EPA-600/3-77-041
May 1977
Ecological Research Series
THE TRANSPORT OF OXIDANT BEYOND
URBAN AREAS
Data Analyses and Predictive
Models for the Southern
New England Study, 1975
ironmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 Special' Reports
9 Miscellaneous Reports
This report has been assigned to the ECOLOGICAL RESEARCH series This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials Problems are assessed for their long- and short-term influ-
ences Investigations include formation transport and pathway studies to deter-
mine the fate of pollutants and their effects This work provides the technical basis
for setting standards to minimize undesirable changes m living organisms in the
aquatic terrestrial and atmospheric environments
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/3-77-041
May 1977
THE TRANSPORT OF OXIDANT BEYOND URBAN AREAS
Data Analyses and Predictive Models for the Southern New England Study, 1975
by
Chester W. Spicer, James L. Gemma, and Philip R. Sticksel
Battelle - Columbus Laboratories
Columbus, Ohio 43201
Contract No. 68-02-2241
Project Officer
Joseph J. Bufalini
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for use.
n
-------
ABSTRACT
The objective of this study has been to use data collected during the 1975
Northeast Oxidant Study to determine the cause of high ozone concentrations in
the Connecticut River Valley and to develop a method for predicting ozone levels
that can be expected in southern New England under various meteorological con-
ditions.
During the summer months, the prevailing southwesterly winds place the
valley directly downwind of the New York/New Jersey/southwestern Connecticut
urban complex (and on some days the Philadelphia and Washington/Baltimore areas),
The ozone formed from the urban emissions (i.e., the urban plume) was observed
on many case study days to move into Connecticut from the southwest in early
afternoon, cross the Connecticut River Valley, and continue into Massachusetts
during the evening. In one case an 0 -rich air mass was tracked as far north
as the coast of Maine. The dimensions of the urban plumes on several days
were found to vary from 30-80 miles in width and 100-175 miles in length,
seemingly depending on wind speed.
Several methods of predicting ozone in southern New England were investi-
gated including regression integrals, simple regression and multiple regressions.
111
-------
CONTENTS
Abstract iii
Figures '.'.". vi
Tables !!!.!..! !v1ii
Acknowledgment !!!.".'!!!!! x
1. Introduction i
Objectives . '. . 2
2. Ozone in the Connecticut River Valley . ." . 3
Airflow in the Connecticut River Valley 4
July 18, 1975 8
July 19, 1975 14
July 23 and 24, 1975 22
August 10, 1975 45
August 13, 1975 53
August 21, 1975 59
The Relationship Between Ozone and Fluorocarbon-11
in Southern New England 65
3. Statistical Analysis of Ozone in Southern New England ... 68
Regression Analysis 69
Results of Regression Analysis '. 70
4. Summary 87
References 89
Appendix
A. Trajectories of Air Arriving at Groton and Simsbury .... 91
-------
FIGURES
Number Page
1 a-c Ozone distribution in southern New England on July 18, 1975. . . 9-11
2 Fluorocarbon-11 profiles for July 18, 1975 15
3 a-c Ozone distribution in. southern New England on July 19, 1975. . . 17-19
4 Fluorocarbon-11 profiles for July 19, 1975 21
5 Forward trajectories for New York City on July 23, 1975 24
6 Forward trajectories for Philadelphia on July 23, 1975 24
7 a-c Ozone distribution in southern New England on July 23, 1975. . . 26-28
8 Fluorocarbon-11 profiles for July 23, 1975 30
9 a-c Ozone distribution in southern New England on July 24, 1975. . . 32-34
10 a Vertical ozone and temperature profiles at Groton, Conn.,
at 1710 EOT, July 23, 1975 35
10 b Vertical ozone profile NE of Putnam, Conn., at 0955 EOT,
July 24, 1975 35
11 Fluorocarbon-11 profiles for July 24, 1975 37
12 a-d Ozone concentrations at 1000 feet AGL from aircraft 39-42
13 Selected New England ground level ozone profiles -
July 23-24, 1975 44
14 a-c Ozone distribution in southern New England on August 10, 1975. . 46-48
15 Fluorocarbon-11 profiles for August 10, 1975 50
16 Ozone (in ppb) and other pollutant results for afternoon
flight conducted on August 10, 1975 51
17 Ozone concentration for cross-section from the Massachusetts-
Connecticut border to south shore of Long Island - approximately
73° 10' longitude 52
VI
-------
FIGURES
(Continued)
Number Page
18 a-c Ozone distribution in southern New England on August 13, 1975. . 55-57
19 Fluorocarbon-11 profiles for August 13, 1975 58
20 Fluorocarbon-11 profiles for August 21, 1975 60
21 a-c Ozone distribution in southern New England on August 21, 1975. . 61-63
22 Average daily 03 versus average daily F-ll at Simsbury 67
A-l-11 Backward trajectories for Groton and Simsbury, Connecticut . . . 92-102
vii
-------
TABLES
Number Page
1 Comparison of Surface and Upper Air Winds in the Connecticut
River Valley for Selected Days 6
2 Winds at 1000 Feet on July 18, 1975 13
3 Winds at 1000 Feet on July 19, 1975 20
4 Winds at 1000 Feet on July 23-24, 1975 23
5 East-West Ozone Cross-Sections Across the Connecticut River
Valley - July 24, 1975 36
6 Winds at 1000 Feet on August 10, 1975 45
7 Winds at 1000 Feet on August 13, 1975 53
£ Winds at 1000 Feet on August 21, 1975 59
= Linear Regression of Fluorocarbon-11 (PPT) on Ozone (PPB) at
Simsbury and Groton ([Ozone] = m[Fluorocarbon-ll] + b) 66
IT Descriptive Statistics 71
1i Regression Analyses-Measures of Ozone Versus Wind Direction ... 72
".I Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = NO 74
":2 Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = N02 75
"•- Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = CO 76
'.: Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = NMHC 77
",: Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = F-ll 78
Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = Sol R 79
viii
-------
TABLES
(Continued)
Number Page
18 Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = Temp 80
19 Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = Rel. Humidity
(Simsbury) - Dew Pt. (Groton) 81
20 Regression Analyses-Measures of Ozone Versus Individual Predictors
and Wind Direction - Predictor Variable Name = Eth/Acy .... 82
21 Multiple Regression Analyses-Measures of Ozone Versus Combinations
of Predictors and Wind Direction 84
22 Cross-Comparison of Ozone Predictions Based on the Simsbury and
Groton Equations from Table 18 (Temperature Regression -
Ozone in PPB) 86
-------
ACKNOWLEDGMENT
We wish to thank the Environmental Protection Agency - Chemistry and
Physics Laboratory for financial support of this program. Helpful discussions
with Drs. J. Bufalini and W. Lonneman are gratefully acknowledged. The assist-
ance of Battelle-Columbus scientist Darrell Joseph in preparing this report
was much appreciated.
-------
SECTION 1
INTRODUCTION
In recent years elevated ozone concentrations have been observed in many
rural areas which were previously thought to be relatively immune from the
symptoms of photochemical smog. The finding of ozone in rural atmospheres,
far from the generally accepted sources of photochemical smog precursors
(i.e., urban areas), has very important implications in terms of the strategies
devised to control smog formation. Since various strategies for the control
of ozone may differ in efficiency, with some having important social and
economic consequences, it is important that the strategies be conceived with
a thorough understanding of the origins of both rural and urban ozone. A
number of recent field investigations have dealt with the sources of ozone and
the impact of ozone transport in the midwest ^"6\ and far west^7"9^. Results
of these and other studies indicate transport of ozone and its precursors
across regional boundaries is an important source of ozone in downwind rural
and even urban areas. Analysis of data collected in the northeastern United
States suggested that transport of ozone into and within the northeast
is a significant factor in determining peak oxidant concentrations in this
area also.
To study these phenomena in the northeast, and continue its long-term
investigations of ozone formation and movement, the EPA organized and funded
the 1975 Northeast Oxidant Transport Study. The study involved the coordina-
tion and participation of a number of research groups including EPA-RTP
(Research Triangle Park), EPA-LV (Las Vegas), EPA-Region I, Washington State
University, Battelle-Columbus Laboratories, and the Interstate Sanitation
Commission. A number of state and local air pollution agencies also provided
invaluable data and assistance during the field study.
Descriptions of the study and preliminary data reports were published
early in 1976 by the major study participants^4"20^. In addition, the pro-
ceedings of a symposium held in January, 1976, dealing with the preliminary
results of the 1975 study will soon be published(21J. These reports should
-------
be consulted for a detailed description of the study design and tabulations of
the data.
Subsequent to the 1975 field investigation, contracts were awarded to
Battelle-Columbus, Stanford Research Institute, and Washington State University
to analyze selected portions of 1975 Northeast Oxidant Study data in an attempt
to answer some of the important questions relating to ozone formation and trans-
port.
The two topics to be addressed by Battelle-Columbus, and the subjects of
this report, concern
(1) The source of high ozone concentrations observed in the
Connecticut River Valley, and
(2) The, development of a predictive model for ozone in southern
New England.
The specific objectives of this study are defined below.
OBJECTIVES
The investigation described in this report had two main objectives. The
first objective was to determine the cause of the high ozone concentrations
observed in the Connecticut River Valley. The second objective was to develop
a method for predicting the ozone levels that can be expected in southern New
England under various meteorological conditions. Some additional topics are
discuss.ed in this report since they are pertinent to the overall question of
oxidant transport.
The question of high ozone levels in the Connecticut River Valley will
be discussed first, followed by a description of the predictive model developed
for the southern New England area.
-------
SECTION 2
OZONE IN THE CONNECTICUT RIVER VALLEY
In recent years, high concentrations of ozone have been observed in the
Connecticut River Valley^10'11'. The high ozone levels have frequently been
found to occur at progressively later times from south to north up the valley,
thus leading to speculation that polluted air masses might be channeled up
(22}
the valleyv '. This and other possible explanations of Connecticut River
Valley ozone will be discussed in this section of the report.
The effect of New York metropolitan area emissions on Connecticut ozone
levels has been discussed by Cleveland, et al. ' and Rubino, et al. .
Cleveland and coworkers compared maximum daily ozone concentrations measured
during the summer of 1974 with wind directions on the same day. They report
that the highest ozone levels at sites throughout Connecticut and Massachusetts
occurred with wind directions from the New York metropolitan area. Sites as
far away as Boston showed this effect. They also found that air entering the
New York area frequently exceeded the federal oxidant standard, but that the
New York area added substantially to the ozone/precursor burden of the air
entering southern New England.
Rubino, et al. ' describe a Connecticut ozone episode which occurred on
June 10, 1974. They suggest "...the advective transport of 0, and 03 precursors
into Connecticut from New York are probably responsible for a significant por-
tion (approximately two-thirds) of the elevated 03 concentrations measured
throughout Connecticut on days when winds are from the south-southwest direction".
It is evident from these two reports that polluted air moving into
Connecticut from the southwest has a definite impact on Connecticut ozone levels.
In the remainder of this section we will use data collected during the 1975
Northeast Oxidant Study to investigate the extent to which transport affects
ozone levels in the Connecticut River Valley. Other possible causes of high
ozone in the valley, such as channeling of polluted air up the valley and local
generation of ozone within the valley will also be discussed. Our investigations
have focused on seven specific days out of the approximately 38 days of data
3
-------
collected .during the main portion of the 1975 study. Preliminary screening of
the chemical and meteorological data indicated that these seven days would be
the most interesting and elucidating in terms of ozone in the Connecticut
River Valley. The days selected for study include July 18, 19, 23, 24, and
August 10, 13, and 21. These days include all of the important Og episode
periods during the 1975 Northeast Oxidant Study; the air flow during some
portion of each of these days was southwesterly to westerly.
A description of the Connecticut River Valley and its potential effect
on local meteorology is instructive and is presented next. The remainder of
this section is dedicated to a discussion of the individual study days.
AIRFLOW IN THE CONNECTICUT RIVER VALLEY
Terrain heights in Connecticut range from zero along the southern coast-
line adjoining Long Island to around 2000 feet in the northwest part of the
state. The southwest quarter and most of the eastern half of the state have
elevations between 300 and 1000 feet. The Connecticut River, which forms the
border between New Hampshire and Vermont in northern New England flows south
through Massachusetts and bisects the state of Connecticut. The Connecticut
River Valley is the broad shallow depression formed by the flowing river over
many thousands of years. The valley is narrower and deeper in northern
Massachusetts and further north. From central Massachusetts south, however,
the valley is broad and shallow to Long Island Sound. South of Hartford the
river no longer flows within its historical valley, but has formed a new
channel to the southeast. The original valley, which is the subject of this
report, continues south-southeast from Hartford to New Haven.
On either side of the Connecticut portion of the river the hilltops are
between 500 and 800 feet above the river. The difference in height between
these hills and the base of the valley causes the hilltop areas to receive
about four more inches of rain a year than the valley^23). These higher
rainfalls are due to increased convective activity and uplifting of moist air
when easterly winds from the Atlantic Ocean blow against the hills on the
west side of the valley. Terrain is also responsible for the lower rainfall
received in the area northeast of Hartford^24). This area is sheltered from
the easterly winds by an intermediate range of hills and as a result experi-
ences subsiding motion and lower rainfall. Thus the terrain along the
Connecticut River Valley can cause local variations in vertical motion when
winds blow across the valley.
4
-------
A narrow and deep river valley can steer wind flows so that the upvalley
and downvalley directions are clearly predominant. For example, winds at
Albany, New York, during the Northeast Oxidant Study showed a high frequency
of south directions paralleling the direction of flow up the Hudson River
between the laconic Range and the Catskill Mountains. Wind observations from
East Hartford and from Bradley International Airport, which lie in the broad
shallow northern Connecticut portion of the Connecticut River Valley, show
only a slight increase in the frequency of winds from the south as opposed to
other directions. Thus the effect of the Connecticut River Valley on wind
direction was minimal. Any channeling was markedly less than at Albany during
the same period. The possibility that a polluted air mass residing over Long
Island Sound could be transported up the Connecticut River Valley to Hartford
and further north by the afternoon sea breeze off the Sound can be discounted.
