United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-79-034a
August 1979
Air
Analysis of High NO2
Concentrations in
1975-1977
-------
EPA-450/4-79-034a
Analysis of High NO2 Concentrations
in California, 1975-1977
by
J.R. Martinez and K.C. Nitz
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
August 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-450/4-79-034a
11
-------
ABSTRACT
During the period 1975-1977, 51 monitoring stations in California
collectively recorded about 1,800 site-days in which hourly nitrogen
dioxide (NC^) concentrations exceeded 0.20 ppm. This work investigates
potential causes of these high N02 events, the physical phenomena
involved in their occurrence, and their spatial and temporal patterns.
In addition, the potential association between emission sources and the
frequency and magnitude of high NC^ levels at the various locations is
analyzed using detailed site-description data compiled in this study.
The relationship between annual maximum hourly levels and annual mean
concentration is explored, and the quality of the NCU data is evaluated.
This report was submitted in fulfillment of Contract 68-02-2835
under the sponsorship of the U.S. Environmental Protection Agency.
iii
-------
CONTENTS
ABSTRACT ill
LIST OF ILLUSTRATIONS vii
LIST OF TABLES ix
ACKNOWLEDGMENTS xi
EXECUTIVE SUMMARY xiii
I INTRODUCTION 1
A* Obj ectives and Background 1
B. Methodology 1
C. Organization of the Report 2
II ANALYSIS OF MONITORING STATIONS 3
A. Geographic Setting 3
B. Description and Classification
of Monitoring Sites 10
C. Potential Source/Receptor Links
in the Los Angeles Area 15
III EVALUATION OF DATA QUALITY 19
A. Data Anomalies 19
B. Measurement Accuracy 22
C. Experimental Methods 23
IV ANALYSIS OF PHYSICAL PROCESSES 31
A. Definition of Physical Phenomena 31
B. Examples of Physical Processes 32
1. Chemical Synthesis 32
2. Titration of NO and 03 36
3. Transport and Point-Source Impacts 41
C. Frequency and Spatial Distribution
of Physical Processes.... 41
V SPATIAL AND TEMPORAL PATTERNS 53
A. Spatial Distribution of High N02 Concentrations 53
B. Seasonal and Diurnal Variations 59
-------
D. Association Between NC^-Monitoring Stations
in the South Coast Air Basin 66
VI ANALYSIS OF PEAK/MEAN RELATIONSHIP 75
A. Introduction 75
B. Peak/Mean Relationship 75
VII SUMMARY OF SPECIFIC FINDINGS 83
A. Monitoring Site Features 83
B. Data Quality 83
C. Physical Processes Linked to N02 Exceedances 83
D. Spatial and Temporal Variation
of NO2 Exceedances 84
E. Peak/Mean Relationship 84
VIII CONCLUSIONS 85
IX RECOMMENDATIONS FOR FURTHER RESEARCH 89
REFERENCES 91
vi
-------
ILLUSTRATIONS
1 Geographical distribution of monitoring sites
recording high NO^ values during 1975-1977 6
2 N02 monitoring sites located in the San Francisco
Bay area « • 7
3 N02 monitoring sites located in the South Coast
Air Basin 8
4 N02 monitoring sites located in the San Diego area 9
5 Sample form for site description data 11
6 Sample map of monitoring site location and its
surroundings • 12
7 Map of stationary-source NO emissions
in the South Coast Air Basin 16
8 Map of statiionary-source SOX emissions
in the South Coast Air Basin 17
9 Scatter diagram of colorimetric as a function
of chemiluminescent N02 at Riverside 25
10 Scatter diagram of colorimetric as a function
of chemiluminescent N02 at San Jose 26
11 Scatter diagram of colorimetric as a function
of chemiluminescent N02 at Upland 27
12 Example of N02 formation by chemical synthesis
at the Los Angeles/San Pedro St. station
on 6 April 1977 33
13 Example of N02 formation by chemical synthesis
at the Temple City station on 8 February 1977 34
14 Example of N02 formation by chemical synthesis
at the Temple City station on 7 December 1977 35
15 Example of synthesis and possible point-source influence
on N02 at Whittier on 28 August 1975 37
16 Example of N02 formation on Sunday and Monday,
14-15 March 1976, at the Los Angeles/Westwood station 38
17 Example of N02 formation by titration
at the San Diego/Island Avenue station on 31 October 1976. 39
18 Example of N02 formation by titration
at the Temple City station on 5 December 1977 40
19 Example of N02 transport at El Toro- station
on 25 January 1975 42
20 Example of potential point source impact at Long Beach
on Saturday, 15 November 1975 43
vii
-------
21 Example of potential point-source impact at Long Beach
on Sunday, 16 November 1975 44
22 Example of potential point-source impact at Whittier
on Monday, 25 August 1975 45
23 Physical processes in high N02 events
classified by site 47
24 Histogram of the frequency distribution
of N02 exceedances aggregated statewide 54
25 Spatial distribution of N©2 concentrations
exceeding 0. 20 ppm 56
26 Diurnal and seasonal variation of N02 exceedance
frequency 60
27 Monthly variation of NO emissions for the period
July 1972-June 1973 for the South Coast Air Basin 65
28 Quarterly variation of NOX emissions
in the South Coast Air Basin of California 67
29 Scatter diagram of annual mean as a function
of maximum NO2 for selected California sites
in 1975 76
30 Scatter diagram of annual mean as a function
of maximum N02 for selected California sites
in 1976 77
31 Scatter diagram of annual mean as a function
of maximum NO2 for selected California sites
in 1977 78
32 Scatter diagram of annual mean as a function
of maximum N02 for selected California sites,
1975-1977 79
33 Frequency distribution of peak/mean ratio of N02
for selected California sites, 1975-1977 81
viii
-------
TABLES
1 California Monitoring Stations Recording High N02 4
2 Classification of N02 Monitoring Sites 14
3 Anomalous NC>2 Data • • 21
4 Monitoring Stations Using Chemiluminescent
Instruments 23
5 Number of High NOo Days Detected
by Two Different Measurement Methods 24
6 Correlation Between Measured Colorimetric
and Chemiluminescent N02 Concentrations 25
7 Data Clusters for Riverside 28
8 Distribution of N02 Exceedances Aggregated
for all Monitoring Stations, 1975-1977 50
9 Effect of Adjusting N02 Data 53
10 Reduction of N02 Exceedance Frequency
at Individual Stations after Data Adjustment 57
11 Comparison of Frequency Distributions of N02 Levels
Occurring in Different Time Intervals 63
12 Quarterly Factors for NOX Emissions
in the South Coast Air Basin 67
13 Estimated 1976 NOX 'Emissions
in the South Coast Air Basin 68
14 Quarterly Variation of NOX Emissions
in the South Coast Air Basin for 1976 68
15 Distribution of 15 Selected Stations
in the South Coast Air Basin
Recording N02 Exceedances the Same Day 69
16 N02 Exceedances Recorded at Two or More
South Coast Air Basin Sites 71
17 Contingency Tables and Measures of Association
Between Selected Stations in the South Coast Air Basin.... 73
18 Parameters of Peak/Mean Regression 75
ix
-------
ACKNOWLEDGMENTS
The authors thank Dr. Harold G. Richter, EPA Project Officer, for
his assistance and encouragement in the preparation of this report.
We also thank Mr- Russell Trudeau and Ms. Joyce Kealoha of SRI for
technical contributions.
xi
-------
EXECUTIVE SUMMARY
Objectives and Background
The objective of this study is to analyze the occurrence and poten-
tial causes of hourly concentrations of nitrogen dioxide (IK^) greater
than 0.20 parts per million (ppm) recorded in California during 1975-
1977. The investigation was performed in support of the current efforts
of the U.S. Environmental Protection Agency to determine whether a new
national short-term ambient air quality standard for N02 should be esta-
blished. (The -current national N02 standard is 0.05 ppm annual arith-
metic mean.)
Summary of Findings
Major topics investigated were:
o Monitoring site features.
o Data quality.
o Physical processes associated with high N02 levels.
o Spatial and temporal variations of high N0£ concentrations.
o Relationship between annual maximum and mean NO2 concentration.
The principal findings can be summarized as follows:
Monitoring Site Features
All the sites reporting N02 > 0.20 ppm (hereafter, N02 levels
greater than 0.20 ppm will be referred to as "NO* exceedances") are
located in urban and suburban areas, with half the sites in the South
Coast Air Basin (Los Angeles, Orange, Riverside and San Bernardino coun-
ties) .
Examination of the local neighborhood of the monitoring stations
revealed that mobile-source emissions predominated at most sites. Local
point sources such as power plants and heavy industries were few in
number. Probable point-source impacts from a steel mill were detected
at one site.
In the South Coast Air Basin, probable source/receptor relation-
ships were identified between stationary NO sources and several moni-
toring stations. In particular, the sites at Long Beach and Whittier
showed the most pronounced stationary—source impacts.
xiii
-------
Data Quality
The investigation of data quality considered data anomalies, accu-
racy, and the comparability of colorimetric and chemiluminescent meas-
urements. Few anomalies were found in the data. Only 62 of approxi-
mately 1800 site-days were judged to contain anomalous N02 measurements
and were eliminated from the data base. This resulted in the deletion
of two of the 51 stations originally reporting NOo exceedances. A third
station was also eliminated because its N02 levels did not exceed 0.20
ppm.
Regarding data accuracy, a recent study by the California Air
Resources Board found that because of calibration procedures used in the
state, actual N02 levels in California are from 10 to 17 percent lower
than measured.
Simultaneous colorimetric and chemiluminescent N02 measurements
were compared at three sites. The two types of measurements were corre-
lated at two sites, with correlation coefficients of 0.71 and 0.86, but
no correlation was evident at the third station. In general, the chemi-
luminescent measurements tended to be higher than the colorimetric
observations.
Physical Processes Associated
with High NO2 Levels
Physical phenomena associated with N02 exceedances were classified
into three categories:
o Chemical synthesis—NO ->N02 conversion by peroxy radicals.
o Titration—NO + 03 ->N02 + 02-
o Other—Transport and point-source effects.
Chemical synthesis was found to be the most common mechanism leading to
N02 exceedances. N02 exceedances resulting from titration were about
two-thirds as frequent as those associated with synthesis and appear to
be ozone-limited, rather than nitric oxide-limited. Titration effects
are most common at downwind sites where transported ozone reacts with
local nitric oxide. Pasadena and Pomona in the South Coast Air Basin
are examples of this type of site. Transport and point-source effects
linked to N02 exceedances were infrequent: Examples of these effects are
found at sites such as Long Beach, Whittier, Barstow and Victorville.
Spatial and Temporal Variations
of High NO2 Concentrations
Over 5,400 site-hours with N02 > 0.20 ppm were recorded statewide
during 1975-1977. The concentrations ranged from 0.21 to 0.62 ppm, with
xiv
-------
a median concentration of 0.24 ppm. About 92 percent of the N02
exceedances were recorded in the South Coast Air Basin, with Los Angeles
County accounting for 80 percent of the statewide total. On a statewide
basis, between 34 and 46 percent of the N02 exceedances would be elim-
inated if the N02 data were adjusted downward to compensate for the 10-
17 percent bias.
N02 exceedances occur most frequently during the period November
through February. The seasonal pattern is evident statewide. In the
South Coast Air Basin, the seasonality of the N02 exceedances appears to
coincide with increased contributions from stationary area sources of
NO , contributions which are largely a result of emissions related to
space heating using natural gas.
The diurnal variation of N02 exceedances appears to be associated
with the traffic cycle in the more densely urbanized areas. The distri-
bution of N0~ concentrations exceeding 0.20 ppm that occurs during the
time intervals 0-0500 and 2200-2300 differs quantitatively from the dis-
tributions that prevail during 0600-1300 and 1400-2100 at six of nine
sites tested. In general, N02 levels during 0-0500 and 2200-2300 were
lower than at the other times. By contrast, seven of nine stations
tested showed no statistically significant differences between the dis-
tribution prevailing during 0600-1300 and that for 1400-2100.
A widespread pattern of interstation correlations was found among
fifteen selected stations in the South Coast Air Basin. The incidence
of high N02 levels recorded at two or more sites the same day surpassed
that of single-site events by better than a 2:1 margin. The pattern of
interstation correlations is consistent with the typical flows that pre-
vail in the South Coast Mr Basin.
Relationship Between Annual Maximum
and Mean NQ_2 Concentration
Peak/mean ratios ranged from 3.3 to 12.9, with a median ratio of
6.3. Peak and mean N02 were found to be linearly correlated. The
correlation was highly statistically significant, the coefficient being
0.82 for the pooled 1975-1977 data. A linear regression equation relat-
ing peak and mean N02 was derived. The equation is: peak =5.5 mean +
0.05, where the concentrations are in ppm and the standard error of
estimate is 0.07 ppm.
Conclusions
In California, high N02 concentrations seem to be primarily an
urban phenomenon. The pattern of N02 exceedances strongly suggests that
a high density of emissions of NOX and hydrocarbons are required for
synthesis to lead to levels of N02 > 0.20 ppm. Titration-related N02
xv
-------
exceedances require high levels of both NO and 03 and appear to be
ozone-limited, rather than NO-limited.
The evidence suggests that N02 is not transported over long dis-
tances and that N02 exceedances are essentially confined to urban areas
and their immediate surroundings. The finding that nighttime levels of
N02 are significantly lower than daytime values indicates that N02 has a
short lifetime, and supports the hypothesis that, unlike 03, high levels
of N02 do not undergo long distance transport.
N02 exceedances occurred most frequently during the period
November-February. Although the more stagnant meteorological conditions
that prevail in California during these months are certainly an impor-
tant contributing factor, the possibility exists that the high N02 lev-
els may be enhanced by increased NO emissions from space heating using
natural gas. The seasonal pattern of NO emissions in the South Coast
Air Basin supports this hypothesis.
The recent discovery by the California Air Resources Board that N02
concentrations in the state are between 10 and 17 percent lower than
indicated by the measurements is particularly important for regulatory
applications. Contemplated regulatory actions must specify the adjust-
ment factor to be applied to the data.
The interstation correlation with respect to same-day N02
exceedances between various pairs of sites in the South Coast Air Basin
suggests that area sources, rather than point sources, are the principal
proximate causes of the elevated N02 levels in the area. This is sup-
ported by the association between the daily traffic cycle and the hourly
fluctuations of the high N02 concentrations.
The derived linear-regression equation relating peak and mean N02
implies that the current California hourly standard of 0.25 ppm is more
restrictive than the national annual standard of 0.05 ppm. Conse-
quently, if a national hourly standard of 0.25 ppm were established, the
hourly rather than the annual standard could become the controlling fac-
tor in abatement efforts. This is the situation that currently exists
in California. Thus, in setting a national hourly standard for N02, the
relationship between peak and mean N02 needs to be considered to ensure
that the two standards reinforce each other.
xvi
-------
I INTRODUCTION
A. Objectives and Background
This study investigates the occurrence and potential causes of
hourly concentrations of nitrogen dioxide (IK^) greater than 0.20 parts
per million (ppm) that were recorded in California during 1975-1977.
The analysis includes examining spatial and temporal patterns of high
NO2 levels, as well as various physical phenomena associated with their
occurrence.
The investigation is motivated by the current efforts of the U.S.
Environmental Protection Agency (EPA) to determine whether a new
national short-term ambient air quality standard for N0£ should be esta-
blished (the current N0« standard is 0.05 ppm annual arithmetic mean):
hence, the emphasis on high hourly NO 2 concentrations. Various concen-
tration thresholds ranging from 0.10 to 0.50 ppm have been mentioned in
connection with an hourly W^ standard (Thuillier and Viezee, 1978);*
this study concerns N02 levels that exceed 0.20 ppm. For comparison, it
should be noted that California currently has an hourly N02 air quality
standard with a threshold level of 0.25 ppm.
B. Methodology
The study examined the following aspects of the problem:
• Characteristics of monitoring sites recording high N0« levels—
the local environment of individual monitoring sites was scru-
tinized to identify factors that may influence the occurrence of
high NO2 levels.
• Quality of the N02 data—the study considered the accuracy of
the measurements and the possible presence of anomalous values
in the data base.
• Physical processes associated with high N02 concentrations—the
primary processes that lead to the high levels of N02» a secon-
dary pollutant, were identified.
• Spatial and temporal patterns of occurrences of high N02
levels—the spatial distribution of high-N02 events and their
seasonal and diurnal variations were examined.
• Relationship between annual maximum hourly values and mean N0«
concentrations—to estimate the potential effect of control
strategies on peak N02 levels, the study derived an equation
that relates the annual NO2 peak to the mean concentration.
References are listed at the end of this report
-------
The text of this report describes in detail the analytical methods
used in the research areas outlined above and presents the conclusions
that were developed from the study.
C. Organization of the Report
Section II describes the geographical coverage of the monitoring
sites that recorded N(>2 levels exceeding 0.20 ppm and contains the
analysis of the site-description data. Data quality considerations are
examined in Section III; Section IV presents the results of the analysis
of physical processes. Spatial and temporal patterns of high N02 events
are discussed in Section V. Section VI describes the investigation of
peak/mean relationships. Conclusions and recommendations are presented
in Section VII.
Two separately bound appendices are part.of this report. Appendix
A contains plots of N02 and other pollutants for all the days with N02 >
0.20 ppm. Appendix B contains detailed site-description data for all
monitoring stations, including site maps.
-------
II ANALYSIS OF MONITORING STATIONS
A. Geographic Setting
During 1975-1977, 51 monitoring stations distributed throughout
California reported at least one hour when N02 exceeded 0.20 ppm.* Table
1 lists the stations, including SAROAD code number, site name, county,
and street address. The sites have been numbered sequentially from 1
through 51 for ease of reference; site numbers are shown in the first
column of Table 1.
Figures 1-4 show the geographical area covered by the 51 stations.
The statewide distribution of the sites is displayed in Figure 1 (site
numbers are keyed to Table 1). Only a few stations have been identified
in Figure 1 because of space limitations; the other sites are identified
in Figures 2-4. Figure 1 shows that the majority of the sites are
clustered around three metropolitan areas:
(1) The San Francisco Bay area, in the north
(2) The Los Angeles area, in the south
(3) The San Diego area, near the southern border of the state.
