SEPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 2771
EPA-450/4-79-035
September 1979
Air
Survey of the Role
of NOX in Nonurban
Ozone Formation
-------
EPA-450/4-79-035
Survey of the Role of NOX
in Nonurban Ozone Formation
by
J.R. Martinez and H.B. Singh
SRI International
333 Ravenswood Avenue
Menlo Park, California 94025
Contract No. 68-02-2835
EPA Project Officer: Harold G. Richter
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
September 1979
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from the
Library Services Office (MD-35) , U.S. Environmental Protection Agency.
Research Triangle Park, North Carolina 27711; or for a nominal fee,
from the National Technical Information Service, 5285 Port Royal Road,
Springfield,Virginia 22161.
This report was furnished to the Environmental Protection Agency by SRI
International, 333 Ravenswood Avenue, Menlo Park, California 94025,
in fulfillment of Contract No. 68-02-2835. The contents of this report are
reproduced herein as received from SRI International. The opinions,
findings, and conclusions expressed are those of the author and not necessarily
those of the Environmental Protection Agency. Mention of company or
product names is not to be considered as an endorsement by the Environmental
Protection Agency.
Publication No. EPA-450/4-79-035
11
-------
ABSTRACT
This study surveys and summarizes current knowledge about the role
of oxides of nitrogen (NO ) in the formation of ozone (Oo) in nonurban
X «J
areas. Project elements include a literature review, a survey of expert
opinion, and analyses of field data. The investigation was motivated by
the U.S. Environmental Protection Agency's concern about the hypothesis
that increased urban NO., emissions could lead to higher levels of nonur-
X
ban 0,.
The results of the study show that present knowledge about N0x/0
-------
CONTENTS
ABSTRACT
LIST OF ILLUSTRATIONS vll
LIST OF TABLES xi
ACKNOWLEDGMENTS xiii
I INTRODUCTION 1
A. Background 1
B. Objectives and Methodology 2
C. Report Organization 2
II LITERATURE REVIEW 3
A. Introduction 3
B. NO Measurement Methods 3
C. Levels of Nonurban NOX 6
D. Sinks of NOV 12
Jv
E. Chemistry of NOX 16
F. Sources of NO 20
X
1. Man-Made Sources 20
2. Natural Sources.. 23
G. Mathematical Models 24
1. Model Type and Intended Application 25
2. Model Verification 29
H. Relationship Between NOX and 03 30
III SUMMARY OF EXPERT OPINION 43
A. Introduction . 43
B. Summary of Results 43
1. Measurement .Methods 43
2. Sources and Sinks of NOX 45
3. Chemical Aspects 49
4. Control Strategies 52
5. Identification of Areas Requiring
Further Research 52
-------
IV ANALYSIS OF DATA FROM THE SULFATE
REGIONAL EXPERIMENT (SURE) 55
A. Introduction 55
B. Description of SURE Data 56
1. Geographical and Temporal Coverage 56
2. Instrumentation 59
C. Data Analysis 59
1. Case Study I: High-Ozone Events
at Montague, Massachusetts (Site 1) 78
2. Case Study II: High-Ozone Events
at Duncan Falls, Ohio (Site 4) .- 92
3. Case Study III: High-Ozone Events
at Giles County, Tennessee (Site 6) 95
4. Case Study IV: High-Ozone Events
at Research Triangle Park,
North Carolina (Site 8) 100
5. Case Study V: High-Ozone Events
at Lewisburg, West Virginia (Site 9) 109
D. Conclusions Ill
V ANALYSIS OF JETMORE, KANSAS, DATA 113
A. Introduction 113
B. Ground Station Data 113
1. Ozone-NQx Measurements 118
2. 03 and 'Be Relationship 124
3. Conclusions 125
VI CONCLUSIONS 127
VII RECOMMENDATIONS FOR FURTHER RESEARCH 131
A. Study I: Analysis of Data from Rural Locations 131
B. Study II: Analysis of Data from the Los Angeles Area. 132
C. Study III: Data Collection and Analysis Program...... ' 133
D. Additional Topics for Investigation 134
REFERENCES 137
vi
-------
ILLUSTRATIONS
1 Actual and Projected Trends in Emissions
of NOX in the United States by Source Categories 22
2 Ozone Concentration as a Function of NOX Level
with NonMethane Hydrocarbon as Parameter 31
3 Location of SURE Monitoring Stations 58
4 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 1 in Montague, Massachusetts,
During August-Dec ember 1977 61
5 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 2 in Scranton, Pennsylvania,
During August-December 1977 62
6 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 3 in Indian River, Delaware,
During August-Dec ember 1977 63
7 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 4 in Duncan Falls, Ohio,
During August-December 1977 64
8 Average Diurnal Variation pf Nitrogen Oxides
and Ozone at SURE Site 5 in Rockport, Indiana,
During August-December 1977 65
9 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 6 in Giles City, Tennessee,
During August-December 1977 66
10 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 7 in Fort Wayne, Indiana,
During August-Dec ember 1977 67
11 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 8 in Research Triangle Park,
North Carolina, During August-December 1977 68
12 Average Diurnal Variation of Nitrogen Oxides
and Ozone at SURE Site 9 in Lewisburg,
West Virginia, During August-December 1977 69
13 Hourly Concentration Variations of NO, N02, andOo
at Site 1 on Wednesday, 3 August 1977 79
14 Hourly Concentration Variations of NO, N02» and 0-j
at Site 1 on Thursday, 4 August 1977 79
15 Hourly Concentration Variations of NO, N02, and 03
at Site 1 on Friday, 5 August 1977 80
16 Hourly Concentration Variations of NO, N02, and 03
at Site 1 on Saturday, 27 August 1977 80
vii
-------
17 Hourly Concentration Variations of NO, N02» and 0-j
at Site 1 on Sunday, 28 August 1977 81
18 Hourly Concentration Variations of NO, N02» and 0-j
at Site 1 on Monday, 29 August 1977 81
19 Hourly Concentration Variations of NO, N02» and 0-j
at Site 1 on Thursday, 1 September 1977 82
20 Trajectories Arriving at Site 1, Montague,
Massachusetts, on 3 August 1977 84
21 Trajectories Arriving at Site 1, Montague,
Massachusetts, on 4 August 1977 84
22 Trajectories Arriving at Site 1, Montague,
Maine, on 5 August 1977 85
23 Trajectories Arriving at Site 1, Montague,
Massachusetts, on 27 August 1977 89
24 Trajectories Arriving at Site 1, Montague,
Massachusetts, on 28 August 1977 89
25 Trajectories Arriving at Site 1, Montague,
Massachusetts, on 29 August 1977 90
26 Trajectories Arriving at Site 1, Montague,
Massachusetts, on 1 September 1977 91
27 Hourly Concentration Variations of NO, N02» and 0-j
at Site 4 on Thursday. 1 September 1977 93
28 Hourly Concentration Variations of NO, N02» and 03
at Site 4 on Friday, 2 September 1977 93
29 Trajectories Arriving at Site 4, Duncan Falls, Ohio,
on 2 September 1977 94
30 Hourly Concentration Variations of 0^ and NOX
at Site 6 on Tuesday, 1 August 1977 96
31 Hourly Concentration Variations of 03 and NOX
at Site 6 on Monday, 22 August 1977 96
32 Hourly Concentration Variations of 0^ and NOX
at Site 6 on Friday, 23 September 1977 97
33 Trajectories Arriving at Site 6, Giles County,
Tennessee, on 2 August 1977 97
34 Trajectories Arriving at Site 6, Giles County,
Tennessee, on 22 August 1977 98
35 Trajectories Arriving at Site 6, Giles County,
Tennessee, on 23 September 1977 98
viii
-------
36 Hourly Concentration Variations of NO, NC^, and 0^
at Site 8 on Monday, 8 August 1977 ........................ 102
37 Hourly Concentration Variations of NO, N02* and 0-j
at Site 8 on Tuesday, 9 August 1977 ....................... 102
38 Trajectories Arriving at Site, Research Triangle Park,
North Carolina, on 9 August 1977 .......................... 103
39 Hourly Concentration Variations of NO, N02, and 03
at Site 8 on Friday, 26 August 1977 ....................... 103
40 Trajectories Arriving at Site 8, Research Triangle Park,
North Carolina, on 26 August 1977 ......................... 10A
41 Hourly Concentration Variations of NO, N02, and 0^
at Site 8 on Wednesday, 31 August 1977 .................... 104
42 Trajectories Arriving at Site 8, Research Triangle Park,
North Carolina, on 31 August 1977 ......................... 106
43 Hourly Concentration Variations of NO, N02» and 0^
at Site 8 on Saturday, 10 September 1977 .................. 106
44 Trajectories Arriving at Site 8, Research Triangle Park,
North Carolina, on 10 September 1977 ...................... 107
45 Hourly Concentration Variations of NO, No2, and 0-j
at Site 8 on Friday, 23 September 1977 .................... 107
46 Trajectories Arriving at Site 8, Research Triangle Park,
North Carolina, on 23 September 1977 ...................... 108
47 Hourly Concentration Variations of NO, NO 2, and Oo
at Site 9 on Saturday, 22 October 1977 .................... 109
48 Trajectories Arriving at Site 9, Lewisburg,
West Virginia, on 22 October 1977 ......................... 110
49 Mean Diurnal Variation in CH, at Jetmore, Kansas,
3 April-20 May 1978 ....................................... 116
50 Mean Diurnal Variation in CO at Jetmore, Kansas,
3 April-20 May 1978 ....................................... 117
51 Mean and Maximum 0-j at Jetmore, Kansas,
3 April-20 May 1978 ....................................... 118
52 Mean Diurnal Variation in Oo at Jetmore, Kansas,
3 April-20 May 1978 ....................................... 119
53 Mean Diurnal Variation in NO at Jetmore, Kansas,
3 April-20 May 1978 ....................................... 120
54 Mean Diurnal Variation in N02 at Jetmore, Kansas,
3 April-20 May 1978 ....................................... 120
55 Mean Diurnal Variation in NO at Jetmore, Kansas,
3 April-20 May 1978 ......... * ............................. 121
ix
-------
56 Mean Diurnal Variations in Wind Speed
and Selected Pollutants at Jetmore, Kansas,
on 27 April 1978 122
57 Diurnal Variations in Wind Speed
and Selected Pollutants at Jetmore, Kansas,
on 28 April 1978 123
58 Diurnal Variations In Wind Speed
and Selected Pollutants at Jetmore, Kansas,
on 15 May 1978- 123
59 Mean Daily 02 and Be at Jetmore, Kansas,
3 April-20 May 1978 125
60 Scatter Diagram for Mean Daily Oj versus Be
at Jetmore, Kansas, 3 April-20 May 1978 126
-------
TABLES
1 Summary of Measurement Methods for NOX
in Ambient Air
2 Concentration of Nitrogen Oxides Observed
in Clean Remote Areas •
3 Concentration of Nitrogen Oxides Observed
in Rural and Suburban Areas 8
4 Concentration of Nitrogen Oxides Observed
in Urban Areas 9
5 Nitric Acid and Nitrate Aerosol Formation Reactions 14
6 Natural Sources of NOX 23
7 Photochemical Models for Nonurban Areas ?6
8 Range of Hydrocarbon Concentrations Measured
Aboard Da Vinci II Balloon 37
9 Average Hydrocarbon Concentrations
in Elkton, Missouri 38
10 Atmospheric Half-Life of PAN 47
11 Location of Monitoring Stations in the SURE 57
12 Number of Available Observations for SURE Stations,
August-December 1977 59
13 Instrumentation Used at the SURE Sites 60
14 Summary of Hourly Concentrations of Nitrogen Oxides
and Ozone Measured at the SURE Sites
During August-December 1977 71
15 NO/NOX Ratios for the SURE Sites,
August-December 1977 73
16 Number of Hours with Ozone Concentrations
above 80, 100, and 120 at SURE Sites,
August-December 1977 75
17 Occurrence of Maximum Daily Ozone Exceeding 100 ppb
at the SURE Sites During August-December 1977 77
18 Data Summary for High-Ozone Days at SURE Site 1
in 1977 87
19 Chemical and Meteorological Parameters Measured
at Jetmore, Kansas 114
20 SRI Mobile Research Laboratory Instrumentation 114
21 Average Concentrations of Selected Species
at Jetmore, Kansas 115
xi
-------
ACKNOWLEDGMENTS
The authors thank Dr. Harold G. Richter, EPA Project Officer, for
his assistance and encouragement in the preparation of this report.
We also express our appreciation to Dr- Chester W. Spicer of
Battelle Columbus, Dr. Basil Dimitriades of EPA, Dr. Thomas E. Graedel
of Bell Laboratories, Dr. Harold Westberg of Washington State Univer-
sity, Dr- John Trijonis of Technology Service Corp., Mr. Francis L.
Ludwig of SRI International, Dr. Jack Fishman of the National Center for
Atmospheric Research, and Dr. John Noxon of NOAA for their participation
in the survey of exoert opinion.
From SRI we thank Dr- Susan Russell, Ms. Joyce Kealoha, and Mr-
Kenneth C. Nitz for technical contributions. Also, Mr. Jack Byrne and
Ms. Shirley Bartels for editorial/publications contributions and Ms.
Shirley Webster for secretarial assistance.
We are grateful for the cooperation of Mr- Dale Coventry of EPA,
who furnished the trajectories used in Section IV of this report, of
Environmental Research and Technology, Inc. for providing the data from
the Sulfate Regional Experiment, and of the Coordinating Research Coun-
cil, Inc., for their permission to use the data obtained in SPI Project
No. 6690.
xiii
-------
I INTRODUCTION
This report describes the results of a study of the role played by
nitrogen oxides (NO ) in the formation of ozone (0.,) in nonurban areas-
X J
The investigation was performed in support of the U.S. Environmental
Protection Agency's review of current strategies for controlling photo-
chemical oxidant pollution.
A. Background
The recognition that the occurrence of elevated 0^ concentrations
is a phenomenon that extends beyond the boundaries of urban areas has
introduced additional complications into the already complex problem of
how best to control photochemical oxidant pollution. Ozone being a
secondary pollutant chemically synthesized in the troposphere from
hydrocarbons (HC) and NO ,* its control allows for various options
X
regarding relative reductions in emissions of HC and NOX. Current
strategies emphasize reducing HC more than NOV emissions, which results
X
in low HC/NOX ratios. Such a scheme is urban-oriented because it tends
to slow down ozone formation near the emissions source. While this may
be satisfactory for urban areas, it appears that this approach could
worsen ozone pollution farther downwind, thereby shifting the air qual-
ity problems from one area to another. This possibility has prompted a
reexaraination of abatement policies, including the determination of
whether more stringent NOX controls are required.
Thus, from a regulatory standpoint, the question of interest is to
determine the extent to which anthropogenic NO affects ozone formation
X
outside urban areas. A primary purpose of this study is to review
available evidence that may help define the nature of NO /Oo
X j
Tropospheric ozone levels are also affected by injection of ozone from
the stratosphere. In this report we are solely concerned with the for-
mation of ozone from its precursors.
-------
interactions in nonurban areas. However, the investigation is not
intended to provide a definitive answer to the question.
B. Objectives and Methodology
The specific objectives of the project are to:
• Survey and summarize current knowledge about NOX and its role in
f>2 formation outside urban areas.
• Identify knowledge and data gaps where further researcjj is
indicated.
• Recommend research approaches that will help to define the rela-
tionship between NOX and nonurban 0^.
To achieve the objectives, the scope of work of the project speci-
fied conducting a survey of the expert opinion of several leading
researchers in the field. This was expanded to include performing an
extensive review of the literature and a preliminary analysis of two new
sources of data on nonurban NO and 0.,. One source of measurements was
J\ J
the Sulfate Regional Experiment (SURE), which was sponsored by the
Electric Power Research Institute. The other data source was a project
conducted by SRI International in Jetmore, Kansas, which was sponsored
by the Coordinating Research Council, Inc.
C. Report Organization
The literature review is described in Section II, and the results
of the survey of expert opinion are summarized in Section III. Sections
IV and V contain the analyses of field data. This is followed by con-
clusions and recommendations in Sections VI and VII, respectively. An
extensive bibliography is provided in Section VIII.
-------
II LITERATURE REVIEW
A. Introduction
The literature review was undertaken to survey current knowledge
regarding the role of NO in nonurban ozone formation. The recent
literature was surveyed extensively, including papers from scientific
journals and technical reports published by public and private concerns.
(Section VIII provides a complete bibliography.) The information gained
was categorized under the following topic headings, each discussed in
greater detail in the subsections that follow:
• NOX Measurement Methods
• Nonurban NOX Concentrations
• NOX Sinks
• Sources of Nonurban NOX
• NOX Chemistry
• Mathematical Models
• NOX/03 Relationship.
B. N0_ Measurement Methods
~^™" Jx ^m^**^^^^f^m^
A number of methods have been used to measure oxides of nitrogen in
the ambient atmosphere; manual and automated continuous methods are sum-
marized in Table 1. Currently, the most commonly used method employs
instruments based on the principle of chemiluminescence, which measures
NO directly with high sensitivity and reliability. The chemiluminescent
reaction
NO + 0 - NO* + 0
J ^ £
NO* ~ N02 + hv
-------
Table 1
SUMMARY OF MEASUREMENT METHODS FOR NOX IN AMBIENT AIR
(Knelson and Lee, 1977)
Method
Type
Description
Comments
Jacobs-Hochheiser
Ma nua1
Sodium arsenite
Triethanolamine
Manual
Manual
Griess-Saltzman
Manual/
automated
Chemiluminescent
Automated
Collection in aqueous NaOH, conversion
of N02 to nitrate ion, which is
measured colorimetrically
Similar to Jacobs-Hochheiser method;
sodium arsenite added to aqueous NaOH
collection solution
Collection in aqueous solution of
triethanolamine containing small
amount of organic compounds; colori-
metric determination of nitrate ion
Diazonium salf formed from the
reaction of NO2 with sulfanilic acid
is coupled with N- (1-naphthol)
-ethylenediamine dihydrochloride which
is measured colorimetrically
NOo quantitatively reduced to NO,
which reacts with ozone to produce
light
Poor removal efficiency (35%);
lack of constancy in correc-
tion factor; dropped as EPA
reference method
Collection efficiency improved
to 85%
Appears to be reliable
Minimal interferences
Measures NO and NOX, not sub-
ject to. interferences from
common air pollutants,
exhibits good precision
-------
produces radiation emission in the 600 to 3000 nm range with a maximum
at about 1200 nm. A photomultiplier tube is used to detect the lumines-
cent radiation and thus the presence of NO.
Since only NO is measured directly, NO must be reduced to NO to
X
allow for the measurement of NOX (NO + N02)• Although N02 has been con-
verted to NO in many ways, most commercial instruments reduce N02 to NO
catalytically at temperatures below 300°C. It is generally assumed that
converter efficiency approaches unity. N0« has also been converted to
NO at much higher temperatures (600°C), but this sometimes results in
reduction to NO of several other interfering species, including ammonia
(NH.,). The lower temperature (< 300°C) conversion of N02 to NO is
therefore preferred.
Instruments that use converters below 300°C do not show interfer-
ence from 0
-------
measurements in rural areas, chemiluminescent analyzers offer the best
option. Slight interferences from PAN, HC1, and HN02 may exist, but are
unlikely to be important in rural environments. Calibrations for NO can
be performed using multiple-dilution techniques and the standards can be
stored for a relatively long period of time. N02 calibrations can
either be done from NO standards by gas phase titration or by using per-
meation tubes. Commercial instrumentation can be modified to achieve
better sensitivities with only limited reliability below 1 ppb for NO.
A column content measurement of tropospheric N0~ has been made by
Noxon (1975, 1978) using absorption spectroscopy. The identification of
N02 with complete spectral scans at 5°A interval between 4350 to 4500°A
was considered to be unambiguous. The method can be used to make NOo
column content measurements in the atmosphere using the sun as the
source. This method was one of the first to suggest tropospheric N02
levels of less than 0.1 ppb, a concentration that is beyond the capabil-
ity of most chemiluminescent analyzers. The method is most suitable for
obtaining spatial averages (rather than point estimates) of concentra-
tion, and is still in a developmental stage.
C. Levels £f Nonurban N0_x
To understand the nonurban NO phenomenon it is necessary to exam-
ine the concentrations of NOX observed at a variety of locations ranging
from clean remote areas to suburban locales directly impacted by emis-
sions from neighboring cities.
Below we discuss NO measurements that have been reported in the
literature. We also include measurements for urban areas to span the
full range of NOX concentrations. In Sections IV and V of this report
we supplement the figures obtained from the literature with data col-
lected in 1977 at several nonurban locations in the United States.
Tables 2 through 4 catalog NOX concentrations observed in remote,
rural and suburban, and urban areas, respectively. In addition to the
NO-, levels, the tables show the location, method, and date of the
j\
-------
Table 2
CONCENTRATION OF NITROGEN OXIDES OBSERVED
IN CLEAN REMOTE AREAS
Location
La ramie, Wyoming
Fritz Peak,
Colorado
Fritz Peak,
Colorado
Northern
Michigan
Tropical Areas
Ireland
Concentration (ppb)
NO
X
O.I - 0.4
0.2 - 0.5
0.3 - 0.5
NO
0.01 - 0.05
0.1 - 0.5
0.3 - 0.4
0.3 - 0.7
0.3 - 0.7
£0.2
N02
<0. 1
0.2 - 0.4
0.3 - 0.5
0.3 - 0.6
0.3 - 0.5
0.2 - 1.5
Measurement
Me thod
Chemi luminescent
Absorption
spectroscopy
Chemi luminescent
Chemi lumi nescen t
Saltzman
Chemi luminescent
Date of
Measurement
Summer 1975
Fall 1974;
Summer -
Spring,
1975-1976
Sep 1977
Jun 1977
1965-1966
Jul-Nov
1974
Reference
Drummond (1977)
Noxon
(1975, 1978)
Ritter et al.
(1978)
Ritter et al.
(1978)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Lodge and Pate
(1966);
Lodge et al.
(1974)
Cox (1977)
Remarks
10-day average
Under forest
canopy
Above forest
canopy
River bank
Seashore
and maritime
Maritime. Maximum
hourly averages
ranged from 0. 3
to 5.0 ppb.
-------
Table 3
CONCENTRATION OF NITROGEN OXIDES OBSERVED
IN RURAL AND SUBURBAN AREAS*
00
Location
Piedmont, North
Carolina
Porton, England
Athens v I lie, Illinois
and Co f man, Missouri
Wilmington, Ohio
McConnelsville, Ohio
Wooster, Ohio
McHenry, Maryland
DuBois, Pennsylvania
Fritz Peak, Colorado
New Carlisle, Ohio
Huber Heights, Ohio
Cold Lake, Alberta,
Canada
Bradford, Pennsylvania
Creston, Iowa
Wolf Point, Minnesota
De Ridder, Louisiana
Concentration (ppb)
NO
X
17
26
NO
1.9
5
8
2
3.8
<1
1.5
N02
5.6
10
1-3
7
6
7
6
10
10
1-2
2.7
2.3
<1
2.6
Measurement
Method
Modified Saltzraan
Modified Saltzraan
Saltzman
Chemi luminescent
Cheml luminescent
Chemi luminescent
Chemi luminescent
Chemi luminescent
Absorption
spectroxcop-y
Chemi luminescent
Chemi luminescent
Absorption
spectroscopy
Chemi luminescent
Chemi luminescent
Chemi luminescent
Chemi luminescent
Date of
Measurement
Nov 1965-
Jan 1966
Oct 1972-
Jan 1973
Oct-Nov 1971
Jun-Aug 1974
Jun-Aug 1974
Jun-Aug 1974
Jun-Aug 1974
Jun-Aug 1974
Fall 1974
Jul-Aug 1974
Jul-Aug 1974
Feb 1977
Jun-Sep 1975
Jun-Sep 1975
Jun-Sep 1975
Jun-Sep 1975
Reference
Ripperton
et al. (1970)
Nash (1974)
Breeding
et al. (1973)
t
t
t
t
t
Noxon (1975)
Splcer et al.
(1976a)
Spicer et al.
(1976a)
Noxon (1978)
t
t
t
t
Remarks
Mean of 13 samples.
Range Is 4-21 ppb.