While the sea breeze is an important feature of Connecticut's climate in the
late spring and summer, it penetrates inland only 5 to 10 miles.
Two effects on wind are caused by friction with the surface. The wind
speed at the ground is less than the speed above the surface where the retard-
ing frictional force is smaller. Secondarily the wind at the ground blows in
a direction toward the left (in the Northern Hemisphere) of the direction
of the upper wind. The gradient wind level, at which the ground's frictional
effects become insignificant, will depend upon the roughness of the under-
lying surface, but will be on the order of 2000 feet. Comparison of the surface
winds recorded by the National Weather Service and FAA observers at Bradley
International Airport and at East Hartford with the pilot balloon soundings
made at West Springfield, Mass., shows the effects of the Connecticut hills
on the wind flow. Well-developed wind flows (wind speeds of 15 knots equal
to 17 miles per hour) at 2000 feet above West Springfield were accompanied
by winds of about one-third to two-thirds of this speed at the surface of
the valley. Directions at the surface were about 20° to the left of those at
2000 feet.
Pilot balloon soundings routinely measure the wind at 325-meter (1000
feet) intervals. Above 1000 feet the wind should be capable of transporting
pollution across southwest Connecticut with only limited blocking by the hills.
The 1000-foot wind apparently provides a key for explaining pollutant trans-
port into the northern Connecticut portion of the Connecticut River Valley.
Table 1 lists surface wind observations made at Bradley International Airport
-------
TABLE 1. COMPARISON OF SURFACE AND UPPER AIR WINDS IN
CONNECTICUT RIVER VALLEY FOR SELECTED DAYS
THE
= ._«-*«-.. -.- -. ...
UpplT
Air
Wpst Sprimif iiil'J, MISS —
0900, l?00 ami 1500 EOT
M loLipn. N Y -2000 EOT
Date
July 18, 1975
July 23, 1975
July 24. 1975
Aug. 10. 1975
Aug. 11, 1975
Aug. 12. 1975
Aug. 13, 1975
Aug 21, 1975
Time
(EOT)
1200
1400
1500
1700
2000
0900
1100
1500
1700
2000
0900
1100 .
1500
1700
2000
0900
nob
1500
1700
2000
0900
1100
1500
1700
2000
0900
1100
1500
1700
2000
0900
1100
1500
1700
2000
0900
1100
1500
1700
200
iooo"
Direction
(degrees)
180
180
_*
250
210
-
230
180
200
290
290
-
J10
260
210
340
300
000
240
220
MSG5
310
220
230
"feet
Speed"
(knots)
2
13
-
7
13
-
13
22
32
9
14
-
9
9
12
5
7
0
4
9
MSG
1
14
23
2000
Direction
(degrees)
280
190
-
260
220
-
240
190
200
320
290
-
310
270
270
340
310
330
80
200
260
320
220
230
Tcet
Speed
(knots)
5
11
-
15
17
-
22
23
30
18
18
-
17
8
11
17
4
12
6
10
26
11
15
26
Windsor Locks,
Bradley l"t' 1
" uTrcefion
(degrees)
120
190
170
200
220
230
210
190
190
200
260
240
200
220
210
310
310
300
70
230
160
200
200
200
Conn.—
Airport
Speed
(knots)
9
12
9
9
11
7
10
15
13
5
7
6
5
7
7
7
6
5
3
9
5
5
9
8
Surface
Cast Hartford
Pentschler
Direction
(degrees)
200
190
190
190
220
180
220
190
180
210
290
230
210
180
160
350
360
320
000
200
190
NA*
M
NA
, Conn.—
Tower
Speed
(knots)
10
7
6
7
10
5
9
10
14
8
7
5
5
5
5
10
10
8
0
10
8
NA
NA
NA
- = "o sounding madp.
S'1SC = Data missing.
MIA - Data not available for this report.
-------
and at East Hartford at times within 2 hours of the launch times of pilot
balloons at West Springfield, Massachusetts, and at Fort Totten , New York.
West Springfield balloons were generally launched at 0900 and 1500 EOT. Fort
Totten generally also made a wind sounding at 2000 EOT, so its sounding is
included in Table 1 to augment the evening data at West Springfield. From
this table several observations can be made about wind flow during the North-
east Oxidant Study.
(1) When the 1000-foot wind at Springfield is 10 knots (11.5 mph) or
more, the surface winds in Hartford will reflect the direction of the upper
winds, although the surface wind direction may be as much as 30° to the left
of the direction at 1000 feet. The surface winds will also be somewhat slower
than the upper winds. Such a coupling of surface and upper winds is an important
consideration in pollutant transport.
(2) If the Springfield wind at 1000 feet is less than 10 knots, the winds
in the Hartford area may show little resemblance to the upper air winds. Even
the wind directions at Bradley Airport and at East Hartford may disagree. When
the 1000 foot wind speed exceeded 10 knots, the wind direction at Bradley Air-
port and at East Hartford was always similar.
(3) Generally the afternoon (1500 EOT) winds in the valley and at 1000
feet were organized (i.e., the speed at 1000 feet was greater than 10 knots).
This results from the establishment of momentum exchange between the upper air
and the surface. This exchange is missing during the more stable atmospheric
conditions in early morning and night. When the surface and upper winds are
organized, general air movement across the state results. This condition is
well suited to pollutant transport.
In terms of pollutant transport across Connecticut, the foregoing discussion
leads to several general conclusions:
The topography and the wind data suggest that channeling of polluted air
up the Connecticut River Valley is unlikely. If the wind speed in the Connecticut
River Valley is greater than 10 knots, winds are organized and transport of
pollution is likely. With a persistent southwest wind of 10 knots the emissions
from metropolitan New York can reach central Connecticut in about 8 to 10 hours.
Obviously, wind speeds in excess of 10 knots can result in even more rapid
pollution transport.
-------
**
The remainder of this section of the report will be devoted to discussions
of the 7 days selected for detailed analysis.
JULY 18, 1975
July 18, 1975, was a hot, humid, hazy day in the Connecticut River Valley.
A high pressure ridge begain to build over New England during the day and a
weak upper level subsidence set in. Wind flow below 5000 feet was from the
Southwest throughout most of the region. Figures A-l and A-2 in Appendix A
show calculated air mass trajectories* for Simsbury and Groton, Connecticut,
for several hours during this day. A description of the techniques used to
derive these trajectories may be found elsewhere^ 5'. A discussion on inter-
pretation of the trajectories is included in Appendix A. It is clear from
the trajectories that the flow throughout the day was from the southwest.
The air arriving at Groton during the daylight hours had generally passed over
the northern New Jersey-New York City area within the previous 5-7 hours. The
air arriving at Simsbury passed well to the north of the New York metropolitan
area during most of the day, but by evening air arriving at Simsbury had
passed very near the urban boundaries.
In order to discuss the origin of high ozone within the Connecticut River
Valley, we have derived maps which display the ozone distribution throughout
southern New England. The maps for July 18 are shown in Figure 1 a-c. The
ozone data shown by these maps are primarily from the more than 30 ground-
level monitoring stations in Connecticut, Massachusetts, and Rhode Island.
The data from these stations are shown in the maps in ppb. Where avaiable,
aircraft ozone data have been used to fill in questionable portions of the
maps. Since the aircraft data were taken at 1000 feet above ground, there
is a risk of inconsistency; to minimize this risk we have generally used
aircraft data taken over nonurban areas. Ozone data from New York State has
also been used in deriving the maps when the ozone patterns in the western
portion of the region were unresolved.
trajectories courtesy of Stanford Research Institute.
Simsbury and Groton were two of the special study sites during the Northeast
Oxidant Study. Approximately 38 days of continuous chemical and meteorological
data were collected at these sites by heavily instrumented mobile laboratories.
8
-------
J JLV 13, 1 '75
1 2 Z C E 5 ~
D ZG r.-E < c-rZ )
) i Density
i> D 0-50ppb
I
i [jx] 5O-IOOppb
3 I00-I50ppb
3 I50-200ppb
^ 200-250ppb
^ 250-300ppb
• >300ppb
Figure la. Ozone distribution in southern New England on July 18, 1975.
-------
I Density
I
Q 0-50ppb
j£3 50-lOOppb
9 I00-I50ppb
3 I50-200ppb
3 2OO-250ppb
3 250-300ppb
• >300ppb
Figure Ib. Ozone distribution in southern New England on July 18, 1975.
10
-------
7
J '_ L y 16, 1975
2100 E ST
0 ZO NE < PPB )
[ Density
0-5Oppb
jxjil 50-IOOppb
j Q IOO-l5Oppb
!50-20Oppb
200-25Oppb
250-300 ppb
>300ppb
Figure 1c. Ozone distribution in southern New England on July 18, 1975.
11
-------
The-map shading shows ozone concentrations in 50 ppb increments, from
zero to >300 ppb. A legend which categorizes the shading is provided on each
page. For the most part the data shown on the maps for each station was followed
rigorously in determining the concentration contours; however, where data looked
suspicious or led to highly complex patterns, some smoothing was done. The
need to smooth the contours by ignoring a station's data was very rare with
three exceptions. The Springfield, Mass., and Litchfield, Conn., results were
often completely inconsistent with the rest of the data, and have been ignored
in drawing the contours. In addition, the results from Windsor, Conn., were
frequently much lower than the surrounding area. To avoid highly complex
contour patterns some of the Windsor data were excluded.
Ozone data over Long Island Sound and the Atlantic Ocean were not usually
available so that shading in these areas is by inference. In most cases the
shading patterns have been terminated in these areas (extreme right side and
bottom of the maps); lack of shading in these areas does not indicate low
ozone but rather lack of data.
The ozone distribution plots for July 18, 1975, are shown in Figure 1.
Plots were constructed for every third hour from 0900 to 2400 EST. Referring
to the maps, the concentration of 03 along the southern New England coast at
0900 EST was generally about 50 ppb, with inland concentrations of 50-80 ppb.
By noon some areas of very high ozone had developed in Connecticut. The
trajectories mentioned earlier, and the general meteorological data both
suggest that the air mass moving into Connecticut at noon was located over
the New York metropolitan area during the 6-9 a.m. rush hour period, when
ozone precursor emissions are highest. By 1500, areas of high ozone extend
from southern and eastern Connecticut through Rhode Island and up into the
vicinity of Boston. The area of particular concern to this study, the
Connecticut River Valley, shows high levels of 03 from New Haven up to about
Springfield, Mass. However, there is nothing unique about the river valley
in terms of high 0-j concentrations; high levels of ozone also exist outside
of the valley. This is a very important point in terms of whether a
"channeling" effect occurs in the valley. If ozone concentrations outside the
valley are similar to those within, it is very difficult to build a case that
the valley funnels pollution into itself or is in any way unique. As mentioned
earlier, channeling is one mechanism which has been suggested to explain high
Oo in the valley.
12
-------
By 1800 EST elevated 03 concentrations exist from west-central Connecticut
to the Boston area, with the highest values extending eastward about 35 miles
from Hartford. At 2100 the area of high 03 is in eastern Massachusetts and by
midnight only the northeast portion of the region has ozone levels >50 ppb.
It is apparent from these maps that an air mass of high ozone/precursor
concentration entered southern New England from the southwest on the morning
of July 18 and moved through the region from southwest to northeast during the
remainder of the day. Emissions from the urban areas of Connecticut undoubtedly
contributed to this air mass and may be largely responsible for the elevated
ozone in eastern Connecticut, Rhode Island, and east-central Massachusetts at
1500 EST. In terms of the Connecticut River Valley situation, it seems clear
that a polluted air mass moving across the valley from the southwest resulted
in high valley 03, and that nothing unique to the valley's topography or meteor-
ology led to the high values. The lack of evidence for a channeling effect is
consistent with our analysis of the valley's topography and meteorology pre-
sented in a previous section.
It is initially surprising that an air mass moves rapidly enough to
traverse the approximately 200 miles between New York and Boston during the
course of a single day, as these maps seem to suggest. Table 2 shows the
1000-foot wind directions and speeds at several locations during July 18.
Higher altitude winds are generally even higher in speed. During much of the
day the winds averaged at least 15 mph within the important surface-to-5000-
foot transport layer. At this speed, a polluted air mass could easily travel
200 miles over a single day.
TABLE 2. WINDS AT 1000 FEET ON JULY 18, 1975
Time
0700
1300
1430
1900
Location
Chatham, Mass.
Springfield, Mass.
Putnam, Conn.
Chatham, Mass.
Springfield, Mass.
Putnam, Conn.
Avery Point, Conn.
Chatham, Mass.
Putnam, Conn.
Speed,
mph
8
2
8
23
15
14
17
29
14
Direction,
degrees
210
180
271
250
177
243
260
245
220
13
-------
Some additional interesting features of the ozone patterns on July 18 are
apparent from the shaded maps. The pattern of rapid 03 depletion in south-
western Connecticut between 1500-2100 EST is suggestive of the rapid scavenging
of 03 by NO emitted in the evening traffic rush. We have observed this effect
elsewhere1 • '. The NO responsible for 03 depletion in this case is probably
a combination of local emissions and emissions transported from the New York
metropolitan area.
Another interesting feature is the seemingly anomalous behavior of ozone
in the Simsbury area. The 03 remains comparatively low during the afternoon
at Simsbury, even while concentrations of 200 ppb were recorded only 15 miles
to the southeast in Hartford. By 2100 EST however, the higher concentrations
of 03 had reached Simsbury. Recall that the trajectories showed Groton receiving
urban air most of the day but Simsbury not until evening. The late arrival of
the urban air at Simsbury may explain the unusual 03 behavior. A plot of the
fluorocarbon-11 profiles at Simsbury and Groton shown in Figure 2, confirms
this.* The F-ll concentration increases to very high levels during late
morning and early afternoon at Groton, indicating a direct influx of polluted
urban air. As might be expected, the concentration of 03 also reached very
high values by 1500 that afternoon.** Based on the F-ll profile in Figure 2,
Simsbury did not receive its infusion of urban air until 1900-2000 EST. For
this reason the Simsbury 03 peak did not occur until well after dark (132 ppb
at 2000 EST). Since the reactions forming 03 are completed by this time, the
peak 03 and F-ll concentrations occur simultaneously at 2000 EST at Simsbury.