(It is hardly surprising that most of the sites of interest are located
around the three population centers in the state.) Seven sites are
sprinkled throughout central California, ranging from Sacramento (Site
37) in the north to Barstow (Site 4) and Victorville (Site 50) in the
south/central desert. Three other stations (Sites 7, 31, and 43) are
located near the coast, northwest of Los Angeles.
Figures 2 through 4 provide enlarged views of the disposition of
the monitoring sites near San Francisco (Figure 2), Los Angeles (Figure
3), and San Diego (Figure 4); nine sites are found in the San Francisco
Bay area, 26 in the Los Angeles area (including Orange, Riverside and
San Bernardino counties), and six in the San Diego area.
The maps (Figures 1-4) provide a broad picture of the placement of
the monitoring stations reporting N02 exceedances. From these illustra-
tions, it may appear that few, if any, parts of the state are immune
from high N02 levels. However, our research shows that N02 exceedances
are by no means uniformly distributed geographically. In fact, over 90
percent of all the N02 exceedances recorded during 1975-1977 occurred in
the Los Angeles area. The data review (see Section III) revealed that
three sites (Sites 3, 25, and 31) did not experience any N02
exceedances, which leaves some gaps in the N02 distribution shown in
Hereafter, the term "N02 exceedance" will be used exclusively to denote
hours with N02 > 0.20 ppm.
-------
Table 1
CALIFORNIA MONITORING STATIONS RECORDING HIGH N02
Site
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
SAROAD Code
050230001101
050500002101
050520003F01
050580001101
050900002101
050920002101
051030003101
051300001101
051360001101
051600001101
051740001101
052220002101
052390001101
052460002101
052680001101
052780001101
052800005F01
053620001101
053900001101
054100002101
054180001101
054180002101
054200001101
054260001101
05458000 1F01
Name
Anaheim
Azusa
Bakers fie Id
Bar stow
Burbank
Burl ing ame
Camarillo
Chino
Chula Vista
Concord
Costa Mesa
El Cajon
El Toro
Escondido
Fontana
Fremont
Fresno
La Habra
Lennox
Long Beach
Los Angeles
Los Angeles
Los Angeles
Lynwood
Merced
County
Orange
Los Angeles
Kern
San Bernardino
Los Angeles
San Mateo
Ventura
San Bernardino
San Diego
Contra Costa
Orange
San Diego
Orange
San Diego
San Bernardino
Alameda
Fresno
Orange
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Merced
Address
1010 S. Harbor Blvd.
803 Lor en Ave.
225 Chester Ave.
200 E. Buena Vista
228 W. Palm
1229 Burlingame
Elm Dr.
Central & Riverside
80 E. "J" St.
991 Treat Blvd.
2631 Harbor Blvd.
110 E. Lexington
23022 El Toro Rd.
600 E. Valley Pkwy.
14838 Foothill Blvd.
40733 Chapel Way
3250 E. Olive
621 W. Lambert
11408 La Cienega
3648 N. Long Beach
434 S. San Pedro
2351 Westwood Blvd.
18330 Gault St.
11220 Long Beach
Eighteenth & "S" St.
-------
Table 1 (Concluded)
Site
No.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
SAROAD Code
055120001101
055300004F01
055320003101
055760004101
056040001101
056080001101
056200001101
056240001101
056300003101
056400003F01
056535001101
056580003F01
056680001101
056800004101
056800006101
056860003101
056980004A05
057200004F01
057670001101
058040002F01
058080001101
05822000 1F01
058440003101
058440004F01
058510001101
058720001101
Name
Newhall
Oakland
Oceanside
Pasadena
Pomona
Port Hueneme
Red lands
Redwood City
Richmond
Riverside
Rubidoux
Sacramento
San Bernardino
San Diego
San Diego
San Francisco
San Jose
Santa Barbara
Simi Valley
Stockton
Sunnyva le
Temple City
Upland
Upland
Victorville
Whittier
County
Los Angeles
Alameda
San Diego
Los Angeles
Los Angeles
Ventura
San Bernardino
San Mateo
Contra Costa
Riverside
Riverside
Sacramento
San Bernardino
San Diego
San Diego
San Francisco
Santa Clara
Santa Barbara
Ventura
San Joaquin
Santa Clara
Los Angeles
San Bernardino
San Bernardino
San Bernardino
Los Angeles
Address
24811 San Fernando
Jackson St.
100 S. Cleveland
1196 E. Walnut St.
924 N. Garey Ave.
Naval Civil Eng. Lab.
216 Brookside Ave.
897 Barren Ave.
1144 Thirteenth St.
7002 Magnolia Ave.
5888 Mission Blvd.
1025 "P" St.
172 W. Third St.
1111 Island Ave.
5555 Overland Ave.
(Kearney /Mesa)
939 Ellis St.
1208 N. Fourth St.
831 State St.
5400 Cochran St.
1601 E. Hazelton
251 S. Murphy Ave.
Las Tunas Dr.
155 "D" St.
1350 San Bernardino
1556a Eighth St.
(County Bldg. 1557)
14427 Leffingwell
-------
-114
Figure 1. Geographical distribution of monitoring sites recordmg
high NQj values during 1975-1977.
Site numbers are keyed to Table 1.
-------
38.4 —
37.0
-123
Figure 2. NO2 monitoring sites located in the San Francisco Bay Area.
She numbers are keyed to Table 1.
-------
34.4
• i
26
34.2
34.0
o
l-
33.8
00
33.6
33.4
-11.9
• 4
• 23
.5
29<
• 2
•47
• 21
18
I
30
48^.49
•8
15 38
36
35
*
•118
-118
LONGITUDE
-117
32
Figure 3. NO2 monitoring sites located in the South Coast Air Basin.
(Los Angeles, Orange, Riverside and San Bernardino Counties).
Site numbers are keyed to Table 1.
-117
-------
33.2
33.1
33.0
g 32.9
32.8
32.7
32.6
32.5
I ' I
I ' I ' I ' I ' I
I
-117
117
117
•116
LONGITUDE
-116
-116
-116
Figure 4. N02 monitoring sites located in the San Diego Area.
Site numbers are keyed to Table 1.
-------
Figure 1. As a result, N02 exceedances seem to be most closely associ-
ated with areas that have a high population density, and appear to be an
essentially urban phenomenon.* More evidence supporting this hypothesis
is presented later in this report.
What sources influence the occurrence of high NC^ levels at these
monitoring stations? The next sections describe the analysis of the
immediate surroundings of individual stations in order to identify
potentially significant local sources, and examine potential
source/receptor relationships in the Los Angeles area on a regional
scale.
B. Description and Classification of Monitoring Sites
Information describing each monitoring site and its surroundings
was tabulated to detect local factors that might influence the
occurrence of NC^ exceedances at each site. For example, the presence
of a power plant or a busy highway near the site would be of special
interest.
The site-description data were obtained from the California Air
Resources Board (GARB). For the most part, these data were provided to
the GARB by the local air-pollution control agencies. The site data
were compiled and supplemented by SRI using forms similar to those shown
in Figure 5. In many cases, the available data were incomplete, and all
the blanks in the form could not be filled. This was especially true of
the traffic data for most sites except those located in the Los Angeles
area.
In addition to the written data, a map showing the station location
and its surroundings (within a five mile radius) was prepared for each
site. An example is shown in Figure 6 for Anaheim; the dot at the
center of the circle marks the station location. The maps provide a
visual indication of the site's environment. It is apparent that the
Anaheim station is surrounded by roads, with no significant point
sources nearby: Disneyland is essentially a large parking lot, which
contributes to the mobile-source emissions from the nearby roads. It is
clear that traffic-related emissions should be the major influence at
this site.
The site-description forms and associated maps for all 51 stations
are included as Appendix B of this report.
*Note that there may also be gaps in the monitoring network.
10
-------
tin IWOUUTION ro« KOI nisi
»• nn mtimricATioii
I. 0.1.
*. ci«»
j. lute
». Let/too*, ct Dm cooce'lute*
». UtOAO ilte coee
t. otete (lie
T. loc«l ilte i
I. t*m»i of aurnt l«erMCttm
•Ireel•:
10. t\o(ck of elte; ootoil:
<•) Mortk 4UMIIM III
(k) UUt peak* •
(c) Ok>lroctlo«i
(I) tcrki* *r*««
(l) tuiot «ru«
(b) >all4t«t« (ladl
•o. ol »torl»)
I. K>UOtid ilttuc* I* •tto-»*ocl«t«4 >>IUtMti.
dot* tttt con»U«t (Untoi Mvrctt vltkli I •!!•• ol
I)M tit*. CodUir ill (Uottd foUt *MIC« Mltttaa
•on tk*t 1500 toiu/rr. ». «ltfcU IB •!!•• of iMltor.
lourco
DIctctlOT Dliupct foU«t«»t«
I. Typ> ol tccllU M«r th« ilt'o:
Iioony
f «kvor
N>Jor ittoo
loeil «t
IMOCMCtl
J. Wi« U •v>[>(>
*. »i«« to «»«r«t«
hlcU (poolf ____^
>. Mut *» nu^Mr ol tr.fflc UMO ol tb>
». Wui «r« MMkor el H'^t'V '•»•• °' th* »<>^»«T
7. T«cril« checicteiiltlci:
llootk _
lough
Figure 5. Sample form for site description data.
-------
SITE NO. 050230001:
1010 S. HARBOR BOULEVARD
ANAHEIM, CALIFORNIA
Figure 6. Sample map of monitoring site location and its surroundings.
Dot marks station location.
12
-------
Using the compiled site descriptions, the monitoring sites were
classified under various categories:
• Site type—Urban, suburban, or rural
• Site representativeness—Street canyon, neighborhood
• Traffic type
- Heavy (more than 24,000 vehicles/day)
- Medium (6,000-24,000 vehicles/day)
- Light (less than 6,000 vehicles/day)
- Intersection
• Stationary sources—Power plant or miscellaneous .
The classification of all the sites is shown in Table 2.
Table 2 shows that all sites are located in urban or suburban
environments. This could imply that rural areas are not impacted by
high N02 levels, but this inference must be tempered by the knowledge
that rural locations are seldom monitored. The data may merely reflect
the fact that the monitors are found in urban and suburban areas rather
than in rural regions. Consequently, judgments about the prevalence of
high N02 in rural areas must be withheld pending further study.
Nevertheless, Table 2 makes it clear that high N(>2 levels are definitely
associated with the conditions that exist in urbanized areas.
The table also shows that a number of monitoring stations are *
located in street canyons; hence, N02 levels should be directly tied to
traffic influences. The "neighborhood" classification denotes commer-
cial or residential areas that may be near large point sources, so
"neighborhood" sites should reflect the ebb and flow of traffic as well
as any point-source impacts.
Five sites, all from the Los Angeles area, are located near heavily
traveled roads. Twelve stations fall in the moderate-traffic category,
while nine are classified as light-traffic sites. Twenty sites are
located near intersections with varying degrees of vehicle activity.
All of these sites should be locally affected by mobile-source emis-
sions.
Power plants are located near five of the sites: Burbank, Chula
Vista, San Diego, and the two Upland stations. At the first three
sites, subsequent analysis revealed that pollutant histories on high N02
days do not indicate any impact from stationary sources as shown by S02
levels. This is not surprising, considering that the plant stacks are
designed to prevent precisely such localized impacts (except perhaps
under conditions of fumigation). Another possible explanation for the
apparent lack of power plant impact is that the plants may have used
natural gas on the days investigated. If this were the case, there
would be no S02 effects and the power plant impact would go undetected.
13
-------
Table 2
CLASSIFICATION OF N02 MONITORING SITES
Site Name
Anaheim
Azuss
SAROAD ID
Number
05023O001
050500002
Bakersfleld | 050520003
Barstow
Burb«nk
Burllngame
Camarlllo
Chlno
ChuU Vista
Concord
Costa Mesa
El Cajon
El Toro
Escondldo
Fontana
Fremont
Fresno
La Habra
Lennox
Long Beach
Los Angela*
Los Angeles
Los Angelas
Lynvood
Merced
Nevhall
Oakland
Oceanside
Pasadena
Pomona
Port Hueneme
Red lands
Redwood City
Richmond
Riverside
Rubldoux
Sacramento
San Bernardino
San Diego
San Diego
San Francisco
San Jose
Santa Barbara
Siml Valley
Stockton
Sunnyvale
Temple City
Upland
Upland
Victor-vine
Whlttlar
050580001
050900002
050920002
OS 1030003
051300001
051360001
051600001
051740001
052220002
052390001
052460002
052680001
052780001
052800005
053620001
053900001
054100002
054180001
054180002
054200001
054260001
054580001
055120001
055300004
055320003
055760004
056040001
056080001
056200001
056240001
056300003
056400003
056535001
056580003
056680001
056800004
056800006
056860003
056980004
057200004
057670001
058040002
058080001
058220001
058440003
058440004
058510001
058720001
Site Type*
1
X
X
X
X
A
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
2
X
X
X
X
X
X
X
X
Jb
X
X
X
X
X
X
X
X
X
X
X
X
3
Repre-
senta-
Ivenesst
4
X
X
X
X
!
X
X
X
X
A
5
X
X
K
X
X
X
A
X
X
X
A
X
X
X
X
X
X
X
Traffic Type1
6
X
X
X
X
X
7
X
X
X
X
X
X
X
X
X
X
X
X
8
X
X
X
X
X
X
X
X
X
9
X
X
X
X
A
A
X
X
A
X
X
X
X
X
X
X
X
X
X
X
Stationary
Sources
Power
Plant
X
X
X
X
X
Misc.
X
X
X
X
X
X
X
X
X
X
K
X
Remarks
31,000 veh/day
800 veh/day
6,000 veh/day
Power plant- 1 block;
2,400 veh/day
Freevay-600 ft
7,000 veh/day
Power plant-2.5 ml
48,000 veh/day
Freeway- 0.5 ml
Freeway* 1 mi
Kaiser steel plant
14,000 veh/day
Los Angeles Int'l Airport
25,000 veh/day
14,000 veh/day
21,000 veh/day
2,000 veh/day; freeway- 1.5 ml
16,000 veh/day
12,000 veh/day
RR switchyard- 1/4 mi
18,000 veh/day
24,000 veh/day
Dock machinery and vehicles;
cargo vessels
5,000 veh/day
Paint spray shop-1 block
Refinery- 2 ml
Norton AFB-2 ml; 9,000 veh/day
Tire recapping; power plant- 1 ml
Naval Air Scat ion- 2 mi
Power plant -7 mi
Steel mill-7.5 ml; airport-3 ml;
power plant-7 mi
Cement plane- 2 ml; 2,000 veh/day
20,000 veh/day
*Slte Type: I • Urben, 2 • Suburben, 3 - Rural
'Representativeness: 4 • Street Canyon, 5 - Neighborhood
'Traffic Type: 6 • Heavy, 7 • Moderate, 8 • Light, 9 • Intersection
14
-------
These plants burn either oil or natural gas, depending on availability,
and used oil most frequently during 1975-1977. Consequently, the
natural gas hypothesis is unlikely, but cannot be ruled out. No power
plant effects could be detected at Upland, because SC^ was not monitored
there.
The Fontana and Upland sites would be expected to show some impact
from the Kaiser steel mill. Fontana showed enhanced S02 levels, which
indicates that the steel plant's emissions probably impact the site.
But no effects were evident at the Upland sites because, as noted above,
S02 was not monitored.
Not shown in Table 2, but evident in the data sheets in Appendix B,
is the fact that many of the stations are located near parking lots, gas
stations, and other small sources. These miscellaneous local sources
must affect the pollutant burden measured at the site, but their indivi-
dual contributions are too small and diffuse to be detected in the air
quality data; they become part of the background pollution.
In summary, the site classifications suggest that mobile sources
are the predominant local emitters at nearly all the stations. Poten-
tially significant local point sources such as power plants and heavy
industries were found to be few in number: There are indications of
impacts from a steel mill at one site (Fontana), but possible power-
plant influences at three other sites could not be detected. Finally, a
variety of small stationary sources (e.g., parking lots, gas stations)
undoubtedly contribute to the general pollution level, but individual
impacts are small and indistinguishable from the overall contamination.
C. Potential Source/Receptor Links
in the Los Angeles Area
This section extends the spatial scale from the local level of the
previous section to the regional level, seeking possible source/receptor
links for stationary sources in the Los Angeles regional area. The mon-
itoring sites have been plotted on maps of stationary-source emissions
of sulfur oxides (SOX) and nitrogen oxides (NOX). Examining source
locations, monitoring stations, and known meteorological patterns for
the area may reveal likely source/receptor relationships •
Figures 7 and 8 are gridded emissions maps of NOX and SOX, respec-
tively, from stationary sources; the N02-monitoring stations are plotted
on each map. (Each grid cell is 10 km square.) The NOX data were
obtained from Bartz et al. (1974) and the SOX data from Hunter and
Helgeson (1976). The numbers shown on the perimeter of the grid are
Universal Transverse Mercator (UTM) coordinates in km. The range of
emissions associated with individual grid squares is indicated by dif-
ferent shadings. The dots mark the monitoring site locations, and the
site numbers shown are keyed to Table 1. These figures also show aver-
age streamline patterns that typify those that occur in the interval
15
-------
DAILY NOXEMISSIONS, TONS/DAY
D *-"
n 6-19.9
20*
UTM Coordinates(km)
0 10 20km
3720
3700
Figure 7. Map of stationary-source NOX emissions in the South Coast Air Basin.
Dots mark location of N02 monitoring sites; the site numbers are keyed
to Table 1. Streamline pattern shown is typical of winter afternoon conditions.
-------
DAILY SOX EMISSIONS, TONS AS SOj/DAY
D °^'9
F3 6-19.9
20+
UTM Coordinates (km)
500 520^
3 /
3780
North
3700
' - 1
0 1 0 20 km
Figure 8. Map of stationary-source SOX emissions in the South Coast Air Basin.