Mean hourly concentra-
tion obtained from con-
tinuous samples
Column average concen-
tration is attributed
to transport from Denver
Twenty-day average
from continuous samples
Suburban location
Column average
All locations are rural unless otherwise specified.
Research Triangle Institute (1976).
T
-------
Table 4
CONCENTRATION OF NITROGEN OXIDES OBSERVED
IN URBAN AREAS
Location
San Francisco
Bay Area ,
California
Los Angeles
County,
California
Boulder,
Colorado
Denver,
Colorado
Edmonton,
Alberta,
Canada
Dayton, Ohio
Concentration (ppb)
NO
X
47
NO
22
N02
22 - 42
35 - 89
<0.5 - 20
1 - 40
4-20
Measurement
Method
Chemi luminescent
and modified
Saltzman
Chemi luminescent
and modified
Saltzman
Absorption
spectroscopy
Absorption
spectroscopy
Absorption
spectroscopy
Chemi luminescent
Date of
Measurement
1977
1977
Feb 1977
Feb 1977
Feb 1977
Jul-Aug
1974
Reference
California Air
Resources Board
(1978)
California Air
Resources Board
(1978)
Noxon (1978)
Noxon (1978)
Noxon (1978)
Spicer et al.
(1976a)
Remarks
Range of annual average
NC>2 for the area. Range
of maximum hourly average
N02 is 110-260 ppb.
Range of annual average
for the area. Maximum
hourly averages ranged
from 240 to 690 ppb.
Column average
Column average
Column average
20-day average. Range
of maximum hourly averages
is 40-451 ppb for NOX
and 12 to 408 ppb for NO.
-------
measurements, as well as the corresponding literature reference. Unless
otherwise specified, all measurements, were made at ground level. In
compiling these tables we have purposely omitted any measurements
obtained using the Jacobs-Hochheiser method, in view of the well-known
problems associated with this procedure (Federal Register, 1973; cf.
Table 1).
Several aspects of the tables require explanation. The first con-
cerns the concentration estimates obtained using absorption spectros-
copy. These concentrations are averages taken over a column of air
whose horizontal and vertical dimensions are of the order of a few
kilometers. Hence, this method smooths out concentration gradients and
in general will yield pollutant estimates that are lower than ground-
level measurements. These considerations are important in urban areas,
where gradients can be steep; they are of less consequence in cleaner
environments, where uniformity is the rule.
Second, the tables give average values of the compounds, when
available, in order to portray long term conditions at the various
sites •
The third and final point entails distinguishing between remote,
rural, and suburban sites. A remote location is one that is far from
populated areas; hence, its pollutant levels are seldom influenced by
anthropogenic sources. However, in the United States, even remote areas
are sometimes affected by pollutants transported from urban areas, as
evidenced by the entries for Fritz Peak, Colorado, in Tables 2 and 3. A
rural site, by contrast, would be one that is more likely to be impacted
by transport from urban areas. Rural sites include small conmunities
that generate pollutants, but in general do not contain any major
sources of pollution. Suburban areas are those that are within commut-
ing distance, (viz., 20 to 25 miles) of an urban area, and thus bear the
brunt of transport from the urban core.
10
-------
Examination of Tables 2 through 4 reveals certain patterns of N0x
levels for the various areas. Table 2 shows that remote sites have NOX
concentrations that are below 1 ppb. Chemiluminescent measurements show
NOX ranging from 0.1 to 0.5 ppb, the bulk of the NOX being N02, whereas
the absorption spectroscopy method suggests that N02 is less than 0.1
ppb. The latter figure has been questioned by Ritter et al. (1978),
whose concurrent ground-level Chemiluminescent measurements of N02 at
Fritz Peak found higher concentrations. The discrepancy probably
results from the estimation procedure used to obtain the volumetric
column average from the molecular density measurements obtained by
absorption spectroscopy. The procedure requires assuming an effective
height for the mixed layer, a quantity that is subject to large uncer-
tainties. Because no such assumption is involved in Ritter's concentra-
tion estimate, we believe it to be the more accurate.
Measurements of NO at remote sites are very sparse, the most reli-
able being those by Drummond (1977); they suggest a mean NO/NO ratio of
X
less than 0.2. The same bound is indicated by data collected by Ritter
et al. (1978) under clean conditions at a site in Michigan. Thus, it
appears that, on the average, N02 is the predominant species in remote
locations. However, either NO or NO- can predominate at any given
instant. As a footnote, it is noted that the values of N02 shown in
Table 2 for maritime sites generally agree with the 0.5 ppb ocean back-
ground level suggested by Robinson and Robbins (1970).
For rural areas, Table 3 shows that mean N02 concentrations range
from 1 to 10 ppb. The data for the single suburban site at Huber
Heights, Ohio, suggests an N02 level in excess of 10 ppb. Although
reported NO measurements are scarce, the available data indicate NO/NO
X
ratios greater than 0.2. This was expected, since these sites are
likely to be influenced by sources such as roads, urban areas, and power
plants whose respective plumes can contain substantial amounts of
unreacted NO.
11
-------
As shown in Table 4, urban areas contain a wide range of N02
values, the means ranging from 1 to 90 ppb, with peak hourly levels
exceeding 600 ppb in Los Angeles. The NO/NOX ratio for Dayton is about
0.5, which should be typical of urban areas. Of course, morning
(0600-0900 LT) levels of N0x in cities are primarily attributable to
mobile sources, and NO is the predominant compound by far in this time
interval.
To summarize, the data obtained from the literature indicate NO
X
levels below 1 ppb in remote areas, with a mean NO/NOV ratio bounded
X
above by 0.2. For rural areas, mean NO levels range from 1 to 10 ppb,
X
with an NO/NOX ratio of 0.2 or greater. In urban areas, mean N02 levels
span the range 1 to 100 ppb, the concentrations varying widely depending
on geographical location. The mean NO/NOX ratio in urban areas probably
exceeds 0.5.
Regarding N02 background levels, it should be noted that the 4 ppb
concentration proposed by Robinson and Robbins (1970) as a "natural
background" level for land areas in northern temperate zones is within
the range of cooncentrations observed in rural areas of the United
States. However, recent N02 measurements at remote locations show N02
being less than 1 ppb, which suggests that the 4-ppb level contains a
sizable anthropogenic component. This implies that previous estimates
of the atmospheric residence time of N02 need to be reevaluated.
D. Sinks ^f NOX
Once in the atmosphere, NO (NO + N02) is subject to various
removal processes. Atmospheric residence times from 1 to 20 days have
been suggested. Two principal removal mechanisms are operational:
• Direct removal of gaseous NOX by dry and wet deposition
processes.
• Conversion of NOX to nitric acid or particulate nitrate, fol-
lowed by removal by dry or wet deposition processes.
12
-------
In the atmosphere, M02 is oxidized to nitric acid, nitric acid
aerosols, and nitrate aerosols via several pathways listed in Table 5.
The quantitative importance of the mechanisms listed in Table 5 is
currently uncertain. In the relatively clean atmosphere, at least half
of the N02 is converted to HNO-j via reaction with HO. The second path
(Reactions 2 through 4), involving conversion of NC^ to NC^, ^Og, and
finally HNOj, would be about a third as fast as the reaction with HO in
rural atmospheres. It is quite possible, however, that Reactions 2, 3,
and 4 may dominate in polluted atmospheres with very high o'zone levels.
The volatility of nitric acid is such that it is not appreciably
taken into water droplets unless the droplets contain neutralizing reac-
tants. It would thus appear that homogeneous nucleation (Reactions 5
through 7) would not constitute an important mechanism for nitrate
formation.
Nitric acid can exist in the gas phase or react with ammonia to
form particulate ammonium nitrate. It has also been postulated that
aerosol formation could result from direct absorption of NO into aque-
X
ous droplets in the presence of ambient ammonia (Paths III and IV).
While both processes may proceed simultaneously, the presence of ammonia
seems to be essential. The atmospheric abundance of ammonia in urban
and rural atmospheres is poorly determined. Moreover, the rate con-
stants (especially that for Reaction 7) are not firmly established.
Thus, while Paths III and IV may be important, it is not possible to'
obtain a quantitative estimate of their contribution. It is predicted
that about half of NOX is converted to nitric acid and nitrates before
being lost. Nitric acid may be an important component of acid rain.
In the lower atmosphere, and where some pollution exists, other
sinks may be possible. For example, NOX in the presence of organics is
converted to organic nitrogen compounds (such as peroxyacetyl nitrates,
PANs) that could then be lost to the ground by dry or wet deposition or
could slowly hydrolyze (Path V). While present in the gas phase, PANs
may not be a final sink for NOX, since they may decompose to regenerate
13
-------
Table.5
NITRIC ACID AND NITRATE AEROSOL FORMATION REACTIONS
Path
Reaction
No.
I Nitrogen dioxide - hydroxyl radical reaction:
N02 + HO* + M - HN03 + M
II Nitric acid anhydride reaction:
N02 + 03 - N03 + 02
N02 -
N2°5
H2° ~* 2HN03
III Homogeneous conversions of nitric acid to aerosol:
HN03 (g) + H20 (g) - HN03 aerosol
^5 (g) + H20 (j|) - HN03 aerosol
NH3 (g) + HN03 (g) - NH4N03 aerosol
NO absorption into aerosol droplets:
A
2N02 + H20 GO - HN03 + HN02 (aqueous solution)
NO + N02 -f H20 (A) - 2HN02 (aqueous solution)
HN02 (aq.) - H + NO" (aqueous solution)
NO, + ir02 (in aerosol solution) - N03
IV
V
N02 + 03 (in aerosol solution) -• N03 + 02
Hydrolysis 6f PANs:
0 0
RCOON02 + HjO
N02 + 02 -*
- RCON -I- 02 + HN02
H+ +
VI Methyl nitrate formation:
CH30"
Methoxy
radical
N02 -
methyl
nitrate
VII Pernitric acid formation:
H0
N02 -
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
14
-------
N02 and organic radicals. The formation of methyl nitrate has also been
postulated, although no conclusive atmospheric identification has been
made. Its atmospheric fate is largely unknown; however, reaction with
HO is expected to provide at least one removal mechanism. Another tem-
porary sink could be pernitric acid (H02N02), but it is likely to be
important only at relatively colder temperatures. At temperatures
greater than 70°F, when most smog occurs, the role of HO2^2 is limited
(Cox at al., 1977).
Gaseous nitrogen species may be used up by surface absorption.
Vegetation and soil are capable of removing significant amounts of NO,
N02, PAN, and other species from the atmosphere (Tingey, 1968; Rogers et
al., 1977; Hill, 1971; Sundareson et al., 1967). Dry deposition of par-
ticulate matter occurs through sedimentation, Brownian motion, and
impaction. Dry deposition of gaseous NO is only about half as effec-
X
tive as the corresponding particulate dry deposition. The rate of
removal is strongly dependent on wind speed and the nature of the depo-
sition surface.
Rainout and washout are the two major wet removal processes. The
former refers to the processes of nucleation occurring within a cloud;
the latter involves removal below the clouds by falling hydrometeors.
Due to high concentration of nitrogen species within the boundary layer,
washout may dominate the removal process. In the free troposphere, how-
ever, both processes may play an important role. Together, wet and dry
deposition are the major ultimate sinks for the nitrogen containing
species.
15
-------
E. Chemistry of N0_x
Nitrogen oxides play a principal role in smog formation. Perhaps
the most central reaction is the NO-NO2 null cycle:
NO- + h*(385 nm < \ < 435 nm) - NO + 0 (1)
0 + 02 -1- M - 03 + M (2)
NO + 03 - N02 -I- 02 (.3)
NET: no reaction
The NO-NO2 null reaction becomes important when the system is per-
turbed by the inclusion of hydrocarbons. In a complex way, the hydro-
carbons induce the oxidation of NO to N02 without involving 02 destruc-
tion. This is accomplished by the peroxy free radicals in Reaction 4,
and results in net 02 production as indicated below
R02 + NO - RO + N02 (4)
N02 + hv - NO + 0
NET: R0 + Q - RO
16
-------
The peroxy free radicals are principally produced from hydrocarbon
attack by free radicals [HO, 0(3P)] and by ozone itself. Of these HO is
known to play a dominant role
RH + HO - R + HO (5)
R + 0 - R0 (6)
where RH denotes a generic hydrocarbon compound.
Other peroxy free radicals, such as RCO,, are also produced by
reaction of HO with alhehydes and that of 03 with olefins• The RO radi-
cals generated in Reaction 4 are oxidized by oxygen to form aldehydes
(RO + 02 -> H02 + aldehydes). In a way similar to R02, the RCO-j radical
can also lead to net 0, formation as follows:
RC03 + NO + 02 - R02 + N02 + C02 (7)
NO + hv - NO + 0
0 + 0 + M - 0
NET: RC03 + 20^- R02 + 0 +
Thus, the RCO- radical not only contributes to net 0, production but
also releases an R02 radical, which can continue the chain reactions.
Thus, in principle, this chain will continue as long as there is NO ,
X
hydrocarbons, and sunlight. By allowing the oxidation of NO to N02 by
processes other than reaction with 03 (Reaction 3), a build-up of Oo
takes place. Almost any hydrocarbon (also CO) that can react with free
radicals can participate in the ozone-forming process.
17
-------
While initially the reactive free radical HO is released from 0*
photolysis, eventually other reactions take over as follows:
03 + hv - 0(LD) + 02 (8a)
0(LD) + H20 - 2HO (8b)
H02 -I- NO - N02 + HO (9)
In the very clean atmosphere, Reaction 8 is the dominant source of
HO, whereas in a highly polluted atmosphere Reaction 9 predominates. In
rural atmospheres, it is possible that both reactions would be important
contributors to ambient HO.
Perhaps the simplest smog cycle is due to CO:
CO + HO - C02 + H
H + 02 + M - H02 + M
H02 + NO - HO + N02
N02 + hv - NO + 0
0+02+M-0_+M
NET: CO + 202 - C02 + 0
However, in suburban and rural areas, where CO levels are low, the
CO-related reactions are generally not significant ozone producers. In
remote areas and in the free troposphere, it is hypothesized that CO
could become important in the presence of NO levels greater than 0.1
ppb.
18
-------
The chain-carrying radicals are also destroyed in a number of ways
In some cases, the radicals provide a sink for NOX. Important chain-
termination reactions are as follows:
HO + N02 + M - HN03 + M
HO + H0 - H0 + 0
02 (PAN)
NO + N0 + H0 - 2HN0
HN02 + hv - HO + NO
N°2 + °3 "* N°3 + °2
N2°5 + H2° "* 2HN°3
M
HO. + NO- ? H00NO-
2 2 22
HO + N02 - HONO -t-
It should be noted that the ratio of the rates of forward and reverse
reactions for PAN and H02N02 is highly-temperature dependent, the rate
of the reverse reaction increasing with temperature (Hendry and Kenley,
1978).
All mechanisms of ozone formation in the troposphere need NO for
ozone synthesis. In the urban centers, NO may build to such high lev-
els as to actually suppress 03 production. This is unlikely to happen
19
-------
in nonurban locations where NO-NC^ levels are relatively low. Since the
rate of NOX removal from the atmosphere is at least of the order of
several hours, fresh emissions allow a residue of NOV to remain that is
3t
sufficient to sustain the photochemical chain reactions that result in
smog formation. It is also possible that 0^ gas phase loss mechanisms
are minimized in nonurban areas arid the smog cycle is more efficient.
It is expected that hydrocarbons of either man-made or natural origin
would be available in nonurban areas to participate in smog chemistry.
In the free troposphere, the important carbon compounds would.be CH^,
C2H^, and CO. In the nonurban atmosphere, other light alkanes would
dominate (butanes, pentanes). Small amounts of alkenes, (e.g., 02^)
may also be present (Singh et al., 1978).
F. Sources of NO
•^^^™~™ ~^*"™ ^™^™X
Both man-made and natural sources contribute to the total atmos-
pheric NOX burden. The natural NOX sources at this time are poorly
understood and are largely speculative*
1. Man-Made Sources
The principal source of nitrogen oxides is the oxidation of atmos-
pheric nitrogen in combustion processes. The oxidation of nitrogen at
high temperatures primarily occurs through the following mechanisms:
02 2 20
0 + N. 2 NO 4- N
N + 02 NO + 0
20
-------
The production of NO is critically dependent on the atomic oxygen
concentration, which in turn is dependent on the combustion temperature.
The most common combustion processes involve:
• Fuels—coal, petroleum, wood, natural gas, refuse
• Processes—ppwer generation, industry, domestic heating,
refinery production, transportation, controlled burning (forest,
sugar cane), and incineration, among others.
Figure 1 shows the actual and projected trends in emissions of NOX in
the United States by source categories. The global man-made emissions
have been estimated to be three times those of the United States
(Soderlund and Svensson, 1976). Thus, it is possible to estimate that
currently about 10 Mt(N)/yr of NOX is emitted in the United States and
the estimate of global emission rate is 30 Mt(N)/yr-
Not all NO., is released in a primary mode. Combustion sources,
X
animal wastes (feedlots and pastures), and volatilization of nitrogen
fertilizer release significant quantities of ammonia, which can be oxi-
dized to NO or N0« in the atmosphere. This conversion process is poorly
understood, and its occurrence has been questioned (Cox et al., 1975a).
No complete NHj emissions inventory is available for the United States.
Best estimates are that man-made activities result in the release of 50
to 80 Mt(N)/yr of ammonia (National Academy of Sciences, 1978). Burning
of coal accounts for 4 to 12 Mt, volatilization from animal and human
waste account for 20 to 35 Mt, and inefficiencies in handling and appli-
cation account for about 30 Mt of NH*. Thus, in principle, man-made
releases of ammonia can be larger than direct releases of NO . Best
X
estimates are that no more than 3 to 4 percent of ammonia would be con-
verted to NOX (National Academy of Sciences, 1978). Thus, emissions of
ammonia would contribute about 10 percent to the NO burden.
21
-------
12
10
I
3
to
2 6
to
I
Ul
x
O
z
Total Emiuions
Industrial
and Miscellaneous
1940 1950 I960 1970 1980
YEAR
Source: National Academy of Sciences, 1978
FIGURE 1 ACTUAL AND PROJECTED TRENDS IN EMISSIONS OF NOX
IN THE UNITED STATES BY SOURCE CATEGORIES
1990
22
-------
2. Natural Sources
Natural sources also release NOX and NHj via a variety of mecha-
nisms. The quantitative estimates of fluxes are poorly understood, how-
ever. Table 6 shows the various natural processes that result in a
release of either NOX or NH-j.
Table 6
NATURAL SOURCES OF NO
Source Type
Soil: decomposition and volatilization
of organic matter and animal wastes
under alkaline conditions
Chemodenitrification in acidic swamps
and soils
Forest fires
Volcanoes
Lightning
Inflow from the stratosphere (from
photolysis of N20)
Ma j o r
Species
NH,
NO
NH NO (?)
J X
NO
NO
Since NO is unreactive in the troposphere,
it is not included.
Of all the processes listed in Table 6, the biological and chemical
transformations of nitrogen compounds in the soil appear to be an impor-
tant source NOX- Although this source is poorly quantified, a global
flux of 3 to 60 Mt(N)/yr can be estimated (Ratsch and Tingey, 1978;
Soderlund and Svensson, 1976). Inflow from the stratosphere (from N20
decomposition and photolysis) is not expected to exceed 0.5 Mt(N)/yr and
23
-------
is negligibly small. Soderlund and Svensson (1976) estimate that an NOV
X
source of 20 to 90 Mt/yr is necessary to balance the NOX cycle. The
difference between the soil source and total NO required may be
explained by NOX generated from lightning. Estimates of global produc-
tion by lightning range from 10 to 20 Mt(N)/yr (CAST, 1976; Delwiche,
1970; Holland, 1973). More recently, however, Chameides et al. (1977)
estimate that global production of NOX during lightning is 30 to 40
Mt/yr. Although uncertainties abound, it is possible that a global
natural source of NOX is 20 to 90 Mt(N)/yr and is largely composed of
NOX produced by lightning and biological and chemical transformations in
the soil.
Emissions of NHj, however, are expected to be much larger. Esti-
mates of Robinson and Robbins (1971) suggested a natural NOj source of
870 Mt(N)/yr- Soderlund and Svensson (1976) have improved this estimate
and suggest that the natural sources of NH, lie between 80 and 200
Mt/yr- The available data are extremely poor and subject to major revi-
sions. Together, both natural and man-made sources of NH-, would result
in the formation of 2 to 8 Mt/yr of NO...
A
G. Mathematical Models
Efforts to model photochemical pollution date back to the late
1960s, their main thrust being the simulation of urban pollution.
Urban-oriented models continue to be developed, improved, and refined; a
useful summary of the state of the modeling art for urban areas circa
1976-1977 is given by Seinfeld and Wilson (1977). More recently, a
number of models have been developed that attempt to describe nonurban
pollution. The discussion that follows reviews these models. Because
we are concerned with models that attempt to simulate the physical
processes governing photochemical air pollution, our review will not
include statistical or curve-fitting approaches to the estimation of
pollutant levels.
24
-------
The review will be concerned with the following aspects of the
models:
• Model type
• Domain of applicability of the model
• Verification tests.
The first aspect, model" type, encompasses the structure and contents of
the model. The second aspect specifies those conditions under which the
model can be applied. The last, verification tests, examines the extent
to which the models have been tested against experimental data. Appli-
cations of the models will be discussed in the section that examines the
relationship between NOX and 02 in nonurban areas.
1. Model Type and Intended Application
All the models reviewed fall into one of two categories:
• Those that have chemistry only
• Those that contain both chemistry and meteorology.
Some models operate under steady-state assumptions, and others simulate
time-dependent conditions. Table 7 lists and classifies the models
reviewed; each model is identified by the name of the developer(s). The
table also shows the type of application for which each model is
intended•
The models listed in Table 7 have as their centerpiece a set of
chemical reactions that describe the photochemistry of nitrogen oxides,
ozone, and hydrocarbons. The kinetic modules of the various models are
similar in many ways, which implies that there is general agreement on
the major features of the chemistry. In particular, the description of
the gas phase chemistry of NOX differs only slightly in the various
models. The major chemistry differences involve the hydrocarbons.
Thus, models that simulate clean environments do not include the same
hydrocarbons that appear in models intended for more polluted areas. In
general, the models emphasize chemistry at the expense of transport and
25
-------
Table 7
PHOTOCHEMICAL MODELS FOR NONURBAN AREAS
Model
Identification
Classification
Intended
Application
References
Chang and Weinstock
Chameides
Liu
Fishman and Crutzen
Hov and Isaksen
Graedel and Allara
Chemical; time-dependent
Steady-state; chemistry
and parameterized
vertical transport
Time-dependent; chemistry
and parameterized
vertical transport
Time-dependent; chemistry
and vertical transport
and diffusion
Time-dependent; chemistry
and parameterized
dilution
Time-dependent; chemistry
and instantaneous mixing
within a variable volume
Rural areas
influenced by
urban transport
Remote areas
Remote areas
Remote areas
Rural areas
influenced by
urban transport
Rural areas
influenced by
urban transport
Chang and Weinstock
(1967a,b)
Chameides (1978);
Chameides and Stedman
(1977)
Liu (1977);
Liu et al. (1976)
Fishman and Crutzen
(1977)
Isaksen et al. (1978a,b);
Hov'et al. (1978)
Graedel and Allara (1976)
All the models are one-dimensional.
-------
diffusion. Moreover, the models are one-dimensional, i.e., they con-
sider a single column of air as it moves over a region.
Chang and Weinstock's model employs 57 reactions, including reac-
tions for methane (CH4), other alkanes, and aldehydes. The model
assumes instantaneous mixing and considers the air column to be a moving
reactor.
The models devised by Chameides and by Liu use the same chemical
description, which contains 47 reactions. Being intended for simulating
remote environments, the chemistry does not include aldehydes or any
primary hydrocarbons other than CH,. Thus, CH4 oxidation is the princi-
pal source of organic radicals in these models. The two models differ
only in the numerical values assigned to some of the reaction rate con-
stants. Chameides' model assumes steady state conditions, and thus is
useful only for estimating bounds on pollutant levels. Liu's model
operates under dynamic conditions, thereby yielding concentration his-
tories for the species of interest. Neither model can be used to assess
the impact on ozone -of hydrocarbons and NO transported from urban
A.
areas. It may be possible to study the effect of NOX transport on NOX
and ozone levels using these models.
