EPA-600/3-78-019
February 1978 Ecological Research Series
cNVlRONMENT/\Jk
PROTECTION
AGSr.'CY
DALLAS . ifXA8
YERIFICATWH OF THE ISOPLETH METHOD
FOR RELATIN6 PHOTOCHEMICAL
OXIDANT TO PRECURSORS
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 2771!
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8 "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems are assessed for their long- and short-term influ-
ences. Investigations include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-78-019
February 1978
VERIFICATION OF THE ISOPLETH METHOD FOR RELATING
PHOTOCHEMICAL OXIDANT TO PRECURSORS
BY
J. Trijonis and D. Hunsaker
Technology Service Corporation
2811 Milshire Boulevard
Santa Monica, CA 90403
Contract No. 68-02-2299
Project Officer
Basil Dimitriades
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for pub-
lication. Approval does not signify that the contents necessarily re-
flect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
ii
-------
ABSTRACT
Historical trend data for the Los Angeles region are used to check
the ozone isopleth method that has been proposed as a replacement for the
Appendix J model. Using the median 6-9 AM NMHC/NOX ratio measured during
the summer as an input to the isopleth model, significant discrepancies
are found between the isopleth predictions and actual oxidant trends.
Most of these discrepancies are statistically significant considering
statistical errors in the actual oxidant trends and potential errors in
our estimates of precursor trends. Using a range in the NMHC/NOX ratio,
in particular a low value for the ratio, much better agreement is found
between the predicted and actual trends. Potential explanations for the
discrepancies and possible improvements to the isopleth model are discussed.
111
-------
CONTENTS
ABSTRACT Hi
FIGURES vi
TABLES ix
1. INTRODUCTION AND SUMMARY 1
Validating the Isopleth Method 3
Summary of Conclusions 4
Recommendations for Future Work 5
2. HISTORICAL OXIDANT AIR QUALITY TRENDS 7
Selection of Monitoring Sites and Air Quality Indices . . 7
Error Bounds on Oxidant Trends 11
Oxidant Trend Data 12
3. DATA ON HISTORICAL PRECURSOR TRENDS AND THE NMHC/NOX RATIO . . 19
Basinwide Analysis for Los Angeles 19
Analysis of Individual Locations 36
Analysis of Critical Assumptions 53
4. VALIDATION OF THE ISOPLETH
METHOD AGAINST HISTORICAL TREND DATA 63
Validation of Basinwide Isopleths 63
Validation of Isopleths for Fixed Irradiation Times ... 75
5. DISCUSSION OF RESULTS 98
Factors Not Accounting for the Disagreement 98
Possible Explanations for the Disagreement 100
REFERENCES 109
APPENDICES
A. Table of oxidant trend data 112
-------
FIGURES
Number Page
1 Sensitivity of Maximum Afternoon Ozone Concentrations
to Precursor Concentrations 2
2 Oxidant Monitoring Sites Studied in the Los Angeles Region. . . 8
3 Trends in the Basinwide Second Maximum 13
4 Trends in the Second-Highest One-Hour and the 95th
Percentile of Daily Maxima at Azusa 14
5 Trends in the 95th Percentile of the Daily Maxima at DOLA ... 15
6 Trends in the 95th Percentile of the Daily Maxima at Anaheim. . 16
7 Trends in the 95th Percentile of the
Daily Maxima at San Bernardino 17
8 Streamlines for the Westerly Flow Pattern 21
9 Streamlines for the Diurnal South Pattern 22
10 Streamlines of Most Frequent Surface Winds During July 23
11 Approximate Source Area Affecting Basinwide
Oxidant Maximum in the Los Angeles AQCR 24
12 Total NOX Emission Trends in the Los Angeles Basin 25
13 Total RHC Emission Trends in the Los Angeles Basin 26
14 Geographical Distribution of Percent Changes in Population
in the Los Angeles Basin, 1965 to 1975 28
15 Best Estimates of Historical Precursor Trends in the
Source Region for the Basinwide Oxidant Maximum 33
16 Approximate Source Area Affecting the Oxidant Maximum
at Downtown Los Angeles 37
17 Best Estimates of Historical Precursor Trends in
DOLA Source Region 41
18 Approximate Source Area Affecting the Oxidant Maximum at
Anaheim 43
19 Best Estimates of Historical Precursor Trends
in the Anaheim Source Region 47
20 Approximate Source Area Affecting the Oxidant Maximum
at San Bernardino 49
21 Best Estimates of Historical Precursor Trends
in the San Bernardino Source Region 52
22 Frequency Distribution of Vector-Averaged Wind Direction
(7 AM - 2 PM) at Downtown Los Angeles 55
vi
-------
Number Page
23 Frequency Distribution of Vector-Averaged Wind Speed
(7 AM - 2 PM) at Downtown Los Angeles 56
24 Frequency Distribution of 6-9 AM NMHC/NOX Ratio at
Downtown Los Angeles 61
25 Prediction of Oxidant Trends for the 95th Percentile
at Azusa Using Basinwide Isopleths 65
26 Oxidant Trends in the 95th Percentile of the Daily Maxima
at Azusa, Predicted for 7:1 Ratio vs. Actual 67
27 Oxidant Trends in the 95th Percentile of the Daily Maxima
at Azusa, Predicted for 12:1 Ratio vs. Actual 68
28 Oxidant Trends in the 95th Percentile of the Daily Maxima
at Azusa, Predicted for 23:1 Ratio vs. Actual 69
29 Summary of Oxidant Trends in the 95th Percentile
at Azusa, Predicted vs. Actual 70
30 Oxidant Trends in the Second Maximum for Azusa,
Predicted for 7:1 Ratio vs. Actual 71
31 Oxidant Trends in the Second Maximum for Azusa,
Predicted for 12:1 Ratio vs. Actual 72
32 Oxidant Trends in the Second Maximum for Azusa,
Predicted for 23:1 Ratio vs. Actual 73
33 Summary of Oxidant Trends in the Second Maximum for Azusa,
Predicted vs. Actual 74
34 Oxidant Trends in the Basinwide Second Maximum,
Predicted for 7:1 Ratio vs. Actual 76
35 Oxidant Trends in the Basinwide Second Maximum,
Predicted for 12:1 Ratio vs. Actual 77
36 Oxidant Trends in the Basinwide Second Maximum,
Predicted for 23:1 Ratio vs. Actual 78
37 Summary of Oxidant Trends in the Basinwide Second Maximum,
Predicted vs. Actual 79
38 Oxidant Trends in the 95th Percentile of the Daily Maxima
at DOLA, Predicted for 7:1 Ratio vs. Actual 80
39 Oxidant Trends in the 95th Percentile of the Daily Maxima
at DOLA, Predicted for 12:1 Ratio vs. Actual 81
40 Oxidant Trends in the 95th Percentile of the Daily Maxima
at DOLA, Predicted for 23:1 Ratio vs. Actual 82
41 Summary of Oxidant Trends in the 95th Percentile at DOLA,
Predicted with 5-Hour Isopleths vs. Actual 83
vii
-------
Number Page
42 Oxidant Trends in the 95th Percentile of the Daily Maxima at
Anaheim, Predicted for 7:1 Ratio vs. Actual 85
43 Oxidant Trends in the 95th Percentile of the Daily Maxima at
Anaheim, Predicted for 12:1 Ratio vs. Actual 86
44 Oxidant Trends in the 95th Percentile of the Daily Maxima at
Anaheim, Predicted for 23:1 Ratio vs. Actual 87
45 Summary of Oxidant Trends in the 95th Percentile at
Anaheim, Predicted with 5-Hour Isopleths vs. Actual 88
46 Oxidant Trends in the 95th Percentile of Daily Maxima
at Azusa, Predicted for 7:1 Ratio vs. Actual 89
47 Oxidant Trends in the 95th Percentile of Daily Maxima
at Azusa, Predicted for 12:1 Ratio vs. Actual 90
48 Oxidant Trends in the 95th Percentile of Daily Maxima
at Azusa, Predicted for 23:1 Ratio vs. Actual 91
49 Summary of Oxidant Trends in the 95th Percentile at Azusa,
Predicted with 7-Hour Isopleths vs. Actual 92
50 Oxidant Trends in the 95th Percentile of Daily Maxima at
San Bernardino, Predicted for 7:1 Ratio vs. Actual 94
51 Oxidant Trends in the 95th Percentile of Daily Maxima at
San Bernardino, Predicted for 12:1 Ratio vs. Actual 95
52 Oxidant Trends in the 95th Percentile of Daily Maxima at
San Bernardino, Predicted for 23:1 Ratio vs. Actual 96
53 Summary of Oxidant Trends in the 95th Percentile at San
Bernardino, Predicted with 9-Hour Isopleths vs. Actual ... 97
54 Oxidant Trends in the Second Maximum for Azusa, Predicted
for 7:1 Ratio vs. Actual, Predicted Values Based on
DOLA Source Region 104
55 Oxidant Trends in the Second Maximum for Azusa, Predicted
for 12:1 Ratio vs. Actual, Predicted Values Based on
DOLA Source Region 105
56 Oxidant Trends in the Second Maximum for Azusa, Predicted
for 23:1 Ratio vs. Actual, Predicted Values Based on
DOLA Source Region 106
57 Summary of Oxidant Trends in the Second Maximum for Azusa,
Predicted vs. Actual, Predicted Values Based on DOLA
Source Region 107
viii
-------
TABLES
Number Page
1 Average Yearly Values and Year-to-Year
Deviations for Oxidant Air Quality Indices 10
2 Percent Occurrence of Air Flow Patterns During
July to September in Los Angeles 20
3 Trends in Ambient NOX in the Source Area for the
Los Angeles Basinwide Oxidant Maximum 30
4 Trends in Ambient NMHC in the Source Area for the
Los Angeles Basinwide Oxidant Maximum 31
5 Best Estimates of Precursor Trends in the Source Area
for the Basinwide Oxidant Maximum 32
6 Ambient 6-9 AM NMHC/NOX Ratios 35
7 Trends in Ambient NOX in the Source Area for
Downtown Los Angeles 38
8 Trends in Ambient NMHC in the Source Area for
Downtown Los Angeles 39
9 Best Estimates of Precursor Trends for the DOLA
Source Region 40
10 Trends in Ambient NOX in the Source Area for Anaheim 44
11 Trends in Ambient NMHC in the Source Area for Anaheim 45
12 Best Estimates of Precursor Trends for the Anaheim
Source Region 46
13 Trends in Ambient NOX in the Source Area for San Bernardino. . 50
14 Trends in Ambient NMHC in the Source Area for
San Bernardino 51
15 Best Estimates of Precursor Trends for San Bernardino
Source Area 53
16 Comparison of Alternative Ambient Trend Indices for NMHC. ... 58
17 Comparison of Alternative Ambient Trend Indices for NOX 59
18 Summary of Actual and Predicted Oxidant Changes
1965 to 1974 (NMHC/NOX Ratio of 12:1) 99
19 Summary of Actual and Predicted Oxidant Changes,
1965 to 1974 (NMHC/NOX Ratio of 7:1 and 12:1) 101
ix
-------
CHAPTER 1
INTRODUCTION
In recent years Appendix J to Chapter IV, Part 51, Title 42 of the
Code of Federal Regulations has been used to estimate the amount of hydro-
carbon control needed to attain the National Ambient Air Quality Standard
for photochemical oxidant. The Appendix J model is based on an upper-limit
curve relating maximum afternoon oxidant concentrations to morning hydro-
carbon concentrations observed at the same location. In spite of its
widespread use in the past, however, Appendix J has come under increasing
criticism for its limitations [l, 2]:
t The role of NOX in oxidant formation is neglected.
• Relating oxidant and hydrocarbons at the same location
neglects transport of the air mass.
• The observed relationship between oxidant and hydrocarbons
may be distorted by unaccounted for meteorological variables.
• The upper-limit curve is not statistically well defined and
no error bounds are provided.
• Background levels of oxidant and hydrocarbons are ignored.
• Emissions occurring after 9 AM are neglected.
• The effect of the spatial/temporal distribution of emissions
is not accounted for.
Because of these shortcomings, various alternatives to Appendix J have
been proposed. One of the most attractive alternatives involves oxidant
isopleths derived by the Empirical Kinetic Modeling Approach (EKMA)[l]. The
EKMA isopleth method offers several advantages over the Appendix J procedure.
First, the EKMA isopleths are based on a chemical-kinetic model calibrated
to smog chamber data and hence represent a cause-and-effect relationship be-
tween oxidant and precursors. Second, the effect of transport is implicitly
included. Third, NOX is explicitly considered as an oxidant precursor.