Dots mark location of NO2 monitoring sites; the site numbers are keyed
to Table 1. Streamline pattern shown is typical of winter afternoon conditions.
-------
1200-1700 (PST) in January (De Marrais et al., 1965). There is very
little difference between streamline patterns in winter and those for
other times of year.
Figure 7 shows that sites at Long Beach (Site 20) and Rubidoux
(Site 36) are located in grid squares that have a high density of NO
emissions from stationary sources. In addition, the stations at Burbank
(Site 5), Los Angeles/San Pedro (Site 21) and Whittier (Site 51) are in
squares of moderate emissions density. These five sites should experi-
ence the strongest impact from stationary sources of NO. (In fact,
Section IV will show that Long Beach and Whittier exhibit such impacts.)
The streamlines show that many other monitoring stations are downwind of
the major sources, and thus are likely candidates for stationary-source
impacts: examples are Anaheim (Site 1), La Habra (Site 8), Lennox (Site
19), Los Angeles/Gault St. (Site 23), Lynwood (Site 24), and San Bernar-
dino (Site 38). Note also that Whittier (Site 51), in addition to being
in a square of medium emissions density, is downwind of the coastal NO
sources, and is thereby exposed to a multiplicity of sources. It can
also be seen in Figure 7 that the Fontana site (Site 15) is downwind of
squares with high emissions density, consistent with the earlier obser-
vation regarding evidence of stationary-source impact at Fontana.
Since SOX is primarily associated with stationary sources, its
presence identifies areas that may be impacted by such sources. It is
evident from Figures 7 and 8 that grid squares with high NOX and SO
emissions coincide in most cases, enhancing the use of SOX as an indica-
tor of stationary-source effects for NOX, Figure 8 shows that Long
Beach (Site 20) is located in an area of high SOX emissions density and
that Whittier is located in an area of moderate emissions density.
Since Whittier is in a downwind location, as shown in Figure 7, it is
thus likely to receive additional pollutants from the coastal SO
sources. The impacts of SOX emissions will be discussed in greater
detail in Section IV.
The pattern of streamlines in Figure 8 suggests that other sites
may also be influenced by SOX sources: Anaheim (Site 1), Fontana (Site
15), La Habra (Site 18), Lennox (Site 19), Los Angeles/San Pedro (Site
21), and Lynwood (Site 24) are likely candidates. However, except for
Fontana, the levels of SOX observed at these sites were generally low,
indicating that small effects from stationary sources may be the rule.
Fontana did show appreciable levels of S02» (Plots of S02 concentra-
tions for those sites where SO2 was measured are shown in Appendix A of
this report.)
Thus, Figures 7 and 8 provide graphic evidence linking stationary
sources and emissions levels at certain monitoring stations. The flow
field also suggests that levels at many of the stations may be corre-
lated, since several sites lie on the path of the same air trajectories.
This theme is again addressed in Section V, which examines the correla-
tion between station pairs.
18
-------
Ill EVALUATION OF DATA QUALITY
An important concern of the study was to assess the quality of the
NOj data. The evaluation considered:
• Data anomalies
• Measurement accuracy
• Experimental methods*
The procedures used in the data-quality assessment and the results
obtained are described below.
A. Data Anomalies
In this study, detecting anomalous data meant identifying N02
exceedances that could have been erroneous. Thus, no effort was made to
detect errors for N02 concentrations in the range 0-0.20 ppm. It is
emphasized that while a number of N02 exceedances were found to be ques-
tionable and were subsequently excluded from our analyses, the original
data records (which were unavailable to us) must be examined to deter-
mine whether or not these data are in fact invalid.
The error-detection procedure consisted of four steps:
* (1) For each site, plot concentration histories for N02 and other
pollutants for all days that report at least one NO2
exceedance.
(2) Define error-detection criteria.
(3) Visually scan each plot and identify questionable data by
applying the error-detection criteria.
(4) Check the questionable data against the original data record
and accept or reject the data.
The plots generated in Step 1 are found in Appendix A of this
report. In addition to N02, the graphs include NO, Oj, CO, and S02
whenever these data were available. Because of the chemical links
between NO, N02, and Og, the adequacy of the N02 data could be evaluated
in relation to NO and 03. CO and S02 are relatively inert chemically
compared to the others, and can be helpful in judging the effect of
sources and meteorological conditions. Thus, a very high N02 level that
occurs on a day when CO is low can be flagged as suspect, pending
further checking of the data.
Pollutant data were also plotted for the days immediately preceding
and following the date of the N02 exceedance. The extended data record
enriched the context in which the N02 data were judged. For example, if
the N02 instrument was operating during only part of the day when an
19
-------
exceedance occurred, the record of the instrument's behavior on the
preceding and following day provides information about the likelihood of
erratic behavior when the exceedance was recorded.
The following attributes served to detect questionable data:
(1) Sudden increases (spikes) or drops in the NOj data.
(2) High NC>2 concentrations occurring immediately before or after
an instrument malfunction.
(3) Very high NC^ concentrations (apparent outliers).
(4) N(>2 exceedances preceded and followed by gaps in the data
record (isolated exceedances).
(5) Simultaneous existence of high NO and Og.
After verification that the flagged data were real, and not the
cesult of a plotting error, the data were usually rejected if they
satisfied any of the above criteria. However, not all data that satis-
fied the first three criteria were rejected: Such cases were examined in
relation to other pollutants, and a subjective judgment was made about
whether the instrument was working properly. The "data gap" criterion
led to immediate rejection if the length of the data record was short
compared to the length of the gaps; that is, if the degree of isolation
was large. Otherwise, a subjective judgment was made about the likeli-
hood of erratic instrumental behavior, and the data accepted or
rejected. The last criterion always resulted in rejection of the data,
since the coexistence of NO and 0- for any extended period is chemically
untenable. It is possibile that the 0^ instrument, rather than the
NO/N02 equipment, was malfunctioning, but in those few instances when
this criterion was applied, it seemed more likely that the NO/N02
instrument was operating erratically.
In general, these decisions were conservative and data was rejected
more often than not. Yet, even though the net was fine, the catch was
small. Of approximately 1,800 site-days when N02 exceedances were
recorded, only 62 were rejected. Table 3 lists the sites and dates when
data were rejected, along with the applicable rejection criteria. The
plots corresponding to these days are found in Appendix A of this
report. The table shows that the "data gap" criterion was invoked most
frequently, followed by the "spike" and "instrument malfunction" cri-
teria. The third criterion, presence of apparent outliers, was used
infrequently, and always in combination with some other criterion. As
noted in Table 3, some of the outliers are decimal-point errors. The
fifth criterion, the coexistence of high values of NO and Og, was
invoked only six times.
The five error-detection criteria were not strictly applicable to
the rejected data for 9 September 1976 at Chino and for 15 July 1977 at
Fontana. At Chino, the data were rejected because they appear to con-
tinue the pattern of erratic behavior observed on 8 and 10 September.
20
-------
T«ble 3
ANOMALOUS N02 DATA
Sice
Anaheim
Bakersfleld
Barstow
Burbank
Camarilla
China
Cost* Mesa
Fontana
U Habra
Lennox
Long Beach
Lo« Angelee/San Pedro
Lot Angeles /Westvood
Los Angeles /Gault
Lynwood
Merced
Oceanalde
Pasadena
Riverside
Sacramento
San Diego/Island
San Jose
Temple City
Whlttler
Dace
16 Sep 75
23 Apr 77
10 Oct 77
26 Oct 77
3 Jul 76
19 Jul 76
3 Jun 77
22 Mar 76
19 Jan 77
IS Jan 75
3 Sep 76
8 Sep 76
9 Sep 76
10 Sep 76
13 Sep 76
14 Sap 76
15 Sep 76
23 Sep 75
3 Nov 76
27 Aug 76
28 Aug 76
15 Jul 77
13 Jan 76
24 Nov 76
8 Dec 76
2 Dec 75
19 Jan 76
21 Jan 76
19 Apr 76
28 Dec 76
30 Nov 77
17 Oct 75
7 Mar 77
14 Nov 77
27 Feb 76
16 Aug 77
25 Oct 77
26 Oct 77
3 Nov 77
22 New 77
22 Dec 75
27 Jan 76
2 Feb 76
25 Jun 76
19 Jan 77
5 Apr 77
8 Mar 77
16 Oct 75
26 Feb 77
24 Nov 76
8 Feb 77
6 Apr 77
30 Jan 76t
15 Nov 77
31 Dec 77
30 Aug 76
23 Jul 75*
21 Dec 76*
30 Dec 77
24 Aug 76
16 Aug 77
16 Nov 77
Data Rejection Criteria*
I
X
X
X
X
X
X
X
X
Jl
X
X
X
X
X
X
X
X
X
X
2
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
3
X
X
X
X
X
4
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
5
X
X
X
X
X
X
Remarks
Apparent decimal-point error
Apparent decimal-point error
Erratic behavior
Strange NO patterns
Apparent decimal-point error
Apparent decimal-point error
Apparent decimal-point error
I
Data spike, 2 - Instrument malfunction, 3 - Outliers, 4 - Data gaps, 5 - Simultaneous NO and 03
Colorimetrie method
Chemiluminescent method
21
-------
At Fontana, the data for 15 July were rejected because they duplicate
the pattern observed on 27 and 28 August, when the fifth criterion was
invoked. However, the ozone data were missing on 15 July, and the fifth
criterion was not strictly applicable. The Fontana data show strong
indications of point-source impacts. Thus, what appears to be a
strange, and possibly erroneous, pattern in the data could be an accu-
rate reflection of actual conditions. In the absence of additional
information, the data were rejected.
We consider a data-rejection rate of about 3 percent to be quite
low, and indicative of the fact that the data underwent extensive edit-
ing and checking before being entered in the data bank. Therefore, we
conclude that the presence of anomalously high N02 concentrations is not
cause for concern in subsequent analyses.
The data check resulted in eliminating three stations from the list
of those reporting NOj exceedances: Bakersfield, Merced, and Port
Hueneme. The exceedances recorded at Bakersfield and Merced were few,
and all but one appear to be decimal-point errors. The N02 data for
Port Hueneme displayed no anomalies: NOo at Port Hueneme equalled but
did not exceed 0.20 ppm.
B. Measurement Accuracy
The term "accuracy" refers to systematic errors or bias in the
measurements. To determine the accuracy of experimental data, observa-
tions must be compared with some standard or reference level. Since
only field data were available in this study, an independent estimate of
the accuracy of the NO2 data was not possible. However, a recent study
by the GARB (GARB, 1979) provides the desired information about the
accuracy of N02 data in California.
The principal finding of the GARB study is that the so-called
"Saltzman factor", which characterizes the efficiency of the NO- conver-
sion, is higher than has been assumed in California experimental proto-
cols. The factor was assumed to be 0.72, but the GARB found it to be
10-17 percent higher. This implies that actual N02 levels in California
are from 10-17 percent lower than indicated by the measurements. These
results apply to both colorimetric and chemiluminescent instruments,
because both use the same calibration procedure in California.
It is emphasized that these findings apply only to pre-1980 N02
data for California. Similar studies should be performed by other
states that have used the Saltzman procedure for calibration to deter-
mine what, if any, error may be present under their particular experi-
mental conditions. Accuracy problems caused by using the Saltzman pro-
cedure will be eliminated in 1980 and thereafter because EPA regulations
require that other calibration methods be used beginning in January 1980
(Federal Register, 10 May 1979).
22
-------
In this study, the N02 data have not been adjusted in any way.
However, the implications of the 10-17 percent bias are discussed in the
analyses that follow.
C. Experimental Methods
The N02 data used in this study were obtained using colorimetric
and chemiluminescent instruments. Colorimetric instruments were used at
30 stations. Table 4 lists the 21 sites where chemiluminescent instru-
ments were used, and the years when the instrument was in operation and
recorded N02 exceedances. Since two different kinds of instrument were
Table 4
MONITORING STATIONS
USING CHEMILUMINESCENT INSTRUMENTS
Site
No.
3
8
9
12
13
14
15
17
25
27
28
31
35
36
39
40
42
44
47
49
Name
Bakers fie Id
Chino
Chula Vista
El Cajon
El Toro
Escondido
Fontana
Fresno
Merced
Oakland
Oceans ide
Port Hueneme
Riverside
Rubidoux
San Diego
San Diego
San Jose
Simi Valley
Temple City
Up land
Years*
1977
1976-1977
1975-1977
1976-1977
1975
1975-1977
1975-1977
1976-1977
1977
1975-1976
1975-1976
1977
1975-1977
1976-1977
1975-1977
1975-1977
1975-1977
1977
1976-1977
1975-1977
%
Only those years reporting N02
exceedances are shown.
23
-------
used, it is of interest to examine the comparability of the data
obtained by the two methods. Such analysis is possible since both kinds
of instrument were operated simultaneously at four stations. Table 5
lists the sites and shows the number of days when NCK exceedances were
detected by each method. The table suggests some discrepancy between
the two types of measurement.
Table 5
NUMBER OF HIGH NO2 DAYS DETECTED
BY TWO DIFFERENT MEASUREMENT METHODS
No.
35
36
42
49
Name
Riverside
Rubidoux
San Jose
Upland
SAROAD
Code
056400003
056535001
056980004
058440004
Measurement
Method
Colorimetric
Chemi luminescent
Colorimetric
Chemi luminescent
Colorimetric
Chemi luminescent
Colorimetric
Chemi luminescent
Number of Days
with N02 > 0.20 ppm
1975
13
13
2
N.A.
4
5
7
9
1976
3
14
2
0
6
5
2
5
1977
N.A.*
8
1
4
2
7
N.A.
5
N.A. = Data not available for specified method.
A correlation analysis was performed to examine the relationship
between the N0£ concentrations measured by the two techniques. The data
used in the analysis were the paired concentrations for the hour when
either method measured N©2 > 0.20 ppm. The three possible types of
concentration-pairs are represented by AB, AB, and AB, where A_- _
Colorimetric > 0.20 ppm, B = chemiluminescent > 0.20 ppm, and A and B
denote the complement. The results of the analysis are described below.
No analysis was performed for Rubidoux because the number of concentra-
tion pairs was too small (one).
Figures 9 through 11 are scatterplots of Colorimetric concentra-
tions as a function of chemiluminescent concentrations for Riverside,
San Jose, and Upland. The correlation statistics are shown in Table 6.
As the table indicates, only Upland shows the expected high correlation.
The correlation for Riverside is low but statistically significant, but
Sari Jose shows no significant correlation.
Figure 9 shows that the Riverside data contain two clusters,
labeled A and B, that dominate the correlation. Table 7 shows the
24
-------
ho
u»
.30
.28
.20
.22
.20
|
1
0«.17
Z
o
Su .IB
1
g
8 .12
.10
.07
.00 1
+
•
X^"\B
/ *}
i . y
i ^s
*^_^*^
f
.11 .13 .16 .18
* *
*
« <
3 2
• • * *
• 22*
2
»
* 4 *
3
>
•
•
+ w+ + + + +
.21 .23 .20
* H
• * *
2
2 >
«
•
«
•
i
H
.28 .30 .33
h
h
•
•
•
•
•
h
•
•
K
IB
CHEMILUMINESCENT NO2 - ppm
Figure 9. Scatter diagram of colorimetric as a function of chemiluminescent NO2 at Riverside.
Number of points plotted is 53. See text for explanation of the data labeled A and B.
-------
N>
.26 +
.'26
.88 +
.23 +
.22 +
- .20 *
1
E
8
.19 •<•
• 2
• I <
._-+ + + +..
+
• • *
•
.17
,16
.IB
.13 +
.+ + -f + + * + «• -+ + + + + + + + + + + + +.
.17 .16 .19 .20 .21 .22 .23 .24 .28 .26 .27
CHEMILUMINESCENT NO2 - ppm
Figure 10. Scatter diagram of colorimetric as a function of chemiluminescent NO2 at San Jose.
Number of points plotted is 39.
-------
to
.+ + + + + + + + •¥ + + + + + + + + + + + +.
.26 + *+
.25
.23
.22
u
7 .20
K
o
8
.16
.17 2
,16
,14
. 13 +«
.+ ----
.21
*
*
1
.22
.23 .24 .28 .26 .26
CHEMILUMINESCENT N02 - ppm
+ ---- + ---- 1 ---- + ---- + ---- +
.27 .26 .29
+
+ .
.30
Figure 11. Scatter diagram of colorimetric as a function of chemiluminescent NO2 at Upland.
Number of points plotted is 22.
-------
Table 6
CORRELATION BETWEEN MEASURED COLORIMETRIC
AND CHEMILUMINESCENT N02 CONCENTRATIONS
Site
Name
Riverside
San Jose
Up land
Number of
Data Points
53
39
22
Correlation
Coefficient
0.53
-0.12
0.86
Significance
Level
< 0.006
N.S.*
<0.0001
Not statistically significant
Table 7
DATA CLUSTERS FOR RIVERSIDE
Cluster
A
A
B
B
B
B
Colorimetric
NO 2 (ppm)
0.08
0.05
0.23
0.24
0.25
0.22
Chemi luminescent
NO 2 (ppm)
0.21
0.21
0.13
0.14
0.12
0.11
Date
2 Feb 1976
2 Feb 1976
26 Feb 1976
26 Feb 1976
26 Feb 1976
26 Feb 1976
Hour (PST)
0100
0600
1700
1800
1900
2000
28
-------
cluster contents and the dates and hours that correspond to the data in
the clusters. The table shows that both points in Cluster A were
observed on the same day: 2 February 1976. The situation is similar for
Cluster B: these four points were measured on 26 February 1976. Perusal
of the data record for 2 February reveals some gaps in the chemilumines-
cent data, which suggests that this instrument may have behaved errati-
cally that day. On 26 February the colorimetric measurement increased
abruptly from 0.12 ppra at 1600 to 0.23 ppm at 1700, the level remaining
as shown in Table 7 until 2100 when it decreased to 0.19 ppm. By con-
trast, the chemiluminescent data remain steady throughout. The evidence
suggests, therefore, that Clusters A and B represent anomalous cases
that may be reasonably eliminated from the data comparison.