The meteorological component of Flshman and Crutzen's model is more
elaborate and realistic than in the models previously discussed, since
it includes time-dependent vertical diffusion and transport. Its chemi-
cal module contains 42 reactions and, as in the case of Chameides's and
Liu's models, considers CH^ oxidation to be the principal source of
organic radicals. The model can be used in a qualitative sense to gain
insights about the interactions of the various species.
The Hov and Isaksen model treats chemistry in great detail. Its
chemical module contains 161 reactions for NO and seven hydrocarbons.
The hydrocarbons are associated with urban plumes, and consist of pro-
pene (C3H6), ethylene (C2H4), m-xylene (mCgH10), acetylene (C2H2),
n-butane (nC4H1Q), and n-hexane (nCgH14), as well as CH4. The model
assumes instantaneous vertical mixing within the mixed layer in a moving
27
-------
air column whose horizontal dimensions are assumed to be large enough to
justify neglecting horizontal diffusion. However, the model allows the
mixing volume to change with time.
Hov and Isaksen's model represents an ambitious attempt to simulate
the phenomena associated with photochemical smog formation downwind of
urban areas. This is reflected in the explicit inclusion of kinetic
models of the oxidation of the various hydrocarbons listed above. While
their treatment of C3Hg, C2H4, C2H2, C4H10, nC6H14, and CH4 follows well
• ••*
known lines (see, e.g., Demerjian et al., 1974, and Hendry et al.,
1978), the formulation of m-xylene oxidation differs significantly from
other descriptions of the kinetics of aromatic compounds (Hendry, 1978).
The net thrust of Hov and Isaksen's approach seems to cause m-xylene to
induce a higher rate of NO oxidation. In effect, m-xylene is made to
appear to be more reactive than is thought to be the case, which implies
that ozone production attributed to m-xylene would be overestimated.
However, the reports of applications of Hov and Isaksen's model do not
discuss the contribution of individual hydrocarbons to ozone formation;
hence, we cannot establish the importance of m-xylene in the overall
mechanism. All that can be said at this time is that there may be a
potential problem associated with m-xylene, which could alter any infer-
ences drawn from applications of this model.
The model developed by Graedel and Allara is similar to the others,
in that it is a chemical model that allows the mixing volume to change.
The species are assumed to be instantaneously and uniformly mixed within
the volume; in essence, it is a box-type model. The chemical module is
highly detailed, containing over 300 reactions involving more than 200
chemical species. A unique feature of this model is the inclusion of
reaction sets for alpha-pinene and isoprene. It also includes CH^,
hydrogen sulfide, ammonia, and aldehydes. Thus, the model is intended
for studying the contribution of local photochemistry and sources on
nonurban ozone levels. It can also be used to examine the impact on
ozone of urban NOX.
28
-------
2. Model Verification
None of the models has undergone adequate verification tests
against field data; this is partly due to the sparseness of the data
available. Chang and Weinstock's model has been tested against smog
chamber data in a limited fashion. We found no evidence of verification
attempts by Chameides, Liu, or Fishman and Crutzen. Hov et al. (1978)
compared the predictions of Hov and Isaksen's model against 0
-------
H. Relationship Between NO and CU
^•«^^^M^MI^B«M^HMM^^B^^^ ^*^—*—i^^*~^~^— «^WJ£ MB^H^^ "^J_
Several hypotheses have been proposed to explain the rural ozone
phenomenon. Reduced to essentials, the hypotheses are:
• 0^ is locally produced from existing and transported precursors.
• Ozone is formed in urban plumes and transported to nonurban
areas.
• Rural ozone results from natural causes, such as stratospheric
intrusion.
It should be recognized, however, that rural ozone is due to a combina-
tion of factors, rather than to a single mechanism. Consequently, the
hypotheses are complementary rather than mutually exclusive. Thus, the
problem becomes one of attempting to find a predominant mechanism among
a variety of causes. For our purposes, we examine below various aspects
of the first two postulates.
Before proceeding to review the literature, we note that (for regu-
latory purposes) a basic question about the role of NO in ozone forma-
X
tion in nonurban areas is whether such ozone production is N0x-limited.
Ozone formation is said to be NO-_-limited when addition of NOV results
A X
in higher ozone concentrations. However, the NO -limited condition is a
A
function not only of NO , but also of the hydrocarbon concentration.
The situation is illustrated in Figure 2, which plots 03 as a function
of NOV with nonmethane hydrocarbon (NMHC) as a parameter. The figure
X
shows that 0«j increases with NOX until it reaches a maximum, decreasing
thereafter with further increases in NOX. Whether ozone increases or
decreases with increasing NOX is determined by the NMHC/NOX ratio. The
figure shows that for NMHC/NOX ratios greater than about 4.8, ozone will
be in the N0x-limited condition, and vice versa. The precise NMHC/NOX
ratio that forms the boundary of the NO -limited region will vary
Jw
depending on the type of hydrocarbons involved (the curves in Figure 2
are for a mixture of propylene and n-butane), so no special significance
attaches to the ratio of 4.8. Nevertheless, atmospheric values of
30
-------
40
36
32
28
6
a,
* 24
O
20
o
z
O
o
O 18
12
NMHC
- 10:1
NMHC
4.8:1
NMHC - 0.8
NMHC - 0.8
NMHC • 0.4
I
I
I
I
8 12 16 20
NO- CONCENTRATION (pphml
24
28
FIGURE 2 OZONE CONCENTRATION AS A FUNCTION OF NOX LEVEL
WITH NONMETHANE HYDROCARBON (NMHC) AS PARAMETER
31
-------
NMHC/NOX > 10 generally lead to N0x-limited conditions for 03 formation.
Thus, a high NMHC/NOX ratio is a necessary condition for ozone to be in
the N0x-limited region.
If Q.J formation in nonurban areas is N0x-limited, then the possi-
bility exists that increases in NO emissions from urban areas may
A
enhance ozone levels in nonurban regions downwind of the NOV source.
X
Hence, to reduce nonurban ozone, it would be necessary to control urban
NOX emissions. On the other hand, the case for controlling NOj emis-
sions is less compelling if nonurban ozone levels are not NO -tiraited.
A
This is the fundamental problem that motivates the literature review
described below.
The literature contains conflicting reports on whether adding NO
X
enhances, inhibits, or has no effect on ozone formation. The most
direct evidence of enhancement due to addition of NO has been recently
A
reported by Miller et al. (1978), who used an airplane to track a power
plant plume over Lake Michigan for four hours—a distance of about 90
miles. As expected, ozone in the NO -rich plume initially was below
Jv
ambient levels, the latter being about 90 ppb and the former approxi-
mately 60 ppb. After about one hour, ozone in the plume began to sur-
pass ambient levels, and continued to accumulate. After four hours,
ozone levels in the plume reached about 150 ppb compared to 100 ppb for
the ambient. The authors offer plausible evidence that the apparent
cause of this phenomenon was the addition of NOX to an atmosphere with a
NMHC/NOV ratio that was very high »30/1) initially. With such high
3C *
NMHC/NOX ratios, ozone production is N0x-limited. Hence, as the NOX
from the plume mixes with the surrounding air, the extra NOX enhances
ozone formation. It seems reasonable to expect that the type of hydro-
carbons present would determine the threshold that the NMHC/NOX ratio
has to exceed for the system to be in the N0x-limited region of ozone
formation. Unfortunately, no hydrocarbon composition data are given in
the published report. Lacking such data, it is risky to attempt to for-
mulate any general rules. Nevertheless, the fact remains that ozone
enhancement did occur, and that it appears to be directly related to the
32
-------
addition of NOX. Thus, under the proper conditions, similar phenomena
can occur in nonurban areas impacted by urban NO .
Jt
Jeffries et al. (1976) provide indirect evidence in support of the
concept of N0x-enhanced ozone production in nonurban areas. Their
investigation used an outdoor smog chamber irradiated by sunlight. Some
of the experiments involved extended irradiations over several days in
succession. Their experiments indicate that first-day ozone production
is dominated by the initial hydrocarbon concentration. However,
second-day ozone levels can be as high as the first day's, even though
most of the alkenes were consumed the first day. The implication is
that an aged air mass essentially devoid of alkenes can retain the
capacity to produce substantial amounts of ozone given a small amount of
NO . They estimate that the amount of NO,, required to produce such
A Jt
behavior is of the order of 5 ppb. However, we have seen in Table 3
that 5 ppb is typically observed in rural areas. Hence, if their con-
clusions are correct, the chemical conditions may already exist that
favor local ozone generation in nonurban areas.
To investigate the effects of dilution, Jeffries et al. (1976) per-
formed a three-day experiment in which the initial charge of the chamber
was irradiated and diluted with purified air until only 5 percent of the
original mass remained. The remainder was irradiated by sunlight for
two successive days, and produced significant amounts of ozone, but gen-
erally less than in the first day. These results lend credence to the
hypothesis that ozone production can occur in an air mass that has been
transported over a long distance, in spite of the substantial dilution
that the pollutants undergo.
An independent study by Ripperton et al. (1976), also using an out-
door smog chamber, supports the conclusions of Jeffries et al- (1976)
mentioned above. Ripperton et al. concluded that NO,, levels in the
j&
1- to 5-ppb range can generate substantial amounts of ozone. Their
results also parallel Jeffries with regard to dilution effects.
33
-------
Although highly suggestive, these results should be interpreted
with great caution owing to the uncertainties associated with wall
effects in smog chambers. Thus, it is generally recognized that the
walls can act as a source of free radicals that act to initiate the
hydrocarbon oxidation chain, thereby inducing a level of photochemical
activity that may not be representative of atmospheric conditions.
While the previous studies suggest that low levels of NOV can sus-
X
tain, and that additional NOX may enhance, local ozone production in
rural areas, a report by Lonneman (1976) concludes that ozone transport
from urban areas may be the dominant factor in determining rural ozone
levels. Lonneman discusses ozone, NO , and hydrocarbon measurements in
Jl
Wilmington, Ohio, a rural site surrounded by three metropolitan areas.
The key evidence indicating that transport predominates is that the high
ozone values occurred in the late afternoon and early evening. This
certainly suggests that a polluted air mass arrived at the site, since
local ozone generation is tightly coupled to the sunlight cycle. Thus,
a locally generated ozone maximum would occur around noon, at the time
of the peak in solar radiation intensity, rather than in the late after-
noon or early evening. It would have been interesting to see whether
the days containing the late ozone peak also displayed an earlier peak
around noon, but this is not discussed by Lonneman. In Section IV of
this report we examine some recent field data that show this double-peak
phenomenon.
That ozone is transported over hundreds of kilometers is by now a
well-documented phenomenon. The literature on this subject is extensive
and growing; some references of interest are Blumenthal et al. (1974),
Cox et al. (1975b), Decker et al. (1977), White et al. (1976), Westberg
et al. (1978 a,b,c), Spicer et al. (1976a), Research Triangle Institute
(1976), and Kauper and Niemann (1976). The reader should consult the
references for descriptions of the transport mechanism; here we summa-
rize the effects associated with transported ozone. In general, it
appears that transported ozone combines additively, rather than syner-
gistically, with local ozone to increase the overall concentration at
-------
the ground. This higher level can exceed the ambient standard; in some
cases, the transported ozone by itself exceeds the standard. Ozone
transport can occur over several days, but the evidence suggests that
single-day transport may be a predominant influence because of losses
that occur as ozone is subjected to chemical scavenging and dry deposi-
tion processes near the ground. Hence regeneration of ozone seems to be
required on a daily basis to sustain transport of high ozone concentra-
tions over several days. Thus, while transported ozone can be responsi-
ble for high downwind ozone levels observed within a travel time of 24
to 36 hours, the availability of ozone precursors may be necessary to
extend the transport range. These precursors may well be carried along
with the ozone itself, since the evidence suggests that an "aged" or
"spent" polluted air mass can continue to form ozone. Alternatively.
the precursors may be present already at the downwind location, for (as
we have seen) NOX levels of 1 to 10 ppb prevail in rural areas and such
levels are sufficient to form significant quantities of ozone, provided
that the necessary hydrocarbons are also available. Thus, in examining
the role of NO in nonurban ozone formation, the concentrations and
A
types of hydrocarbons that are present at the rural sites must also be
considered. A full review of this aspect of the problem is beyond the
scope of this report, and we shall only take a brief look at it.
Both natural and anthropogenic hydrocarbons are found in rural
areas of the United States. The role of natural hydrocarbons (NHC) in
the formation of nonurban ozone is a subject of considerable controversy
(cf. Coffey and Westberg, 1977; Lonneman et al., 1978 and 1979; Sculley,
1979; Ludlum and Bailey, 1979). Some researchers believe that NHC are
unimportant, while others assert that further research is needed to
prove or disprove this hypothesis. Others have used statistical argu-
ments to postulate that NHC are significant contributors to urban and
suburban smog, a contention that has been hotly debated (Sandberg et
al., 1978 and 1979a,b; Bufalini, 1979; Miller et al., 1979). A recent
study by Arnts and Gay (1979) examined the photochemistry of selected
NHC, including isoprene, p-cymene, and six monoterpenes. The investiga-
tion was aimed at estimating the ozone-forming potential of these NHC in
35
-------
the presence of NOX in a controlled laboratory environment. They con-
clude that tnonoterpenes do not permit ozone to accumulate because of the
fast reaction between ozone and monoterpenes, which agrees with sugges-
tions previously advanced by Westberg (Coffey and Westberg, 1977).
Isoprene was found to produce significant amounts of ozone, but not at
high carbon/NOx ratios, where some ozone suppresion was observed. How-
ever, the hydrocarbon concentrations used in the experiments were higher
than have been observed in the atmosphere, and the results of the study,
while suggestive, can not be readily extrapolated to the atmospheric
milieu. In view of the wide difference of opinion, it is clear that
nothing definitive about the role of NHC can be said at this time, which
leaves the anthropogenic component to be dealt with.
Light alkanes (e.g., butanes, pentanes, and ethane) and alkenes
(e.g., ethylene) have been observed in rural areas. Spicer et al.
(1976a) report 3-hour average NMHC levels at a rural site downwind of
Dayton, Ohio, ranging from 0.67 to 0.72 ppmC. Ethane, ethylene, and
acetylene were also detected at the same site, the latter being an indi-
cator of anthropogenic influence since it is primarily emitted by auto-
mobiles. Decker et al. (1977) measured hydrocarbons aloft in the St.
Louis urban plume that was tracked with a balloon. While their data
contain some anomalies (e.g., one sample was reported to contain 1016
ppbC propane and 112.6 ppm of CH^), they do provide an indication of the
hydrocarbon levels that are present in a plume under transport condi-
tions. Table 8 shows the concentration range of several compounds dur-
ing the day and night portions of the balloon's flight. It is evident
that the alkenes and propylene were present in small but significant
quantities during the overnight portion of the flight. During this
period, ozone remained essentially constant at about 125 ppb; hence,
there appears to have been very little ozone lost through O^-alkene
reactions. The general meteorological conditions consisted of a sub-
sidence inversion aloft and a strong radiative ground-based inversion,
conditions that favor long-range transport. Thus, not only is ozone
transported, but the hydrocarbons are also carried along, and will be
available to participate in the photochemical smog process the next day,
36
-------
Table 8
RANGE OF HYDROCARBON CONCENTRATIONS MEASURED
ABOARD DA VINCI II BALLOON*
Compound
Propane
Isobutane
n-Butane
Isopentane
Acetylene
Propylene
1-Butene
Trans-2-Butene
Concentration Range (ppbC)
Daytime
8 Jun 1976
1100 - 1900
2.9 - 3.9
1.0 - 3.1
6.4 - 11.8
1.8 - 39.8
0.9 - 6.1
1.5 - 8.8
Of - 3.2
Of - 2.0
Nighttime
8-9 Jun 1976
2000 - 0600
2.5 - 10.5
1.1 - 8.8
5.0 - 20.6
Of - 12.2
2.2 - 4.8
0.8 - 3.4
O1" - 5.4
Of - 2.0
Daytime
9 Jun 1976
0600 - 0800
3.8 - 40.0
3.7 - 5.1
10.6 - 65.2
5.4 - 8.3
3.3 - 5.4
1.5 - 13.7
0.5 - 34.6
Of - 8.4
Adapted from Decker et al. (1977). Four out of 22 samples
have been excluded due to anomalous concentrations.
Zero concentration denotes levels below limit of detection.
when all the species undergo fumigation upon the breakup of the
inversion.
Additional evidence of the presence of hydrocarbon precursors at a
rural site is given in Table 9, which lists average concentrations of
several compounds measured in Elkton, Missouri (Rasmussen et al., 1977).
It is apparent that acetylene, propylene, and 1-Butene are at the lower
end of the concentration range observed in the St. Louis plume (cf.
Table 8). The same is true for the butanes and isopentane. However,
propane levels are comparable to the St. Louis data. Obviously, Elkton
37
-------
Table 9
AVERAGE HYDROCARBON CONCENTRATIONS
IN ELKTON, MISSOURI
Compound
Propane
Isobutane
n-Butane
Isopentane
n-Pentane
Ethane
Acetylene
Propylene
1-Butene
Ethylene
Concentration (ppbC)
5
2
3
3
2
5
1
1
0
2
Adapted from Rasmussen et al. (1977)
is cleaner than an urban plume, but it still shows evidence of anthropo-
genic influence. This suggests that some hydrocarbon precursors of
ozone are as ubiquitous as N0_.
X
Two very useful studies by Ludwig et al. (1976; 1977) and one by
Meyer et al. (1976) investigated the relationship between rural ozone
and transported precursors. Ludwig examined high ozone concentrations
at several rural sites by analyzing the emission and meteorological his-
tory of the air masses arriving at the sites. In general, it was found
that NOX emissions were positively correlated with ozone. The overall
correlation was low (0.27 and 0.32 for two different time lags) but
highly statistically significant, and one site showed a correlation of
0.64. By contrast, hydrocarbon emissions showed no significant correla-
tion with ozone. The analysis showed that NO., emitted within 12 hours
38
-------
and between 24 and 36 hours prior to the arrival of the air mass at the
site had the most influence on ozone. These time lags respectively
reflect the impact of nearby emissions (0 to 12 hours) and long range
transport (24 to 36 hours). A half-life of NOV of at least one day is
X
suggested by these time intervals, which is consistent with the range of
estimates previously mentioned in Section II-D. From Ludwig's work, we
can infer that ozone levels at rural sites are probably enhanced by
addition of NOX, which implies that ozone production is N0x-limited.
This further suggests that high NMHC/NCL ratios exist at these rural
X
sites, and that the prevailing concentrations of NMHC are sufficient to
promote the formation of substantial amounts of ozone. While these
inferences cannot be considered definitive, they are consistent with the
findings of other studies.
Meyer et al. (1976) independently studied ozone levels at urban and
rural sites using a trajectory approach similar to Ludwig's. Their
results parallel Ludwig's, showing that NO emissions with a 6-hour time
Ji
lag are positively correlated with ozone at rural sites. The overall
correlation is low (0.23), but statistically significant. Moreover, it
is noteworthy that lagged NOX emissions were not significantly corre-
lated with ozone at the urban sites, whereas the unlagged emissions were
correlated. This latter result is precisely what would be expected in
view of the disparate distance scales that prevail between sources and
receptors in urban and rural areas. Meyer's analysis also showed that
hydrocarbon emissions with lags of 24 and 30 hours showed a positive
correlation with ozone at the rural sites of the same magnitude as for
NO . This is rather puzzling, as one would expect to see hydrocarbon-
A
ozone correlations with short time lags as well; it also contradicts
Ludwig's results. Thus, these results indicate that at the rural sites
NOX emissions from sources that are relatively close to the site are
more important than from sources that are far away, while the opposite
is indicated for hydrocarbons. This may be a consequence of the geogra-
phy of the sites analyzed, and further investigation is required to con-
firm this effect. For NOX, the positivity of the correlation indicates
that ozone levels will tend to be enhanced by the addition of NO , which
39
-------
is the same qualitative effect indicated by Ludwig's analysis. Thus,
Meyer's study supports the hypothesis that ozone formation in rural
areas may well be NO -limited.
2t
Various attempts have been made to model rural ozone formation.
Chameides (1978) used his model to examine the effect of the rate con-
stant of the reaction H02 + NO -> N0~ + HO, which has been recently
revised upward (Howard and Evenson, 1977). He concluded that with the
higher rate constant (approximately 1.2 x loVppm/min levels of NO
<•> x
below 1 ppb are sufficient to promote the formation of substantial
amounts of ozone. However, his work was limited to steady-state esti-
mates using CH^ as the sole source of organic radicals; hence we must
discount the quantitative aspects of this work and consider only the
qualitative effects. Nevertheless, it is generally agreed that the fas-
ter reaction HO, + NO -> N09 + HO enhances the ability of NO,, to gen-
£ fm X
erate ozone since in all photochemical smog mechanisms this reaction is
one of the key steps (another being R02 + NO -> RO + N02) that oxidize
NO to N02.
Graedel and Allara (1976) report that increasing NOX emissions
would result in a small decrease in peak ozone. This is contrary to the
hypothesis that ozone production in rural areas may be N0x-limited.
However, their model used the lower H02 + NO rate constant; hence, their
result may be subject to revision. Liu (1977) also reported a decrease
in ozone due to increased NO emissions, but he also used the outdated
rate constant for H02 + NO.
Recently, Cleveland and Graedel (1979) have examined the impact of
changing NO emissions on ozone downwind of urban areas in the northeast
United States. They conclude that increasing NO will initially depress
urban ozone levels, but is likely to enhance ozone concentrations
downwind. Conversely, reductions in urban NO emissions will increase
ozone within the source region, but will probably lead to lower ozone
levels downwind. Their results also indicate that, in the northeast,
controlling NO emissions is more effective than hydrocarbon controls as
40
-------
an abatement strategy for photochemical pollutants. These results
address directly the problem that concerns us, and imply that NO emis-
X
sions should not be allowed to increase in the northeast because nonur-
ban ozone pollution will become worse. However, the conclusions of
Cleveland and Graedel are based on extrapolations of model predictions
and require further work to test their validity. Nevertheless, their
results agree qualitatively with studies previously discussed, notably
Ludwig et al- (1976, 1977) and Meyer et al. (1976), which also suggest a
positive relationship between NO,, emissions and nonurban ozone.
• *t
Isaksen et al. (1978a) examined the problem of ozone production in
an urban plume using NMHC/NO^ ratios of 4 and 0.01 with a constant total
X
emission strength• These ratios are at the lower end of the range of
ratios usually found in urban areas, which range from 2 to 50 with a
median ratio of 10. (In fact, ratios of 4 and 0.01 are seldom observed
in urban areas.) As expected, the higher ratio yielded more efficient
ozone production per molecule of NO.,, and resulted in higher ozone con-
J£
centrations downwind. This implies that increased NO., emissions would
X
lower ozone downwind, provided that hydrocarbon emissions are not aug-
mented in the same ratio. From these results one might conclude that
nonurban ozone production is not NO -limited, but NMHC-limited instead.
X
While this may be true in the short term, the conditions can change
drastically in the long term. Thus, with high NO emissions and a low
X
initial NMHC/NOX ratio, the hydrocarbon consumption is very slow (this
is, of course, the cause of the low ozone). This results in a substan-
tial residue of unreacted hydrocarbons, accompanied by a gradual but
steady increase in the NMHC/NOV ratio. The net result is that the plume
X
can retain its ozone production potential for several hours, and the
eventual high NMHC/NOX ratio turns the situation around: from NMHC-
limited to N0x-limited conditions. The situation is thus analogous to
that described by Miller et al. (1978) in the power plant plume over
Lake Michigan. Thus, contrary to Isaksen's results, the possibility
remains that adding NOX can enhance ozone production downwind, which is
consistent with the results of Cleveland and Graedel (1979) previously
discussed.
41
-------
Ill SUMMARY OF EXPERT OPINION
A. Introduction
In assessing the state of knowledge about nonurban NOX/03 interac-
tions, SRI contacted leading researchers to obtain information about
recent developments in the field. Such a survey of expert opinion helps
not only to update the published record, but also identifies areas of
uncertainty where further research is required. The latter point is
helpful in establishing research priorities.