Fourth, estimates of error bounds are possible. Fifth, the EKMA isopleth
model can be modified, if necessary, to account for the effects of back-
ground oxidant, background precursors, and post 9 AM emissions.
Figure 1 presents a series of EKMA isopleths for the basinwide oxidant
maximum during a 9-hour irradiation. Note that these isopleths are not
1
-------
to
ID
0)
U
O
o
I/)
i-
U
9)
Q.
0)
o
o
9)
O
IM
O
c
0)
E
•r—
X
I/)
(U
oo
O)
s_
3
D)
(Wdd) XON
-------
intended to be used in an absolute sense; rather, they should be interpreted
as representing the sensitivity of oxidant maxima to changes in presursor
concentrationsf1]. The isopleths can be used to predict future oxidant con-
centrations based on percent changes in presursor concentrations.
In order to predict future oxidant maxima, the isopleth approach re-
quires three basic inputs: the present oxidant maximum (or second maximum),
the present NMHC/NOV ratio, and the future degree of hydrocarbon and NOV
X A
control. For instance, as shown in Figure 1, for a present oxidant maximum
of 0.50 ppm, NMHC/NOV ratio of 9.5, hydrocarbon decrease of 40%, and NO
X A
increase of 20%, the predicted regionwide maximum would be 0.42 ppm.
VALIDATING THE ISOPLETH METHOD
The EKMA isopleth method should be subjected to validation studies
before it is accepted as an accurate method for evaluating oxidant control
strategies. Since the method is used in a relative sense to estimate the
sensitivity of oxidant to changes in precursors, the most appropriate
validation tests would involve historical changes in air quality, i.e.
historical trend data. This report tests the EKMA model by "predicting"
historical oxidant maxima based on past changes in precursors and comparing
these predictions to actually observed oxidant maxima.
Testing the model against trends requires several years of historical
data on oxidant and precursors. Also, since the location of the regionwide
oxidant maximum may change with time, good spatial coverage is necessary in
the historical air quality data. Based on these criteria, we chose the Los
Angeles basin and the time period 1964-1975 for the analysis. Only for this
region and time period can one find high-quality, long-term trend data with
excellent spatial resolution.*
In this report, the regionwide isopleth model is tested against trends
in the basinwide oxidant maximum for Los Angeles. To provide greater gener-
__
It was originally planned that Denver and Chicago be included in the
study. These sites were subsequently excluded because of the sparsity of
trend data for emissions and air quality, and because of uncertainties in
these data.
-------
ality in validating the isopleth approach and to increase the number of test
cases, four individual sites are also chosen for analysis: Downtown Los
Angeles (DOLA), Anaheim, Azusa, and San Bernardino. The trends at the four
individual locations are tested against isopleths that are specific to the
time of occurrence of maximal oxidant at those locations.
All of the validation studies cover the time period 1965-1974, with
tests made at every three-year interval. In order to provide robust data
sets for the analysis, three-year averages (1964-1966, 1967-1969, 1970-1972,
and 1973-1975) of air quality data are used.
SUMMARY OF CONCLUSIONS
Based on a variety of data sources, we are fairly confident that the
(6-9 AM summertime) ambient NMHC/NOX ratio was approximately 12:1 in 1965.
The validation studies using this ratio indicate significant discrepancies
between historical air quality trends and the predictions of the EKMA isopleth
method. The basic disagreement is that the isopleth method underestimates
the historical reductions that have occurred in maximal oxidant basinwide
and in oxidant at DOLA, Anaheim, and Azusa. Considering the statistical
errors in actual oxidant trends and the potential errors in our estimates of
precursor trends, the discrepancies between actual and predicted trends
(for a 12:1 ratio) are significant.in four of the seven situations
analyzed.
If we consider a range in the NMHC/NOX ratio, in particular the pos-
sibility that the ratio may have been as low as 7:1 in 1965, most of the
discrepancies become statistically insignificant. Using a 7:1 ratio, good
agreement is found between actual and predicted trends in all cases except
Anaheim.
We have investigated several factors which might contribute to the
discrepancies between the isopleth predictions and actual oxidant trends.
Some of these factors have been eliminated as plausible explanations for the
disagreement. Factors that apparently do not account for the disagreement
include the following: (1) the median NMHC/NOX ratio is slightly greater
on high oxidant days than on all summertime days; (2) our data for yearly
-------
average NOX trends underestimate the historical increase in 6-9 AM summer-
time NOX concentrations; and (3) historical trends may have been affected
by changes in monitoring practices.
There are several factors that could account for the discrepancies
between isopleth predictions and historical oxidant trends. The three most
likely explanations are as follows:
• A 12:1 atmospheric NMHC/NOX ratio may be equivalent to a
lower ratio in the isopletn model. In particular, a given
level of ambient NMHC may be equivalent to a lower level of
NMHC in the isopleths. This would be the case if am-
bient NMHC were of lower reactivity (per ppmc) than the
isopleth NMHC mix.
• The present versions of the EKMA isopleths neglect the effect
of emissions after 9 AM. Adding post 9 AM emissions
to the EKMA model might change the shape of the isopleths and
produce better agreement in predicting historical trends.
We would expect better agreement because the inclusion of
post 9 AM emissions would give greater emphasis to the ozone
inhibition role played by NOX emissions increases.
• The isopleth method may underpredict the actual oxidant im-
provement from 1964-1966 to 1973-1975 because of meteorological
bias in the actual oxidant trends. There is some evidence
that pollution potential in Los Angeles appeared to be lower
in 1973-1975 than in 1964-1966.
A fourth possible explanation is that our source areas have not been pro-
perly defined. The historical precursor changes of consequence to maximal
oxidant may be the precursor changes in the sub-areas of greatest emissions
density (which have low growth rates) rather than the precursor changes
throughout the entire upwind source areas. Limited spatial coverage of
oxidant monitoring sites is another factor that could account for some of
the discrepancies in the tests involving the basinwide isopleths .
RECOMMENDATIONS FOR FUTURE WORK
There at least three analyses that should be performed to isolate the
cause of the observed discrepancies and, possibly, the improve the isopleth
method:
-------
The reactivity of ambient 6-9 AM NMHC in Los Angeles should
be compared to the reactivity of the isopleth NMHC mix. This re-
activity comparison should consider both the number of moles
per ppmc and the oxidant producing potential per mole of
hydrocarbons.
EKMA isopleths should be prepared which include post 9 AM
emissions. The verification studies should be repeated with
these new isopleths.
It would be useful to normalize the actual oxidant trends in
Los Angeles for meteorological variance. This would eliminate
meteorological bias in the trends and would also decrease the
statistical error bounds on the actual oxidant trends, result-
ing in a more finely-tuned validation study.
-------
Variance in Air Quality Indices
As an aid in selecting air quality indices for measuring oxidant trends,
we conducted an analysis of the year-to-year statistical variance in alterna-
tive oxidant indices. Our goal was to identify an air quality index that is
representative of high oxidant days but has low relative year-to-year variance.
The left hand column of Table 1 lists the oxidant air quality indices
that we considered. For each index, and for each of the eleven monitoring
sites, we computed the mean value of the index and a de-trended standard
deviation*using data from 1965 to 1975. Table 1 lists the mean value of
each index and the year-to-year deviation, both averaged over all eleven sites.
Table 1 indicates that the single ./early maximum value exhibits the
highest relative deviation from year-to-year, - 17.0%. An index that is
representative of high oxidant days, but which has a relatively low variance,
is the 95th percentile of daily maximum one-hour concentrations. Many of our
analyses will be based on this latter index.
Basinwide Maximum
The EPA isopleth procedure [1] calls for the use of the second-highest
yearly one-hour oxidant at the station(s) under consideration. This convention
has been adopted because of the form of the oxidant standard which prohibits
more than one violation each year. In order to test the EPA isopleth procedure
in its conventional form, we will conduct the basinwide verification using the
second-highest oxidant concentration each year.
Among the eleven monitoring sites, Azusa exhibited the greatest second-
highest one-hour oxidant in nine of the twelve years (1964-1975). The string
of Azusa worst-cases is broken only by Downtown Los Angeles (DOLA) in 1965,
Pomona in 1968, and DOLA in 1973. In each of those three years, Azusa ranks
as the second-worst station.
Basinwide trends in the second-highest one-hour oxidant can be studied
in two ways. First, we could select, each year, the specific station which
exhibited the greatest second-hiighest oxidant, i.e., Pomona in 1968, DOLA in
1965 and 1973, and Azusa in all other years. Second, recognizing that Azusa
*
This is the standard error away from a least-squares trend line from
1965-1975, adjusted for degrees of freedom.
9
-------
Table 1. Average Yearly Values and Year-to-Year
Deviations for Oxidant Air Quality Indices.
OXIDANT AIR QUALITY INDEX
MEAN VALUE OF INDEX
FOR 11 LOCATIONS
DURING 1965-1975
IPPhm)
DE-TRENDED YEAR-
TO-YEAR DEVIATION
AVERAGED OVER
11 LOCATIONS
(as % of mean value
for the index)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
99th Percenti le of
All Hours
95th Percent! le of
All Hours
Annual Mean of All
Hours
Yearly Maximum 1-Hour
Second Highest 1-Hour
99th Percent! le of
Daily Maximum 1-Hour
95th Percent!' le of
Daily Maximum 1-Hour
90th Percent!' le of
Daily Maximum 1-Hour
3rd Quarter Mean of
Daily Maximum 1-Hour
Yearly Mean of Daily
Maximum 1-Hour
17.4
10.9
3.0
33.3
30.8
26.9
20.3
17.0
13.2
8.6
- 11.5%
- 12.1%
- 11.1%
- 17.0%
- 16.5%
- 13.0%
- 11.5%
±11.9
- 14.2%
- 10.6%
10
-------
is typically the worst-case station, we could use data for Azusa only. Both
methods will be tried in this report.
We will also test the basinwide isopleths against trends in the 95th
percentile of daily maximum one-hour concentrations at Azusa. Using this
air quality index should decrease the statistical error in the oxidant
trends.
Individual Locations
The basinwide isopleths are based on maximal one-hour oxidant concen-
trations observed anytime in a nine-hour irradiation. Isopleths are also
available for oxidant concentrations observed exactly at certain irradiation
times:5 hours, 7 hours, or 9 hours (Personal communication with Gary Whitten,
Science Applications, Inc., San Rafael, CA, August 1977). The 5-, 7-, and 9-
hour isopleths will be checked against trend data at DOLA, Anaheim, Azusa, and
San Bernardino. These locations typically experience maximal oxidant concen-
trations around 1:00 PM, 1:30 PM, 2:30 PM, and 4:00 PM respectively. Thus,
they approximately correspond to 5, 5, 7, and 9 hour irradiations from a
7:30 AM start time.
In checking the isopleths for fixed irradiation times, only one oxidant
trend index will be used: the 95th percentile of daily maximum one-hour con-
centrations. As indicated in Table 1, this index is representative of high
oxidant days; yet it exhibits relatively low year-to-year variance. Another
reason for selecting the 95th percentile of daily maxima (rather than, say,
the second highest one-hour) is that we need to check the isopleths for fixed
irradiation times against typical high oxidant days rather than worst-case
high oxidant days. Worst-case days in downtown Los Angeles occur when the ef-
fective irradiation time is more than five hours; these days would not be ,
appropriate for validating the 5-hour isopleths.
ERROR BOUNDS ON OXIDANT TRENDS
When comparing the actual oxidant trends to the oxidant trends predicted
by the isopleth model, we would like to place error bounds on the actual
trends to represent the variance due to meteorological fluctuations. Table 1
provides data relevant to this issue. However, two modifications must be made
on the results in Table 1 to arrive at appropriate error bounds for our trend
study.
11
-------
Table 1 lists the year-to-year deviation (standard error) for various
oxidant air quality indices. In our trend analysis, we will be working with
3-year averages of oxidant air quality, i.e., 1964-1966, 1967-1969, etc. The
standard error of these three year averages will be lower than the single-year
standard error by a factor of /3".
In validating the isopleths, we will take the base year (1964-1966) con-
ditions as given and will examine changes relative to those conditions. The
error of interest will be the error in the difference between base-year air
quality (1964-1966) and air quality for subsequent periods (e.g., 1967-1969
or 1973-1975). To obtain the error in this difference we must multiply our
*
(3-year) standard errors by /2.
The two oxidant air quality indices that will be used in this study are
the yearly second-highest one-hour concentration and the 95th percentile of
daily maximal one-hour concentrations. Table 1 indicates that the year-to-
year deviations for those two indices are - 16.5% and - 11.5%, respectively.
