 1,219 and 1,981 ra (4,000-6,500 ft) was also indicated by
rawinsonde data. A humidity increase of approximately 10 percent is observed
at this level between 0915 CST and noon. The noontime profile suggests that
the layering present at 0915 CST persisted until noon. This is puzzling in
view of the well-mixed conditions indicated by both the ozone and B . noon-
scat
time vertical profiles.
The NO concentration is expected to decrease with altitude due to the
preponderance of ground-based NO sources. Nitric oxide and NOa concentration
A
data were collected on the NOAA aircraft. In both vertical flights, the NO
concentration was relatively constant with altitude at approximately 0.01 ppm.
The mean NO concentration between 762 and 1,981 m (2,500-6,500 ft) was 0.009
± 0.001 ppm in the 0915 flight and increased only slightly to 0.010 ± 0.002 ppm
by noon. The invariant behavior of the NO with altitude is somewhat surprising
in view of the apparent layering of other chemical species and may reflect in-
strument error. The 0900 and 1100 CST hourly average ground-level NO concen-
trations at the nearest RAMS stations were below the minimum detection limit of
0.005 ppm. The reported NO concentrations aloft are also near the instrumental
minimum detection limit and therefore may be somewhat suspect.
The vertical N02 profiles shown in figure 24 are inconsistent with the
RAMS data and may also be suspect. Both profiles are internally consistent
and show decreasing NOa concentrations with altitude . Although behavior may
be indicative of upward mixing from ground-based sources, the persistence of
the vertical gradient at noon is contrary to the well-mixed conditions indi-
cated by the 0^ and B vertical concentration profiles noted earlier. Over
S C£k£
the 2.5-h interval between flights, a net increase of 0.014 to 0.018 ppm N02
was observed aloft. Although upward mixing of N02, an ozone precursor, may
facilitate enhancement and replenishment of Og levels aloft, conflicting evi-
dence and a limited data base prevent this speculation in the present study.
Although they are not pictured in this section, rawinsonde temperature
data indicate that the base of the subsidence inversion was at 2,743 to 3,048 m
(9,000-10,000 ft). A sharp decline in the rawinsonde dewpoint data at 1,981
to 2,133 m (6,500-7,000 ft) suggests the existence of a stable layer beneath
the subsidence inversion at that level.
Vertical profiles of 03, B , and RH also show sharp declines at 1,981
S C3L
70
-------
to 2,133 in (6,500-7,000 ft). In the noontime profile, the mean ozone concen-
tration between 2,133 and 2,743 m (7,000-9,000 ft) was reduced by a factor of
2.4 in comparison to that between 762 and 1,981 m (2,500-6,500 ft). In addi-
tion, relatively high water vapor concentrations existed below 1,981 m (6,500
ft). These observations do not suggest that stratospheric intrusion was re-
sponsible for the elevated ozone concentration between 762 and 1,981 m (2,500-
6,500 ft).
4.3.5.2 Acoustic Sounder
Data from AN! acoustic sounders are presented in figure 26. These data
suggest that Da Vinci II was launched prior to complete dissipation of the
nocturnal radiation inversion.
On the morning of 8 June the inversion was destroyed by 0920 GST; this
correlates well with the first thermal experienced by Da Vinci II between
0937 and 0956 CST (see table 5). This is some 2 to 3 h after mixing had been
established within the lowest 30 m (100 ft) as indicated by RAMS temperature
gradient measurements. This, however, is not inconsistent because acoustic
sounders provide information on the thermal layering to altitudes of 1 km
(3,280 ft). It is likely that an additional 2 h of solar heating were required
before the inversion was completely dissipated.
On the afternoon of 8 June, establishment of the nocturnal radiation in-
version was observed to commence during the 1800 CST hour. This coincides well
with RAMS temperature gradient measurements.
The reduced thickness of the nocturnal radiation inversion before midnight
coincides well with the nighttime ozone peak observed by the RTI-EML. This
suggests that the resistance to downward mixing was minimized at this time and
is consistent with the previous interpretation of the nighttime 03 peak ob-
served at the RTI-EML on the evening of 8 June.
4.3.5.3 Temperature, Relative Humidity, and Condensation
Nuclei on Da Vinci II
Air temperature, relative humidity, and condensation nuclei as measured on
Da Vinci II are presented in figures 27, 28, and 29. Sandia and ASL air tem-
perature data plotted in figure 27 are in good agreement. Comparison of these
data with the altitude profile shown previously in figure 7 indicate the ex-
pected inverse relationship between altitude and temperature. During the day,
when Da Vinci II was above 1,219 m (4,000 ft), the temperature was low. After
71
-------
DA VINCI H JUNE 8-9,1976
ARGONNE NATIONAL LABORATORY
km
1.0'JiJ!
O.5-
0-
04 05 06 07 08 09
HOURS CST
10
12
km.
0-
o * I, •$• • L i • fc» "»*> i- [ i'i':<
13 14 15 16 17 18 19 20
HOURS CST
21
22
kmv
1.0-
0.5-
0-
M
P Q
23 00 OJ 02 03 04
HOURS CST
S T
V tf
. I
05 06 07
Figure 26. ANL acoustic sounder data along the flight track of Da Vinci II, 8 and 9 June 1976.
72
-------
o
CO
CD
CO
cc
o
LU
(XL-
Co ' 3cJniUcJ3dW31
73
-------
o
o
o
00
CM
o
o
CM
LU
cr
U o
O
vO
o h-
CO
O
O
00
o
CM
o
o
o
o
o
00
o
CM
—
00
74
-------
CO
o
eo
III
U313WUN30 oiano
NO
75
-------
a stable altitude was achieved at 2000 CST, the temperature was relatively
constant.
As indicated in figure 28, the ASL and the Sandia relative humidity data
are in good agreement. A direct relationship between RH and altitude is indi-
cated on comparing figures 29 and 7. When Da Vinci II was above 1,219 m (4,000
ft), elevated RH was observed. This is consistent with the aircraft data in
figure 23 and rawinsonde data. As with temperature, the relative humidity pro-
file is relatively stable after 2000 CST.
Condensation nuclei data are presented in figure 29. In the mean, exclud-
ing major excursions, a slight trend toward decreasing CN prevailed over the
day. Selected CN data that coincide with altitude excursions have been identi-
fied by the number of the corresponding altitude excursion shown in figure 7.
The highest CN level is associated with a small descent (3b to 4a) at 640 m
(2,100 ft). Two of the lower midday CN concentrations are associated with two
of the higher altitude excursions, 5b and 8b. In general, however, the limited
frequency of CN measurements prevents comparison with other data.
4.3.5.4 Ozone and Sulfur Dioxide on Da Vinci II
4.3.5.4.1 Ozone and Sulfur Dioxide Profiles
Both ASL and Sandia ozone data are presented in figure 30. The agreement
for these data is very good. In 28 cases of near-simultaneous measurements,
ASL data are 5.5 ± 1.2 percent higher than Sandia data. To allow a visual
assessment of temporal O2one behavior aloft, the ASL data have been replotted
in figure 31 with the points connected. Major concentration excursions that
coincide with altitude excursions have also been identified by the numbers of
the corresponding altitude excursions from figure 7.
Sandia SC-2 data are presented in figure 32. Although S02 excursions are
not associated with major altitude excursions, increased 80% levels coincide
with small-scale altitude reductions in several cases.
4.3.5.4.2 Sulfur Dioxide Excursions
Sulfur dioxide concentrations were below 0.02 ppm (50 |Jg/m3) for the first
12 h of the flight. A marked increase in SC>2 concentration occurred at 2035
CST, approximately 2 h after passing near the Sioux Power Plant and 45 min
after passing near the Wood River Power Plant in the industrialized East Alton
area. Three major S02 excursions occurred at 2047, 0005, and 0503 CST. The
mean magnitude of these peak values is 0.134 ± 0.022 ppm. Smaller peaks asso-
76
-------
LO 0 LO 0 LO 0
CM CM «-• *H O O
• »•••*
O 0 O O O OO
0
1 1 ' 1 ' 1 ' 1 '
JE
r"~
- CE
0
UJ
o +•$.
~~ M : •*
0 "...•+.
• *V« *
HH -Hr
M '•*+ .
0 <**"
5 'V->.
__ *^ . • __
•~~ «^ . • ^^
cc
Q
•.+•.* *
' ". •• *
* « •
• to • •" *
+ *•
•
4 * ../•• '
__ +,.'..-' _
^*-
S
.-f+H-
•+•'• •.'..*. +
. . iL i 1 1 i^Jl 4- *4t>
+****Pl|*|l** * ^
s. -
4- .• '
^Mt
4" ' +
4ff
, 1,1.1.1.
CM
•-«
O
vD
O
O
O
§ 1-
CO
o
flb
£
l-l
0 *~
00
«— 1
0
o
CM
*— 1
o
o
\D
LO 0 LO 0 LO 00
CM CM ~H *H 0 0
1
CO
i 00
•
Q-g
§"
o
N
O
8
O
o o
Wdd '3NOZO
77
-------
LO
CM
o
CM
o
*—I
•
o
LO
O
00
o
CM
O
LU
CC
O
O
O
CO
O
o
GO
CM
o o
LO
CM
O
CM
If)
LO
O
O
o
vD
CO
£
§.
LL
Wdd '3NOZO
78
-------
CO
o
•o
wdd '301x010
79
-------
elated with the major S02 excursions were also observed. Two small peaks were
noted that were not associated with the major excursions.
For the three major S02 excursions, corresponding reductions of ozone
occurred. The mean reduction of 03 is 0.051 ± 0.010 ppm. In addition, reduc-
tions of ozone were observed at 2115 and 2210 CST for the two secondary SC>2
peaks associated with the first major SOz excursion.
The timing of the excursions and the corresponding position of Da Vinci II
on the flight track suggests that the first two major S02 peaks were contrib-
uted by sources located in the greater St. Louis area. The only major SC>2
source located downwind from St. Louis that could have contributed to the 0503
CST excursion was the Coffeen Power Plant. The currently available data, how-
ever, do not allow reconciliation of the SOz excursions with specific sources.
Sulfur dioxide does not react appreciably with ozone. However, both S02
and NO are emitted in high concentrations from both power plants and petroleum
refineries (see tables 6 and 7). If the SC>2 excursions are caused by Da Vinci
II"s entering a portion of such plume, then correspondingly high NO levels
A
would also have been present. Ozone reacts quite rapidly with NO and somewhat
more slowly with N02- For example, in the presence of a constant 0.15 ppm of
03, NO has a half-life of 12 s. Molar NO /S02 ratios from power plant emis-
X
sions calculated from the emissions inventory data range from 0.25 to 0.50.
The ratio for petroleum refineries also falls within this range at 0.45. Thus,
0.034 to 0.067 ppm of NO could be simultaneously present with an SOa peak con-
centration of 0.134 ppm. If it is assumed that this NO is completely consumed
A
by reaction with Os, then a net reduction of 0.034 to 0.067 ppm is expected in
the QZ concentration. The mean observed Oa decrease falls within the expected
range.
4.3.5.4.3 Ozone Behavior and Excursions
Between 0900 and 1100 CST, hourly average ozone levels on Da Vinci II in-
creased from 0.095 to 0.115 ppm, a net increase of 0.020 ppm. This is in ex-
cellent agreement with the net increase of 0.020 ppm calculated from adjusted
NOAA data averaged between 762 and 1,981 m (2,500-6,500 ft).
Between 0900 and 1700 CST, ozone levels aloft generally increased. Hourly
average concentrations increased from 0.095 ppm at 0900 CST to the flight maxi-
mum value of 0.153 ppm at 1600 CST, a net increase of 0.058 ppm. Thus, a net
increase of 0.058 ppm ozone was observed aloft at or near Da Vinci II on 8
June.
80
-------
The mean 03 concentration aloft during the flight was 0.125 ± 0.018 ppm.
On the average, slightly higher 63 concentrations were observed aloft at night
than during the day. The 1100-1500 CST mean value on 8 June was 0.121 ± 0.018
ppm in comparison to the 0100-0600 CST mean of 0.135 ± 0.005 ppm on the morn-
ing of 9 June. This is probably due to the differences in vertical movement
and the resulting highly variable ozone concentrations measured during the
first half of the flight.
During the flight, many excursions occurred in the ozone concentration.
This suggests that Da Vinci II failed to behave as a perfect Lagrangian marker.
Several ozone excursions were associated with altitude movement (see figures
31 and 7). The highest altitude excursion, 8b, is associated with a distinct
reduction of ozone concentration. Da Vinci II may have penetrated the stable
layer at 1,981 m (6,500 ft) and entered the ozone-deficient regime shown in the
vertical ozone profiles, figures 21 and 25. The highly erratic vertical move-
ment of Da Vinci II between 1500 and 1800 CST was not reflected in the ozone
profile.
The altitude profile stabilized at 2000 CST, but the ozone profile became
highly erratic. This behavior is closely associated with Da Vinci II's passing
near the Sioux and Wood River power plants and the Wood River-Roxana petro-
chemical complex. The first of several coincident nighttime excursions of 802
and QS occurred at 2047 CST. The next two 03 reductions, occurring at 2115 and
2210 CST, were coincident with two secondary SQ% peaks associated with the
first major SC-2 excursion.
In addition to being associated with a major increase of SC>2 at 0005 CST,
the negative 63 excursion between 2330 and 0145 CST is associated with a de-
crease in altitude of between 150 and 270 m (500-900 ft). As noted earlier,
elevated ozone was observed between 2315 and 0115 CST on the ground at the RTI-
EML (see figure 20). The coincident downward movement of the balloon, reduced
ozone aloft, and elevated ground-level ozone concentrations provide tempting
opportunities for associations with downward mixing.
The third ozone excursion occurred at 0320 CST and was not associated with
an altitude excursion. Sulfur dioxide was not measured during a 27-min period
encompassing this event and thus prevented a comparison of 0$ and 862 behavior
during this period.
The final major ozone excursion occurred at 0453 CST. This reduced ozone
behavior was associated with a major 802 excursion. These coincident excur-
81
-------
sions are difficult to assign to any specific source. As noted previously,
this behavior was not caused by chemical reaction of ozone with S02- It is
thought to have been caused by the chemical reaction of ozone with the NO
A
that was emitted along with 80% by various sources. Atmospheric layering can
allow several plumes as well as a balloon to travel long distances in parallel
before vertical turbulence promotes interaction and mixing.
4.3.5.4.4 Ozone Half-Lives
It is of interest to speculate on the half-life of ozone under various
conditions in the atmosphere. In the absence of the 0$ precursors (HC and
NO ), ozone in a mixture of air and water vapor will decay slowly in the
x f
dark—half-lives in excess of 100 h have been observed. Under exposure to
sunlight, ozone will, in the same system, decay more rapidly and will exhibit
a half-life of approximately 10 h. This suggests that if 03 and water vapor
were present in an HC- and NO -free atmosphere, half of this ozone would be
naturally destroyed during a single sunny day by photoinitiated processes. In
the troposhere, however, both anthropogenic and natural ozone precursors are
present. Photochemical synthesis from these precursors can permit Og to
accumulate during the day.
In the ambient atmosphere, the HC and NO that act as ozone precursors
A
during the day, act as ozone-destructive agents at night. A question then
arises concerning the behavior of ozone during nighttime hours in the presence
of HC and NO . The nighttime decay rate to a large extent defines the maximum
impact that ozone, which is generated on one day in an urban area, will have
on downwind air quality on the next day.
The ideal approach for answering this question would involve in situ
measuring of [03] aloft under transport conditions. This would require a
Lagrangian platform that could monitor ozone concentrations within an undiluted
air parcel as it travels overnight. Data from the Da Vinci II experiment
provide an opportunity to do just this. Two caveats must be emphasized,
however: Da Vinci II was not a perfect Lagrangian marker, as shown in figure 31,
and the extent of dilution experienced by the air parcel cannot be quantified
for this experiment. With these shortcomings in mind, the ASL ozone data
between 2000 CST on 8 June and 0500 CST on 9 June were examined to determine a
nighttime ozone decay rate (see figure 31).
Ozone data were employed to define an envelope of maximum nighttime ozone
concentrations. A least-squares regression of In [03] versus time permitted
82
-------
the calculation of a nighttime ozone half-life of 116 h. This approach ne-
glects non-Lagrangian behavior of Da Vinci II and negative 0& excursions pre-
sumably caused by isolated pockets of species that act as ozone scavengers.
An early half-life estimate of approximately 30 h was based on a preliminary
limited set of raw data in which negative excursions were difficult to dis-
tinguish. Refinement of the nighttime half-life to our current estimate of
116 h has been permitted by an enlarged data base.
Nitrogen oxides and HC can act as both ozone precursors and ozone-destruc-
tive agents. The lack of NO measurements on Da Vinci II prevents an assess-
A
ment of NO concentrations aloft. The examination of Da Vinci II HC data that
A
is discussed in a subsequent section of this report suggests that the concen-
tration of ozone-destructive and other HC species aloft was low during the
flight. Thus, except for apparent encounters with remnants of industrial
plumes, the conditions for which the above ozone half-life estimate was made
is thought to be representative of suburban-to-urban air. The dark phase
half-life of ozone under these conditions of presumed low levels of both NO
A
and HC is entirely sufficient to allow ozone from one urban area to be trans-
ported overnight to another populated area without significant diminishment
due to decay.
4.3.5.5 Hydrocarbons on Da Vinci II
Concentration profiles for many of the hydrocarbons are presented in
figures 33a-33k. Each of these data points is representative of the few sec-
onds of the flight during which each corresponding grab sample was collected.
In view of this, a comprehensive picture of events for the flight cannot be
prepared. Nevertheless, HC data were examined to determine the extent to
which Da Vinci II was traveling in the urban plume and to document the behavior
of Da Vinci II's environment for the flight. The approaches employed a mean
value analysis, an excursion analysis, and a trend analysis.
4.3.5.5.1 Mean Value Analysis
Mean concentrations and concentration ratios, excluding excursions, were
calculated for each species. These results were compared with ground-level
values typical of urban, suburban, and nonurban RAMS sites.
Detailed HC data were not collected at ground level during the flight of
Da Vinci II. To provide a basis for comparison, however, mean concentrations
were calculated from 3 to 4 days of data collected at three RAMS sites during
g
August 1976. These data are presented in table 13 along with data collected
83
-------
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0600
33a OflVINCI II RTI CflRBON MONOXIDE -
1200
1800 0000
TIME. CST
0600
3.0
2.5
2.0
1.5
1.0
0.5
0.0
1200
120
100
80
60
to
20
0600
1200
33 b jbflVINCI II RTI METHflNE -
1800 0000
TIME. CST
0600
12C
IOC
80
60
HO
20
1200
10.0
(383)
8.0
6.0
N t.O
2.0
0.0
0600
33 c DfWINCI II RTI HCETYLENE .
1200
1800 0000
TIME. CST
0600
10.0
8.0
6.0
1.0
2.0
1200
0.0
too
300 -
(1095)
100 -
0600
1200
1800 0000
TIME. CST
0600
too
- 300
- 200
- 100
1200
(IOK5)
to
0600
1200
1800 0000
TIME. CST
0600
1200
70
60
50
to
30
20
10
I ' I ' I '
33 f DHVINCI II RTI ft-flUTHNE
0600
1200
1800 0000
TIME, CST
0600
70
60
50
to
30
20
10
1200
Figure 33. Pollutant concentrations as determined by RTI from grab
samples collected on Da Vinci 11,8 and 9 June 1976.
-------
I
*
V
70
60
50
to
30
20
10
0600
33g DRVINCI II RTT ISOBUTflNE
1200
1800 0000
TIME. CST
0600
70
60
50
10
30
20
10
1200
a
V
to
35
30
25
20
15
10
5
0
0600
I ' I ' I '
33 h OflVINCI II RTI ISOPENTflNE
1200
1800 0000
TIME. CST
0600
to
35
30
25
20
15
10
S
1200
20
15
. 10
0
0600
\ ' I ' I ' I '
33 i DFWINCI II RTI PROPYLENE
1200
1800 0000
TIME. CST
0600
20 50
15
10
1200
to
30
o 20
10 -
0600
33j DRVINCI II RTI 1-BUTENE -
1200
50
30
20
10
1800 0000
TIME. CST
0600
1200
1000
800 -
600 -
too -
200 -
0600
(2173)
I '(
33k DHVINCI II RTI SUMMED NMHC -
1200
1800 0000
TIME, CST
0600
1000
- 800
- 600
too
- 200
1200
Figure 33. (con.)
85
-------
tfl
co
C
o
•H
4J
U
co
co
2
"8
4J
0)
rH
01
co
4J
Cfl
"d
CO
M
M
•H
U
C
•H
CO
O
4-1
Cfl
CO
a
o
•H
4J
(0
4-1
0)
CJ
C
O
o
CJ
w
c
CO
01
S
00
(11
w
f^
cfl
H
M
I—I
•H
O
a
•rH
Cfl
O
CM
,—4
g
K
rH
to
2
3
rH
0
rH
to
*H
M
§
•rl
4-1
Cfl
CJ
O
rJ
C
^
CT\
rH
0)
a
t— 1
J
CTS
oo"
c
o
rH
0
r^
o
c
o
m
i — i
o
o
rH
, |
C
o
rH
1
O
rH
rH
ol
C
^
rOl
6
O
O
CJ
O
CO
0)
•H
CJ
0)
ft
W
"flJl
i — i >a-vo
r-i--cMi-~invoooo
CM CN rH
"sj" *vf *"3" *^ *«^ *^ "«^ *>^
oo vo cr»
tOOO^vD • • •
• • • • •**3"OOrO
v^ iO CO 00 CM iH
+1 +1 +1 +1 +1 +1 +1 +1
rHmr^-rHincMinoo
rHCMOOrHOO^CM
CM rH CM rH UO
r**.
+1
00 r--
r^ oo
vO
CN
f^
rH
^J-
+1
m CM
rH r^*
in
rH
CN
O CN
CJ CJ
^
CJ
en
01 01
> 4-1
4-1 CO
• crj
4J 4J C
d C CO
01 01 ^0
0 W l-l
•H Q) 3
& fa CJ
0) ft O
B1" c
Q) 13
01 C
M U3 CO
•H rH C^
O 3 cfl
a o ,0
•H J3 )H
> co 3
cfl T3 3
a c co
a)
,d vo C
4-1 1^- Cfl
ON ,O
<4-l rH )^ .
O 3 CJ
4-1 rO
13 CO 4-1 ft
o 3 co a.
•H 60
U ^4 dJ 'H
ft -u
60 CJ CU
01 G Q) M
^ -H ft cfl
3 0) CO
H T3 C
O CO O
IM t3 C -H
a) o 4-1
a) 4-1 -H co
i— i a 4-1 >_i
,O CU CO 4-1
Cfl rH M C
rH rH 4-1 01
•H O C CJ
cO cj ol C
> cj o
cfl CU C cj
H 0
4J 0) CJ i-H
O > rH
(3 CJ 4H >-.
0 rH
CO
rl C
O) Cfl
t~i
0 CD
3 o)
C rH
ft
III 0
CO
C to _
CJ| T)|
•
CO
01
CO
K*^
rH
Cfl
C
Cfl
Pi
* ^3
a> cj
C 25
co
rH B
>~> • O
& Pi V-t
4J •< <4H
01 O
!s cfl
CO 4J
3 >, CO
rH r^ *"O
ft
TJ O
O) Ol U
a N
cfl >^ C
XI rH Cfl
4-J Cfl 0)
Q) C B
CO
CO C
•rl CO O
Ol
0) rH TJ
3 ft 01
rH 0 CO
cO CO cfl
> to PQ
CU| UH| 60|
86
-------
aloft during the flight. RAMS station 101 is considered to be urban, station
114 is a suburban site, and station 124, located well south of St. Louis, is a
nonurban site. At the urban site, data from early morning samples are higher
than those from samples collected during the well-mixed portion of the day by
factors of from 2 to 4. The decrease of HC concentration with increased dis-
tance from the CUA that was noted previously in figure 18 is also evident from
the data in table 13.