In view of the above, Clusters A and B were eliminated and the
correlation was recomputed. The new correlation, 0.71, is statistically
significant (p < 0.0001); it is also in line with the correlation seen
at Upland. The regression line for the modified data set is y = 0.83x +
0.001, where y » colorimetric N02 and x » chemiluminescent N02« The
standard error of estimate for this regression is 0.03 ppm. The slope
of the line indicates that, on the average, colorimetric N02 is 17 per-
cent lower than chemiluminescent N02» when the latter exceeds 0.20 ppm.
The remarkable agreement between this 17 percent factor and that
discovered by the GARB is purely coincidental.
The diagram for San Jose (Figure 10) shows the large amount of data
scatter that led to the insignificant correlation between the two types
of measurement. In contrast to Riverside, the San Jose data show no
pattern or data clusters that exert undue influence on the correlation.
The only plausible conclusions seem to be that the instruments were
either measuring different things or were improperly operated, or both.
The behavior of the Upland data is displayed in Figure 11. It is
apparent that a linear relationship exists between the two quantities.
The regression line is y » -0.10 + 1.19x, where y and x are colorimetric
and chemiluminescent N02, respectively, and the standard error of esti-
mate is 0.02 ppm. The figure shows that colorimetric levels are always
lower than chemiluminescent measurements, yet the slope of the line is
greater than one. The relatively large negative intercept compensates
for the slope and indicates that the two data types are offset by a sig-
nificantly high constant difference in favor of chemiluminescent NC>2»
Thus, the chemiluminescent instrument reports NO 2 exceedances more fre-
quently than the colorimetric instrument. Differences in zero-setting
procedures could account for the offset.
In summary, two of the three sites analyzed (Riverside and Upland)
showed a reasonably linear relationship between colorimetric and chemi-
luminescent measurements. However, both sites exhibited systematic
biases between the instruments. In general, it seems that the chemi-
luminescent instrument registers higher readings than the colorimetric
equipment. This may be a result of differences in calibration ot to
29
-------
positive interferences in the chemiluminescent instrument (no signifi-
cant interferences are expected for the colorimetric instrument).
The remaining site (San Jose) showed no correlation between the two
instrument types. In view of the results for the other stations, it
appears that the instruments at this location were either poorly
adjusted or were not colocated, or both.
30
-------
IV ANALYSIS OF PHYSICAL PROCESSES
A. Definition of Physical Phenomena
The aim of this analysis is to identify chemical and other
phenomena associated with NOo exceedances. The main physical processes
are defined as follows:
• Synthesis—NO -*N02 conversion by peroxy radicals
• Titration—NO + 03 -*N02 + 02
• Other—Transport and point-source effects
The synthesis process is the backbone of the photochemical-smog
cycle. The oxidation of NO to N0« by peroxy radicals such as RO^ and
HOo results in a rapid depletion of NO, and leads to a high N02/NO
ratio, setting the stage for the formation of 0^ via N02 photolysis.
Ozone accumulation usually follows; its level is controlled by meteoro-
logical conditions. In urban areas, this process takes place in a few
hours between 0600 and 1300; most of the NO is gone by 0900 or 1000.
The initial charge of NO is of course a result of the morning-traffic
rush.
In the absence of NO, 02 lingers in the atmosphere for several
hours until more NO is injected to react with 0^ in the titration pro-
cess, or meteorological factors cleanse the air, or a combination of the
two. On weekdays in urban areas, the early-evening traffic emits fresh
NO that yields N02 via titration. Thus, we expect that the titration
process generally occurs in the evening.
Other factors may also account for the occurrence of NO 2
exceedances. Transport effects are indirect manifestations of synthesis
and/or titration, since N02 is a secondary pollutant. Thus, N02 tran-
sported to a monitoring station was chemically produced along the way.
Nevertheless, for our purposes we define transport as a separate effect
to differentiate between nonlocal and local impacts. In the discussion
that follows, transport will be associated with nonlocal impacts, while
synthesis and titration will be linked to short-term effects. In terms
of hours, an exceedance of N02 is defined as short-term if it is the
result of synthesis occurring between 0600 and 1300, and titration
between 1600 and 1900. N02 exceedances occurring at other times are
more likely to be related to transport.
Point-source effects in this study are identified by relating N02
and S02 concentration histories. The tacit assumption is that S02 is an
indicator of large point sources, such as power plants (unless they are
gas-fired) and certain heavy industries, and that simultaneous S02 and
NO2 peaks signal the presence of a point-source impact. As has been
noted, only a few sites showed unambiguous evidence of these effects
(see Section II). In a point-source plume that is initially rich in NO,
31
-------
N02 would be produced via thermal oxidation by the three-body reaction
2NO -i- 02 ->2N02. This reaction is slow, however, and NO concentrations
must be on the order of several hundred ppm to make it significant. As
the plume mixes with the ambient air, it may entrain peroxy radicals
that oxidize NO via the synthesis process described earlier. It is also
possible that a plume rich in NO may undergo fumigation or otherwise
impact the ground downwind of the source and lead to high N02 levels via
the titration of NO and O-j. Whatever the cause of the N02 in the plume,
or of N02 associated with a plume, point-source impacts will be identi-
fied separately.
The next section presents several examples of the physical
processes defined above, and Section IV-C describes the physical
phenomena associated with all the monitoring stations.
B. Examples of Physical Processes
1. Chemical Synthesis
Figures 12 through 14 contain typical examples of chemical-
synthesis phenomena at two sites in Los Angeles County. The three days
shown are weekdays and display the typical morning and evening peak in
NO that is due to the traffic. Although it is evident that mobile
sources influence these two sites, there are some differences in the
time phasing of the NO peak. Figure 12 shows that the NO maxima occur
before 0800 at the Los Angeles/San Pedro site, but after 0800 at Temple
City (Figure 13). This is consistent with the location of these sta-
tions; Temple City is farther inland than Los Angeles/San Pedro. The
delayed appearance of the NO maximum at Temple City also suggests that
transport influences this site.
The shape of the NO and N02 curves in Figures 12 and 13 is typical
of the N02 synthesis process. Thus, the fast decay of NO is accompanied
by an equally rapid increase in N02. Note, however, that the N02 peak
occurs at 1000 at Los Angeles/San Pedro and at 1300 at Temple City.
This suggests that local effects are the primary influence in Los
Angeles, but some minor transport influence is likely at Temple City.
Figure 14 exhibits N02 concentrations exceeding 0.20 ppm that are
the result of several factors. The significant feature of this figure
is the high level of N02 that is present during the period midnight-
0900. This phenomenon is termed "carryover" because it is presumed to
be caused by pollution from the previous day that remains in the area (a
situation that would prevail under stagnant meteorological conditions).
Note that the nighttime N02 level hovers about 0.20 ppm: thus, it is
easy for the threshold to be exceeded when N02 synthesis comes into
play. This "carryover" phenomenon was observed on a number of occasions
at Los Angeles County sites, but not elsewhere. The phenomenon was seen
most frequently at the Los Angeles/San Pedro, Los Angeles/Westwood, Len-
nox
32
-------
0.8
.3
0.7
.2
0.6)
0.5
.1
oc
z
til
o
o
u
0.4
0.3
0.2
0.1
12
TIME - hours
16
20
24
Figure 12.
Example of NO2 formation by chemical synthesis at the
Los Angeles/San Pedro St. station on 6 April 1977.
CO levels are scaled by 0.01.
33
-------
0.8
0.7
0.6
O
cc
UJ
O
I
0.5
0.4
0.3
0.2
03
3
0.1
12
TIME - hours
16
20
24
Figure 13. Example of N02 formation by chemical synthesis at the
Temple City station on 8 February 1977.
34
-------
0.8
0.7
.2
CO
0.6
0.5
i
z
o
GC
Z
ill
u
0.4
- A ./
/N
/ %
0.3
0.2
0.1
\ *
I
12
TIME - hours
16
20
Figure 14. Example of NO2 formation by chemical synthesis at the
Temple City station on 7 December 1977.
Note the high nighttime N02 level carried over from previous
day. CO levels are scaled by 0.01.
24
35
-------
and Long Beach stations, so it may be related to a land/sea breeze
effect.
A potential point-source effect acting jointly with chemical syn-
thesis is shown in Figure 15 for the Whittier station. Note that the
S02 and the N02 peaks coincide, and that the S02 concentration is sub-
stantial (this concentration is rather unusual in the Los Angeles area).
The NO and N02 curves follow the typical synthesis pattern, but the
coincidence of S02 and N02 peaks suggests that some fraction of the N02
concentration may be associated with a point source plume.
An example of the so-called "Sunday effect" is depicted in Figure
16, which shows pollutant histories for Sunday and Monday, 14-15 March
197b, at the Los Angeles/Westwood site. Comparison of the NO and CO
curves for the two days clearly shows that mobile-source influences on
Sunday differ from those on Monday. Note that the morning traffic peak
is much smaller on Sunday than on Monday. Interestingly, 03 levels are
quite similar on these two days, but the 02 peak occurs later on Sunday
than on Monday, suggesting that the hydrocarbon/NOx ratio may be lower
on Sunday or that transport may be responsible. N02 does not seem to be
affected very much by the reduced Sunday emissions, peaking at noon on
both days. However, the figure shows that the N02 maximum on Sunday is
lower than on Monday. If true in general, this would be qualitatively
similar to the Sunday/weekday differences found for Og, but the effect
for N02 remains to be confirmed.
2. Titration of_ NO and £3
N02 exceedances due to titration of NO and 03 are portrayed in Fig-
ures 17 and 18. If the reaction NO + 03 -»N02 + 02 were stoichiometric,
the increase in N02 concentration would equal the decrease in 03 level.
However, stoichiometric conditions in the atmosphere can only be approx-
imated; the examples given are reasonable approximations. Figure 17 is
a typical titration example. The San Diego/Island Avenue site is in
downtown San Diego, and thus subject to traffic influences. However, 31
October 1976 is a Sunday, which accounts for the lack of pronounced CO
and NO peaks during the day and for the enhanced mobile source activity
that is suggested by the increase in NO and CO levels that begins around
1800. Even though it is Sunday, 03 reaches a maximum level of 0.15 ppm
(which is above the standard of 0.12 ppm). The increase in NO coincides
with the decrease in 03, and the increment in N02 between 1800 and 2100
is about 0.12 ppm while 03 decreases by 0.10 ppm. Hence the titration
is not stoichiometric by about 20 percent. However, it comes closer to
being stoichiometric if N02 is reduced by 10-17 percent (see Section III
for a discussion of these adjustment factors).
In the titration example shown in Figure 18 for Temple City, N02
levels are well above 0.20 ppm by 1600, when NO levels begin to
increase. Thus, additional N02 produced by titration exacerbates an
36
-------
0.8
0.7
0.6
0.5
i
i
o
1 °-4
H
Ul
o
z
o
u
NO
0.2
0.1
12
TIME - hours
16
20
24
Figure 15. Example of synthesis and possible point source influence
on NO2 at Whittier on 28 August 1975.
CO levels are scaled by 0.01.
37
-------
0.8
0.7
0.6
I
0.4
i
K
(C
Ul
u
8 0.3
0.2
0.1
SUNDAY
I 'I
MONDAY
.3
.2
.1
Figure 16. Example of N02 formation on Sunday and Monday,
14—15 March 1976, at the Los Angeles/Westwood
station. CO is scaled by 0.01.
38
-------
0.8
1.3
0.7
.2
0.6
0.5
•x°3
\
.1
I
O
DC
Z
u>
u
O
0
0.4
0.3
0.2
0.1
12
TIME - hours
16
20
Figure 17. Example of NO2 formation by titration at the
San Diego/Island Avenue station on 31 October 1976.
CO has been scaled by 0.01.
24
39
-------
0.8
.3
0.7
0.6
I
O
01
u
O
o
0.5
0.4
0.3
0.2
0.1
.2
.1
12
TIME - hours
16
20
Figure 18. Example of NO2 formation by titration at the
Temple City station on 5 December 1977.
CO has been scaled by 0.01.
24
40
-------
existing problem. From 1600 to 1800, N02 increases by 0.11 ppm while 03
decreases by 0.10 ppm, which agrees closely with stoichiometric condi-
tions. Adjusting N02 downward by 10 percent improves the agreement, but
a 17 percent does not.
3. Transport and Point-Source Impacts
An example of N02 transport at the El Toro station is shown in Fig-
ure 19; the day shown is a Saturday. Note that N02 peaks relatively
late, at 1500, and that its peak coincides with the 03 maximum. The low
NO levels indicate that most of the NO in the air mass has been con-
verted to N02. The double peak in the 03 curve often indicates tran-
sport effects: the noon peak is associated with local effects and the
second peak with transport from more distant areas.
Probable point-source impacts at Long Beach and Whittier are shown
in Figures 20 through 22. Figures 20 and 21 depict a Saturday/Sunday
sequence at Long Beach. Figure 20 exhibits N02 carryover and synthesis
in addition to the putative point-source effect. Note that during the
day the N02 and S02 curves track each other. Both N02 and S02 show
peaks of decreasing magnitude at 1200, 1500, and 1900, suggesting that a
point-source plume is advecting both pollutants into the site. S02 con-
centrations are higher in Figure 21, and the N02 does not exceed 0.20
ppm in this case, but the N02 and S02 peaks coincide and there is no
evidence of synthesis or other processes being at work. We conclude,
therefore, that the N02 increase is most likely caused by a point
source. The example in Figure 21 .was chosen because it suggests an
intriguing possibility: Since N02 did not exceed 0.20 ppm despite sub-
stantial S02 levels, the N02 contribution from point sources at this
site is insufficient by itself to cause N02 exceedances. However, the
point-source contribution could well result in N02 exceedances when com-
bined with N02 resulting from other processes. The pattern of Figures
20 and 21 is apparent in Figure 22 for Whittier. Both S02 and N02 track
each other; their peaks coincide at 1200. Evidence for other processes
that form N02 is lacking, hence a point-source impact seems indicated.
In this case, however, N02 exceeds 0.20 pprt. Further investigation is
required to establish the relative contribution of point sources to the
overall N02 burden when N02 exceedances occur.
C. Frequency and Spatial Distribution
of Physical Processes
What is the relative frequency of the three physical processes
associated with N02 exceedances? Does this frequency vary with the
location of the monitoring site? Is there a statewide pattern? This
section provides answers to these and other questions.
During this study, the plots of the data contained in Appendix A of
this report were analyzed, and each N02 exceedance was assigned to one
41
-------
0.8
.3
0.7
.2
0.6
0.5
.1
o
p
(T
0.4
u
o
0 0.3
0.2
0.1
NO.
12
TIME - hours
16
20
24
Figure 19. Example of N02 transport at El Toro station
on 25 January 1975. CO has been scaled by 0.01
42
-------
0.8
.3
0.7
0.6
.2
a
Z
o
Z
Ol
o
Z
o
0
0.5
0.4
03
0.2
i
i
i
i
t
i
i
0.1
•^ _ — * •» ^
1 , *h*
12
TIME - hours
16
20
Figure 20. Example of potential point source impact at
Long Beach on Saturday, 15 November 1975.
CO has been scaled by 0.01.
24
43
-------
0.8
.3
0.7
0.6
E °'5
a
i
O
< 0.4
GC
Z
ui
0
O
° 0.3
SO.,
.1
0.2
0.1
\NO
I
12
TIME - hours
16
20
Figure 21. Example of potential point source impact at
Long Beach on Sunday, 16 November 1975.
CO has been scaled by 0.01.
24
44
-------
0.8
.3
0.7
.2
0.6
0.5
I
t
O
DC
Z
UJ
O
8
0.4
0,
0.2
0.1
12
TIME - hours
16
20
24
Figure 22. Example of potential point source impact at
Whittier on Monday, 25 August 1975.
CO has been scaled by 0.01.
45
-------
of the three categories previously defined in Section IV-A: synthesis,
titration, and "other" (where the last category contains both transport
and point-source effects). The resulting classification for each site
is displayed in Figure 23.
Figure 23 shows both the frequency and spatial distributions of the
physical phenomena. The size of each dot indicates the number of cases
associated with each category (a case may consist of more than one
hour): The small dot corresponds to the range 1-10 cases, the medium-
sized dot to 11-50 cases, and the large dot to more than 50 cases. The
order of the monitoring sites follows an approximate north/south axis,
running from Sacramento in the north to Chula Vista near the Mexico-
California border. Within the north/south orientation, the stations
have been grouped by geograhical region, which often coincides with
county boundaries. The major regions are:
• San Francisco Bay area—begins with Concord and ends with San
Jose.
• The South Coast Air Basin (SCAB)
- Los Angeles County—begins with Los Angeles/Gault St. and ends
with Whittier.
- San Bernardino County—begins with Upland and ends with Chino.
- Riverside County—contains Riverside and Rubidoux.
- Orange County—begins with La Habra and ends with El Toro.
• San Diego area—begins with Oceanside and ends with Chula Vista.
A small number of stations that do not belong to these three major
regions are shown either separately or in small clusters.
Figure 23 indicates that in general, chemical synthesis is the
predominant process associated with N02 exceedances. This is followed
by titration and "other," in that order. Overall, synthesis is approxi-
mately 44 percent more frequent than titration, and titration is roughly
eight times more frequent than transport and point-source influences.
However, the relative dominance of the three categories can change
drastically depending on the location of the monitoring station. These
and other spatial effects are discussed below.
The spatial distribution of the N02 exceedances and their associ-
ated physical processes can be inferred from Figure 23. It is evident
that the majority of the exceedances occur in the South Coast Air Basin,
where Los Angeles County predominates. The San Diego area is second to
the SCAB in number of exceedances, followed by the San Francisco Bay
area. In addition to these major groupings, considerable variation
exists among the various sites. Within the Bay area, Sunnyvale shows
some predominance of synthesis over titration and no apparent transport
or point-source effects. By contrast, Oakland shows some impact from
all physical processes. Burlingame, on the other hand, shows no
46
-------
FREQUENCY RANGE
CtMmial Synthoii
•
•
•
*
•
•
»
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
*
•
•
•
•
•
•
•
•
0 4- 03 Trtntion
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
*
•
•
•
•
•
•
•
Othtr
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
*
•
SITE NAME
SACRAMENTO
STOCKTON
CONCORD
RICHMOND
SAN FRANCISCO
OAKLAND
BURLINGAME
FREMONT
REDWOOD CITY
SUNNYVALE
SAN JOSE '
SAN JOSE t
FRESNO
BARSTOW
VICTORVILLE
SANTA BARBARA
CAMARILLO
SIMI VALLEY
NEWHALL
LOS ANGELES/Giult St.