From the line of reasoning presented in the preceding paragraphs, the error
bounds that we will use in our verification study will be as follows:
— (- 16.5%) = - 13.5% for the second-highest
/3~ one hour concentrations
-— (- 11.5%) = - 9.4% for the 95th percentile
/? of daily maximum concentrations
OXIDANT TREND DATA
Figures 3 to 7 present the oxidant trend data that will be used in
the verification study[3l. Figure 3 presents trends in the basinwide
second highest one-hour concentration. Figure 4 presents trends in the
second highest one-hour and the 95th percentile of daily maxima at Azusa.
Figures 5, 6, and 7 present data on the 95th percentile of daily maxima at
DOLA, Anaheim, and San Bernadino, respectively. The data in Figures 3 to 7
The standard error in the sum or difference of two variables, each with
the same standard error (a), is
12
-------
50
40
E
Q.
Q.
O
O
O
O
30
§
f-H
X
O
20
10 _
3 YEAR AVERAGE
YEARLY VALUES
i I r in r i i i I i i
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 3. Trends in the Basfnwide Second Maximum.
13
-------
60 I
50 ~
40
Q.
Q.
CHL
LU
o
30
SECOND
MAXIMUM
95
th
PERCENTILE
§
i—i
X
o
20
10-
YEARLY VALUES
3 YEAR AVERA6E
I I I I I I I I I I I I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 4. Trends in the Second-Highest One-Hour and the 95th
Percent!le of Dally Maxima at Azusa.
14
-------
.c
CL
Q.
30 -
o
o
§
I—t
X
o
\
20 -
10 -
YEARLY VALUES
--• 3 YEAR AVERAGF
I | | I I I 1 I I I I I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 5. Trends in the 95th Percent!le of the Daily Maxima at DOLA.
15
-------
30-i
E
Q.
20-
S 10.
I
I—I
X
o
3 YEAR AVERAGE
I I I r I ri I i i i r
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 19.74 197.5
YEAR
Figure 6. Trends in the 95th Per«e»^il% of the Daily Maximaj. at Anaheim.
16
-------
a.
Q-
30 -
o
CJ
§
1—t
X
o
20
10 ~
3 YEAR AVERAGE
YEARLY VALUES
I i IT II ii
1964 1965 1966 1967 1968 1969 1970 1971
YEAR
1972
I I I
1973 1974 1975
Figure 7. Trends In the 95th Percentlle of the Daily
Maxima at San Bernardino.
17
-------
are listed in tabular form in Appendix A.
As explained in the discussion of monitoring sites and air quality
indices, the data in Figures 3 and 4 will be used to check the basinwide
isopleths. The data in Figures 4, 5, 6 and 7 will be used to verify the
isopleths for individual irradiation times.
18
-------
CHAPTER 3
DATA ON HISTORICAL PRECURSOR TRENDS AND THE NMHC/NOX RATIO
In order to validate the EKMA isopleth model, information is required
concerning historical precursor trends and the ambient NMHC/NOX ratio. This
chapter provides that information for several source areas within the Los
Angeles basin: the source area corresponding to the basinwide oxidant maximum
and the source areas corresponding to oxidant maxima at Downtown Los Angeles,
Anaheim, Azusa, and San Bernardino.
First, the source area for each location under study is defined. Next,
historical trends in the photochemical precursors from 1965 to 1974 are es-
timated for each source area by considering both emission data and ambient
data. The NMHC/NOX ratio in 1965 (the base year for the validation study)
is estimated from present ambient data on the ratio and from historical pre-
cursor trends. The chapter concludes with a sensitivity analysis of three
critical assumptions inherent in our treatment of the precursor data.
BASINWIDE ANALYSIS FOR LOS ANGELES
The first test of the isopleth method will involve trends in the basin-
wide oxidant maximum for Los Angeles. This section defines the source area
for the basinwide maximum and provides data on historical precursor trends
and on the ambient NMHC/NOX ratio for that area.
Definition of Source Area
From our discussion of historical oxidant data for the Los Angeles
region in Chapter 2, we conclude that the source area for the basinwide
maximum can be considered as the source area affecting the Azusa monitoring
site. To define the boundaries of this source area, we rely on a study of
source/receptor situations for the Los Angeles basin performed as part of a
recent Technology Service Corporation project for the California Air Resources
Board [4J . The TSC study reviewed various wind trajectory and streamline
analyses [5-14] and concluded that the following wind patterns occur rather
consistently during the summer smog season in Los Angeles:
• during the night and early morning hours—variable wind or near
stagnation
19
-------
• during the late morning—west or southerly sea breeze
• during the afternoon and evening—dominant westerly sea breeze
A small sample of the evidence supporting these conclusions is provided
in Table 2 and Figures 8 through 10. Table 2 lists the frequency of
occurrence (by time of day) of various air-flow patterns during the summer
smog season [5] . The two most prevalent patterns (west and south) are
illustrated in Figures 8 and 9. Figure 10 presents the most frequent stream-
line pattern during the month of July [71 .
Table 2. Percent Occurrence of Air Flow Patterns During
July to September in Los Angeles
TIME OF DAY WEST SOUTH EAST ALL OTHERS
4 AM
10 AM
4 PM
10 PM
44%
43
83
62
19%
38
13
28
19%
3
0
5
19%
18
5
5
From TSC's source/receptor analysis [41, we conclude that the source
area typically affecting oxidant at Azusa is as shown in Figure 11. The
source area covers most of the southwestern part of Los Angeles County.
Estimates of Historical Precursor Trends
Two types of data can be used to estimate historical trends in NOX
and reactive hydrocarbons: emissions data and ambient precursor data. Both
are examined below in order to arrive at best estimates of precursor trends
for the source area affecting the basinwide oxidant maximum in Los Angeles.
The trend estimates are made at three-year intervals, 1965, 1968, 1971, and
1974.
Emission Trends
A recent report of the Caltech Environmental Quality Laboratory [15]
provides emission trend data for the Los Angeles region. Figures 12 and 13
summarize the EQL estimates of basinwide emission trends for NOX and RHC,
respectively. Basinwide NOX emissions increased by 35% from 1965 to 1974,
while basinwide RHC emissions decreased by 18%. Nearly all of the NOX
20
-------
1X5
C
cu
fC
O.
0)
0)
s-
o
(U
c
to
oo
CO
O)
21
-------
c
01
+J
••->
CO
O-
O
t/1
(O
C
=1
s_
O
Ol
-p
CO
CT)
3
O5
22
-------
t
,
(U f—
C 3
J- J-
•«-» 3
t/J Q
O)
CD!
23
-------
c
(T3
•a
x
o
c
a>
O)
u
J-
3
O
CO
O)
x
g ^
Q. X
a. to
OJ
24
-------
1600 -i
1400
1200
YEARLY 1000
AVERAGE
TONS/DAY
(CUMULATIVE)
800
600
400
200
LIGHT DUTY VEHICLES
OTHER STATIONARY SOURCES
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974
Figure 12. Total NOX Emission Trends in the
Los Angeles Basin [15].
25
-------
2100 -,
1800 -
1500 -
YEARLY
AVERAGE
TONS/DAY
(CUMULATIVE)
1200 -
900
600
100
LIGHT DUTY VEHICLE,
EVAPORATIVE AND CRANKCASE
LIGHT DUTY VEHICLE EXHAUS
GASOLINE HEAVY DUTY VEHICLE
OTHER MOVING SOURCES
ORGANIC CHEMICAL
ORGANIC FUEL 6 COMBUSTION
>65 196(>
1967 1968
I960 [970 1971 1972 1973 1974
Figure 13. Total RHC Emission Trends in the
Los Angeles Basin [15].
26
-------
increase and most of the RHC decrease resulted from changes in
emissions from gasoline-powered motor vehicles.
The EQL report also documents emission trends on a county by-county
basis. Because of low growth rates in Los Angeles County (see Figure 14),
Los Angeles County emissions decreased relative to basinwide total emis-
sions. Los Angeles County emission changes were +25% for NOX and -24% for
RHC from 1965 to 1974 [151. As was the case with basinwide RHC emissions,
the decrease in RHC emissions for Los Angeles County was rather continuous
over the nine-year period. Unlike basinwide NOX emissions which peaked
in 1973, Los Angeles County NOX emissions reached a maximum around 1970-1971.
The emission trends for the source area of interest (Figure 11) should
be similar to, but not exactly the same as the emission trends for Los
Angeles County. A slight difference will arise because the source area is
a lower growth area than the county as a whole (see Figures 11 and 14).
Estimating emission trends specific to the source area involves educated
guesswork based on relative growth rates (Figure 14) and the spaHal dis-
tribution of various source types [161. Judging from the results of the
EQL trend study, we estimate that emissions in the source area changed as
follows from 1965 to 1974:
Estimated NOX Estimated RHC
Emission Increase Emission Decrease
Year
1965
1968
1971
1974
Relative to 1965
0%
14-18%
19-26%
13-23%
Relative to 1965
0%
3-11%
16-22%
24-33%
Note that the error range in our emission estimates increases with time
because some of the uncertainties are compounded over time.
Ambient NOX Trends
An alternative method of estimating precursor trends is to examine
ambient data. To minimize statistical fluctuations in the trend estimates,
27
-------
-------
a large sample of air quality data should be used. Table 3 summarizes trends
in ambient NOX for the source area; these trends are based on changes in an-
nual mean NOX and the yearly average of daily maximum NOX for each three-year
period from 1964-1966 to 1973-1975 [3]. All the listed changes are relative
to the 1964-1966 ambient NOX level.
The trends indicated by both the means and medians in Table 3 agree
quite well with the NOX emissions trend discussed previously. The trend of
the medians in Table 3b most closely follows the emissions trend, which
basically is a pattern of increasing values in the period 1964-1972 followed
by decreasing values in the 1973-75 time period.
Ambient NMHC Trends
Long-term ambient trend data for total hydrocarbons (THC) are available
at Downtown Los Angeles and Azusa. A partial history of THC trends is avail-
able at Burbank, Lennox, and Whittier. Estimating historical changes in
NMHC concentrations with the THC data is a tenuous procedure. Ambient hydro-
carbon measurements are considerably more error prone than other monitoring
data [41. Also, conceptual difficulties arise in translating THC trends
into NMHC trends. Using a very simple procedure to calculate NMHC from THC
levels , approximate estimates of ambient NMHC trends can be derived; these
trends are summarized in Table 4.
The trends in the median percent changes of ambient NMHC concentrations
agree fairly well with the estimated trends in RHC emissions. The trends
in the average percent changes of ambient NMHC concentrations don't agree
as well with emissions, perhaps because the average of a given sample is
inherently more susceptible to extreme values, such as the data from Burbank
and Azusa. The discrepancies at Burbank and Azusa most likely arise from
errors in the ambient trends. The basic trend for both emissions and am-
bient concentrations has been one of steadily decreasing values in the
period 1965-1974.
*
NMHC trends are estimated from THC trends using the relation
NMHC = (THC - 1 ppm)/2. The accuracy of this formula changes as rel-
ative THC and NMHC levels alter with time. This leads to a basic con-
ceptual difficulty in estimating NMHC trends from THC trends.
29
-------
Table 3. Trends in Ambient NO in the Source Area
/\
for the Los Angeles Basinwide Oxidant Maximum [31
Table 3a. Percent Changes in Annual Mean NOV Relative to 1964-1968
X
YEAR
1964-66
1967-69
1970-72
1973-75
STATION
DOLA
: 0%
: +8
i
:+22
! +1
i
LENNOX
Q%*
+21
+23
+5
WEST
L.A.
0%
+8
+19
+9
BURBANK
0%
+39
+37
+7
LONG
BEACH
0%
+17
+10
-16
AZUSA
0%
+16
+54
+46
AVG. OF
PERCENT
CHANGES
0%
+18
+28
+9
MEDIAN OF
PERCENT
CHANGES
0%
+16
+22
+6
Table 3b. Percent Changes in Yearly Average of Daily One-Hour
Maximum NOV Relative to 1964-1966
A
YEAR
1964-66
1967-69
1970-72
1973-75
DOLA
0%
+7
+22
0
STATION
LENNOX
0%*
+22
+30
+14
WEST
L.A.