In seven of eight cases, the mean HC concentrations aloft are smaller than
those presumed to be representative of suburban St. Louis. This is consistent
with the previous observations based on comparisons of ozone concentrations on
Da Vinci II with RAMS data. These data indicate that, in the mean, Da Vinci II
was within a suburban-to-nonurban air parcel.
9
Kopczynski et al. reported a mean CO/C2H2 ratio of 73 in St. Louis road-
way samples and respective urban and nonurban ratios of 48 and 90. The more
recent results from St. Louis sites presented in table 13 suggest representa-
tive urban and nonurban CQ/Cz#z ratios of 80 and 200. The mean values from
Da Vinci II data range between 120 and 153. This provides additional evidence
suggesting that Da Vinci II was in suburban-to-nonurban air.
4.3.5.5.2 Excursion Analysis
The hydrocarbon data analyzed by RTI were examined for apparent excursions
from their mean behavior. The mean values and the type and timing of the ex-
cursions are identified in table 14. Comments concerning coincident excursions
of altitude, 63, and SOz are also presented.
The WSU acetylene, ethane, and ethylene data cited in the footnote on
table 14 were not considered in this analysis due to a possible calibration
discrepancy. The respective mean values for these species, 13.2, 18.9, and
34.7 ppbC, are approximately a factor of 4 higher than the corresponding RTI
mean values. The ratios of ethane plus ethylene to acetylene, however, com-
pare closely: 4.06 (WSU) vs. 4.10 (RTI), suggesting calibration differences.
In addition, comparison of the WSU acetylene and ethane concentrations and the
carbon monoxide to acetylene ratio with the corresponding RAMS data in table 13
would indicate that for the whole flight, Da Vinci II was exposed to air char-
acteristic of that trapped in an urban area beneath the nocturnal radiation
inversion. Based on the flight track and the altitude profile of Da Vinci II,
it seems unlikely that the air was contaminated to this extent, thus providing
further support for the hypothesis of calibration differences.
87
------- An error occurred while trying to OCR this image.
-------
In general, no clear associations are evident between the times of the
hydrocarbon and altitude excursions. However, increased S02, reduced 63, and
HC excursions occurred at 2158, 0126, and 0441 CST.
During the first 13 h of the flight only two HC excursions occurred that
were widespread across many spe'cies. In the 0919 CST sample, excursions oc-
curred for 7 of 12 species, and 10 of 12 species exhibited excursions in the
1145 CST sample.
The frequency of widespread excursions is substantially greater during
the last half of the flight, the nighttime period. Six such excursions oc-
curred during this period at 2158, 0126, 0441, 0535, 0556, and 0603 CST. This
is also the interval of the flight during which generally high S02 levels were
observed. Da Vinci II passed near the industrialized East Alton-Wood River
area at approximately 2000 CST. The nighttime hydrocarbon excursions may be
associated with various plumes from this area.
In five of the eight cases of widespread hydrocarbon excursions, elevated
concentrations of both paraffins and olefins occurred. The remaining three
cases displayed excursions of paraffinic hydrocarbons only. Interestingly,
both Oa and SC>2 exhibited excursions at these times. This behavior suggests
that the HC excursions that occurred at 2158, 0126, and 0441 CST originated
from different sources than those for the remaining widespread excursions.
Excluding excursions, mean methane concentrations of 1.53 and 1.77 ppm
were determined by RTI and NCAR. These data are slightly higher than the
global background methane concentration of 1.4 ppm. The four excursions in
the RTI methane data exceed the background by factors of 4 to 80. These peaks
coincide with excursions of ethane and ethylene, propane, and isobutane. The
ratios of ethane plus ethylene to methane (~0.01), to propane (~1.0), and to
isobutane (~20) for the samples at 2158, 0441, and 0556 CST suggest common HC
sources. Acetylene is usually considered to be an ideal tracer of automotive
emissions. Because neither methane nor propane are associated with automotive
emissions and because no acetylene excursions occurred in these samples, it is
unlikely that the sampled air parcels were heavily contaminated by automotive
emissions.
The first methane excursion (1145 CST) is associated with the major acet-
ylene excursion. This suggests, at least in this case, that two sources of
hydrocarbons contributed to the overall composition of the sampled air parcel.
Propane excursions coincide with excursions for the sum of butanes. This
89
-------
tends to suggest a similar source for both categories of emissions. Excursions
of the individual butanes, normal and isobutane, however, do not coincide uni-
versally with propane peaks. This suggests at least two sources--both emitting
propane, one emitting butane, and the other emitting isobutane.
The alkane hydrocarbon, isopentane, comprises a large fraction of gasoline
vapor. Isopentane excursions at 1145 CST may be associated with automotive
evaporative losses. The peaks at 0126 and 1441 CST may be associated with
•*
evaporative emissions from a petrochemical complex. The largest excursion of
isopentane occurred at 1244 CST. No corresponding C$ or C4 paraffin peaks were
apparent, although CO exhibited a coincident excursion. Implications of this
behavior are not clear.
Condensation nuclei data were compared to CO and C2H2 data. Theoretical-
ly, CN should be indicative of combustion sources; however, the limited data
base prevented a time-resolved analysis of both CN and the appropriate HC that
might conclusively identify various combustion sources.
The limited data base prevents detailed reconciliation of sources from
the HC data. Analysis of emissions from major sources is required to assign
fractional contributions of specific sources to each sampled air parcel. In
a study of this type, however, it may not be possible to obtain HC data from
the major sources. Nevertheless, a major shortcoming of the current data base
was the lack of coincident HC samples for comparing concentration behavior with
air quality parameters, altitude, and source location. More frequent HC sam-
ples should therefore be collected in future studies to provide a more compre-
hensive data base.
4.3.5.5.3 Trend Analysis
The behavior of a chemical species in the atmosphere is governed by emis-
sions, horizontal advection, dilution due to dispersion and to increased mixing
height, and chemical reaction. It is difficult to quantify this information
for the flight of Da Vinci II with the current data base. However, trends in
hydrocarbon behavior were examined for the following relatively unreactive
($\
species: CO, C2H2, C2H4 + C2H6, C3H8, iC4H10, nC4H10, iC5H12, Freon 11, and
Freon® 12.
Visual inspection of concentration-time profiles for these species sug-
gests that concentrations declined between 1300 and 2100 CST. Linear regres-
sion of concentrations with time were performed with the excursions excluded.
Predicted concentration changes and r2 values were determined for each species.
90
-------
Concentration reductions were indicated for eight of the nine species examined.
Declines ranging from 38 to 58 percent were found for the butanes, isopentane,
jg»
and the Freons. The r2 values ranged from 0.42 to 0.67 for these five spe-
cies. Since this behavior is not expected to occur for both hydrocarbons and
®
Freons solely from chemical reaction in the atmosphere, dispersion processes
may be implicated.
As noted in a previous section, Da Vinci II may have entered a plume at
2100 on 8 June that was enriched in HC and may have remained in HC-rich air
for the remainder of the flight. Mean concentrations of the nine species
considered above were calculated for two time intervals: 1300-2100 and 2100-
0700 CST. Ratios of data for the period after 2100 to that before 2100 were
all greater than one. Mean concentrations of ethane plus ethylene and of the
®
Freons were enhanced by only 2 to 5 percent in comparison to the greater than
20 percent enhancement calculated for the remaining six species. A statisti-
cal test with the null hypothesis of no difference between the means was per-
formed for each species. The hypothesis was rejected (« = 0.05) for CO, C2H2,
i-C4Hio> n~c4Hio> and i-C5H12. This provides strong evidence supporting an
earlier contention that Da Vinci II entered an air parcel on the evening of
8 June that was enriched in hydrocarbons.
4.3.6 Air Quality Aloft and at the Ground
4.3.6.1 Sulfur Dioxide
Profiles of hourly average SC>2 concentrations measured on Da Vinci II, on
the RTI-EML, and at the RAMS stations nearest Da Vinci II are presented in fig-
ure 34. Agreement among these profiles is poor.
Sulfur dioxide concentrations at RAMS sites were low. The daily RAMS net-
work maximum of 0.04 ppm occurred at 1100 CST at station 115, although along
the flight track, concentrations measured at RAMS stations were below 0.005
ppm. This is in qualitative agreement with concentration data from Da Vinci
II.
It has been noted that the S(>2 monitors employed on Da Vinci II and on
the RTI-EML were subject to positive interference by hydrocarbons. The RTI-
EML traveled on roadways and sampled air that was contaminated to an unde-
fined extent with vehicular exhaust that was rich in hydrocarbons. The RTI-
EML S02 data are therefore considered to be somewhat suspect.
The SOg monitor aboard Da Vinci II was subject to the same interferences.
91
-------
s
8
o>
s
I
3
£
?
il
wdd '301x010
92
-------
Analyses of grab samples collected on Da Vinci II, however, have shown the
hydrocarbon concentrations aloft to be low. On this basis, the SC>2 profile
measured on Da Vinci II is considered to be more reliable than data from the
RTI-EML. The SOg data collected on Da Vinci II were examined previously and
will not be discussed further.
4.3.6.2 Ozone
Profiles of hourly average ozone concentrations measured on Da Vinci II,
on the RTI-EML, and at the RAMS stations nearest Da Vinci II are presented in
figure 35. As noted previously, the RTI-EML data and the interpolations from
isopleths of RAMS data are in good agreement.
Early morning ozone concentrations aloft on Da Vinci II were greater than
those at the ground. This is presumably due to reduced concentrations of 03-
destructive agents above the nocturnal radiation inversion. It should be noted
that QS levels at the RTI-EML and the westernmost RAMS stations, while reduced
in comparison to levels aloft, were significantly higher than levels in down-
town St. Louis. This reflects the differences in concentrations of Os-destruc-
tive agents in nonurban and urban air beneath the nocturnal radiation inver-
sion.
As the surface-based nocturnal inversion was destroyed, ground-level ozone
concentrations approached those aloft. Concentrations at these two locations
became almost indistinguishable by the 1100 CST hour. This is also illustrated
by the adjusted vertical ozone profiles in figure 25. The stratification that
was apparent in the 0915 CST profile had disappeared by the time the 1148 CST
vertical flight was conducted.
Sunrise occurred at 0436 CST on 8 June. Temperature gradient data suggest
that the inversion breakup began to occur through the first 30 m (100 ft) dur-
ing the 0600 CST hour. According to acoustic sounder records, the inversion
was destroyed by 0920 CST. The ozone profiles in figure 35 suggest that well-
mixed conditions prevailed from the 1100 to the 1700 CST hour. The 1100-1500
CST mean concentration on Da Vinci II of 0.121 ± 0.018 ppm is in excellent
agreement with the corresponding 0.123 ± 0.005 ppm from the RTI-EML.
Based on the RTI-EML ozone data, the first indication of the reestablish-
ment of two ozone regimes occurred during the 1800 CST hour. On 8 June, sun-
set occurred at approximately 1924 CST, and temperature gradient data show the
first indications of the formation of the radiation inversion during the 1800
CST hour. Acoustic sounder data show similar behavior during the 1800 CST hour.
93
-------
CO
o
to"
0)
O)
•o
i
|c
II
!-.
5*
^ UJ
o _-
Sz
o
X
£
Wdd '3NOZO
94
-------
At the 1900 CST hour and throughout the night, the Da Vinci II and RTI-
EML data clearly demonstrate existence of two regimes of ozone. Aloft, ozone
remains insulated from ground-based destructive agents, and the concentration
remains relatively constant except for periodic encounters with industrial
plumes. Beneath the inversion, ozone concentrations are reduced both by sur-
face deposition and by chemical reaction with destructive agents emitted from
local sources that are trapped within this layer. The relative importance of
these two mechanisms remains to be defined.
Just prior to midnight, the inversion was penetrated as a result of me-
chanical turbulence, which brought air containing high concentrations of ozone
down to the surface. The simultaneous reduction in ozone concentration aloft
was coincident with SC>2 excursions and hence the associated, but unmeasured,
NO excursions that accompany SC>2 in power plant plumes. The extent that up-
A
ward mixing of ozone-deficient air from the surface perturbed concentrations
aloft at the 762-m (2,500-ft) level remains to be determined.
Sunrise occurred at approximately 0427 CST on 9 June. This correlates
with the increase in ground-level ozone measured on the RTI-EML during the
0600 CST hour. Similar behavior is noted in figures 12 and 35 on 8 June at
RAMS stations. This increase in ozone is presumably due both to mixing of
high concentrations of ozone to the surface from aloft during the dissipation
of the nocturnal radiation inversion and to photochemical synthesis.
Ozone concentrations at the RTI-EML increased on the morning of 9 June as
if to converge with 0^ levels aloft as measured on Da Vinci II. On the basis
of the QS behavior on 8 June in St. Louis, convergence is expected on the morn-
ing of 9 June with the establishment of well-mixed conditions. Termination of
the experiment prevented an assessment of this hypothesis.
On the afternoon after Da Vinci II had landed, additional ozone measure-
ments were conducted aloft by ICFAR. Ozone concentrations measured at 457 m
(1,500 ft) AGL from 225 km northeast to 290 km southwest of Indianapolis are
presented in figure 36. Between 1527 and 1656 CST on 9 June, ozone concentra-
tions aloft ranged between 0.14 and 0.19 ppm. These data coupled with those
from Da Vinci II, the NOAA aircraft, RAMS stations, and ground stations in the
northeastern quadrant of the country provide strong evidence for the existence
of high regional-scale concentrations of ozone aloft during the period encom-
passing the Da Vinci II experiment.
95
-------
H
W
U
o
M
H
u
W
«
M
c
s
CM
CM
O
CM
00
iH
CD
VO
iH
CO
CM
o
H
o
oo
o
vO
O
2
O
M
H
U
W
H
O
O
o
O
IT)
CM
O
O
CM
H
GO
H
§
CO
O
O
w a S
0160.
g O CO
°£ S
ri ^
c
'§
CM
°
10
o o o o
mdd 'auozo
o
o
CM
O
o
z
O .S
•50
96
-------
4.3.6.2.1 Transported Ozone
Evidence for the regional-scale existence of high concentrations of ozone
aloft during the study period have been presented. The mean of the 1100-1500
CST hourly average ozone concentrations at an upwind site was used as an esti-
mate of ozone transported into the study area. The 1100 to 1500 CST interval
is considered because it occurs during the portion of the day when the bound-
ary layer is well mixed, and it occurs late enough in the day for synthesis of
ozone from transported precursors to have occurred as well.
On 8 June the winds were generally from the west. RAMS station 125 is
the westernmost site. The mean of the 1100-1500 CST hourly average ozone con-
centrations at this site was 0.108 ± 0.007 ppm. Thus, a considerable amount
of ozone, approximately 0.108 ppm, was transported into St. Louis on flight
day.
The mean of the 1100-1500 CST hourly average ozone concentrations observed
on Da Vinci II was 0.121 ± 0.018 ppm. This is only slightly higher than the
0.108 ppm transported into the area. The maximum mean 1100-1500 CST ground
level ozone concentration within the RAMS network was 0.177 ppm. Comparison
of these data (also see figure 17a) suggests that the balloon was not in the
center of the urban plume as defined by ozone concentrations.
4.3.6.2.2 Ozone Synthesis
During the morning, ground-level ozone concentrations were less than con-
centrations aloft. The observed increase of ozone concentrations aloft, there-
fore, could not have resulted from upward mixing of ozone from the ground, but
was probably due to photochemical synthesis aloft.
The magnitude of synthesis aloft on 8 June is estimated to have been 0.058
ppm. This estimate is based on the difference between the 0900 CST and the
daily maximum hourly average measured on Da Vinci II.
A net increase of 0.020 ppm has been noted on Da Vinci II between the 0900
and 1100 CST hourly average ozone concentrations. Between 0915 and 1148 CST,
net increases ranging from 0.020 to 0.032 ppm have been found based on averaged
ozone data collected between 762 and 1,981 m (2,500-6,500 ft) MSL on the NOAA
aircraft. During this same period, ozone levels at the RTI-EML and RAMS sites
nearest the flight track increased by 0.042 and 0.056 ppm. Thus, the net in-
crease of ozone at the ground was approximately twice the net increase that
occurred aloft.
Well-mixed conditions prevailed from noon until some time in the after-
97
-------
noon. During this period, synthesis is presumed to be a major contributor to
increases in the observed ozone level. On Da Vinci II, the ozone increase be-
tween the 1100 CST hour and the daily maximum value was 0.038 ppm. The corres-
ponding increases at the nearest RAMS station and at the RTI-EML were 0.026 and
0.024 ppm. These results suggest that between noon and the end of the day, a
net amount of approximately 0.030 ppm of ozone was synthesized along the flight
track.
Based on ozone isopleths drawn from RAMS data and presented in figure 19,
the buildup and movement of an ozone "hot spot" was documented for 8 June and
was attributed to precursors released in the St. Louis urban area. A specific
source was not identified.
4.3.7 Considerations for Future Programs
The current study was successful at consolidating data collected by many
separate investigators and at providing a cohesive interpretation of the re-
sults. During the data analysis effort, however, several points became appar-
ent that should be considered in future programs of this type.
1. The limited area encompassed by the RAPS monitoring network in com-
parison to the scale of air movement that can occur during a 1-day
period prevented detailed definition of the St. Louis urban plume.
A comprehensive measurement program conducted on an aircraft in sup-
port of future balloon-borne experiments could fill this gap by de-
fining the vertical and aerial extent of the urban plume, the influ-
ence of major sources on that plume, and the maximum ozone concen-
tration within the plume. The lack of such an effort hampered the
data analysis effort in the current study.
2. Nitrogen oxides are ozone precursors and are instrumental in atmos-
pheric photooxidation processes. They are also emitted by both au-
tomotive and industrial sources. In future studies, balloon-borne
measurements of NO could enhance the quality of the data base by
permitting assessment of the impact of ground-based automotive
sources or industrial plumes on NO concentrations aloft.
r x
3. The atmosphere behaves in a dynamic fashion. The behavior of atmos-
pheric chemical species is also highly variable. Continuous measure-
ment of such species as Og, NO , S02, and CN as well as frequent CO
and EC sampling are desirable.
4. Detailed RAPS emissions inventory data were not available in the cur-
rent study. High-quality data of this type are needed to relate am-
bient concentrations of chemical species to their emission rates.
5. An assessment of the degree to which a balloon behaves as an ideal
Lagrangian marker is important. Relative air velocity data should
be examined in future studies to permit this evaluation.
98
-------
4.4 Findings
The following observations have been drawn from the atmospheric chemistry
analysis:
1. Conditions of elevated ozone concentrations persisted aloft for the
study period. These conditions were widespread and extended for
several hundred kilometers.
2. Ozone and water vapor concentrations determined in vertical aircraft
flights do not suggest that stratospheric intrusion was responsible
for the elevated ozone concentrations that existed aloft on 8 June.
3. Air containing high concentrations of ozone, approximately 0.11 ppm,
was transported into the study area on the flight day.
4. Da Vinci II was launched into air that had an immediate history in
nonurban environs, although the total experiment was conducted
within a stagnant, polluted, high pressure system.
5. Da Vinci II did not travel in the heart of the St. Louis urban plume
as defined by ozone concentrations. Examination of the hydrocarbon
concentrations sampled on the balloon suggests that the flight oc-
curred in air characteristic of suburban-to-nonurban areas.
6. During the morning, ground-level ozone concentrations were less than
concentrations aloft. The observed increase of ozone concentrations
aloft, therefore, could not have resulted from upward mixing of
ozone from the ground, but was probably due to photochemical synthe-
sis.
7. The impact of anthropogenic emissions on ambient concentrations
within the urban area is variable and depends strongly on the time
of day. Ground-level CO and NO were diluted by factors of 5 to 9
between morning and afternoon. This behavior reflects the signifi-
cant increase in the mixed volume that occurs with the dissipation
of the surface-based radiation inversion and the establishment of
well-mixed conditions.
8. As vertically well-mixed conditions were established during the
morning, both downward mixing and photochemical synthesis contrib-
uted to the observed net increase in ground-level ozone concentra-
tion. The ground-level net increase was approximately twice that
aloft.
9. Between the 1100 and 1700 CST hours on flight day, a near zero ozone
concentration gradient existed from the ground into the mixed layer.
This is, to a large extent, based on the close comparison of ozone
measurements determined on the balloon with those beneath the bal-
loon at ground level.
10. After well-mixed conditions were established and until they began to
dissipate, the increase in ozone concentrations aloft and at ground
level were approximately equal and probably resulted from photochem-
ical synthesis.
11. The stratification of the daytime mixed layer that occurred at night
with the establishment of the nocturnal radiation inversion resulted
in the formation of two regimes of ozone concentrations.
99
-------
12. Ozone concentrations beneath the radiation inversion were much re-
duced compared to levels above it. This is presumably due to de-
struction by surface deposition and reaction with ozone-destructive
agents trapped beneath the inversion.
13. The dark-phase stability of ozone at the Da Vinci II level above the
nocturnal surface-based radiation inversion suggests that transport
of ozone can occur aloft over long distances at night without signif-
icant diminishment. It is possible for the ozone to be transported
overnight several hundred kilometers and be mixed to the ground on
the next day with a significant impact on ground-level air quality.
14. Penetration of the nocturnal radiation inversion occurred frequently
during May and June of 1976 and resulted in the mixing of high
levels of ozone to the ground and of precursors from the ground to
aloft. This was suggested by the sharp nighttime ozone peaks and
the associated declines of NO and CO observed at selected ground
stations. This phenomenon was widespread and occurred over several
hundred kilometers. It may also provide mechanisms for increasing
nighttime ozone destruction aloft and for enhancing early morning
ozone synthesis by distributing ozone precursors aloft above the
inversion before sunrise.
15. The balloon may have entered an air parcel that was enriched in
hydrocarbons after 2100 CST on 8 June and traveled within this par-
cel for the remainder of the flight.
16. Sharp reductions in ozone concentrations that occurred during the
nighttime portion of the flight were associated with increases in
concentration. Both S02 and NO are emitted by power plants and
petroleum refineries. Although NO was not measured in this study,
the observed ozone behavior is proBably the result of destruction by
reaction with NO .
x
17. The influence of ozone precursors emitted in the urban area were
manifest at ground-level by ozone concentrations enhanced over non-
urban concentrations. The net magnitude of this enhancement was
0.06 to 0.11 ppm. Both the buildup and movement of the region of
enhanced ozone were documented, although detailed definition of the
extent and magnitude were severely hampered by the lack of a compre-
hensive supporting measurement program conducted on an aircraft.
18. The feasibility of balloon-borne experiments for obtaining the data
needed to address atmospheric chemistry problems has been demonstra-
ted. However, this study has also shown that detailed interpreta-
tion of the results from such a study requires data from several
complementary platforms: the Lagrangian marker, a ground-level
chase vehicle, a network of ground stations, and a chase aircraft.
The omission of any one of these facets severely limits the inter-
pretation effort.
100
-------
5.0 MESOMETEOROLOGICAL AND AIR POLLUTION STUDY RELATIVE TO DA VINCI II
5.1 Introduction
The objective of section 5.0 is to characterize the three-dimensional
meteorological and air pollution distribution on 8 June 1976 in St. Louis,
during the period Da Vinci II was in close proximity to the city. For the
study, ozone was used as the primary trace gas for air pollution. Upper level
(but in the boundary layer) concentrations of ozone were obtained by the Da
Vinci II system, which was launched at 0756 CST on 8 June from the Arrowhead
Airport, some 24 km west of St. Louis. In the period 0756 to 2100 CST, the
balloon drifted in and around the immediate vicinity of St. Louis.
In the period of interest, the meteorology in St. Louis was dominated by
a strong, daytime heat island circulation. It is normally considered that the
most intense urban heat island circulation should be found when the urban heat
island is most intense. Past evidence as well as current evidence shows that
the urban heat island is generally most intense at night. Past data have
also shown that at times the urban heat island persists during the day, partic-
ularly in the winter, which is believed to be due to anthropogenic heating.
Previously, there has only been scattered evidence that a daytime heat island
persists in the summertime.