BURBANK
LOS ANGELES/Wonwod
LOS ANGELES/S. Pkdro
LENNOX
LYNWOOD
LONG BEACH
PASADENA
AZUSA
TEMPLE CITY
POMONA
WHITTIER
UPLANO/0 St.
UPLANO/S. B. St. *
UPLAND/S. B St.
FONTANA
SAN BERNARDINO
REOLANDS
CHINO
RIVERSIDE *
RIVERSIDE f
RUBIDOUX *
RUBIDOUX T
LA HA8RA
ANAHEIM
COSTA MESA
EL TORO
OCEANSIOE
ESCONOIDO
SAN DIEGO/Kfwmy
SAN OlEGO/ltlmd
EL CAJON
CHULA VISTA
LEGEND: '1-10 .11-50 •> SO CASES f
* Colorimnric method
t Chemilumintjcinl mithod
f A caj» may bt man thin oni hour
Figure 23. Physical processes in high NO2 events classified by site.
47
-------
evidence of synthesis or titration. Only synthesis is important in Con-
cord, Richmond and San Francisco. The other sites in the Bay area are
influenced by synthesis and titration, but show no transport or point-
source impacts. (The San Jose station will be discussed later.)
Barstow and Victorville are located inland in relatively sparsely
inhabited areas. Both are impacted primarily by transport. One possi-
ble explanation is that these sites are affected by ozone transport from
urban areas, which when combined with local NO emissions yields high N02
levels. In this case, what appears to be transport is the result of
titration. Alternatively, N02 transported from urban areas or from
point sources could account for the exceedances. The diurnal variation
of the N02 exceedances at these two sites (described in Section V-B)
suggests transport or point-source impacts, rather than local effects.
However, we were unable to identify point-source impacts because S02 was
not monitored.
Several interesting differences are evident among the sites in Los
Angeles County. Synthesis predominates at Burbank, Los Angeles/ West-
wood, Los Angeles/San Pedro, Lennox, Long Beach, and Whittier. These
stations are located in areas that can be considered to be source
regions, i.e., areas with a high emissions density* It seems reason-
able, therefore, to conclude that the prevalence of synthesis at these
stations is associated with short-term effects directly related to
mobile sources in their general vicinity. By contrast, the Los
Angeles/Gault St. station experiences about equal influence from syn-
thesis and titration. This station is located near a traffic corridor
in the northwestern part of the country. Consequently, high N02 is
associated both with synthesis due to short-term causes and with the
titration of ozone formed during the day reacting with NO emitted by the
evening commuter traffic•
Pasadena, Azusa, and Pomona are known to be recipients of ozone
transported from the central basin. Thus, we expect the ozone to arrive
at these sites in time to react with the evening NO peak. Accordingly,
titration should play a significant role in N©2 production. This is
precisely what Figure 23 shows for Pasadena and Pomona, where titration
predominates. In fact, the titration/synthesis ratio is 4/1 at Pasadena
and 6/1 at Pomona. Azusa has less traffic than either Pasadena or
Pomona, and the dominance of titration tends to be reduced. Although
Figure 23 shows that the number of cases of synthesis and titration at
Azusa fall in the same frequency range (11-50 cases), the titration/
synthesis ratio is in fact 2/1. The station at Temple City is located
approximately between Azusa and Pasadena, and has appoximately a one-
to-one titration/synthesis ratio: In this respect, it appears to be
similar to the Los Angeles/Gault St.site.
Long Beach and Whittier are the only stations that show any signi-
ficant impact from stationary sources. As Figure 23 shows, the fre-
quency of transport/point-source effects is in the same range as the
titration influence. Actually, titration occurs somewhat more
48
-------
frequently: the ratio of transport/titration is 0.6 at Long Beach and
0.8 at Whittier. It is not surprising that these sites show the influ-
ence of stationary sources: Section II showed that both sites are close
to major clusters of stationary sources.
The San Bernardino County stations—Upland through Chino in Figure
23—are located downwind of Los Angeles County and are considered to be
receptors of transported pollutants. Upland, especially, frequently
exhibits some of the highest 0^ levels in the basin. However, as Sec-
tion V will demonstrate, N02 exceedances at these sites are much less
frequent than in Los Angeles County. When they occur, titration is more
frequent than synthesis, as expected for ozone receptors. Chino and
Redlands, in particular, experience only titration.
Data for Riverside and Rubidoux seem to be similar to that of the
San Bernardino County sites. Thus, titration is the predominant effect
at both sites. Riverside, but not Rubidoux, exhibits some impact due to
synthesis.
La Habra and Anaheim exhibit N(>2 exceedance frequencies that are
similar to those seen at Los Angeles County sites. Synthesis and titra-
tion are approximately equally important at these two stations, in keep-
ing with their dual source/receptor character (see Section II-C for a
description of wind patterns affecting these sites). Both sites also
experience some minor impacts due to transport or point-source effects.
The frequency of NO2 exceedances at Costa Mesa and El Toro is much lower
than at the other two sites (see Section V). Both sites are in suburban
locations. The Costa Mesa station is near the coast, close to a heavily
traveled highway (see Table 2): Figure 23 shows that synthesis predom-
inates here, which is consistent with the highway influence. By con-
trast, the El Toro station is farther inland in an area of low popula-
tion density. Titration and transport are of about equal influence at
El Toro.
Two of the San Diego area sites—San Diego/Kearney and Chula
Vista—are classified as suburban, and the others are in urban loca-
tions; the San Diego/Island site is downtown. Titration predominates at
the suburban sites, and also at Oceanside and Escondido, but synthesis
is most important at El Cajon and San Diego/Island. The prevalence of
synthesis at the San Diego/Island site is consistent with similar
behavior observed at source-dominated locations elsewhere. The fre-
quency of occurrence of N02 exceedances at this site is also the highest
of all the San Diego area stations (see Section V). The El Cajon site
also seems to be located in a source-oriented area, judging by the domi-
nance of the synthesis process.
Sites containing measurements made using both colorimetric and
chemiluminescent techniques have been shown separately in Table 8. The
figure clearly shows the tendency of the chemiluminescent observations
to be higher than the colorimetric data, as discussed in Section III.
Thus, San Jose shows a somewhat higher frequency of chemical synthesis
49
-------
Table 8
DISTRIBUTION OF N02 EXCEEDANCES AGGREGATED
FOR ALL MONITORING STATIONS, 1975-1977*
Ul
o
Concentration
(ppm) t
0.21
0.22
0.23
0.24
0.25
0.26
0.27
0.28
0.29
0.30
0.31
0.32
0.33
0.34
0.35
0.36
0.37
0.38
0.39
0.40
0.41
Number of
Hours$
942
909
631
604
448
368
260
252
156
169
92
104
81
61
58
53
40
36
32
25
15
Cumulative
Percentage
17.3
34.1
45.7
56.8
65.1
71.9
76.6
81.3
84.2
87.3
89.0
90.9
92.4
93.5
94.6
95.5
96.3
96.9
97.5
98.0
98.3
Concentration
(ppm) t
0.42
0.43
0.44
0.45
0.46
0.47
0.48
0.49
0.50
0.51
0.52
0.53
0.54
0.55
0.56
0.57
0.58
0.59
0.60
0.61
0.62
Number of
Hours$
14
17
9
11
10
6
7
2
4
1
2
2
2
1
2
0
1
0
2
0
1
Cumulative
Percentage
98.5
98.8
99.01
99.21
99.39
99.50
99.63
99.67
99.74
99.76
99.80
99.83
99.87
99.89
99.93
99.93
99.94
99.94
99.98
99.98
100.00
Where simultaneous chemiluminescent and colorimetric measurements were made, only
colorimetric data have been used in this table.
Median =0.24 ppm
*Total hours = 5430
t
-------
cases measured by chemiluminescence than by colorimetry. The apparent
difference at Riverside is due to different lengths of the period of
simultaneous measurements (see Table 5). A similar effect causes the
apparent discrepancy in the titration frequency at the Upland/San Ber-
nardino St. Station. Rubidoux does not show any apparent disagreements
because^the frequency of N02 exceedances is very low (and is below the
diagram's ability to resolve it). Section V contains a further discus-
sion of these discrepancies.
51
-------
V SPATIAL AND TEMPORAL PATTERNS
Previous sections discussed some facets of the spatial variation of
N02 exceedances. This section continues the examination of spatial pat-
terns and considers diurnal and seasonal effects.
A. Spatial Distribution of High N0_2 Concentrations
It is useful to view the pattern of N02 exceedances from a state-
wide standpoint as well as in detail. Thus, we begin by examining the
spatially aggregated frequency distribution of N02 exceedances, and then
study their spatial distribution.
Table 8 displays the frequency distribution of N02 concentrations
greater than 0.20 ppm, aggregated for all monitoring sites. In prepar-
ing this table, only colorimetric data were used for those sites that
employed both colorimetric and chemiluminescent instruments simultane-
ously (see Table 5). A histogram of the distribution is shown in Figure
24. As might be expected of a tail distribution, the histogram has the
shape of a decaying exponential function.
Section III-C presented the recent finding that the N02 data are
from 10-17 percent too high. Table 8 allows us to examine the effect of
adjusting the data to reflect this. The adjustment is applied using the
formula Cn = CQ/(l4a), where CQ » old unadjusted N02, Cn - new adjusted
N02, and a - adjustment factor, 0.10 £ a £ 0.17. The adjustments can be
considered in either of two equivalent ways: raise the old threshold
level of 0.20 ppm without altering the N02 data, or adjust the N02 data
by lowering them by a factor of l/(l+a) and retain the 0.20 ppm thres-
hold. Both methods produce the same results and the choice of method is
a matter of convenience. Table 9 presents the effect of the adjustment
from the standpoint of the first method above. The table makes it clear
that data adjustments in this range would reduce the number of
exceedances between 34 and 46 percent on a statewide basis, a substan-
tial impact. In view of this significant effect, it is apparent that
regulations pertaining to an hourly N02 standard must specify the
adjustment factor to be used.
Table 9
EFFECT OF ADJUSTING N02 DATA
Adjustment
Factor
0.10
0.17
Old
Threshold
(ppm)
0.20
0.20
New
Threshold
(ppm)
0.220
0.234
Reduction in
Exceedance Frequency
(7.)
34.1
45.7
53
-------
70
60
SO
40
z
UJ
o
oc
UJ
0.
30
20
10
21-25 26-30 31-35 36-40 41-45 46-50 51-55 >55
NO2 CONCENTRATION - pphm
Figure 24. Histogram of the frequency distribution
of N02 exceedances aggregated statewide.
54
-------
Although Table 9 shows that data adjustments have a significant
impact on a statewide basis, the effect on individual stations can be
even greater. This point will be considered in the examination of the
spatial distribution of N02 exceedances.
Figure 25 shows the range of concentrations measured at all the
monitoring stations. (The figure retains the order of the sites used in
Figure 23.) In addition to the concentration range, Figure 25 displays
the number of hours for each site when N02 > 0.20 ppm (given by the
number at the end of each bar). When the number of hours is greater
than 10, the figure also shows the median concentration of the ensemble
of N02 exceedances (marked by a vertical line with a dot in its center) •
The fact that the N02 exceedances occur most frequently in the SCAB
is clear in Figure 25. Of the 5430 total exceedances (not counting the
simultaneous chemiluminescent data), 4977 are reported in the SCAB,
which accounts for about 92 percent of the statewide total. Moreover,
Los Angeles County alone accounts for approximately 80 percent (4368
hours) of the statewide exceedances, and for 88 percent of the SCAB's
total. The Orange County sites are second to Los Angeles's with 472
hours, a 9:1 ratio in favor of the latter. It is also apparent that the
sites at La Habra and Anaheim are similar to those in Los Angeles
County, while the other SCAB stations differ markedly from those in Los
Angeles.
A clear division exists between the northern and central California
sites and those in Los Angeles and points south. The southern stations
show not only a higher exceedance frequency, but also higher maximum N02
concentrations. Thus, except for Stockton, all the sites from
Sacramento to Newhall exhibit a rather narrow range of concentrations,
from 0.21 to 0.30 ppm. This range is exceeded by all twelve sites in
Los Angeles County.
The stations in the San Diego area show considerable variation.
Only the site located in the central business district (San
Diego/Island) exhibits an exceedance frequency and a concentration range
that are similar to those found in the SCAB.
It should be noted that the numerical predominance of the Los
Angeles County stations is somewhat misleading because many of the sites
are correlated, which results in some degree of multiple counting. More
detailed investigation, which is beyond the scope of this work, is
required to unravel fully the interstation dependencies. However, Sec-
tion V-D presents the results of an analysis that sheds some light on
this problem.
As noted earlier, adjusting the data can have a very large impact
on some individual sites. For example, Figure 25 suggests that a 17
percent adjustment would eliminate several stations from among those
reporting N02 exceedances. The percent reduction in the exceedance fre-
quency for individual sites is listed in Table 10 for both 10 and 17
55
-------
SITE NAME
SACRAMENTO
STOCKTON
CONCORD
RICHMOND
SAN FRANCISCO
OAKLAND
BURLINGAME
FREMONT
REDWOOD CITY
SUNNYVALE
SAN JOSE '
SAN JOSE T
FRESNO
BARSTOW
VICTORVILLE
SANTA BARBARA
CAMARILLO
SIMI VALLEY
NEWHALL
LOS ANGELES/Guilt St.
BURBANK
LOS ANGELES/Wmmod
LOS ANGELES/S. Pidro
LENNOX
LYNWOOO
LONG BEACH
PASADENA
AZUSA
TEMPLE CITY
POMONA
WHITTIER
UPLAND/D St.
UPLAND/S. B. St '
UPLANO/S. B St. T
FONTANA
SAN BERNARDINO
REDLANDS
CHINO
RIVERSIDE *
RIVERSIDE '
RUBIOOUX '
RUBIDOUX T
LA HABRA
ANAHEIM
COSTA MESA
EL TORO
OCEANSIDE
ESCONDIDO
SAN DIEGO/KMrnfv
SAN DIEGO/lilind
EL CAJON
CHULA VISTA
MAXIMUM CONCENTRATION - ppm
.25 .30 .35 .40 .45 .50 .55 .60 .85
~n 1
IB
3-,
I"
5
" I X
|6
| 43
I 26
I 28
I 6
» Ms
1
1
I*
J'
| 205
» I 386
828
633
I 297
] 163
483
526
|166
|264
| 231
386
|13
I"
|38
rf..
19
l«
131
|82
] 212
|206
139
1 16
|24
5
120
|135
167
_l»
i t i i i I i I 1
* Colonniitric method
t Chtmilumintscent method
Figure 25. Spatial distribution of NO2 concentrations exceeding 0.20 ppm.
Number at end of bar indicates hours exceeding 0.20 ppm. Vertical line
with a dot marks the location of the median concentration.
56
-------
Table 10
REDUCTION OF N02 EXCEEDANCE FREQUENCY
AT INDIVIDUAL STATIONS AFTER DATA ADJUSTMENT
Ui
Site Name
Sacramento
Stockton
Concord
Richmond
San Francisco
Oakland
Burlingame
Fremont
Redwood City
Sunnyvale
San Jose*
San Joset
Fresno
Bar stow
Victorville
Santa Barbara
Camarillo
Simi Valley
Newhall
Los Angeles/Gault St.
Burbank
Los Angeles/Westwood
Los Angeles /San Pedro
Lennox
Lynwood
Ajustment Factor
10%
0
25
50
0
43
53
100
58
50
58
46
36
50
62
40
100
100
0
0
42
36
23
25
38
42
177.
100
38
100
100
57
68
100
65
83
67
65
57
67
71
67
100
100
40
100
58
49
32
35
49
56
Site Name
Long Beach
Pasadena
Azusa
Temple City
Pomona
Whittier
Up land/ D St.
Upland/San Bernardino St.*
Upland/San Bernardino St.t
Fontana
San Bernardino
Red lands
Chino
Riverside*
Riverside*
Rubidoux*
Rubidouxt
La Habra
Anaheim
Costa Mesa
El Toro
Oceanside
Escondido
San Diego/Kearney
San Diego/Island
El Cajon
Chula Vista
Adjustment Factor
10%
33
32
39
31
34
33
23
53
50
50
55
22
40
45
38
40
63
40
37
36
0
38
100
30
46
61
63
17%
42
43
49
42
47
46
31
76
58
75
91
33
60
65
49
60
63
54
49
49
13
54
100
30
56
76
88
Colorimetric method
Chemiluminescent method
-------
percent adjustment factors. The table shows that a 17 percent adjust-
ment eliminates nine sites, whereas the 10 percent factor eliminates
four. For the San Francisco Bay area stations, it is apparent that
either adjustment would mitigate considerably the magnitude of the high
N02 problem. The same applies to other sites outside the SCAB, as well
as to the stations in San Bernardino and Riverside counties. Thus,
adjusting the data to improve its accuracy has the effect of further
limiting the occurrence of high N02 levels to the SCAB in general and to
Los Angeles County in particular.
Figure 25 provides further evidence of the high incidence of N02
exceedances that is associated with source-dominated areas. As men-
tioned in Section IV-C, Burbank, Los Angeles/Westwood, Los Angeles/San
Pedro, Lennox, Long Beach, and Whittier are areas with high traffic den-
sity, where chemical synthesis is the most common mechanism leading to
NO2 exceedances. Figure 25 shows that these sites also have some of the
highest exceedance frequencies and experience a wide range of concentra-
tions. Pasadena is the only other site that is in the same class with
respect to number of exceedances, and it is a hybrid which combines the
characteristics of both source and receptor sites. The association
between high exceedance frequency, N02 levels, and source density is
also apparent at La Habra, Anaheim and San Diego/Island. By contrast,
those sites which are located beyond areas of high source density, e.g.,
Azusa, Riverside, Chino, and El Cajon, show lower exceedance counts and
N02 concentrations.