The occurrence of such high levels of 03 after dark in a rural area like
Sijnsbury, combined with the simultaneous peak in the urban air tracer, F-ll,
is very strong evidence of the transport of urban pollution to rural areas.'
JULY 19, 1975
The meteorological situation on July 19, 1975, was very similar to July 18.
Wind flow was still from the southwest throughout most of the day, but the
wind speed was somewhat greater. The Simsbury and Groton trajectories from
-------
500
400 -
9
12 13 14 15 16 17 18 19 20 21
Time of Day, (EST)
Figure 2. Fluorocarbon-11 profiles for July 18, 1975.
22 23 24
15
-------
Figures A-2 and A-3 in Appendix A are also very similar to the trajectories on
July 18. Flow into Connecticut during most of the day was from the southwest;
the direction of the New Jersey-New York urban complex.
The ozone shading maps for July 19 may be found in Figure 3 a-c. The
ozone patterns at 0900 EST still show the effect of the previous evening's
high ozone in the northeastern part of the region. It is uncertain how much
of this ozone is truly residual, that is, surviving from the previous evening,
and how much is a result of an increase in the morning rate of ozone formation
due to a more favorable N02/N0 ratio caused by nighttime reaction of residual
03 with NO. The diurnal ozone profiles from such sites as Fitchburg and
Lowell, Mass.*, indicate that significant surface concentrations of 03 did
exist overnight at these locations. Since the overnight concentration of
ozone within stable layers aloft was probably even higher than the surface
concentration, it seems plausible that much of this morning ozone is actually
left over from the previous evening.
The concentration of 03 within the Connecticut River Valley at 0900 is
generally low. By noon however, there is a definite intrusion of ozone-rich
air into Connecticut from the southwest. The greatest concentrations exist
around Bridgeport, but it is clear that the entire southern portion of the
Connecticut River Valley from New Haven to Hartford is affected. At 1500
the highest levels of 03 are in the vicinity of Hartford, about 40 miles
northeast of Bridgeport. However, high concentrations of 03 (>100 ppb)
exist within a band from southwestern Connecticut to northeastern Massachusetts.
The fact that wind speeds averaged more than 30 mph throughout the day is
entirely consistent with the hypothesis that this band of high ozone represents
the 'smeared-out urban plume from the urban complexes in New Jersey, New York,
and southwestern Connecticut. Table 3 shows representative wind information for
July 19. The band of high 03 at 1500 EST extends across the Connecticut
River Valley, with high levels found within and on either side of the valley.
As on July 18, there is no indication of any unique feature of the valley
which accounts for the observed high 03; rather the high 03 in the valley seems
to relate to the geographical location of the valley downwind of major urban
centers.
The band of high 03 is somewhat smaller and has moved north and slightly
eastward by 1800 EST. Simsbury data show a shift in the wind from southwesterly
to southerly starting around 1600 EST. This probably explains the northward
16
-------
j ULV 19, i =, r;
•» 0 0 E S T
0 ZO NE ( r PB >
J -J L V 15, 1 r 7 5
0 C E I '
Q ZO NE (
i Density
„"?. i
=r~ J, i D 0-50ppb
' 13 50-100 ppb
100-150 ppb
I50-200ppb
200-25Oppb
250-300 ppb
>300ppb
Figure 3a. Ozone distribution in southern New England on July 19, 1975.
17
-------
V, : Density
0-5Oppb
50-100ppb
IOO-l50ppb
I50-200ppb
200-250 ppb
250-300ppb
>300ppb
Figure 3b. Ozone distribution in southern New England on July 19, 1975.
18
-------
JULY 19, 1975
2 ! 0 0 E S T
0 ZO N'E ( P F 8 >
! Density
H 0-50ppb
: E3 50-100 ppb
33 I00-I50ppb
3 !50-20Oppb
3 20O-25Oppb
•J3 250-30Oppb
| >300ppb
Figure 3c.
Ozone distribution in southern New England on July 19, 1975.
19
-------
TABLE 3 . WINDS AT 1000 FEET ON JULY 19, 1975
Time
0700
0900
1700
1900
Location
Chatham, Mass.
Avery Point, Conn.
Avery Point, Conn.
Chatham, Mass.
Speed,
mph
30
25
33
41
Direction,
degrees
250
247
247
230
movement. The highest 03 at this time is at Fitchburg, Mass. Since the photo-
chemical reactions which generate 03 are essentially terminated by this time
in the evening, some scavenging and decay of 03 are undoubtedly occurring;
thus the band of high concentrations is shrinking. By 2100, all of the surface
stations report concentrations less than 100 ppb. The highest concentrations
in the region are in the rural areas of central Massachusetts and northwestern
Connecticut, and are clearly the residue of the ozone band observed entering
these areas at 1800. By midnight, concentrations throughout the region are
less than 60 ppb.
The daily average fluorocarbon-11 concentration at Simsbury was the same
on the 19th as on the 18th. However, the profiles differ markedly. The
fluorocarbon-11 concentration on the 19th, as shown in Figure 4, was virtually
constant throughout the day, as opposed to the late evening surge which occurred
on July 18. The constant F-ll profile suggests a steady influx of dilute urban
air throughout the day at Simsbury. This is consistent with the 03 shading
maps, which show no unusual 03 contours or gradients.near Simsbury on the
19th (as were observed on the 18th). Judging from the F-ll patterns, the
input of urban 03 precursors to Groton is less than Simsbury on the 19th and
considerably less than Groton on the 18th. This undoubtedly explains the much
lower 03 at Groton on this day.
In summarizing the 03 patterns of July 19 in terms of the Connecticut
River Valley situation, it is important to emphasize the cross-valley nature
of the high 03 area (especially visible at 1500 and 1800 EST). It seems
apparent that the high 03 in the valley on this day was not related to the
valley itself, but the fact that a portion of the valley was located directly
downwind of a major emissions complex to the southwest. Under the meteorological
conditions which existed on this day, a smeared-out urban plume extended across
20
-------
50O
400
300
200
„_ 100
a.
o.
§
.a
i—
o
o
o
500
400
300
200
100
Simsbury
Groton
1
12 14 16 18 20
Time of Day (EST)
Figure 4. Fluorocarbon-11 profiles for July 19, 1975.
21
22
24
-------
the valley; high levels of ozone were observed both within and on either side
of the valley. The 03 levels within the urban plume are superimposed on the
03 which would form in the absence of the plume (about 70 ppb judging from the
concentrations on either side of the plume). The ozone which exists outside
the urban plume may result from local precursor emissions, long range transport
(i.e., associated with a high pressure system), natural sources or some com-
bination of the three. The interaction and superposition of 03 and precursors
from these sources was discussed in our 1974 midwest study report^4'5).
Some feeling for the spatial extent of urban plumes can be gained from
the ozone maps for July 18 and 19. On July 18 the maximum diameter (per-
pendicular to the wind flow) of the urban plume (03 >100 ppb) was between
50 and 80 miles. The length of the plume was in excess of 100 miles. On
the 19th, with considerably higher wind speeds, the plume diameter was only
30-40 miles, with a length greater than 175 miles. Of course, the spacing
between the urban centers contributing to this plume undoubtedly contributes
to its overall length.
JULY 23 AND 24, 1975
July 23 and 24 will be treated as a single episode here for clarity. A
high pressure system moved eastward through the region on July 23 causing a
southwest to west-southwest surface flow pattern. By July 24 the high began
to diffuse and winds became more southerly. Representative wind data at 1000
feet are presented in Table 4.
Air mass trajectories for these 2 days are shown in Figures A-4 - A-5 in
Appendix A. The air parcels arriving in Simsbury and Groton during the early
part of July 23 had passed through the fairly rural south-central areas of
New York State. By midafternoon however, the flow had shifted to the south-
west, and the 1900 EST trajectory at Groton passed directly over the New Jersey-
New York-Connecticut urban complex known as the New York metropolitan area.
The air arriving in Simsbury passed well north of the urban complex. A set
of forward trajectories for New York City and Philadelphia were reported by
Wolff, et al. and are included here as Figures 5 and 6. The New York City
trajectories show that air leaving the city after 0800 moves into Connecticut
from the southwest. Air leaving the metropolitan area during the morning
peak traffic period (trajectory B) arrives in the Bridgeport-New Haven area
by midafternoon. Air from Philadelphia (Figure 6) has little impact on
22
-------
TABLE 4. WINDS AT 1000 FEET ON JULY 23-24, 1975
Time
0700
0800
0822
0830
1300
1355
1400
1900
0700
0809
0183
1340
1400
1900
Location
Winds at 1000 Feet on
Chatham, Mass.
Springfield, Mass.
Putnam, Conn.
Avery Point, Conn.
Chatham, Mass.
Avery Point, Conn.
Springfield, Mass.
Chatham, Mass.
Winds at 1000 Feet on
Chatham, Mass.
Springfield, Mass.
Putnam, Conn.
Avery Point, Conn.
Springfield, Mass.
Chatham, Mass.
Speed,
mph
July 23,
14
8
7
16
17
17
17
20
July 24,
28
15
13
15
25
36
Direction,
degrees
197S
270
255
267
w V /
305
250
262
205
260
1975
250
228
230
240
184
205
Connecticut during July 23, but does enter the state early on the morning of
the 24th after passing through the New York metropolitan area.
Figure 7 a-c shows the ozone distribution maps for July 23. Ozone at
0900 is less than about 60 ppb throughout southern New England. At noon, very
high levels of 03 are observed entering the region from the southwest. By
1500 a plume of elevated 03 (>100 ppb) is found from southwestern Connecticut
to eastern Massachusetts. The highest concentrations, and these were the
highest levels observed during the 1975 Northeast Oxidant Study, are found
near New Haven. Recall that the New York forward trajectories predicted that
the air over the city during the morning rush hours would arrive in New Haven
by midafternoon. This corresponds precisely with the New Haven ozone maximum.
Note also that the band of high 03 extends across the Connecticut River Valley,
with very high levels of 03 found within and on either side of the valley. It
seems very likely that the high 03 in the Connecticut River Valley results
from the geographical fact that the valley lies downwind of a complex of major
urban emissions sources, and at an optimum downwind distance which permits
extensive photochemical reaction and 03 generation within the pollutant-laden
morning air mass. The combination of these factors allows extensive 03 formation
23
-------
MASSACHUSETTS
/ .
i —-—T A
I CONNECTICUT
\ R. /.
1
Trajectory Starting Time (EST)
}
}
1 r
/ A
/
B
C
0100
0700
1300
1900
O
Figure 5. Forward trajectories for New York City on July 23, 1975.
24
-------
MASSACHUSETTS
-—\
n
Trajectory Starting Time (EST)
A 0100
B 0700 \
C 1300 *
D 1900 *
Figure 6. Forward trajectories for Philadelphia on July 23, 1975.
25
-------
JULY 23, 1975
900
0 ZONE ( PPB)
Density
0-50ppb
|{!ix| 50-IOOppb
IOO-l5Oppb
I50-200ppb
20O-250ppb
250-300ppb
>300ppb
Figure 7a. Ozone distribution in southern New England on July 23, 1975.
26
-------
JULY 23, 1975
1500 H5T
C ZO NE ( = F- B )
J LILY 2 3, 197;
1 8 & C
0 ZD tj-: < r ? 6 ;
Density
0-50ppb
50-IOOppb
I00-I50ppb
I50-200ppb
200-250ppb
250-300ppb
>300ppb
Figure 7b. Ozone distribution in southern New England on July 23, 1975.
27
-------
J ULV 23, 1975
2100 EST
0 ZONE < PPB )
JULY 23,
2"CC
CZQNE
Density
0-50ppb
50-100 ppb
IOO-l50ppb
!5O-2OOppb
2OO-25Oppb
250-3OOppb
> 30Oppb
Figure 7c. Ozone distribution in southern New England on July 23, 1975.
28
-------
to occur by the time the air parcel reaches the river valley. Obviously the
03 and precursors moving into the source area (New Jersey-New York) in the
morning contribute to the ultimate ozone burden within the urban plume. On
some days this contribution can be significant, for example, when the Philadelphia
urban plume overlaps the New York metropolitan area. The trajectories suggest
this is not occurring on the afternoon of July 23. This topic will be discussed
again shortly.
Other sources of 03 and precursors such as regional 03 associated with
high pressure cells also contribute to the ultimate concentration within the
plume. However, judging from the 03 levels outside of the urban plume during
the afternoon of July 23 (70-85 ppb), it is clear that the dominant source of
the very high (>200 ppb) 03 within the plume is upwind urban emissions.
By 1800 (7 p.m. local time) the photochemical reactions producing 03 are
terminated, yet extensive portions of Connecticut, Rhode Island, and parts of
Massachusetts are still experiencing high Og. The center of the 03 distri-
bution extends across the Connecticut valley from New Haven to the northeast
corner of the state. Note that Groton experiences high levels of 03 during
the afternoon, levels about twice those found at Simsbury. These observations
are consistent with the trajectory analysis, which showed Groton receiving
direct input of urban air, in contrast to Simsbury which experienced relatively
clean rural emissions.
By 2100 EST the urban plume extends from the eastern Connecticut valley
up to northeastern Massachusetts. Three hours later, at midnight, high con-
centrations exist only in eastern Massachusetts north and west of Boston.
This is about the distance that the wind speed data from Table 4 predict
the morning New York air mass would travel in the intervening 15-16 hours.
We will shortly use aircraft data to track this air mass further north
and east of Massachusetts. Before doing so however, we will first discuss
the fluorocarbon data for July 23 and the July 24 03 distribution maps.
The fluorocarbon-11 profiles for Simsbury and Groton are presented in
Figure 8. The F-ll concentration in Simsbury was near the tropospheric
background level (.90-120 ppt) during most of the day, confirming the absence
of urban air influx to Simsbury. After 1800 there apparently was an infusion
of urban air; at about this same time 03 reached its maximum at Simsbury.