Q%
+8
+17
+10
BURBANK
0%
+34
+36
+7
LONG
BEACH
0%
+18
+20
-8
AZUSA
0%
+17
+56
+47
1 AVC. or
PERCENT
CHANGES
0%
+ 18
+ 30
+ 12
'MEDIAN' OF'
PERCENT
CHANGES
0%
+18
+26
+9
based on two-year average
30
-------
Table 4. Trends in Ambient NMHC in the Source Area for the
Los Angeles Basinwide Oxidant Maximum [3]
Table 4a. Percent Changes in Annual Mean NMHC Relative to 1964-1966
YEAR
1964-66
1967-69
1970-72
1973-75
-- STATION
DOLA
0%
-15
-8
-38
LENNOX
tf%
-12f
-24
-35
WHITTIER
Of%
-18f
-29
-41
BURBANK
*
0 %
+3A
+7
-7
AZUSA
*
0 %
+11
+33
+39
AVG. OF
PERCENT
CHANGES
0%
-6
-4
-16
MEDIAN OF
PERCENT
CHANGES
0%
-12
-8
-35
Table 4b. Percent Changes in Yearly Average of Daily One-Hour
Maximum NMHC Relative to 1964-1966
YEAR
1964-66
1967-69
1970-72
1973-75
DOLA
0%
-15
-23
-46
STATION
LENNOX
Of%
_14t
-31
-48
WHITTIER
Of%
-15f
-29
-47
BURBANK
*
0 %
+2A
+5
-13
AZUSA
*
0 %
+11
+35
+12
'AVG.' OF
PERCENT
CHANGES
0%
-6
-9
-28
"MEDIAN OF
PERCENT
CHANGES
0%
-14
-23
-46
t
based on extrapolation of 1970-1975 trend
based on two-year average
A-,,
'linear interpolation between 1970-72 and 1964-66
31
-------
Best Estimates of Precursor Trends
By considering both emission trend data and ambient trend data, one
can arrive at reasonable estimates of precursor trends in the source area.
Both the emission estimates and the ambient precursor data are subject to
potential errors from several factors. The principal factors affecting the
ambient data are:
• characteristics of sampling site
t meteorological fluctuations
t uncertainties in analytical methodology (especially NMHC)
Emissions data are generally affected by the following:
• growth rate of source area
• changes in source emission rates
In spite of potential errors from all these factors, the general agreement
between emissions estimates and ambient data for both precursors gives one
confidence in ascertaining the historical trends of NMHC and NOY. Table 5
A
which was constructed using both ambient and emissions data for NMHC and NOX,
summarizes our best estimates of precursor changes relative to 1964-1966.
These data are shown graphically in Figure 15. Also included in the table
and figure are error bounds based on a subjective analysis of the uncertainties,
including the agreement or disagreement between emission trends and ambient
trends.
Table 5. Best Estimates of Precursor Trends in the
Source Area for the Basinwide Oxidant Maximum
Year
1964-66
1967-69
1970-72
1973-75
NOV Change
A
0%
+17% + 3%
+24% + 5%
+15% +7%
NMHC Change
0%
-10% + 3%
-16% + 6%
-30% + 6%
32
-------
o
co
DC: uj
o >
CO LlJ
o cx:
LU -
<_>
1—I
o
CO
I—I
n
u_
o
o
CQ
LO
<-D
CTl
NO.
NMHC
1965
1968
1971
1974
YEAR
Figure 15. Best Estimates of Historical Precursor Trends in
the Source Region for the Basinwide Oxidant Maximum.
33
-------
Estimates of Ambient 6-9 AM NMHC/NOX Ratio
Considerable data on ambient NMHC and NOX concentrations in Los Angeles
are available for the early and mid-1970's. Based on these data, 6-9 AM
NMHC/NOX ratios are calculated for various locations in the source region
and for various times in the period 1969-1976. These results are sum-
marized in Table 6.
With two exceptions, the NMHC/NOX ratio is very consistent in spite
of the spatial and temporal variations in the monitoring of the precursors.
The two data sources that aren't in agreement with the others are the APCD
data, which tend to give low NMHC/NOX ratios, and the 1974 ARB data, which
tend to give high ratios. The nature of this disagreement is thought to
be due to the method of monitoring the NMHC. The flame ionization detection
method used by both the ARB and APCD has been shown to give unaccountably
poor readings [4]. Furthermore, the accuracy of the GC separation technique
employed by the ARB is strongly dependent upon operator skill [4].
The ratios from the other data sources did not fluctuate much over
the years 1970-76 because the ambient concentrations of both precursors
were simultaneously decreasing. Based on the data in Table 6, our best
estimate of the NMHC/NOX ratio in the 1970's is the following:
Median: 8
10th Percent!le: 5
90th Percentile: 15
The NMHC/NOX ratio for the 1970's, together with the best estimate of
the precursor trends from 1965-75 will now be used to estimate the NMHC/NOX
ratio for 1965. Since NMHC have decreased about 20% and NOX has increased
about 20% from 1965 to the early 1970's, the 1965 NMHC/NOX ratio was there-
fore about 1.5 times the ratio in the 1970's. Consequently, the best esti-
mate of the 1965 NMHC/NOX ratio is the following:
Median: 12
10th Percentile: 7
90th Percentile: 23
We are using a range of ratios for two basic reasons. First, the low
quality of the ambient NMHC data introduces uncertainty concerning the real
ambient ratio. Second, the NMHC/NOX ratio appears to fluctuate considerably
34
-------
Table 6. Ambient 6-9 AM NHMC/NOx Ratios
DATA SOURCE
NMHC [1]
NOX [1]
NMHC [2]
NOX [2]
NMHC [3]
NOX [3]
NMHC [4]
NOX [3]
NMHC [4]
NOX [3]
NMHC [5]
NOX [5]
TIME PERIOD
9 summer days
0600-0900 1976
19 summer days
0600-0900 1976
6 days per
year in April -
Sept. 0600-
0900 1969-74
30 summer days
0600-0900 1974
30 summer days
0600-0900 1971
13 days 9/73-
10/73 0800-1000
LOCATION
Riverside
Temple City
Azusa
DOLA
Lennox
Pomona
MEASURE
MEDIAN
8
8
5
3
4
3
i j
Azusa
DOLA
Azusa
DOLA
Central Los
21
15
8
8
7
Angeles Area
) NMHC/NOV RATIO
A
10th%
6
4
0
0
0
0
- - - -
15
11
7
4
4
90th%
20
10
14
7
8
6
30
28
13
11
14
1. Statewide Air Pollution Research Center, GC/FID Hydrocarbon Measurements.
2. Air Resources Board, GC/FID Hydrocarbon Measurements.
3. APCD Data Base 1969-1974, FI Hydrocarbon Measurements.
4. "Atmospheric Hydrocarbon Concentrations June-Sept. 1974," "Distribution
of Hydrocarbons in the Los Angeles Atmosphere, Aug.-Oct. 1971," Air
Resources Board, GC/FID Hydrocarbon Measurements.
5. LARPP, Semi-permeable CH membrane + FID Hydrocarbon Measurements.
35
-------
from day-to-day [2, 171, possibly due to specific wind trajectories and
associated stationary source areas, or to variations in motor vehicle
NMHC/NOX emission ratios because of temperature and relative humidity
fluctuations.
ANALYSIS OF INDIVIDUAL LOCATIONS
Downtown Los Angeles, Anaheim, Azusa, and San Bernardino were selected
for testing the model at specific sites and hence validating the isopleths
for individual irradiation times.
Downtown Los Angeles
This section defines the source area for Downtown Los Angeles and
presents the historical precursor trends and the ambient NMHC/NOX ratio
for that location.
Source Area
Based on an analysis of wind flow patterns in the Los Angeles basin
(see earlier discussion) we conclude that the source area affecting oxidant
in Downtown Los Angeles is as shown in Figure 16. The source area
essentially consists of the southwest quadrant from Downtown Los Angeles
to the coastline.
Historical Precursor Trends
Emission trend estimates for the source area affecting Downtown
Los Angeles are derived by modifying the results of the EQL emission trend
study [151. These modifications are based on relative growth rates (Figure
14) and the spatial distribution of various source types [16]. Our estimates
indicate that emissions changed as follows in the source area from 1965 to 1974:
Estimated NOX Estimated RHC
Year
1965
1968
1971
1974
Emission Increase
0%
10-15%
14-21%
7-16%
Emission Decrease
0%
10-14%
20-28%
30-42%
36
-------
-------
Table 7. Trends in Ambient NOX in the Source Area
for Downtown Los Angeles [3]
Table 7a. Percent Changes in Annual Mean NOX Relative
to 1964-1966
YEAR
1964-66
1967-1969
1970-1972
1973-1975
STATION
DOLA
0%
+8
+22
+1
WEST
LENNOX L.A.
0%* 0%
+21 +8
+23 +19
+5 +9
LONG
BEACH
0%
+17
+10
-16
AVG. OF
PERCENT
CHANGES
0%
+14
+18
0
MEDIAN OF
PERCENT
CHANGES
0%
+12
+20
+3
Table 7b. Percent Changes in Yearly Average of Daily
One-Hour Maximum NOX Relative to 1964-1966
YEAR
1964-66
1967-69
1970-72
1973-75
STATION
WEST LONG
DOLA LENNOX L.A. BEACH
0% 0%* 0% 0%
+7 +22 +8 +18
+22 +30 +17 +20
0 +14 +10 -8
AVG. OF MEDIAN OF
PERCENT PERCENT
CHANGES CHANGES
0% 0%
+14 +13
+22 +21
+4 +5
based on two-year average
38
-------
Table 8. Trends in Ambient NMHC in the Source Area
for Downtown Los Angeles [3]
Table 8a. Percent Changes in Annual Mean NMHC Relative
to 1964-1966
YEAR
1964-66
1967-69
1970-72
1973-75
STATION
DOLA LENNOX WHITTIER
0% 0%f 0%f
-15 -12 f -18 f
-8 -24 -29
-38 -35 -41
AVG. OF MEDIAN OF
PERCENT PERCENT
CHANGES CHANGES
0% 0%
-15 -15
-20 -24
-38 -35
Table 8b. Percent Changes in Yearly Average of Daily One-
Hour Maximum NMHC Relative to 1964-1966
YEAR
1964-66
1967-69
1970-72
1973-75
STATION
DOLA LENNOX WHITTIER
0% 0%f 0%f
-15 -14 + -15 +
-23 -31 -29
-46 -48 -47
AVG. OF MEDIAN OF
PERCENT PERCENT
CHANGES CHANGES
0% 0%
-15 -15
-28 -29
-47 -47
t
based on extrapolation of 1970-1975 trend
39
-------
The trends in ambient NOX for the Downtown Los Angeles source area
are presented in Table 7. Downtown Los Angeles, Lennox, West Los Angeles,
and Long Beach were selected as being most representative of the source
region's NOX trends. The ambient NOX trends agree fairly well with our
estimates of NOX emission changes for the DOLA source area.
Table 8 presents the trends in ambient NMHC for the Downtown Los
Angeles source area. DOLA, Lennox, and Whittier were selected as being
most representative of the ambient NMHC trends in the source area for DOLA.
The basic hydrocarbon trend has been one of steadily decreasing concentrations
over the years 1964-1975.
After considering both emission trend data and ambient trend data,
our best estimates of historical precursor trends for the Downtown Los
Angeles source region are as presented in Table 9. The results are ex-
preosed as percent changes in precursors relative to 1964-1966. We attach
good confidence to these results because the ambient and emissions data
agreed quite well for the Downtown Los Angeles source area. The data
presented in Table 9 are shown graphically in Figure 17.
Table 9. Best Estimates of Precursor Trends for
the DOLA Source Region
Year
1964-66
1967-69
1970-72
1973-75
NOV Change
X
0%
+13% + 2%
+18% + 3%
+ 7% + 5%
NMHC Change
0%
-13% + 2%
-24% + 3%
-38% + 5%
40
-------
o
oo
O LU
OO >
C£ LU
CTi
NO.
NMHC
1965
1968
1974
YEAR
Figure 17. Best Estimates of Historical Precursor Trends
in the DOLA Source Region
41
-------
Ambient NMHC/NOX Ratio
The 6-9 AM NMHC/NOX ratio for the Downtown Los Angeles source area
is assumed to be the same as that for the basinwide-maximum source area.
In 1965 the ratio is estimated to be as follows:
Median: 12
10th Percentile: 7
90th Percentile. 23
Anaheim
This section discusses the source area, historical precursor trends,
and ambient NMHC/NOX ratio for the validation study at Anaheim.
Source Area
Our analyses of wind-flow patterns in the Los Angeles basin (see
earlier discussion) indicates that the source area affecting oxidant in
Anaheim is as shown in Figure 18. The area includes the northwest part
of Orange County and the southern coast of Los Angeles County.