However, differences in emissions density alone cannot explain the
statewide pattern of high N02 events. The Bay area sites are located in
high traffic-density areas, yet the incidence of N02 exceedances and the
maximum N02 levels reached are well below those for Los Angeles.
Clearly, the north-south differences are heavily influenced, if not con-
trolled, by the marked variation in climate encountered in these two
parts of the state. The north is generally cooler, windier, and has
cleaner air: the south, of course, has a warmer climate, and more smog.
Figure 25 shows differences in exceedance frequency and in maximum
concentration at those sites where both colorimetric and chemilumines-
cent measurements were obtained. We caution, however, that the
discrepancy in exceedance frequency at Upland/San Bernardino St., River-
side, and Rubidoux is explained by the fact that one instrument operated
three years, and the other two (see Table 5). The two techniques were
used for three years at San Jose; there is a slight difference in the
exceedance frequency as well as in the maximum N02 concentration reached
at this site.
58
-------
B. Seasonal and Diurnal Variations
The analysis presented here seeks to answer the following ques-
tions:
• How does the N02 exceedance frequency change with season and
with time of day?
• Do the concentrations of N02 that exceed 0.20 ppm vary with time
of day? If so, what is the variation?
The seasonal and diurnal fluctuations of NC^ exceedances should
provide some indications about the forcing function, i.e., the emis-
sions, that drives the source/atmosphere system. They should also fur-
nish evidence about the links between the physical processes discussed
in Section IV and the exceedance frequency. The clues obtained from the
study of seasonal variations prompted the investigation of seasonal pat-
terns in stationary-source emissions in the SCAB which is described in
Section V-C.
The second question is aimed at determining the relative strength
of the various physical processes leading to NOn exceedances. What
differences, if any, exist between the NO2 concentrations associated
with synthesis, titration, and other influences such as transport? This
is studied by comparing the distribution of NO2 levels that exceed 0.20
ppm stratified for three time intervals, on the assumption that the
various physical processes follow some regular pattern of diurnal varia-
tion. This analysis is conducted at several selected stations, because
the assumption is more likely to be satisfied at those sites and because
the respective data bases are extensive.
Seasonal and diurnal changes in the NO2 exceedance frequency are
displayed for all the sites in Figure 26, where the order of the moni-
toring stations is the same as in Figure 23 (Section IV-C). The sea-
sonal fluctuations are shown in the right side of Figure 26. Note that
each column corresponds to a month, and that the calendar has been
arranged so that December and January appear next to each other in the
center of the panel. The figure shows that N02 exceedances have a
strong seasonal component; the exceedances occur most frequently during
November-February. It is clear that the phenomenon occurs statewide,
although the effect is less pronounced outside southern California owing
to the lower exceedance frequency. Meteorological conditions are
largely responsible for the seasonal variation, since it is well known
that the mixing layer is shallower in the winter, thereby leading to
higher levels of nitrogen oxides (NOX). However, it is possible that
higher NOX emissions from increased space heating and other sources con-
tribute to the problem in the winter months. Since this may be impor-
tant for regulatory purposes, Section V-C examines the seasonal com-
ponent of NOX emissions in the SCAB.
Diurnal fluctuations in the exceedance frequency are displayed in
the left side of Figure 26. The columns of the panel correspond to the
59
-------
DIURNAL VARIATION
OF HIGH N02 CONCENTRATIONS
SEASONAL VARIATION
OF HIGH NO2 CONCENTRATIONS
FREQUENCY
00
01
•
12
03
*
*
• • .; • •
• • I •
i
i
M
OS
•
•
»|07
• •
• 0
08
•
09
10
.
• fj
••
Afl
• t]
::
•
•
•
•
*
OF OCCURRENCE
11
.
12)13
ups
• •
••
A 1*
?:
••
••
•;
••
£•
••
**
•
• •
•
"
6
17
•
•
8
19
20
21
•
•
•
•
P^^^^^r^ ^^r^ r^r*
m
•
•
•
•
• 0
• •
. •
::
•
•
*
••
••
••
-
•
• ;
•
•
•
•
22
23
*•
•
•
SITE NAME
SACRAMENTO
STOCKTON
CONCORD
RICHMOND
SAN FRANCISCO
OAKLAND
BURLINGAME
FREMONT
REDWOOD CITY
SUNNYVALE
SAN JOSE *
SAN JOSE f
FRESNO
BARSTOW
VICTORVILLE
SANTA BARBARA
CAMARILLO
SIMI VALLEY
NEWHALL
LOS ANGELES/Gwilt St.
BURBANK
LOS ANGELES/Wotwood
LOS ANGELES/3. Pidro
LENNOX
LYNWOOO
LONG BEACH
PASADENA
AZUSA
TEMPLE CITY
POMONA
WHITHER
UPLAND/0. St
UPLANO/S. B. SL*
UPLANO/S. B.Str
FONTANA
SAN BERNARDINO
REDLANOS
CHI NO
RIVERSIDE *
RIVERSIDE f
RUBIDOUX '
RUBIDQUX T
LA HABRA
ANAHEIM
COSTA MESA
EL TORO
OCEANSIDE
ESCONOIDO
SAN OlEGO/Knrmy
SAN DIEGO/lshnd
EL CAJON
CHULA VISTA
FREQUENCY OF OCCURRENCE
JiH
•
•
•
Aug
•
•
•
•
Sip
•
•
:
'
•
Oet
•
'
* •
Nov
*
•
i
•
OK
*
*
?
Jin
*
•
*
?
Fib
•
•
*
•"
Mir
•
Apr
•
Miy
ft
*
Jun
•
•
LEGEND : -1-5
• 6-20
>>20 hours
LEGEND
1-10
»11-SO
)>50 hours
* Colorinutric mtthod
t Chimiluminncint mtthod
Figure 26. Diurnal and seasonal variation of NO2 exceedance frequency.
60
-------
24 hours of the day, running from 0 to 23, one pair of hours to a
column. The figure shows some clustering for the hours 0800-1300 and
1600-2000. In general, sites where synthesis predominates have clusters
in the morning hours; stations where titration is dominant have clusters
in the 1600-2000 period. The Bay area stations contain several examples
of the morning-clustering effect, as do Lennox and San Diego/Island in
the south. Pasadena, Azusa, Temple City, and Pomona display clusters
during the evening, in keeping with the fact that titration is important
at these sites. The hybrid source/receptor character of Pasadena is
evident in the double cluster it exhibits: one cluster for 0800-0900
corresponding to synthesis and one for 1600-2100 corresponding to titra-
tion.
Many other similar patterns can be observed. Comparing Figure 26
with Figure 23 yields more insights about the diurnal variation of the
various physical processes.
Figure 26 shows that the N02 exceedances recorded in Stockton, Bar-
stow, and Victorville occur only at night. This suggests that transport
or point sources or both are probably responsible for the high N0£
observed at these locations. Examination of the plots of NOo concentra-
tion for these two sites (see Appendix A) reveals that many exceedances
are sudden pulses of short duration, which is an attribute commonly
found in point-source impacts. However, such occurrences cannot be con-
fidently attributed to point sources because SO2 was not monitored.
Are N02 levels associated with synthesis greater, smaller, or equal
to those linked to titration or other physical causes? To answer this
question, we divided the calendar day into three 8-hour intervals (0-
0500 and 2200-2300,* 0600-1300, and 1400-2100) and compared the fre-
quency distributions of N02 levels corresponding to each time interval.
The rationale for this approach is that synthesis is generally a morning
phenomenon, titration usually occurs in the afternoon and early evening,
and other processes are more common at night. Consequently, we associ-
ate synthesis with the period 0600-1300, titration with the interval
1400-2100, and other factors with the hours 0-0500 and 2200-2300.
Although this approach is only an approximation, it is a practical means
for tackling this problem, since the volume of data precludes an hour-
by-hour classification of each NO2 exceedance.
Nine monitoring stations were selected for the analysis because
each has an extensive data base and because the correspondence between
physical processes and time of day is likely to be satisfied in the
majority of cases. The nine sites are: Anaheim, Azusa, Lennox, Long
Beach, Los Angeles/San Pedro St., Los Angeles/Westwood Blvd., Pasadena,
The first interval is designated and 0-0500 and 2200-2300, rather than
2200-0500, to emphasize that there is no crossover between calendar
days.
61
-------
Whittier, and San Diego/Island Avenue. The first eight stations are in
the SCAB and the last is in the city of San Diego. The prevalence of
various physical processes among the stations is another criterion for
selecting these stations.
The frequency distributions of N02 levels for the three time inter-
vals were tested statistically for homogeneity. The test is based on
the chi-square statistic and indicates any statistically signficant
differences between the distributions (c.f. Dixon and Massey, 1957).
The distributions are said to be homogeneous if no statistically signi-
ficant difference can be detected. The null hypothesis is that the dis-
tributions are homogeneous, and the alternative hypothesis is that they
are not. The null hypothesis is rejected or not depending on the value
of the chi-square statistic. If the null hypothesis is not rejected
(i.e., if it is "accepted"), the implication is that the physical
processes at work at various times of day yield N02 concentrations that
are similar. Conversely, if the null hypothesis is rejected, we infer
that significant differences exist in the NO^ levels associated with the
various time intervals (and by assumption, with the physical factors
corresponding to those intervals).
The frequency distributions for the nine stations and corresponding
chi-square statistics are shown in Table 11. In several instances, all
three distributions could be compared simultaneously, but in some cases
only the 0600-1300 and 1400-2100 distributions could be compared because
certain technical assumptions of the chi-square test were not satisfied
when the night distribution was included (see Dixon and Massey, 1957).
Table 11 shows that the night distribution was excluded from the compar-
ison at Anaheim, Azusa, Whittier, and San Diego/Island Avenue. This may
in itself be significant, since it suggests that the processes operating
at night at these sites differ qualitatively from the daytime and
early-evening processes.
It can be seen in Table 11 that the homogeneity hypothesis is
rejected for the 0600-1300 and 1400-2100 distributions at Anaheim and
Los Angeles/Westwood Blvd. N02 exceedances are more common during
1400-2100 than during 0600-1300 at Anaheim, while the opposite is true
at Los Angeles/Westwood Blvd. However, at both sites there are more
high values (>32 pphm) during the period 0600-1300, suggesting that syn-
thesis yields more high N0£ levels than titration at these two sites.
No statistically significant difference (with significance level of
0.05) was detected among the three distributions at Lennox, Los
Angeles/San Pedro St., and Pasadena. Thus, at these three sites the
various physical processes yield statistically indistinguishable N0£
concentrations, although the frequency of occurrence differs substan-
tially among the three time intervals.
Long Beach and Los Angeles/Westwood Blvd. exhibit significant
differences when all three distributions are considered. However, the
homogeneity hypothesis cannot be rejected at Long Beach for the periods
62
-------
Table 11
COMPARISON OF FREQUENCY DISTRIBUTIONS OF N02 LEVELS
OCCURRING IN DIFFERENT TIME INTERVALS
Ch
Site Name/
Time Interval (PST)
Anaheim
0-0500 and 2200-2300
0600-1300
1400-2100
Azusa
0-0500 and 2200-2300
0600-1300
1400-2100
Lennox
0-0500 and 2200-2300
0600-1300
1400-2100
Long Beach
0-0500 and 2200-2300
0600-1300
1400-2100
Los Angeles /San Pedro St.
0-0500 and 2200-2300
0600-1300
1400-2100
Los Angeles/Westwood Blvd.
0-0500 and 2200-2300
0600-1300
1400-2100
Pasadena
0-0500 and 2200-2300
0600-1300
1400-2100
Whittler
0-0500 and 2200-2300
0600- 1300
1400-2100
San Diego /Is land Ave.
0-0500 and 2200-2300
0600-1300
1400-2100
Hours in Specified Concentration Range
21-23
pphm
9
26
66
4
14
64
33
85
28
47
101
54
52
94
73
25
114
63
21
54
153
15
87
74
10
24
42
24-26
pphm
2
13
34
5
9
41
13
54
15
20
71
33
32
75
50
19
99
46
16
30
94
10
41
52
4
14
18
27-29
pphm
0
6
21
0
6
12
4
25
5
9
37
17
12
39
31
14
59
40
4
21
46
1
17
22
0
7
8
30-32
pphm
0
6
7
0
2
3
1
8
5
7
31
10
12
34
19
3
33
15
1
7
27
0
16
11
0
4
1
>32
pphm
0
10
6
0
0
6
5
13
3
4
36
6
17
68
25
0
81
17
0
9
43
0
25
15
0
3
0
Total
Hours
11
61
134
9
31
126
56
185
56
87
276
120
125
310
198
61
386
181
42
121
363
26
186
174
14
52
69
Statistical Analysis
Chi-Square
t
10.30
10.30
t
2.37*
2.37*
7.66
7.66
7.66
15.69
7.43
7.43
13.63
13.63
13.63
29.48
13.97
13.97
12.83
12.83
12.83
t
6.02
6.02
t
7.74
7.74
Degrees of
Freedom
4
4
3
3
8
8
8
8
4
4
8
8
8
8
4
4
8
8
8
4
4
4
4
Homogeneity*
Hypothesis
I
1
2
2
2
2
2
3
3
3
2
2
2
4
4
4
2
2
2
2
2
2
2
"Notes: a = significance level, 1 » rejection with a - 0.05; 2 = acceptance with a = 0.05; 3 - rejection for all three distribu-
tions but not for 0600-1300 and 1400-2100 with a » 0.05; 4 = rejection for all three distributions and for 0600-1300 and 1400-
2100, a - 0.05.
Distribution excluded.
Last two concentration ranges combined for comparison purposes.
-------
0600-1300 and 1400-2100. This indicates that nighttime conditions at
Long Beach differ quantitatively, and possibly qualitatively, from those
that exist at the other time intervals. Recalling that Long Beach is
situated in an area with a high density of stationary sources, it is
possible that the difference revealed in Table 11 may be caused by the
predominance of mobile sources during the day and early evening and of
stationary sources at night. We saw earlier that at Los Angeles/
Westwood Blvd. the homogeneity hypothesis was rejected for the 0600-1300
and 1400-2100 distributions. Thus, all three distributions differ sig-
nificantly at this station, which indicates that the physical processes
operating here are well-differentiated with time of day: This is the
only station that showed this feature.
Azusa, Whittier, and San Diego/Island Avenue did not exhibit any
statistically significant differences between the 0600-1300 and 1400-
2100 distributions. Evidently, the physical processes operating during
these periods yield similar NO2 levels.
C. Seasonal Emissions Patterns
in the South Coast Air Basin
The previous section discussed the strong seasonal component in the
frequency of N02 exceedances. Specifically, it was noted that the
exceedance frequency increases during the period November through Febru-
ary. This phenomenon was most strongly manifested at the SCAB stations.
Here we examine seasonal effects in NOX emissions in the SCAB to see
whether they follow a pattern similar to that of the N©2 exceedances.
NOX emissions from stationary sources have been studied in detail
by Bartz et al. (1974). Figure 27 shows the estimated monthly variation
in NOX emissions for the period July 1972-June 1973. Note that the
emissions reach a maximum in December and January, which coincides with
part of the period when the number of NO2 exceedances increases. Most
of the December-January peak is due to increased emissions from residen-
tial, small commercial, and small industrial sources rather than to the
large point sources. This may pose some problems in the formulation of
strategies for controlling high N02 levels.
The top curve of Figure 27 also shows monthly emission factors nor-
malized to mean monthly emissions. For example, June shows a factor of
0.98, which means that in June emissions are 98 percent of the monthly
mean. Thus, December and January respectively contain 129 and 139 per-
cent of mean monthly emissions.
The next step is to include the effect of mobile sources, since
they contribute between 60 and 70 percent of total NOX emissions on an
annual basis. The seasonal component of mobile-source emissions was
estimated from the literature for both light- and heavy-duty vehicles
(Goodman et al., 1977; Arledge and Tan, 1978). Unfortunately, the
available seasonal data for mobile sources were not given for each
64
-------
20
(NATURAL GAS AND REFINERY
GAS) COMBUSTION
0.98 0.79 0.83 0.86 0.99
18
16
14
to
O
10
Ul
h\
1.02
1.29
1.39
1.02
1.04
0.88
0.92
0.98
NORMALIZED MONTHLY
EMISSION FACTOR
... .
W
•-f—~4—- r
NOX FROM NATURAL
GAS COMBUSTION
NOX FROM OIL
COMBUSTION
NOx FROM
REFINERY GAS
RESIDENTIAL. SMALL
COMMERCIAL, AND
SMALL INDUSTRIAL
COMMERCIAL. INDUS-
TRIAL, AND INDUSTRIAL
REFINERIES
) ELECTRIC UTILITIES
JUN JUL AUG SEPT OCT NOV DEC
JAN
FEB
MAR
APR
MAY
JUN
Figure 27. Monthly variation of NOX emissions for the period July 1972—June 1973
for the South Coast Air Basin. (Adapted from Bartz, et al., 1974)
-------
month, but were instead aggregated by quarter. The stationary-source
data were similarly aggregated, and the quarterly factors for the three
source categories are given in Table 12 and compared in Figure 28. Note
that mobile-source emissions are out of phase with the stationary
sources. This enhances the impact of the latter in the winter months.
To estimate the composite effect of stationary and mobile sources,
we obtained emission estimates for the SCAB for 1976, applied the quar-
terly emission factors, and computed for each quarter: (1) the composite
factor for total N0x emissions; (2) the ratio of stationary to total
emissions; (3) the ratio of stationary to mobile-source emissions. The
estimated mean NO emissions for the three source types are given in
Table 13. Items (l)-(3) above are contained in Table 14. The fourth
column of Table 14 shows that emissions during January-March are about 5
percent higher than mean levels . It is apparent from the second and
third columns of the table that this slight enhancement is due to sta-
tionary sources. The table also indicates that two of the other quar-
ters exhibit hardly any change, and that there is a decrease of 5 per-
cent during July-September that is also driven by stationary sources.