If the heat island exists during the day, there is the potential to
produce a stronger heat island circulation than during the night. During the
daytime period, the boundary layer is generally characterized, particularly in
the summertime, by a deep adiabatic or superadiabatic layer. The presence of
this layer will allow heating from the urban surface to be mixed through a
deeper layer. From basic hydrostatics, it can be shown that if the heating is
distributed through a deeper layer, a greater pressure perturbation will be
produced at the surface over the city. This will lead to greater horizontal
and vertical accelerations. The general motion associated with the urban heat
12
island circulation is complex and may have a marked effect on the air pollu-
tion distribution in and around the city.
One of the purposes of the Da Vinci II system was to characterize the
time history of air pollution in the urban plume as it moved downstream from
St. Louis. However, due to synoptic and local flow characteristics, the
balloon drifted around the city for approximately 13 h yielding a unique
opportunity to determine three-dimensional structure of air pollution where a
101
-------
marked heat island circulation existed. Following are the results of the data
analyses.
5.2 Data Sources
One of the basic sources of data for this analysis was the Regional Air
Pollution Study (RAPS). The location of RAPS surface and upper air stations
operating on 8 June are shown in figure 37. Five of the RAPS stations are not
shown in the figure because they were outside of the region of immediate
interest. However, these stations were used in the analysis of data.
For the analysis, hourly averaged temperature, winds, and ozone concen-
trations from the RAPS surface stations were used for the period 0300 CST,
8 June, to 0500 CST, 9 June. The surface wind data were used as input for an
objective analysis algorithm, which produced an interpolated wind speed field
on a 4-km by 4-km grid over the area of interest. The radiosonde data provided
vertical profiles of potential temperature, wind speed, and wind direction for
0345 CST, 0945 CST, 1545 CST, and 2145 CST at two upper air stations. One of
the upper air stations is located in downtown St. Louis and the other in the
rural region to the west (figure 37).
The principal source of ozone data aloft was the Da Vinci II system. The
Da Vinci II system made a variety of environmental and gas chemistry meas-
urements; however, of particular importance to this study were the ozone
measurements.
Surface ozone concentrations were also obtained by the RTI-EML. The
RTI-EML followed the track of Da Vinci II as closely as possible from takeoff
to landing (figure 38) making a number of environmental measurements. Only
the ozone measurements were used in this study. A detailed description of the
Da Vinci II system and the RTI-EML is given in section 3.0.
The ozone data from the Da Vinci II system and the RTI-EML were obtained
in the period 0756 CST, 8 June, to 1000 CST, 9 June 1977. However, for this
study, only the data for the period 0800 CST, 8 June, to 0600 CST, 9 June,
were used. Between 0756 and 2100 CST, 8 June, the Da Vinci II system and the
RTI-EML were in the immediate vicinity of St. Louis, Missouri.
5.3 Surface Temperature Distribution
Hourly surface temperature from twenty-five RAPS stations were used to
delineate the St. Louis urban heat island on 8 June. Figure 39 represents the
diurnal variation of the temperature difference between the urban and rural
102
-------
I 142
RAMS Station
RAWINSONDE Station'
\
I I 1111
0 5km
Figure 37. The area of immediate interest around St Louis showing the locations
of the RAPS stations and the two upper air sounding stations.
103
-------
Da Vinci II Gondola
—Q— RTI Van
5km
Figure 38. The ground track of the Da Vinci II gondola and the track of the
RTI-EML from lift-off at Arrowhead Airport through 2000 CST.
The RTI-EML maintained visual contact with the gondola for the
entire mission.
104
-------
5 —i
4 -
Qi
1—1 "3
o 3 —
UJ
ct:
o:
1 -J
i i r r i i i i i i i i i i i * i i
04 06 08 10 12 14 16 18 20
TIME (CST)
Figure 39. The hourly temperature difference between the St. Louis
center city and the rural/suburban area some 15 mi away.
region for the period 0300 to 2000 CST on the 8th of June. Temperature at the
surface in the urban region was represented by the average of four RAPS sta-
tions in the center of the city (stations 101, 105, 106, and 107). The rural
surface temperature was represented by the average, using data from stations
117, 118, 119, and 120. The diurnal variation is typical of that found by
*u *-• *. Hil3
other investigators.
The data show that at 0300 and 0400 CST, the heat island intensity (meas-
ured by the urban-rural temperatures difference) was, on the average, slightly
more than 3° C. By 0500 CST, warming began and the magnitude of the urban
heat island intensity began to decrease (sunrise occurred at 0445 CST). The
heat island existed throughout the daytime period, reaching a minimum at 0900
CST of approximately 0.6° C. Afterwards, it increased somewhat, probably due
105
-------
to increased anthropogenic activity. The marked increase in the intensity of
the urban heat island after 1500 CST is probably due to the fact that the
solar intensity reaching the surface had decreased at this time, the infrared
outgoing radiation was overbalancing the incoming radiation, and the rural
regions were cooling at a greater rate than the urban regions.
The spatial distribution of the urban heat island is shown in figures 40
and 41. At 0500 CST (figure 40), the actual heat island intensity was somewhat
greater than 3.5° C (the heat island intensities shown in figure 39 are based
upon averaged temperatures). The extension of the heat island to the north is
not to be considered a result of advection since it will be shown that the
flow at this time was from the northwest to the southeast at the surface. It
is believed that the extension of the heat island to the north was due in part
to intensified anthropogenic activity in that region.
At 0900 CST, the surface analysis showed a weak horizontal contrast in
temperature (figure 40) on the order of 1.5° to 2.0° C. There was a region of
low temperature along the northwest boundary of the urban complex, which was
slightly over 1° C lower than the center of the city to the east and 0.5° C
lower than the rural areas to the north, south, and west.
The surface warming trend continued uniformly over the entire study area
for the next 5 h. During this period, the pocket of slightly lower surface
temperature along the western edge of the urban complex remained identifiable
although it was never more than 0.5° C lower than the surrounding rural re-
gions. During the early afternoon, the center of the urban heat island shifted
back and forth between RAPS station 107 (approximately the geographic center
of the St. Louis downtown area) and station 105 (some 6 km to the south south-
east near the west bank of the Mississippi River). It was located at station
107 at 1000 CST to 1100 CST, but shifted to station 105 for the next 3 h. It
then shifted back to station 107 at 1400 CST when the highest surface tempera-
ture of the day (30.2° C) was recorded (figure 40). At 1400 CST, the horizon-
tal contrast in temperature had increased somewhat but not to a great extent.
The cooler area to the northwest was still evident.
From about 0800 to approximately 1600 CST, the area east of the river and
to the northeast of the center of the city remained warmer than the suburban
areas to the north, west, south, and southeast of the city. Stations 103,
108, 109, and 115 consistently recorded temperatures within a few tenths of a
degree of the temperatures found in the center of the city and nearly a degree
106
-------
0500 CST
0900 CST
1400 CST
Figure 40. Spatial distribution of the St. Louis urban heat island [°C] at
0500, 0900, and 1400 CST on 8 June. (Dashed isotherms
indicate half-degree increments.)
107
-------
higher than the other suburban areas. This apparently was due to a transport
of heat from the urban center.
At 1500 CST, the outgoing radiation apparently overbalanced the incoming
solar radiation because the cooling process began in which the rural and
suburban regions cooled at a greater rate than the center of the city did.
This caused an increase in the intensity of the heat island as mentioned
earlier. That there is a greater cooling rate in the rural regions relative
14
to the urban regions has been demonstrated by Oke and East and Oke, Maxwell,
and Yap.
By sunset (1900 CST), the intensity heat island had increased to a value
between 2.5° and 3.0° C (figure 41, 1800 CST). It was approximately at this
time that the cool pocket along the northwestern boundary of the urban complex
disappeared. At 2000 CST, the heat island was centered at station 105 and had
increased in intensity to between 5° and 5.5° C (figure 41). Note that the
isotherm distribution again shows an extension of the heat island to the north
of the city.
5.4 Surface Wind Distribution
An interpolation model and the surface wind observations from the RAPS
stations were used to produce analyses of the surface wind field over the
study area for the 8th of June (figures 42 through 43). The length of the
wind vectors appearing on a 4-km square grid are proportional to the wind
speed and indicate the direction toward which the wind is blowing away from
the grid point. Hourly averaged wind data were used in the interpolation
model. These data were obtained at the 10-m level above the surface, and the
hourly average represented 60 1-min observations from the previous hour.
Just over an hour after sunrise, the surface flow over the St. Louis
urban complex was from the northwest and typically light for early morning
(approximately 1 m/s or less) (figure 42, 0500 CST). An increase of the wind
speed was noted in the western part of the city; but in the low areas east of
the Mississippi River, the winds were calm. This region of calm winds down-
wind of the city may be associated with the relatively intense heat island
(figure 40) that existed over the St. Louis urban complex at this time.
This same general flow pattern continued for several hours, strengthening
in wind speed to about 3 m/s by 0800 CST, which may be due to the increase in
mixing as the boundary layer became less stable through daytime heating. As
108
-------
1800CST
2000 CST
Figure 41. Spatial distribution of the St. Louis urban heat island [°C]
at 1800 and 2000 CST on 8 June. (Dashed isotherms indicate
half-degree increments.)
109
-------
0500 CST
0900 CST
5km
1000 CST
1400 CST
0 5km
5km
Figure 42. Interpolated surface wind field [m/s] at 0500, 0900, 1000, and 1400
CST on 8 June. Vectors are proportional in length to wind speed
and extend from the grid point in the direction of the flow.
110
-------
1500 CST
1800 CST
5km
2000 CST
5km
Figure 43. Interpolated surface wind field [m/s] at 1500, 1800, and 2000 CST
on 8 June. Vectors are proportional in length to wind speed and
extend from the grid point in the direction of the flow.
Ill
-------
the intensity of the urban heat island decreased and the mixing increased,
measurable air movement began in the low-lying regions.
At 0900 CST (figure 42), a transition in the direction of the surface
mesoscale flow began to take place. At 0800 CST, the wind was from the west
northwest; however, by 1000 CST, the wind direction was from the west south-
west. This factor, combined with the influence of the effects on the momentum
field by the urban heat island, produced a complex flow pattern varying in
direction from northerly to the north of the city and westerly to the south of
the city at 0900 CST.
The wind field at 1000 CST (figure 42) showed the first convergence
effects of the daytime urban heat island circulation. A region of inflow was
apparent over and downwind of the city. This flow pattern persisted and at
1400 CST (figure 42) the center of convergence was north and east of its
1000 CST position. Note in both the 1000 CST and the 1400 CST analyses, the
winds north of the city were more westerly than those to the south and west.
This was probably due to the pressure gradient accelerations associated with
the heat island.
The intense effect of the heat island circulation persisted to at least
1600 CST (figure 43, 1500 CST). At that time, the zone of convergence began to
weaken considerably. The time 1500 CST marked the beginning of the intensifi-
cation of the urban heat island through differential cooling when the suburban
and rural regions were cooling at a greater rate than the urban region.
Low-level cooling suggests the initiation of a surface-based stable layer.
Stability in the boundary layer will weaken the heat island circulation. '
By 1700 CST, when the effects of surface cooling were more evident, the
heat island circulation had diminished considerably resulting in a fairly
uniform west-southwesterly flow of around 3 m/s over the entire area. There
was a zone of weak convergence evident downwind of the city's center. By 1900
CST (figure 43, 1800 CST), no evidence of the urban heat island circulation
could be determined in the flow even though the heat island intensity had
increased markedly by this time (figures 39 and 41). At 2000 CST (figure 43),
the overall wind speeds were diminishing to about 1 to 2 m/s accompanied by a
slight backing. The decrease in the wind speed and the backing of the wind
direction are indicative of decreased mixing in the boundary layer. The urban
heat island circulation was not evident in the 10-m wind field, even though a
5° C surface temperature difference existed between the center of the city and
112
-------
the surrounding rural regions. In light of the results of Vukovich and Dunn ,
this has to be attributed to the dampening effects of a strong surface-based
inversion.
5.5 Temperature and Wind Profiles
The temperature and wind profiles at RAPS upper air stations 141 and 142
were examined to determine the diurnal variability of the potential temper-
ature distribution and the wind distribution with height. Figure 44 shows the
diurnal variability of potential temperature with height at station 141 in the
center of the city. At 0346 GST, a surface-based inversion was noted with the
top at approximately the 200-m level. Above the 200-m level to approximately
2.7 km, the boundary layer was characterized by weak stability. By 0949 CST,
a shallow superadiabatic layer developed between the surface and the 100-m
level. Above the 100-m level, the atmosphere was quasiadiabatic to 1.2 km.
Stability characterized the layer above 1.2 km. As the analysis of the surface
temperature distribution indicates, after 1500 CST, surface cooling began to
occur. The 1545 CST temperature profile shows a stable layer had developed in
the urban region from the surface to the 200-m level, confirming the earlier
suggestion that the surface cooling was producing low-level stability. Above
the 200-m level, the boundary layer was superadiabatic to 400 m, quasiadiabatic
in the layer from 400 m to 1.2 km, weakly stable from 1.2 km to 3.0 km, and
strongly stable above the 3.0-km level. At 2151 CST, the shallow surface-based
stable layer had strengthened but its top was still at the 200-m level. Above
the 200-m level, the boundary layer was characterized by weak stability up to
2.4 km and strong stability above the 2.4-km level.
In the rural regions (station 142) at 0358 CST (figure 44), a shallow
inversion was also evident from the surface to the 300-m level. Weak stabil-
ity characterized the region from 300 m to 2.8 km and strong stability was
found above the 2.8-km level. Comparing the temperature profiles at 0346 CST
at station 141 and 0358 CST at station 142, it can be seen that the urban
region was warmer than the surrounding rural regions in the layer from the
surface to the 400-m level.
A shallow superadiabatic layer developed from the surface to the 200-m
level by 0945 CST in the rural regions. Weak stability was found in the layer
from 200 m to 2.0 km, and strong stability was noted above the 2-km level. At
this time, the urban region was warmer than the rural region in the layer from
the surface to about the 1.0-km level.
113
-------
ST LOUIS UPPER AIR STATION 141
6 JUNE 1976
ST, LOUIS UPPER AIR STATION 142
8 JUNE 1976
4 r
3 -
<2 2
UJ
X
RIVER
or
UJ
(E
UJ
>
RIVER
0358 CST
0945 CST
1548 CST
2151 CST
15
280
285 290 295 300 305
POTENTIAL TEMPERATURE (K1
280 285 290 295 300 305
POTENTIAL TEMPERATURE (K°)
3IO
Figure 44. Diurnal variability of the potential temperature [°K] profile on
8 June over the St. Louis center city (Station 141) and over
the rural area southwest of St. Louis (Station 142). Figure 37
shows the exact locations of these two sounding stations.
At 1548 CST, a shallow superadiabatic layer existed from the surface to
the 200-m level in contrast to the weak stable layer found in the same region
over the urban area. No explanation could be found for the difference.
Perhaps mixing coupled with surface heating had momentarily removed the stable
layer at station 142 and produced the superadiabatic layer. Above the 200-m
level, a quasiadiabatic region was noted up to 1.7 km with strong stability
above that level. At this time, the urban region was warmer than the rural
region up to a height of 2.2 km.
A shallow layer of strong stability had developed by 2151 CST in the
rural regions from the surface to the 200-m level. Above 200 m, the boundary
layer was quasiadiabatic up to 1.7 km and weak to strong stability was found
above that region. At this time, the urban heat island extended up to approxi-
mately 600 m according to the radiosonde data.
114
-------
The wind speed profile over the center of the city (station 141) indi-
cated the presence of a low-level jet above the inversion top (200-m level) at
0346 CST (figure 45). The wind direction (figure 45) through the low-level
jet was fairly uniform with winds out of the west. Above 800 m, the wind
direction changed markedly, retaining some uniformity in the layer from 1.2 km
to 2.4 km with winds from the south to southwest. In this layer, a secondary
speed maximum was found at the 2.1-km level.
The low-level jet was also evident over the rural region at 0358 CST
(figure 46) but was centered at a higher elevation (300-m level). The wind
direction (southwest) was also uniform through the layer where the low-level
jet was found in the rural region (figure 46). Above 700 m, wind direction
underwent marked changes as it did over the urban regions. There was a layer
from 1.2 km to 3.8 km of relatively uniform wind direction (south to south-
east). In that layer, a secondary wind speed maximum was noted from 2.4 to
2.9 km.
At 0949 CST, the low-level jet was still present over the rural region.
It is suggested that the relatively high wind speeds found at low levels over
the urban areas were not remnants of the low-level jet but were the results of
pressure gradient accelerations associated with the urban heat island circu-
lation.
There was a change in the wind direction in the surface layer around 1000
CST. It is noted that the wind direction over the urban area at 0945 CST was
rather uniform from the surface to the 1.0-km level with winds out of the
northwest. In the same layer over the rural region, wind direction varied
from north to northwest to northeast. A number of changes in the wind direc-
tion occurred above the 1.0-km level over both regions, but there was a layer
of relatively uniform wind direction from the south to southeast from 1.2 to
2.8 km.
Wind speeds were small at all levels over the urban and rural regions at
around 1545 CST. In the urban region, the strongest wind speeds were found
near the surface, suggesting effects associated with the heat island circula-
tion. The wind direction over the urban and rural regions underwent marked
changes with height. The region of relatively uniform wind direction was
noted over the rural region in the layer from 400 m to 2.4 km. In that layer
the wind direction was from the south. No such layer was evident over the
115
-------
ST LOUIS UPPER AIR STATION 141
WIND SPEED (mps)-8- JUNE 1976
3r-
> 2
t-
x
RIVER
3r
2 -
I -
2 4 6 8 10
0346 CST
3r
2 -
2 -
2468
1545 CST
2 4 6 8 IO
2147 CST
ST. LOUIS UPPER AIR STATION 141
WIND DIRECTION-8-JUNE 1976
3r
E
tr
^-
i i
RIVER
3r
00 09 18 27 36
0346 CST
00 09 18 27 36
0949 CST
00 09 18 27 36
1545 CST
00 09 18 27 36
2147 CST
Figure 45. Diurnal variability of the wind speed [m/s] and wind direction
profiles over the St. Louis center city on 8 June.
116
-------
RIVER
ST LOUIS UPPER AIR STATION 142
WIND SPEED (mps)- 8-JUNE 1976
2 4 6 8 10
0358 CST
2 4 6 8 10
0945 CST
246 8 IO
1548 CST
2 4 6 8 10
2151 CST
ST. LOUIS UPPER AIR STATION 142
WIND DIRECTION - 8-JUNE 1976
6
— 2
0009 18 2736
0358 CST
3 T
0009 18 27 36
0945 CST
00 09 18 27 36
1548 CST
00 09 18 27 36
2151 CST
Figure 46. Diurnal variability of the wind speed [m/s] and wind direction
profiles over the rural area southwest of St. Louis on 8 June.
117
-------
urban region. It is interesting to note that over the urban region, the wind
direction at the 1.0-km level (winds from the east) was almost 180° out of
phase with the wind direction at the surface (winds out of the west southwest).
Later in the evening (around 2147 CST), wind speeds increased at most
levels over both the urban and rural locations relative to the 1545 CST pro-
files, and were relatively uniform above the surface to about the 3.5-km
level. The wind direction was relatively uniform over both the urban and
rural regions in the layer from the surface to approximately the 2.6-km level.
In this layer the winds were from the southwest. Above that layer the wind
direction underwent marked changes.
5.6 Simulation Results
The analysis of the observed data yields at most a two-dimensional view
, jo
of the flow. A primitive equation model was used to simulate the time-
dependent behavior of the heat island and the heat island circulation in order
to obtain some insight into the three-dimensional structure of the flow over
the city associated with the heat island circulation and into the effect of
the perturbation on the local air pollution distribution in the period 0900 CST
to 2100 CST when Da Vinci II was in close proximity to the city. Initially
for the model, the boundary layer was adiabatic from the surface to the 1.0-km
level. Above the 1.0-km level, the temperature lapse rate was identical to
the standard atmosphere lapse rate. The initial wind was zero at the surface
and was 2.0 m/s through the adiabatic layer, increasing to approximately 5 m/s
at the 4.0-km level (the top of the model). The initial wind direction was
from the west throughout the domain of the model, which generally characterizes
the geostrophic wind direction at low levels (below 1.0 km) through the period
of interest. The maximum allowable temperature perturbation at the surface
due to the urban heat island in the daytime period (the period from 0900 to
1500 CST, which marked the beginning of the differential cooling) was 2.4° C.
Figure 47 shows the near-surface, simulated flow reminiscent of the
period around 1000 CST on 8 June. Though the initial wind direction was from
the west, frictional effects produced a general flow from the southwest.
Figure 47 shows the strong zone of convergence over the central area of the
city and a region of inflow over and downwind of the city associated with the
urban heat island circulation similar to that shown in the analysis of the
RAPS wind data for 1000 CST (figure 42). The strongest wind speeds were found
near the central regions of the city upstream of the center of convergence,
118
-------
100m
1,000 m
600 m
Figure 47. Simulated horizontal flow at 1000 CST is shown at 100 m above the
surface of the river (upper left) and at 1,000 m above the river
(upper right). Vectors are proportional in length to wind speed
and extend from the grid point in the direction of the flow.
Horizontal distribution of the vertical velocity [m/s] at 1000 CST
is shown for the 600-m level (lower left) with a section line
along which a vertical cross section of the vertical velocity profile
is shown (lower right). Positive values indicate upward motion.
119
-------
where the pressure gradient accelerations associated with the urban heat
island were in the direction of the synoptic flow.
The magnitude of the simulated wind vectors were not precisely comparable
to those shown in the 1000 CST analysis of the observed data (figure 42).
This was due in part to the fact that the interpolation routine used to create
the analysis of the RAPS wind data produced considerable smoothing (actual
wind speeds near the region of convergence were as large as 3 to 4 m/s in many
cases). Furthermore, the simulated flow in figure 47 is characteristic of the
60-m level over the city rather than the 10-m level, as is the RAPS data.
At the 1.0-km level, which was not influenced by surface friction, the
flow remained from the west (figure 47). Over the central part of the city
the flow was characterized by a strong directional and speed divergence.
Linking the low-level flow with the upper-level flow was a zone of upward
vertical motion centered about the city with a maximum speed on the order of
0.5 m/s (figure 47). The top of the urban heat island circulation (the region
where the vertical velocity approximately goes to zero or is negligible) was
in the vicinity of the 2.0-km level (figure 47).
At this time, the temperature perturbation near the surface according to
the simulated results was approximately 1.0° C. Figure 48 shows an extension
of the heat plume to the northeast similar to that shown in the observed
results. The temperature cross section indicates that the top of the heat
island was at about the 400-m level (figure 48). The depth of the heat island
is generally smaller than that for the heat island circulation.
Figures 49 and 50 depict the simulated results for the period around 1300
to 1500 CST. Figure 49 shows the near-surface flow, which is characterized by
strong convergence near the center of the city, but the inflow characteristics
shown in the solution for the 1000 CST period are not as evident in this
solution. The 1.0-km level was characterized by strong divergence over the
center of the city (figure 49). Note that the winds at the 1.0-km level over
the downtown area were about 180° out of phase with the near-surface flow in
the same region, a factor noted in the observed data. Positive vertical
velocities were centered over the central portions of the city with magnitudes
on the order of 1.0 m/s (figure 49). Top of the heat island circulation was
still found around the 2-km level (figure 49).
The near-surface temperature perturbation associated with the urban heat
island was on the order of 1.5° C in the central area of the city (figure 50).
120
-------
100m
1 '•*
I „
HORIZONTAL DISTANCES Ikrnl
Figure 48. Horizontal distribution of the potential temperature [°K] perturbation
at the 100-m level above the river surface at 1000 CST on 8 June (above).
Below is the vertical cross section of the potential temperature [°K] field
constructed along the section line shown on the horizontal distribution
map.
121
-------
100 m
1,000 m
100m
600 m
Figure 49. Simulated horizontal flow at 1400 CS7 is shown at 100 m above the
surface of the river (upper left) and at 1,000 m above the river (upper
right). Vectors are proportional in length to wind speed and extend
from the grid point in the direction of the flow. Horizontal distribution
of the vertical velocity [m/s] at 1400 CST is shown for the 600-m level
with a section line along which a vertical cross section of the vertical
velocity profile is shown (lower right). Positive values indicate upward
motion.