In this survey, a number of investigators representing government,
private, and academic institutions were contacted. Of those who indi-
cated interest in participating, eight subsequently submitted replies to
our questions; their names are listed elsewhere in this document. How-
ever, no individuals are quoted in the text. The responsibility for the
interpretation of the replies is ours.
The survey questionnaire contained 35 questions covering five gen-
eral categories: measurement methods, sources and sinks of NOX, chemical
aspects, control strategies, and identification of knowledge and data
gaps. Not all respondents answered all the questions. In what follows,
we summarize the replies received, and provide some additional comments
and explanatory remarks•
B. Summary of Results
1. Measurement Methods
In this category, the participants were asked to evaluate the qual-
ity of routinely collected NOX data for various concentration ranges.
The replies are shown below:
43
-------
Range of NO
X
Levels (ppb) Assessment of_ Data Quality
<1 Poor
1-5 Fair to poor
6-10 Good to fair
>10 Excellent to good,
This evaluation makes it apparent that the NO data in the higher con-
X
centration ranges is generally reliable. While this is encouraging for
urban data, it warns that rural and remote measurements below 6 ppb
should be treated with caution.
Respondents offered the following comments about NO data:
X
• Use of wet chemical methods should be discontinued for routine
monitoring.
• NO data are generally better than N02, but can be poor below 10
ppb because of difficulties with zero-setting techniques.
• NO2 data are subject to interference from other gases. This
problem is especially important for late afternoon measurements.
A3 a postscript, we note that wet chemical (i.e., colorimetric)
instruments are being gradually replaced by chemiluminescent detectors
for routine monitoring applications. Recent work at SRI compared simul-
taneous chemiluminescent and colorimetric N©2 measurements made at three
urban sites in California during 1975-1977 (Martinez and Nitz, 1979);
the comparison was performed for N02 levels that exceeded 200 ppb.
Colorimetric and chemiluminescent measurements were found to be linearly
correlated at two of the three sites, the correlation coefficients being
0.71 and 0.86. At these two sites it was also found that the chemi-
luminescent data tended to be greater than the corresponding
colorimetric concentrations. The third site showed no correlation
between the two instruments. For the concentration range considered,
44
-------
these results show that N02 data obtained using colorimetric and chemi-
luminescent techniques may not be strictly comparable; additional work
is required to test this conclusion for N02 levels below 200 ppb.
2. Sources and Sinks of NO
The questions in this category were aimed at:
• Obtaining estimates of the half-life of NOX
• Identifying NOX removal mechanisms
• Evaluating the importance of indirect sources of NOX;
• Obtaining a qualitative assessment of the relationship between
NOX sources and atmospheric concentrations.
The questions and replies are first given below, and our comments
follow.
• Estimates of. half-life o_f NOX
- Six to twenty-four hours in the boundary layer -
- One to two days in the free troposphere.
• Important removal mechanisms
- Heterogeneous removal through dry and wet deposition may be
the most important mechanism.
- Gas phase reactions HO + N02 -> HN03 and RC03 + N02 <-> PAN.
• Indirect Sources of NO,
- Anthropogenic NH-j emissions are considered to be trivial as an
indirect source of NOX through conversion of NH-j to NOX.
- All indirect sources (including NH^), are generally unimpor-
tant factors in contributing to the total tropospheric NOX
burden.
Assessment pf^ PAN as a reservoir for NOX. It is likely that PAN
can act as a reservoir for NOX. ~~
Assessment of the relationship between NOX emissions and atmos-
pheric levels;
- In suburban atmospheres, NOX emissions and ambient levels are
directly related. However, atmospheric levels tend to be
lower than expected from estimated emission rates. Local
(suburban) sources can predominate.
45
-------
- In rural atmospheres, NOX levels are weakly influenced by
urban sources. Local sources are most important determinants
of ambient concentrations of NO .
3t
- In remote atmospheres, NOX concentrations are generally unaf-
fected by man-made sources, because of the short half-life of
NOX. Infrequent high concentrations are due to transport from
anthropogenic sources.*
Regarding the half-life of NO , the range of estimates given above
X
is at the lower end of the 1- to 20-day estimate mentioned in Section
II-D. This is consistent with suggestions that residence time^estimates
should be revised downward in light of the discovery that background
levels are below 1 ppb (rather than several ppb) as previously believed
(cf. Section II-C and Ritter et al., 1978).
Residence time estimates are related to the removal mechanisms.
Thus, it is instructive to note that the half-life of NO due to gas
X
phase losses only (primarily OH attack) is estimated to be between seven
and fifteen hours. For dry deposition, velocities of about 0.3 cm/s
have been suggested. Under the appropriate circumstances, (viz., a
shallow mixing layer and moderate mixing within the layer), such a
velocity can result in a half-life as low as three hours, thereby
overwhelming gas phase losses. For comparison, the ozone half-life has
been estimated to be about 30 hours (Ripperton et al., 1976). However,
near the ground, ozone deposition velocities as high as 1 cm/s have been
estimated, (Harrison et al., 1978), which can lead to a half-life as
short as one hour. Thus, when the aggregate effect of all removal
processes is considered, it is probable the half-life of NOX within the
mixed layer is comparable to that of ozone.
In Section II-F we discussed the potential impact of indirect
sources as contributors to the overall NOX burden. In that discussion,
*While it may seem contradictory to state that "remote" areas are im-
pacted by anthropogenic sources, such a statement merely reflects the
fact that few, if any, areas in the United States are totally free from
air pollution.
46
-------
conversion of ammonia to NOV was estimated to constitute about 10 per-
X
cent of the total burden. The reply to our questionnaire could be
interpreted as being at variance with the literature, because 10 percent
is not necessarily a trivial contribution. However, as noted in Section
II-F, the NH-j -> NOX conversion process is poorly understood, and its
occurrence has been questioned. The replies to the questionnaire
apparently reflect the latter view.
The reversible reaction RC03 + N02 <-> PAN suggests the possibility
that PAN can act as a reservoir for N02 by (in effect) "storing" N02,
only to return it later upon decomposing. The reply to the question
indicates that this process is now considered to be likely to occur- If
this is correct, the transport range of NO would be affected since PAN
A
may be longer-lived than N0« under the proper conditions. However, the
atmospheric half-life of PAN depends strongly on temperature and on the
N02/N0 ratio. Table 10 shows estimates of the half-life of PAN obtained
by Hendry and Kenley (1978). Based on these data, it is conceivable
Table 10
ATMOSPHERIC HALF-LIFE OF PAN
(Hours)
Temperature
(°C)
12
22
32
NO /NO Ratio
0.1
4.5
0.9
0.2
L.O
5.5
1.1
0.24
10
17
3.5
0.8
100
140
28
6.2
Adapted from Hendry and Kenley (1978).
-------
that in an urban plume undergoing transport at night, the NO^/NO ratio
and the temperature may be such that PAN can have a half-life of several
hours, and thus could travel a reasonably long distance. The next day,
PAN would decompose rapidly as the temperature increased, thereby acting
as a source of N02 and RC03. The latter could participate in the oxida-
tion of NO (given a low N02/N0 ratio), which would eventually result in
enhanced ozone levels. Thus, PAN may be a pool not only of N02, but
also of RC02, and both of these compounds help to promote Oo formation.
While the participants in the survey believe that this process is likely
to occur, many uncertainties remain and further investigation is needed.
In particular, we need to estimate the concentration of PAN that might
survive after transport and dilution. Finally, it should be noted that
other organic peroxynitrates may undergo similar processes. One respon-
dent commented that organic peroxynitrates are potential reservoirs of
odd nitrogen, and that their formation and stability should be studied
further.
The replies to the question about the relationship between emis-
sions and ambient levels of NOX seem to reflect the view that transport
of NOX from urban areas is important only in suburban locations. This
view is compatible with the belief that the half-life of NOX is too
short to allow any significant transport to occur. On the surface, it
appears that this view rules out the hypothesis that local ozone syn-
thesis in rural areas is influenced by precursors transported from urban
regions. This is not necessarily the case, however, because (as we have
seen) the evidence suggests that low levels of NOV are sufficient to
li
produce substantial amounts of 0^ in rural areas (assuming the necessary
NMHC is available). Thus, even though dispersion may obscure or weaken
the link with urban emissions, the small amounts of urban-generated NOX
that reach the countryside could still play a significant role in local
ozone production. This argument is not meant to imply that local rural
sources of NOX are not important — they are — but rather that the
residue from urban emissions cannot be dismissed as a contributing fac-
tor in local ozone synthesis.
48
-------
3. Chemical Aspects
The questions posed in this category were intended to elicit judg-
ments about a few specific topics, rather than to range over a broad
panorama of chemical kinetics• The questions and answers are given
below and are followed by our comments.
• Photochemical Models for Nonurban Areas
- Reliability and verification: Photochemical models for nonur-
ban environments were judged to be essentially unreliable.
' Respondents also view the models as being basically
unverified.
- Critical Chemical Parameters: The following reactions were
suggested as being critical in the sense that model perfor-
mance would be significantly affected if their rate constants
were modified:
* Surface reactions
* R02 4- Oj
* R02 + NO
* H02 + NO
* H02 + N02
* HO + N02.
• Use of_ smog chambers for determining NX)X/J33 interactions in
nonurban areas; Smog chambers were jud"g~ed~~to be fairly adequate
for establishing the involvement of NOX in 0^ formation in
nonurban areas•
• Relationship between NO and nitrateg; NOX and nitrates are
related, but the relationship is probably nonlinear. However,
there is no relationship between coexisting concentrations.
HNOj and some organic nitrates may serve as pools for NOX, which
can be transported long distances; this would likely yield low
NO and N02 levels in remote areas.
• NOX reactions important in^ nonurban atmospheres; The following
reactions or processes were judged to be more important in a
nonurban environment than in the urban atmosphere:
- H02 + NO -> HO + N02
- Formation of HNOg.
• Role £f natural hydrocarbons in rural £3 formation; This ques-
tion yielded several answers, demonstrating a split opinion on
this topic. The answers are shown below. (The order in which
the replies are listed has no particular significance.)
49
-------
- Natural hydrocarbons are probably very important, but emission
data are needed for a complete assessment.
- If present, natural hydrocarbons should contribute to 0-j
formation.
- Some natural hydrocarbons, e.g., terpenes, consume 0.,, and the
net effect may be that 03 is destroyed.
- Natural hydrocarbon levels in ambient air are too low to play
a significant role in ozone formation.
• Effect of ozone intrusion on photochemistry in nonurban atmo-
spheres; The question asked for a judgment about the likelihood
that ozone intrusion from an external source (e.g., the strato-
sphere) might accelerate the photochemistry in nonurban atmo-
spheres. Two contradictory judgments emerged, some respondents
replying that it would be unlikely, and some that it would be
likely.
We included the questions about photochemical models partly because
several hypotheses advanced about nonurban ozone rest on results
obtained from models, and partly because some of the participating
scientists have developed and applied models and their judgment would be
especially valuable in assessing the reliability of the various models.
The response to the questions of reliability and verification agrees
with our earlier remarks in Section II-G. The implication is that con-
tinued research on model development and verification is necessary in
order to improve reliability. This is imperative because models can be
extremely useful in unraveling the complex interactions•
While photochemical models in their present state of development
were judged to be essentially unreliable for application to nonurban
areas, smog chambers were considered to be fairly adequate for a similar
use. These two responses appear to be inconsistent, since the chemical
component of the models is normally tested using smog chamber data. We
believe that there is no inconsistency, in that the judgment about the
models is probably based on the perception that the verification tests
of the models have not gone far enough.
Somewhat akin to the application of models is the use of smog
chambers to elucidate chemical effects: Both have their problems, but
must be used because of the complexity of the problem. The qualified
50
-------
judgment of adequacy bestowed on smog chambers for application to a
nonurban environment is encouraging, both for model development and for
the assessment of chemical effects based on smog chamber data.
Regarding the relationship between NOX and nitrates, the replies
reiterate a theme mentioned earlier about the potential of some nitrates
to serve as reservoirs for NOX. Given that the same suggestion was made
independently by several individuals, it appears that this possibility
merits further investigation. In the context of nitrates and other
products of NOX, the reader should consult the work reported by Spicer
(1976; 1977) and Spicer et al. (1976b; 1977).
The divergent replies to the question about the role played by
natural hydrocarbons in rural ozone production reflect the current
debate that surrounds this topic, a glimpse of which was seen in Section
II-H. It is apparent that the last word on this subject has not been
written.
The question about the effect of ozone intrusion was prompted by
suggestions in the literature (e.g., Hathorn and Walker, 1976) that such
additions of ozone may help trigger ozone episodes by speeding up the
photooxidation processes, hence the term "acceleration." If it happens
at all, such a phenomenon would be more likely in an urban setting where
unreacted alkenes exist. Assuming that the concentration of the
injected ozone is high enough, the 0^ + alkene reactions would yield
peroxy radicals that in turn oxidize NO, a process that could eventually
result in a fast ozone buildup. However, such favorable conditions are
unlikely to exist in nonurban areas, where alkene levels tend to be low.
On this basis, it seems reasonable to conclude that Oo intrusions in
nonurban locations would simply add to the total ozone load, instead of
promoting additional local synthesis.
51
-------
4. Control Strategies
This category contained questions intended to obtain judgments
about the effectiveness of using NOX controls as an ingredient of an
oxidant control strategy. In addition, we inquired about the appropri-
ateness of applying a popular methodology, the so-called isopleth
method, in a nonurban setting* Questions, replies, and comments follow.
• Estimate the likelihood that control of NO^ emissions will sub-
stantially reduce photochemical air pollution in urban and
nonurban locations! Although the replies were diverse,"the con- •
sensus was that NOX controls would be generally ineffective in
controlling ozone in urban regions, but could be moderately
effective in nonurban areas.
• Can the isopleth method for estimating £3 from N0_ and NMHC pre-
cursors be extended for use in nonurban settings? The response
was negative.
Regarding the first question, one participant also commented that
introduction of NOX into rural areas might well trigger 0-j problems
where there are none, and could aggravate existing problems. Thus, the
view seems to be that anthropogenic NOX emissions are at least partly
responsible for the elevated ozone levels that are observed in nonurban
areas. This is consistent with some of the findings of Section II-H.
(cf. Ludwig et al., 1976; 1977; Meyer et al., 1976; Cleveland and
Graedel, 1979).
5. Identification of Areas Requiring Further Research
The participants were asked to name subjects where, in their view,
important knowledge or data gaps exist with regard to the relationship
between NOX and 0^ in suburban, rural, or remote locations. They are
shown below grouped in an approximate order of importance. It should be
noted that the topics within a particular group should be considered
equally important, i.e., it is the order of the four groups that is of
some consequence, not that of individual topics.
52
-------
IMPORTANT KNOWLEDGE OR DATA GAPS
NOX sinks
Natural sources of NOV
X
Role of natural hydrocarbons
NOX measurements
NOX residence time in the atmosphere
Natural sources of Og
Atmospheric residence time of Oo
Basic NO chemistry
A
Rate constants
Og sinks
The order of the groups reflects rather faithfully those areas
which were described earlier as being generally uncertain (cf. Section
II). Thus, the first group includes natural hydrocarbons, whose rather
controversial status has already been mentioned. As regards priorities,
we would go further and aggregate the first six topics into one group,
and the last four into another. However, it should be recognized that
the differences in priorities are slight — all the topics need further
research. It must also be realized that the suggested priorities do not
necessarily coincide with regulatory objectives. Thus, the list of
topics requires additional screening from a regulatory standpoint.
53
-------
IV ANALYSIS OF DATA
FROM THE SULFATE REGIONAL EXPERIMENT (SURE)
A. Introduction
The SURE is an extensive air quality monitoring and analysis pro-
gram sponsored by the Electric Power Research Institute (EPRI) designed
to establish the relative importance of emissions from various source
categories in contributing to regional ambient sulfur dioxide (S02),
particulate sulfate, and NO (Mueller et al., 1977). To this end, a
X
regional network of monitoring stations was established to obtain a data
base of pollutant concentrations. The overall program is being con-
ducted for EPRI by the firm of Environmental Research and Technology
Inc. (ERT), whose cooperation in providing SRI with monitoring data is
gratefully acknowledged.
The monitoring network established for the SURE contains 54 sta-
tions covering a large area whose western boundary consists of all or
parts of Minnesota, Iowa, Missouri, Arkansas, and Louisiana, and extends
to the Eastern Seaboard. Nine of these stations, located in nonurban
regions in eight states, monitored NOX, NO, N02, and Oo, as well as
other parameters. More details about the geographical and temporal cov-
erage of the network and the experimental methods used will be given
below.
The data base obtained in the SURE is unique because it is the most
extensive set of observations collected in nonurban areas. By contrast,
the majority of the data cited earlier in Tables 2 through 4 were
obtained during special studies of short duration. Moreover, the SURE
data were measured exclusively using the latest monitoring methods and,
because the data are of recent vintage, portray better the contemporary
air quality conditions that exist in nonurban areas of the United
States. Consequently, this data resource merits analysis to supplement
the data of Tables 2 through 4.
55
-------
Before proceeding, it must be noted that the SURE data provided to
SRI have undergone only preliminary validation by ERT, and have been
released at this early date for purposes of scientific investigation
only. We recognize that the data are preliminary, and assume full
responsibility for any conclusions and interpretations that may be drawn
from the data. We have closely scrutinized the data in the course of
our analysis and have detected few inconsistencies; those data that
appear to be spurious have been brought to the attention of ERT and have
not been considered in the analysis. »
B. Description of SURE Data
1. Geographical and Temporal Coverage
The geographical coverage of the monitoring network is displayed in
Figure 3, which shows the locations of the nine sites. Table 11 lists
the site names, their UTM coordinates, the cities nearest to each site,
and any local sources that may affect pollutant levels at the site.
The continuous measurements run from August through December 1977.
(The monitoring program will run through 1978, but only the August
through December 1977 data were available for release when our study was
undertaken.) The data consist of hourly averages of N0x, NO, and Oj; NO2
was estimated from the NOX and NO data. The number of data points
available for each parameter for each site is given in Table 12, which
shows that the data capture rate (the percent of the maximum possible
number of observations for which data are available) for N0x ranges from
72 percent at Site 3 to 98 percent at Site 4. The data capture rate for
NO varies from a low of 59 percent at Site 3 to 98 percent at Site 4.
For 0-j, the range of the data capture rate is from 82 percent at Site 3
to 99 percent at Site 6. Thus, except for NO at Site 3, the data set is
quite complete.
56
-------
Table 11
LOCATION OF MONITORING STATIONS IN THE SURE
Station
No.
01
02
03
04
05
06
07
08
09
Name
Montague
Scranton
Indian River
Duncan Falls
Rockport
Giles County
Ft. Wayne
Research
Triangle
Park
Lewisburg
(Greenbrier
Airport)
UTM Coordinates (km)
X
702.88
410.64
476.16
424.30
494.80
508.90
639.05
695.54
558.72
Y
4,715.55
4,604.80
4,270.48
4,411.39
4,192.40
3,904.48
4,535.65
3,973.00
4,180.80
UTM
Zone
18
18
18
17
16
16
16
17
17
Nearest City
Amherst,
Massachusetts
Greenfield,
Massachusetts
Scranton,
Pennsylvania
Millsboro, Delaware
Zanesvllle, Ohio
Rockport, Indiana
Owensboro,
Kentucky
Columbia, Tennessee
Huntsville, Alabama
Pulaski, Tennessee
Ft. Wayne, Indiana
Durham, North
Carolina
Raleigh, North
Carolina
Lewisburg, West
Virginia
White Sulfur
Springs, West
Virginia
Miles1
15
5
25
1
8
0.5
8
24
42
10
15
7
10
5
2
Direction
N
SE
NW
E
SE
W
N
SSE
NE
NE
SW
SE
NW
E
SW
Local Sources
Power plant 1 ml SE
Power plant 0.5 mi N
Metal processing plant
3.2 mi NE
Power plants 8.5 mi S
and 13.5 mi WNW
Aluminum recycling plant
5 mi NE
Ouarry (sant) 1.5 mi NE
Heavy traffic
Adapted from Mueller et al. (1977).
Distance between city and site.
'Direction is from the city to the site.
t
-------
FIGURE 3 LOCATION OF SURE MONITORING STATIONS
58
-------
Table 12
NUMBER OF AVAILABLE OBSERVATIONS
FOR SURE STATIONS
AUGUST-DECEMBER 1977
Site No.
1
2
3
4
5
6
7
8
9
NO
X
3330
2817
2629
3616
3408
3506
3216
3477
3303
NO*
3349
2814
2168
3616
3405
2415
3195
3482
3303
"3*
3419
3410
3017
3441
3462
3632
3438
3495
3459
Each observation is an hourly
average. Maximum possible num-
ber of observations is 3672.
2. Instrumentation
Table 13 shows the instruments uded to measure the various com-
pounds. The sensitivity limit of the NOX and 03 instruments is 2 ppb,
based on manufacturer's specifications. A detailed discussion of the
experimental methodology used in the SURE is given in Mueller et al.
(1977).
C. Data Analysis
The analysis will first be concerned with establishing the charac-
ter of each site, (i.e., suburban, rural, or remote), as revealed by the
aerometric data. Although all the monitoring stations were carefully
59
-------
Table 13
INSTRUMENTATION USED AT THE SURE SITES
Parameter
NO, NO
X
°3
Instrument
Monitor Labs 8440
Monitor Labs 8410
Method
Chemi luminescent
Chemi lumi ne see nt
From Mueller et al. (1977), pp. 54 and 86.
located outside urban areas to maximize exposure to regional-scale
phenomena, the sites are susceptible to some degree of influence from
nearby sources, and this should be reflected in the data record. We
will also analyze temporal and spatial concentration patterns, and will
investigate relationships between the various pollutants.
A qualitative indication of the type of source affecting a site can
be obtained from the daily cycle of concentration variations. Figures 4
through 12 display average diurnal curves for NO, NO, NOo, and Oo for
Sites 1 through 9, respectively. Figures 5, 6, 8, 10, and 11 (respec-
tively, Sites 2, 3, 5, 7, and 8) show evidence of influence from mobile
source emissions. This influence is reflected in the plots of NOX and
NO, which have shapes typically associated with daily traffic patterns.
A sign of mobile source influence is the presence of a morning NOX peak,
which mainly results from increased NO. Such a pattern is particularly
strong in Figures 5, 6, and 11; it is weaker but still clear in Figures
8 and 10. The effect is especially pronounced at Site 8 (Figure 11),
which is located in Research Triangle Park, North Carolina. This area
contains a large U.S. Environmental Protection Agency laboratory and a
sizable complex of private companies, but is outside the urban core.
Hence, it reflects the influence of commuting patterns. Thus, Site 8
can be classified as suburban.
60
-------
30
20
oc
K-
o
a
10
T T
NO,
— £.--
I I
NO
*
I J —- 1 --- —I -- — t
50 -
40
z 30
O
5
cc
I-
01
a ~
Z 20
I I
10 _ ^
I
8
12
HOUR (EST)
16
20
23
FIGURE 4 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 1 IN MONTAGUE, MASSACHUSETTS.
DURING AUGUST-DECEMBER 1977
61
-------
30
.a
a
z
o
20 -
oc
z
LU
" 10
8
1 1 I
I I I I I I
MO
7~
NO
50
40
30
Ol
8
20
°3
y
^-
12
HOUR (EST)
16
20
23
FIGURE 5 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 2 IN SCRANTON, PENNSYLVANIA,
DURING AUGUST-DECEMBER 1977
62
-------
30
•a
_a
z 20
g
$
oc
LU
g 10
o
u
i i i i i i
_L 1 1 I
1 i
50
40
Z 30
O
t-
20
10
I I I T I I I I
/
\
\.
X
\
I i i i
J
I
J I
I I
8 12
HOUR (EST)
16
20
23
FIGURE 6 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 3 IN INDIAN RIVER, DELAWARE,
DURING AUGUST-DECEMBER 1977
63
-------
30
Z 2°
O
P
oc
z
01
o 10
8
i r
i i i
i i i r
NO.