Historical Precursor Trends
Estimates of emission trends for the Anaheim source area are derived
according to the procedures described earlier. Net changes in emissions
relative to 1965 are approximately as follows:
Estimated NOX Estimated RHC
Year Emission Increase Emission Decrease
1965
1968
1971
1974
0%
25-35%
35-50%
40-60%
0%
2-8%
4-13%
6-18%
Trends in ambient NOX for monitoring sites within or near the Anaheim
source region are presented in Table 10. The average and median percent
changes among the three monitoring sites are consistent with the estimated
emission changes for the source area. There is of course an obvious dif-
ference between the low growth parts of the source area (e.g. Long Beach)
and the high growth parts of the source area (e.g. Anaheim and La Habra).
42
-------
o>
O)
O
3
O
to
-------
Table 10. Trends in Ambient NOX in the Source Area for
Anaheim [3]
Table lOa. Percent Changes in Annual Mean NOX Relative
to 1964-1966
YEAR
1964-1966
1967-1969
1970-1972
1973-1975
STATION
LONG
BEACH
0%
+17
+10
-16
LA
HABRA
0%f
+19*
+26
+60
ANAHEIM
0%
+75
+85
+78
AVG. OF
PERCENT
CHANGES
0%
+37
+40
+41
MEDIAN
PERCENT
CHANGES
0%
+19
+26
+60
Table lOb. Percent Changes in Yearly Average of Daily One-Hour
Maximum NOX Relative to 1964-1966
YEAR
1964-1966
1967-1969
1970-1972
1973-1975
STATION
LONG
BEACH
0%
+18
+20
-8
LA
HABRA
0%*
+3
+14
+44
ANAHEIM
0%
+92
+93
+86
AVG. OF
PERCENT
CHANGES
0%
+38
+42
+41
MEDIAN
PERCENT
CHANGES
Q%
+18
+20
+44
based on two-year average
h
based on extrapolation
t
44
-------
Table 11. Trends in Ambient NMHC in the Source Area for
Anaheim [3]
Table lla. Percent Changes in Annual Mean NMHC Relative
to 1964-1966
YEAR
1964-66
1967-69
1970-72
1973-75
ANAHEIM
0%*
-12
- 6
-24
Table lib. Percent Changes in Yearly Average of Dally
One-Hour Maximum NMHC Relative to 1964-1966
YEAR
1964-66
1967-69
1970-72
1973-75
ANAHEIM
0%*
- 7
- 3
-20
based on two-year average
45
-------
There is only one monitoring site (Anaheim) providing data on ambient NMHC
trends for the Anaheim source region. As shown in Table 11, the ambient
NMHC decrease at Anaheim is slightly greater than the estimated RHC emission
decrease. For the Anaheim source region, we place greater confidence in the
RHC emission trend than in the ambient NMHC trend because only one monitoring
site is available.
By considering both the emission trend data and the ambient trend data,
we arrive at best estimates of historical precursor trends for the Anaheim
source area. These best estimates are listed in Table 12 and illustrated
in Figure 19.
Table 12. Best Estimates of Precursor Trends for
the Anaheim Source Area
Year
1964-66
1967-69
1970-72
1973-75
NOX Change
0%
+29%±8%
+47%±12%
NMHC Change
0%
-6%±4X
.j.
Ambient NMHC/NOX Ratio
The 6-9 AM NMHC/NOX ratio for the Anaheim source area is assumed to be
the same as that for the basinwide source area. For the base year 1965, our
estimates for the ratio are as follows:
Median: 12
10th Percentile: 7
90th Percentile: 23
Azusa
The source area for oxidant at Azusa is assumed to be the same as the
source area for the basinwide oxidant maximum. Thus, the historical pre-
cursor trends and ambient NMHC/NOX ratio for the Azusa source area are as
presented in the section on the basinwide maximum.
San Bernardino
This section defines the source area for San Bernardino and presents
the historical precursor trends and the ambient NMHC/NOX ratio for that source area.
46
-------
1.5 —,
NOV
NMHC
1965
1968
1971
1974
YEAR
Figure 19.
Best Estimates of Historical Precursor Trends
1n the Anaheim Source Region.
47
-------
Source Area
After a study of the wind flow patterns in the Los Angeles Basin,
(see section on basinwide analysis), we conclude that the source area govern-
ing oxidant concentrations in San Bernardino is as shown in Figure 20. The
area extends from the coast to San Bernardino, encompassing parts of Los
Angeles, Orange, San Bernardino, and Riverside Counties.
Historical Precursor Trends
The emission trend estimates for the source area affecting San Bernardino
are derived following procedures discussed previously. We estimate that emis-
sions changed as follows in the source area from 1965 to 1974:
Estimated NOX Estimated RHC
Year Emission Increase Emission Decrease
1965 0% 0%
1968 17-21% 6-9%
1971 23-30% 11-17%
1974 25-35% 15-24%
The trends in ambient NOX for the San Bernardino source region are
presented in Table 13. The 10 cities listed in the table were chosen be-
cause of data availability and geographical location. For both the daily
maximum and hourly average NOX concentrations, the ambient trends, averaged
over the sites, are similar to the emission trends for the source area.
Table 14 presents the trends in ambient NMHC for the San Bernardino
source area. The overall pattern of ambient NMHC has been one of steadily
decreasing concentrations in agreement with emission trends; however, as
one can see by examining Table 14 some stations deviated drastically from
this overall pattern.
Table 15 presents our best estimates of historical precursor trends
for the San Bernardino source area. These data are shown graphically in
Figure 21.
48
-------
CO
J
'x
c
(O
•o
•r-
X
o
o>
-p
o
OJ
«c
ro
0)
O
3 •
O O
CO C
•r-
fl) "C
ij *^
'x o>
O CQ
Q. C
Q. (O
O
CM
-------
0
c:
-a
S-
fO
c
$_
cu
CQ
c
to
to
s_
o
n-
(O
cu
S-
•r™
4->
fO
i—
CU
fy
X
O
^r
^~
c
LU 3C
eC Q. 0
v:
7*
<:
1
CQ
O
Z
1— 1
Q
C£
<
z a:
tO CQ
s:
f-H
LU
31
<
<
a:
CQ
-
&s
o o
CM
s«
O LO
CM
+
&«
o cn
CO
%«
O CM
CM
+
fc«
O LO
•^^
o en
r—
&«
o r».
T
*
%•£
o •—
CM
+
*
»«
o cn
CM
+
*«
0 VO
"+
S*
O CO
*e
O CO
i— r^
CO •—
vo <*
CO CM
+ +
—
r^ i^
CO +
VO 1^
9 ¥
LO CO
+ •+
vo o
CM VO
+ +
O vo
f"* r—
+ 1
CO LO
CM +
"Kf LO
T ¥
-
VO
C ID
•1- I
•*
(/> VO
ai cn
o
CO
O
a)
a.
ai
-Q
01
5C
§
-^
tC
~r»
_
u.
o
f- to
•~r •— & i , i
k^ *C. LU
«C LU CD
1-1 CJ Z
o a: «a:
LU LU OI
s: | | | -T-
fcAJ r A_
< a. o
^
z
•sC
CQ
C£
CQ
O
Z
1— 1
Q
or
<
?•" fV
—i
LU
1
a?
ca
-
— — — 1
^^
O CO O O
i— CO CM
+ 4. •!•
T^ T^
*5
O LO f^ VO
CM CO CM
+ + +
'SS.
O «* VO f^
CO CO +
+ +
*
*s>
o •* o r>«
r- CO CM
+ + +
&S.
O CM CO VO
cn en op
+ + +
*
t~) ff\ ^- ^
0 CO J J
3^
0 CO O CO
r- CM 1
+ +
*
fc*
O CM O •*
CM CO r—
+ + +
V
O <* r- CM
CO LO CO
+ + +
*s
o i^ vo r*.
r— LO «*•
+ + 4-
*s
o co r^ o
4-i — r—
+ +
&«
o r->. CM o
4- CM
II
vo cr\ oj tn
MD co r*^ r*1**
i i i i
*f fx. O CO
UD ID r^. r^-
oS o*> o> 01
r— • r"~ ^™ r~*
CD
C7)
a
s-
(O
(O
(U
>>
O
3:
o
O
Q.
+J
X
o
T3
O)
O)
to 10
(O (O
X) .a
* -f-
-------
CO
t
•=c
S o
^ I
o z
oo j-
5 1
-M S
c r~i
•1- -J
yr~ ! iio
^,°O| <; vo
Z ^
o c T
4-> C '•- J.
C M- ,-?*
*"2 Sen
•i * c "~
< | J=S
=
^f) f—
fO _Q
| (XJ
i—
z
o
1 — 1
1—
oo
u_
o
h- 0-
Z Z U.
< LU CE
•-i O Z
Q ce <
LU LU I
z: Q- c_
LU h- CO
O Z LU
LU C£
• 0 Z
CD CC <;
^* * ' f T
ff r\ cj
0
1— 1
o
Z C£
CO Z
LU
00
BURBANK
<
c
D_
z:
LL
n:
-
o en f*-
i i
v^
O CO O
1
*0
o LO r^
1 i
*
£•5 Q
CD JM •«,
1 1
•K
r— CO
•K
&« -t-
o CM cn
i— CM
1 |
4-
O CM <3-
r— CsJ
1 1
O LO CO
1 '
vo cn CM
i i i
<|?^ r*^ c^
^o io r*H
O*t CTt CT>
1
VO
7 '
*
vo
1
r-.
LO
10
CM
i
cn
co
f—
^~
LO
co
1
CO
CO
1
LO
""I*
CO
^^
cn
1
(U
o
>>
^™
to vo
O (O
M-2
O 1
**•
cn
£-
§!S
<
(U
>>>
r™ 'r*
S- +J
rO (O
CD r—
^C?
C
'f* C^
X
Percent Changes
Hour Maximum MM
.
f">
^"
T— H
oj
-Q
ftJ
1
z
=c
CO
U_
0
1— V
< LU CS
•-I O Z
o a: <
LU LU n:
S Q- CJ
U. H- CO
CD ?"* LLJ
LU cn
> LU 3C
< (X O
o
1— 1
o
z a:
CO Z
C£
CO
.
1 •_. I—
1 1
•K
° ^ CO °
1 1 1
O r-t LO CM
r^ co 'T
+.
&« -H
o LO o, r^
T CN1 i
1 i '
H-
O •* CO
T °° "f
O LO (T) VO
T ~ T
us en CM 10
Lf^ yj ^^ fx^
i i i i
•sf r^ o co
cn cn cn cn
a;
LO
r>.
en
o
vo
cn
r—
-o
c
to
CM
cn
r—
O
cn
cn
to
r—
O
Q.
X
V
o
T3
10 0)
S- JO
Ol
> C
ro O
•r-
S_ 4->
ro CO
O> •—
I CX.
O J-
c
C i-
o
T3 «O
a; a>
M C
(O •!-
51
-------
1.5 -,
CJ
UJ
O 5>
OO LU
UJ
eC OO
O <
i—i CO
C£
O LO
\— ^o
OO CTl
O
O
1.0
0.5
NMHC
1965
1968
1971
1974
YEAR
Figure 21. Best Estimates of Historical Precursor Trends
1n the San Bernardino Source Region
52
-------
Table 15. Best Estimates of Precursor Trends for San
Bernardino Source Area
Year
1964-66
1967-69
1970-72
1973-75
N0y Change
0%
+20% ± 3%
+30% ± 5%
+25% ± 7%
NMHC Change
0%
-8% +• 3%
-13% - 4%
-20% - 5%
Ambient NMHC/NOX Ratio
The 6-9 AM NMHC/NOX ratio for the San Bernardino source area is
assumed to be the same as that for the basinwide source area in 1965:
Median: 12
10th Percentile: 7
90th Percentile: 23
ANALYSIS OF CRITICAL ASSUMPTIONS
Several assumptions are implicit in our treatment of the precursor data
for use in the EKMA isopleth model. This section assesses the validity of
three assumptions that may be particularly critical to the isopleth verifi-
cation study. The issues addressed are as follows:
1. The source areas have been selected based on the predominant
wind flow pattern during the summer smog season in Los
Angeles. Does this wind flow pattern also predominate on days
with extreme oxidant (the days of interest in the isopleth
validation study)?
2. Ambient precursor trends have been examined using two air
quality indices: annual mean concentrations and yearly
average of daily one-hour maximum concentrations. Are the
trends in these indices representative of trends in 6-9 AM
summertime concentrations (the precursor averaging time of
interest in the isopleth validation study)?
3. The median NMHC/NOX ratio has been estimated from ambient
data for the entire summer smog season. Is the median ratio
the same on days with extreme oxidant (the days of interest
in the isopleth validation study)?