122
-------
HORIZONTAL (HITANCI (km)
1,000 m
Figure 50. Horizontal distribution of the potential temperature [°K] perturbation
is shown at the 100-m level above the river surface at 1400 CST on
8 June. In the upper right is the vertical cross section of the potential
temperature [°K] field constructed along the section line shown on the
horizontal distribution map. Below is the horizontal distribution of the
potential temperature [°K] perturbation at the 1,000-m level above the
river at 1400 CST.
123
-------
There is an extension of the heat plume to the northeast. At the 1.0-km level
(figure 50), temperature perturbation was on the order of 0.5° C. It is
interesting to note that the heat plume at the 1.0-km level extended to the
southeast. This is associated with the strong southeastward flow produced by
the upper level divergence. The temperature cross section (figure 50) suggest-
ed the top of the heat island was approximately at the 1.6-km level.
After 1500 CST, the surface temperature in the model was forced to cool
differentially (the surface temperature in the rural area cooled at a greater
rate than that in the urban area) so as to produce a heat island comparable to
that found in the observed data at 2000 CST. Figure 51 yields the distri-
bution of the simulated winds near the surface at 1600 CST and shows that the
zone of convergence had weakened (stream lines do not unite). At the 1.0-km
level, the divergence aloft had also become weaker. The vertical velocity had
decreased to a value of 0.5 m/s, and the zone of updraft had been displaced
downstream of its position found at 1400 CST. The depth of the heat island
circulation remained at approximately the 2.0-km level. (See figure 51.) The
thermal perturbation also began to decrease and move downstream (figure 52).
An inversion had developed over the suburban and rural regions, but not over
the urban region.
The near-surface convergence zone found downstream of the urban area at
1600 CST continued to weaken according to the 1800 CST results (figure 53).
The 1.0-km level was characterized with speed divergence but directional di-
vergence was not evident. The maximum vertical velocity was located over East
St. Louis, Illinois, downstream of St. Louis, and the magnitude of vertical
velocity had decreased substantially to approximately 0.10 m/s. There were
secondary centers of updraft located west of the downtown section of the city
and north and south of the center of maximum updraft. The weaker updraft found
upstream of the city extended to about the 2.0-km level (figure 53), whereas
the updraft center over East St. Louis only extended to the 1.5-km level. The
thermal perturbation associated with the urban heat island (figure 54) had de-
creased substantially but apparently there still was a flux of heat to the
100-m level from the surface since there was an extension of the heat plume
over the city. It would have been expected that if there were no heat flux to
the 100-m level from the city, the heat plume would have moved downstream by
this time, and there would be no extension of the heat plume over the city.
Vertical cross section (figure 54) indicates a weak stable layer from the sur-
124
-------
fir
100m
1,000m
600m
Figure 51. Simulated horizontal flow at 1600 GST is shown at 100 m above the
surface of the river (upper left) and at 1,000 m above the river (upper
right). Vectors are proportional in length to wind speed and extend
from the grid point in the direction of the flow. Horizontal distribution
of the vertical velocity [m/s] at 1600 CST is shown for the 600-m level
(lower left) with a section line along which a vertical cross section of
the vertical velocity profile is shown (lower right). Positive values indicate
upward motion.
125
-------
100m
J.400-
1,200-
1,400
; 1200.
: 1 000-
HORIZONTAL OIITANCI
Figure 52. Horizontal distribution of the potential temperature [°K] perturbation
at the 100-m level above the river surface at 1600 CST on 8 June
(above). Below is the vertical cross section of the potential temperature
[°K] field constructed along the section line shown on the horizontal
distribution map.
126
-------
100m
1,000 m
600 m
Figure 53. Simulated horizontal flow at 1800 CST is shown at 100 m above the
surface of the river (upper left) and at 1,000 m above the river (upper
right). Vectors are proportional in length to wind speed and extend from
the grid point in the direction of the flow. Horizontal distribution of the
vertical velocity [m/s] at 1800 CST is shown for the 600-m level (lower
left) with a section line along which a vertical cross section of the vertical
velocity profile is shown (lower right). Positive values indicate upward
motion.
127
-------
100m
HIZONTAL DISTANCE (km)
Figure 54. Horizontal distribution of the potential temperature [°K] perturbation
at the 100-m level above the river surface at 1800 CST on 8 June (above).
Below is the vertical cross section of the potential temperature [°K] field
constructed along the section line shown on the horizontal distribution map.
128
-------
face to approximately the 150-m level over the urban area with an inversion in
approximately the same layer over suburban and rural regions surrounding the
city.
Simulation results that characterize 2100 CST show little indication of
the urban heat island circulation (figure 55). The wind speed had increased
at most locations over and surrounding the city. Furthermore, there was no
indication of divergence aloft, but there was a center of upward motion (ap-
proximately 2 cm/s) centered over the city (figure 55). The top of the region
of upward motion was at about the 800-m level (figure 55). Stronger stability
had developed over the urban region (figure 56) but the lifting of the isopo-
tentialtherms over and downwind of the city indicate upward mixing of heat.
5.7 Surface Ozone Distribution
At 0500 CST (figure 57), the ozone distribution was fairly uniform to the
west of the city, averaging around 0.03 ppm. In the downtown region, along
the river, and immediately to the east, ozone was barely detectable, which is
indicative of a strong ozone-destructive process that occurs in a surface-
based nocturnal temperature inversion layer in the inner city. Ozone trapped
in such a stable layer would be destroyed by contact with surface objects and
by reactions with ozone-destructive agents such as NO and N02 emitted by
combustion processes and certain hydrocarbons emitted by petrochemical plants
located along the river.
From 0500 through 0800 CST, the lowest ozone in the area consistently was
recorded in the east central portions of the downtown business districts.
Higher readings were noted to the north and west of the city. By 0900 CST
(figure 57), the ozone had become fairly uniform over the study area with
slightly lower readings still occurring over the downtown area. Higher ozone
concentrations were found in an area to the southeast of the city that had
been downwind of the center of the city for most of the morning, suggesting
photochemical generation of ozone from precursors transported downwind from
the city.
After 0900 CST, the wind direction changed from northwesterly to south-
westerly. The location of the highest ozone concentration began to move from
southeast of the city to northeast of the city and, in the period 1000 to
1400 CST, appeared immediately downwind of the center of the city in the
approximate location of the zone of convergence associated with the urban heat
129
-------
100m
1,000m
600 m
i_
•'.' f
Figure 55. Simulated^ horizontal flow at 2100 CST is shown at 100 m above the
surface oft the, river (upper left) and at 1,000 m above the river (upper
right). Vectors are proportional in length to wind speed and extend
from the grid point in the direction of the flow. Horizontal distribution
of the vertical velocity [m/s] at 2100 CST is shown for the 600-m level
(lower left) with a section line along which a vertical cross section of the
vertical velocity profile is shown (lower right). Positive values indicate
upward motion.
130
-------
100m
2400 -
2,200
1000 -
Figure 56. Horizontal distribution of the potential temperature [°K] perturbation at
the 100-m level above the river surface at 2100 CST on 8 June (above).
Below is the vertical cross section of the potential temperature f°K] field
constructed along the section line shown on the horizontal distribution
map.
131
-------
0500 CST
0900 CST
1300 CST
Figure 57. Surface ozone distribution ;Jppm] in the study area at 0500, 0900, and
1300 CST on 8 June. (Dashed line represents .005 ppm.)
132
-------
island circulation (data not shown). Areas of lower ozone concentration were
found upwind of the city. An example of this distribution is shown in figure
57, which is the ozone distribution at 1300 CST.
By 1600 CST, a rather broad urban plume was found along a southwest-north-
east axis through the center of the city (figure 58). At this time, the maxi-
mum ozone concentration was displaced farther downstream than previously. The
downstream displacement of the maximum ozone concentration occurred at about
the same time that the strong convergence characterized by inflow into the
city downwind of the city and associated with the urban heat island weakened
and the inflow region no longer existed. This general pattern for the ozone
distribution persisted, and by 1800 CST the maximum ozone in the plume decreased
from 0.22 at 1600 CST to 0.15 ppm, while the minimum value in the surrounding
suburban and rural regions dropped by about 0.03 ppm to around 0.09 ppm in the
same period. This indicates the presence of more ozone-destructive agents in
the urban plume than in the surrounding suburban and rural regions.
By 1900 CST (figure 58), several distinct features appeared in the pat-
tern of ozone distribution: highs were noted to the northeast and southeast.
The high to the northeast was associated with the still distinct urban plume
stretching out in that direction from the center of the city. The fairly high
ozone noted to the southeast of the city was probably due to the lack of
ozone-destructive agents in that area relative to the western suburban and
rural region.
By 2100 CST (figure 58), the nocturnal destruction of ozone had markedly
affected the ozone distribution. Although the urban plume was still evident,
the suburban areas to the west of the city through the southern downtown area,
and east into Belleville, showed very low ozone concentrations. This was most
likely due to destructive effects of NO and N02 emitted by the rush hour and
early evening traffic into the stable surface layer. Areas to the south and
southeast of the city still showed relatively higher ozone concentration at
this hour.
5.8 Da Vinci II and RTI-EML Ozone Data
In the period 0800 to 2100 CST on 8 June, Da Vinci II and the RTI-EML
collected ozone data in the immediate vicinity of St. Louis. Figure 38 shows
the track of Da Vinci II and the RTI-EML. Figure 59 yields the time variation
of the height of Da Vinci II in the period of interest.
133
-------
1600 CST
1900 CST
2100 CST
Figure 58. Surface ozone distribution [ppm] in the study area at 1600, 1900, and
2100 CST on 8 June. (Dashed line represents .005 ppm.)
134
-------
Da VINCI II GONDOLA ALTITUDE PROFILE
2r-
UJ
Q
ID
I I
J I
I
08 09 10 II 12 13 14 15 1617 18 19 20 21 22 23 00 01 02 08 04 05 06 07 08
8 JUNE 1976 9 JUNE 1976
TIME (CST)
Figure 59. Variation of altitude with time for the flight of Da Vinci II.
Da Vinci II was launched at 0756 CST on 8 June from Arrowhead Airport,
which is approximately 24 km west of St. Louis. It took the balloon approxi-
mately 30 min to reach flight altitude. In the period 0830 to 1030 CST, the
balloon drifted southwestward at an approximate altitude of 700 m MSL. Winds
at flight altitude were from the northwest at approximately 1 to 2 m/s. Be-
cause of severe thermal convection encountered by the balloon at around 1030
CST, the balloon changed flight altitude. The transition took place between
1030 and 1130 CST. From 1130 CST to approximately 1500 CST, the balloon
cruised at an approximate altitude of 1,700 m, where the winds were from the
south to southeast at 2 to 3 m/s and the balloon moved to the northwest. As
the stability began to develop in the boundary layer between 1500 and 1600 CST,
the severe convection ceased, and the altitude of the balloon lowered to 700 m.
The balloon remained at 700 m for the rest of the period of interest (1600 to
135
-------
2100 CST). At this time, the balloon moved to the east as the winds at flight
altitude were from the west at 3 m/s.
Figure 60 shows the data from the Da Vinci II system, the RTI-EML, and
also some additional ozone data derived from the RAPS network. The solid line
is the ozone concentrations obtained aboard Da Vinci II. The long dashed line
is the ozone concentrations obtained by the RTI-EML at the surface. The short
dashed line is the surface concentration of ozone under Da Vinci II determined
using the path of Da Vinci II relative to time and the analyses of ozone data
at the surface derived from the RAPS data. The dash-dot line is the surface
ozone maximum in the plume downwind of St. Louis. The latter analysis was
also derived from the RAPS data.
The data from Da Vinci II indicate that at the time Da Vinci II reached
flight level, the ozone concentrations were approximately 0.080 ppm. The
ozone increased from 0800 CST, reaching a maximum of approximately 0.152 ppm
at 1600 CST. Vertical profiles of ozone obtained at 1145 CST, which will be
given in the next section, indicate that the ozone distribution decreased with
height frgm approximately 1 to 2 km. The fact that the ozone continued to
increase in the time period 1030 CST to 1130 CST, when the Da Vinci II under-
went its altitude changes from 400 to 1,700 m, suggests that some overriding
factor, possibly synthesis, was influencing the time variation of ozone during
that period.
After 1500 CST, the balloon descended. The major portion of its descent
occurred between 1500 and 1600 CST. Afterwards, there was a rapid decrease in
ozone (1700 to 2100 CST). At the time of the rapid decrease of ozone, the Da
Vinci II system was over the Alton-Wood River Industrial Region. It is sug-
gested that the balloon entered into a region where there were significant
amounts of ozone-destructive agents, the apparent source of which was the
Alton-Wood River Industrial Region. Since the decrease in ozone ceased at
1900 CST and since the ozone increased from 1900 CST, reaching a value of
approximately 0.145 ppm at 2300 CST, it is suggested that most, if not all, of
the destructive agents were removed through reaction with ozone by 1900 CST,
and that the increase in ozone that followed was due to mixing with the sur-
rounding environment.
The data from the RTI-EML also indicated that ozone increased initially
reaching a maximum at 1600 CST also. The maximum concentration of ozone was
0.149 ppm. After 1600 CST, the ozone decreased rapidly, reaching a minimum at
136
-------
.20
.15
UJ
z
o
N
O
.10
.05
DaVinci II Gondola (ASL)
RTI Van (Bendix)
Surface under Gondola
Surface maximum in urban plume
\
10 12 14 16 18 20 22 00 02
TIME (CST)
Figure 60. Hourly average ozone concentrations [ppm] measured along the
Da Vinci II flight track, including measurements taken from the
gondola, the RTI van, and the RAPS monitoring network.
137
-------
2000 CST. It was noted earlier that after 1400 CST, a surface-based inversion
began to develop. The inversion would trap ozone-destructive agents emitted
from surface sources. Furthermore, during this period the solar intensity was
decreasing. Evidently a destruction of ozone overbalanced the synthesis of
ozone, causing the decrease in the ozone concentration.
The RTI-EML data indicated a rapid increase, followed by a rapid decrease
in the ozone concentration in the period 2000 CST to 0100 CST. At this time,
the RTI-EML was northeast of St. Louis (see figure 38 and figure 58) in the
approximate location of the remnants of the major surface ozone plume emitted
from St. Louis.
Variation of the ozone concentration at the surface under Da Vinci II as
determined from the RAPS data, indicated a pattern similar to that found in
the Da Vinci II data in the period 0800 to 1500 CST. There was an increase in
ozone in the period, and after 1100 CST, the values at the surface under Da
Vinci II were approximately equal to those observed on Da Vinci II. Maximum
ozone concentration was approximately 0.154 ppm at 1500 CST. From 1500 to
2000 CST, the surface ozone under the gondola had a pattern similar to that
observed by the RTI-EML; namely, the ozone concentration decreased rapidly.
The maximum ozone concentration at the surface in the plume downwind of
the city increased from 0800 CST to 1600 CST. The maximum ozone concentration
observed in the plume was 0.22 ppm at 1600 CST. From 0900 CST to 1900 CST,
the ozone concentration in the plume downwind of the city was in excess of the
values observed by the Da Vinci II system, the RTI-EML, and surface values
under Da Vinci II.
After 1600 CST, the surface maximum ozone concentration in the plume
decreased in a manner similar to that observed by the RTI-EML and similar to
the surface ozone concentration under Da Vinci II. It is interesting to note
that the surface ozone concentration under Da Vinci II approached the maximum
ozone concentration in the plume around 1900 CST, but the RTI-EML was observ-
ing lower concentrations until 2300 CST. Between 2100 and 0100 CST, the RTI-
EML experienced an increase in ozone. It is also important to note that the
ozone concentration in the plume northeast of the city (RAPS station 115) was
considerably smaller than that observed by the RTI-EML in that period. The
RTI-EML was farther to the northeast of the city (see figure 38) at this time.
5.9 Aircraft Ozone Profiles
Two vertical profiles of ozone were obtained by a NOAA aircraft within
138
-------
2.0 km of Da Vinci II on the morning of 8 June (figure 61). The profiles were
based upon ozone concentrations recorded during the ascending portions of the
aircraft missions, the first from 0915 to 0933 CST and the second from 1148 to
1211 CST. Both of these missions were conducted during the time Da Vinci II
was moving rather slowly through an area about 20 km southwest of the center
of downtown St. Louis (figure 38).
The two profiles suggest the mixed layer extended up to about 6,500 ft
MSL. The ozone destruction that occurred within the surface layer during the
previous night was still in evidence at 0915 CST even though the inversion
had, by this time, dissipated. Examination of the two profiles reveals evi-
dence of synthesis of ozone in the layer during the hour and a half between
ascents. Over the period, there was an average increase in ozone in the layer
762-2,286 m of approximately 0.035 ppm. Vertical mixing would produce a
fairly uniform ozone concentration throughout the mixed layer of about 0.145
ppm based on the 0915 CST profile. As the 1148 CST profile clearly shows, the
average ozone concentration in the layers was around 0.180 ppm. These data
suggest that a significant amount of ozone was generated in the layer between
0915 and 1148 CST. Precursors of ozone may have been mixed upward from the
urban region; but it is also possible that they were mixed upward upstream of
the urban region, horizontally transported into the region, and through hor-
izontal mixing, affected the ozone concentration observed by Da Vinci II.
5.10 Summary and Conclusions
The analysis indicated that the surface ozone distribution during the
period 1100 to 1700 CST was affected by an intense urban heat island circula-
tion that existed during that period. It was shown that from approximately
1000 CST to 1400 CST, the highest concentration of ozone was found immediately
downwind of the downtown region of St. Louis over the East St. Louis, Illinois,
region and parts of north St. Louis. It was also shown that a convergence
zone characterized by wind reversals and inflow associated with the urban heat
island circulation was immediately downwind of the city, and in the same
location where the highest concentration of ozone was found. This suggests
that the convergence prevented rapid downwind transport of ozone and ozone
precursors.
After 1500 CST, the heat island circulation began to dissipate. Dif-
ferential cooling intensified the urban heat island, but also produced a
surface-based stable layer that was responsible for the weakening and eventual
139
-------
CO
<
h-
U-
o
o;
1211 CST
1148 CST
.05
10
.15
.20
OZONE CONCENTRATION [PPM]
Figure 61. Vertical ozone profiles flown by a NOAA aircraft in the vicinity
of the Da Vinci II gondola on the morning of 8 June 1976. Sur-
face ozone concentrations measured beneath the gondola are
also plotted and are projected up to the aircraft level.
140
-------
dissipation of the urban heat island circulation. During the dissipation
stage, the near-surface zone of convergence found immediately downwind of the
city dissipated in stages. Initially, the region of inflow over and downwind
of the city was no longer apparent. In the final stages of dissipation, no
directional or speed convergence was found at the surface and the results of
the primitive equation model suggested that there were no perturbations aloft.
During this period the highest ozone concentration in the plume moved signifi-
cantly farther downstream than previously. Dissipation of the urban heat
island circulation and the accompanying convergence zone allowed the ozone
precursors and ozone to be transported farther downstream.
Movement of Da Vinci II was governed by changes in the wind direction
associated with synoptic weather changes and with changes in the flight level
of Da Vinci II. At approximately 1000 CST, wind direction at the surface
changed from northwest to southwest. From 1100 CST and until 1500 CST, Da
Vinci II was located at approximately the 1,700-m level where radiosonde data
indicated that the winds were from the south to southeast.
However, even if Da Vinci II could have moved eastward toward St. Louis,
the results from the primitive equation model suggested that the strong diver-
gence field over the city associated with the urban heat island circulation
would not have allowed Da Vinci II to pass directly over the city. If it had
moved eastward, Da Vinci II would only have come in contact with the major
urban plume from the city if it passed south of the center of the city.
Strong upward motion in the center of the surface-based convergence zone would
have transported air pollutants aloft. Using the heat perturbation as a
tracer, the results indicated that the upper plume would have moved south-
eastward immediately downstream of the city. The surface observations and
simulated results indicated the major surface plume moved to the northeast.
This suggests that the major upper plume and the major surface plume could
have been as much as 90° out of phase in their respective directions of mo-
tion.
It was also noted that during the period 0900 to 1700 CST, the ozone
concentration observed by Da Vinci II increased. Since Da Vinci II is a
Lagrangian system, changes in the ozone concentrations measured aboard Da
Vinci II can only be due to turbulent transport, synthesis, and destruction of
ozone and only turbulent mixing and synthesis could account for increases in
the ozone concentration. Existing vertical profiles of ozone (figure 61)
141
-------
indicate the region above 2.0 km to have lower concentrations of ozone than
that observed at the flight level of Da Vinci II (below 2.0 km). The theory
of turbulent transport indicates that under those circumstances, if there were
mixing between the region above 2*. 0 km and the region around the flight level
of Da Vinci II, ozone would be transported from the region below 2.0 km to the
region above -2.0 km, and the region below 2.0 km would experience a decrease
in the ozone concentration, if there were no means of compensation for the
effects of this particular exchange process, which is in contrast to what was
observed.
From 0800 to about 1200 CST, the ozone concentration at the surface was
smaller than that at the flight level of Da Vinci II (figure 60). In this
case, turbulent flux of ozone would be downward, and mixing would increase the
ozone concentration at the surface at the expense of the ozone aloft; i.e.,
the ozone concentration at the flight level of Da Vinci II should decrease.
This again is in contrast to what was observed. From about 1300 to 1700 CST,
the ozone concentrations at the flight level of Da Vinci II were, for all
practical purposes, equal to that at the surface. Therefore, the turbulent
flux (proportional to the vertical gradient) was zero, or any mixing of ozone
between the surface and the flight level of Da Vinci II could not influence
the ozone concentration one way or another.
These arguments indicated that in no way could mixing of ozone account
for an increase in the ozone concentrations aloft. This only leaves synthesis
as the potential mechanism. Precursors may come from the city because strong
vertical convection occurred during the daytime period over St. Louis, or may
be mixed upward from surface sources upstream of the city and transported
horizontally over the city. Furthermore, a great percentage of the ozone
observed at the time of maximum ozone must be considered background ozone
since 0.080 ppm ozone existed at flight level when Da Vinci II reached it.
After 1500 CST, the cooling process began. Differential cooling in-
creased the intensity of the urban heat island. It also developed a surface-
based stable layer over the city and the rural region. It was noted that a
surface-based stable layer was present over the city at 1545 CST. Observed
data and the results using the primitive equation model indicated that though
the urban heat island intensity increased, the urban heat island circulation
v
weakened and dissipated due to the surface-based stability, which was the
overriding factor influencing the urban heat island circulation. Stability
142
-------
also trapped ozone-destructive agents near the surface causing the ozone
concentration at the surface to decrease rapidly in the vicinity of the RTI-
EML and under Da Vinci II; but the ozone concentrations aloft remain at levels
above the NAAQS.
The ozone concentration observed by Da Vinci II decreased markedly in the
period 1800 to 1900 CST. Though the change was great, the concentration of
ozone remained at a level above the NAAQS. Da Vinci II was over the indus-
trial regions of Alton-Wood River at this time, and it is suggested that the
gondola may have lowered into the remnants of a plume from that region or
there was injection of air pollutants from smoke stacks that penetrated through
the surface inversion into the air mass that influenced Da Vinci II. Ozone
concentration observed by Da Vinci II increased as Da Vinci II moved farther
downstream, suggesting mixing with the surrounding environment.
At 2100 CST, Da Vinci II and the RTI-EML moved into a region where rem-
nants of a major surface ozone plume from St. Louis may have been located;
i.e., northeast of St. Louis. The ozone data from the RTI-EML suggested that
it was in a region of elevated ozone at the surface between 2100 CST and
0100 CST. No such indications were evident in the Da Vinci II data. The
radiosonde data indicated that the stability would have trapped the surface
ozone plume in the layer below 200 m (the top of the inversion). In any case,
the ozone data from the Da Vinci II and the RTI-EML indicated that the mixing
would have been downward since the ozone concentrations aloft were greater
than that in the plume at the surface. The fact that the ozone concentration
in the plume was significantly lower than that observed by the RTI-EML suggests
that downward mixing through a break in the inversion was responsible for the
elevated ozone observed by the RTI-EML.