NO
NOg
-t 1 <- -t
50 -
40 -
30
z
IU
20
10
I T I
i i i r r i i
,
v
\
\.
J I
I I I I I
8 12
HOUR (EST)
16
20 23
FIGURE 7 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 4 IN DUNCAN FALLS, OHIO,
DURING AUGUST-DECEMBER 1977
64
-------
30
t
z M
o
H
or
z- KV
IU
« 10
8
oc
h-
z
01
8
i i i i i i
50
40
10
/ ' \
/
30
03
\
\
I
0 4 8 12 16 20 23
., HOUR (EST)
FIGURE 8 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 5 IN ROCKPORT, INDIANA,
DURING AUGUST-DECEMBER 1977
65
-------
30
z
UJ
o
i i i i i i i r
a
|
° \ ^"O,
'<
EC
N02
V
NO
\
I I I I "I »- J 1 ^~* I I
50
40
Z 30
Z
LU
z 20
8
10 -
i i i i
/
\
/
I I I I I | I I 1 I
0 4 8 12 1C 20 23
HOUR (EST)
FIGURE 9 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 6 IN GILES COUNTY,
TENNESSEE, DURING AUGUST-DECEMBER 1977
66
-------
•a
z
LU
O
O
o
Z
O
£
-------
40 -
Z
o
o
Z
o
a
i i i i i i i
40
a
CL
a
Z 30
O
Ul
8
20
10
/
•
•
\
\
I
I
12
HOUR (EST)
16
20
23
FIGURE 11 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 8 IN RESEARCH TRIANGLE PARK,
NORTH CAROLINA, DURING AUGUST-DECEMBER 1977
68
-------
30
20
ai
O
I 10
u
i r
i i i r
NO,
NO
NO?^ >
4 J 1 1_J 1 1 1 1 ^. [ 1^,
50
40
Z 30
g
OC
H-
Z
cS
20
10
i i i i i i i i i r
X
X
/
\
\
\
\
\
j i
J I
J I
8
12
HOUR (EST)
16
20
23
FIGURE 12 AVERAGE DIURNAL VARIATION OF NITROGEN OXIDES
AND OZONE AT SURE SITE 9 IN LEWISBURG. WEST VIRGINIA,
DURING AUGUST-DECEMBER 1977
69
-------
At Sites 3 and 8 (Figures 6 and 11) the NOX peak occurs about 0700,
whereas it occurs later at Sites 2, 5-, and 7 (Figures 5, 8, and 10).
This suggests that Sites 3 and 8 are located close to the mobile
sources, but that Sites 2, 5, and 7 are somewhat removed from the com-
muter traffic, yet feel its impact. Thus, it seems appropriate to clas-
sify Sites 2, 3, 5, and 7 as suburban.
Sites 1, 4, and 9 (Figures 4, 7, and 12, respectively) show no dis-
tinctive diurnal variations in nitrogen oxides; certainly nothing that
-j^
resembles traffic-related concentrations. The essentially steady level
of NOX displayed at these sites suggests that, on the average, they are
not particularly influenced by any one source. Thus, we conclude that
regional-scale influences are the dominant factors at Sites 1, 4, and 9.
Accordingly, these sites can be considered to be rural- The NO levels
X
observed at these sites are such that none can be considered to be in
the remote category.
The diurnal variation of NO at Site 6 displays a very slight NO
A,
peak at 0700, the NO decaying rapidly thereafter- -This NO peak is
largely reflected in the NO concentration, and it is probably not indi-
3t
cative of significant traffic influence. On the other hand, the night-
time levels of NO are second only to Site 8, which suggests consider-
A
able anthropogenic influence. The most probable cause is transport from
urban areas, since there are no local sources nearby. Thus, Site 6 may
be classified as a rural site that is a receptor for transported pollut-
ants of urban origin.
Table 14 summarizes the data on nitrogen oxides and ozone at the
nine sites. The table shows that, as might be expected, maximum NO lev-
els are lower at rural sites than in the suburbs, except for Site 6.
This is not the case for N02» however, some suburban sites having lower
maxima than rural locations. The ozone maxima exceed 100 ppb at the
four rural sites, but only at one of the five suburban locales (viz.,
Site 8).
70
-------
Table 14
SUMMARY OF HOURLY CONCENTRATIONS OF NITROGEN OXIDES AND OZONE
MEASURED AT THE SURE SITES DURING AUGUST-DECEMBER 1977
Site
No.
1
2
3
4
5
6
7
8
9
Site
Type
Rural
Suburban
Suburban
Rural
Suburban
Rural
Suburban
Suburban
Rural
Maximum (ppb)
NO
X
109
173
151
94
233
112
101
263
49
NO
78
70
114
53
184
96
99
249
33
N02
73
64
48
43
59
55
35
145
28
°3
153
77
99
107
99
117
80
118
106
Mean (ppb)
NO
X
10
14
7
9
13
15
9
23
5
NO
3
3
3
1
3
5
3
10
1
N02
7
11
5
8
10
11
7
13
4
°3
21
23
30
29
25
27
20
25
35
Mean Daily Maximum (ppb)
NO
X
24
26
29
21
29
41
24
69
11
NO
11
8
17
4
11
24
14
49
3
N02
16
20
17
18
20
24
15
28
9
°3
44
35
49
49
46
52
39
50
54
fr
Minimum value of ail the species is zero at all sites.
-------
It can be seen from Table 14 that mean values of NO are equal or
lower at rural sites than at suburban locations. Once again, Site 6 is
seen to be impacted by high NO levels, as evidenced by the mean. How-
ever, these high levels occur mostly at night, as shown in Figure 9.
Mean N02 levels range from 4 to 13 ppb, and all the sites show compa-
rable values. Mean ozone levels follow the same pattern as N02 except
at Site 9, which shows both the lowest mean N02 and the highest mean
ozone.
«*
Because many concentrations are zero, the means are not necessarily
the best indicators of pollutant levels at a site. Thus, we use the
mean daily maximum (MDM) as a descriptor of the ensemble of worst case
conditions. Table 14 shows that for NO., the MDM differs only slightly
j£
with location for Sites 1 through 5 and 7, with Sites 6, 8, and 9
differing substantially from the other sites. Site 8 has the highest
MDM for NOX, NO, and N02, and the third highest for 03, whereas Site 9
shows the lowest MDM for the nitrogen compounds but the highest for
ozone. Site 6 follows 8 in oxides of nitrogen, and resembles both 8 and
9 with respect to ozone. Thus, Site 9 is the "cleanest" in terras of
NOX, but contains significant amounts of 03. Site 8 has the dubious
distinction of being heavily polluted on all counts, as is Site 6. How-
ever, these data for Site 6 can be misleading, since the NO maxima are
X
a nighttime phenomenon, whereas this is not the case at Sites 8 and 9.
It appears that Site 9 is a receptor of aged pollutant plumes, while at
Site 6 local ozone synthesis predominates, the precursors having been
transported or otherwise retained overnight. At Site 8, it is likely
that a significant fraction of the secondary pollutants is locally pro-
duced, thereby adding to the pollution transported from the surrounding
urban areas•
It should be noted that almost half of the oxides of nitrogen data
for Site 3 consist of zeros. This may be an indication of data prob-
lems, and Site 3 data should be treated with caution. The preponderance
of zeros will tend to lower the means of NOX, NO, and N02 shown in Table
14, which is why mean NOV is lower at Site 3 than at any other station
72
-------
except Site 9. No other site exhibits as large a fraction of zeros in
the NOX data as does Site 3.
It is of interest to examine the NO/NOX ratio at the various sites,
as it provides some insight regarding the presence of unreacted NO.
Thus, as was seen in Section II-C, one expects a high NO/NOX ratio in
urban and suburban areas, and lower ratios in rural and remote loca-
tions. Table 15 lists the ratio of mean NO to mean NOV for the various
Jv
sites.
Table 15
NO/NOX RATIOS FOR THE SURE SITES
AUGUST-DECEMBER 1977
Site No.
1
2
3
4
5
6
7
8
9
NO /NO
X
0.30
0.21
0.43
0.11
0.23
0.33
0.33
0.43
0.20
Table 15 shows that Sites 4 and 9, two rural sites, have the lowest
ratios, as expected. However, Sites 1 and 6, which are also rural, have
ratios that are closer to those exhibited by the suburban stations. On
this basis, it appears that Sites 1 and 6 may be transitional suburban-
rural locations. Sites 2, 5, and 7 also seem to be hybrid suburban-
rural locations, since their respective NO/NOX ratios are lower or equal
73
-------
to the ratios at Sites 1 and 6. However, as previously noted, the
hourly fluctuations of NO and NO at Sites 2, 5, and 7 indicate the
X
presence of local mobile sources (cf. Figures 5, 8, and 10), which was
the reason for classifying them as suburban instead of rural. Sites 3
and 8 show the highest ratios, further indication of the heavy impact of
mobile sources at these two locations.
Ozone concentrations greater than 80 ppb occurred infrequently or
not at all at the various sites. Table 16 shows the number p.f^ hours
when 0-j exceeded 80 and 100 ppb at each monitoring station. Exceedances
of 80 ppb range from zero at Sites 2 and 7 to a maximum of 81 hours at
Site 8. The 100 ppb threshold was surpassed at Sites 1, 4, 6, 8, and 9.
All the high 02 concentrations were recorded during August-October. One
concentration of 82 ppb was observed at 1500 on 16 November at Site 3.
It is apparent from Table 16 that Site 1 exhibits the largest
number of exceedances of 100 ppb. The other rural sites, Sites 4, 6,
and 9, also show several instances when ozone surpassed 100 ppb, but the
frequency is much lower than at Site 1. Thus Site 1 is subject to
influences that differ qualitatively from those affecting the other
rural stations.
Site 8 has the highest number of exceedances of 80 ppb, but rela-
tively few of those exceed 100 ppb. The probable cause of this
phenomenon is that the preponderance of NO at this site (the NO'/NOX
ratio is 0.43, cf. Table 15) tends to reduce the maximum level that
ozone can reach. Moreover, as will be seen shortly, exceedances of 100
ppb at Site 8 tend to occur late in the day, and thus are likely to be
associated with transport from surrounding areas.
Table 16 shows that the new ozone standard of 120 ppb was exceeded
only at Site 1. The number of exceedances of the standard recorded at
Site 1 was 21, with ozone levels ranging from 121 to 153 ppb. Below we
investigate these events in detail.
74
-------
Table 16
Ul
NUMBER OF HOURS WITH OZONE CONCENTRATIONS
EXCEEDING 80, 100, AND 120 ppb AT SURE SITES
AUGUST-DECEMBER 1977
Site
No.
1
2
3
4
5
6
7
8
9
Total
Hours
Observed
3419
3410
3017
3441
3462
3632
3438
3495
3459
Number
of Hours
0- > 80 ppb
60
0
29
52
17
63
0
80
23
Number
of Hours
0_ > 100 ppb
33
0
0
2
0
5
0
10
3
Number
of Hours
0_ > 120 ppb
21
0
0
0
0
0
0
0
0
Maximum Hourly 0
Concentration
(ppb)
153
77
99
107
99
117
80
118
106
-------
It is noteworthy that Sites 1, 4, 6, and 9, which are considered to
be rural, account for a sizable fraction of the exceedances of 80 ppb.
Nevertheless, in comparison to urban areas, these sites (indeed, all the
sites) are relatively clean with respect to ozone. Thus, although ozone
maxima reach levels comparable to those found in urban areas, the fre-
quency of occurrence of these high values is quite low. This suggests
that local ozone generation is probably not a main contributor to
elevated ozone concentrations. Instead, the low frequencies associated
with high ozone values may indicate that an infrequent event such as the
wind blowing from a specific direction is required to increase ozone
beyond 80 or 100 ppb.
Those days when the hourly ozone maximum exceeded 100 ppb will be
investigated below in a series of case studies. The case studies
include descriptions of prevailing meteorological conditions and trajec-
tory analyses. It would .have been useful to examine ozone levels
recorded in neighboring communities on these occasions. However, we
were unable to do so because of constraints of time and data
availability.
The maximum daily ozone exceeded 100 ppb on 16 days at four sites.
The exceedances are listed in Table 17 by site, together with the time
of occurrence of the peak. The table shows that Sites 1 and 8 accounted
for 11 of the 16 high-ozone days. Except as detailed below, all the
maxima are single-day events.
Site 1 appears to have experienced a three-day episode during 3-5
August, a maximum concentration of 97 ppb on 3 August (not shown in
Table 17) having preceded the peaks for 4 and 5 August listed in Table
17. The 3 August peak occurred at 1500 local time. Another three-day
episode may have taken place in the period 27-29 August at this site.
The 9 August maximum of 109 ppb at Site 8 was preceded by a peak of
83 ppb on 8 August, and was followed by maxima of 87 and 90 ppb on 10
and 11 August, respectively. This suggests a moderate four-day episode.
Similarly, at Site 9, the peak on 22 October was followed by maximum
76
-------
Table 17
OCCURRENCE OF MAXIMUM DAILY OZONE
EXCEEDING 100 ppb AT THE SURE SITES
DURING AUGUST-DECEMBER 1977
Site
No.
1
1
1
1
1
1
4
6
6
6
8
8
8
8
8
9
Date
4 Aug
5 Aug
27 Aug
28 Aug
29 Aug
1 Sep
2 Sep
2 Aug
22 Aug
23 Sep
9 Aug
26 Aug
31 Aug
10 Sep
23 Sep
22 Oct
Day
of Week
Thu
Fri
Sat
Sun
Mon
Thu
Fri
Tue
Mon
Fri
Tue
Fri
Wed
Sat
Fri
Sat
0- Concentration
(ppb)
139
142
147
153
140
102
107
115
101
117
109
106
118
102
115
106
Time of Occurrence
(Local Time)
1700
1300
1800
2000
1300
1700
1300
1900
1300
1500
1700
1600
1500
1600
1700
1600
levels of 88 and 87 ppb on 23-24 October, which could be evidence of a
three-day episode. Likewise, the peak of 2 September at Site 4 was
bracketed by a maximum of 85 ppb on the previous day, and peaks of 74,
91, and 87 ppb on the three days immediately following.
The ozone peak of 115 ppb observed at Site 6 on 2 August was brack'
eted by maxima of 86 and 80 ppb on 1 and 3 August, respectively, which
suggests a mild three-day episode. A four-day interval of high ozone
levels occurred during 21-24 August. In this period, the 101 ppb max-
imum of 22 August was preceded by a peak of 88 ppb on 21 August, and
77
-------
followed by two consecutive days with peaks of 94 and 95 ppb,
respectively.
Note that all the rural sites (viz., 1, 4, 6, and 9) appear in
Table 17. The relatively late time of occurrence of most of the ozone
peaks at these sites suggests that, on these days at least, transport is
the primary contributor to oxidant pollution at the rural locations.
The influence of transport is also evident at Site 8. Thus, the high-
ozone events may yield some clues about the effect of transport on oxi-
dant pollution in outlying areas. Case studies of these high^ozone days
are provided below.
1. Case Study !_: High-Ozone Events
at Montague. Massachusetts (Site _1)
Figures 13 through 19 display pollutant histories for the high-
ozone days at Site 1. Figures 13 through 15 depict the three-day
episode that occurred during 3-5 August, and Figures 16 through 18 por-
tray a similar incident occurring on 27-29 August. Figure 19 illus-
trates the single-day event of 1 September.
Analysis of surface weather data revealed that on 3 August a low
pressure front had recently passed over New England, and that moderate
high pressure prevailed in the Northeast. The high-pressure system
remained over the area on 4 and 5 August, being replaced by low pressure
and precipitation over New York, Massachusetts, and other parts of New
England. Thus, the classic meteorological elements of an air pollution
episode prevailed during 3-5 August 1977. A similar situation occurred
on 27-29 August, except that the high-pressure system was already well
established by 26 August, and spread from North Carolina to Maine. No
significant precipitation was reported during this period, but a cold
front that spread rain over a large area was approaching on 29 August.
It passed over the area on 30 August, ending the episode. A high-
pressure system was reestablished by 31 August, and remained over the
area through 1 September.
78
-------
0.16
10 15
LOCAL TIME
20
FIGURE 13 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND O3
AT SITE 1 ON WEDNESDAY, 3 AUGUST 1977
C.'C
10 15
LOCAL TIME
20
FIGURE 14 HOURLY CONCENTRATION VARIATIONS OF NO, NO2/ AND 03
AT SITE 1 ON THURSDAY, 4 AUGUST 1977
79
-------
OJ6
10 15
LOCAL TIME
20
FIGURE 15 HOURLY CONCENTRATION VARIATIONS OF NO. N02, AND 03
AT SITE 1 ON FRIDAY, 5 AUGUST 1977
10 IS
LOCAL TIME
20
FIGURE 16 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 1 ON SUNDAY. 28 AUGUST 1977
80
-------
5 10 15
LOCAL TIME
FIGURE 17 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 1 ON SUNDAY, 28 AUGUST 1977
0.16
10 IS
LOCAL TIME
20
FIGURE 18 HOURLY CONCENTRATION VARIATIONS OF NO. N02 AND 03
AT SITE 1 ON MONDAY, 29 AUGUST 1977
81
-------
10 15
LOCAL TIME
20
FIGURE 19 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 1 ON THURSDAY, 1 SEPTEMBER 1977
A striking feature of Figures 13 and 16 through 18 is the presence
of a double peak in the ozone curve. This phenomenon is often observed
at locations impacted by pollutants transported from other areas. The
first ozone peak is generally associated with local photochemical
activity and the second, which occurs in the late afternoon or in the
evening, is attributed to transport. The figures show that the two
ozone peaks are separated by four to seven hours. In Figure 13, the
first peak has a magnitude of 97 ppb and occurred at 1400; the second
occurs at 1800 and its magnitude is 70 ppb. The second peak is higher
than the first in Figures 16 and 17, but not in Figure 18. Six hours
separate the two peaks on 27 August (Figure 16), increasing to seven
hours on 29 August (Figure 18). The difference is only four hours on 3
and 28 August (Figures 13 and 17). In all four cases, the second ozone
82
-------
peak Is accompanied by slight but noticeable increases in NC^- Levels
of NO at the time of the second peak on all four days are less than 1
ppb, which is below the minimum level of detectability of the
instrument.
The ozone buildup during the episode of 3 to 5 August is shown in
Figures 13 through 15. Ozone concentrations remain rather high until
about 2000 on 3 August, and N02 levels increase during the day and early
evening, peaking at 2000. However, there is little or no ozone carry-
over into the next day. Ozone begins to accumulate very early in the
morning of 4 August, the accumulation being preceded by a buildup of
N02» which is almost certainly due to transport. N02 peaks at 0500, and
decays slowly but steadily on 4 August until about 1000, remaining
essentially constant thereafter. Meanwhile, ozone reaches a late peak
of 139 ppb at 1700. Afterwards, ozone remains high until about 0200 on
5 August, when it decays rapidly. The accumulation of ozone begins
again after 0500 on 5 August, reaching a peak at 1300, decaying quickly
thereafter. NO levels are very low throughout the whole episode.
Figures 20 through 22 show estimated trajectories arriving at Site
1 during 3 to 5 August 1977. Each figure contains at most four individ-
ual trajectories which arrive at 0200, 0800, 1400, and 2000 (local
time). The trajectories were computed using a procedure devised by
Heffter and Taylor (1975). The nodes of each trajectory are separated
by six hours.
The trajectories for 3 August are depicted in Figure 20. The air
arriving at 0200 and 0800 originated in southern Canada and has passed
over areas of precipitation on its way to Site 1. Consequently, it is
probably relatively clean. By 1400, the path has shifted, the air pass-
ing over industrialized areas of Pennsylvania and New York before
arrival, essentially the same track being followed by the air arriving
at 2000. It is interesting that these two trajectories coincide respec-
tively with the ozone maximum at 1400 and a peak in N02 at 2000. Since
the 1400 and 2000 trajectories coincide, it is probable that between
83
-------
FIGURE 20 TRAJECTORIES ARRIVING AT SITE 1, MONTAGUE, MASSACHUSETTS,
ON 3 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
FIGURE 21 TRAJECTORIES ARRIVING AT SITE 1, MONTAGUE, MASSACHUSETTS,
ON 4 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
8A
-------
FIGURE 22 TRAJECTORIES ARRIVING AT SITE 1, MONTAGUE, MASSACHUSETTS,
ON 5 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
1400 and 2000 the air continued to pass over the same pollution sources,
which may be responsible for the ozone Increase between 1600 and 1800.
The pattern of air movement seen in the afternoon and evening of 3
August is essentially repeated on 4 and 5 August, as shown in Figures 21
and 22, respectively• The air flows from a southwesterly direction into
the site, passing over southeastern Pennsylvania and the southern tip of
New York along the way. These air masses originated entirely within a
high-pressure system with its attendant reduction in the ability to
disperse pollutants. Thus, it is highly probable that on 4 August the
air masses could accumulate high pollutant levels, which were trans-
ported to the site. This process continues through the night of 4
August, and is likely responsible for the carryover of Oj and N02 into 5
August. From the larger spacing between the nodes of the trajectories
85
-------
for 5 August (cf. Figure 22), it can be inferred that the wind speed of
the 1400 and 2000 trajectories has increased, which may explain the
afternoon drop in ozone concentration on 5 August (cf. Figure 15).
The above discussion suggests that it is very likely that transport
is responsible for the ozone levels observed on 3 and 4 August.
Although the ozone maximum occurred at 1300 on 5 August, it does not
appear likely that it is due to local chemical activity. The reason is
that NO and N02 change very little throughout the day, which tends to
discount the possibility that local precursors are responsible for the
ozone. In fact, the prevailing very low level of NO and the concomitant
high N02/NO ratio suggest that the ozone was formed on the way to the
site. Thus, it is probable that transport is responsible for the ozone
peak. A potential contributing factor to the total ozone burden at Site
1 on 5 August is ozone fumigation that accompanies the breakup of a noc-
turnal radiation inversion. Analysis of three-hour surface weather maps
indicated that a ground-based inversion was present in the early morning
of 5 August. This is consistent with the fast decay of ozone levels
that occurs between midnight and 0500, since such a rapid change is
associated with gas-phase and surface reactions that deplete the ozone
present in a shallow mixed layer. The weather data also show that the
inversion has lifted by about 1000, which suggests that the ozone
increase up to that time may result from downward transport of ozone
stored aloft. (Note.that the ozone concentration at 1000 matches that
previously measured at midnight.) In view of the very light to calm wind
conditions that prevail in the area, the level of ozone present at mid-
night can be interpreted as an indication of the concentration of ozone
stored aloft. If so, this concentration will be restored upon lifting
of the inversion, which is precisely what happens by 1000. Thus, down-
ward ozone transport may have contributed up to 60 ppb during the early
stages of the ozone buildup.
Data on pollutant levels are given in Table 18 for the high ozone
days at Site 1. The table and the plots show that the level of NO
86
-------
Table 18
DATA SUMMARY FOR HIGH-OZONE DAYS
AT SURE SITE I IN 1977
Date
3 Aug
4 Aug
5 Aug
27 Aug
28 Aug
29 Aug
1 Sep
Mean 0-
(ppb)J
49
77
57
59
73
64
44
Mean NO,
(ppb)
6
8
8
8
9
9
7
Mean NO
(ppb)
<1
<1
<1
<1
<1
<1
<1
Maximum 0«
(ppb)
97
139
142
147
153
140
102
Time of
Occurrence
of Maximum
(Local Time)
1400
1700
1300
1800
2000
1300
1700
for the high ozone days is very low. Comparing the mean NO level from
Table 18 with the five-month mean of 3 ppb for Site 1 (cf. Table 14)
shows that although this site usually experiences NO concentrations
greater than 1 ppb, on high-ozone days essentially all the NO has been
converted to NO2 thereby enhancing net ozone production. Table 18 By
contrast, mean N02 levels on all high ozone days range from 6 to 9 ppb,
and are comparable to the five-month mean of 7 ppb. Thus, NO2 levels
are only slightly affected on these high-ozone days, which is further
indication that high concentrations of ozone are probably due to tran-
sport, rather than to local influences.