The evidence points to enhanced NOX emissions driven by stationary
sources during the winter months. It is probable that the increase
would have been greater than 5 percent had the months been treated indi-
vidually or been aggregated differently, e.g., by keeping December and
January together. Unfortunately, we were constrained to use the quar-
terly aggregation used for the mobile sources. While 5 percent does not
seem significant, this is an average figure for the entire basin that
masks the impact at individual stations. Since the emissions are not
distributed uniformly, it is quite possible that certain areas of the
SCAB, particularly those with high residential and commercial density,
may experience wintertime increases in NOX emissions that are greater
than 5 percent.
More research is needed to elucidate the role of enhanced NO emis-
sions in inducing high N02 levels. The analysis described above pro-
vides a qualitative indication that the seasonal component of the emis-
sions is compatible with its counterpart for the N0 exceedances.
D . Association Between N02-Monitoring Stations
in the South Coast Air~Basin
Are NO 2 exceedances reported in the SCAB linked to individual sta
tions, or are they a manifestation of a regional phenomenon? This sec
tion attempts to answer this question by investigating the degree of
association between various stations.
Fifteen stations were selected for the analysis: Anaheim, Azusa,
Burbank, La Habra, Lennox, Long Beach, Los Angeles/San Pedro St., Los
Angeles /Gault St., Los Angeles/Westwood Blvd., Lynwood, Pasadena,
Pomona, Temple City, Riverside, and Whittier. The sites were selected
66
-------
Table 12
QUARTERLY FACTORS FOR NOX EMISSIONS
IN THE SOUTH COAST AIR BASIN
Period
Jan- Mar
Apr-Jun
Jul-Sep
Oct-Dec
Stationary
Sources*
1.15
0.93
0.83
1.10
Light-Duty
Vehiclest
1.00
1.04
1.01
0.95
Heavy- Duty
Vehicles*
0.93
1.03
1.07
0.97
-------
Table 13
ESTIMATED 1976 NOX EMISSIONS
IN THE SOUTH COAST AIR BASIN
Source
Category
Emissions
(tons/day)
Stationary
Light-duty vehicles
Heavy-duty vehicles
460*
6001
1801
Estimated from Bartz et al. (1974)
California Air Resources Board,
personal communication, 18 May 1979.
Table 14
QUARTERLY VARIATION OF NOX EMISSIONS
IN THE SOUTH COAST AIR BASIN FOR 1976
Period
Jan- Mar
Apr-Jun
Jul-Sep
Oct-Dec
Annua 1
Stationary/Total
Emissions Ratio
0.41
0.34
0.32
0.41
0.37
Stationary /Mobile
Emissions Ratio
0.68
0.53
0.48
0.68
0.59
Composite
Quarterly
Factor
1.05
1.00
0.95
1.01
—
on the basis of spatial coverage, their locations relative to prevailing
wind flows (see Figure 7), and quantity of data available.
The analysis begins by examining the pattern of occurrence of two
or more stations reporting at least one N02 exceedance the same day.
The degree of association between certain pairs of stations is then
estimated*
68
-------
exceedances were reported at one or more of the fifteen sites
on 351 different days during 1975-1977, which is 32 percent of the total
number of days in the three-year period. Some of these days had more
than one station recording an exceedance the same day. This is illus-
rated in Table 15, which lists the number of days when one or more sta-
tions reported an exceedance. The table shows that single-station
exceedances were reported on 105 days. Hence, multiple-station
exceedances occurred about 70 percent of the time (i.e., 246 days).
Thus, multiple reports of exceedances are more common than single-site
reports by a ratio of 2:1. Table 15 also shows that, as expected, the
frequency of multisite reports generally decreases, but the overall
decrease is somewhat erratic, showing increases when the number of sites
Table 15
DISTRIBUTION Of 15 SELECTED STATIONS
IN THE SOUTH COAST AIR BASIN
RECORDING N02 EXCEEDANCES THE SAME DAY
Number of
Stations
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Number of
Days
105
58
33
40
19
19
15
15
13
12
16
5
0
1
0
69
-------
is 4 and 11. The five days when 12 sites recorded exceedances are: 25
January 1975, 4 November 1975, 17 December 1975, 16 March 1976, and 14
February 1977- The day when 14 of the 15 sites reported exceedances was
23 November 1976. Note that these six days occur in the fall and
winter, which suggests that regional episodes of high N02 levels are
confined to this time of year. However, further analysis revealed that
days with single-site exceedances are frequent during the winter as
well, with 53 of 105 days falling during the months November through
March. This is, of course, a reflection of the seasonal pattern dis-
cussed in connection with Figure 26. It would be of interest to study
these six days to examine the weather patterns that resulted in so many
stations reporting N02 exceedances the same day, but we must leave this
for a future study.
Table 16 shows how frequently the various sites reported
exceedances on the same day as at least one other site; the table also
contains the total number of days when each site reported an exceedance.
It is evident that most of the sites reported exceedances in combination
with one or more stations. Only the sites at Los Angeles/Westwood
Blvd., Los Angeles/San Pedro St., Pasadena, and Temple City reported
more than ten days each when the exceedance did not coincide with any of
the other 14 stations. Thus, a significant degree of dependence among
the sites is apparent.
In connection with the question of which sites are most closely
associated, we examined the correlation between pairs of sites using 2x2
contingency tables as shown below.
2X2 CONTINGENCY TABLE FOR SITES A AND B
SITE B
SITE A
811
a21
a12
a22
The symbols E and E respectively denote the presence or absence of an
exceedance on a given day. The entry aj^ represents the number of days
when both sites A and B recorded an exceedance the same day, and entry
a22 contains the days when neither site recorded an exceedance the same
70
-------
Table 16
N02 EXCEEDANCES RECORDED
AT TWO OR MORE SOUTH COAST AIR BASIN SITES
Site Name
Los Angeles/Westwood Blvd.
Los Angeles/San Pedro St.
Pasadena
Long Beach
Whittier
Burbank
Lennox
Pomona
Los Angeles /Gault St.
Lynwood
Temple City
Azusa
La Habra
Anaheim
Riverside
Number of Days
with Multisite
Exceedances
144
140
142
129
113
111
99
71
68
63
44
58
53
47
14
Total Days
167
161
158
134
118
117
106
73
70
65
59
58
54
47
14
day. Entries aj^ and a.^ contain the number of days when one site, but
not the other, reported an exceedance.
Of the 105 possible site pairs, 22 were selected for study based on
their locations relative to the typical wind flows shown in Figure 7.
Contingency tables were obtained for the 22 pairs and tested for inde-
pendence using a standard chi-square test with one degree of freedom
(c.f. Hosteller and Rourke, 1973; Dixon and Massey, 1957). Another
measure of association revealed whether knowledge about the reported
exceedances at Site A gave information about exceedances at Site B. In
other words, given that Site A is in state E, what is the probability
71
-------
that Site B is in state E, and vice versa? This measure of association
is denoted by M, and is defined as follows:
M = PH/P]^
where
PU = au/N
Pl = (all + ai2)/N
P2 = (au + a21)/N
N =» aj^ 4- a^2 4- a2i 4- a22
Thus, M is the ratio of the probability that both A and B are in state
E, to the product of the probabilities that either site is in state E.
If Sites A and B are independent, then M * 1. If they are negatively
correlated, then M < 1, and if positively correlated, then M > 1. How-
ever, chance alone may cause M to exceed 1. Hence, for each contingency
table we computed a value f such that if the observed Pji > f, then we
can say that P^ is not due to chance with a confidence level of 95 per-
cent or better. (This is known as Fisher's test, see Langley, 1970.)
The computation of f was performed while holding Pj and P2 constant. We
also calculated a "bond coefficient" defined by:
bond coefficient « 100 x (Pu - f)/f
The bond coefficient measures the information gained about the probab-
bility of one site being in state E, given that the other is in state E.
The higher the bond coefficient, the higher the association between the
two sites. If the bond coefficient is zero or negative, then knowing
that one site is in state E yields no information about the state of the
other site.
Results are shown in Table 17, which contains the entries a^j of
the contingency tables, the chi-square statistic, the coefficient M, and
the bond coefficient. In the table, the station pairs are arranged in
descending order of the bond coefficient. Regarding the chi-square
statistic, note that for one degree of freedom the null hypothesis of
independence is rejected at a 95 percent confidence level if chi-square
is greater than 3.84.
Table 17 shows that the chi-square statistic leads to rejecting
independence for 20 of the 22 station pairs. Thus, Los Angeles/San
Pedro St. and Los Angeles/Gault St. appear to be independent, as do Lyn-
wood and Temple City. Note that the bond coefficient for each of these
two pairs of sites is zero and negative, respectively. By and large,
the bond coefficient, M, and the chi-square statistic agree on the pres-
ence of some association, but the bond coefficient quantifies it best.
72
-------
Table 17
CONTINGENCY TABLES AND MEASURES OF ASSOCIATION
BETWEEN SELECTED STATIONS IN THE SOUTH COAST AIR BASIN
Site A
La Habra
Long Beach
Pasadena
Lynwood
Azusa
Burbank
Long Beach
Burbank
Los Angeles/Westwood
Los Angeles /San Pedro
Los Angeles/San Pedro
Lennox
Burbank
Pasadena
Los Angeles/Westwood
Los Angeles/San Pedro
Los Angeles/San Pedro
Los Angeles/Westwood
Los Angeles/Westwood
Los Angeles/San Pedro
Los Angeles/San Pedro
Lynwood
Site B
Whittier
Anaheim
Azusa
Whittier
Lynwood
Los Angeles/Gault
Whittier
Pasadena
Lennox
Azusa
Lynwood
Lynwood
Los Angeles /San Pedro
Temple City
Pasadena
Long Beach
Pasadena
Los Angeles/Gault
Azusa
Los Angeles/Westwood
Los Angeles/Gault
Temple City
Contingency Table Entries
(Number of Days)
all
39
39
51
43
24
44
74
84
80
40
43
31
72
36
94
76
92
44
35
86
39
13
a!2
13
94
107
20
33
73
58
31
85
121
119
74
46
122
74
82
71
123
132
72
121
52
*21
74
7
7
74
41
26
42
74
26
16
18
32
89
21
62
55
68
24
21
77
30
46
a22
220
210
187
213
255
209
174
162
154
171
167
210
142
172
118
132
119
154
160
107
157
241
Bond
Coefficient
(%)
70
63
55
54
50
47
42
40
36
25
23
19
14
13
12
12
11
10
6
2
0
-19
M
2.30
2.23
1.96
2.04
2.29
1.89
1.68
1.62
1.58
1.54
1.51
1.63
1.32
1.40
1.25
1.27
1.23
1.35
1.30
1.14
1.23
1.19
Chi-
Square
47.65
46.94
49.95
39.98
23.55
32.90
47.80
52.62
45.28
15.81
15.71
12.02
15.00
8.20
15.39
11.92
13.35
8.65
4.96
4.90
3.25
0.35
U)
-------
The two sites most closely associated are La Habra and Whittier, fol-
lowed by Long Beach and Anaheim. Figure 7 demonstrates the plausibility
of the associations since La Habra (Site 18) and Whittier (Site 51) are
close to each other and on the path, of the same air trajectory, as are
Long Beach and Anaheim (although the association is less obvious).
Similar associations can be seen in Figure 7 for Pasadena (Site 27) and
Azusa (Site 2), Lynwood (Site 24) and Whittier (Site 51), and Azusa
(Site 2) and Lynwood (Site 24). The lack of association between Los
Angeles/San Pedro (Site 21) and Los Angeles/Gault (Site 23) is also evi-
dent in Figure 7, since the typical flows do not pass over both sites.
In general, the plausibility of all the associations quantified in Table
18 is confirmed by Figure 7.
It is interesting to note that upwind sites such as Los Angeles/
Westwood (Site 22) show only a weak association with downwind locations
such as Pasadena (Site 29) and Azusa (Site 2). This report has shown
that the latter are dominated by titration, whereas synthesis is most
frequent at the upwind sites; this suggests that NO? transport from
upwind to downwind sites is relatively unimportant In causing
exceedances at the downwind stations.
74
-------
VI ANALYSIS OF PEAK/MEAN RELATIONSHIP
A.
Introduction
The current national ambient air quality standard for NC^ is based
on the annual mean concentration. In considering an hourly standard, it
would be useful to know whether any relationship exists between extreme
values (such as the annual maximum) and the mean.
This section investigates the functional dependence between the
annual maximum and mean HO* using regression methods. The data used in
the analysis were obtained from publications issued by the GARB (1976;
1977a,b). All 51 original stations were considered in the analysis, but
some stations were excluded for individual years because of missing
data. The analyses were performed for individual years during 1975-1977
and for the three years combined.
B. Peak/Mean Relationship
The correlation between peak and mean values was calculated, and
linear regression equations were obtained for the peak as a function of
the mean. Table 18 contains the parameters of the regression. Scatter
diagrams of the calculated values are shown in Figures 29 through 31 for
the individual years and in Figure 32 for the pooled data.
Table 18
PARAMETERS OF PEAK/MEAN REGRESSION
Year
1975
1976
1977
1975-1977
Correlation
Coefficient*
0.85
0.80
0.83
0.82
Regression Line^
A
6.1
4.6
5.8
5.5
Intercept
B
(ppm)
0.04
0.08
0.02
0.05
Significance
p < 0.05
p < 0.002
None
p < 0.002
Number
of Points
49
47
41
137
Standard
Error of
Estimate
(ppm)
0.07
0.06
0.08
0.07
t
Significant for p < 0.0001
Regression line is Peak
B = intercept.
A * Mean + B, where A = slope and
75
-------
11 . 00
10.OO
.00
e.oo
O
UJ
1
7.00
6.00
8.00
4.00
a. oo
2.00
2 « *
* *
*
>
1.00 +s
.+ + +
9.OO 18.00
*
+ ---- -f ---- + ---- + ---- + .
•»•
+ + + * + + + + + + + + + + + + +.
21.OO 27.00 33.OO 39.OO 49.00 81.00 87.00 63.OO 69.OO
ANNUAL MAXIMUM NO2 — pphm
Figure 29. Scatter diagram of annual mean as a function of maximum NO2
for selected California sites in 1975. Number of points plotted is 49.
-------
.+ ---- + ---- + ---- +
11.00 *
10.00
9.00
.00
a
a
i 7.00
O
z e.oo
<
in
5
Z
8.00
4.0O
9.00
e.oo
«
• *
s
1.00 +
.+ ----
10.00
+ ---- + ---- + ---- + ---- + ---- * ---- + ---- + ---- + ---- + ---- + ---- +.
+
*
+ ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- * ---- + ---- +.
16.00 22.00 28.00 34.00 40.00 46.00 82.00 06.00 64.00 70.00
ANNUAL MAXIMUM NO2 — pphm
Figure 30. Scatter diagram of annual mean as a function of maximum NO2
for selected California sites in 1976. Number of points plotted is 47.
-------
. •»• ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- -f ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + ---- + .
•vl
oo
11 . 00
10.00
0.OO +
8.00
E
4
a
CM
o
Z
UJ
* s
*
« «
7.OO
6.00
8.00 +
4.0O
3.00 +
2.00
1.00 +« +
.+ + + + + + * + ••• + + + + + + + + + + + + .
10.0O 16.00 22.00 26.00 34.00 40.00 46.00 B2.OO S8.OO 64.00 70.00
ANNUAL MAXIMUM NO2 — pphm
Figure 31. Scatter diagram of annual mean as a function of maximum NO2
for selected California sites in 1977. Number of points plotted is 41.
-------
.+ •»• + * + + 1 + 1 * •»•---- + 1 1 + * + + 1 + •»•.
II.00 +
IO.OO
a. oo
a.oo
E
I
I 7.00
(M
O
2 6.00
IU
S
B.OO
4.00
3.00
2.00
• • *
• « •
21
13 2
• •
• » «
• 2 •
222 * •
• » • *
• 3 • •
• • « « 2 2
* • •» • •
* 2« 2
2* *
• 2 •
•
* •
I.00 +* •
.+ » -f •»• •»• + + + + + 1 1 1 1 ••• •»• 1 + +
9.OO 19.OO 21.OO 27.OO 33.00 39.OO 4B.OO 01.00 57.00 63.00
ANNUAL MAXIMUM NO2 — pphm
Figure 32. Scatter diagram of annual mean as a function of maximum N02
selected California sites, 1975-77. Number of points plotted is 137.
•»•
69.00
-------
It is evident from the scatter diagrams that a strong linear rela-
tionship links the peak and the mean. As Table 18 shows, the correla-
tions for individual years range from 0.80 in 1976 to 0.85 in 1975, and
for the pooled data it is 0.82. The narrow range of correlations and
slopes for the individual years indicates that the variability of the
data is small, so it appears to be reasonable to pool the annual data to
obtain an overall regression.
A question might be raised about the effect of data adjustments on
the regression. The correlation coefficient and the slope will not be
affected if the same adjustment factor were applied to all the data, but
the intercept and the standard error of estimate will be scaled by the
adjustment factor. On the other hand, if different adjustments are
applied to individual peak/mean pairs, the correlation coefficient and
the regression line will be affected. We cannot tell in general what
effect such nonuniform data adjustment might have on the regression: The
effects of nonuniform adjustments can be evaluated only by recomputing
the regression after the data have been adjusted.
Using the derived regression equation for the pooled data, if the
mean equals the national air quality standard of 0.05 ppm, the estimated
peak value will be 0.33 ± 0.07 ppm, apparently violating the California
hourly standard of 0.25 ppm, a limit not to be equalled or exceeded.
Thus, in California the hourly standard of 0.25 ppm is more restrictive
than the annual norm of 0.05 ppm, since satisfying the latter does not
imply compliance with the hourly limit. Adjusting the regression param-
eters by 10 or 17 percent changes the predicted peak value only slightly
to 0.32 ± 0.06 ppm, assuming that the 0.05 ppm mean value used in the
calculation represents an accurate measurement.