Analysis of the data and the simulation results indicate a complicated
structure for the three-dimensional ozone distribution over the city of St.
Louis. This structure may also exist for other pollutants. Because of this
complex structure, changes in the synoptic weather distribution, and changes
in the flight level of Da Vinci II, Da Vinci II never came in contact with the
major air pollution plume emanating from the city of St. Louis when it was in
the vicinity of the city.
143
-------
-------
6.0 METEOROLOGICAL ANALYSIS FOR DA VINCI II
In order to understand the atmospheric chemistry measurements made aboard
Da Vinci II, it is also necessary to understand the physical environment where
those measurements were made. The environment can be examined in several
ways: At a given time and altitude, what is the history of the air which
currently surrounds Da Vinci II? Has that air been in the St. Louis urban
area? Was it affected by emissions from St. Louis? How does the understand-
ing of the meteorological processes, as revealed by measurements, help to
interpret the air chemistry measurements? These general questions have been
investigated by attempting to define the three-dimensional plume emanating
from St. Louis, using a trajectory analysis approach. That approach is neces-
sary since there were no measurements normal to the airflow at Da Vinci II
altitudes. Secondly, the vertical structure of the atmosphere along the
flight track was examined for the effects of wind shear and atmospheric sta-
bility, which can be derived from standard measurements. The atmospheric
stability of the lowest kilometer of air was examined by analyses of the
acoustic sounder records made during the field program. These analyses are
interpreted to assess the impact of vertical diffusion and transport upon
contents of the air. Finally, in recognition of the complex behavior of the
atmospheric boundary layer, a simple one-dimensional model of that layer,
developed by other investigators, was tested and applied to a portion of the
Da Vinci II flight.
6.1 Plume Identification
6.1.1 Introduction
The "urban plume" is an analogy to the isolated plume emanating from a
point source of material into the atmosphere. The concept is that the total
plume, composed of the combination of area and point sources of all types of
pollutants, behaves as an isolated plume would behave. This is not totally
the case because of the multiple scales of motion affecting the urban emis-
sions. When dealing with the concept of an urban plume, it is extremely
important to recognize that the transport wind speed and direction varies
horizontally, vertically, and with time. Consequently, emissions from one
portion of the city may take different paths than emissions from another part
of the city, while air at one altitude will take a different direction and
speed than the air at another altitude. Because of this inhomogeneous motion,
145
-------
the emissions from the urban area may spread over a large depth and width of
the atmosphere and be at different locations at different altitudes although
they were initially within the same column of air. Nevertheless, individual
plumes within the urban plume may maintain their identity.
Da Vinci II did not operate in a totally Lagrangian mode. Soon after
launch it became apparent that controlling the altitude of the balloon was far
more difficult than anticipated for this particular situation. After the
onset of strong convection, the balloon was permitted to seek its own altitude.
Later in the day it was maneuvered to near the planned altitude. During the
evening and into the next morning, Da Vinci II remained at a relatively stable
altitude between 700 and 850 m. Because of the changes in altitude and the
wandering path, it is difficult to appraise whether Da Vinci II was actually
in the effluent of the city of St. Louis during the flight. The multiple
level analysis presented attempts to reconcile the movement of Da Vinci II at
various altitudes within the limits of the available data.
The objective of section 6.1 is to develop streak lines of air movement
for up to 24 h at five altitudes--10, 100, 500, 1,000, and 2,000 m--beginning
at 0700, 1300, 1900, and 0100 CST hours on 8-9 June for selected locations
around St. Louis. This analysis will help to define the three-dimensional
boundaries of the St. Louis "urban plume." These objectives were slightly
modified in terms of altitudes, beginning times, and duration to make the
analyses more pertinent to defining the three-dimensional plume.
6.1.2 Data Sources
Single theodolite pilot balloon (pibal) observations were taken at 17
different locations, 42 different times before and during the flight of Da
Vinci II. (See figure 62 and table 15). The two mobile crews taking the
observations tried to anticipate the movement of Da Vinci II and to be in the
position to make a sounding when Da Vinci II moved over. These data were
taken to assist with flight operation and with postflight diagnostic studies.
Because of the unusual flight tract, the pibals were not always taken in the
immediate vicinity of the gondola. These wind data at 50-m increments above
the ground were obtained through Mr. John Bujnoch, the meteorologist for the
Da Vinci II field program.
6.1.3 Method of Computation
A streak line is the instantaneous locus of points at time tf (air par-
146
-------
Table 15. Pilot balloon observations
Date
6/7/76
6/7/76
6/7/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/8/76
6/9/76
6/9/76
6/9/76
6/9/76
6/9/76
6/9/76
Release #
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Time
(CST)
0242
2330
2345
0000
0030
0100
0130
0300
0430
0500
0530
0600
0610
0630
0700
0700
0730
0730
0801
0830
1022
1025
1120
1130
1223
1300
1600
1646
1730
1730
1820
1830
2000
2020
2115
2340
0200
0230
0440
0620
0630
0720
Location Map
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Jet. 170 and Lucas and Hunt Rd.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Jet. 170 and Lucas and Hunt Rd.
Jet. 170 and Lucas and Hunt Rd.
Arrowhead Airport, Mo.
Arrowhead Airport, Mo.
Bi State Airport, 111.
5 Miles South of Dupo, 111.
Fairview Heights, 111.
5 Miles South of Dupo, 111.
Fairview Heights, 111.
3 Miles North of Red Bud, 111.
St. Libory, 111.
Jet. Route 61 and P, Mo.
Old Monroe, Mo.
Bowling Green, Mo.
Grafton, 111.
Grafton, 111.
Bowling Green, Mo.
Bowling Green, Mo.
Grafton, 111.
Grafton, 111.
Edwardsville, 111.
Greenville, 111.
Vandalia, 111.
Flora, 111.
Flora, 111.
Samsville, 111.
Flora, 111.
Site #
1
1
1
1
1
1
1
1
1
1
1
1
2
1
1
2
2
1
1
6
5
9
5
9
11
12
4
7
17
8
8
17
17
8
8
10
13
14
15
15
16
15
147
-------
MO
Figure 62. Pilot balloon launch sites.
PkUME flREflS 650M MSL BEGIN 0800 CST
Figure 63. Typical analysis of air leaving St. Louis during Da Vinci II flight.
Da Vinci II positions every third hour are shown.
148
-------
eels), having passed a fixed point at some previous time. Mathematically
stated, the streak line position vector, S, is given by
t
-> s- ->-
S = / V dt
'f
where V is the velocity experienced by the parcel. A trajectory is the path
taken by a given point (air parcel) to achieve its present position. Mathe-
matically stated, the trajectory position vector, T, is given by
T = I V dt
t
o
At a time, tf, the position vectors of a parcel leaving the fixed point at t =
t are equal in length and opposite in direction, T = - S.
Given the type of available data, a trajectory is easier to compute than
a streak line. Streak lines can be determined from the trajectories. The
position of points at a given time having a common space origin but different
"starting times" can be found from the trajectory analysis. The locus of
those points is the streak line. In this analysis, the trajectory approach is
generally taken so that the fate of the air in the St. Louis urban area at a
particular time and altitude can be examined.
The trajectories at 50 m and 300 m above ground level and at 650 m, 1,000
m, and 1,400 m MSL were computed for air parcels leaving four locations approx-
imately surrounding the metropolitan St. Louis area (see figure 63). Those
locations form a rectangle that encloses many primary emission sources of the
St. Louis urban area. The western and southern points coincide with the
release point and southernmost points of the flight track.
The computational procedure for a trajectory began by specifying the
latitude and longitude, altitude, and time of a given air parcel. For that
altitude, a u and v component of the wind were interpolated to that location
using the Barnes' space and time interpolation technique (appendix D). The
parcel is then moved with that speed such that the latitude, A4>, and longitude,
AA, changes (in radians) are given by:
A* = uAt/R£
AA = vAt/(R,, cos *)
t
149
-------
where R_ is the radius of the earth (6.4 x 106 m). This procedure allows for
convergence of the meridians and the change of u and v with location on the
map. Throughout the computations, an Albers equal area projection on specifi-
able scale was used to convert from spherical to Cartesian coordinates. For
each altitude, each trajectory was integrated for 18 h beginning at 0800,
1100, 1400, 1700, 2000, and 2300 CST hours, 8 June. These computations were
performed using k = 38 km, v = 1.5 h, y = 2.0, and V.,, = -5.0 (parameters of
Barnes1 method) at a scale of 1:5,000,000. These parameters were chosen so
that two or more observations would contribute to each interpolated value.
Consequently, the procedure smoothed the analyses in data-rich areas to obtain
analyses in data-sparce areas.
6.1.4 Results
The trajectory computations were carried out to show these features of
the air flow:
1. the effects of wind shear on a plume,
2. the identification of a low altitude plume, and
3. the position of the plume relative to Da Vinci II at the altitude of
Da Vinci II.
A typical interpretation of the computations is given in figure 63. The
States of Missouri, Illinois, and Indiana are outlined. The Da Vinci II track
is indicated by a heavy line with the hourly or three-hourly Da Vinci II
position indicated by an inverted triangle. The positions of the air moving
from the four locations are shown every third hour. The air that had previous-
ly been over St. Louis is indicated by the striped area. The effects of the
convergence (divergence) and deformation of the wind field are easily shown as
the area changes size and orientation. There was no attempt to add a diffusive
spreading to the analyses, although diffusion certainly occurs. The air path
is an estimate, not an exact determination of position. The errors of position
are proportional to errors of the wind interpolation. If the wind speed is
off by 1 m/s, a positional error of 3.6 km occurs after 1 hour. The Da Vinci
1 Q
II position error is known to about 2 km. Thus the typical error in the
parcel displacement is about equal to the expected diffusive component.
6.1.4.1 Wind Shear
The effect of wind shear upon the dispersal of the urban air at launch
time is clearly shown in figure 64a. In the lower altitudes, the flow is
150
-------
(a)
Figure 64. Dispersion of air, initially, over St. Louis at indicated
altitudes (km) and times, resulting from wind shear.
151
-------
2000 CST
(C)
Figure 64. (con.)
Figure 65. Movement of air at 50 m from St. Louis between 0800 and 2000 CST,
8 June 1976 and for 18 h of transport. Envelope of air at 50 m pass-
ing over St. Louis between 0800 and 2000 CST, 8 June, and transported
for 18 h.
152
-------
generally from the northwest, but at 1.4 km, the flow is from the southwest.
At 0800, the wind speeds were low at all levels, so the directional shear was
predominant in dispersing the air over St. Louis. Since the winds were dis-
organized, the thermally induced motion played a stronger role in the dis-
persion of the pollutants. The change in the thermal structure became most
noticeable as the nocturnal inversion was lifted and destroyed by vertical
turbulence (and diffusion) arising from surface heating. Thus the surface-
based emissions were easily mixed through a deepening surface layer. The
development of the nocturnal low level jet moved the 650-m air much farther
than the air at 50 or 1,000 m, emphasizing the effect of wind speed shear.
The relatively weak, disorganized flow persisted into the early after-
noon. The analysis of figure 64b shows the shearing and dispersion of the
column of air over St. Louis at 1400 CST. By 2000 CST (figure 64c), the flow
became more organized throughout the column. The air at all levels (below
1,400 m) moved toward the northeast. The column of air was not so strongly
subjected to shear by direction as by speed (figure 64c).
These results agree with the cross-sectional analysis (section 6.2) and
emphasize the effects of horizontal and vertical shear of the wind upon the
dispersal of the St. Louis urban plume.
6.1.4.2 Low Altitude Plume
The air nearest the ground is expected to contain the highest concen-
tration of pollutant species and affects the greatest number of people. The
superposition of the trajectory analyses at 50 m beginning at 0800 CST and
extending through 2000 CST of 8 June (figure 65) clearly shows the area of
impact of the emissions from St. Louis. The emissions nearly saturate the
area just to the east and northeast of the St. Louis urban region because of
the low wind speeds during the first 8 h of the flight. The RTI-EML main-
tained visual contact with Da Vinci II and stayed within about 2 km of the Da
Vinci II position. The RTI-EML was subjected to only a small portion of the
major urban plume between 0800 and 2000 CST.
After 1600 CST, Da Vinci II began to move eastward. The RTI-EML followed
closely. The emissions associated with the city moved east northeasterly
until the RTI-EML appeared to encounter that air about 2300 CST and remained
in it until approximately 0100 CST. Thereafter, the low altitude movement of
the air was slightly slower and from a more northerly direction than the air
aloft, thereby moving from the area of operations of the RTI-EML.
153
-------
6.1.4.3 The Plume at Altitude
The Da Vinci II balloon-gondola system is very nearly a Lagrangian plat-
form insofar as horizontal motion is concerned. Turbulent eddies with sizes
of the order of Da Vinci II or larger will affect the balloon motion. Just
after launch, Da Vinci II rose above the level of thermally induced turbulence
near 500 m. As the depth of mixing increased, overtaking Da Vinci II, flight
altitude became more difficult to control and maintain. During this time Da
Vinci II drifted southeastward about the outskirts of the urban emissions, the
ozone concentration increased with time at altitude and at the ground, and the
air at 650 m moved southeastward as indicated in figure 66a.
By 1100 CST, turbulence had increased to such intensity that Da Vinci II
was permitted to rise to about 2 km where the turbulence was less intense and
less altitude control was required or exercised. Da Vinci II drifted northward
during the next 3 h. Its relationship to the air (at 1.4 km) over St. Louis
during that time is shown in figure 66b. Pilot balloon data did not extend
above 1,500 m for logistical reasons; although Da Vinci II went to higher
altitudes.
Over the next 3 h (1400 to 1700 CST), Da Vinci II was generally descending
to 600 m. Choosing the 1 km winds as representative of the mean winds encoun-
tered during this period, the air initially over St. Louis at 1400 CST moved
northeastward, roughly paralleling the Da Vinci II horizontal motion (figure
66c).
Between 1700 and 2000 CST, Da Vinci II, operating in the 800- to 500-m
altitude range, began a slow eastward motion that gradually accelerated. The
analysis (figure 66d) indicates that the air at 650 m over St. Louis moved
northeastward, so that the midpoint of the emissions area was co-located with
Da Vinci II at 2000 CST.
For the remainder of the flight, Da Vinci II stayed near 650 m and the
air that was over St. Louis between 2000 and 2300 CST moved with Da Vinci II
(figure 66e).
6.1.5 Interpretation
There are two apparent contradictions in this analysis. During the
1700-2000 CST period, the air at 650 m moved northeastward while Da Vinci II,
near the same altitude, moved eastward. If Da Vinci II is a Lagrangian plat-
form, then horizontal motion cannot bring the air from St. Louis into the Da
Vinci II path, as the analysis indicates. Thus either Da Vinci II is non-La-
154
-------
05 (0
155
-------
§
tO
to
I
B>
156
-------
grangian or the analysis is incorrect, or both. Furthermore, the analysis
suggests that the air at 650 m is moving about 1 m/s faster than Da Vinci II,
which is very near the same altitude. This appears to contradict the Lagrang-
ian characteristics of Da Vinci II.
The probable reason for these contradictions comes from the pilot balloon
observations, their spatial and temporal variations, and the interpolation
procedure for the winds. Single theodolite measurements require assuming a
constant rise of the pilot balloon, so that altitude is directly proportional
to time since release. Studies of sequential wind profiles with double theod-
olite measurements of pilot balloons show a variable rise rate and, more im-
portantly, show that significant changes in speed and direction occur between
19
observations. Both of these factors play a role in the discrepancy. Pilot
balloon observations (single theodolite) made southwest of St. Louis (RAMS
site 142) and in downtown St. Louis (RAMS site 141, see figure 37) between
1800 and 2300 CST show a substantial difference in the wind direction at 650 m
MSL. (Those observations were not included in the analysis.) In the downtown
measurements (table 16), the wind was principally from the west-southwest
while the southwest measurements indicated a southerly wind. This type of
variability is difficult to incorporate into an objective analysis over several
hours and tens of kilometers. Effectively, the wind data are not dense enough
to describe the air motion at the same scale as Da Vinci II. The 1 m/s differ-
ence of the Da Vinci II speed and the interpolated wind speeds gives a real-
istic estimate of the resolution of the interpolation procedure.
The rise and fall of the altitude of Da Vinci II during the nocturnal
portion of the flight may also contribute to the discrepancy betwen analysis
and observation. During most of this portion of the flight, Da Vinci II was
above the 650-m level and, according to the cross section analysis, experi-
enced slower air speeds, just above the developing low level jet.
6.1.6 Conclusions
This analysis has identified the probable location of the envelope of the
area containing emission from the St. Louis urban area, the St. Louis "plume,"
at selected times and altitudes. These analyses suggest that Da Vinci II did
not encounter a significant portion of that envelope during the daytime and
that Da Vinci II may have encountered some of the plume during the nighttime
portion of the flight.
157
-------
Table 16. Winds at 650 m MSL from RAMS pilot balloon stations
Time
(CST)
1800
1900
2000
2100
2300
Location
Downtown
direction/speed (m/s)
221/3.1
254/2.3
260/4.4
253/4.1
249/4.2
Southwest
216/1.5
172/1.1
170/2.2
183/3.5
194/4.5
There is insufficient evidence from this portion of the analysis to
identify the cause of the increase in surface layer ozone concentration during
the period 2200 to 0100 CST.
6.2 Cross Section Analyses
6.2.1 Introduction
In the lower atmosphere, two predominant processes affect the vertical
distribution of atmospheric variables and in particular the distribution of
pollutants near the ground. A continual downward flux of momentum replaces
the momentum lost to the ground by friction. This process results in a verti-
cal profile of the wind velocity that generally increases in speed and changes
direction as altitude increases, in the familiar Ekman spiral. The heat flux
at the air-ground interface adds or extracts heat from the atmosphere resulting
from solar heating or radiational cooling. This heat transfer near the ground
plays an important role in the scaling of atmospheric turbulence and thus the
rate at which the momentum and heat are mixed within the lower atmosphere.
By scaling variables, it is possible to show that vertical diffusion of
materials is more significant to the time rate of change of concentration than
is the horizontal diffusion. This is especially true considering a Lagrangian
type measurement platform the size of Da Vinci II.
Evidence is mounting that ozone, generated during the day and distributed
through the mixed layer, is transported long distances at night aloft, sepa-
rated from destruction by ground contact or destructive agents at ground level
by a nocturnal radiative inversion. Concurrently, surface-based emissions of
158
-------
precursors into the radiative inversion layer destroy the ozone there and tend
to accumulate precursors. As the solar insolation increases the following
morning, precursors are available to produce ozone in that lowest layer. When
the inversion layer is raised and dissipated by the insolation, there is an
abundant reservoir of ozone aloft available to mix downward, giving a rapid
increase in surface concentrations. The observed concentrations aloft are not
significantly depleted by mixing the shallow "below inversion" air with the
much deeper reservoir aloft. Since urban areas such as St. Louis are the
principal sources of anthropogenic ozone precursor materials, it is reasonable
to expect that a similar phenomenon might be observed downwind of the St. Louis
area. Da Vinci II monitored the ozone at altitude, the RAMS stations monitored
ozone continuously at fixed ground stations, and the RTI-EML measured ozone
near the ground while following the balloon. Other sections of this report
discuss those data in some detail. The combination of the Da Vinci II data at
altitude and near the ground support the idea of the separated layers of ozone
at night. The data also show some interesting discrepancies from the scenario
outlined above. Therefore, in this section the role of the vertical structure
of the lower atmosphere and its diurnal changes upon the measured concentra-
tions of ozone will be examined.
6.2.2 Analysis Approach
The vertical flux of heat in the St. Louis area is related to time of
day. Therefore, this analytical approach quantifies the vertical structure of
the atmosphere within the column of air containing Da Vinci II at hourly
increments of time along the path of Da Vinci II. Since Da Vinci II did not
move at a constant speed, the distance traveled in these intervals varied sub-
stantially. The position of Da Vinci II at each hour was interpolated from
RTI's time and position data.
The set of 42 pilot balloon observations used to compute the analyses of
section 6.1, a special set of six-hourly upper air observations at the regular
rawinsonde stations of the National Weather Service, and a set of acoustic
sounder data prepared from operations of Dr. E. Miller of Argonne National
Laboratory were the principal sources of data for this analysis.
6.2.2.1 Analysis of Winds Aloft
Pilot balloon data were interpolated to 50-m increments above the ground
along the flight track for each hour during the Da Vinci II flight using
159
-------
Barnes' technique. These data are displayed as wind vectors in figure 67.
The wind vectors are drawn using the standard meteorological connotation,
i.e., showing the direction from which the wind is blowing. The speed is
indicated by the length of the vector. North is at the top of the figure.
During the first hour after launch (0800 CST), the winds in the lowest
300 m increased in speed before decreasing wind direction turned slightly more
northerly through the first kilometer. Thereafter, the winds became quite
light, turning to a southwesterly flow. Over the next few hours, wind speeds
at lower levels continued to decrease while remaining northwesterly, whereas
winds farther aloft became a little stronger from the southwest. Three hours
after launch, Da Vinci II ascended to an altitude well above the operation
level of the pilot balloon data. During the time that Da Vinci II was aloft,
winds below it were quite light and variable.
About 9 h after launch (1700 CST), the wind flow throughout the entire
area became more organized. The winds, initially light and from the southwest,
began to increase in speed from the ground upward. Winds near the very top of
the layer (1.5 km) showed a little more southerly flow. As nighttime pro-
gressed, the winds in the lower altitudes increased in speed and became more
westerly, the greatest increases in speed occurring near 250 to 300 m above
the ground. With the advent of morning (21 h after launch), the lower level
winds continued to turn to a northwesterly and then a northerly flow through-
out the first kilometer. These interpolations, of course, agree quantita-
tively with the trajectory analyses and also with the movement of the balloon,
since the same data set and interpolations were used.
The air flow was disorganized for the first 8 to 10 h of the flight
because of the stagnant anticyclone centered in the St. Louis area. Toward
the later part of the day, the surface high pressure system drifted southwest-
ward, slightly increasing the pressure gradient and developing the flow near
the ground. This turning of the wind and the increase in wind speeds overnight
accompanied a redevelopment of a low level jet just above the nocturnal inver-
sion layer. This jet was clearly evident on the night before the launch and
the remnants of it were apparent in the first 2 h of the Da Vinci II flight.
At 2300 CST (15 h after launch), a substantial increase in wind speeds occurred
in the lowest 150 m; a minimum in the wind speed occurred about 19 h after
launch (0300); and a wind speed increase and wind direction shift (more north-
westerly) occurred during the remainder of the flight. This change of the
160
-------
o
O
or
o
LU
o
or
_j
LU
cc:
CO
a
\/
\\ \ \ \
\ \ \ \ \ 1 1 / / I//////////////'
\\\\\\\ 1 I I/I//////////////
\ \ \ \ \ \
\ \uii\\\\\\\\
\ \ \ \ \
V V V V V V V V V X t •
VV V X V V X
If!
o
'*-•
CD
0)
V)
.a
o
ro
J2
*^
Oi
'a.
E
o
O
(O
a
01
CO
•»->
i—
_O
CO
C
2
o
0)
"D
r*.
(O
m
1.50 1.25 1.00 0.750 0.500 0.250
RLT1TUDE. KM
161
-------
wind characteristic suggests that there may have been an additional momentum
transferred down to the ground near midnight and later in the early morning of
the flight. A similar trend in surface wind speeds and directions was also
observed at Terre Haute, Indiana, Champaign, Illinois, and Decatur, Illinois,
during the nighttime and early morning hours of 9 June. The coincidence of
increase in the wind speed and the higher ozone concentrations at the ground
suggest that a downward mixing of ozone-rich air occurred.