The episode of 27-29 August exhibits the double ozone-peak
phenomenon on all three days. The second peak undoubtedly results from
the arrival of an aged air mass loaded with pollutants. As Figures 16
through 18 show, these late ozone peaks are accompanied by increases in
N02 but not in NO, which would be expected in an air mass that has con-
verted essentially all the NO to N02» The occurrence of the late ozone
maximum on all three days is in keeping with the fact that a high
87
-------
pressure system was stationed over the eastern United States on these
three days.
Figure 16 shows that on 27 August Site 1 experienced NO levels of
the order of 15 ppb from 0600 to 0800. This indicates the arrival of
air loaded with emissions from the morning traffic. HoweveT, the NO
disappears quickly, being undetectable by 1000. On 28 August (cf. Fig-
ure 17) ozone levels remain high well into the evening and into the
morning of 29 August, which was the same phenomenon observed during 4-5
August (Figures 14 and 15). Ozone decays rapidly from 0100 to 0400 on
29 August, building up again after 0600. On all three days, ozone con-
centration always exceeded 80 ppb between 1100 and 2100, indicating a
persistent pollution condition; such was not the case in the episode of
3-5 August.
Trajectories arriving at Site 1 during 27-29 August are displayed
in Figures 23-25. The pattern is remarkably similar to that seen ear-
lier in Figures 20-22. Thus, high ozone levels occurring in the late
afternoon and early evening are associated with southwesterly flows into
the site. Moreover, the lack of ozone carryover into the early morning
of 27 August corresponds to trajectories originating in southern Canada
in regions where emissions may be lower, and to less severe stagnation
conditions such as prevailed on this day. However, on 28 and 29 August
trajectories originate in the southeastern United States and Figures 17
and 18 show evidence of overnight ozone carryover from the 28th into the
29th. As was seen earlier during the events of 4-5 August, the carry-
over ozone is quickly depleted between midnight and 0400, followed by a
rapid ozone buildup that peaks at 1300. As previously suggested, this
could indicate fumigation of transported ozone "stored" aloft in a ther-
mally stratified atmosphere.
The high-ozone event on 1 September (Figure 19) shows a very late
ozone buildup that begins at 1200 and peaks at 1700. One implication of
this condition is that the NO/NC^/Og null reaction is dominant until
about 1000, resulting in no net ozone production; it also suggests that
88
-------
FIGURE 23 TRAJECTORIES ARRIVING AT SITE 1, MONTAGUE, MASSACHUSETTS.
ON 27 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
FIGURE 24 TRAJECTORIES ARRIVING AT SITE 1, MONTAGUE, MASSACHUSETTS,
ON 28 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
89
-------
FIGURE 25 TRAJECTORIES ARRIVING AT SITE 1, MONTAGUE. MASSACHUSETTS,
ON 29 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
the sky may have been overcast since ozone chemistry is so tightly cou-
pled to sunlight intensity. The late peak is again indicative of ozone
that has been transported to the site. It is interesting, however, that
even after sunset the ozone tends to linger and actually increases late
at night. This is consistent with the fact that the incoming air has
very little NO to scavenge ozone.
Air masses arriving at Site 1 on 1 September are tracked in Figure
26, which shows the path of the air arriving at 0200 and 0800. (It was
not possible to compute the other two trajectories because of data prob-
lems.) The trajectories are quite different from those seen earlier.
This time the air takes a circuitous route to the site, implying that
stagnation conditions and low wind speeds prevail. The latter may also
be inferred from the node spacing in Figure 26. Part of the path of
90
-------
FIGURE 26 TRAJECTORIES ARRIVING AT SITE 1. MONTAGUE. MASSACHUSETTS,
ON 1 SEPTEMBER 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
each trajectory is over the ocean, which may cause ozone levels to
increase in the absence of NO sources (Graedel and Farrow, 1977). The
convoluted path of the trajectories also suggests that the same air is
being circulated about, which could account for the steady ozone level
that prevails from midnight to 1100 (cf. Figure 19). If this meandering
pattern were to continue throughout the day, it could also explain the
relatively high nighttime ozone levels; however, we cannot confirm this.
In closing, it seems likely that most, if not all, the ozone peaks
observed at Site 1 are attributable to transport. In addition to the
lateness of some 0-j maxima, the evidence suggests the presence of a
thermally stratified atmosphere, possibly leading to fumigation and an
early afternoon peak the next day. The single-day episode of 1 Sep-
tember offers no evidence of thermal layering, but of a stagnant
91
-------
atmosphere where the air is being recirculated. This phenomenon is con-
sistent with a transport explanation of the observations at Site 1.
While ozone transport seems to be the best explanation of these
observations, this does not preclude the possibility of ozone enhance-
ment by increased levels of NOX. In all cases, NO levels are very low,
but N02 shows slight fluctuations about 10 ppb, regardless of the ozone
concentration. Is it possible, therefore, that increasing NO., would
X
lead to higher NC^, and hence higher ozone? What would happen if the
#•
"steady" NC^ concentration were 20 ppb instead of 10 ppb? We shall keep
these questions in mind as we review the data for the other sites.
2. Case Study II; High-Ozone Events
at Duncan Falls. Ohio (Site ^)
Ozone reached a maximum of 107 ppb at 1300 and 1400 on 2 September.
H.
This event was preceded by a maximum of 85 ppb on 1 September. A high-
pressure system prevailed over the eastern United States, extending from
North Carolina to Maine. The site appears to be located in what may be
termed the back side of this high-pressure system. On 1 September a
front was stationed west of Duncan Falls, with precipitation being
reported in northern Illinois and western Michigan. This front moved
closer to the site on 2 September, causing rain in northern Ohio. How-
ever, high-pressure conditions prevailed at the site on this day. The
front passed over the site on the afternoon and evening of 2 September,
and high-pressure conditions were reestablished by the morning of
3 September-
Figures 27 and 28 show the hourly variations in Oo, NO, and N02 at
Site 4 on 1 and 2 September, respectively. Both days are characterized
by ozone carryover during the night. Although nighttime ozone is low
(about 30 ppb), the concentrations did not vanish, suggesting a rela-
tively deep mixing layer and little or no chemical scavenging. NO and
N0£ were very low on these two days, with very little evidence of N02
carryover. Figure 27 displays an increase in 0-j between 1800 and 2000,
92
-------
OJ6
10 15
LOCAL TIME
FIGURE 27 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 4 ON THURSDAY, 1 SEPTEMBER 1977
10 15
LOCAL TIME
20
FIGURE 28 HOURLY CONCENTRATION VARIATIONS OF NO, NO2. AND 03
AT SITE 4 ON FRIDAY, 2 SEPTEMBER 1977
93
-------
which is undoubtedly due to transport. On 2 September, the evidence
also suggests transport effects, even though 0.» peaks early. Transport
may be inferred, given the rural location and character of the site, but
equally important is the fact that NO and NO 3 hardly change during the
day, which indicates that the oxidation of NO to N02 is essentially com-
plete in the incoming air mass.
Four trajectories arriving at Site 4 on 2 September are depicted in
Figure 29. The air masses follow a semicircular path into the^site,
arriving from the southwest. Some of the trajectories (particularly
those arriving between 0800 and 2000) probably passed over Cincinnati,
which could account for the high ozone concentrations at Site 4.
FIGURE 29 TRAJECTORIES ARRIVING AT SITE 4, DUNCAN FALLS, OHIO,
ON 2 SEPTEMBER 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
94
-------
An interesting feature of the data in Figures 27 and 28 is the very
low level of NOX that prevailed on these days. Once again we see sub-
stantial increases in Og, but little change in NOX- This would be
expected to occur in a spent air mass. It is noteworthy that not only
is there a low frequency of occurrence of high ozone levels, but also
that the high concentration was below 120 ppb. Thus, while the "spent"
air mass may be capable of producing ozone in excess of 80 ppb as sug-
gested by smog chamber results (cf. Section II-H), it may be unable to
yield levels above 120 ppb.
3. Case Study III; High-Ozone Days
at Giles County, Tennessee (Site Ji)
Ozone levels greater than 100 ppb were observed on 2 and 22 August
and 23 September. A high-pressure system prevailed over the area on the
first two days. However, a front was approaching on the 22nd. Mild
weather was the norm on 23 September, a weak warm front being present
over Indiana and Ohio.
Figures 30 to 32 show the daily variations of 0-j and NOX on the
three days of interest. We have shown NOX instead of NO and N02 as in
other figures because data for NO are missing. The trajectories that
impact the site on these three days are depicted in Figures 33 through
35.
Transport influence is evident in Figures 30 and 32. On both occa-
sions, the ozone peak occurs somewhat late, at 1900 and 1500 respec-
tively, and its advent is sudden rather than gradual. Local photochem-
istry is apparent during the early morning buildup. These two days also
have relatively high ozone levels late at night, reaching a value of
about 50 ppb•
Trajectories arriving at Site 6 on 2 August and 23 September are
displayed in Figures 33 and 35, respectively. The figures show that the
early paths of both sets of trajectories have a similar north-south
95
-------
10 IS
LOCAL TIME
20
25
FIGURE 30 HOURLY CONCENTRATION VARIATIONS OF 03 AND NOX
AT SITE 6 ON TUESDAY, 1 AUGUST 1977
10 is
LOCAL TIME
FIGURE 31 HOURLY CONCENTRATION VARIATIONS OF 03 AND NOX
AT SITE 6 ON MONDAY, 22 AUGUST 1977
96
-------
10 IS
LOCAL TIME
23
FIGURE 32 HOURLY CONCENTRATION VARIATIONS OF 03 AND NOX
AT SITE 6 ON FRIDAY, 23 SEPTEMBER 1977
FIGURE 33 TRAJECTORIES ARRIVING AT SITE Q. GILES COUNTY, TENNESSEE,
ON 2 AUGUST 1977
The number shown at one extreme of a trajectory is the local time of arrival at the site.
97
-------
FIGURE 34 TRAJECTORIES ARRIVING AT SITE 6, GILES COUNTY. TENNESSEE.
ON 22 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
FIGURE 35 TRAJECTORIES ARRIVING AT SITE 6. GILES COUNTY, TENNESSEE,
ON 23 SEPTEMBER 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
98
-------
orientation. However, the trajectories for 2 August (Figure 33) are
much longer than for 23 September (Figure 35), which indicates that
meteorological conditions were less stagnant on 2 August than on 23
September. The convoluted path of the trajectories shown in Figure 35
is further indication of the presence of stagnant, low-wind conditions
on 23 September.
Examination of the path of the trajectories suggests that the ozone
peak that occurred at 2000 on 2 August (Figure 30) may be due to precur-
sors that originated in the Nashville area, which is about 100 km
northeast of Site 6. On 23 September, the ozone peak may be associated
with an urban plume from Huntsville, Alabama, located about 40 km south
of the site. It appears, therefore, that short-range transport is the
primary influence at work on these two days.
The ozone maximum observed on 22 August is probably the result of
local photochemistry. Figure 31 shows high NOX in the morning and its
rapid decay is accompanied by a buildup of,ozone that peaks at 1300; NOX
remains low and essentially constant between 1300 and 2000, when its
concentration increases. This day displays a sustained high ozone con-
centration that lasts from 1200 to 1800. In this interval, ozone con-
centrations do not differ much from the peak of 101 ppb. While the max-
imum is tied to local chemistry, the broadening of the ozone curve is
most likely induced by transport.
Figure 34 shows that the trajectories on 22 August differ radically
from those for 2 August and 23 September. On 22 August the wind flow is
consistently from the west and southwest, and the trajectories are
clustered over northern Mississippi and southwestern Tennessee. The
convoluted track of the trajectories is symptomatic of low wind speeds
and relatively stagnant conditions. In contrast to the other two days,
22 August shows no sudden increase in the ozone peak, but displays
instead a broadening of the ozone maximum lasting several hours, and the
peak ozone level is lower than on the other two days (cf. Figure 31).
This discrepancy in pollutant history is undoubtedly a manifestation of
99
-------
the different wind flow patterns present on 22 August, but the origin of
the transported ozone is unclear - No large urban areas appear to be on
the path of the trajectories on this day, which may account for the
lower ozone level and the absence of a sudden ozone increase.
Carryover ozone from the previous night is not significant on any
of the three days. The same is true of NO on 2 August (cf. Figure 30).
X
As the figures show, ozone builds up more rapidly when high levels of
N0x are present in the morning (cf. Figures 31 and 32) than when morning
N0x is low (cf. Figure 30). This implies that local photochemistry is
most important during the early stages of ozone accumulation, but that
its influence wanes as the day wears on. Consequently, the ozone max-
imum may or may not be due to local effects. At this site, it seems
that local ozone production cannot result in very high ozone levels, but
that transported ozone combined with local ozone can approach the
120-ppb ozone standard.
4. Case Study IV; High-Ozone Days
ait Research Triangle Park, North Carolina (Site j})
Site 8 presents an interesting contrast to the monitoring stations
previously considered: It is in a suburban location heavily impacted by
traffic, and thus has the highest NOX levels. Since the site is sur-
rounded by three medium-sized cities, it is also a receptor for urban
pollution.
Ozone exceeded 100 ppb on five days. One maximum occurred at 1500,
and the others at 1600 and 1700. General weather conditions on the
high-ozone days were variable. A high-pressure system was present on 9
and 26 August. High pressure also prevailed on 30 August, which pre-
ceded the high-ozone event of the 31st. A low-pressure center was sta-
tioned off the coast of Virginia on 10 September, low pressure and pre-
cipitation having prevailed in North Carolina for several days previ-
ously. On 23 September, a weak warm front was observed across the
midwestera states, while stagnant to light wind conditions prevailed
100
-------
throughout the southeast* Thus, overall weather patterns were quite
unlike those that accompanied the elevated ozone levels observed at Site
1. This partly accounts for the fact that four of the five cases are
single-day episodes, the lone exception being the event of 9 August, and
that ozone concentrations are not as high as were seen in Site 1.
Figures 36 and 37 show the pollutant variations on 8 and 9 August.
The high ozone level of 9 August is characterized by considerable carry-
over from the previous day. The ozone maximum occurred late on both
days, suggesting transport as the cause. NO and NC^ are more abundant
than at the other sites. In view of the abundance of precursors, it is
probable that the early ozone peak is locally produced, rather than
resulting from transport. The circulation patterns for 9 August are
illustrated in Figure 38, the trajectories for 8 August being almost
identical to these. As in other cases, the tracks are semicircular with
a clockwise orientation typical of anticyclonic circulation. Because
these trajectories do not appear to traverse any heavily polluted areas,
it is likely that the transported pollutants originate in the Durham-
Chapel Hill area, which is several miles west of Site 8.
Conditions on 26 August are depicted in Figure 39. As before,
there is substantial carryover from the previous night. However NO
X
levels are low during the day, which indicates the arrival of an aged
mass of polluted air* As seen in Figure 40, the trajectories are again
typical of anticylonic circulation, with the air passing over the
heavily congested Washington D.C. area. Wind speeds near the site are
relatively low, and the southeasterly direction points to Raleigh as the
source of pollutants.
On 31 August (Figure 41) a typical NO traffic peak is evident in
the morning. The NO decays rapidly, but with no attendant increase in
N02; in fact, N02 decreases sightly in spite of the rapid disappearance
of NO, thereby suggesting that, in addition to chemical reaction,
101
-------
10 15 20
LOCAL TIME
FIGURE 36 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 8 ON MONDAY, 8 AUGUST 1977
E
a
z
o
0.16
0.14
0.12
0.10
< 0.08
cc
z
O
u
0.06
0.04
0.02
0
10 15
LOCAL TIME
20
FIGURE 37 HOURLY CONCENTRATION VARIATIONS OF NO, N02. AND 03
AT SITE 8 ON TUESDAY, 9 AUGUST 1977
102
-------
FIGURE 38 TRAJECTORIES ARRIVING AT SITE 8, RESEARCH TRIANGLE PARK,
NORTH CAROLINA, ON 9 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
5 10 (5
LOCAL TIME
FIGURE 39 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 8 ON FRIDAY, 26 AUGUST 1977
103
-------
r
\
L. -'
\r
<
FIGURE 40 TRAJECTORIES ARRIVING AT SITE 8, RESEARCH TRIANGLE PARK,
NORTH CAROLINA. ON 26 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
OJ6
O.W
0.12
10 IS
LOCAL TIME
FIGURE 41 HOURLY CONCENTRATION VARIATIONS OF NO. N02, AND 03
AT SITE 8 ON WEDNESDAY. 31 AUGUST 1977
104
-------
dispersion is causing some of the NO and NO2 decay. The jaggedness in
the ozone curve may result from fluctuations in radiation intensity.
The sudden appearance of the ozone peak and the relatively late time of
its occurrence suggest that transport plays a significant role.
Trajectories arriving at Site 8 on 31 August are depicted in Figure
42. It is evident that between 0200 and 1400 prevailing winds are from
the west. Hence, pollutants originating in the vicinity of Durham are
probably responsible for the ozone peak that occurs at 1500. The wind
has shifted to an easterly direction by 2000, coinciding with an
increase in NO2 that begins at 1900. Since Raleigh is located east of
Site 8, it is likely that the increased N02 is associated with an urban
plume from this city.
Figure 43 illustrates the pollutant history of 10 September.
Recall that weather conditions on this day included a low-pressure
center off Virginia. Cyclonic conditions are in evidence in Figure 44,
which shows air trajectories for this day. Clearly, the trajectories
are oriented counterclockwise and meander considerably. The early morn-
ing air is probably clean, having come off the Atlantic. The air arriv-
ing at 1400 appears to have passed over the Durham area. The indica-
tions are that the morning ozone buildup is from local causes, and that
the high concentrations between 1300 and 1600 are sustained by transport
from nearby communities. The lack of hourly fluctuations in NO is prob-
ably explained by the lower level of traffic that would prevail on a
Saturday.
The pollutant histories on 23 September, displayed in Figure 45,
show that very high NO and NOo levels were present. The large early NOX
concentration suggests a low inversion that allowed them to accumulate.
In contrast to the other days at Site 8, there is considerably carryover
of NO and NO2 from the previous day. Since NO is so high, ozone essen-
tially vanishes. NO peaks at 0800 and quickly disappears thereafter,
allowing 0 to accumulate. Ozone continues to build up until 1700, when
105
-------
FIGURE 42 TRAJECTORIES ARRIVING AT SITE 8, RESEARCH TRIANGLE PARK,
NORTH CAROLINA, ON 31 AUGUST 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
10 15
LOCAL TIME
20
FIGURE 43 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 8 ON SATURDAY, 10 SEPTEMBER 1977
106
-------
FIGURE 44 TRAJECTORIES ARRIVING AT SITE 8, RESEARCH TRIANGLE PARK.
NORTH CAROLINA, ON 10 SEPTEMBER 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
C.I6
O.W
0.12
- 0.10
z
o
< 0.08
s.
Z
o
0
0.06
0.0«
0.02
T
T
NO
. i _ _
10 IS
LOCAL TIME
20
FIGURE 45 HOURLY CONCENTRATION VARIATIONS OF NO, N02, AND 03
AT SITE 8 ON FRIDAY, 23 SEPTEMBER 1977
107
-------
there is a break in the data. Meanwhile, N02 stays at about 10 ppb
until 1700, when it begins to increase.
Figure 46 displays the trajectories arriving at Site 8 on 23 Sep-
tember. Between 0200 and 1400 the air travels over the Washington, D.C.
area on its way Co Site 8, which suggests that pollutants from this
region impact Site 8. A long- or medium-range ozone transport
hypothesis is consistent with the appearance of the ozone peak at 1700
(cf. Figure 45^. Thus, while local factors are no doubt responsible for
the ozone buildup between 0800 and 1100, transport effects are the
likely causes of the higher ozone levels that occur after 1100.
FIGURE 46 TRAJECTORIES ARRIVING AT SITE 8, RESEARCH TRIANGLE PARK,
NORTH CAROLINA, ON 23 SEPTEMBER 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
108
-------
The investigation of ozone maxima at Site 8 reveals a mixture of
local and transport effects. This is. not unexpected for a suburban
location where automobile emissions can be a dominant factor. However,
if we could separate local and transport impacts, we would hypothesize
that the higher ozone levels, which generally occur late in the day, are
associated with transport from both Durham and Raleigh, depending on
wind direction. The lower values are more likely to be of local origin.
5. Case Study V: High Ozone
at, Lewisburg. West Virginia (Site _9)
A maximum hourly ozone level of 106 ppb was observed at Site 9 on
22 October. Figure 47 illustrates pollution conditions on this day:
some ozone carryover from the previous night, an Oj buildup that begins
late, at 0900, and very little NO. Ozone peaks late, at 1600, remains
X
high for the rest of the day, then decays slowly until midnight, when
its concentration reaches 60 ppb•
10 iS
LOCAL TIME
20
FIGURE 47 HOURLY CONCENTRATION VARIATIONS OF NO. N02. AND 03
AT SITE 9 ON SATURDAY, 22 OCTOBER 1977
109
-------
Meteorological patterns on 22 October, a Saturday, include a high
pressure system off the coast of North Carolina, the system covering
most of the eastern United States and reaching well beyond West
Virginia. However, a front approximately aligned across the United
States from New Mexico to New York is moving toward the southeast, pass-
ing over Site 9 on the morning of 23 October, bringing behind it another
high-pressure system. The presence of a high-pressure system on the
22nd is manifested in the trajectories shown in Figure 48. While the
trajectories follow a slight anticyclonic pattern, this is weakened by
the front. The air masses arrive at the site from directions ranging
between west and southwest. It is unlikely that the air masses passed
over any major urban areas within 12 hours before arriving at Site 9.
\
f
FIGURE 48 TRAJECTORIES ARRIVING AT SITE 9, LEWIS8URG, WEST VIRGINIA,
ON 22 OCTOBER 1977
The number shown at one extreme of a trajectory is the local time
of arrival at the site.
110
-------
The pollutant histories shown in Figure 47 imply that the high
ozone observed at Site 9 is due to transport. However, the origin of
the air is unclear from the trajectory analysis. Once again, NOX levels
are very low, yet 63 remains at about the same level for several hours.
Thus, ozone is neither made nor destroyed; NOX appears to be superfluous
in this case•
D. Conclusions
Ozone levels greater than 100 ppb were observed on a total of six-
teen days at five sites, which included all four rural sites (Sites 1,
4, 6, and 9) and one suburban location (Site 8). Five of the high-ozone
days occurred at Site 8, six at Site 1, three at Site 6, and one each at
Sites 4 and 9. It is significant that all the locations that are least
susceptible to localized sources exhibited high ozone levels, whereas
four of the five suburban locales did not. This phenomenon is typically
observed in urban regions, where ozone levels are often lower than in
areas with a lower source density.
Influence of transport was evident on all the occassions when ozone
was greater than 100 ppb. Transport was the most likely cause of the
high ozone on eleven of the sixteen days, and both local chemistry and
transport were present on the other five days. Local effects are indi-
cated on 27 August at Site 1 (cf. Figure 16),. 22 August and 23 September
at Site 6 (cf. Figures 31 and 32), and 31 August and 23 September at
Site 8 (cf. Figures 41 and 45). Figure 41 shows that on 31 August the
local chemistry at Site 8 is dominated by traffic-related precursors,
evidenced by the NO peak that occurs at 0800. Figure 45 also exhibits a
traffic-induced NO peak at 0800, but there is evidence of a substantial
amount of precursors present before 0500. In the five cases where local
chemistry is Important, local effects lead to a rapid ozone buildup,
which usually reaches levels below 100 ppb. After ozone reaches a max-
imum, transport effects act to lengthen the time when ozone levels
remain high. Transported ozone also combines with the local product to
exceed 100 ppb.
Ill
-------
An interesting aspect of the high-ozone days at the SURE sites is
that the variations in the concentrations of NOX are remarkably small
compared to those of ozone. Moreover, the absolute levels of NO are
quite low during the day. At Site 1, for example, NO is less than 1 ppb
most of the time, and N02 fluctuates between 5 and 15 ppb, while ozone
variations span a range of 100 ppb or more. The same is true at Sites 4
and 9, and is often the case at Sites 6 and 8. Using the available
data, it is difficult to attribute this phenomenon to a particular fac-
tor. Thus, it is apparent that ozone fluctuations are heavily J.nflu-
enced by the sunlight cycle, and that there is sufficient NOX to produce
the observed ozone. But we have no data on the NMHC/NO ratio at these
X
sites, and the NO levels are so low that the N02/NO ratio can fluctuate
over a wide range owing to uncertainty in the NO measurements. Thus,
basic data are lacking to characterize the relationship between NO and
Jt
On at these sites. Recommendations to remedy this situation are pro-
vided in Section VII of this report.