The F-ll concentrations at Groton were higher than at Simsbury during most
of the day, consistent with the trajectory analysis and the higher 0, levels
at Groton. 3
29
-------
500
40O -
300 -
ex
Q.
2OO -
100
S 500
o
i_
o
3
400
300
200
100
t.i I - ? '
8
I
10
I
I
12
Groton
I
20
14 16 18
Time of Day (EST)
Figure 8. Fluorocarbon-11 profiles for July 23, 1975.
22
24
30
-------
Turning our attention now to the following day, the Simsbury and Groton
air mass trajectories for July 24, 1975, are shown in Figure A-5 in the
Appendix. The trajectories inoicate that the air arriving at Simsbury during
most of the day had passed over the major metropolitan complexes of New York
and Philadelphia, and to some extent even Washington and Baltimore. The air
arriving in Groton during the daytime and evening hours has passed near Baltimore,
up the Atlantic coast of New Jersey and across eastern Long Island. Groton
apparently receives very little input from New York and Philadelphia on July 24.
The forward trajectories from Philadelphia, shown earlier in Figure 6, are in
substantial agreement. Air which left Philadelphia at 1400 on July 23 passes
over New York in the middle of the night and enters Connecticut in the early
morning hours of July 24. On a day such as this it is very possible that an
overlapping of urban plumes within the Washington-Boston corridor could occur.
Thus, 0-j entering southern New England on July 24 could be the result of
emissions from several upwind urban areas. These overlapping urban plumes
will be superimposed on any regional 03 which might result from emissions
several hundred miles upwind (e.g., the midwest).
The ozone distribution maps for July 24 are pictured in Figure 9 a-c.
At 0900 we can see what are probably the remains of the previous day's high
ozone band. Ozone from one day frequently survives overnight by being trapped
aloft above the nocturnal inversion away from scavenging surfaces and surface-
based scavenging emissions. This ozone aloft then fumigates the surface after
the inversion breakup the next day. That this phenomenon occurred on July 23-
24 can be seen from verticle 03 profiles during the evening of July 23 and
the morning of July 24. These profiles are shown in Figure 10 a and b. Figure
lOa shows a very concentrated layer of 03 at 2000 feet MSL above Groton at
1600 hours. The 03 values in this layer aloft (elevated urban plume) are
nearly twice the surface concentrations. Thus a reservoir of O-j exists aloft
for possible isolation and survival overnight. A profile obtained at 0855
the next morning over Putnam, in the remote northeast corner of Connecticut,
is shown in Figure lOb. It is evident that a reservoir of Og still exists
aloft (2000-2500 feet MSL) and, because of the early hour, this 0- must have
survived from the previous day. It is highly probable that the high 03 con-
centrations observed in east-central Massachusetts at 0900 on the morning of
July 24 are the remains of the previous day's urban plume, which was "stored"
aloft overnight.
31
-------
u_v
90C EST
0 ZO HE ( PPB)
JULY 2 "» , 1
1200 EST
0 ZC F
Density
Q O-50ppb
50-100 ppb
100-150 ppb
l50-20Oppb
2OO-25Oppb
250-300 ppb
>300ppb
Figure 9a. Ozone distribution in southern New England on July 24, 1975.
32
-------
J ULV 2H, 19
1500 EST
0 ZO NE ( PPB )
JULY 21, 19
1 80 0 ES
0 ZC r;E < P?B
Density
0-50ppb
m 50-100 ppb
100-150 ppb
j £3 I50-200ppb
200~25Oppb
250-300ppb
>300ppb
Figure 9b. Ozone distribution in southern New England on July 24, 1975.
33
-------
J ULV 21,
2100 EST
0 ZO NE ( PPB )
J ULV 2 t,
2^00 EST
OZONE (PPB)
Density
Q 0-50 ppb
^ 50-IOO ppb
g lOO-l50ppb
H I50-200ppb
3 200-250ppb
|g 250-300 ppb
• >3OOppb
Figure 9c. Ozone distribution in southern New England on July 24, 1975.
34
-------
8
7
^ 6
i
if 5
ro
O
-------
Referring to the 03 distribution maps, we see that by noon on the 24th,
high concentrations of 03 exist throughout western Connecticut and southwestern
Massachusetts. This is consistent with the trajectories and our earlier obser-
vation of a shift toward more southerly flow on the 24th. High concentrations
of ozone are found within the Connecticut River Valley and also to the west
of the valley (e.g., Torrington and Danbury).
By 1500 the area of high ozone concentration has spread further northward,
with fairly remote areas of northwest Massachusetts exceeding the federal
standard by wide margins. At 1800 only far western Massachusetts and part of
eastern New York showed 03 in excess of 100 ppb. Subsequent to 1800 EST, the
03 levels throughout the region dropped to well below the 80 ppb standard.
As on the previous days that we have discussed, no special effect of the
Connecticut River Valley is evident from the 03 distribution data. Levels of
03 found within the valley were no greater than those found outside. A
tabulation of ozone concentrations across the valley illustrates this point
and is included in Table 5. Once again the high concentration of ozone found
within the valley in early afternoon and in remote northwestern Connecticut
and western Massachusetts later in the day seems to originate from a source
upwind (southwest) of the region. As discussed earlier, the source of pre-
cursors could be any one of several upwind urban areas. More likely several
upwind urban areas each contributed in different degrees to the ultimate
burden of 03 in the air mass.
TABLE 5 . EAST-WEST OZONE CROSS-SECTIONS ACROSS THE
CONNECTICUT RIVER VALLEY - JULY 24, 1975
West of Valley Within Valley East of Valley
Time (Station) (Station) (Station)
1200 145 ppb (Danbury) 120 ppb (Hamden) 100 ppb (Groton)
1200 140 ppb (Torrington) 125 ppb (Windsor) 95 ppb (Eastford)
1500 168 ppb (Pittsfield) 135 ppb (Greenfield) 120 ppb (Aircraft)
The fluorocarbon profiles from Groton and Simsbury on July 24, shown in
Figure 11, are not very revealing due to lack of structure.
A somewhat different perspective on 03 transport during the July 23-24
episode may be gained from combining the aircraft data collected by Battelle-
36
-------
500
400 -
300 -
200 -
100 -
o
-Q
500
4OO
3OO
2OO
100
Groton
I
I
I
I
8
10
12
Figure 11
20
22
14 16 18
Time of Day (EST)
Fluorocarbon-11 profiles for July 24, 1975
37
24
-------
Columbus and Washington State on these days. Using area averages to represent
different segments of the flight patterns, we obtain the maps of Og distribution
at 1000 feet AGL shown in Figure 12 a-d. Referring to these maps we see that
O-j concentrations were moderate throughout the region during the morning of
July 23. By the afternoon of the 23rd the concentration of 0, over southern
Connecticut and Long Island Sound reached extremely high values (Figure 12b),
in agreement with the surface Og distribution maps shown earlier. On the
morning of July 24,Figure 12c shows the Og-rich air mass residing over northern
and northeastern Massachusetts. We have already demonstrated the likelihood
that this morning's 03 resulted from the previous afternoon's urban plume which
survived aloft overnight.
The aircraft data on the afternoon of July 24 are found in Figure 12d and
are quite interesting. Recall from the Og maps in Figure 9 that a new urban
plume has formed during the afternoon on the 24th, causing high levels of Og
in western Connecticut and Massachusetts. The aircraft data do not show the
new urban plume because the planes were still tracking the plume from the
previous day. It is clear from Figure 12 that high concentrations of 03 had
moved northeastward up the Atlantic coast to areas northeast of Portland, Maine.
Vertical profiles during the flight over the ocean showed the maximum Og was
concentrated in a fairly thin layer about 2000 feet above the sea. With south-
west to south-southwest winds of about 20 mph existing throughout the region on
July 24 (Table 4) the Og-rich air mass residing over northeastern Massachusetts
in the morning would require about 8 hours to reach the area northeast of
Portland in late afternoon. Eight hours is about the time difference between
Figure 12c and Figure 12d. Thus it seems likely that the high Og in the
air mass found off the coast of Maine in late afternoon was from the same
Og-rich air observed in northeast Massachusetts in the morning and was partly
the remnants of the metropolitan New York urban plume from the previous day.
Urban areas along the trajectory, most notably Boston, probably contributed
significantly to the precursor burden of the air mass (the effect of overlapping
urban plumes discussed earlier). However, the existence of high levels of Og
early in the morning in northeastern Massachusetts, observed before locally
emitted precursors could have generated significant Og, suggests that the
contribution of the previous day's urban plume is significant. The inter-
action between the aged plume from the previous day and fresh local emissions
can also be an important factor in the rate of Og generation (e.g., by affect-
ing the N02/NOX ratio) as the overlapping plumes move downwind.
38
-------
Morning
July 23, 1975
(03 in ppb)
Figure 12a. Ozone concentrations at 1000 feet AGL from aircraft.
39
-------
Afternoon
July 23, 1975
(03 in ppb)
Figure 12b. Ozone concentrations at 1000 feet AGL from aircraft.
40
-------
Morning
July 24, 1975
(03 in ppb)
Figure 12c. Ozone concentrations at 1000 feet AGL from aircraft.
-------
60
O
HARTFORD
Afternoon
July 24, 1975
(03 in ppb)
Figure 12d. Ozone concentrations at 1000 feet A6L from aircraft.
42
-------
One final demonstration of ozone transport during this 2-day episode can
be obtained from the diurnal ozone profiles from the surface stations. Profiles
of 0-j from stations along the southwest-to-northeast air mass trajectory between
southwestern Connecticut and northeastern Massachusetts should show progressively
later Og maxima. Profiles from several ground stations have been plotted together
in Figure 13. Stations from southwestern Connecticut (Bridgeport at the top of
the figure) to northeastern Massachusetts (Salem near the bottom of the figure)
are shown, along with a background site (Pittsfield) from western Massachusetts
which is well out of the path of the July 23 urban plume (Pittsfield is in the
path of the July 24 plume, as shown earlier). Lines have been drawn through
each profile at the 80 ppb level for reference. It is clear from the figure
that there is a progression in the time of maximum Og along a southwest to
northeast trajectory. Since these are surface data, the nighttime readings
may be misleading due to scavenging below the nocturnal inversion. However,
the occurrence of the Oj maximum early on the morning of July 24 in Framingham,
Cambridge, and Salem, suggests that this Og originated on the previous day and
was transported into the area.
To summarize briefly, the July 23-24 ozone episode in southern New England
can be attributed to the combination of (1) urban plumes entering the region
from the southwest, (2) regional 0., associated with the residing high pressure
system, and (3) Oj generated from local emissions. During this particular
period, urban plumes seem to be the predominant contributor. No special effects
of the Connecticut River Valley were apparent; high ozone in the valley resulted
from the addition of 03 generated by local emissions to Og in the urban plume
crossing the valley. Evidence was also presented for the overnight survival
of Og aloft, with subsequent fumigation of the surface when the inversion
broke the next day. An 03-rich air mass which entered the region from the
southwest early in the afternoon of July 23 was observed the following morning
over northeast Massachusetts and late that same afternoon off the coast of Maine.
These observations suggest that transport of 03 within urban plumes over dis-
tances of nearly 400 miles may be possible. It was not possible to define
the role played by the overlapping of urban plumes in such long distance
transport.
43
-------
.0
a
250
200
150
100
50
0
300
250
200
150
100
50
0
200
150
IOO
50
0
150
100
50
0
8 150
100
50
0
150
IOO
50
0
IOO
50
0
150
IOO
50
0
IOO
50
0
/\
Bridgeport
Middletown
Eastford
Enfield
Worcester
Cambridge
6 8 10 12 14 16 18 20 22 24 2 46 8 10 12 14 16 18 20 22 24
7-23-75
Time of Day
7-24-75
Figure 13. Selected New England ground level ozone profiles-July 23-24, 1975,
44
-------
AUGUST 10, 1975
August 10, 1975, was a hot hazy day in the Connecticut River Valley. A
high pressure system centered near North Carolina and a low northeast of the
region influenced conditions on this day. A surface trough extended along
much of the east coast of the United States. Figures A-6 and A-7 in Appendix
A show the calculated trajectories for Simsbury and Groton. Flows were
generally from the west to northwest. The speed and direction of the winds
at 1000 feet for several representative locations within the region are given
in Table 6. These data indicate northwesterly winds in the morning shifting
to westerly sometime before noon.
TABLE 6. WINDS AT 1000 FEET ON AUGUST 10, 1975
Time
0700
0755
0807
0815
0830
1110
1408
1425
1900
Location
Chatham, Mass.
Boston, Mass.
Springfield, Mass.
Avery Point, Conn.
Putnam, Conn.
Avery Point, Conn.
Springfield, Mass.
Avery Point, Conn.
Chatham, Mass.
Speed ,
mph
22
17
10
23
13
12
16
23
29
Direction,
degrees
295
329
290
292
307
267
•285
261
260
The ground level ozone distributions are presented in Figure 14 a-c. The
concentrations at 0900 EST throughout the region were similar and moderately
high for this time of day, suggesting the possibility that the high pressure
system over the region may be having a region-wide impact on ozone concentra-
tions. This will be discussed shortly.
At noon high concentrations of 03 exist throughout western Connecticut
and Massachusetts; indeed the entire portion of the Connecticut River Valley
within Connecticut and Massachusetts is experiencing high levels of 03- By
1500 EST high 03 is observed all along the southern New England coast and as
far north as Hartford. Ozone levels throughout the rest of the region are
generally in excess of the federal standard. Throughout the remainder of
the day the area of high 03 remained along the southern coast.
With westerly and northwesterly winds it seems likely that the high
03 along the coast (and over Long Island and the Sound, as we will show shortly)
45
-------
fiL'GUST 10, 1975
9 C C E S T
0 ZD NE ( r P :. >
\ Density
0-50ppb
50-IOOppb
100-150 ppb
I50-200ppb
20O-25O ppb
250-30Oppb
•">300ppb
Figure 14a. Ozone distribution in southern New England on August 10, 1975.