53
-------
Wind Flow Patterns on High Oxidant Days
Our selection of source areas was based on the southwesterly (sea
breeze) wind pattern that predominates in Los Angeles during daytime hours
in the summer smog season. Since the isopleth validation studies involve
days of extreme oxidant (either the second maximum or the 95th percentile
of daily maxima), it is important to examine wind patterns on episode days.
The source areas will be appropriate for the verification studies only if
the southwesterly pattern also dominates on days of highest oxidant.
Figure 22 illustrates the frequency distribution of vector-average
wind direction (7 AM to 2PM) at Downtown Los Angeles. The prevalence of
the southwesterly pattern during the smog season is obvious in the upper
graph, representing all days from May to October in the years 1971-1975.
The lower graph, representing the 50 days of highest oxidant at Azusa
during the May-October/1971-1975 period, indicates that the southwesterly
pattern is even more consistent on days of extreme oxidant.
Figure 23 presents the frequency distribution of vector-averaged wind
speed (7 AM to 2 PM) at Downtown Los Angeles. As was the case with wind
direction, wind speeds are more concentrated around "normal" conditions
on days of high oxidant. The median wind speed on high oxidant days (5.5
mph) is slightly greater than the median wind speed on all days (4.9 mph).
Wind speeds of 4 to 7 miles per hour from the southwest are especially
prevalent on days of high oxidant at Azusa because this wind pattern
promotes transport from the source-intensive central-coastal parts of
the basin toward Azusa.
The foregoing analysis demonstrates that the southwesterly wind
pattern indeed predominates on days of excessive oxidant (at least at Azusa).
Although possibly not as dominant as in the case of Azusa, we would expect
that the general sea-breeze pattern also prevails for typical high oxidant
days (95th percentiles) at the other locations under study. This is partially
evidenced by high correlations between daily maximum oxidant at Azusa and
daily maximum oxidant at the other locations (.78 with DOLA, .80 with
Anaheim, and .86 with San Bernardino [4]), implying that high oxidant days
at Azusa tend to be high oxidant days elsewhere.
54
-------
40-
ac.
LU
Q.
o-
UJ
0£.
30-
20—
10—
ALL DAYS
(May-October, 1971-1975)
N I NNE I NE I ENE E ESE SE SSE S SSW SW WSW W WNW I NW I NNW
WIND DIRECTION
o
cc.
50-
40—
30—
50 DAYS OF HIGHEST
OXIDANT AT AZUSA
(May-October, 1971-1975)
10—
N 'NNE ' NE 'ENE E ESE SE SSE
SSW SW WSH W WNW NW NNW
WIND DIRECTION
Figure 22. Frequency Distribution of Vector-Averaged Wind Direction
(7 AM - 2 PM) at Downtown Los Angeles.
55
-------
30
o
ce.
o-
o-
LU
cc.
20-
£ 10-
ALL DAYS
(May-October, 1971-1975)
1 2 3 4 5 6 7 8 9 10 11 12
WIND SPEED (MPH)
ou—
40_
30_
20-
10-
1 1
50 DAYS OF HIGHEST
OXIDANT AT AZUSA
(May-October, 1971-1975)
1 1 1 1
1 234 56 7 8 9 10 11 12
WIND SPEED (MPH)
Figure 23. Frequency Distribution of Vector-Averaged Wind Speed
(7 AM - 2 PM) at Downtown Los Angeles.
56
-------
In future work it might be worthwhile to substantiate our conclusions
further by repeating the analysis (i.e. Figures 22 and 23) for oxidant at
other locations. However, the present evidence suggests that selecting
source areas according to the prevalent sea-breeze wind pattern is ap-
propriate for the isopleth validation studies.
6-9 AM Summertime Precursor Trends
We have estimated historical precursor trends by analyzing both emis-
sions data and ambient precursor data. The ambient precursor trends were
based on two air quality indices: annual mean concentrations and yearly
average of daily maximal concentrations. In general, we found that the
trends in these two air quality indices agreed fairly well with one another
and with the emission trends. A question remains, however, as to whether
these trends are consistent with trends in 6-9 AM summertime precursor con-
centrations which are most relevant in applying..the isopleth model.
The California Air Resources Board has compiled data concerning trends
in 6-9 AM summertime precursors for several locations over the period 1963
to 1972 [18]. Tables 16 and 17 compare the net changes in 6-9 AM summer-
time precursors during that period to corresponding changes in the precursor
air quality indices we have used.
Table 16 indicates that trends in 6-9 AM summertime NMHC are basically
very similar to trends in annual mean NMHC and yearly average of daily max-
imal NMHC. If we had used 6-9 AM summertime concentrations as the ambient
NMHC trend indicator, our conclusions concerning historical NMHC trends in
each source area probably would not have changed substantially.
Table 17 reveals a discrepancy between trends in 6-9 AM summertime
NOX versus trends in annual mean NOX and yearly average of daily maximal
NOX. From 1965 to 1971, the 6-9 AM summertime concentrations appear to have
increased 10 to 25% more than the other two air quality indices. The dif-
ferences in the trends may be explained by the temporal distribution of NOX
emissions. Specifically, automotive emissions are relatively more important
to 6-9 AM concentrations, and automotive emissions are relatively more im-
portant during the summer. Since large increases in NOX emissions from
57
-------
Table 16. Comparison of Alternative Ambient
Trend Indices for NMHCt
LOCATION
TIME PERIOD
NET PERCENT CHANGE IN NMHC CONCENTRATIONS
6-9 AM
Yearly Average Concentrations
Annual Mean of Daily Maxima July-September
Anaheim
Azusa
Burbank
DOLA
Riverside
San Bernardino
1965-66 to 1970-72
1963-64 to 1970-72
1963-64 to 1971-72
1964-66 to 1970-72
1965-66 to 1970-71
1965-66 to 1970-72
AVERAGE OF PERCENT CHANGES
MEDIAN OF PERCENT CHANGES
-5%
+47%
+10%
-8%
-27%
-7%
+2%
-6%
-2%
+43%
+3%
-23%
-20%
-8%
-1%
-5%
+12%
| +35%
1 +4%
' -16%
1
-18%
I -11%
+1%
1 -4%
i
Calculated from THC condentrations as explained previously
Although often constrained by data availability, we have basically
attempted to use changes in 3-year averages from 1964-66 to 1970-72
58
-------
Table 17. Comparison of Alternative Ambient
LOCATION
Trend Indices for NOV
TIME PERIOD
NET PERCENT CHANGE IN NOX CONCENTRATIONS
6-9 AM
Yearly Average Concentrations
Annual Mean of Daily Maxima July-September
Anaheim
Azusa
Burbank
DOLA
Lennox
Long Beach
Pomona
Reseda
1964-66 to 1970-72
1964-66 to 1970-72
1964-66 to 1970-72
1964-66 to 1970-72
1965-66 to 1970-72
1964-66 to 1970-72
1965-66 to 1970-75
1965-66 to 1970-72
San Bernardino 1965-66 to 1970-72
West L.A.
AVERAGE OF
MEDIAN OF
1964-66 to 1970-72
PERCENT CHANGES
PERCENT CHANGES
+89%
+54%
+37%
+22%
+23%
+10%
+44%
+47%
+33%
+17%
+38%
+35%
+93%
+56%
+37%
+22%
+30%
+20%
+51%
+48%
+25%
+17%
+40%
+34%
+69%
+64%
+61%
+27%
+19%
+54%
+61%
+56%
+62%
+42%
+52%
+59%
Although sometimes constrained by data availability, we have basically
attempted to use changes in 3-year averages from 1964-1966 to 1970-1972
59
-------
motor vehicles were the basic cause of the overall NOX increase, it is not
unreasonable that the rise in NOX is more evident in 6-9 AM summertime
concentrations.
The isopleth verification analyses will proceed using the "best estimate"
of NOX changes that were derived earlier in this chapter. That these "best
estimates" may understate the increase in NOx could have a significant ef-
fect on our results. The implications of this possible underestimate will
be discussed in Chapter 5.
Ambient NMHC/NOX Ratio on High Oxidant Days
Our estimate of the ambient NMHC/NOX ratio was based on data for 6-9 AM
precursor concentrations during the entire summer season. The latest pro-
cedural guidelines for the EKMA isopleth model [19] indicate that the NMHC/
NOX ratio on days of highest oxidant should be used. To test whether the
ratio we are using is appropriate, we should compare it with ratios on
extreme oxidant days.
Figure 24 presents frequency distributions of the 6-9 AM NMHC/NOX ratio
at Downtown Los Angeles based on APCD data. The NMHC concentrations have
been computed from THC concentrations using an empirical formula derived by
the California ARB [181.* As shown in the upper graph, the median ratio for
the entire smog season during the early 1970's is approximately 8:1 (in exact
agreement with our earlier conclusions **).
The lower graph in Figure 24 indicates that the median ratio on high
oxidant days (8.9) is slightly greater than the median ratio on all summer
days (8.1). It is also interesting to note that there is less spread in the
frequency distribution on high oxidant days; the 10th and 90th percentiles
of the ratio are 5.7 and 12.7 for high oxidant days and 4.5 and 13.6 for all
summer days.
This formula is NMHC = .7 (THC - 1.3 ppm).
In an earlier section we examined several sources of data and concluded
that the median ratio during the smog season in the early 1970's was
8:1. We then used historical precursor trends to calculate a median
ratio of 12:1 for the 1965 base year.
60
-------
15 __
10
cc
LU
a.
a:
U-
ALL DAYS
(May-October, 1971-1975)
Median Ratio =8.1
12 3456
15 —
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21+
6-9 AM NMHC/NOX RATIO
10—
LU
O
Q.
>-
cr
1 1
1
1
1
1
50 DAYS OF HIGHEST
OXIDANT AT AZUSA
(To smooth the distribution,
frequencies have been averaged
over 2 integer intervals)
ill I 1 1 1 1
1 2 34 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21+
6-9 AM NMHC/NOX RATIO
Figure 24. Frequency Distribution of 6-9 AM NMHC/NOX Ratio at
Downtown Los Angeles.
61
-------
To check the conclusion that the NMHC/NOX ratio tends to be slightly
higher than normal on extreme oxidant days, we acquired recent data on 6-9
AM NMHC and NOX concentrations from the ARB monitoring program at Temple
City. For nineteen days selected at random during the 1976 smog season,
the median ratio was 8. For the nine days of highest oxidant during the
1976 smog season, the median ratio was 10.
Our verification study will proceed using a median ratio of 12:1 for
the 1965 base year (corresponding to a ratio of 8:1 in the early 1970's).
The significance of slightly underestimating the ratio which is appropriate
to high oxidant days will be discussed in Chapter 5.
62
-------
CHAPTER 4
VALIDATION OF THE ISOPLETH METHOD
AGAINST HISTORICAL TKEfflT DATA '
Chapters 2 and 3 described the two basic types of information needed
to validate the EKMA isopleth method against historical trends. Chapter 2
presented the actually observed oxidant trends for the locations under study;
Chapter 3 discussed the ambient precursor trends and the NMHC/NOX ratio for
the study areas. The present chapter uses the isopleth method to predict
historical oxidant trends and compares the predicted trends with actual
trends to assess the accuracy of the method.
The validation study is conducted for the basinwide oxidant maximum and
for the oxidant maxima at four individual locations (DOLA, Anaheim, Azusa,
and San Bernardino). In the following pages, four basic types of isopleths
are referred to, and their descriptions are as follows:
(1) basinwide isopleths: corresponding to the maximum oxidant
during 0 to 9 hours of irradiation.
(2) five-hour isopleths: corresponding to oxidant after five
hours of irradiation; not neccessarily the maximum oxidant.
(3) seven-hour isopleths: as in (2) above, with seven hours
of irradiation.
(4) nine-hour isopleths: as in (2) above, with nine hours of
irradiation.
VALIDATION OF BASINWIDE ISOPLETHS
This section begins with a detailed description of the validation
procedure using the basinwide isopleths and the 95th percentile of daily
maxima at Azusa. The results of other validations with the basinwide isopleths,
using the Azusa yearly second maximum and the basinwide yearly second maximum,
are then summarized.
95th Percentile of Daily One-Hour Maxima at Azusa
Three types of input data are required to compute predicted oxidant
trends: the oxidant value for the base year (1965, or actually 1964-1966),
the 6-9 AM NMHC/NOX ratio for the base year, and historical precursor trends
63
-------
for the source area. For Azusa, the 95th percent!le of daily oxidant maxima
in 1964-1966 was 0.33 ppm. The 1965 NMHC/NOV ratios chosen for all sites in
A
this study are the following:
Median: 12
10th Percent!le: 7
90th Percentile: 23
The historical precursor trends for the Azusa source area are summarized
in Table 5 and Figure 15.