112
-------
V ANALYSIS OF JETMORE, KANSAS, DATA
A. Introduction
SRI recently conducted a study for the Coordinating Research
Council (CRC), the main objective of which was to measure concentrations
of natural 0, in the free troposphere and determine mechanisms by which
stratospheric Og was transported into the troposphere. An aircraft was
equipped with 03, NO/NOX, and meteorological instrumentation to make
these measurements to an altitude of 7 km. To complement aircraft data,
a ground station was set up in a rural location.
The ground station operated between 4 April and 18 May 1978 at
Jetmore, about 20 miles north of Dodge City (38°N, 100°W), Kansas. The
site was specifically chosen because it is a clean site removed from
sources of contamination. The ground station continuously measured CH^,
CO, THC, NO, N02» 03, and various meteorological parameters. Other
species of interest were measured in a batch mode. Table 19 provides a
list of chemical and meteorological parameters measured at the ground
station. The air quality instrumentation used is listed in Table 20.
In the following sections we discuss the characteristics of the air
quality data.
Because of the high quality of the measurements and the rural
nature of the monitoring site, the data are useful for the purposes of
this study.
B. Ground Station Data
Before proceeding with the data analysis, it is useful to note that
the Jetmore site was a clean site only infrequently and marginally
affected by urban transport. An indication of the cleanliness of the
air can be obtained from Table 21, which lists the average concentra-
tions of several chemicals measured during the study. The average
methane and CO concentration of 1650 and 175 ppb, respectively, are
113
-------
Table 19
CHEMICAL AND METEOROLOGICAL PARAMETERS
MEASURED AT JETMORE, KANSAS
Chemical Parameters
Meteorological Parameters
3
NO, NO
C, to Cc hydrocarbons:
CH4, C2H6, C3Hg, C3H&
i"C4H10' n~C4H10' i"C5H12
and
Temperature at two levels
Wind speed at two levels
••**0i»
Wind direction at two levels
Solar flux
Relative humidity
Fluorocarbon 12
CO
PAN
Beryllium-7 (7Be)
Table 20
SRI MOBILE RESEARCH LABORATORY INSTRUMENTATION
Parameter
Instrument
Ozone
NO/NO
X
C -C Hydrocarbons
Fluorocarbon 12
CO-CH.-THC
4
PAN
Beryllium-7 (7Be)
Bendix 8002 Chemiluminescent
analyzer
Monitor Labs Model 8440E
(specially modified for greater
sensitivity)
Perkin Elmer 3920 GC with FID
Perkin Elmer 3920 GC with ECD
Bendix 6800
Dual ECD Coulometer
GIHA high-volume sampler
Ge (Li) detector
114
-------
Table 21
AVERAGE CONCENTRATIONS
OF SELECTED SPECIES
AT JETMORE, KANSAS
Compound
CO
C2H6
C3H8
n-C4H10
SH6
Fluorocarbon 12
PAN
N02
NO
Concentration
1650 ppb
175 ppb
20.7 ppbC
9.9 ppbC
3.2 ppbC
8.4 ppbC
3.2 ppbC
3.6 ppbC
0.8 ppbC
0.8 ppbC
240 ppt
asO.4 ppb
4 ppb
1 ppb
close to their expected background levels. The average concentration of
20*7 ppbC for C^Hg was higher than expected. Propane, butanes, and pen-
tanes were all present at concentrations of a few ppbC, as shown in
Table 21. The NMHC have a concentration of about 50 ppbC. Table 21
makes it clear that alkenes are present at extremely low concentrations.
Figure 49 shows that CH^ has a steady mean diurnal concentration.
The HC levels are typical of clean sites at ground level. It is impor-
tant to note that all alkenes added up to only 1.6 ppbC. On the other
hand the alkanes (€2 and above) approached 50 ppbC. Even if one
115
-------
3000
2500
2000
i
o
1600
WOO
600
I . I
10 15
LOCAL TIME
20
FIGURE 49 MEAN DIURNAL VARIATION IN CH4 AT JETMORE, KANSAS,
3 APRIL-20 MAY 1978
excluded ethane as being relatively unreactive, the concentration of C^
to Cc alkanes was roughly 20 times that of the alkenes. The concentra-
tions of CH^ and Cflb are higner than what is typically encountered in
rural or remote locations. This, however, is not entirely surprising
because of the large number of gas fields in the region; such fields are
known to be significant sources of CH^ and Cfl^' Furthermore, while we
only measured light hydrocarbons, it appears that there were no other
organic compounds of higher molecular weight present here. The total HC
measurement was nearly identical to the CH^ level. Attempts made by
Washington State University to measure aromatics at this site found
these to be virtually nonexistent.
Figure 50 shows the mean diurnal variations of carbon monoxide
(CO) at Jetmore. CO is not released from gas fields and is therefore a
good indicator of urban transport. The average CO concentrations of
116
-------
900
400
BOO
200
too
I
I
10 15
LOCAL TIME
20
25
FIGURE 50 MEAN DIURNAL VARIATION IN CO AT JETMORE, KANSAS,
3 APRIL-20 MAY 1978
175 (± 40) ppb is typical of background levels at this latitude. It is
also clear from Figure 50 that no diurnal variation exists, which is
another indication of the lack of impact from urban sources.
An analysis of light HC data suggests that the C3 to C5 alkanes and
CO levels are typical of clean rural environments. Fluorocarbon-12
(F-12) is an ideal indicator of any urban source of contamination, since
it is completely inert and almost exclusively emitted in urban areas.
The overall average concentration of about 240 ppt (Table 21) is almost
identical to the prevailing geophysical background of F-12. In addi-
tion, the variability of F-12 was found to be small (about 5 percent) .
These observations support the view that there is no significant urban-
type contamination above the background levels in the region of
operation.
117
-------
A third indicator of local photochemistry is PAN. The concentra-
tion of PAN was found to be very low and did not exceed I ppb. On
nearly 75 percent of the days, the mean PAN concentration level was less
than 0.5 ppb. The daily average PAN concentration was never less than
0.1 ppb. As we shall see later, the very low PAN concentration is con-
sistent with the precursor mix present in this area.
1.
Ozone-No^ Measurements
Figure 51 shows the variation of maximum 1-hour daily 03 and the
daily 03 average during the study period (3 April to 20 May). It is
clear from Figure 51 that the maximum one-hour 0-j never exceeded 70 ppb,
while the daily mean never exceeded 55 ppb. Figure 52 shows the average
Z
O
GC
I-
Z
til
U
Z
O
u
IOC
80
60
40
20
Oailv Maximum O*
I
I
I
10 15 20 25 30
NUMBER OF DAYS
35
40
45
50
FIGURE 51 MEAN AND MAXIMUM 03 AT JETMORE, KANSAS,
3 APRIL-20 MAY 1978
US
-------
too
80
S 60
S 40
20
I
I
10 IS
LOCAL TIME
20
25
FIGURE 52 MEAN DIURNAL VARIATION IN 03 AT JETMORE, KANSAS,
3 APRIL-20 MAY 1978
diurnal variation of 03 and the standard deviation associated with this
diurnal variation* It is apparent that the 0, maximum appears somewhat
late (1600-1800 LT).
Figures 53 to 55 show the mean diurnal variation of NO, N02, and
NOX» Average levels of NO, N02, and N0x are 1.2 ppb, 4.3 ppb, and 5.5
ppb, respectively. The NO/NOX ratio of 0.22 is comparable to that
observed at some rural sites in the SURE data (cf. Table 15). Although
the mean NO levels throughout the day are essentially constant, N02
shows a slight decline in the afternoon. The sensitivity of NO and NO
X
measurements was better than 1 ppb; therefore, these data should be
reliable. The source of the NOX in this region is currently unclear.
Agricultural sources of NOX cannot be ruled out and should be further
studied.
119
-------
25
20
i 10
0 5 10 IS 20
LOCAL TIME
FIGURE 53 MEAN DIURNAL VARIATION IN NO AT JETMORE, KANSAS,
3 APRIL-20 MAY 1978
25
25
20
IS
«>
I
_L
10 IS
LOCAL TIME
20
25
FIGURE 54 MEAN DIURNAL VARIATION IN N02 AT JETMORE. KANSAS
3 APRIL-20 MAY 1978
120
-------
20
16
I
10 IS
LOCAL TIME
20
25
FIGURE 55 MEAN DIURNAL VARIATION IN NOX AT JETMORE, KANSAS,
3 APRIL-20 MAY 1978
Both hydrocarbons and NOX are present at relatively low concentra-
tions at this site. Excluding the essentially unreactive CH4 and C2H6,
there are only about 30 ppbC of alkanes and 1.6 ppbC of alkanes. These
low levels of HCs seem to be consistent with the very low levels of PAN
at this location. It should also be added that aircraft data taken in
the vicinity of Jetmore showed 0^ levels aloft comparable to or larger
than the hourly maximum observed at ground level. Therefore, while 03
undergoes a diurnal variation at ground level, we feel it is largely
dictated by 0^ loss at the ground and mixing from aloft. Not only are
the HCs here largely slow reactive (alkanes), but they are also present
in low concentrations. A small amount of local 03 production, however,
cannot be ruled out•
121
-------
The days where the one-hour 0^ maximum was found to be in excess of
58 ppb were studied. Figures 56 to 58 display the 0-j, CO, CH^ and NOX
data for these days, together with the wind speed data. Figure 57 shows
somewhat higher than average CO levels in the morning that rapidly
decline in the afternoon. The decline in CO and NO is associated with
an increase in 0.,. This is indicative of mixing processes that are
I
8
•»
I
I "
i
40
20
WS+50
\
CO
\
NO,
CH4
-1.
10 16
LOCAL TIME
20
25
FIGURE 56 MEAN DIURNAL VARIATIONS IN WIND SPEED AND SELECTED POLLUTANTS
AT JETMORE, KANSAS. ON 27 APRIL 1978
122
-------
I
4
8
40
20
WS+50
10 IS
LOCAL TIME
20
25
FIGURE 57 DIURNAL VARIATIONS IN WIND SPEED AND SELECTED POLLUTANTS
AT JETMORE, KANSAS, ON 28 APRIL 1978
£ IOO
1
§
i "
2
3 An
ww
1
1 40
S
*
s
1
x o
i • i • i ' i
WS+50 S N\
«-i»i ^^~~ ^**^ ^^^-
-
°3 xx \
- --.^ s~' v--.. -
CO
i .'\ CH4 , ,
NOX
10 IS
LOCAL TIME
20
25
FIGURE 58 DIURNAL VARIATIONS IN WIND SPEED AND SELECTED POLLUTANTS
AT JETMORE. KANSAS, ON 15 MAY 1978
123
-------
replenishing 03 in the boundary layer from aloft. On 27 April (Figure
56), the CO levels are high throughout the day and may indicate some
contamination. The wind speed in the afternoon was high (> 30 mph);
therefore, the 0-^ maximum, once achieved, was maintained by constant
mixing from aloft.
The CO levels on 15 May were somewhat higher, but are within the
range of variability of clean continental background of CO. The wind
speed was less than 10 mph. It is possible that some Oo transport may
have occurred•
Thus, it appears that this location is probably deficient in HCs
and that little photochemical activity is going on. The On is largely
brought down to the ground from the free tropospheric reservoir.
2. £3 and 7Be Relationship
Since the 0, at Jetmore did not appear to be photochemically gen-
erated, we tested the data to see whether a correlation between 0., and
Be existed. The Be data were averaged over 24 hours, and these were
compared with daily average 0-j. Figure 59 shows the daily average con-
centrations of 0-j and 'Be at Jetmore. Typically, average 'Be concentra-
tions varied between 80 and 400 pci/1000 SCM. In one instance, levels
as high as 640 pci/1000 SCM were measured. It is clear from Figure 59
that fairly large variations in 'Be levels occur- It also appears that
peaks in 0, coincide with corresponding peaks in Be, at least some of
the time. Figure 60 shows a scatter diagram of 0, versus Be. As indi-
cated by the scatter in the data, the linear correlation was very low
(R » 0.25). The daily one-hour 0., maximum is more representative of
free tropospheric 0^, and it was tested to see whether any correlation
existed with daily average 'Be. This, too, failed to show a substantial
correlation although a small improvement (R =• 0.31) was observed. When
we considered only those days when Be concentration was greater than
124
-------
300 pel/100 SCM, this correlation showed significant improvement
(R = 0.29).
3. Conclusions
It appears that a location such as the Jetmore, Kansas, site is low
in hydrocarbons abundance, and even the HCs present are alkanes which
react slowly compared to alkenes. The NMHC (less C2H6) was present at a
concentration of 30 ppbC along with an alkene concentration of less than
700
3 500
3
400
300
63 Daily Mean
10
20 30
NUMBER OF DAYS
40
50
FIGURE 59 MEAN DAILY 03 AND 7Be AT JETMORE, KANSAS,
3 APRIL-20 MAY 1978
Ozone mean obtained for interval 1700-1700.
125
-------
2 ppbC and an NO concentration of 5 ppbC. It does not appear that
A
local photochemical activity that could result in significant 0., forma-
tion was occurring. The indications are that On concentration was
largely controlled by mixing from the free troposphere and ground level
destruction.
700 f
600
500
.c
a
a
400
Z 300
UJ
O
z
o
o
200
00
••
100
200
300
400
500
600
700
7Be - pci/1000 SCM
FIGURE 60 SCATTER DIAGRAM FOR MEAN DAILY 03 VERSUS 7Be
AT JETMORE, KANSAS, 3 APRIL-20 MAY 1978
126
-------
VI CONCLUSIONS
A primary objective of this investigation was to survey and summa-
rize the current state of knowledge about nonurban N0x/0-j interactions.
Our results show that present knowledge can be characterized as fragmen-
tary. Consequently, there is currently no direct quantitative evidence
that answers the question that motivated this study, namely, whether
increases in urban NOV emissions will enhance, reduce, or have no effect
X
on nonurban ozone. Relevant previous research efforts have not been
aimed at this specific problem and the inferences that can be drawn from
them are at best qualitative and suggestive, rather than quantitative
and conclusive. Therefore, additional research is needed to establish
the relationship between NO and nonurban ozone. Pursuant to another
X
objective of the project, three research programs are proposed in
Section VII that respond to this need. The available evidence allows
some tentative conclusions to be reached nonetheless, and these are dis-
cussed below.
NOX transported from urban to nonurban areas affects ozone forma-
tion by participating in two categories of chemical processes: local and
nonlocal. In the local domain, transported NOX increases the nonurban
NO level, and thus participates in the local chemical reactions that
X
form ozone. In a nonlocal mode, NOX and hydrocarbons of urban origin
react to form ozone in an air mass on the way to a nonurban area. Ozone
thus arrives ready-made, and combines with existing ozone to increase
the concentration at the nonurban locale. The time of occurrence
differs for these two effects. Local chemical effects tend to predom-
inate in the morning, which implies that transported NOX must arrive
during the night or early in the morning in order to participate in the
local photochemical process. Nonlocal effects require several hours to
develop, since the air mass must travel distances of the order of tens
of miles from source to receptor. Consequently, enhanced levels of
nonurban ozone due to nonlocal causes tend to occur in the late after-
noon and early evening, and sometimes at night.
127
-------
We consider the following two Issues to be important:
• Which effect, local or nonlocal, is most frequently associated
with high nonurban ozone?
• What is the likely impact of nonurban ozone if N0x levels
increase either in a local or nonlocal mode?
Regarding the first question, the analysis of field data collected
in the SURE (see Section IV) indicates that nonlocal effects predom-
inated when ozone levels exceeding 100 ppb occurred in rural and subur-
ban areas. Of the nine SURE sites investigated, those exhibiting at
least one hour when ozone exceeded 100 ppb included all four rural sites
(Sites 1, 4, 6, and 9) and one suburban location (Site 8). Site 8 is
influenced by emissions of ozone precursors from commuter traffic, but
is also surrounded by several medium-sized cities. Hence, it is also a
receptor for pollutants that originate in these urban centers. The four
rural sites, by contrast, show no influence from traffic or other
specific sources.
Ozone exceeded 100 ppb on 16 days at these five sites. Nonlocal
effects are the probable cause of the high ozone on 11 days, and both
local and nonlocal factors are evident on the remaining 5 days. It is
apparent, therefore, that nonlocal effects are closely associated with
all the high ozone events at the SURE sites. While it is risky to gen-
eralize based on this small sample, the evidence does suggest that ozone
formation in an air mass en route to a nonurban site is the principal
mechanism that leads to high ozone levels, and that ozone synthesized
from local precursors (which may be affected by transported NO and
X
hydrocarbons) plays a lesser role. In particular, it appears that
locally formed nonurban ozone does not by itself lead to concentrations
exceeding 100 ppb, although the local component of the total ozone bur-
den can be substantial depending on time of day. The analysis further
suggests that transport times of the order of six to eight hours are
most likely to be associated with the observed high ozone.
128
-------
As noted earlier, it is not possible to answer confidently the
question about the likely impact of increased NO on nonurban ozone.
X
The review of the literature reveals a convergence of several indepen-
dent lines of research that suggest that the likely impact of increases
in NOX emissions will be to enhance nonurban ozone. This judgment
agrees with the consensus response to a question on NO control stra-
X
tegies contained in the survey of expert opinion (see Section III-B-4)•
However, this estimate of likely impact is only qualitative, and cannot
serve as a basis for regulatory decisions. Moreover, the impact of NOX
on nonurban ozone is undoubtedly a function of geographical location
since ozone production can be NO -limited at some sites, but not at oth-
X
ers. An example of the latter is the site in Jetmore, Kansas (see
Section V), where local ozone production appears to be limited by hydro-
carbons, rather than by NO . Hence, it is inappropriate to attempt to
X
state a general rule that applies equally to all locations. Neverthe-
less, although the signs are uncertain, they seem' to point in the direc-
tion of deleterious effects, adding a note of urgency to the need to
perform more research to test the validity of these preliminary
indications.
If confirmed by more detailed studies, the conclusions discussed
above imply that more stringent NOV controls may be necessary to allevi-
X
ate (or at least not worsen) the problem of high ozone levels at some
nonurban areas. Because of the geographically widespread nature of the
problem, control strategies with a regional scope should be the pre-
ferred approach to a solution. However, the indications are that not
all nonurban areas would benefit from such measures, and that it would
be more efficient to tailor controls to specific areas.
129
-------
VII KECOttfEHDATIOHS FOR FURTHER RESEARCH
The previous discussions have made it dear that additional study
is needed to characterize quantitatively (and perhaps mechanistically)
the relationship between MOX and 0, in nonurban areas. This can best be
done in die context of a program specifically designed for this purpose.
Three such programs are outlined below. Two are retrospective analyses
of existing data that would be relatively easily performed, whereas the
third program entails data collection as veil as analysis. These stud-
ies are ->TI«-«»TW*«»H to determine die role BO frgg played in oxidant forma-
tion in the recent past, and to define the impact of future changes in
BOX on the formation of nonurban ozone.
A- Study Ii Analysis of Data fron Rural Locations
In many rural locations (as distinct from remote locations), the
•Ox levels vary from 5— to 50-ppb levels. The reliability of data col-
lected in this concentration range is fair to good. Over the last five
years, a number of stations have been operated to collect both short-
and long-term BOX data on an essentially continuous basis • (The short-
term implies a period of one to three mondis, while die long tern could
be diree months to several years.) We have in die past looked at a lim-
ited amount of long-term K>x data (Singh et al., 1977) and are assured
of die availability of much more extensive data bases around die
country*
Because of less stringent controls on the emissions of SOX compared
to HC emissions, the amount of BOX released to -he atmosphere has
steadily increased over the past five years. We suggest that a
comprehensive analysis of BOX and 0^ data in rural locations within the
United States be conducted widi die primary view of obtaining answers to
the following questions
131
-------
• Has the atmospheric abundance of NOX Increased over the last
five years?
• Is this increase (if any) consistent with regional increases in
NO emissions?
Jt
• Have the corresponding 03 levels registered any changes? If so,
what kind of changes?
• Are the 03 variabilities in any way, statistically or mechanist-
ically, related to the variabilities of NOX? If so, what is the
nature of this relationship?
It is not our intent here to develop a detailed plan for such a
study; rather, we are attempting to outline the nature of analysis that
would utilize existing data, obtained at great cost, to study NO /O^
JS -j
relationships under actual atmospheric conditions. The effort to con-
duct such a study should be a small one (six to eight months) and the
potential utility of results would more than justify the cost.
B. Study II; Analysis of Data from the Los Angeles Area
This program would yield direct evidence about the kind of downwind
impact that can be expected when urban NO emissions increase. The pro-
Ai
gram entails a retrospective analysis of air-quality data collected in
the 1960s in areas downwind of the Los Angeles basin. It is well known
that NO emissions from automobiles increased substantially in Los
Angeles in the late 1960s* The increases in NO resulted from attempts
A
to control automobile hydrocarbon emissions. At the time, the techno-
logical solution for controlling hydrocarbons consisted of adjusting the
gasoline/air mixture for more complete combustion; the desired reduction
in hydrocarbon emissions was achieved at the expense of increased NOX«
The situation was corrected in the early 1970s, when limits were placed
on emissions of both hydrocarbons and NO . Thus, in the late 1960s, the
A
Los Angeles area experienced precisely the condition that concerns the
present study, namely, increased NO and reduced hydrocarbons. An
Jv
analysis of the air quality and meteorological data before and during
the time of increased N0_ would answer the question: What happens to
Jv
ozone downwind of an urban area when NO increased and hydrocarbons
Ju
decrease?
132
-------
The method of "intervention analysis" developed by Box and Tiao
(1975) would be especially useful for performing the suggested study.
In fact, the method has been applied to the problem of determining the
impact of the aforementioned engine adjustments on ozone levels in down-
town Los Angeles (Tiao et al., 1975). However, the impact on ozone at
downwind locations was not examined.
Effects of increased NO emission on NO concentrations at various
X
sites in Los Angeles were recently examined by Phadke et al. (1978),
while Trijonis (1978) analyzed long-term trends in N0«. Neither of
these studies addresses the issue that concerns us here, since they are
restricted to NO and NCU, and no attempt was made to examine the
interactions between NOX and 0-j.
C. Study III; Data Collection and Analysis Program
One of the major uncertainties in assessing the role of NOX in
rural locations is simply the lack of available information on which to
base such as assessment. No-models exist that are adequately validated
for rural environments. Besides, there is no reason to believe that
most rural areas would have identical NO^Oj relationships; on the con-
trary, the precursor compositions in various rural locations may result
in completely different NO^Oj relationships.
Recognizing that we are in the preliminary stages of defining the
rural NO^Og relationships, we present here a field program that would
provide a great deal of useful information about these relationships,
and would lay the groundwork for more extensive field studies in the
future, should they be needed. The proposed field program would provide
an essential insight into the NO-/Oo relationships in several rural
locations. The basic field program can be outlined as follows:
• Select five to ten representative rural sites in the country.
These should include rural sites downwind of large and small
urban centers of varying characteristics, downwind of power
plant plumes, refineries, isolated rural centers, and the like.
133
-------
• Equip a mobile environmental laboratory with reliable 03, NOX
and HC instrumentation. Here we would expect that NOX analysis
sensitivity be about 1 ppb.
• Go to Site 1 (say in July) with the mobile laboratory and a
dozen Teflon bags of 500-liter volume each. On a given day.
fill five bags simultaneously between 0600 and 0900 LT. Analyze
them for HC distributions, 0, and NO .
• Spike four bags with NO levels of 2 NOX, 3 NOX, 5 NOX, and 10
NOV, where 1 NOV would be the N0_. concentration in the unspiked
XX X
bag.
• Monitor these bags for HCs, NOX, and 03 for the next 36 hours.
• Quantify and collect air quality and meteorological information
(e.g., solar flux).
• Repeat above experiments on another day.
It is expected that normal scientific procedures (e.g., conduct
replicate experiments) be used to ensure reliability. The above steps
could then be repeated at other sites.