46
-------
UGUST 10,
500 E ST
0 ZONE < PFB )
i ec& E ST
C ZO NE < PP6 )
Density
0-50 ppb
H 50-100ppb
IOO-l50ppb
1 !50-200ppb
2OO-250 ppb
2 50-300 ppb
> 300ppb
Figure 1%. Ozone distribution in southern New England on August 10, 1975.
47
-------
A UGUST 10, I <»7!
2100 EST
0 ZONiE ( PPB )
j Density
Jj )Q 0-50 ppb
^ 50-100 ppb
13 I00-I50ppb
P 150-ZOOppb
53 200-250 ppb
250-3OOppb
>300ppb
Figure 14c. Ozone distribution in southern New England on August 10, 1975.
48
-------
is part of the metropolitan New York urban plume. As on previous days, there
is no evidence for a unique source of 03 in the Connecticut River Valley.
The 03 concentrations in Simsbury and Groton were quite different on
August 10. Simsbury experienced only moderate 03> consistent with the tra-
jectories which showed a lack of major upwind sources. Groton experienced
rather high 03 concentrations on this day. Early in the day Groton lay
directly downwind of Hartford, while later it should be on the fringe of the
New Jersey-New York-southern Connecticut urban plume. The fluorocarbon data
in Figure 15 also suggest this possibility. Early in the day, the level of
F-ll at Groton was more than twice that at Simsbury, and later in the after-
noon and evening F-ll at Groton reached very high values consistent with a
major urban plume.
Further exploration of the 03 distribution is possible through the air-
craft data shown in Figure 16. This plot represents the results of a cross-
sectional pattern flown during the early afternoon of August 10, 1975. Four
vertical profiles and horizontal data at 1000 feet AGL were obtained along
the approximately north-south line from the south shore of Long Island to
the Massachusetts-Connecticut border. With westerly winds the southern portion
of this flight should overlap the metropolitan New York urban plume. The
vertical and horizontal data along this flight path have been used to construct
the computer-derived distribution map of 03 shown in Figure 17. Horizontal
and vertical distances are shown on the abscissa and ordinate, respectively.
The ozone concentration is represented by the shading density. The effect
of the urban plume is shown quite clearly in the lower right-hand corner of
the plot. This part of the study area is directly downwind of metropolitan
New York. A similar pattern flown almost simultaneously approximately 70 miles
further east (downwind)also shows the highest levels of 03 just off the
southern shore of Long Island"5'. The maximum concentration of 140 ppb was
observed at the surface. It is interesting that the concentration of 03 within
the urban plume on these two flights was quite similar (140-170 ppb) even
though the flight patterns were separated by about 70 miles. It may be that
additional generation of 03 between the two flight paths nearly balanced the
dilution and scavenging processes.
Another interesting feature shown in Figure 17 is the region of high
ozone concentration aloft at about 3000 feet. This 03 aloft was also observed
during the flight 70 miles to the east and was reported at several locations
49
-------
5OO
40O
3OO
2OO
IOO
o.
Q.
g
2 500
o
40O
3OO
200
IOO
I
10
I
12
I
20
Simsbury
Groton
14 16 18
Time of Day (EST)
Figure 15. Fluorocarbon-11 profiles for August 10, 1975.
50
22
24
-------
OIRfCTIOH OF FlICHTj
1
-v-
Figure 16. Ozone (in ppb) and other pollutant results for afternoon flight
conducted on August 10, 1975.
51
-------
6000
r- 5000
_ 4000
—3000
C
Q-
re
CD
ro
_ 2000
1000
Figure 17. Ozone concentration for cross-section from the Massachusetts-
Connecticut border to south shore of Long Island - approximately
73° 10' longitude.
52
-------
in New Jerseyv '. It is possible that this layer of 03-rich air is associated
with the high pressure cell influencing the entire northeastern United States.
Ozone throughout the area was high on August 10, with values in New Jersey,
Delaware, eastern Pennsylvania, and southern New England exceeding 100 ppb.
Such widespread regional 0^ has been related in the past to long distance
IA H\
transport within high pressure systems^'5'. The trajectories discussed earlier
show that the air influencing the northeast on August 10 had passed over the
industrialized midwest in the previous 2-3 days. The high precursor concen-
trations picked up at that time may be responsible for the regional 03 observed
on August 10. Obviously any 03 formed in urban plumes is superimposed on the
regional 03, leading to the very high concentrations observed downwind of urban
centers.
AUGUST 13, 1975
Another day which exhibited high levels of 03 in the Connecticut River
Valley was August 13, 1975. Sunny and hazy conditions prevailed on this day, as
a weak north-south trough was centered over southern New England. Wind direction
varied considerably over the region, but the general flow was from the western
quadrants. Wind data from some representative locations in southern New England
are listed in Table 7. These data show that a fairly stable southwesterly flow
was established by afternoon.
TABLE 7. WINDS AT 1000 FEET ON AUGUST 13, 1975
Time
0700
0800
0820
1300
1405
1414
1900
Location
Chatham, Mass.
Springfield, Mass.
Avery Point, Conn.
Chatham, Mass.
Avery Point, Conn.
Springfield, Mass.
Chatham, Mass.
Speed ,
mph
2
5
7
5
12
10
16
Direction,
degrees
045
024
064
180
231
216
220
The trajectories for Simsbury and Groton are shown in Figures A-8 and A-9
in Appendix A. Until just after noon the air arriving in Simsbury and Groton
was from the northwest and had passed diagonally across New York State. It
should be noted that this trajectory places Groton directly downwind of Hartford.
53
-------
By early afternoon the winds had shifted and Groton began receiving air from
the metropolitan New York area. At this time Simsbury's air was from north-
western New Jersey and southern New York State.
Maps showing the patterns of ozone distribution in southern New England
on August 13 are included in Figure 18 a-c. Two regions of elevated 0, existed
at noontime; one from southwest to central Connecticut and one in southeast
Massachusetts. By 1500 EST, after the wind shifted to the southwest, a mass
of O-j-enriched air is observed entering Connecticut from the southwest. High
ozone is found both within and to the west and south of the Connecticut River
Valley. Another area of high Og exists in southeastern Massachusetts. Much
of the data shown there are from aircraft flights and therefore represent
the concentrations at 1000 feet. The morning wind patterns suggest that this
high 0.3 may be due to an intermingling of the Boston and Providence plumes,
although firm evidence for this is not available.
By 1800 the area of high 03 in Connecticut has moved further eastward
and now encompasses Groton. Further eastward and northeastward motion is
discernable in the map for 2100 hours, as eastern Connecticut and Rhode Island
stations all show increases in the 03 concentration between 1800-2100 EST.
Since these increases occur after dark it is clear that transport rather than
photochemistry has caused the increase. The concentrations of 03 at the
Rhode Island and eastern Massachusetts stations increase even further between
2100 hours and midnight, providing further evidence of transport.*
The fluorocarbon-11 results for August 13 are shown in Figure 19. Until
2200 EST values of F-ll were close to background at Simsbury. At about 2200
the F-ll concentration increased markedly and the maximum 03 for the day was
observed simultaneously. From the ozone distribution maps (Figure 18) it
appears that a kind of "backwash" of urban air moved into Simsbury at this
time. The concentration of F-ll peaked at 1100 EST in Groton, possibly due
to emissions from the Hartford area, and increased again later in the evening
*
While not pertinent to our discussion of the Connecticut River Valley situation,
it is nevertheless noteworthy that a layer of very high 03 concentration existed
aloft during the afternoon over a large portion of southern New England and even
into New Hampshire. The altitude of this layer varied between 1000-3000 feet
depending on location. It is also interesting that several of the vertical
profiles taken during the late morning show ozone concentration increasing
with altitude between 6000-11,000 feet. The complex behavior of ozone on this
day should be investigated in future studies.
54
-------
A UGUST 13,
900 E ST
0 ZONE ( PPB )
a J G u S T ! 3 , I 5 7 5
1203 E S T
C ZONE ( PPB )
I Density
0-5Oppb
>3 50-IOOppb
IOO-l5Oppb
H I50-200ppb
200-250 ppb
250-300ppb
>300ppb
Figure 18a. Ozone distribution in southern New England on August 13, 1975.
55
-------
ft U3UST 13, 1975
15CO EST
0 ZO NE ( PFB)
ft L 5 I! S 7 13, 1 *
! e 'j 0 E S '
0 21 '.-. ( -
Density
D 0-50ppb
H 50-100 ppb
I00-I50ppb
I50-200ppb
200-250 ppb
250-300 ppb
>300ppb
Figure 18b. Ozone distribution in southern New England on August 13, 1975.
56
-------
AUGUST L 3 ,
2100 E ST
0 20 N£ ( F-PB )
c UGUST 13, 1 IT.
iiCO E ST
0 ZC NE '. PP9 1
Density
0-50 ppb
50-100 ppb
I00-I50ppb
I50-200ppb
20O-250ppb
250-3OOppb
>300ppb
Figure 18c. Ozone distribution in southern New England on August 13, 1975.
57
-------
5OO
400 -
300 -
o.
o.
I
200 -
100 -
500
40C
Groton
30O
200
100
I
I
I
I
10
12
20
14 16 18
Time of Day (EST)
Figure 19. Fluorocarbon-11 profiles for August 13, 1975.
58
22
24
-------
at about the time when the 03 distribution maps showed 03 moving through the
Groton area. The F-ll concentration is still high at midnight, indicating
that urban air is still moving into Groton. The 03 level has decreased by
this time, no doubt due to the lack of sunlight to initiate reactions in the
air mass leaving the urban source area after sundown.
As has been found on other case study days, the origin of high 03 in the
Connecticut River Valley is in the emissions of upwind urban areas, with New
Jersey, New York, and southwestern Connecticut all contributing to the ultimate
03/precursor burden of the air crossing the valley under southwesterly winds.
AUGUST 21, 1975
August 21, 1975, was cool, clear and sunny in the Connecticut River Valley,
as the entire region was dominated by a strong high pressure system. The wind
data in Table 8 indicate moderate morning winds from the northwest with stronger
afternoon and evening winds from the southwest. The trajectories in Figures
A-10 and A-11 in the Appendix show air arriving at both Simsbury and Groton
from the northwest until early afternoon, at which point the southwesterly
flow sets in. The air arriving in Simsbury from the northwest originated in
Canada north of Lake Ontario and crossed some rather remote stretches of New
York State; it should be rather clean. Morning air in Groton may have crossed
Hartford and may not be so clean. The fluorocarbon-11 profiles in Figure 20
also suggest this possibility.
TABLE 8. WINDS AT 1000 FEET ON AUGUST 21, 1975
Time
0700
0800
0800
1300
1400
1400
1900
1900
Location
Chatham, Mass.
Springfield, Mass.
Putnam, Conn.
Chatham, Mass.
Springfield, Mass.
Putnam, Mass.
Chatham, Mass.
Albany, N.Y.
Speed,
mph
10
13
10
12
16
20
30
20
Direction,
degrees
295
320
314
230
219
254
235
195
The ozone concentrations in southern New England throughout the day are
shown in the distribution maps in Figure 21 a-c. The concentrations were low
at 0900 and only moderate at most locations at noon. The one exception is
59
-------
500
4OO
300
Simsbury
2OO
(OO
CL
CL
O
500
4OO
30O
200
IOO
Groton
I
I
I
I
I
10
Figure 20.
12
20
14 16 18
Time of Day (EST)
Fluorocarbon-11 profiles for August 21, 1975.
60
22
24
-------
A IIGUST 21, 1975
900 EST
0 ZO ME ( PPB )
',\
I
*!; i Density
0-50 ppb
£3 50-100ppb
100-150 ppb
!5O-200ppb
200-250 ppb
250-300 ppb
>300ppb
Figure 21a. Ozone distribution in southern New England on August 21, 1975.
61
-------
ft UGUST 21, 19
1500 EST
0 ZO ME ( PPB )
Density
Q 0-50ppb
£3 50-100 ppb
03 I00-I50ppb
(H I50-200ppb
£3 200-250 ppb
g 250-300ppb
• >300ppb
Figure 21b. Ozone distribution in southern New England on August 21, 1975.
62
-------
P. US'JST 21, 197!
2 1 CO E ST
0 ZO ME ( P PB )
it
I ! Density
0-50 ppb
j[3 50-100ppb
IOO-l5Oppb
!5O-2OOppb
200-250ppb
250-300 ppb
>300ppb
Figure 21c. Ozone distribution in southern New England on August 21, 1975.
63
-------
Groton, which reached 115 ppb at noon. Both the fluorocarbon data and the
trajectories imply that morning emissions in Hartford may be responsible.
Later in the afternoon a band of Og-rich air was evident along the
southern half of the region, with maximum values occurring in southern
Connecticut and over Long Island Sound at 1800 EST. The dividing line between
low 03 in the northern portion of the region and moderate to high 03 to the
south is quite distinct. The trajectories shown earlier reveal the reason for
the sharp separation; during the afternoon the air arriving at Groton and
southern Connecticut has passed over or near several major urban centers, while
Simsbury and more northerly locations experienced air which has passed through
predominantly rural areas. Indeed the F-ll concentration in Simsbury was very
close to the tropospheric background level during much of the day.
Since both the air mass trajectories and the Simsbury F-ll results demon-
strate that the northerly portions of the region experienced clean rural air
during much of the day, we should be able to derive some information on 03
formation in rural air using the data from these northerly locations. It is
clear that the photochemical conditions for generating 03 exist on this day,
as witnessed by the high levels observed in the southern portion of the region.
The ozone concentrations found in the rural air of northern Connecticut and
western Massachusetts should be the sum of tropospheric background 03 and
any 03 synthesized from local or upwind rural emissions. The ozone distri-
bution maps through 1800 EST demonstrate that these combined sources of 03
yield concentrations on the order of 25-40 ppb. After 1800 EST the F-ll con-
centration in Simsbury increased, indicating an influx of urban emissions.
During this same period the 03 maps show that the band of 03-enriched air
to the south moved northward, so that by 2100 it encompassed Simsbury. The
simultaneous increases in 03 and F-ll at night are strongly indicative of
transport of urban air to this rural location.