In assessing the utility of the regression function relating peak
and mean NC>2 concentration, whose output is in essence an average, it is
useful to examine the variability of the peak/mean ratio observed in the
data. Figure 33 shows a digital histogram of the ensemble of peak/mean
ratios for the pooled data. The ratios range from 3.3 at Bakersfield to
12.9 at Oceanside, with a median ratio of 6.3. The figure shows that
the ratios cluster tightly about a ratio of 6, with 96 of 137 ratios
ranging from 5.0 to 7.6.
For a mean value of 0.05 ppm, the regression equation relating peak
and mean NC^ yields an expected peak/mean ratio of 6.5 ± 0.6; the uncer-
tainty is computed from the standard error of the coefficients of the
regression equation. How does the estimated ratio compare with the
observed values? Examination of the data revealed that there are 11
peak/mean ratios where the mean is in the interval 0.045 _< mean <^ 0.054
ppm. These ratios are: 3.3, 5.2, 5.4, 5.6, 5.7, 5.8, 5.8, 6.2, 6.2,
7.4, and 9.3. Comparison of the estimated and observed peak/mean ratios
shows that the estimates, which are in the range 5.9 to 7.1, lie at the
upper end of the set of observed ratios. This implies that for a mean
80
-------
oo
o
3
1
o
H
1
PL,
CJ
O
a
3
4
5
6
7
8
9
10
11
12
3
2
0
0
0
0
1
3
0
6
6
4
0
0
0
2
1
4
4
9
5
0
0
1
2
1
4
5
6
0
0
1
2
3
9
7
DECIMAL DIGIT OF RATIO
Bakersfleld, .Fresno
67999
11112233334444555556777788888889999
0112222222222 3^3J3 3444445567788899
11111223344455556
4456788
455
7 9
Costa Mesa, Oceanside
Row Count;
( 2)
( 9)
(39)
(36)
(21)
(11)
( 7)
( 4)
( 6)
( 2)
Total
(137)
Figure 33. Frequency distribution of peak/mean ratio of NO2 for selected California
sites, 1975-1977.
Interpretation: 12|6 9 corresponds to ratios of 12.6 and 12.9 at Costa Mesa
and Oceanside, respectively. Similarly, 3J3 6 stands for ratios of 3.3 and 3.6
at the sites listed. Sites are not shown for other rows. The boxed digit
corresponds to the median ratio of 6.3
-------
of 0.05 ppra the regression, equation may tend to overestimate the peak
since the majority of the observed peak/mean ratios are lower than the
estimates. This conclusion applies only when the mean is 0.05 ppm, and
could change if another mean value were used.
82
-------
VII SUMMARY OF SPECIFIC FINDINGS
The study analyzed NO 2 concentrations exceeding 0.20 ppm recorded
in California during 1975-1977. Principal findings in the major inves-
tigative topics of the project are presented in the following sections.
A. Monitoring Site Features
Fifty-one monitoring stations originally reported NC>2 exceedances
during 1975-1977. This number was reduced to 48 after the data were
reviewed. All the sites are located in urban and suburban areas. Half
the stations are in the South Coast Air Basin.
Mobile-source emissions predominate at most monitoring sites.
Potentially significant local point sources such as power plants and
heavy industries were few in number. Probable effects from a steel mill
were identified at one site (Fontana). Probable source-receptor links
between stationary sources of NOX located outside the immediate vicinity
of the sites and several monitoring stations in the South Coast Air
Basin were identified. The sites at Long Beach and Whittier showed the
most pronounced stationary-source impacts•
B. Data Quality
The NO- data were found to contain very few anomalies. The data
for only 62 of approximately 1800 site-days were considered to be
anomalous and were eliminated from the data base. Based on the low
incidence of data errors noted above, data quality can be considered to
be good. However, according to a recent study by the California Air
Resources Board, the N02 levels are from 10 to 17 percent lower than the
measurements indicate.
Comparison of simultaneous colorimetric and chemiluminescent NO2
measurements at three sites revealed correlations of 0.71 and 0.86 at
Riverside and Upland, respectively, but no correlation at San Jose. The
colorimetric/cherailuminescent comparison also revealed a tendency for
the latter data to be higher than the colorimetric observations.
C. Physical Processes Linked to NOp Exceedances
N02 synthesis is the most common mechanism leading to high N02 in
densely urbanized areas. NO + Og titration also causes N02 exceedances,
but is about 2/3 as frequent as synthesis. Titration appears to be
ozone-limited, rather than NO-limited. N02 exceedances due to titration
are especially important at downwind sites where transported ozone
reacts with NO emitted by local sources. Pasadena and Pomona are exam-
ples of this type of site.
83
-------
Transport and point-source effects related to NOo exceedances were
relatively infrequent. Long Beach and Whittier exhibited the best exam-
ples of stationary source impacts.
D. Spatial and Temporal Variation
of NOo Exceedances
On a statewide basis, N02 levels exceeding 0.20 ppm ranged from
0.21 to 0.62 with a median concentration of 0.24 ppm. Over 5,400 site-
hours exceeding 0.20 ppm were recorded statewide during 1975-1977.
About 92 percent of the N02 exceedances were recorded in the South Coast
Air Basin, with Los Angeles County accounting for 80 percent of the
statewide total.
However, it was determined that between 34 and 46 percent of the
exceedances would be eliminated statewide if the NO2 data were adjusted
downward to compensate for the 10-17 percent bias. Such an adjustment
often has a much greater impact on individual stations.
NO2 exceedances have a strong seasonal component, occuring most
frequently during the period November through February. The seasonality
is evident statewide. Seasonal patterns of NOX emissions in the South
Coast Air Basin showed increased contributions from stationary area
sources during the winter months.
N©2 concentrations occurring during the time interval 0-0500 and
2200-2300 differs quantitatively from the distributions that prevail at
other times at six of nine sites examined. The N02 levels were lower
during 0-0500 and 2200-2300. Seven of nine stations tested showed no
statistically significant differences between the distribution of N02
exceedances occurring during 0600-1300 and that for 1400-2100.
The diurnal variation of N©2 exceedances appears to be associated
with the traffic cycle in the more densely urbanized areas.
Analysis of 15 selected stations in the South Coast Air Basin
revealed that the incidence of exceedances recorded at two or more sites
on the same day surpasses that of single-site exceedances by better than
a 2:1 margin. As a corollary to the preceding finding, a widespread
pattern of interstation correlations was found among the selected sta-
tions in the South Coast Air Basin. The correlations are consistent
with the typical flows that occur in the Basin.
E. Peak/Mean Relationship
Annual maximum and mean N02 were found to be correlated. The
correlation coefficient is 0.82 and is highly statistically significant.
The peak/mean ratio ranged from 3.3 to 12.9, with a median ratio of 6.3.
84
-------
VIII CONCLUSIONS
High N02 concentrations seem to be primarily an urban phenomenon.
This conclusion is reached bearing in mind that N02 monitors are
clustered in urban areas, which raises the possibility that high N02
levels in rural areas may go undetected. Nevertheless, the pattern of
N02 exceedances throughout California strongly suggests that a high den-
sity of emissions of NO and hydrocarbons are required for synthesis to
result in N02 exceedances. Furthermore, since N02 exceedances due to
titration require high levels of both NO and 03, titration effects will
probably be insignificant in rural areas where either NO or 0^ or both
tend to be low. In fact, although titration-related exceedances in
urban areas generally are ozone-limited, they would tend to be NO-
limited in rural areas. N0« transport from urban areas and from point
sources is thus the sole remaining mechanism that could lead to N02
exceedances in rural areas.
The evidence of this study suggests that N02 from urban areas is
not transported over long distances, certainly not over distances as
long as those associated with ozone transport. Examples of this situa-
tion are found in southern California when one considers well-known
ozone receptors such as Upland, Riverside, and Palm Springs. Upland and
Riverside show relatively few N02 exceedances, and Palm Springs shows
none, yet all are heavily impacted by ozone transport from the Los
Angeles area. Point sources of NOX in rural areas could help to promote
N02 exceedances if the NO-rich plume from the point source encounters an
ozone-laden urban plume, thereby producing N02 by titration of NO and
Og. However, being conditioned on the vagaries of meteorological condi-
tions, such an occurrence is likely to be infrequent. Nevertheless, it
remains a distinct possibility whose chances would be enhanced if point
sources of NO were to proliferate in rural areas. On balance, however,
the evidence examined indicates that high N02 concentrations are essen-
tially confined to urban areas and their immediate surroundings.
A second item of interest from a regulatory standpoint is the
seasonality of the frequency of N02 exceedances. The seasonal pattern
was apparent on a statewide basis, consisting of an increased exceedance
frequency during the period November-February. Although the more stag-
nant conditions that prevail in California during these months are cer-
tainly a contributing factor, the possibility exists that the high N02
levels may be enhanced by increased NOX emissions from space heating
using natural gas. Analysis of seasonal patterns of NOX emissions in
the South Coast Air Basin, where the seasonality of the N02 exceedances
is most pronounced, tends to support this hypothesis. However, more
research is needed to define better and to quantify the relationship
between seasonal fluctuations of NO emissions and N02 concentrations.
Recommendations in these areas are given in the next section.
Few anomalies were found in the data base; only about 3 percent of
the site-days containing exceedances were rejected on the basis of
85
-------
suspected errors. Thus, in this respect the data is of good quality.
It is strongly recommended that data-screening procedures employing
graphical methods such as were used in this study continue to be applied
in the future, since this increases the confidence in the results
obtained.
More important than the presence of anomalous data is the recent
discovery by the California Air Resources Boar4 that N02 levels in the
state are between 10 and 17 percent lower than indicated by the measure-
ments. On a statewide basis, adjusting the data by these factors elim-
inates between 34 and 46 percent of all the exceedances. It also has
the effect of further confining the problem of high NOj levels to the
South Coast Air Basin in general and to Los Angeles and Orange Counties
in particular. Thus, any contemplated regulatory actions must specify
the adjustment factor to be applied to the data.
A disturbing result of the data quality assessment was discovered
in the comparison of simultaneous colorimetric and chemiluminescent
measurements. The analysis indicated that the correlation was 0.71 at
Riverside and 0.86 at Upland, which is respectable but perhaps not quite
as high as one might expect. One site, San Joge, showed no significant
correlation between the measurements. More serious is the fact that, in
general, the, chemiluminescent measurements tended to be greater than the
colorimetric observations, which leads one to ask which one is correct.
The question cannot be answered from the data available to us, but it is
imperative tjhat an answer be obtained in view of the regulatory implica-
tions of the choice of measurement method.
Pursuant to federal regulations (Federal Register, 10 May 1979),
NO2 measurements in California as of January 1980 will be performed
using chemiluminescent instruments only. (All states must discontinue
using colorimetric methods for Np2» but the compliance dates vary.) The
discrepancies between colorimetric and chemiluminescent measurements
noted above suggest that these two types of data may not be strictly
comparable. Since implementation plans are baaed on analyses of histor-
ical data, ^.ack of data comparability makes it difficult to monitor pro-
gress toward compliance with ambient NO2 standards. Thus, additional
efforts must; be devoted to establishing the degree of correspondence
between colorimetric and chemiluminescent NO2 measurements.
The comparison of the distributions of N0« concentrations for vari-
ous times of day revealed that N02 levels observed during the hours 0-
0500 and 2200-2300 (PST) were significantly different from concentra-
tions measured during 0600-2100. In general, the nighttime levels were
lower than £he daytime values. The lower concentrations that exist at
night indicate that NO2 has a short lifetime, further evidence pointing
to the relative lack of importance of N02 transport from urban areas as
a contributor to elevated N0£ in rural areas.
The In^erstation correlation with respect to same-day N02
exceedances between various pairs of sites in the South Coast Air Basin
86
-------
suggests that short-term effects dominate the formation and transport of
high N02 levels. Moreover, the correlation indicates that area sources,
rather than point sources, are the principal proximate causes of the
elevated NO2 levels in the Basin.
The derived linear-regression equation relating annual maximum and
mean NC>2 implies that the current California hourly standard of 0.25 ppm
is more restrictive than the national annual standard of 0.05 ppm, since
meeting the annual standard does not necessarily guarantee satisfying
the hourly limit. Consequently, if a national hourly standard equal to
California's 0.25 ppm were established, then the hourly rather than the
annual standard would become the controlling factor in abatement
efforts, assuming that the annual standard remains unchanged. Thus, in
setting a national hourly NC^ standard, it is necessary to consider the
relationship between peak and mean N02 to ensure that the two standards
reinforce each other.
87
-------
IX RECOMMENDATIONS FOR FURTHER RESEARCH
In addition to the specific results reported in the previous sec-
tion, this study revealed other areas that should be the subject of
further investigation:
• A more detailed investigation of the relationship between sea-
sonal emissions of NO and elevated NO2 levels is required. In
particular, it could be fruitful to attempt to relate degree-
days and N02 levels, correlate degree-days with NOX emissions,
and compare the two.
• Regional-scale episodes of elevated N02 levels occurred in the
South Coast Air Basin on six days that are identified in the
text of the report. Meteorological patterns on these days
should be investigated to identify the weather conditions that
led to such episodes.
• About a third of all the N0« exceedances reported in the South
Coast Air Basin were associated with single stations. It would
be of interest to examine a sample of these cases to establish
why the elevated N0~ was observed at a given station, and at no
other station, on a given day.
• The data suggest that there may be weekend/weekday or
Sunday/weekday differences in N02 analogous to those found for
ozone. It is recommended that this topic be investigated
further•
• Several cases of high NOo related to point sources were identi-
fied at Whittier and Long Beach. Such instances often combined
synthesis and point-source effects. It is recommended that
attempts be made to establish the relative contributions of the
various processes to the overall N02 level in order to isolate
point-source effects.
• More analyses and comparisons of simultaneous colorimetric and
chemiluminescent measurements of N02 should be made at the high
end of the concentration range and under field, rather than
laboratory, conditions.
• The frequency distributions of N02 levels exceeding 0.20 ppm
occurring in summer and winter should be compared to determine
whether significant differences exist between them. Further
study is also required to establish the seasonal variation, if
any, of physical processes such as chemical synthesis and titra-
tion.
• A more extensive investigation should be undertaken of peak/mean
relationships for N02. Such an investigation should include
data from other states.
89
-------
REFERENCES
Arledge, K.W. and R.L. Tan, 1978: "A Heavy-Duty Vehicle Emission Inven-
tory System," Final Report, Contract No. A6-051-87, TRW Environmen-
tal Engineering Division, Redondo Beach, California (March)
Bartz, D.R., K.W. Arledge, J.E. Gabrielson, L.G. Hays, S.C. Hunter,
1974: "Control of Oxides of Nitrogen from Stationary Sources in the
South Coast Air Basin of California," Report No. KVB-5800-179, KVB
Inc., Tustin, California, NTIS No. PB-273688 (September).
California Air Resources Board. 1976: "Summary of 1976 Air Quality
Data—Gaseous Pollutants," Annual Summary, Vol. VIII.
California Air Resources Board, 1977a: "Three-Year Summary of California
Air Quality Data, 1973-1975" (January).
California Air Resources Board, 1977b: "Summary of 1977 Air Quality
Data—Gaseous and Particulate Pollutants," Annual Summary, Vol. IX.
California Air Resources Board, 1979: "California Air Quality Data,"
Vol. IX, No. 1.
De Marrais, G.A., G.C. Holzworth, and C.R. Hosier, 1965: "Meteorological
Summaries Pertinent to Atmospheric Transport and Dispersion Over
Southern California," Technical Paper No. 54, U.S. Weather Bureau,
' Washington, D.C.
Dixon, W.J. and F.J. Massey, 1957: Introduction to Statistical Analysis,
2nd edition, McGraw-Hill Book Company (New York, New York).
Federal Register, 1979: Vol. 44, No. 92, pp. 27558-27604 (10 May).
Goodman, H.S., G.E. Abercrombie, K.W. Arledge, and R.L. Tan, 1977: "A
Mobile Source Emission Inventory System for Light Duty Vehicles in
the South Coast Air Basin," Final Report, Contract No. ARE 4-1236,
TRW Environmental Engineering Division, Redondo Beach, California
(February).
Hunter, S.C. and N.L. Helgeson, 1976: "Control of Oxides of Sulfur from
Stationary Sources in the South Coast Air Basin of California,"
Report No. KVB-5802-432, KVB, Inc., Tustin, California (June), NTIS
No. PB 261754.
Langley, R., 1970: Practical Statistics. Dover Publications (New York,
New York).
Mosteller, F. and R.E.K. Rourke, 1973: Sturdy Statistics. Addison-Wesley
Publishing Company (Reading, Massachusetts).
91
-------
Thuillier, R.H. and W. Viezee, 1978: "Air Quality Analysis in Support of
a Short-Term Ambient Air Quality Standard for Nitrogen Dioxide,"
Final Report, SRI Project 6780, EPA Contract 68-02-2835, SRI Inter-
national, Menlo Park, California (February).
92
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-450/4-79-034a
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Analysis of High NG*2 Concentrations
in California, 1975-1977
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
J. R. Martinez and K. C. Nitz
8. PERFORMING ORGANIZATION REPORT NO.
6780-12 and -13
g. 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
Envirpnmental 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. Rienter
16. ABSTRACT
During the period 1975-1977, 51 monitoring stations in California collectively
recorded about 1,800 site-days in which hourly nitrogen dioxide (N02) concentra-
tions exceeded 0.20 ppm. This work investigates potential causes of these high
N02 events, the physical phenomena involved in their occurrence, and their
spatial and temporal patterns. In addition, the potential association between
emission sources and the frequency and magnitude of high N02 levels at the
various locations is analyzed using detailed site-description data compiled in
this study. The relationship between annual maximum hourly levels and annual
mean concentration is explored, and the quality of the N02 data is evaluated.
Appendix A (unpublished) is a compilation of data curves used in the analysis.
Appendix B (unpublished) is a compilation of descriptions of the sites from
which the data came. Neither of these is required to understand the basic
report text.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI E;ield/Group
Nitrogen Dioxide
California
Short-Term Standard
N02, Spatial and Temporal Variations
N02, Causes of High Concentrations
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
108
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
------- |