Figure 25 presents the basinwide isopleths and illustrates the pre-
diction of oxidant values for the Azusa validation; for reasons of simplic-
ity, only the 12:1 ratio is shown. The intersection of the isopleth corres-
ponding to the 1965 oxidant level and the appropriate NMHC/NOX ratio line
defines the reference point to which the changes in precursors are applied,
thus arriving at the predicted oxidant values for the years 1968, 1971, and
1974.* The point labeled "1965" is the reference point; it was found by the
intersection of the 0.33 ppm oxidant isopleth and the 12:1 ratio line. The
NOX and NMHC concentrations corresponding to the base year are read from
the respective axes. In this particular diagram, the base-year values for
Azusa are NMHC =1.5 ppmC and NOX = 0.125 ppm. The precursor trends pre-
sented in Table 5 are then applied to these levels to give the coordinates
of points for each successive three year period. For example, in the period
1967-1969 for the Azusa source area, NOX increased 17% ± 3%, and NMHC de-
creased 10% * 3%. Thus, the point labeled "1968" is determined. The process
is repeated to yield the points for 1971 and 1974.
The ellipses surrounding each point represent the uncertainties in pre-
cursor trends. These ellipses are drawn through four points: two from the
uncertainty in NOX (± 3% in 1968) and two from the uncertainty in NMHC (± 3%
in 1968). The error bounds in the predicted oxidant for each year are ob-
tained by taking the isopleth range that is covered by each ellipse.
The final step in the validation of the isopleths is to plot the pre-
dicted oxidant trends, with error bounds, on the same graph as the actually
*
Actually, these predictions are for the 3-year periods 1967-1969,
1970-1972, and 1973-1975.
64
-------
IO
<^
i. £
h- -P
CO
U
•i- CT
-a c
O) -r-
t- V)
Q. =3
LT>
CM
65
-------
observed oxidant trends, with error bounds. These graphs are shown in
Figures 26 through 28, each corresponding to a single NMHC/NOX ratio. The
overall agreement between the predicted trend lines and the actual trend
lines appears to be best for the 7:1 ratio and worst for the 23:1 ratio.
Figures 26 to 28 indicate that, for all three ratios, the isopleth
model tends to underpredict the net reduction in oxidant from 1965 to 1974.
The underprediction is very small for the 7:1 ratio, moderate and not
statistically significant for the 12:1 ratio, and fairly large and statis-
tically significant for the 23:1 ratio.
Figure 29 presents predicted trends for the 12:1 ratio and the maximum
possible error bounds based on both the errors in the precursor trends and
the range in the NMHC/NOx ratio. In other words, for any given year (1968,
1971, or 1974), the bottom error bound was found by taking the lowest error
bound for any ratio; similarly, the upper error bound was found by taking
the highest error bound for any ratio. Figure 29 indicates that the net
oxidant and precursor changes are too small, and the potential errors in
the analysis are too large, to arrive at a conclusive test of the isopleth
method. Figures 27 and 29 do raise some doubt concerning the predictions
of the method, especially if we accept 12:1 as the appropriate NMHC/NOX
ratio for 1965. However, considering the error bounds, we conclude that
the isopleth predictions are not inconsistent with historical trends in
the 95th percentile of daily maxima at Azusa.
Yearly Second Maximum One-Hour at Azusa
Figures 30 to 32 (corresponding to ratios of 7:1, 12:1, and 23:1) pre-
sent the results of the validation study for the basinwide isopleths using
the second highest yearly one-hour oxidant values at Azusa. Figure 33
summarizes the results, presenting the predicted trends for a 12:1 ratio
and the maximum possible error bounds based on errors both in the precursor
trends and in the range of the NMHC/NOx ratio.
The overall agreement between the actual trend line and the predicted
trend line is fair for the 7:1 ratio (Figure 30) and the 12:1 ratio (Figure
31), and very poor for the 23:1 ratio (Figure 32). Again, the predictions
66
-------
60-i
NMHC/NOX =7:1
50-
40-
.c
Q.
Q.
O
-------
60
50 -
40
£
Q.
Q.
o:
O
o
o
X
o
30
20 -
10 ~
NMHC/NOX =
12:1
A -r
Predicted
Oxidant
Trend
Actual
Oxidant
•*• Trend
A: Statistical error in ambient
ox1dant trends
B: Error in precursor trends
1
1964
1
1965
1
1966
1
1967
1
1968
1
1969
YEAR
1
1970
I
1971
1
1972
1
1973
1
1974
1
1975
Figure 27. Oxidant Trends in the 95th Percent!le
of the Daily Maxima at Azusa, Predicted
for 12:1 Ratio vs. Actual.
68
-------
60-,
NMHC/NOV = 23:1
50-
40-
A-r
Q-
CL
30-
(_>
o
-------
NMHC/NOX = 12:1 with a range of 7:1 to 23:1
40 -
30 _
Q-
Q.
o
-------
60—,
50—
40—
OL
CL
30_
X
o
20—
NMHC/NOx =7:1
A T
Predicted
Actual
Oxidant
Trend
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
i f i r | T r | i i ~] I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 30. Oxidant Trends in the Second Maximum for Azusa,
Predicted for 7:1 Ratio vs. Actual.
71
-------
60
50 _
40
e
Q.
Q.
NMHC/NOX =12:1
M-
^~^^^
L^^^
q — ^
-LB
T
1
J%
p>
T Predicted
' Oxidant
•^ Trend
Actual
(Oxidant
Trend
30 _
o
o
20 _
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
X
o
10
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 31. Oxidant Trends in the Second Maximum for Azusa,
Predicted for 12;1 Ratio vs. Actual
72
-------
60 _
NMHC/NOX =23:1
B
50-
40-
Q.
Q.
Predicted
-Q Oxidant
! Trend
Actual
Oxidant
Trend
5 30-
UJ
o
o
X
o
20-
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
10-
1 I ' 1 I ' ' I ' ' I '
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 32. Oxidant Trends in the Second Maximum for Azusa,
Predicted for 23:1 Ratio vs. Actual.
73
-------
60-1
NMHC/NCL « 12:1 with range of 7:1 to 23:1
A
50-
40'
Q.
Q.
T1
I
l
I
I
1 Predicted
1 Oxidant
Trend
Actual
Oxidant
Trend
30
O
•z.
O
O
Q
i—i
X
O
20'
A: Statistical error in ambient oxidant trends
B: Range in NMHC/NOX ratio and error in
precursor trends
10-
I T I I 1 I T I I I I I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 33. Summary of Oxidant Trends in the Second Maximum for
Azusa, Predicted vs. Actual.
74
-------
for all three ratios underestimate the net reduction in oxidant from 1965 to
1974. This underestimate is statistically insignificant for the 7:1 ratio,
marginally statistically significant for the 12:1 ratio, and very significant
for the 23:1 ratio.
Figures 30 to 33 do not provide a definitive test of the isopleth model.
Considering the potential errors, including the possible range of the NMHC/NOX
ratio, we conclude that the isopleth predictions are not inconsistent with
historical trends.
Yearly Second Maximum One-Hour, Basinwide
Figures 34 to 37 summarize the validation study for the basinwide isopleths
using the basinwide second maximum one-hour oxidant. The agreement between
predicted trends and actual trends is good for the 7:1 ratio (Figure 34),
fair for the 12:1 ratio (Figure 35), and very poor for the 23:1 ratio (Figure
36). Overall, the patterns and conclusions are similar to those for Figures
30 to 33.
VALIDATION OF ISOPLETHS FOR FIXED IRRADIATION TIMES
This section tests isopleths for fixed irradiation times against trends
at individual locations. Data for Downtown Los Angeles and Anaheim are used
with 5-hour isopleths; data for Azusa and San Bernardino are used with 7-hour
and 9-hour isopleths, respectively.
95th Percentile at Downtown Los Angeles (DOLA)
Figures 38 to 41 summarize the validation of 5-hour isopleths with trend
data for Downtown Los Angeles. The agreement between the predicted trend line
and the actual trend line is good for the 7:1 ratio (Figure 38) and poor for
the 12:1 and 23:1 ratios (Figures 39 and 40). The tendency noted before,
that the isopleth method underestimates historical oxidant reductions, is
even more evident here for the 12:1 and 23:1 ratios.
In the summary graph for the 12:1 ratio (Figure 41), we see that the
discrepancies between predicted and actual are within the overall error
bounds. The overall error bounds include statistical errors in ambient
oxidant trends, errors in estimated precursor trends, and the potential range
75
-------
60-,
50-
40-
E
Q.
•ZL
O
I—H
t—
-------
60-1
NMHC/NOX « 12:1
50-
40-
Q.
CL
Predicted
Oxidant
Trend
Actual
Oxidant
Trend
t 30-
o
o
O
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
X
o
20-
10 —
. | I I | I I J T I -p I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 35. Oxidant Trends in the Basinwide Second Maximum, Predicted
for 12:1 Ratio vs. Actual.
77
-------
NMHC/NOX =23:1
60-
50-
40-
A •*•
Predicted
Oxidant
Trend
Actual
Oxidant
Trend
30_
20-
10-
A: Statistical error 1n ambient
oxidant trends
B: Error in precursor trends
r 7 -i r | i i | i i | i
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 36. Oxidant Trends in the Basinwide Second Maximum, Predicted
for 23:1 Ratio vs. Actual.
78
-------
60
50 -
40-
Q.
CL
NMHC/NOX =12:1 with range of 7:1 to 23:1
B
T
I
i
Predicted
Oxidant
Trend
Actual
Oxidant
Trend
cc.
*z 30
o
o
o
a
i—i
X
o
20-
A: Statistical error in ambient oxidant trends
B: Range in NMHC/NOX ratio and error in precursor
trends
10
IT I I T I I II 1 I I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 37. Summary of Oxidant Trends in "the Basinwide Second
Maximum, Predicted vs. Actual.
79
-------
40 —
30 —
NMHC/NO¥ =7:1
A
Q.
Q-
o
C_5
Predicted Oxidant Trend
Actual Oxidant Trend
-••A
X
O
10 —
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
_, ! , , j , , r , , j ,
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 38. Oxidant Trends in the 95th Percentile of the Daily Maxima
at DOLA, Predicted for 7:1 Ratio vs. Actual.
80
-------
NMHC/NOv
12:1
40-
£
CL
o.
30-
LU
O
o
§
h—H
X
o
20-
Predicted
Oxidant
Trend
Actual
Oxidant
Trend
10 _
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
-! 1 . . 1 1 1 1 1 1 | I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 39. Oxidant Trends in the 95th Percentile of the Daily Maxima
at DOLA, Predicted for 12:1 Ratio vs. Actual.
81
-------
40-
CU
CL
30-
NMHC/NOX =23:1
o
o
o
o 20-
X
o
T Predicted
-O Oxidant
Trends
Actual
Oxidant
-1- Trends
10-
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 40. Oxidant Trends in the 95th Percentile of the Daily Maxima
at DOLA, Predicted for 23:1 Ratio vs. Actual.
82
-------
NMHC/NO,
= 12:1 with range of 7:1 to 23:1
40-
30-
Q.
CL
y 20-
o
o
Q
t—t
X
O
10-
A:
B:
Tl
Predicted
Oxidant
Trend
Actual
Oxidant
Trend '
Statistical error 1n
ambient oxidant trends.
Range 1n NMHC/NDX ratio
and error 1n precursor
trends.
I I I I I I I I I I | I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 41. Summary of .Oxidant Trends in the 95th Percentile at
DOLA, Predicted with 5-Hour Isopleths vs. Actual,
83
-------
in the NMHC/NOX ratio. If we do not consider the potential range in the
ratio (i.e. as in Figure 39), the discrepancies between actual and predicted,
for the 12:1 ratio, become very significant statistically. This raises sub-
stantial doubts concerning the consistency between historical oxidant trends
and the isopleth predictions (for the median ratio of 12:1).
We questioned whether the disagreement might be due to the specific
air quality index used. The validation precedure was repeated using the
90th percentile of daily maximum one-hour concentrations. No significant
improvement was obtained in the agreement between actual and predicted trends
for the 12:1 ratio.
95th Percentile at Anaheim
Figures 42 to 45 present the results of the validation study using the
5-hour isopleths with trend data for Anaheim. The tendency for the isopleth
method to underestimate the historical oxidant reductions in Los Angeles is
very evident here. The agreement between predicted and actual trends is fair
to poor for the 7:1 ratio (Figure 42), very poor for the 12:1 ratio (Figure
43), and very poor for the 23:1 ratio (Figure 44).
As shown in the summary graph (Figure 45), the differences between
predicted and actual trends are significant even if we consider
all three potential sources of error: statistical error in ambient oxidant
trends, error in precursor trends, and possible range in the NMHC/NOX ratio.