The 25-40 ppb background rural ozone concentration derived above is a
useful value to compare with 03 from other sources, such as urban plumes and
regional blankets of 03 associated with high pressure systems. In the
Connecticut River Valley, where ozone concentrations reached 200-300 ppb
on some of our case study days, it is clear that this background 03 makes
a relatively small contribution to the maximum 03 in the valley. As stated
before, urban emissions transported into the valley from the southwest, super-
imposed on regional 03 concentrations, together create the high levels of 03
observed in the valley.
64
-------
The high concentration of ozone observed over Long Island Sound and the
lower half of southern New England on August 21 is clearly the result of
transport of 03 and precursors into the area on southwest winds, judging from
both the trajectories and the 03 distribution maps. The possibility of an
overlap of urban plumes on this day is also suggested by the trajectories.
This possibility has been discussed in a report by Wolff, et alJ20\ which
should be consulted for further details.
THE RELATIONSHIP BETWEEN OZONE AND
FLUOROCARBON-11 IN SOUTHERN NEW ENGLAND
Measurements of F-ll were made during the Northeast Oxidant Study to pro-
vide a means of distinguishing between urban and rural air. The sources of
F-ll emissions are population oriented, so that air passing over urban areas
has higher concentrations of F-ll than rural air. Since F-ll is essentially
inert in the lower atmosphere, it can be used to trace the movement of urban
air. In the case studies just discussed, we frequently referred to the F-ll
data for this purpose.
Another possible use of the data, and the subject of this brief discussion,
involves the derivation of information on background 03 using extrapolations
of the F-ll results. This possibility was explored through the use of linear
regression techniques on the F-ll and ozone data from Simsbury and Groton.
Regression equations were derived for three categories of data:
(1) Hourly Averages
(2) Daily Averages
(3) Daily Maxima.
The results of these regressions are listed in Table 9.
The tabulated data show that the hourly averages are not well correlated,
which is not surprising considering that both daytime and nighttime data were
included. During the night 03 formation ceases and the concentration at the
surface decays, while high levels of F-ll can persist. This alone would
greatly reduce the correlation coefficient.
The daily average results are much more enlightening. The equations for
Simsbury and Groton are
65
-------
TABLE 9. LINEAR REGRESSON OF FLUOROCARBON-11 (PPT) ON
OZONE (PPB) AT SIMSBURY AND GROTON
([Ozone] = m[Fluorocarbon-ll] + b)
Simsbury -
Groton -
Hourly Averages
Daily Averages
JDaily Maxima
Hourly Averages
Daily Averages
_Daily Maxima
Slope
0.02
0.21
0.19
0.10
0.27
0.27
Intercept
29
0.75
22
27
-7.7
5.8
Correlation
Coefficient
0.03
0.68
0.48
0.21
0.65
0.70
Simsbury [03]Avg - 0.21 [F-H]Avg + 0.75
Groton [03]Avg = 0.27 [F-ll]Avg - 7.7.
A plot of the daily average Simsbury data is shown in Figure 22. If we make
an assumption about the average F-ll concentration in clean tropospheric air,
we can calculate an average 03 concentration for clean tropospheric air based
on the above equations. Clean air concentrations of F-ll in 1975 in the
(26)
northern hemisphere were reported* ; to vary between about 80-100 ppt, so
that an assumed average of 90 ppt seems reasonable (the lowest daily average
concentration in Simsbury was 92 ppt). Using this assumption, the average
background tropsopheric 03 concentration is calculated to be 20 ppb from the
Simsbury data and 17 ppb from the Groton results.
We can compare these daily average background 03 values with the average
ozone at Simsbury on August 21. This day was described earlier in the case
studies as showing minimal urban emissions up to 1800 EST, based on the F-ll
data. The average 03 concentration up to 1800 was 18 ppb, in very good agree-
ment with the background value derived from the F-ll extrapolations.
The regressions based on the maximum daily value for 03 and F-ll will be
useful in predicting maximum 03 and will be discussed further in the next
section on predictive modelling.
66
-------
en
100
90
60
70
O' 60
CL
rO
8 so
CD
0>
2 40
0)
30
20
10
50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250
Average Daily Fluorocarbon- II, ppt
Figure 22. Average daily 03 vs average daily F-ll at Simsbury.
-------
SECTION 3
STATISTICAL ANALYSIS OF OZONE IN SOUTHERN NEW ENGLAND
The objective of this section of the report is to develop a method for
predicting the ozone levels that can be expected in Southern New England under
various meteorological conditions. Data collected during the Northeast Oxidant
Study at Simsbury and Groton, Connecticut, have been employed for these analyses,
since these two sites provide a long and continuous record of diurnal pollutant
and meteorological conditions.
PRELIMINARY ANALYSIS
In order to predict ozone levels likely, to occur in southern New England
under ya.rious meteorological conditions,, a method of statistical analysis
known, as regression integral analysis was first attempted. This method can
be used to predict a single value of a dependent variable from a sequence of
values of an independent variable (it can also be used with multiple sequences
of independent variable values), and has proven to be an appropriate tool for
predicting an ozone level (such as the afternoon ozone maximum) from the
distribution of precursor and meteorological variables during the morning
hours^4'. The method proved fruitful only in a limited sense in this study,
however. While developing regression integrals for various measures of
afternoon ozone levels, it was noted that the only statistically significant
components of fit were the zero order components of the polynomials describing
the morning distributions of the independent variables. Since these zero
order components are proportional to the means of the independent variables
over the distribution period, these regression integrals proved to be equivalent
to regressions with ozone level as dependent variable and morning averages of
precursors and meteorological variables as independent variables.
In retrospect, one can see that the changes in average values of predictor
variables dominate the relationship with ozone level. In order to fine tune
this model and see what effect differences in distributions of predictor
68
-------
variables have, one needs data with roughly the same average level of predictor
variables but differing distributions over the time period under consideration.
Under these conditions, one should be able to detect differences caused by
distributional changes.
However, this discovery about regression integrals also has a positive
aspect; namely, some rather good statistical relationships were discovered
between ozone levels and some predictors even though significance was found
only in the zero order relationship. Thus we decided to abandon the regression
integral approach and to concentrate on simple regressions and multiple
regressions between ozone levels and averages of precursor and meteorological
variables taken over a suitable time period.
REGRESSION ANALYSIS
Having decided to concentrate on regression analysis, consideration was
now given to what measures of ozone level should be used, what should be the
time period over which the precursor and meteorological variables would be
averaged, and how wind speed and wind direction should be incorporated into
(A)
this analysis. Based on some previous researchv ' and on the preliminary
regression integral analysis, we selected the 0600 through 1300 EST averages
of the following variables (where available): NO, N02, CO, nonmethane hydro-
carbon, fluorocarbon-11, solar radiation, temperature, relative humidity, dew
point, and ethylene/acetylene ratio. To incorporate wind speed and wind
direction into the analysis, vector averages of the 0600 through 1300 EST
wind speed—wind direction vectors were obtained. These average vectors
were then classified by the following scheme into one of five categories:
(Calm) - wind speed below 1 m/s (2.237 mi/h)
(NE) - wind speed above 1 m/s and wind direction
in the northeast quadrant
(SE) - wind speed above 1 m/s and wind direction
in the southeast quadrant
(SW) - wind speed above 1 m/s and wind direction
in the southwest quadrant
(NW) - wind speed above 1 m/s and wind direction
in the northwest quadrant.
69
-------
By using dummy variables for categories NE, SE, SW, and NW, regressions
made including these variables as predictors will be composed of five parallel
hyperplanes, one hyperplane corresponding to each of the five categories.
Formulated in this way, if the regression coefficient for any of the four
directions is statistically significant, that indicates that the ozone level
for this category is significantly different from the ozone level in the Calm
category.
Finally, we decided to try three measures of ozone level. These were,
respectively, average ozone (1300-1800 EST), mid-max ozone (maximum ozone in
the time period 1300-1800 EST) and maximum ozone (1300-2400 EST). These
measures were selected for the following reasons: the mid-max ozone was
chosen as a measure of the maximum ozone photochemically generated from the
morning precursors, the average ozone was chosen as a more stable measure of
the ozone level generated by morning precursors, and the maximum was chosen
in order to have a measure of late arriving transported ozone.
Two types of regression analyses were performed using these data and
constructed variables. The first type consists of simple regressions of each
of the measures of.-ozone level versus individual predictors, but incorporating
the wind speed-wind direction categories. The second type is multiple regres-
sion of the measures of ozone level versus all combinations of predictors,
including the wind speed-wind direction categories. These analyses were
performed on the hourly ground station data collected at Groton from July 15,
1975, through August 22, 1975, and at Simsbury from July 15, 1975, through
August 21, 1975.
RESULTS OF REGRESSION ANALYSIS
Table 10 contains descriptive statistics of the variables included in
the regression analysis. These are instructive as they show quite a difference
in environmental conditions between Simsbury and Groton, the latter site
having markedly higher levels of pollutants and a high proportion of days in
the SW category of wind speed-wind direction.
Table 11 contains the regression analysis of the three ozone measures
versus the wind speed-wind direction categories. Note that the Simsbury
data show no days in the NE category and the Groton data contain no days
in the SE category. There is only one observation in the Simsbury SE category
and even that one observation is misleading since it occurred on July 18 when
70
-------
TABLE 10. DESCRIPTIVE STATISTICS
Variable
Name
Avg. 03, ppb
Mid-Max. 03, ppb
Max. 03, ppb
NO, ppb
N02, ppb
CO, ppm
NMHC, ppmC
F-ll, ppt
NSJDL R*
Temp, °C
Rel. Hum., %
Dew Pt., °C
NE
SE
9LJ
On
NW
Calm
Eth/Acy
Mean
54.9
64.6
67.6
3.0
3.5
.19
.40
138.5
2.98
23.8
68.0
—
—
—
—
—
—
.63
Sims bury
Std. Dev.
22.3
27.0
28.4
1.8
2.4
.11
.24
38.6
1.16
3.4
13.7
—
—
—
— •-
—
—
.99
No. of
Cases
38
38
38
38
38
32
33
37
38
38
38
—
0
1
11
11
15
14
Mean
76.1
88.7
94.9
10.0
17.5
.69
.23
218.7
.66
23.6
_
18.1
__
—
—
—
—
^—
Groton
Std. Dev.
36.6
42.2
45.1
3.4
8.7
.38
.31
64.3
.23
2.3
__
3.3
_^
—
—
_ .
_
—
No. of
Cases
38
38
38
35
36
35
35
38
38
38
-ma_
38
5
0
27
1
5
—
*Different instruments were used to monitor solar radiation at the two sites.
71
-------
TABLE 11. REGRESSION ANALYSES-MEASURES OF OZONE VERSUS WIND DIRECTION
PO
Regression Coefficients
Site NE
Simsbury -
Groton -6.5
Simsbury -
Groton -5.8
Simsbury —
Groton -8.4
SE SW
Dependent
22.9 16.2
33.6
Dependent
25.7 23.6
42.1
Dependent
68.7 23.3
42.4
NW
Variable
-13.4
39.9
Variable
-14.4
58.0
Variable
-16.1
51.4
Const.
= Average
53.4
52.1
= Maximum
61.3
58.0
= Maximum
63.7
64.6
R
Ozone
.54
.46
•Ozone
.5.7
.49
Ozone
.66
.47
y Residual
R^ Std. Dev.
(1300-1800
.29
.21
(1300-1800
.32
.24
(1300-2400
EST)
19.63
34.02
EST)
-23.18
38.48
EST)
.43 22.26
.22 41.58
Significance
Level
.008
.046
.004
.025
.000
.036
-------
the air flow throughout the region was from the southwest. The SE category
at Simsbury will therefore be ignored in the remaining discussion. There is
also only one observation in the Groton NW category. Hence in these and
ensuing regressions we will be primarily concerned with the Simsbury NW and
SW categories as opposed to the Calm category, and the Groton NE and SW as
opposed to Calm. In this regression table and in following ones it is gen-
erally true that wind speed-wind direction is more of a factor at the Simsbury
site, and that the greatest impact of wind speed-wind direction as a predictor
is on the maximum ozone variable.
As an example of the use of the results in Table 11, one equation will
be analyzed further. We will choose the Simsbury site with maximum ozone as
dependent variable, since this is the strongest result in the table (43 percent
of the variance in maximum ozone is explained by wind speed-wind direction
alone as indicated by an R2 of 0.43). The significance level of this regression
is zero to three places. The constant of 63.7 indicates that the average
maximum ozone in the Calm category is 63.7 ppb. A value of -16.1 for NW
indicates that the average maximum ozone in the NW category is 47.6 ppb
(63.7 - 16.1). Similarly the average maximum ozone in the SW category is
87 ppb. These ozone levels are all different from the Calm level at the .1
level of significance. As stated earlier, the results in the Simsbury SE
category are misleading and should be ignored.
Table 12 through 20 show the results of regression analyses of ozone
levels against individual predictors, including wind speed-wind direction
categories. The results of these regressions can be interpreted similarly
to those in Table 11. For example, consider the regression of maximum ozone
vs fluorocarbon-11 at Simsbury (Table 16). Since the average value of
fluorocarbon-11 during the study was 138.5 ppt, the prediction of the maximum
ozone level for this level of fluorocarbon-11 if the wind is calm would be
(.48)038.5) + 4.0 = 70.5 ppb.
Similarly, if the wind is from the SW, this equation predicts 73.9 for the
maximum ozone level. Note that this simple regression explains 74 percent
of the variation in maximum ozone. Note also the consistent values of the
slopes of the predictor variable at the Simsbury and Groton sites.
One might get the impression from the above discussion that this regres-
sion equation predicts only about 3-4 ppb difference in 03 between calm winds
73
-------
TABLE 12. REGRESSION ANALYSES-MEASURES OF OZONE VERSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME.= NO
Site
Simsbury
Groton
Simsbury
Groton
Simsbury
Groton
Predictor
.87
.84
2.4
1.6
2.2
1.0
Regression
NE SE
Dependent
22.2
3.6
Dependent
21.2
6.9
Dependent
63.5
3.4
Coefficients
SW
Variable =
16.8
38.3
Variable =
25.3
47.2
Variable =
25.0
48.2
NN
Averaoe
-13.4
40.2
Maximum
-14.3
58.4
Maximum
. -16.0
51.7
Const.