6.2.2.2 Rawinsonde Analyses
A set of special rawinsonde observations were taken in support of the Da
Vinci II flight program at the National Weather Service Offices at the times
given in table 17. These data were made available to RTI as computer print-
outs of the altitude, pressure, temperature, dew point temperature, wind speed
and direction at 50-mb increments and at significant levels. The data were
ordered by altitude and interpolated to 100-m increments from the ground to
3.5 km MSL. The rawinsonde data offers a distinct advantage in understanding
the dynamics of the lower atmosphere because air temperature and humidity are
measured in addition to the wind speed and direction and data are available to
3.5 km of higher. Time-altitude cross sections were developed using the
Barnes' technique outlined in appendix D. These analyses were done for wind
speed, regardless of direction, and for potential temperature, 0. Since 0 is
conserved in adiabatic processes, the vertical gradient of 0 is proportional
to atmospheric stability, and atmospheric stable layers are easily identified.
It is superior to temperature as an analysis variable.
The dominant factor in the resulting analyses was the set of observations
made at Salem, Illinois, approximately 75 km east of St. Louis and about 20 km
south of the midnight position of Da Vinci II. Since the data are more sparse
than the pibal data, the analyses are not sensitive to short-term or local
influences. These analyses do not include rawinsonde data taken in support of
the RAMS program.
6.2.2.2.1 Potential Temperature
The time-altitude cross section of the potential temperature along Da
Vinci II's path is given in figure 68. Throughout the flight, a very stable
layer persisted just below 3 km resulting from atmospheric subsidence in the
high pressure system. It had been evident for the past 3 days in the St. Louis
area. This layer represents the upper limit of any vertical motion that may
have occurred in the planetary boundary layer during that time.
162
-------
Table 17. Rawinsonde data used in analyses
Station
Omaha, Nebr.
Monett, Mo.
Topeka, Kans.
Peoria, 111.
Salem, 111.
Flint, Mich.
Nashville, Tenn.
0600
X
X
X
X
X
X
N/A
June 8
1200
X
X
X
X
X
N/A
X
Time, CST
June 9
1800
X
X
X
X
X
X
N/A
0000
X
X
X
X
X
N/A
X
0600
X
N/A
N/A
X
X
X
N/A
N/A: Not available.
On the morning of the flight, the analysis of figure 68 shows static
stability increasing from the ground upward. The destabilization of the lower
1.5 to 2.0 km of air occurs by midday, as a result of the solar heating. The
destabilization was enhanced by very weak wind patterns within the lower
atmosphere. Without a wind shear to transport momentum downward, the surface
heat flux plays a dominant role in the vertical transport of heat and momen-
tum. Thus, during the afternoon, it appears that a free convection regime
could easily be established. In that circumstance the convection motion is
predominant and vertical mixing takes place very quickly and freely over a
wide area. The analysis suggests that the lowest 2.0 to 2.5 km are well
mixed.
By evening a superadiabatic layer is indicated near 1.5 km. Over the
duration of the night, the feature tends to rise to higher altitudes, retain-
ing its tendency for instability. At these altitudes at night, the atmosphere
loses or gains heat by radiation or advection rather than from surface proces-
ses. It is difficult to envision the physical processes that would maintain a
superadiabatic layer in the absence of a heat source. Thus, since the air was
unstable in the late afternoon, the air aloft remained very nearly unstable
with a potential for good vertical mixing, but lacked the impetus to initiate
the exchange.
Significant radiational cooling began at the ground at about 1700 CST.
This cooling continued into the night, giving a shallow, stable inversion
163
-------
it — -^ ^^ i "- • rf »
mil L, ,
u
c
re
Q
D)
JO
ra
o>
E-8
f S
re c
u
£
I
U.
'aannnv
164
-------
layer by midnight. Thereafter, the stability of that lower layer tended to
decrease until by morning (0600 CST), a warming process was clearly taking
place. By 0800 CST of 9 June, the vertical structure of the air was similar
to but slightly less stable than that observed the morning before.
This analysis of potential temperature is further substantiated by the
rawinsonde observations from the St. Louis area and from the soundings at
Salem, Illinois. Both sets of data indicate a moist layer of air having a
slowly decreasing water vapor mixing ratio from the ground to approximately
2 km throughout the flight. Above 2 km, the mixing ratio decreased up to the
subsidence inversion at 2.9 km and decreased more rapidly in the superior air.
The aircraft vertical profiles, discussed in section 4.3.5.1, also indicate
that the upper extent of the high ozone concentration was near 2 km with a
substantial decrease above. Vertical mixing of surface effluents effectively
extended up to that level but did not extend all the way to the subsidence
inversion about 500 to 800 m above. Assuming that the oxidant and the water
vapor result from emissions near the ground and that they are susceptible to
the same transport processes, the effective mixing depth of the atmosphere
during the daytime must have been 2.0 to 2.5 km.
6.2.2.2.2 Wind Speeds
The vertical distribution of wind speeds as interpolated to the hourly
positions of Da Vinci II through the first 3.5 km of the atmosphere from the
rawinsonde data is shown in figure 69. The lack of detail in wind speeds in
this analysis is compensated for by the added depth of the atmosphere, reveal-
ing further insight into the behavior of the winds during the day. This
analysis clearly shows that the disorganized, low wind speed pattern in the
lower 1.5 km extends throughout the depth of the atmospheric boundary layer.
The wind speeds reached their minimum just below the upper level inversion
layer. After sunset (2000 CST), a low level jet began to form near 500 m MSL
(300 m above ground), just above the surface radiative inversion. The jet
built in intensity, reaching a maximum speed near 0100 CST. A secondary wind
maximum that occurred in the 2.0- to 2.5-km layer was not related to the lower
level jet but resulted from a slight southerly shift of the high pressure
pattern at 850 mb, increasing the pressure gradient aloft.
6.2.2.3 Acoustic Sounder
A pair of acoustic sounders were operated by Dr. Ed Miller of Argonne
165
-------
o
(O
a
TB
21
V fl>
^X ^O
tn c
•o o
Is
"5 i
II
a>
(O
166
-------
National Laboratory in support of the Da Vinci II program. These mobile units
were operated in a leap-frog fashion, sampling the Da Vinci II environment.
An acoustic sounder operates on the principle of backscattering of a trans-
mitted sound pulse by a change in the temperature of the atmosphere. The
pulse of sound is transmitted vertically approximately every 20 s. Reflected
sound energies are then recorded as a function of time, which is translatable
to altitude. Figure 26 was prepared by Dr. Miller and given to RTI as prelimi-
nary data. Since there have been no other data forthcoming, it was assumed
that this portrayal of the data has become final. RTI's interpretation fol-
lows .
During the morning hours preceding and shortly after launch, the acoustic
sounder showed a stable low level inversion layer or layers in the lowest
200 m. The stable layer was gradually rising and weakening, until by 0920 CST
the low level inversion had dissipated. Thereafter, until approximately
1730 CST, the record indicates numerous thermals.
During the early evening a stable layer developed and intensified. At
2300 CST the thin, wavy inversion layer appeared only in the first 50 m above
the ground. This is a smaller thickness than was indicated in figure 68. By
0200 CST the layer was still less than 100 m thick; by 0300 to 0400, a little
thicker and perhaps a little more intense inversion layer had developed. At
0530 CST, wave motion on the top of the inversion was strongly indicated,
which suggests that destabilization was beginning and would probably persist.
By 0700 CST, the weaker but thicker inversion layer was indicated.
6.2.3 Conclusions
The morning breakup of the inversion layer on 8 June 1976 was completed
by about 0930 CST. A free convection regime seemed to have developed since
the winds were light and disorganized. This combination of events eventually
produced a well-mixed layer to about 2.2 km where the air began to stabilize,
limiting the vertical mixing.
There seemed to be a breakdown in the low level, nocturnal stability on
the night following the Da Vinci II launch. The acoustic sounder record and
the development of the low level jet seemed to confirm the event. Thus it
appears that during the period of 2300 to 0100 CST, there might have been a
breakdown in the low level inversion, permitting the ozone-rich air from aloft
to be brought to the ground.
167
-------
6.3 Modeling Atmospheric Chemistry and Physics
6.3.1 Need for Modeling
The assimilation of meteorological and air chemistry data into a cohesive
picture of the events of the Da Vinci II flight requires the analyst(s) to in-
corporate the ongoing physical processes and their qualitative contribution to
"explain" the various nuances of the data. Both the air chemistry and the at-
mospheric physics are time- and space-dependent processes. Some proceed rapid-
ly; others proceed slowly. Quantitative rather than qualitative results give
far more insight into the processes, but are far more difficult to attain.
The Da Vinci II program has an extensive data base in both atmospheric
chemistry and physics to support a quantitative analysis. As good as the data
base may be, it is still insufficient to provide the detailed relationships
among the variables. The program is undersampled and unable to fully describe
the ongoing atmospheric processes. While it is not a panacea, a model of the
atmospheric chemistry and physics can lend insight into pollutant transport,
diffusion, and interaction. Atmospheric modeling of emissions, transport, and
diffusion in urban areas into nonurban areas is not well developed at the pres-
ent time. Such models also require an extensive emissions inventory, including
a mobile source and fixed sources throughout the urban area. A steady, pre-
dictable wind field is usually required to simplify the model. The Da Vinci II
program and the St. Louis area do not lend themselves to application of the
large photochemical models. The analysis of the cross section data indicates
significant vertical shear within the mixed layer and thus does not fit the
other models. The models are incompatible with the space scale of the Da Vinci
II flight. It soon became apparent that it would be necessary to adapt a sim-
ple, existing meteorological model and couple it with a simplified existing
air chemistry model. In reality, neither model is simple. RTI's experience
with an air chemistry model to simulate smog chamber studies appeared a proper
choice for the air chemistry model. The one-dimensional planetary boundary
19
layer (PEL) model of Busch, Chang, and Anthes was chosen to simulate the
atmospheric physics (e.g., meteorological conditions), since it incorporated
the surface heating and cooling, friction, and a driving geostrophic wind in
a nonactuating environment, which are the principal factors of the Da Vinci II
environment.
168
-------
6.3.2 Results of Model Application
The model was applied to the data of the morning of the Da Vinci II
flight. The vertical profiles of u, v, and 0 from the cross section analysis
at 0800 CST were used to initialize the meteorological fields at 100-m vertical
increments. Hourly temperature observations at Lambert Field, St. Louis, were
used to compute the surface heat flux. Geostrophic winds of 190° @ 2.0 m/s at
the top of the model (3,210 m) and 325° @ 0.8 m/s at the ground were used. An
inert material (T = 0) with constant concentration (C=l) throughout the atmos-
phere with a constant source (w'c1 = 0.05) was modeled, using 5-min time steps.
A test run of 4 h was attempted. The results are shown in figure 70.
The test run was successful until about 1100 CST (see figure 70). Between
1000 and 1100 CST, the values of L became extremely small (^-0.1) because the
friction velocity u^. was very small, and the surface heat flux was large but
not excessive. The domination of heat flux over surface friction characterizes
the free convection mode of the atmosphere. The atmospheric scaling used in
the model does not apply to free convection, thereby invalidating further
modeling results of the daytime convection. The small values of L, arising
from the weak geostrophic flow, give large, negative values of z/L, which re-
sulted in large eddy diffusivities (K, ^ 105 m2/s) in the mixed layer. Such
diffusivities are encountered only at a highly turbulent, convective atmos-
phere.
Before the onset of the free convection, the boundary layer behaved as
expected. The slightly stable layer gradually became adiabatic. The mixing
depth increased with time. Very little change was observed in the concentra-
tion profile since it was initially uniform.
These model results agree very well with the evidence gathered and ex-
periences of the Da Vinci II flight. The time of onset of the free convec-
tion corresponds to the acoustic sounder records, the cross section analysis,
and the reasons for letting the balloon go to higher altitudes during the day-
time .
Future simulations of the Da Vinci II PEL with the current model must be
restricted to the late afternoon and nocturnal phases of the flight to avoid
the free convection situation. Considerations of nonuniform profiles of ma-
terials and their elementary interactions should help the understanding of the
measurements made during the flight. Coupling of the atmospheric physics model
with the air chemistry model awaits further development work.
169
------- An error occurred while trying to OCR this image.
-------
7.0 PRINCIPAL FINDINGS AND CONCLUSIONS
7.1 Principal Findings
Selected data collected by various participants during the flight of Da
Vinci II were consolidated, summarized, and segmented into general subject
areas for analysis and interpretation. These subject areas include atmos-
pheric chemistry and meteorological analyses. Data were analyzed and inter-
preted according to the objectives for each analysis and have been incorpor-
ated to combine both a chemical and meteorological interpretation of results.
Principal findings from these studies are summarized below.
1. Conditions of high ozone concentrations persisted aloft for the
study period. These conditions were widespread and extended for
several hundred kilometers. Air containing high concentrations of
ozone, approximately 0.11 ppm, was being transported into the study
area on the flight day.
2. Da Vinci II was launched into air that had an immediate fetch over a
nonurban area, although the total experiment was conducted within a
stagnant, polluted, high-pressure system. Air parcel trajectories
suggest that Da Vinci II was launched into air that 72 h earlier had
been in eastern Kentucky and the Ohio Valley.
3. Da Vinci II did not travel in the heart of the St. Louis urban plume
as defined by ozone concentrations. Examination of the hydrocarbon
concentrations sampled on Da Vinci II suggests that the flight oc-
curred in air characteristic of suburban-to-nonurban areas.
4. The ozone concentration at the flight level (700 m) of Da Vinci II
was greater than 0.08 ppm when Da Vinci II reached that altitude.
Since it reached that level prior to 1030 CST when deep convective
mixing was experienced in the urban boundary layer, a major portion
of the upper level ozone concentration must have come from sources
other than St. Louis on the morning of 8 June.
5. During the morning, ground-level ozone concentrations were less than
concentrations aloft. The observed increase of ozone concentrations
aloft, therefore, could not have resulted from upward mixing of
ozone from the ground, but was probably due to tropospheric photo-
chemical synthesis.
6. Vertical profiles show substantial ozone concentrations aloft be-
tween 762 and 1,981 m (2,500-6,500 ft) on the morning of 8 June with
a sharp decline in ozone concentration above 2,133 m (7,000 ft).
Mean ozone concentrations between 2,133 and 2,743 m (7,000-9,000 ft)
were less than those measured between 762 and 1,981 m (2,500-6,500
ft) by a factor of 2.4. These observations suggest that stratospher-
ic intrusion was not responsible for the elevated ozone concentra-
tions that existed aloft on 8 June.
7. The impact of anthropogenic emissions on ambient concentrations
within the urban area is variable and depends strongly on the time
171
-------
of day. Ground-level CO and NO were diluted by factors of 5 to 7
between the morning and afternoon. This behavior probably reflects
the significant increase in the mixed volume that occurs with the
dissipation of the surface-based radiation inversion and the estab-
lishment of well mixed conditions.
8. As vertically well mixed conditions were established between 0500
and 1100 CST, both downward mixing and photochemical synthesis
contributed to the observed increase in ground level ozone concen-
tration. This ground level increase exceeded that aloft by a factor
of approximately two.
9. Between 1100 and 1700 CST on flight day, a near zero ozone con-
centration gradient existed from the ground into the mixed layer.
This is based, to a large extent, on the close comparison of ozone
concentrations determined on Da Vinci II with those beneath Da Vinci
II at ground level.
10. After well mixed conditions were established and until the nocturnal
inversion began to form, the increases in ozone concentrations aloft
and at ground level were approximately equal and probably resulted
from photochemical synthesis.
11. The airflow in the immediate vicinity of St. Louis on 8 June was
governed by the synoptic scale flow and an intense circulation
produced by the urban heat island. Da Vinci II responded to both of
these scales of motion.
12. A major air pollution plume (defined by the ozone distribution) was
found at the surface. The maximum ozone concentration in this plume
was immediately downstream (within 5 km) of the center of downtown
St. Louis between 1000 and 1500 CST. Simulation results (meteorolog-
ical model) suggest that if a plume existed in the upper portions of
the boundary layer, it would have moved in a different direction
than the surface plume.
13. Between 1000 and 1500 CST, the maximum ozone concentration in the
major air pollution plume at the surface (as defined by ozone distri-
bution) was found in a region where there was a zone of horizontal
convergence in the wind field associated with the urban heat island
circulation.
14. After 1500 CST, the maximum ozone concentration in the air pollution
plume (as defined by the ozone distribution) moved farther downstream
(greater than 15 km) when the heat island circulation and its accom-
panying convergence zone was dissipating.
15. Within the layer that had been well mixed during the daytime, strat-
ification which occurred at night with the establishment of the
nocturnal radiation inversion resulted in the formation of two
regimes of ozone concentrations. Ozone concentrations within the
radiation inversion were much reduced compared to levels above it.
This is presumably due to destruction by surface deposition and
reaction with ozone-destructive agents emitted into and trapped
within the inversion.
16. Ozone data obtained aboard Da Vinci II were used to calculate a
nighttime ozone half-life of 116 h. Examination of available hydro-
172
-------
carbon precursor data obtained aboard Da Vinci II suggests that
concentrations of ozone destructive and other HC species aloft were
low during the flight. The dark phase half-life of ozone under the
conditions of presumed low levels of precursors is sufficient to
allow transport of ozone from an urban area overnight to another
area without significant diminishment due to decay.
17. Penetration through the nocturnal radiation inversion resulted in
the mixing of high levels of ozone to the ground and precursors from
the ground to aloft. This was suggested by the sharp nighttime
ozone peaks and the associated declines of NO and CO observed at
selected ground stations. This phenomenon occurred frequently in
May and June and was observed at ground stations covering an area of
several hundred kilometers. It may also provide mechanisms for in-
creasing nighttime ozone destruction aloft and for enhancing early
morning ozone synthesis by distributing ozone precursors aloft above
the inversion before sunrise.
18. The balloon may have entered an air parcel that was enriched in
hydrocarbons after 2100 CST on 8 June and traveled within this par-
cel for the remainder of the flight.
19. Sharp, short-term reductions in ozone concentrations occurred aloft
during the nighttime portion of the flight and were coincident with
increases in SO. concentration. Both SO and NO are emitted by
power plants. Although NO was not measured in this study, the
observed ozone behavior is probably the result of destruction by
reaction with NO .
x
20. The influence of ozone precursors emitted in the urban area was
manifested at ground level by ozone concentration enhancement over
nonurban concentrations. The magnitude of this enhancement was 0.06
to 0.11 ppm. Both the buildup and movement of the region of enhanced
ozone were documented, although detailed definition of the extent
and magnitude were severely hampered by the lack of a comprehensive
supporting aircraft measurement program.
21. High ozone concentrations measured aloft aboard Da Vinci II on the
morning of 9 June in Southwestern Indiana are attributed to long
distance transport of ozone. High ozone at the surface is influenced
by synthesis and mixing downward of the ozone aloft in the morning.
After a well-mixed layer is established, changes resulting from
mixing are minimized. Further increases in ozone concentration at
the surface can be attributed to synthesis.
22. The use of a Lagrangian system to document ozone transport and
address atmospheric chemistry problems has been shown to be feasible.
Interpretation of these data are severely limited, however, unless
supporting data that define vertical and horizontal pollutant distri-
butions are also available.
7.2 Conclusions
The conclusions listed below are derived from the analysis of the Da
Vinci II data relative to the objectives of this project.
173
-------
1. The synoptic wind flow pattern as modified locally by the urban heat
island circulation prevented Da Vinci II from passing over the
St. Louis urban area and taking up a position in the urban plume.
2. During the nocturnal phase of the flight, changes in ozone con-
centrations observed aboard Da Vinci II could not be directly attri-
buted to the downtown St. Louis urban plume.
3. General meteorological conditions--a subsidence inversion aloft and
a strong radiative inversion based at the ground--were nearly ideal
for long-distance transport aloft of ozone at night, while keeping
ozone separated from ground level emissions and destruction and
limiting vertical mixing.
4. The dark-phase stability of ozone above the nocturnal surface-based
radiation inversion suggests that transport of ozone can occur aloft
over long distances at night without significant diminishment. It
is possible for the ozone to be transported overnight several hundred
kilometers and be mixed to the ground on the next day with a signifi-
cant impact on ground-level air quality.
5. The occurrence of high ozone concentrations aloft on the morning of
9 June, in a rural area in southwestern Indiana, is attributed to
long-distance transport of ozone.
6. Balloon-borne experiments are a feasible approach for addressing
atmospheric chemistry problems, provided supporting data from ground
stations, a ground-level chase vehicle, and an instrumented aircraft
are available to complement the interpretation. The omission of any
one of these measurement platforms severely limits the interpreta-
tion effort.
174
-------
8.0 REFERENCES
1. Research Triangle Institute, Ambient Monitoring Aloft of Ozone and Pre-
cursors in the Vicinity and Downwind of a Major City, Environmental Pro-
tection Agency Report No. EPA-450/3-77-009.
2. Research Triangle Institute, Study of the Formation and Transport of
Ambient Oxidants in the Western Gulf Coast and North-Central and North-
east Regions of the United States, Environmental Protection Agency Report
No. EPA-450/3-76-033.
3. Research Triangle Institute, Investigation of Rural Oxidant Levels as Re-
lated to Hydrocarbon Control Strategies, Environmental Protection Agency
Report No. EPA-450/3-76-035, March 1975.
4. G. T. Wolfe, personal communication, Interstate Sanitation Commission,
New York, 1977.
5. W. H. White, J. A. Anderson, D. L. Blumenthal, R. B. Husar, N. V. Gillani,
J. D. Husar, and W. E. Wilson, Jr., "Formation and Transport of Secondary
Air Pollutants," Science, 194:187, 1976.
6. J. E. Sickles, II, "Ozone Precursor Relationships of Nitrogen Dioxide
Isopentane and Sunlight Under Selected Conditions," Doctoral Dissertation,
Department of Environmental Sciences and Engineering, University of North
Carolina, Chapel Hill, North Carolina, 1976.
7. J. E. Sickles, II, L. A. Ripperton, and W. C. Eaton, "Oxidant and Precur-
sor Transport Similation Studies in the Research Triangle Institute Smog
Chambers," in Proceedings of International Conference of Photochemical
Oxidant Pollution and its Control, Environmental Protection Agency Publi-
cation No. EPA 600/3-77-OOla, p. 319, 1977.
8. S. L. Kopczynski, personal communication, Environmental Protection Agency,
Research Triangle Park, North Carolina, 1977.
9. S. L. Kopczynski, W. A. Lonneman, T. Winfield, and R. Seila, "Gaseous
Pollutants in St. Louis and Other Cities," J. Air Pollution Control Assoc.,
25(3):251, 1975. ~~
10. H. Levy, II, "Photochemistry of Troposphere," Advances in Photochemistry,
9, J. N. Pitts, Jr., G. S. Hammond, and K. Gollnick, Eds., John Wiley
Interscience, New York, p. 369, 1974.
11. J. M. Mitchell, Jr., "The Thermal Climate of Cities," Symposium of Air
Over Cities, Cincinnati, Ohio, November, 6-7, 1961, U.S. Public Health
Service Report No. A62-5, Cincinnati, Ohio, November 1961, pp. 131-145.
12. F. M. Vukovich, J. W. Dunn, III, and B. W. Crissman, "A Theoretical Study
of the St. Louis Heat Island: The Wind and Temperature Distribution,"
J. Appl. Meteor., 15(5):417-440, May 1976.
13. K. D. Hage, "Nocturnal Temperatures in Edmonton, Alberta," J. Appl. Meteor.,
1_1(1): 123-129, February 1972.
14. T. R. Oke and C. East, "The Urban Boundary Layer in Montreal," Boun. Layer
Meteor., H4):4ll-437, April 1971, D. Reidel Publishing Company, Dordrecht,
Holland.
175
-------
15. T. R. Oke, D. Yap, and G. B. Maxwell, "Comparison of the Urban/Rural Cool-
ing Rates at Night," Proceedings of the Conference on Urban Environment
and 2nd Conference on Biometeorology, October 31-November 2, 1972, Phila-
delphia, Pennsylvania, 1972, pp. 17-22.
16. F. M. Vukovich, "A Study of the Effect of Wind Shear on a Heat Island
Circulation Characteristic of an Urban Complex," Hon. Wea. Rev., 103(175):
27-33.