A five-site experiment of this type could be conducted over a five
to eight week period with minimal cost. It should be added that the
spiking principle is well known and has been used in the past to obtain
valuable information. However, the bags used have been too small, and
no attempt at a systematic characterization of NOX/03 dependence has
been made to date. The suggested systematic approach would be rela-
tively inexpensive, and would yield useful information about NOx/03
relationships. For example, the program would provide data that would
answer the question of whether NOX/03 relationships at various represen-
tative rural locations are similar or different. In addition, the
potential impact of changes in NO levels would be demonstrated. Such a
field program would benefit from the simplicity of the smog chamber
approach and the realism of field conditions.
D. Additional Topics for Investigation
The following emerged in the course of the study as Important
topics that require further research to characterize N°x/03 interactions
in the nonurban atmosphere.
134
-------
• Estimate the tropospheric half-life of NCL..
X
• Establish the role of low-reactive and natural hydrocarbons in
nonurban ozone formation.
• Define the significance of organic nitrates as reservoirs of NOX
and organic radicals*
• Establish the identity and relative importance of sinks and
natural sources of NO .
X
The first item is important in estimating the time (and conse-
quently distance) during which NO can be transported in significant
quantities. Thus, a short half-life would limit the zone of influence
of an urban source and vice versa. To be sure, a number of the studies
reviewed, and the expert commentary, suggest that the half-life is of
the order of 24 hours. However, many uncertainties remain. Thus, the
question of sinks and sources of NO in the nonurban environment is one
X
that must be resolved as part of the process of estimating the half-life
of NOX- Data needs related to the issue of the half-life of NOX include
nighttime measurements of NOX aloft. Preferably, these should be per-
formed in a Lagrangian mode. Also required are highly sensitive mea-
surements of levels NOX as low as 0.10 ppb.
The zone of influence of urban areas can be expanded if longer-
lived organic nitrates were to act as reservoirs of NO and organic rad-
X,
icals• Current knowledge about this topic is sketchy and somewhat
speculative. To resolve this question, we need to obtain accurate mea-
surements of organic nitrates, expecially PAN, in nonurban areas. From
such data it may be possible to estimate the quantity of NO contributed
X
by this source.
No assessment of the role of NOX in ozone formation can be complete
without considering the third dimension, namely, the hydrocarbons. To
do so, we need to obtain more detailed hydrocarbon measurements at rural
sites, including aldehydes and natural hydrocarbons. The discussions in
Sections II and III make it apparent that the role of natural hydrocar-
bons is currently the subject of considerable controversy; more data are
needed to settle the argument.
135
-------
Investigation of these topics would require new data to be gathered
and analyzed. Such investigations could be pursued independently of the
three projects outlined above.
136
-------
REFERENCES
Arnts, R.R. and B.W. Gay, 1979: "Photochemistry of Naturally. Emitted
Hydrocarbons: Isoprene, p-Cymene and Selected Monoterpenes," Draft
Report, U.S. EPA (April).
Blumenthal, D.L., W.H. White, R.L. Peace, and T.B. Smith, 1974: "Deter-
mination of the Feasibility of the Long-Range Transport of Ozone or
Ozone Precursors," EPA-450/3-74-061 (November).
Box, G.E.P., and G.C. Tiao, 1975: "Intervention Analysis with Applica-
tions to Economic and Environmental Problems," ^J. Amer. Stat.
Assoc., Vol. 70, No. 349, pp. 70-79.
Breeding, R.J., J.P. Lodge, J.B. Pate, D.C. Sheesley, H.B. KLovis, B.
Fogle, J.A. Anderson, T.R. Engberg, P.L. Haagenson, R.B. McBeth,
A.L. Morris, R. Pogue, and A.F. Wartburg, 1973: "Background Trace
Gas Concentrations in the Central U.S.," JGR, Vol. 78, No. 30, pp.
7075-7064.
Bufalini, J.J., 1979: Correspondence, Science, Vol. 203, p. 81
(January).
California Air Resources Board, 1978: Summary of 1977 Air Quality Data,
Vol. 9, pp. 23-27.
Chameides, W.L., 1978: "The Photochemical Role of Tropospheric Nitrogen
Oxides," Geophysical Research Letters, Vol. 5, No. 1, pp. 17-20
(January).
Chameides, W.L. and D.H. Stedman, 1977: "Tropospheric Ozone: Coupling
Transport and Photochemistry," ^J. Geophys. Res., Vol. 82, pp.
1787-1794.
Chameides, W.L., D.H. Stedman, R.R. Dickerson, D. W. Rusch, and R.J.
Cicerone, 1977: NOX Production by Lightning," _J. Atmos. Sci., Vol.
34, pp. 143-149.
Chang, T.Y., and B. Weinstock, 1976a: "Net Ozone Formation in Rural
Atmospheres," International Conference on Photochemical Oxidant
Pollution and its Control, Raleigh, North Carolina (12-17
September).
Chang, T.Y., and B. Weinstock, 1976b: "Methane and Nonurban Ozone," Air
Pollution Control Association, 69th Annual Meeting, Portland,
Oregon (June).
Cleveland, W.S., and T.E. Graedel, 1979: "Photochemical Air Pollution in
the Northeast United States," Science. Vol. 204, No. 4399, pp.
1273-1278 (22 June).
137
-------
Coffey, P.E., and K.H. Westberg, 1977: "The Issue of Natural Organic
Emissions," EPA-600/3-77-116 (October).
Council for Agricultural Science and Technology (CAST), 1976: Effect of
Increased Nitrogen Fixation on Stratospheric Ozone," Report No. 53,
Council for Agricultural Science and Technology, Ames, Iowa.
Cox, R.A, 1977: "Some Measurements of Ground Level NO, N02> and Og.
Concentrations at an Unpolluted Maritime Site," Tellus, Vol. 29,
pp. 356-362.
Cox, R.A., R.G. Derwent, and P.M. Hold, 1975a: "The Photooxldation of
Ammonia in the Presence of NO and NO*, Chemosphere, No. 4, pp.
201-205.
Cox, R.A., A.E.J. Eggleton, R.G. Derwent, J.E. Lovelock, and D.H. Pack,
1975b: "Long-Range Transport of Photochemical Ozone in Northwestern
Europe," Nature. Vol. 255, No. 5504, pp. 118-121 (8 May).
Cox, R.A., R.G. Derwent, and A.J.L. Hutton, 1977: "Significance of
Peroxynitric Acid in Atmospheric Chemistry of Nitrogen Oxides,"
Nature, Vol. 270, pp. 328-329 (24 November).
Decker, C.E., J.J.B. Worth, C.A. Ripperton, and W.D. Bach, 1977:
"Ambient Monitoring Aloft of Ozone and Precursors Near and Downwind
of St. Louis," EPA-450/3-77-009 (January).
Delwiche, C.C., 1970: The Nitrogen Cycle, Scientific American. Vol. 223,
No. 3, pp. 137-146.
Demerjian, K.L., J.A. Kerr, and J.G. Calvert, 1974: "The Mechanism of
Photochemical Smog Formation," Advances in. Environmental Science
and Technology, Vol. 4, pp. 1-262.
Drummond, J.W., 19.77: "Atmospheric Measurements of Nitric Oxide Using a
Chemiluminescent Detector," Ph.D. Dissertation, University of
Wyoming, Laramie.
Federal Register, 1973: Vol. 38, No. 110, pp. 15174-15175, (8 June).
Fishman, J., and P.J. Crutzen, 1977: "A Numerical Study of Tropospheric
Photochemistry Using a. One-Dimensional Model," ^J. Geophys» Res.,
Vol. 82, No. 37, pp. 5897-5906 (20 December). [A correction of
this paper appeared in the March 20, 1978 issue of the journal,
Vol. 83, No. C3, p. 1383.]
Graedel, T.E., and D.L. Allara, 1976: "The Kinetic Ozone Photochemistry
of Natural and Perturbed Nonurban Tropospheres," Int'l Conference
on Photochemical Oxidant Pollution and Its Control, Raleigh, North
Carolina, pp. 467-473 (12-17 September).
138
-------
Graedel, I.E., and L.A. Farrow, 1977: "Kinetic Photochemistry Downwind
Over Water from Urban Areas," JGH, Vol. 82, No. 31, pp. 4943-4946
(20 October).
Harrison, R.M., C.D. Holman, H.A. McCartney, and J.F.R. Mcllveen, 1978:
"Nocturnal Depletion of Photochemical Ozone at a Rural Site,"
Atmos. Env, Vol. 12, pp. 2021-2026.
Hathorn, J.W., and H.M. Walker, 1976: "A Texas-Size Episode Tracked To
Its Source," Proc. Int/1 Conf. Photochemical Oxidant Pollution and
Its Control. Vol.. 1, Raleigh, North Carolina, pp. 353-380 (12-17
September).
Heffter, J.L., and A.D. Taylor, 1975: "A Regional Continental Scale
Transport, Diffusion and Deposition Model, Part I: Trajectory
Model." NOAA Tech. Memo ERL ARL-50.
Hendry, D.G., 1978: "Reactions of Aromatic Compounds in the Atmosphere,"
Conference of Chemical Data Needs for Modeling the Lower Tropo-
sphere, Reston, Virginia (15-17 May).
Hendry, D.G., A.C. Baldwin, J.R. Barker, and D.M. Golden, 1978: "Com-
puter Modeling of Simulated Photochemical Smog," EPA 600/3-78-059
(June).
Hendry, D.G., and R.A. Kenley, 1978: "Importance of Organic Peroxyni-
trate in Photochemical Smog," paper presented at the meeting of the
American Chemical Society, Anaheim, California (12-17 March).
Hill, A.C., 1971: "Vegetation: A Sink for Atmospheric Pollutants,"
J.' Air Pollut. Control Assoc. Vol. 21, p. 341.
Holland, H.D., 1973: "Ocean Water, Nutrients, and Atmospheric Oxygen,
Proc. Internat. Symp. Hydrogeochim* Biogeochen., ed. by E.
Ingerson, Vol. 1, pp. 68-71, Clarke County, Washington, D.C.
Hov, 0., E. Hesstvedt, and I.S.A. Isaksen, 1978: "Long-Range Transport
of Tropospheric Ozone," Nature, Vol. 273, pp. 341-344 (June).
Howard, C.J., and K.M. Evenson, 1977: "Kinetics of the Reaction of H02
with NO," Geophys. Res. Letters, Vol. 4, No. 10, pp. 437-440.
Isaksen, I.S.A., 0. Hov, and E. Hesstvedt, 1978a: "Ozone Generation over
Rural Areas," Env. Sci. and Technology. Vol. 12, No. 12, pp. 1279-
1284 (November).
Isaksen, I.S.A., E. Hesstvedt, and 0. Hov, 1978b: "A Chemical Model for
Urban Plumes: Test for Ozone and Particulate Sulfur Formation in
St. Louis Urban Plume," Atmos. Env., Vol. 12, pp. 599-605.
139
-------
Jeftries, H.E., J.E. Sickles, and L.A. Ripperton, 1976: "Ozone Transport
Phenomena: Observed and Simulated," Paper #76-14.3, 69th Annual
Meeting of the Air Pollution Control Assoc., Portland, Oregon, (27
June-1 July).
Kauper, E.K., and B.L. Niemann, 1976: "Transport of Ozone by Upper-Level
Land Breeze—An Example of a City's Polluted Wake Upwind From Its
Center," Proc. International Conference on Photochemical Oxidant
Pollution and Its Control, Vol. I, pp. 227-236, Raleigh, North
Carolina (12-17 September).
Knelson, J.H., and R.E. Lee, 1977: "Oxides of Nitrogen in the Atmo-
sphere: Origin, Fate and Public Health Implications," Ambio. Vol.
6, No. 2-3, pp. 126-130.
Liu, S.C., 1977: "Possible Effects on Tropospheric 02 and OH Due to NO
Emissions," Geophysical Research Letters, Vol. 4, No. 8, pp. 325-
328 (August).
Liu, S.C., T.M. Donahue, R.J. Cicerone, and W.L. Chameides, 1976:
"Effect of Water Vapor on the Destruction of Ozone in the Strato-
sphere Perturbed by C1X or NOX Pollutants," J_. Geophys. Res., Vol.
61, p. 3111.
Lodge, J.P., Machado, P.A., Pate, J.B., Sheesley, D.C., and Wartburg,
A.F, iy/4: "Atmospheric Trace Chemistry in the American Humid Tro-
pics," Tellus. Vol. 26, No. 1-2, pp. 250-253.
Lodge, J.P., and J.B. Pate, 1966: "Atmospheric Gases and Particulates in
Panama," Science. Vol. 153, pp. 408-410 (22 July).
Lonneman, W.A., 1976: "Ozone and Hydrocarbon Measurements in Recent Oxi-
dant Transport Studies," Proc. International Conference on Photo-
chemical Oxidant Pollution and Its Control, Vol. I, pp. 211-223,
' Raleigh, North Carolina (12-17 September).
Lonneman, W.A., R.L. Seila, and J.J. Bufalini, 1978: "Ambient Air Hydro-
carbon Concentrations in Florida," Environ. Sci. j^ Technol., Vol.
12, No. 4, pp. 459-463 (April).
Lonneman, W.A., and J.J. Bufalini, 1979: Correspondence, Environ. Sci.
and Technol.. Vol. 13, No. 2, pp. 237-239 (February).
Ludlum, K.H., and B.S. Bailey, 1979: Correspondence, Environ. Sci. and
Technol.. Vol. 13, No. 2, pp. 235-236 (February).
Ludwig, F.L., W.B. Johnson, R.E Ruff, and H.B. Singh, 1976: "Important
Factors Affecting Rural Ozone Concentration," Proc. Int'l Confer-
ence on Photochemical Oxidants and Its Control, Vol. 1, pp. 425-
538, Raleigh, North Carolina (12-17 September).
140
-------
Ludwig, F.L., E. Relter, E. Shelar, and W.B. Johnson, 1977: The Relation
of Oxidant Levels to Precursor Emissions and Meteorological
Features, EPA-450/3-77-022a.
Martinez, J.R., and K.C. Nitz, 1979: Analysis of High NO2 Concentrations
in California, 1975-1977. Draft Final Report, SRI International
Project No. 6780-12-13 (22 June).
McConnell, J.C., 1973: "Atmospheric Ammonia," _J. Geophys. Res., Vol. 78,
pp. 7812-7820.
Meyer, E.L., W. P. Freas, J.E. Summerhays, and P.L. Youngblood, 1976:
"The Use of Trajectory Analysis for Determining Empirical Relation-
ships Among Ambient Ozone Levels and Meteorological and Emissions
Variables," Proc. Int'l Conference on Photochemical Oxidant and Its
Control. Vol. 1, pp. 9"03-913, Raleigh, North Carolina (12-17
September)•
Miller, D.F., A.L. Alkezweeny, J.M. Hales, and R.N. Lee, 1978: "Ozone
Formation Related to Power Plant Emissions," Science, Vol. 202, pp.
1186-1188 (15 December).
Miller, P.R., J.N. Pitts, and A.M. Winer, 1979: Correspondence, Science,
Vol. 203, pp. 81-82 (5 January).
Mueller, P.K., G.M. Hidy, and E.Y. Tong, 1977: "Implementation and Coor-
dination of the Sulfate Regional Experiment (SURE) and Related Pro-
grams," Environmental Research and Technology, Inc., Report on
Research Project 682 (October).
Nash, T., 1974: "Nitrous Acid in the Atmosphere and Laboratory Experi-
ments on its Photolysis," Tellus. Vol. 26, No. 1-2, pp. 175-179.
National Academy of Sciences, 1978: Nitrates; An Environmental Assess-
ment , Washington, D.C.
Noxon, J.F., 1978: "Tropospheric N02," JGR, Vol. 83, C6, pp. 3051-3057
(20 June).
Noxon, J.F., 1975: "Nitrogen Dioxide in the Stratosphere and Troposphere
Measured by Ground-Based Absorption Spectroscopy," Science, Vol.
189, pp. 547-549 (15 August).
Phadke, M.S., M.R. Grupe, and G.C. Tiao, 1978: "Statistical Evaluation
of Trends in Ambient Concentrations of Nitric Oxide in Los
Angeles." Env. Sci. ^ Tech. Vol. 12, No. 4, pp. 430-435 (April).
Rasmussen, R.A., R.B. Chatfield, and M.W. Holdren, 1977: "Hydrocarbon
Species in Rural Missouri Air," College of Engineering, Washington
State University (January).
141
-------
Ratsch, H.C., and D.T. Tingey, 1978: "Tampa Bay Area Photochemical Oxi-
dant Study: Emission Rate for Biogenic NOX," Appendix C, Final
Report, EPA 904/9-77/036 (January).
Research Triangle Institute, 1975: "Investigation of Rural Oxidant
Levels as Related to Urban Hydrocarbon Control Strategies,"
EPA-450/3-75-036 (March).
Research Triangle Institute, 1976: "Formation and Transport of Oxidants
Along Gulf Coast and in Northern U.S.," EPA-450/3-76-033 (August).
Ripperton, L.A., L. Kornreich, and J.J.B. Worth, 1970: "Nitrogen Dioxide
and Nitric Oxide in Nonurban Air," JAPCA. Vol. 20, No.. 9, pp. 589-
592 (September).
Ripperton, L.A., J.E. Sickles, and W.C. Eaton, 1976: "Oxidant-Precursor
Relationships During Pollutant Transport Conditions-An Outdoor Smog
Chamber Study," EPA-600/3-76-107 (November).
Ritter, J.A., D.H. Stedman, and T.J. Kelly, 1978: "Ground Level Measure-
ments of NO, N02» and 0^ in Rural Air," ACS Symposium on Atmo-
spheric Nitrogen Compounds, Anaheim, California (12-17 March).
Robinson, E., and R.C. Robbins, 1970: "Gaseous nitrogen compound pollut-
ants from urban and natural sources," JAPCA, Vol. 20, pp. 303-306.
Robinson, E., and R.C. Robbins, 1971: Emissions, Concentrations and Fate
of Gaseous Atmospheric Pollutants, Air Pollution Control, Part II,
pp. 1-93, W. Strauss, ed., John Wiley and Sons (New York).
Rogers, H.H., H.E. Jeffries, E.P. Stahel, W.W. Heck, L.A. Ripperton, and
A.M. Witherspoon, 1977: "Measuring Air Pollutant Uptake by Plants:
A Direct Kinetic Technique," ^J. Air Pollut. Control Assoc., Vol.
27, pp. 1192-1197.
Sandberg, J.S., M.J. Basso, and B.A. Okin, 1978: "Winter Rain and Summer
Ozone: A Predictive Relationship," Science, Vol. 200, pp. 1051-
1054.
Sandberg J.S., M.J. Basso, and B.A. Okin, 1979: Correspondence, Science.
Vol. 203, pp. 82-83 (5 January).
Sandberg, J.S., R.E. DeMandel, M.J. Basso, and B.A. Okin, 1979b: "Veri-
fication of the First Seasonal Ozone Forecast," Proc. 4th Symp. on
Turbulence, Diffusion, and Air Pollution, Reno, Nevada, pp. 614-619
(15-18 January).
Sculley, R.D., 1979: Correspondence, Environ. Sci. and Technology. Vol.
13, No. 2, pp. 234-235 (February).
142
-------
Seinfeld, J.H., and K.R. Wilson, 1977: International Conference on Oxi-
dants, 1976—Analysis of Evidence, and Viewpoints. Part VI: The
Issue of Air Quality Simulation Model Utility, EPA-600/3-77-118
(November)•
Singh, H.B., F.L. Ludwig, and W.B. Johnson, 1977: "Ozone in Clean Remote
Atmospheres: Concentrations and Variabilities," Final Report SRI
Project 5661, SRI International (June).
Singh, H.B., J.R. Martinez, D.G. Hendry, R.J. Jaffe, and W.B. Johnson,
1978: "Estimates of the Oxidant Forming Potential of Light
Saturated Hydrocarbons," SRI International, Draft Final Report, SRI
Project 6395 (September).
Spicer, C.W, 1976: "The Fate of Nitrogen Oxides in the Atmosphere,"
EPA-600/3-76-030 (March).
Spicer, C.W, 1977: "Photochemical Atmospheric Pollutants Derived from
Nitrogen Oxides," Atmos. Env., Vol. 11, pp. 1089-1095.
Spicer, C.W., J.L. Gemma, D.W. Joseph, P.R. Sticksel, and F.G. Ward,
1976a: "The Transport of Oxidant Beyond Urban Areas," EPA-600/3-
76-018 (February).
Spicer, C.W., J.L. Gemma, P.M. Schumacher, and G.E. Ward, 1976b: "The
Fate of Nitrogen Oxides in the Atmosphere," Second Year Report,
PB-267784 (August).
Spicer, C.W., D.W. Joseph, and J. Kouyoumjian, 1977: "The Fate of Nitro-
gen Oxides in the Atmosphere-Summary of an Exploratory Study in
Phoenix, Arizona," Final Report, PB-276355 (March).
Soderlund, R., and B.H. Svensson, 1976: "The Global Nitrogen Cycle,"
Nitrogen, Phosphorus and Sulfur—Global Cycles, edited by B.H.
Svensson and R. Sodurlund, SCOPE Report 7, Ecological Bulletin
No. 22 (Stockholm, Sweden), pp. 23-73.
Sundareson, B.B., C.I. Harding, F.P. May, and E.R. Henrichson, 1967:
"Absorption of Nitrogen Oxides from Waste Gas," Environ. Sci. and
Technol.. Vol. 1, p. 151.
Tiao, G.C., G.E.P. Box, and W.J. Hamming, 1975: "Analysis of Los Angeles
Photochemical Smog Data: Statistical Overview," ^J. Air Poll.
Control Assoc. Vol. 25, pp. 260-269.
Tingey, D.T. 1968: "Foliar Absorption of Nitrogen Dioxide," M.A. Thesis,
Department of Botany, University of Utah.
Trijonis, J., 1978: "Empirical Relationship Between Atmospheric Nitrogen
Dioxide and Its Precursors," EPA-600/3-78-018 (February).
143
-------
Westberg, H., K. J. Allwine, and E., Robinson, 1978a: "The Transport of
Oxidant Beyond Urban Areas/Light Hydrocarbon and Oxidant Data-New
England Study, 1975," EPA-600/3-78-006 (January).
Westberg, H., K. Allwine, and E. Robinson, 1978b: "Measurement of Light
Hydrocarbons and Oxidant Transport—Houston Study, 1976,"
EPA-600/3-78-062 (July).
Westberg, H., K. Allwine, E., Robinson, and P. Zimmerman, 1978c: "Light
Hydrocarbon and Oxidant Transport Studies in Ohio-1974,"
EPA-600/3-78-007 (January).
White, W.H.,. J.A. Anderson, D.L. Blumenthal, R.B. Husar, N.v. Gillani,
and J.D. Husar, 1976: "Formation and Transport of Secondary Air
Pollutants: Ozone and Aerosols in the St. Louis Urban Plume,"
Science. Vol. 194, pp. 187-189 (October).
144
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-450/4/79-035
I. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Survey of the Role of NOX in Nonurban Ozone Formation
5. REPORT DATE
September 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
J. R. Martinez and H. B. Singh
8. PERFORMING ORGANIZATION REPORT NO.
6780-8
9. PERFORMING ORGANIZATION NAME AND ADDRESS
SRI International
333 Ravenswood Avenue
Menlo Park, California 94025
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2835
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer - Harold G. Richter
16. ABSTRACT
This study surveys and summarizes current knowledge about the role of oxides of
nitrogen (NOX) in the formation of ozone (03) in nonurban areas. Project elements
include a literature review, a survey of expert opinion, and analyses of field data
The results of the study show that present knowledge about NOX/03 interactions in
nonurban areas is fragmentary, and that there is no direct quantitative description
of the link between urban NOX and nonurban 03.
A preliminary analysis of nine rural and suburban sites indicates that transport is
the principle mechanism associated with ozone levels that exceed 100 ppb. The
analysis further suggests that transport times of the order of six to eight hours
are most likely to be associated with the high ozone observed at these sites.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Nitrogen Dioxide
Ozone Formation from N02
Nitrogen Dioxide Transport
Rural Ozone
18. DISTRIBUTION STATEMEN1
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
158
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
------- |