Ozone
50
43
Ozone
53
42
Ozone
55
54
R R2
Residual Significance
Std. Dev. Level
(1300-1800 EST)
.7 .54 .30
.6 .50 .25
(1300-1800 EST)
.8 .59 .34
.5 .54 .30
(1300-2400 EST)
.9
.5
.67 .46
.52 .27
19.
31.
23.
35.
22.
38.
86
18
14
06
15
10
.018
.06
.006
.028
.000
.047
-------
TABLE 13. REGRESSION ANALYSES-MEASURES OF OZONE VERSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME = NO-
01
Regression Coefficients
Site
Simsbury
Grcton
Simsbury
Groton
Simsbury
Groton
Predictor NE SE SM
Dependent Variable
-.15 - 24.3 16.5
2.0 4.3 - 28.0
Dependent Variable
-1.1 - 29.2 25.9
2.2 6.5 - 37.5
Dependent Variable
-.13 - 68.7 23.5
2.0 2.3 - 38.1
NW Const
= Average Ozone (
-13.5 54.9
23.5 4.3
Residual Significance
R R2 Std. Dev. Le^el
1300-1 COO EST)
.54 .29 19.92
.69 .48 27.00
.020
.000
= Maximum Ozone (1300-1800 EST)
-15.1 64.7
39.4 25.9
. .57 .33 23.39
.71 .51 30.53
.009
.000
= Maximum Ozone (1300-2400 EST)
-16.2 64.1
35.1 36.5
.66 .43 22.59
.66 .43 34.85
.001
.001
-------
en
TABLE 14. REGRESSION ANALYSES-MEASURES OF OZONE VERSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME = CO
Regression Coefficients Residual !
Site Predictor NE SE SW NW Const. R R* Std. Dev.
Dependent Variable = Average Ozone (1300-1800 EST)
Simsbury 73.4 - - 18.2 -9.7 38.4 .63 .39 19.07
Groton 38.1 -6.9 - 27.7 30.5 31.0 '.59 .34 32.85
Dependent Variable = Maximum Ozone (1300-1800 EST)
Simsbury 70.7 - - 24.1 -9.6 45.6 .63 .39 21.21
Groton 40.1 -6.05 - 35.6 48.2 35.7 .59 .35 37.66
Dependent Variable = Maximum Ozone (1300-2400 EST)
Simsbury 84.6 - - 24.2 -11.0 45.6 .69 .48 19.40
Groton 39.4 -10.3 - 35.0 40.2 44.3 .56 .31 41.22
Significance
Level
.003
.011
.003
.011
.000
.021
-------
TABLE 15. REGRESSION ANALYSES-MEASURES OF OZONE VIRSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME = NMHC
Reqression Coefficients n Residual S
Site
Simsbury
Groton
Simsbury
Groton
Simsbury
Groton
Predictor
45.4
12.5
51.8
13.3
48.1
10.2
" NE SE SW
Dependent Variable =
13.0 10.4
-5.2 - 34.1
Dependent Variable =
13.3 14.1
-4.3 - 42.3
Dependent Variable =
56.7 13.4
-8.6 - 41.9
NW Const. R R£ Std. Dev.
Averaqe Ozone (1300-1800 EST)
-12.0 37.1 .70 .49 16.83
41.5 49.2 .47 .22 35.81
Maximum Ozone (1300-1800 EST)
-13.5 42.7 .71 .50 19.04
59.8 54.8 .49 .24 40.54
Maximum Ozone (1300-2400 EST)
-15.1 46.4 .78 .61 17.75
51.5 63.5 .47 .22 43.86
ignificance
Level
.001
.107
.000
.073
.000
.104
-------
00
TABLE 16. REGRESSION ANALYSES-MEASURES OF OZONE VERSUS INDIVIDUAL PREDICTORS AND WIND DIREClION
PREDICTOR VARIABLE NAME = F-ll
Site
Simsbury
Groton
Simsbury
Groton
Simsbury
Groton
Predictor
.39
.41
.46
.46
.48
.47
Regression
NE SE
Dependent
5.1
-3.5
Dependent
3.4
-2.4
Dependent
45.1
-4.9
Coefficients
SW
Variable =
-.04
13.7
Variable =
4.4
19.5
Variable =
3.4
19.1
NW
Averaqe
-12.8
14.4
Maximum
-13.5
29.1
Maximum
-15.3
21.6
Const.
R
R2
Residual Significance
Std. Dev. Level
Ozone (1300-1800 EST)
5.2
-22.2
.78
.80
61
64
15
23
.05
.15
.000
.000
Ozone (1300-1800 EST)
3.9
-26.2
.79
.81
63
66
17
26
.58
.12
.000
.000
Ozone (1300-2400 EST)
4.0
-22.3
.86
.78
74
61
15
29
.53
.72
.000
.000
-------
1O
TABLE 17 . REGRESSION ANALYSES-MEASURES OF OZONE VHRSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME = SOL R
Site
Simsbury
Groton
Simsbury
Groton
Simsbury
Groton
Predictor
6.4
71.9
6.8
82.9
5.9
92.1
Regression
ME SE
Dependent
22.5
24.7
Dependent
24.3
30.1
Dependent
67.0
31.5
Coefficients
SW
Variable =
16.2
39.0
Variable =
23.6
48.3
Variable =
23.3
49.3
NW
Averane
-13.3
51.7
Maximum
-14.2
71.6
Maximum
-15.9
66.5
Const.
Ozone
34.
-3.
Ozone
41.
-6.
Ozone
46.
-6.
R
(1300-1800
4
6
.63
:59
(1300-1800
1
2
.64
.61
(1300-2400
2
7
.70
.61
R2
EST)
.40
.34
EST)
.41
.37
EST)
.49
.37
Residual
Std. Dev.
18.
31.
22.
35.
21.
37.
32
40
00
37
41
99
Significance
Level
.002
.006
.001
.003
.000
.004
-------
00
o
TABLE 18. REGRESSION ANALYSES-MEASURES OF OZONE VERSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME = -T:EMP.
Reqression Coefficients _, Residual S
Site
Simsbury
Groton
Simsbury
Groton
Simsbury
Groton
Predictor
3.1
5.8
3.7
5.5
3.5
6.3
ME SE SW
Dependent Variable =
8.7 10.5
-6.2 - 20.5
Dependent Variable =
7.3 16.7
-5.5 - 29.8
Dependent Variable =
50.8 16.7
-8.1 - 28.1
NW Const. R R^ Std. Dev.
Averaqe Ozone (1300-1800 EST)
-11.8 -17.2 .69 .48 17.12
22.9 -76.2 .56 .32 32.07
Maximum Ozone (-1300-1800 EST)
-12.3 -24.3 -71 .51 20.02
42.0 -62.0 .56 .31 37.17
Maximum Ozone (1300-2400 EST)
-14.2 -17.4 .77 .59 19.32
32.9 -74.6 .55 .30 39.85
lignificance
Level
.000
.012
.000
.014
.000
.015
-------
00
TABLE 19 . REGRESSION ANALYSES-MEASURES OF OZONE VERSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME = REL. HUM. (SIMSBURY)
DEW PT. (GROTON)
Rearession Coefficients _, Residual S
Site
Simsbury
Groton
Simsbury
Groton
Simsbury
Groton
Predictor
-.60
.94
-.64
.65
-.57
-.13
• NE SE SW
Dependent Variable
27.2 21.3
-9.2 - 29.1
Dependent Variable
29.3 29.1
-9.6 - 39.0
Dependent Variable
71.4 28.1
-8.0 - 43.0
NW Const. R R^ Std. Dev.
= Averaqe Ozone (1300-1000 EST)
-13.2 92.4 .64 ,41 18.10
35.0 38.8 .46 ,21 34.42
= Maximum Ozone (1300-1800 EST)
-14.1 102.9 .65 .42 21.77
54.5 48.8 .49 .24 39.01
= Maximum Ozone (1300-2400 EST)
-15.9 101.10 .71 .50 21.14
52.1 66.4 .47 .22 42.20
ignificance
Level
.001
.088
.001
.055
.000
.078
-------
TABLE 20. REGRESSION ANALYSES-MEASURES OF OZONE VERSUS INDIVIDUAL PREDICTORS AND WIND DIRECTION
PREDICTOR VARIABLE NAME = ETH/ACY
Regression Coefficients
Site PredictorNESE SW NW Const. R
Residual
Std. Dev.
Significance
Level
Simsbury
Groton
-3.7
oo
ro
Simsbury
Groton
-5.7
Simsbury
Groton
-6.4
Dependent Variable = Average Ozone (1300-1800 EST)
22.5 -16.4 54.6 .62 .38 16.7
Dependent Variable = Maximum Ozone (1300-1800 EST)
29.3 -16.4 63.0 .66 .43 17.56
Dependent Variable = Maximum Ozone (1300-2400 EST)
27.3 -17.6 65.2 .68 .46 16.91
.173
.118
.094
-------
and winds-from the SW quadrant. This is true if the value of fluorocarbon-11
is the same for each of these categories (i.e., no or constant input of urban
emissions). However, it is likely that the average fluorocarbon-11 is higher
in the SW quadrant due to urban emissions, making a comparison at the same
fluorocarbon-11 concentration of less value. The use of actual morning average
F-ll values when making predictions should be much more informative.
These regressions provide a basis for predicting ozone levels based on
an average morning concentration of a given predictor variable, and the category
to which the morning wind speed-wind direction belongs. Some of these variables,
such as fluorocarbon-11, provide a rather good base for predicting ozone levels.
Tnese predictions can be strengthened somewhat by adding combinations of
variables in the regression analyses. However, since correlations among
predictors make the resulting regression coefficients more uncertain and
also mask any causal or correlative significance of these coefficients, the
effectiveness of these multiple regression equations as predictors of ozone
level may be limited. With these cautionary notes, Table 21 contains multiple
regression analyses of ozone levels versus all predictors. Since these equa-
tions are based on the numbers of observations in common to all variables, they
have fewer observations than many of the simple regressions, and hence their
effectiveness cannot be directly compared by, for example, considering the
2 2
values of R in individual regressions as compared to R in the multiple
regressions.
It should be noted that the use of the multiple regression results in
Table 21 for predicting ozone should only be attempted when values for all of
the predictor variables are available. If one or more predictor variables are
missing, the effectiveness of the multiple regression equation is greatly
reduced and it may even yield misleading results. Since data on all the
predictors used in Table 21 will rarely be available except during special
studies, the utility of the multiple regressions in Table 21 for routine
predictions is marginal. Indeed we cannot even cross-compare the multiple
regression models at the two sites since one of the variables (solar radiation)
is not directly comparable. For purposes of routine predictions then, the
individual regressions in Tables 11-20 are much more useful. Our studies
suggest that the regression of greatest utility for simple predictions based
only on £enerally available meteorological data is the temperature regression
shown in Table 18. We find very little gain in predictive capability when
83
-------
TABLE 21. MULTIPLE REGRESSION ANALYSES-MEASURES OF OZONE VERSUS COMBINATIONS OF PREDICTORS AND WIND DIRECTION
Regression Coefficients
Site
Simsbury
Groton
Simsbury
Groton
Simsbury
Groton
NO
-1.9
-2.5
-.93
-2.3
-1.0
-2.8
N02
-4.9
-
-5.0
-.57
-4.0
-1.0
CO
76.7
6.5
58.9
4.1
61.5
-10.2
NMHC
20.0
-21.4
20.7
-25.1
17.7
-15.9
F-ll Sol R
Dependent
.23 -1.5
.39 57.9
Dependent
.32 -2.0
.51 79.1
Dependent
.33 -2.4 -
.59 43.6
Temp
Variable
1.2 -
-3.0
Variable
.97 -
-6.0
Variable
.71 -
.56
RH DP NE SE SW NW
= Average Ozone (1300-1800 EST)
.36 -- - - 8.6 -11.5
2.9 19. 6; - 23.3 21.9
= Maximum Ozone (T300-1800 EST)
.42 - - 14.3 -12.8
5.0 27.5 - 31'. 5 43.0
= Maximum Ozone (1300-2400 EST)
.42 - - 12.1 -13. 7
-2.1 22.4 - 46.7 47.3
Const.
23.4
-21.9
31.5
14.1
37.9
-15.5
R
.87
.86
.86
.87
.88
.86
R2
.76
.74
.75
.76
.77
.73
Residual Significance
Std. Dev. Level
13.6
21.7 »
15.8
24.6
14.36
27.7
.000
.000
.001
.000
.000
.001
-------
other meteorological variables such as relative humidity and solar radiation
are included, since temperature is highly correlated with such variables.
Tnus the simplest and most convenient predictions can be made using morning
temperature and wind direction/speed vector data, along with the equations
from Table 18. Increased accuracy is obtained by using the fluorocarbon-11
regression or the multiple regressions, but the necessary data are not gen-
erally available.
A comparison of ozone predictions based on the temperature and wind
vector regression results (Table 18) for each site and across sites is
given in Table 22. Predictions for calm and southwest winds are included;
data are not sufficient in the other wind categories to permit cross-comparison.
The average temperature over the entire study at each site was used for these
predictions, rather than the average morning temperature. The table shows
that the two equations make similar predictions when the winds are calm,
although the Simsbury model predictions are slightly lower. Under southwest
wind conditions the Groton model predicts considerably higher 03- Predictions
for rural southern New England areas similar to Simsbury should probably be
made with the Simsbury model; predictions for less rural and/or coastal areas
should employ the model derived from the Groton data. Obviously it would be
useful to have many more data and data under a greater variety of conditions
(e.g., greater variation in wind direction) in order to derive more meaningful
predictive equations. The equations presented in this section can be used
with the appropriate cautions to predict ozone, but the user must bear in mind
the limited data base upon which the equations have been derived and the
statistical nature of the predictions. The equations should help to character-
ize the "average" behavior of ozone based on a limited set of predictor vari-
ables; extreme values caused by unusual combinations of conditions or conditions
which do not affect a predictor variable will not be accurately predicted.