17. F. M. Vukovich, and J. W. Dunn, III, "A Theoretical Study of the St. Louis
Heat Island: Some Parameter Variations," paper submitted to the J. Appl.
Meteor., 1977.
18. B. D. Zak, personal communication.
19. N. E. Busch, S. W. Chang, and R. A. Anthes, "A Multi-Level Model of the
Planetary Boundary Layer Suitable for Use with Mesoscale Dynamic Models,"
J. Appl. Meteor., 15:909-919, 1976.
176
-------
APPENDIX A
EVALUATION OF SAMPLE COLLECTION CONTAINERS
FOR HYDROCARBON SAMPLING
177
-------
APPENDIX A
EVALUATION OF SAMPLE COLLECTION CONTAINERS
FOR HYDROCARBON SAMPLING
A-l Introduction
During the flight of Da Vinci II, grab samples (2- to 3-min sample dura-
tion) were collected in stainless steel containers and Tedlar bags (protected
from sunlight) using a common sample pumping system for analysis by Washington
State University (WSU) , National Center for Atmospheric Research (NCAR) , and
Research Triangle Institute (RTI). Listed in table A-l are participants, sam-
pling container type, and components analyzed by the respective organizations.
From table A-l, comparisons could be made between RTI and NCAR for carbon
®
monoxide and methane and between RTI and WSU for Freon 11 and 12, acetylene,
and ethane/ethylene.
Preliminary evaluation of the Da Vinci II ambient hydrocarbon data showed
disagreement between absolute concentrations of various compounds that could
be compared among the three participating organizations.
Figure A-l shows a plot of acetylene concentrations reported by WSU and
RTI for samples collected during the Da Vinci II flight. Samples were col-
lected within 3-5 rain of each other and since acetylene is a relatively un-
reactive compound, the difference between the two sets of numbers becomes
significant.
Table A-l. Hydrocarbon sampling and analysis: Da Vinci II
Participant
Sample Container Type
Components Analyzed
RTI
WSU
NCAR
Tedlar® bags covered with
Scotchpak
Stainless steel
containers
Stainless steel
containers
CO, CH4, Freon® 11, Freon® 12,
ethane/ethylene, acetylene,
and C3~Cs hydrocarbons
Freon 11, Freon® 12, methyl
chloride, carbon tetrachlo-
ride, nitrous oxide, ethane,
ethylene, and acetylene
CO, CH4, N20, hydrogen,
carbon tetrachloride
178
-------
.a
S
N >
CM <
CM O
OC
CQ
O
o>
c
•o
1
o
o.
E
§
'€
iri cj o r^ iri
(Oqdd) N08UV3 NOI11I8 H3d SlUVd
•o
re
CO
.o
Q.
0)
I
iZ
179
-------
The differences could come from several sources, the most obvious being
the use of different materials for sample collection and problems in the prep-
aration or use of primary calibration standards. Before data analysis could
begin, it was apparent that an investigation was needed to evaluate the prop-
erties of various materials routinely used in the collection and storage of
hydrocarbons in the C2-Cs range and certain halocarbons. Secondly, the labo-
ratory procedures and quality control practices were examined for all partici-
pants involved in hydrocarbon/halocarbon measurements for Da Vinci II. Results
of that evaluation are summarized in appendix B. The discussion and results
presented in the following paragraphs describe RTI's evaluation of glass
® ®
bulbs, Tedlar and Teflon bags (protected from light), and treated stainless
steel containers (purchased from WSU) for storage of ambient concentrations of
various hydrocarbons (C2~C5 range) and certain halogenated compounds.
A-2 Sample Container Evaluation Program
Hydrocarbon measurements at the ambient concentration level have been
made on a routine basis by numerous research groups. The analytical method-
ology, techniques, and approaches used are inconsistent from one research
group to another. In many cases, interlaboratory comparisons have shown large
discrepancies in the results. Each research group is confident of its approach
and frequently has incorporated no quality control practices or procedures
into its methodology.
A variety of sample collection techniques has been used during the past
10 yr to collect samples; however, a systematic evaluation of the stability of
compounds in the sample container after the sample has been collected has not
been performed. There are basically two types of devices for collection of
^2~Cjo hydrocarbons and halogenated compounds in ambient air, namely polymeric
films (Tedlar and Teflon bags) and hard surface containers, such as stainless
steel and glass vessels. Each container has its own unique properties with
advantages and disadvantages.
Controlled experiments were performed to evaluate both the containers and
the stability of the various hydrocarbons at ambient concentration levels.
Ambient air monitoring sample integrity has to be maintained to prevent sample
degradation or sample contamination. To study these effects, a matrix of
tests was designed. Table A-2 shows a list of the tests performed on four
types of containers. Four containers of the same type were employed for each
180
-------
Table A-2. Evaluation tests
Test #
1
2
3
4
5
(4) Stainless
Zero Air
HC's + Zero
HC's + NO
HC's + N02
HC's + 03
Steel (4) Glass
Zero Air
Air HC's + Zero Air
HC's + NO
HC's + N02
HC's + 03
(4) Teflon®
Zero Air
HC's + Zero Air
HC's + NO
HC's + N02
HC's + 03
(4)Tedlar®
Zero Air
HC's + Zero Air
HC's + NO
HC's + N02
HC's + 03
test. Two of the containers were spares in case the two containers evaluated
developed leaks or unforeseen problems were encountered.
The hydrocarbon mixture employed in the matrix was composed of ethane
plus ethylene, acetylene, propane, propylene, and n-pentane. Each test was
conducted over an 18-day period with more intensive analyses during the ear-
lier part of the test. Samples were withdrawn from the containers on day 0,
1, 2, 3, 4, 8, 13, and 18. After 4 days, only two containers of each type
were tested for the extended part of the stability test. From day 1 to day 4
samples were analyzed on a random basis (see table A-3) to minimize any sys-
tematic error in the analyses.
A-3 Zero Air Evaluation Results
The zero air experiment provided data on the contamination of the sample
stored in the four containers. The zero air had been previously analyzed for
trace hydrocarbons and was found to contain only trace quantities (~20 ppb) of
ethylene. It was later discovered that the ethylene was due to a small piece
of Teflon tubing (FEP) in between the sampling containers and the zero air
source. This problem was solved by replacing the Teflon tubing with flexible
stainless steel tubing. The zero air was introduced inside the containers
after it had passed through a double dilution system, which was later employed
in the rest of the study to generate low concentrations of hydrocarbons and
other pollutants. Two different types of sampling containers were filled
simultaneously to insure some cross reference between sample containers.
The zero air test samples collected from the containers were analyzed for
both C2~Cs and C5-C10 hydrocarbons. The data collected demonstrate clearly
the difficulties that had been experienced previously by RTI and others.
181
-------
Table A-3. Analysis schedule
Time
08:00
09:00
10:00
11:00
12:00
13:00
14:00
15:00
Day 1
A!
BI
Ci
DI
EI
A2
B2
C2
Day 2 Day 3
A2 A!
B2 Bj
C2 d
D2 D!
E2 Ej
AI A2
BI B2
c,
Day 4
A2
B2
C2
D2
E2
A!
B!
Cl
Note: Suffix number refers to container number. Numbers 3 and 4 of each
type were spares.
A = stainless steel cylinders (2)
B = stainless steel cans (2)
C = glass bulbs (2)
D = Tedlar- bags with Viton quick connect (2)
E = Teflon bags with stainless steel fitting (2)
Stainless steel containers showed a slight decrease in the ethylene concen-
tration from 20 ppb to 12 ppb (18th day). After 18 days, no other components
were detected in the analyses. It should be pointed out that the stainless
steel containers had been pressurized to 40 psig with the use of an MB151
Metal Bellows pump. The absence of contamination over such an extended period
is important. Since in many cases, sampling containers are transported in
contaminated atmospheres, it is imperative that the samples retain their in-
tegrity even though the containers are exposed to pollutant levels several
orders of magnitude higher than the collected sample.
Glass containers appeared to be as good as the stainless steel containers,
but some disadvantages were encountered. Since the glass bulb could not be
pressurized to any great extent (10 psi), retrieval of samples was more diffi-
cult and possible contamination could occur unless a great deal of care was
taken in maintaining leak-free seals during transfer of samples from the con-
tainers to the gas chromatograph. Since only limited amounts of sample could
182
-------
be drawn from the bulb, it was found that some contamination did occur when
the vacuum in the container became significant (~400 mmHg). Leaks occurred at
both the Teflon stopcocks or at the connections between the stopcock and the
GC, drawing lab air inside the sampling bulb. Absolute conclusions could not
be made about the glass containers except that without extreme care, the
samples could be easily contaminated.
®
Results of zero air stability for Tedlar bags, a commonly used polymer
for the collection of air samples, reaffirmed our belief that the sampling
container is adequate for sampling Cj^-Cs hydrocarbons. No increase in any
hydrocarbon species was shown for up to 8 days. Trace quantities (~0.2 ppb
[v/v]) of acetylene, propane, and propylene were observed by the 18th day;
however, significant concentrations (1-2 ppraC) of FID responsive species were
observed in the C$-C10 range.
®
Teflon (5 mil FEP) sampling bags that had been treated with zero air
showed unexpected results. After 4 days, significant concentrations of vari-
ous hydrocarbons were observed. The results of this test are shown below in
table A-4.
®
The concentration of the CS-CJQ hydrocarbons in Teflon bags ranged from
Table A-4. Zero air test results for
Teflon bags (ppb (v/v))
Days in Container
Hydrocarbon Compounds 124
Ethane + Ethylene 4.8 6.0 9.5
Acetylene >0.1 >0.1 >0.1
Propane 0.5 0.8 1.2
Propylene 0.5 1.1 1.4
Isobutane >0.1 >0.1 >0.1
Butane >0.1 0.3 0.3
1-Butene >0.1 >0,1 >0.1
Isopentane >0.1 >0.1 >0.1
Trans-2-Butene >0.1 >0.1 >0.1
n-Pentane >0.1 >0.1 >0.1
8
19.6
0.3
2.9
2.8
>0.1
0.8
>0,1
1.9
>0.1
>0.1
13
111.2
0.6
3.6
2.5
>0.1
2.9
>0.1
2.3
>0.1
>0.1
18
201.7
3.1
4.8
2.3
>0.1
2.2
>0.1
3.0
>0. 1
>0. 1
183
-------
1 ppmC oa day 1 to 2 ppraC by day 18. The problem encountered with the Teflon
film sampling bag has been experienced by other researchers.* Severe treat-
ment of the TFE film in some cases can ameliorate its performance. For exam-
ple, repeated exposure to high ppm concentrations of ozone can improve per-
formance. It has also been found that generation of ethylene by Teflon film
can be batch dependent.*
In summary, it was found that stainless steel containers could retain
samples, uncontaminated, up to 18 days, while other containers such as glass
bulbs are too fragile and require extreme care in order to retrieve the col-
®
lected sample. Tedlar sampling bags exhibited good results for Cg-Cg but
®
were inadequate for any other higher molecular weight species. Teflon sam-
pling bags were unsatisfactory for both C2-Cg and Cg-Cjo hydrocarbon analyses.
A-5 Hydrocarbon Evaluation Results (C2~Cg Range)
Test 2 involved low concentrations of hydrocarbons and zero air. In this
test, the analyses provide some idea of the interaction of the container with
an ideal hydrocarbon sample. Previous evaluations have been made at ppm
levels (references), but control testing had not been undertaken at concentra-
tions close to those found in ambient air. The concentrations of the individ-
ual hydrocarbons were approximately 10 ppb (v/v). Measurements at that level
have been shown to have a repeatability of 6 percent. This was experimentally
determined by four successive analyses of the same sample. Results of repeat-
ability tests are shown in table A-5.
*Mike Holdren, private communication, 1977.
Table A-5. Repeatability results for selected
hydrocarbon analyses (ppb (v/v))
Hydrocarbon
Compounds
Ethane + Ethylene
Acetylene
Propane
Propylene
n-Pentane
1
26.9
10.5
12.9
12.3
10.6
2
28.7
11.6
14.4
13.0
11.3
3
26.8
11.4
14,0
12.3
10.0
4
27.3
11.2
15.2
14.7
12.2
Variance %
3.2
4.3
6.8
8.7
8.6
184
-------
It should be emphasized at this point that approximately 100 cm3 of sam-
ple are eryogenically concentrated for these analyses. The variance in the
analysis can be improved by increasing the trapped volume, but it was felt
that a variance of 6 percent was sufficient. This also provided a base from
which statistical evaluation of the data could be made, i.e., assuming that a
variance greater than 6 percent would be reflective of interaction of the sam-
ple with the container.
During the test, all containers were kept inside aluminum suitcases to
minimize any photochemical activity. When the containers were sampled, the
laboratory lights were turned off. Results of the test are shown in table A-6
and are illustrated in figures A-2 through A-5.
In order to compare variance for each container and component analyzed,
the mean of these variances was calculated for a particular test and container.
The means are shown in table A-6. The results are fairly conclusive. The two
®
containers that have the smallest variance are the Tedlar bag and the stain-
less steel container. It should be noted that both of these containers have
®
variances that approach the variance of the analytical procedure. The Teflon
®
bag and glass container have variances approximately twice that of the Tedlar
and stainless steel containers. The high variance of the glass container is
probably not due to the container itself but to the difficulty of recovering
samples from the container. The high variance observed with the Teflon bag
can only be explained by interaction of the container with the sample. In
most cases, the concentration of each component analyzed increased. Observa-
tion during the zero air evaluation reconfirmed the problem of contamination
®
of the Teflon bags.
Table A-6, Hydrocarbon stability results
Hydrocarbon
Compounds
HC
HC
HC
HC
+ Zero
+ NO
+ N02
+ 03
Stainless Steel
10,2
6.9
5.9
5.4
Glass
19.5
13.7
20.8
5.5
Tedlar®
12.4
4.7
3.6
5.4
Teflon®
16.2
13.7
11.1
8.6
Mean Variance 7.1 14.9 6.5 12.4
of All Tests
185
-------
LU
Z 01 CD
UJ C C
I Ol 0) OJ d)
0) i— c i— c
C >-, to >, ro
CO +J Q. S+J
.
-C QJ O O C
•M O L. S.
U
X
a
CD
o
co
a
CM i—
(A/A) add
186
-------
Z , ro >,<0
(O •*-> Q. O--t->
-C
///
A\ , , Jk! J
o o V o> oo mj r-.^fcin^-rocMr— CL
CO
-
CM . S in <* ro CM «— c
O)
ID x^i
(D
3 »
^
^ ja
*~ 'V
2u, 1
i .i
00 LL
VD
»a-
CM
o
J
o oo
CM i—
(A/A) Sdd
187
-------
LU C
I CU
0) r-
C >>
10 4->
-C (U
•u a
LU et
O)
ai
-------
z: , > a. o. +•>
-C O) O O C
4-> U S- S- 0)
UJ
LEI
(A/A)
189
-------
o
o
CM
CO
1
/
bU
i.
CU
c
in
c
•3 '
o
4
\
i i i
O 0 O
• ^^ CO CVJ r~*
CM
(A/A)
O
Z
CO
S-
cu
o
UJ
z cu cu
UJ C C
i cu cu cu cu
cu i— c i— c
c >> us >, 10
n) •*-> Q. Q. i->
JC CU O O C
-!-> U $- S- CU
UJ < O. Q. Q.
^ o D <3 •
o
LO
O
CO
O
C\J
oo
CO
>,
(O
a
U
o
|
"ro
"55
fi
c
'5
CO
CO
CO
LL
Csl
oo
IT)
CO
(A/A)
190
-------
I/I
£
s
CO CM
o in
CM »—
oo r-» vo in
(A/A) Dqdd
ro CM f—
CKJ» o
s-
a
8
o
o
I
QJ O
C C
a> -, > *
*7 a. CL +J
cu o o c
+-' (j S- $_ OJ
LU
>>
00
O
z
o
X
o
'^
CO
OO t~- VO If)
(A/A)
ro
191
-------
s-
(D
= o
it) O
+•> CM
C LU
O
O
O
r-
o o o
r^ m co
00
(A/A)
DO
i-
0,
= 0
10 O
4-> t—
C UJ
O
I O)
1
01
c
Ol Ol
C i— C
<0 >> <0
a. a. 4->
CD O O C
U S- S_ CD
<: Q. Q. o.
o D
o
CTi
O
O
in
O
n
A.Y;
D«0
oo
O c/>
"I
cr>
oo
O
O
3C
c
O
'^
re
_3
re
O)
re
_Q
©
c
_o
"S3
o
CO
O)
oo
(A/A)
192
-------
S-
Qj
•is
4->
C
O
00
(A/A) qdd
O
.o
'^
ra
_3
(0
S-
o,
IS
o
o
•z. cu o>
UJ C C
I Ol Ol Ol O)
O) •— C r— C
c >> « >» *
ro -»-> Q. CL -4-J
_c 01 o o c
J-> U S- S-
------- An error occurred while trying to OCR this image.
-------
El
o
o
<
\
»
\
I o"
O
oo 5.
CD
i
o>
re
CO
3
O
n
-*K-JT
VD IT) «3- n CVJ
(A/A) add
•— o crv co
-------
S_
QJ
U
<
O
o
(U
(O
a
o
s-
a_
a
.
0
OJ
c
O)
n
o
Q.
O
0
m
01
•»->
c
-
^- Q
D
(A/A) add
196
-------
s-
Ol
=
o
o
o
VO
o
CM
o
o
o
00
O -i
oo
<£>
g
00
CM
vo
CM i—
oo
(A/A) 9dd
OJ
o
o
I
o
re
0)
CO
o
u
(A
O
3
oo
• >•
D<
o
VO
o
CM
O
O
o
00
o
UD
VO
(A/A) 8dd
oo
197
------- An error occurred while trying to OCR this image.
-------
• o
J L
-IV-l
o
in
O
PO
o
ev
n cvj
s-
Oi
c
01
QJ
O)
i 01 o>
-------
s-
Oi
O
o
I I I
O O O O
10 in ^- co
o co
CM r—
vo in
PO CM
O CT)
CM •,
-c : QJ
+-» o
ui «t
Q.
o
s-
o a <
o
un
CO
O
00
CO
CO
o
o
I
c
(O
_
_o
"S3
3
Oi
200
-------
o
IT)
O
co
o
CM
00
(A/A) add
I/O
CM >-
O
CO
O OO
CM i—
00
00
(A/A) add
o
o
4J
CD
_3
CD
1
o
o
5
OJ Ol
c c
CU CU 01 O- O.+J
CU O O C
O S- S- 0)
< Q. Q- O-
oD <] •
O
ID
O
co
o
CM
00
(A/A) 9dd
v
•- o
05
O)
I/)
"I
CO
201
-------
A-6 Hydrocarbon Stability Results (Cs-C^p Range)
Further evaluations were performed with higher molecular weight compounds
that included cyclopentane, n-heptane, toluene, and o-xylene. Due to the
® ®
background in the Tedlar and Teflon bags, these containers were not evalu-
ated. Nor were the glass containers evaluated due to their fragility and the
difficulty of rcovering samples from them. Only the stainless steel containers
were evaluated. Unfortunately, these samples were evaluated without heating
the container. The results (see figure A-6) reflect those conditions. Later
experimentation has shown that the containers need to be heated just prior to
analysis to 70° C to minimize adsorption effects.
A-7 Evaluation of Effects of Prescrubber to
Remove Ozone During Samples
During the Da Vinci II flight, the RTI air sample collection system
employed a prescrubber (MnOa filter) to remove ozone. The filter consisted of
a 1/4" stainless steel tube filled with MnOg powder mixed with glass wool.
The Mn02 glass wool plug is held in the 1/4-in. tube with glass wool plugs at
each end. The efficiency of the MnOg was experimentally determined and was
found to be approximately 50 percent effective with 0.1 ppm 03. The effect of
the MnOj on the various hydrocarbons was not known and further evaluation was
required. Four compounds were tested as follows: n-pentane, propylene, pro-
pane, and acetylene.
A known concentration of these compounds was passed through various com-
ponents that made up the filters. The mixture was passed through an empty
Teflon tubing, a piece of Teflon tubing with a glass wool plug, a filtering
unit with 0.033 gm Mn02, and another filtering unit with 0.082 gm Mn02. The
results of these tests are shown in figures A-7 through A-10. These figures
indicate that the error bars for the analyses are as great as the changes in
the concentration that could be attributed to interaction of the filter with
the various compounds. The only conclusion that can be drawn is that the
effect of the Mn02 is minimal on these various compounds evaluated.
202
-------
o-
c
o
o
0)
8
c
a
v>
_C
£
(A/A)
I
CO
s-
o,
t=
D)
I
o
E
CO
2
O)
il
(A/A)
CM r-
203
-------
20
18
16
14
12
JO
a.
6
4
2
I error bar, + one standard deviation
w/o prefilter w/glasswool w/0.033g MN02 w/0.082g MN02
Figure A-7. Prefilter evaluation (n-pentane).
20
18
16
14
I error bar, + one standard deviation
8
6
4
2
w/o prefilter w/glasswool w/0.033g MN02 w/0.082g MN02
Figure A-8. Prefilter evaluation (propylene).
204
-------
20 p i error bar, +_ one standard deviation
18
I
14
12
10
O.
°- 8
6
4
Q.
°- 8
6
4
2
w/o prefilter w/glasswool 2/0,033g MN02 w/0,082g MN02
Figure A-9. Prefilter evaluation (propane).
o Acetylene
18 p A Ethane-ENE
16
~[~ Error bar, +_one standard deviation
14 '
12
w/o prefilter w/glasswool w/0.033g MN02 w/0.082g
Figure A-10. Prefilter evaluation (acetylene).
205
-------
APPENDIX B
EVALUATION OF LABORATORY AND QUALITY CONTROL PRACTICES
FOR HYDROCARBON SAMPLING AND ANALYSIS
(DA VINCI II PARTICIPANTS)
207
-------
APPENDIX B
EVALUATION OF LABORATORY AND QUALITY CONTROL PRACTICES FOR
HYDROCARBON SAMPLING AND ANALYSIS (DA VINCI II PARTICIPANTS)
B-i Introduction
In addition to the experimental work conducted to determine the effects
of sample container construction on hydrocarbon stability it was decided that
a better understanding of the laboratory and quality control practices for
each participant (i.e., WSU, NCAR, and RTI) was warranted. RTI quality con-
trol practices and analysis methodology are described in appendix C and sec-
tion 3.5 of this report, respectively. WSU and NCAR laboratories were visited
by RTI investigators and the following areas relative to hydrocarbon sampling
and analysis were discussed: (1) analytical methodology, (2) calibration pro-
cedure, (3) sampling containers and stability of compounds, (4) preparation of
sampling device, (5) time duration between sampling and analysis of Da Vinci
II samples, and (6) quality control program and estimated quality of the data.
B-2 Review of Laboratory and Quality Control Practices at National
Center for Atmospheric Research (Dr. Leroy E. Heidt)
The first participant laboratory visited was the National Center for
Atmospheric Research (NCAR) in Boulder, Colorado. Dr. Heidt was extremely
cooperative and freely discussed the work done by NCAR for Da Vinci II samples
and conducted a tour of his facilities. Results of these discussions are sum-
marized below and address each of the six areas described above as follows:
1. The analytical methodology used by NCAR was designed to analyze
samples that were collected in the stratosphere. Much care is
taken to insure that the sample is not contaminated during
transfer from the sampling container to the gas chromatograph.
The gas chromatograph used was modified extensively to accom-
modate samples that are at very low pressures (i.e., 0-100
tnillitorr). Methane and carbon monoxide concentrations are
analyzed by flame ionization detection. Carbon monoxide is
methanated prior to analysis. Hydrogen analyses are per-
formed on a gas chromatograph with an RF detector. Neon anal-
yses are performed on a high resolution mass spectrometer.
2. Primary calibrations of the various instruments are performed
on a monthly basis. Calibration gas mixtures are prepared
based on an absolute pressure basis. The pure compound of
interest is diluted through transfer from one stainless steel
container to another. The accuracy of this method should be
as good as the accuracy of pressure measuring devices.