85
-------
TABLE 22. CROSS-COMPARISON OF OZONE PREDICTIONS BASED
ON THE SIMSBURY AND 6ROTON EQUATIONS FROM
TABLE 18 (TEMPERATURE REGRESSION -OZONE IN PPB)
Simsbury 03 Predicted
by Simsbury Model
Avg. 03
Mid-Max. 03
Max. 03
Calm
57
64
66
SW
68
81
83
Groton 03 Predicted
by Groton Model
Avg. 03
Mid-Max. 03
Max. 00
Calm
61
68
74
SW
81
98
102
Simsbury 03 Predicted
by Groton Model
Calm
62
69
75
SW
82
99
103
Groton 03 Predicted
by Simsbury Model
Calm
56
63
65
SW
66
80
82
86
-------
SECTION 4
SUMMARY
The objective of this study has been to use data collected during the
1975 Northeast Oxidant Study to determine the cause of high ozone concentra-
tions in the Connecticut River Valley and to develop a method for predicting
ozone levels that can be expected in southern New England under various
meteorological conditions.
The Connecticut River Valley situation has been investigated by first
examining the meteorological data for special effects of the valley and then
examining in detail selected days during the Northeast Study. Meteorological
data, air mass trajectories, fluorocarbon-11 profiles and ozone distribution
maps were all used to elucidate the Connecticut River Valley situation. The
results of this investigation indicate that the cause of high ozone in the
Connecticut Valley is rooted in the location of the valley. During the
summer months, the prevailing southwesterly winds place the valley directly
downwind of the New York/New Jersey/southwestern Connecticut urban complex
(and on some days the Philadelphia and Washington/Baltimore areas). Typical
winds speeds are in the ideal range to transport the strong morning urban
emissions to the Connecticut Valley with a reaction time long enough to
photochemically generate high concentrations of ozone yet short enough to
avoid extensive dilution. The ozone formed from the urban emissions (i.e.,
the urban plume) was observed on many case study days to move into Connecticut
from the southwest in early afternoon, cross the Connecticut River Valley,
and continue into Massachusetts during the evening. In one case an Oj-rich
air mass was tracked as far north as the coast of Maine. The dimensions of
the urban plumes on several days were found to vary from 30-80 miles in
width and 100-175 miles in length, seemingly depending on wind speed.
Urban plumes were occasionally observed aloft, and firm evidence was
presented for the overnight survival of 03 aloft above the nocturnal inversion.
Vertical mixing the following morning was shown to result in high surface ozone
concentrations early in the morning before photochemical generation of Og
should be significant.
87 "^
-------
Overlapping of urban plumes appeared likely on some case study days and
was discussed briefly. Regional ozone of 50-100 ppb, probably associated
with high pressure systems, was observed on several days outside the urban
pU'me. Concentrations generated within the urban plume are superimposed on
regional and locally generated 03 and result in very high concentrations
(>200 ppb) in southern New England. Fluorocarbon-11 results and air mass
trajectories were used to define clean rural air conditions. Under these con-
ditions on a photochemically active day, average "background" levels of 03 of
about 20 ppb were observed. Daily maximum Og under these clean conditions
was 25-40 ppb. Extrapolation of F-ll vs 03 linear regressions at Simsbury
and Groton to background tropospheric F-ll concentrations (i.e., no urban
input) suggested an average tropospheric "background" ozone concentration
of about 20 ppb.
Several methods of predicting ozone in southern New England were inves-
tigated including regression integrals, simple regressions and multiple
regressions. Both simple and multiple regression equations were derived for
predicting ozone based on a variety of chemical and meteorological predictor
variables. Fluorocarbon-11, the tracer for urban air, has the greatest
predictive capability of the single variables but F-ll data are not routinely
available. The multiple regression equations should be the most comprehensive
in that they incorporate the greatest number of predictor variables, but again
the necessary input data are not generally available. The most generally
useful predictive equations are probably those based on widely available
meteorological data. We have suggested the use of the regression equations
involving temperature and the wind vector categories. Since temperature is
highly correlated with other meteorological variables such as relative humidity
and solar intensity, the regressions involving temperature and wind vector
categories should reflect the effects of these additional meteorological
variables. Thus, the use of the temperature/wind vector regression results
seems to provide the most widely useful means for simple predictions of
expected ranges of ozone concentrations.
-------
REFERENCES
1. EPA Report 450/3-74-034, Research Triangle Institute, May (1974).
2. EPA Report 450/3-75-036, Research Triangle Institute, March (1975).
3. Westberg, H., Allwine, K., and Elias, D. In: Proceedings - Ozone/
Oxidants Interactions with the Total Environment, APCA Publication,
Pittsburgh (1976).
4. Spicer, C.W., Gemma, J.L., Joseph, D.W. , and SticKsel, P.R., Battelle-
Columbus report: EPA-600/3-76-018, February (1976).
5. Spicer, C.W. In: Proceedings - Ozone/Oxidants Interactions with the
Total Environment, APCA Publication, Pittsburgh (1976).
6. Lovelace, D., Kapsalis, T., Bourke, R., and Cook, P., Indianapolis 1974
Summer Ozone Study, report from Indianapolis Center for Advanced Research,
Inc., Indianapolis (1975).
7. Westberg, H., Robinson, E., and Zimmerman, P., Paper 74-54 presented at
67th Annual Meeting Air Pollution Control Assoc., Denver (1974).
8. Blumenthal, D. and White, W., Paper 75-07.4 presented at 68th Annual
Meeting Air Pollution Control Assoc., Boston (1975).
9. Blumenthal, D., Smith, T., and White, W., preprint from "The Character-
istics and Origins of Smog Aerosol, Hidy, G. and Mueller, P., eds.,
MRI Report Pa-1370 (1975).
10. Cleveland, W.S., Kleiner, B., McRae, J.E., and Warner, J.L., Science,
191:179, 1976.
11. Rubino, R.A., Bruckman, L., and Magyar, J., J. Air Poll. Control Assoc.,
26(10):972 (1976).
12. Stasiuk, W.N. and Coffey, P.E., J. Air Poll. Control Assoc., 24(6):564, 1974,
13. Coffey, P.E. and Stasiuk, W.E., Environ. Sci. and Tech., 9(1):59, 1975.
14. Spicer, C.W., Joseph, D.W., and Ward, G., The Transport of Oxidant Beyond
Urban Areas - Compilation of Data, Battelle-Columbus draft report to EPA,
January (1976).
15. Measurement of Light Hydrocarbons and Studies of Oxidant Transport Beyond
Urban Areas - Final Data Report, Washington State University report to
EPA, February (1976).
89
-------
16. Siple, G.W., Zeller, K.F., and Zeller, T.M., Air Quality Data for the
Northeast Oxidant Transport Study, 1975, EPA draft report-EPA-L.V. (1976).
17. Meteorological Data for the Northeast Oxidant Transport Study, 1975,
EPA draft report-EPA-L.V. (1976).
18. Spittler, T.M., Ozone Field Audits for the 1975 Summer Ozone Study,
EPA draft report-EPA Region I, Boston (1976).
19. Wolff, G.T., Lioy, P.J., Wight, G.D., and Pasceri, R.E., An Aerial
Investigation of Photochemical Oxidants Over the Eastern Mid-Atlantic
States. In: Proceedings of EPA Symposium on Ozone Transport, Research
Triangle Park, N.C., 1976 (In press).
20. Wolff, G.T., Lioy, P.J., Meyers, R.E., Cederwall, R.T., Wight, G.D.,
Pasceri, R.E., and Tiaylor, R.S. Anatomy of Two Ozone Transport Episodes
in the Washington, D.C. to Boston, Mass. Corridor. Presented at the 10th
Annual Meeting of the Mid-Atlantic Amer. Chem. Soc., Philadelphia (1976).
21. Proceedings of EPA Symposium on the Northeast Oxidant Study-January 20-21,
1976, Research Triangle Park, N.C., 1976 (In press).
22. Morris, D., "Trends in Levels of Photochemical Oxidants in Southern New
England for 1975, in Ref. 21.
23. Climates of the States-Connecticut, U.S. -Department of Commerce, Weather
Bureau, Climatography of the United States, No. 60-6, Washington, D.C.,
1959, p 9.
24. Ostby, P.P., Atwater, M.A., and Perry, F., A Small-Scale Precipitation
Network Over Central Connecticut, Weatherwise, 22:63, 1969.
25. Ludwig, F.L. and Shelar, E. , Stanford Research Institute draft report
to EPA-EPA-901/9-76-005, October (1976).
26. Halocarbon Measurements in the Troposphere and Lower Stratosphere,
Washington State University final report to Manufacturing Chemists
Association, April (1976).
90
-------
APPENDIX A*
TRAJECTORIES OF AIR ARRIVING
AT GROTON AND SIMSBURY
This appendix shows the calculated trajectories of air arriving at 6-hour
intervals at Groton and Simsbury, Connecticut, during the period from July 16
to August 31, 1975. The methods and data used to calculate these trajectories
are described in the text of this report. Anyone who uses these trajectories
should consult the discussion in the report so that they would be aware of
the assumptions involved and of some of the shortcomings of the results.
Trajectories shown on the following pages were plotted on computer line
printer so that the north-south and east-west axes are of different scales.
This is the reason for the distortions in the maps. Trajectories arriving
at the end point at 0000 GMT (1900 EST of the preceding day) are marked by
the symbol "A". Those arriving at 0600 GMT (0100 EST) are marked by "B",
those arriving at 1200 GMT (0700 EST) by "C", and trajectories arriving at
1800 GMT (1300 EST) are shown by the symbol "D". When two or more trajectories
passed through the same point, an "X" is shown. The positions of the air at
6-hour intervals are marked by numerals. This position 6 hours prior to
arrival at the end point is shown by a "1", 12 hours prior to arrival is
indicated by a "2", and so forth.
Courtesy of Stanford Research Institute.
91
-------
18 JULY 1975
Figure A-l. Backward trajectories for Groton and Simsbury, Connecticut.
92
-------
19 JULY 1975
Figure A-2. Backward trajectories for Groton and Simsbury, Connecticut.
93
-------
20 JULY 1975
. _ ...
\ *.fc • ' I
1 ctdccc" i L
CCCCC ; IIP X
tccc it o
/•>) eccc 3Joo
/ <DD Ml
L J
Figure A-3. Backward trajectories for Groton and Sirasbury, Connecticut.
94
-------
23 JULY 19iS
. \ V
w I r.. -M::::.;...^
'-j J "C?%
Figure A-4. Backward trajectories for Groton and Simsbury, Connecticut.
95
-------
24 JULY 1975
< I , 1£ S^is.
Figure A-5. Backward trajectories for Groton and Simsbury, Connecticut,
96
-------
10 AUGUST 1975
V.. e
Figure A-6. Backward trajectories for Groton and Simsbury, Connecticut.
97
-------
11 AUGUST 1975
CCCCCCCC til
ccccjccccrecc t;>
ccc»<
tu
en
DOCO ODOUOC03)00 C»«
. OOPnl On.D 0000 XI
M
Figure A-7. Backward trajectories for Groton and Simsbury, Connecticut.
98
-------
13 AUGUST 1975
Figure A-8. Backward trajectories for Groton and Simsbury, Connecticut.
99
-------
14 AUGUST 1975
Don
00
ciwta
c cccooo
CC 00
cr. i cccijno
s< ccc i I ccero
CCttu
MM
ccccecccctccc cetec
ee ccccccdccccieccccccccc
•" X
A
Figure A-9. Backward trajectories for Groton and Simsbury, Connecticut.
100
-------
21 AUGUST 1975
Figure A-10. Backward trajectories for Groton and Simsbury, Connecticut.
101
-------
22 AUGUST 1975
Figure A-ll. Backward trajectories for Groton and Simsbury, Connecticut.
102
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
REPORT NO
EPA-6QO/5-77-Q41
3 RECIPIENT'S ACCESSION-NO.
4 TITLE AND SUBTITLE
THE TRANSPORT OF OXIDANT BEYOND URBAN AREAS
Data Analyses and Predictive Models for the Southern
1975
5 REPORT DATE
May 1977
6. PERFORMING ORGANIZATION CODE
Chester W. Spicer , James L. Gemma, and Philip R.
Sticksel
8. PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle - Columbus Laboratories
Columbus, Ohio 43201
10. PROGRAM ELEMENT NO.
1AA603 AJ-04 (FY-77)
11. CONTRACT/GRANT NO.
68-02-2241
12 SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory- RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
For additional information, see technical report EPA-600/3-76-1Q9
16. ABSTRACT ~~ ~~ —
The objective of this study has been to use data collected during the 1975
Northeast Oxidant Study to determine the cause of high ozone concentrations in the
Connecticut River Valley and to develop a method for predicting ozone levels that
can be expected in southern New England under various meteorological conditions.
During the summer months, the prevailing southwesterly winds place the
valley directly downwind of the New York/New Jersey/southwestern Connecticut
urban complex (and on some days the Philadelphia and Washington/Baltimore areas).
The ozone formed from the urban emissions (i.e., the urban plume) was observed
on many case study days to move into Connecticut from the southwest in early
afternoon, cross the Connecticut River Valley, and continue into Massachusetts
during the evening. In one case an 0 -rich air mass was tracked as far north
as the coast of Maine. The dimensions of the urban plumes on several days were
found to vary from 30-80 miles in width and 100-175 miles in length, seemingly
depending on wind speed.
Several methods of predicting ozone in southern New England were investigated
including regression integrals, simple regression and multiple regressions.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
* Air pollution
* Ozone
Meteorological data
* Transport properties
* Mathematical models
* Predictions
New England
13B
07B
04B
12A
8 DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
llNri.ASSTFTFn
21 NO. OF PAGES
113
20 SECURITY CLASS (Thispage)
22 PRICE
EPA Form 2220-1 (9-73)
103
------- |