208
-------
3. Collection devices used by NCAR are constructed from stain-
less steel. Due to the severe treatment of the containers
prior to sampling, Dr. Heidt felt that the stability of some
components, such as carbon monoxide, may not be as good as
for methane. Hydrogen measurements are considered to be
excellent with these containers, but components, such as
carbon tetrachloride, are not stable.
4. Stainless steel containers for sampling hydrocarbons are
prepared by filling with pure oxygen and heating to 350° C
for a period of time. High temperatures are needed to mini-
mize degassing of hydrogen from the stainless steel. Unfor-
tunately, an oxide layer is formed on the interior of the
container, due to the oxidizing atmosphere. The oxide layer
can act as a catalyst and transform some species, such as
carbon monoxide and carbon tetrachloride. After the con-
tainer has been heated under the oxidizing atmosphere, the
containers are evacuated to 10 torr. The containers are
normally used within 4 weeks. The containers used by Sandia
Laboratories were cleaned approximately 2-3 weeks before the
flight. Dr. Heidt was unaware as to how the samples were col-
lected during Da Vinci II.
5. Dr. Heidt was not certain of the duration of the time between
sample collection and analysis, but felt that the samples were
analyzed within 1 week after receipt from St. Louis. Informa-
tion was not readily available as to the date of sample con-
tainer cleanup, date of shipment and receipt of sample, date
of sample analysis, etc.
6. Discussion with Dr, Heidt revealed that NCAR did not have an
active quality control program during the time of the Da Vinci
II project. NCAR, at that time, relied on previous experience
with the analytical methodology to insure the quality of the
data. Dr. Heidt felt that the quality of the reported data
was good.
B-3 Review of Laboratory and Quality Control Practices at
Washington State University (Dr. Rei Rasmussen)
The second participant laboratory visited was Washington State University
in Pullman, Washington. Dr. Rasmussen was extremely cooperative and provided
RTI valuable information relative to the six areas listed in B-l.
1. The analytical methodology for hydrocarbons utilized flame
ionization detection with sample injection from stainless
steel containers using a 100-ml syringe. The sample is con-
centrated at liquid oxygen temperatures. Several 100-ml
injections are sometimes required to concentrate enough
material for analysis. The sample is then volatized onto
the gas chromatographic column using a beaker of hot water.
Data are acquired using an HP 3352 data system. Halocarbon
measurements are performed by flushing a sample loop and
209
-------
direct injection onto the column at atmospheric pressure and
room temperature. Halocarbon analyses are measured using
either peak height of the chromatograra or digital integration
of the peak area.
2. Calibration procedures and protocols are not clearly defined
as to frequency, etc. Halocarbon calibration concentrations
are prepared in large glass carboys by dilution of cylinder
gas purchased from Matheson Gas Products with zero nitrogen.
WSU also utilizes 6-1 air samples collected in remote areas as
standards. These "standard" samples are analyzed from time to
time to check instrument stability.
3. Dr. Rasmussen's opinion was that the stability of compounds
analyzed by WSU was excellent in stainless steel containers.
He did not have supportive data to substantiate this, but
said that WSU experience indicates that hydrocarbons and
halocarbons are stable.
4. Stainless steel containers employed by WSU have brass values
and are connected to the container with NPT fittings and
sealed with Teflon tape. Routine cleaning of the containers
-3
consist of evacuation to 10 x 10 torr, only. According to
Dr. Rasmussen, severe treatment of the container activates
the surface and makes it more susceptible to wall adsorption.
5. Dr. Rasmussen was not aware of the time frame from sample
collection to analysis. Records were not readily available
to document this.
6. Discussions with Dr. Rasmussen indicate that at that time no
quality control program was in use at WSU with respect to
hydrocarbon sampling and analysis. His opinion was that it
was not necessary since the sample containers and methodology
were reliable. Further discussions with Dr. Heidt of NCAR
indicate that NCAR and WSU had exchanged halocarbon samples
and that WSU values were running low. Subsequent modifica-
tions to the original set of halocarbon data submitted by
WSU were made for Da Vinci II samples.
B-4 Review of Laboratory and Quality Control Practices at
Research Triangle Institute
The quality control program utilized by RTI during Da Vinci II is de-
scribed in the final report for EPA Contract No. 68-02-2391 and in appendix C
to this report. Further comments relative to the six areas described for
hydrocarbon sampling and analysis for NCAR and WSU are presented below.
1. The analytical methodology employed by RTI consists of con-
centration of sample by liquid oxygen trapping followed by
detection of compounds by gas chromatographic separation and
flame ionization detection. A modified Perkin Elmer Model 900
gas chromatograph and Hewlett Packard 3352 data system are
210
-------
used in the analysis. Halocarbons are analyzed using the same
chromatograph with electron capture detector. The technique
has been described extensively in the literature and utilized
by RTI and EPA in several previous programs.
2. Calibration procedures and protocols are clearly defined and
documented. Calibration frequency is at least once per month
for primary standardization and weekly for retention time veri-
fication. Primary standards are prepared by dilution of pure
compounds with zero air in stainless steel containers («100 1
in volume). Mixtures are also purchased from Scott Research
Laboratories for use in a double dilution system capable of
generating mixtures in the low ppb range. Comparisons are
always run against primary standards prepared from pure com-
pounds. Interlaboratory comparisons are also run between RTI
and ERSL-EPA on a periodic basis.
®
3. RTI utilized Tedlar bags covered by Scotchpak to prevent
light absorption as sampling containers. The stability of
various hydrocarbons in the Cg-Cg range was documented in
appendix A.
®
4. Tedlar bags were purged with zero air and evacuated at least
three times immediately before the flight of Da Vinci II.
After the flight of Da Vinci II, the samples were returned to
RTI for immediate analyses.
5. Documentation was maintained to show a chain of custody from
the time the Tedlar bags left RTI until sample analysis was
complete. Individual records are kept for all samples such
that the time is known to the nearest hour from sample col-
lection to analysis.
6. RTI has maintained a quality control program for all measure-
ments for the last 5 years. Quality control/quality assurance
has been an integral part of all RTI. programs since 1973.
The results of this qualitative evaluation of laboratory and quality
control practices cannot by themselves be used to make judgments on the qual-
ity of the data generated by various participants with respect to hydrocarbon
sampling and analysis; however, they do point up certain deficiencies in the
respective measurement programs. Although the existence or use of a quality
control program does not guarantee the collection and generation of quality
data, it does place more confidence in the measurements.
211
-------
APPENDIX C
RTI QUALITY CONTROL PROGRAM RELATIVE TO
DA VINCI II FOR HYDROCARBON SAMPLING
213
-------
APPENDIX C
RTI QUALITY CONTROL PROGRAM RELATIVE TO
DA VINCI II FOR HYDROCARBON SAMPLING
A quality control program was implemented to determine whether (if) air
®
samples collected in Tedlar bags for subsequent Cg-Cg hydrocarbon analysis
would experience significant contamination from the bag material or constit-
uent loss (by wall permeation or sorption). The sampling protocol for the
program was as follows:
®
1. All Tedlar bags were purged with hydrocarbon-free air prior
to installation on Da Vinci II;
®
2. Tedlar bags were protected from sunlight by an aluminized
material (Scotchpack);
3. Samples were stored after collection in air-tight aluminum
suitcases and transported to RTI in the RTI-EML;
4. The mean time between sample collection and analysis was 8
days; and
5. Samples were analyzed in random order as they came from the
shipping containers.
The quality control program consisted of sets of experiments to determine
®
the potential for contamination of zero air by the Tedlar film and concentra-
tion losses due to sorption or permeation for hydrocarbon mixtures stored in
®
Tedlar bags. These tests and results are described in the following para-
graphs .
Zero air was analyzed directly from a cylinder purchased from Scott Envi-
ronmental and then introduced into a Tedlar bag (Q.C.I). Q.C.I was analyzed
immediately and then shipped to St. Louis in the RTI-EML. Q.C.2 and Q.C.3
were filled using the same zero air cylinder at the field site on May 20 and
June 7. Q.C. bags were returned to RTI and analyzed gas chromatographically.
Results of this study, includng the dates of preparation and analysis, are
presented in table C-l.
The results in table C-l indicate that contamination of zero air from
®
hydrocarbon permeation from the outside and from the Tedlar film itself was
insignificant and suggest that hydrocarbon contamination of ambient air sam-
ples should also be similar. Contamination due to halocarbons such as Freon 11
and 12 was also considered minimal; however, the data show serious problems
214
-------
Table C-l. Contamination study of zero air stored in Tedlar bags
Condition
Date Filled
Date Analyzed
Elapsed Time (days)
Ethane /ethylenet
Propane
Propylene
Acetylene
Butane
1-Butene
Trans-2-Butene
Isopentane
Freon llf
Freon 12
Tricholoroethane
Carbon Tetrachloride
Tetrachloroethylene
Concentration
Cylinder
Analysis
5-13-76
5-13-76
0
0.5
N.D.
0.6
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
42.8
N.D.
N.D.
Q.C.
1
5-13-76
5-13-76
0
0.6
0.1
0.3
N.D.
0.2
N.D.
N.D.
N.D.
N.D.
N.D.
41.3
N.D.
N.D.
Q.C.
2
6-7-76
6-11-76
4
2.2
0.5
0.7
N.D.
1.5
0.6
N.D.
N.D.
9.9
N.D.
43.4
5.0
417.0
Q.C.
3
5-20-76
6-14-76
25
7.8
1.8
0.7
1.2
0.4
N.D.
N.D.
0.2
9.5
N.D.
35.7
6.3
44.3
Q.C.
1*
5-13-76
6-17-76
37
7.0
0.2
N.D.
0.9
0.4
N.D.
N.D.
N.D.
15.0
N.D.
29.8
5.3
205.0
*Q.C.l was reanalyzed 37 days after filling.
tConcentration = ppbV.
Concentration = pptV.
N.D. = nondetectable, <0.1 ppbV.
with the other halogenated compounds analyzed. Therefore, the only halogens
analyzed and reported in this study are Freon 11 and 12.
To examine the storage capability of the Tedlar bags, a blend of three
hydrocarbon mixtures (acetylene, 1-butene, and trans-2-butene) was used to
fill Q.C. bags in the field. The concentration for each hydrocarbon blended
into the bag by dilution of a standard cylinder containing these hydrocarbons
was approximately 71 ppbV. Q.C.4 was blended at RTI, analyzed, and trans-
215
-------
ported to the field. Six additional Q.C. bags were blended using the same
gases and procedure at the St. Louis field site immediately prior to the
launch of Da Vinci II. It is estimated that the blending accuracy for these
bags under field conditions was ±10 percent. These bags were then returned to
RTI for analysis with elapsed times of 10 to 36 days between filling and
analysis. Results of this study are presented in table C-2.
The data in table C-2 show that the stability of acetylene, 1-butene, and
®
trans-2-butene was quite good in Tedlar bags for up to 23 days. The devia-
tion of the analysis results is within the estimated accuracy for blending of
the mixtures in most cases. It should be noted that the matrix in which these
compounds were blended was zero air that was free of hydrocarbons and other
reactive pollutants. Changes in the matrix, i.e., ambient air for field
samples, may have some effect on stability of collected field samples.
Table C-2. Stability of acetylene, 1-butene, and trans-2-butene in
Tedlar quality control bags
Condition/
Constituent
Date Blended
Date Analyzed
Elapsed time in days
Ethane /e thy lene
Propane
Propylene
Acetylene
Butane
1-Butene
Trans-2-Butene
Isopentane
Concentration*
QC-4
5-13-76
5-13-76
0
1.7
N.D.
0.5
55.3
0.3
74.5
68.6
N.D.
QC-5
6-07-76
6-17-76
10
2.9
0.8
0.5
65.3
0.5
76.4
72.5
N.D.
QC-6
6-07-76
6-17-76
10
2.2
0.2
N.D.
65.8
0.7
72.3
69.8
N.D.
QC-7
6-07-76
6-18-76
11
9.6
0.4
0.6
66.8
1.6
74.1
70.3
N.D.
QC-8
6-07-76
6-30-76
23
6.8
0.4
0.8
58.3
0.9
67.2
64.0
1.0
QC-9
6-07-76
6-30-76
23
1.9
0.4
0.3
55.2
0.4
66.9
61.7
0.4
QC-4**
5-13-76
6-18-76
36
116.6
8.5
2.4
61.4
10.5
69.9
63.9
6.9
*Concentration = ppbV.
tSampling bag leaked.
N.D. = nondetectable.
216
-------
The data obtained in this study substantiate previously reported work
regarding sampling of hydrocarbons using Tedlar bags in the Cj-Cs range.—1-
®
In summation, results of the quality control program indicate that Tedlar
bags were satisfactory for collection of hydrocarbons in the Cg-Cs range and
®
Freon 11 and 12, when samples were protected from the sunlight and analyzed
®
within 2 weeks. Sampling bags fabricated from Tedlar were not satisfactory
for the collection of other halogenated compounds.
REFERENCE
C.I R. B. Denyszyn, L. T. Hackworth, P. M. Grohse, and D. E. Wagoner, "Hydro-
carbon and Halocarbon Measurements: Sampling and Analysis Procedure,"
presented at the International Conference on Photochemical Oxidant Pollu-
tion and Its Control, Raleigh, North Carolina, September 12-17, 1976.
217
-------
APPENDIX D
OBJECTIVE ANALYSIS TECHNIQUE
219
-------
APPENDIX D
OBJECTIVE ANALYSIS TECHNIQUE
The objectives of this research require horizontal, vertical, and tempo-
ral analyses of the air chemistry and meteorological data. Those data were
taken at fixed and mobile platforms, continuously and sporadically. A system-
atic objective to collate and interpolate the observation to a given time and
space scale is needed.
Barnes - used a weighted space-time interpolative procedure to ascer-
tain winds, temperature, and moisture in the vicinity of a moving thunder-
.A
storm. The procedure is fairly simple and straightforward. If <)> = <|>(x ,y ,t )
is the scalar variable to be interpolated from a set of observations <|). =
i,n
<|)(x.,y.,t ), then a weighted interpolation analysis technique is written:
w(r. ,r ,t ,t ) ' .
n i* o* n* o vi,n
where r^ = x2 + y .
Barnes chooses the w, the weight function, as
w(ri,ro,tA,tQ) = a exp { - R2/4k*2 - T2/4v2}/8ll3/2k*v
where •>
R
T = t -t
n o
4k*2 - 4k* (1+3 cos2 *)
6 = V./V* for V* > 0, =0 otherwise,
i|» = the angle between the position vector from r to
r. and the wind velocity vector V.
220
-------
and V* = a characteristic speed of the system moving past
the observing locations.
The parameters k and v are chosen according to the data density and the scale
of motions represented by the analyses. —2— The term a represents the confi-
dence in the data. If the data is suspect, then a is small (<1); whereas with
total confidence in the data, a = 1.0.
Weighted interpolation analyses fall short by smoothing the analysis in a
highly variable scalar and by underestimating large values and overestimating
small ones. Maxima and minima occur only near maxima and minima in the input
field. Barnes improves the initial analysis by interpolating the differences
A
of the observed and interpolated values 6. =
-------
APPENDIX E
PLANETARY BOUNDARY LAYER MODEL
223
-------
APPENDIX E
PLANETARY BOUNDARY LAYER MODEL
E-l The Model
The PEL model is a one-dimensional, time-dependent model that incorporates
surface heat flux and a time-independent geostrophic wind through the plane-
tary boundary layer to drive the motions. The model divides the atmosphere
into two regimes--a surface layer from the ground to 10 m, which is charac-
terized as a constant flux layer, and a free atmosphere above. The equations
of motion and thermodynamics are:
3u _ , . 18 pu'w'
= f (v-V )
at v g' P az
c r \ — rr
-r- = f (u -u) - - -r- pv'w1
3t g p dz
II = _ i 1 pw'T1 (3)
8t p 3z
where the symbols have their usual meteorological meaning. Fluctuations in
density have been neglected.
E.I.I Surface Layer
In the lowest 10 m of the model, the nondimensional gradients of wind
speed and potential temperature are obtained from similarity theory:
e-0
= 0.74 {£n(z/z) - ¥ (z/L)}
"*0
where z is the surface roughness length and L is a Monin-Obukov length,
defined by
_ CP P To UI0 (6)
Jj — -1 ~-.
224
-------
with H , u| being the surface fluxes of heat and momentum respectively. The
functions ¥ and H*. are integrals of universal functions .
The value for K, is obtained from
225
-------
The height, H, of the planetary boundary layer was chosen as the first
altitude at which the local Richardson number exceeds 0.25.
E.I, 3 Initial Conditions
The initial conditions for the PBL are given by specifying the vertical
profile of potential temperature, The geostrophic wind components u and v
o o
are assigned at the upper and lower boundaries, and a linear shear is presumed
to exist. The initial vertical wind profiles of u and v were obtained by
either of two methods. The equations of motion were integrated for a period
of 4 h without permitting sensible heat flux from below while keeping the
height of the inversion fixed. This procedure produces a quasi- steady- state
F— 2
wind profile. A second technique used Schaefer's solution for an Ekman
spiral for a given profile of K and geostrophic wind. From the solution pro-
file a new set of K's were computed. This procedure continued iteratively
until a quasi-steady-state was achieved. There appears to be no substantial
difference in the results of the two approaches, although Schaefer's technique
is faster.
The equations of motion and thermodynamics were integrated in time using
E-3
an implicit technique to integrate the diffusion equation with the ageo-
strophic wind component as a forcing function. With the implicit formulation,
the time step was not constrained by the vertical spacing of grid points.
Tests of the model with a 50-m vertical increment and a 10-min time step gave
substantially the same results as those obtained with a time step of 30 s.
Computations for U, v, and 6 at the surface were derived from equations 4
and 5. Fluxes of momentum and heat in the surface layer were included in the
computation of the variables at the 10-m level,
The heat flux at the ground could be described in two different ways in
£-4
the model. Initially, the approach of Busch, Chang, and Anthes was fol-
lowed by specifying the function
0.25 sin [2 TT (t-to)/24]
-w'T' - MAX _ _, Kms
( -U.Oo
226
-------
in testing the model. For simulations of the Da Vinci II flight, however,
w'T1 is an unknown quantity. Consequently the temperature at 10 m was spec-
ified by the linear interpolation between hourly temperatures. Equation (3)
was solved diagnostically for the surface heat flux required to give that
observed change in temperature,
E.2 Verification Results
The equations of motion were integrated in time. Initially, a constant
geostrophic wind (u = 10 m/s, v = o) throughout a layer 1.2 km thick was
O O
used with the heat flux formulation given in equation (15), Integrations were
done using 0.5-, 1,0-, 2.5", 5.0", and 10,0-min time steps without substantial
differences in the resulting profiles, surface fluxes, U^ , or 6^ .
An unexpected problem arises in the iterative solution for the surface
layer terms when the surface heat flux is negative and L is positive. In that
situation, L can be obtained from the solution to
(L + a)3 - bL2 = 0 (16)
when a and b are parameters of the existing conditions. There may be three
real roots to the equation, but each root may not be physically consistent
with the atmosphere. One root to which the solution tended was very small,
making $, much too large. Businger's flux profile relationship of , as a
function of z/L was developed for -2.5
-------
resulting from time integration of equation (17), with a constant X for a
s
time At. The vertical profiles of A. were very similar to profiles of X , un-
S
like the original modeler's results.
After solving those two problems, the meteorological variables were inte-
grated for 36 h, simulating the PEL from dawn on one day to dusk on the fol-
lowing day. The results were very encouraging. The lower portion of the
boundary layer became unstable, the mixing height increased during the first
day, and the wind speed increased in the mixed layer. A stable atmosphere
developed from the ground upward during the night, and surface wind speeds
decreased markedly. Above the inversion, supergeostrophic winds developed in
the marginally stable remains of the afternoon mixed layer aloft. The v com-
ponent of the wind became negative (from the north) as an oscillatory response
to the loss of surface friction at those levels. A low level jet (~12 ra/s)
developed in, rather than above, the stable layer as a result of the relative-
ly thin afternoon mixed layer (~700 m) and the large amount of momentum
(u = 10 m/s at all altitudes) available to the PBL.
o
With the onset of heating on the next day, the inversion layer was lifted
by heating from below, increasing the mixing depth, and gradually approached
the 0 of the previous day's boundary layer. The wind distribution showed the
effects of the presence and absence of surface friction across the thin stable
layer separating the older from the newer boundary layer. When that layer was
destroyed, the mixing height increased rapidly to the previous day's value and
increased slowly thereafter.
The model was modified to incorporate up to five trace substances of
concentration, c, each having an equation of mass continuity of
* . |_ . Snrr . rc (19)
for constant destruction or depletion, rate, F, and for
-w'c? = K ~ z > 10 m
c 8z (2Q)
= constant z <^ 10 ra
The implicit formulation was used to integrate these equations with
KC = K. , coupling the equation of continuity with the meteorological portion
of the model. The equation of continuity for one unspecified material was
228
-------
integrated for 24 h, beginning in the afternoon with a constant value through-
out the PBL (above and below the upper stable layer); with £ = 0; and w'c' =
0.05 ra/s. The results showed a buildup of material during the night in the
surface inversion layer and only minimal vertical diffusions. Vertical mixing
on the following morning gave a decrease in concentrations at and below 10 m
and an increase within the surface mixed layer as a result of increasing the
volume containing the surface emission. The concentration profile followed
the vertical mixing trend as the day progressed, until a deep well-mixed layer
developed by the following afternoon.
REFERENCES
E-l J. A. Businger, "Turbulent Transfer in the Atmosphere Surface Layer,"
Workshop on Micrometeorology, D. A. Haugen, Ed., American Meteorology
Society, pp. 67-98, 1973.
E-2 J. T. Schaefer, "On the Solution of the Generalized Ekraan Equation,"
Mon. Wea. Rev., 10J.:535-537, 1973.
E-3 D. W. Peaceman and H, H. Rachford, Jr., "The Numerical Solution of
Parabolic and Elliptic Differential Equations," J. Soc. Indust.Appl.
Math., 3:28-41, 1955.
E-4 N. E. Busch, S. W. Chang, and R. A. Anthes, "A Multi-Level Model of the
Planetary Boundary Layer Suitable for Use With Mesoscale Dynamic Models,"
J. Appl. Meteor.. 15:909-919, 1976.
229
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-450/3-78-028
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
Project Da Vinci II: Data Analysis and Interpretation
June, 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
C. E. Decker, J. E. Sickles, II, W. D. Bach,
F. M. Vukovich, and J. 0. B. Worth
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
Research Triangle Park, N.C. 27709
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2568
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Monitoring and Data Analysis Division
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Air quality data observed aboard a manned balloon is reported and analyzed, together
with concurrent data from the St. Louis RAMS monitoring network, and from a mobile
van which was driven beneath the track of the balloon. The study was conducted on
June 8-9, 1976, near and downwind of St. Louis, Missouri, during a period of atmo-
spheric stagnation. During daylight hours, ozone levels measured aboard the balloon
and at ground level were similar. At night, ozone trapped aloft by a nocturnal
inversion remains stable, whereas ozone observed at ground level decays rapidly.
Transport of ozone aloft overnight, for distances of at least 180 km, is documented.
With the weak synoptic flow conditions prevailing on June 8, a daytime heat island
effect is documented. Maximum ozone concentrations were observed at a location
where this complex flow field converges.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Ozone
Photochemical Air Pollutants
Primary and Secondary Pollutants
Long-Range Transport
Atmospheric Stagnations
Balloon-Borne Measurements
Vertical Pollutant Profiles
b.IDENTIFIERS/OPEN ENDED TERMS
Atmospheric Ozone Levels
Ozone Formation and
Transport
c. COSATI Field/Group
Atmospheric
Photochemistry
Air Pollution
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Report/
Unclassified
21. NO. OF PAGES
243
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
230
-------
> ~o O
-
TO
O
CD
2
O
; = o 5
,&-=*s
oo
O
m j> ~D m TI ~o
-DID 3 3 CD ;i < ^ wi
------- |