In the case of Anaheim, the isopleth method distinctly fails to pass the
verification test against historical oxidant trends.
95th Percentile at Azusa
Figures 46 to 49 summarizes the validation study using the 7-hour
isopleths with the 95th percentile of daily maximum oxidant at Azusa. The
overall agreement is excellent for the 7:1 ratio (Figure 46), poor for the
12:1 ratio (Figure 47), and very poor for the 23:1 ratio (Figure 48). For
all three ratios the isopleth model underpredicts the net reduction in oxidant
from 1965 to 1974. This underpredlction is very significant statistically
for the 12:1 and 23:1 ratios.
84
-------
NMHC/NOY =7:1
A
30 —
JT
Q.
Q.
O
i—i
;
i
Predicted
Oxidant
Trend
Actual
Oxidant
Trend
A:
Statistical error in ambient
oxidant trends
B: Error in precursor trends
-rr r- i i i i i i i i i
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 42. Oxidant Trends in the 95th Percentile of the
Daily Maxima at AnaTTeim, Predicted for 7:1
Ratio vs. Actual.
85
-------
NMHC/NOy = 12:1
30 —
e
Q-
O.
o
I—I
-------
30 —
Q.
Q.
20
o
o
X
o
10.
NMHC/NOy =23:1
Predicted
Oxidant
Trend
Actual
Oxidant
Trend
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
-r j r i | i i
1964 1965 1966 1967 1968 1969 1970
YEAR
I ' ' I 1
1971 1972 1973 1974 1975
Figure 44. Oxidant Trends in the 95th Percentile of the
Daily Maxima at Anaheim, Predicted for 23:1
Ratio vs. Actual
87
-------
OXIDANT CONCENTRATION (pphm)
1— > IK> GO
0 O 0
1 1 1
NMHC/NOX = 12:1 with range of 7:1 to 23:1
_£ T_
T
i Predicted
OJ Oxidant
j Trend
j
r Actual
Oxidant
^]* 1 1 rend
A: Statistical error in ambient
oxidant trends
B: Range in NMHC/NOX ratio and
error in precursor trends
i | i i [ i i |
1964 1965 1966 1967 1968 1969 1970 1971
YEAR
1972 1973 1974 1975
Figure 45. Summary of Oxidant Trends in the 95th
Percentile at Anaheim, Predicted with
5-Hour Isopleths vs. Actual.
-------
NMHC/NOX =7:1
40 _
E
Q.
O.
30-
20_
X
o
Predicted
Oxidant
• Trend
Actual
Oxidant
Trend
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
10-
I ' ' I ' ' I ^ T I '
1964 1965 1966 1967 1968 1969 1970 1971 9172 1973 1974 1975
YEAR
Figure 46. Oxidant Trends in the 95th Percentile of Daily Maxima
at Azusa, Predicted for 7:1 Ratio vs. Actual.
89
-------
40 _
Q.
Q.
30-
LU
CJ
O
O
-------
NMHC/NO =23:1
40 -
CL
d.
~ 30 -
UJ
o
o
O
T Predicted
-C| Oxidant
•*• Trend
Actual
Oxldant
Trend
20 -
X
o
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
10 _
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 48. Oxidant Trends in the 95th Percentile of Daily Maxima
at Azusa, Predicted for 23:1 Ratio vs. Actual.
91
-------
NMHC/NOx =12:1 with range of 7:1 to 23:1
40-
30"
Q.
Q.
I Predicted
Oxidant
Trend
Actual
Oxidant
Trend
20~
-------
As shown in Figure 49, the differences between predicted and actual
trends are within the overall error bounds. However, the major factor in
the overall error bounds is the range in the NMHC/NOX ratio, from 7:1 to
23:1. If we did not consider the possibility that 7:1 is the appropriate
ratio, statistically significant discrepancies would appear (as in Figure 47).
95th Percent!le at San Bernardino
The results of the validation study using the 9-hour isopleths with
the 95th percentile of daily maximum oxidant at San Bernardino are presented
in Figures 50 to 53. The agreement between the predicted and actual trend
lines is fair for all three ratios. The actual and predicted changes in
oxidant at San Bernardino are too small for a conclusive test of the isopleth
model.
Zeldin [20] has reported anomalies in the San Bernardino oxidant
that cannot be explained by meteorology. He attributes these anomalies to
instrumentation problems that were not noticeable enough at the time to
warrant exclusion of the measurements from the San Bernardino APCD data base.
Applying Zeldin1s correction factors for the anomalous data affects only
the 1970-1972 point; the actual oxidant point for those years is moved up
slightly to be in better agreement with the predicted points.
93
-------
40-
30 _
NMHC/NOX =7:1
O.
Q.
20_
O
O
Predicted Oxidant Trend
Actual Oxidant Trend
AT
X
O
10-
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 50. Oxidant Trends in the 95th Percentile of Daily Maxima
at San Bernardino, Predicted for 7:1 Ratio vs. Actual.
94
-------
40-
NMHC/NOX =12:1
30 _
-C
CL
CL
o
I—I
<
20-
Predicted Oxidant Trend
*
Actual Oxidant -*•
Trend
o
X
o
10-
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
—i , i , , 1 1 , , , 1 ,
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 51. Oxidant Trends in the 95th Percentile of Daily Maxima
at San Bernardino, Predicted for 12:1 Ratio vs. Actual.
95
-------
40-
NMHC/NOX = 23:1
30-
a.
Q.
o
I—)
I—
•=c
O
•ZL
O
Predicted Oxidant Trend
20-
Actual Oxidant
Trend
X
O
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
I ' ' I n ' I ' ' I '
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 52. Oxidant Trends in the 95th Percentile of Daily Maxima
at San Bernardino, Predicted for 23:1 Ratio vs. Actual.
96
-------
40 _
30
NMHC/NOX = 12:1 with range of 7:1 to 23:1
ex
QL
O
O
O
Predicted Oxidant Trend
20 ~
JL Actual Oxid,an,t
Trend
X
O
A:
B:
Statistical error in ambient
oxidant trends
Range in NMHC/NOX ratio and error
in precursor trends
—] 1 ] 1 ] 1 1 1 1 1 I I
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 53. Summary of Oxidant Trends in the 95th Percentile at
San Bernardino, Predicted with 9-Hour Isopleths vs.
Actual.
97
-------
CHAPTER 5
DISCUSSION OF RESULTS
If one accepts 12:1 as the appropriate NMHC/NOV ratio in 1965 (equi-
A
valent to an 8:1 ratio in the early 1970's), the validation studies indicate
significant discrepancies between historical air quality trends in Los
Angeles and the predictions of the EKMA isopleth method. Considering the
statistical errors in actual oxidant trends and the potential errors in our
estimates of precursor trends, we found significant differences
between the isopleth predictions (for a 12:1 ratio) and actual oxidant
trends in four of the seven situations that were analyzed. Only if we con-
sider a range in the NMHC/NOX ratio, in particular the possibility that the
ratio may have been as low as 7:1, do most of these discrepancies become
statistically insignificant.
The disagreement for a 12:1 NMHC/NOXratio is highlighted in Table 18 which
lists the actual and predicted changes in oxidant from 1964-1966 to 1973-
1975. Although the isopleth method usually predicts the right direction of
the change, it always underpredicts the magnitude of the change. In the
central parts of the Los Angeles basin (all stations but San Bernardino),
the isopleth method substantially underpredicts the reductions in oxidant
that have actually occurred; this underprediction is especially large in
the tests involving isopleths for fixed irradiation times.
This chapter discusses possible reasons for the disagreement and
potential improvements in the isopleth method. First we eliminate those
factors which would not account for the observed discrepancies; then we
list the factors which may be the cause of the discrepancies and describe
possible modifications to the isopleth method.
FACTORS NOT ACCOUNTING FOR THE DISAGREEMENT
We have investigated several factors which might contribute to the
discrepancies between the isopleth predictions and actual oxidant trends.
Before describing the factors that are the most likely explanations for
the disagreement, it is useful to discuss the factors that we have been
able to eliminate as plausible reasons for the disagreement.
Ambient NMHC/NOX Ratio on High Oxidant Days
Our estimate of the ambient NMHC/NOX ratio is based on 6-9 AM data
for the entire summer smog season. It would be more appropriate to use the
98
-------
Table 18. Summary of Actual and Predicted Oxidant Changes
1965 to 1974 (NMHC/NOX Ratio of 12:1)
VALIDATION STUDY
ACTUAL %
OXIDANT CHANGE,
1965-1974
PREDICTED %
OXIDANT CHANGE,
1965-1974
Basinwide Isopleths, Azusa
95th Percentile
Basinwide Isopleths, Azusa
2nd Maximum
Basinwide Isopleths, Basinwide
2nd Maximum
5-Hour Isopleths, DOLA
95th Percentile
5-Hour Isopleths, Anaheim
95th Percentile
7-Hour Isopleths, Azusa
95th Percentile
9-Hour Isopleths, San
Bernardino, 95th Percentile
-18%
-21%
-18%
-28%
-29%
-18%
+ 9%
- 9%
*
- 8%
- 8%
-14%*
+ 5%*
*
- 1%
+ 6%
significant difference based on potential errors in
actual oxidant trends and in estimates of precursor trends,
6-9 AM ratio on days of highest oxidant. A sensitivity analysis (see Chapter
3) reveals that the median ratio on high oxidant days is 10 to 20% higher
than the median ratio on all summer days.
The foregoing consideration indicates that a median ratio of approx-
imately 14:1 in 1965 might be more appropriate than a median ratio of 12:1.
This, however, would not explain the discrepancies in the validation studies.
In fact, using a 14:1 ratio would slightly increase the disagreement be-
tween the isopleth predictions and actual oxidant trends.
6-9 AM Summertime Precursor Trends
The "best estimates" of precursor trends that we have used in the
validation studies are essentially based on yearly average changes in pre-
cursor emissions and ambient precursor concentrations. A more appropriate
precursor trend index for testing the isopleth method would be changes in
99
-------
ambient 6-9 AM summertime concentrations. A sensitivity analysis (Chapter
3) indicates that our "best estimates" are representative of ambient trends
in 6-9 AM summertime NMHC but may underestimate the increase in 6-9 AM sum-
mertime NOX by 10 to 25%.
Using a greater historical NOX increase would affect our verification
study in two ways. First, in extrapolating the present NMHC/NOX ratio (8:1)
backwards to 1965, we would arrive at a higher median ratio (say 14:1 or
15:1 instead of 12:1). As noted earlier, this would worsen the discrepancies
in the validation study. Second, the NOX level of the predicted points on
the isopleth model would be increased. In the cases involving the median
NMHC/NOX ratio, this would increase the predicted oxidant levels, again
making the discrepancies greater. Thus, if we increased our estimate of the
historical rise in NOX to be representative of 6-9 AM summertime trends,
we would only worsen the discrepancies in the validation studies.
Monitoring Changes
An obvious factor that could lead to disagreement between the isopleth
predictions and actual oxidant trends would be errors produced by monitor-
ing changes. Such errors could be introduced either in the ambient precursor
trends or the actual oxidant trends. We expect, however, that such errors
will be minimal for the following reasons:
• None of the monitoring stations changed location during
the period.
• The same analytical methods were used throughout the period
(flame ionization for hydrocarbons, colorimetric for NOX, and
colorimetric for oxidant).
t The trends were continual over the period and were usually
consistent among stations located in the same part of the
basin.
• The trends at DOLA and Anaheim provide an independent check
on changes in monitoring practices since the data are col-
lected by two separate monitoring agencies.
POSSIBLE EXPLANATIONS FOR THE DISAGREEMENT
There are several factors that could account for the discrepancies be-
tween the isopleth predictions and historical oxidant trends. These factors
100
-------
are discussed in the paragraphs that follow. To determine which of these
factors is most critical would require additional research effort (see
recommendations for future work in Chapter 1).
Atmospheric NMHC Versus Isopleth NMHC
One possible reason for the observed discrepancies could be that the
median ratio of 12:1 is inappropriate. In Chapter 4 (Figures 26 to 53), we
found better agreement between predicted trend lines and actual trend lines
for the 7:1 ratio than for the 12:1 ratio. As evidenced by Table 19, the
7:1 ratio leads to a much better prediction of the net oxidant changes from
1965 to 1974.
Table 19. Summary of Actual and Predicted Oxidant Changes,
1965 to 1974 (NMHC/NOX Ratio of 7:1 and 12:1)
VALIDATION STUDY
ACTUAL %
OXIDANT CHANGE,
1965 to 1974
PREDICTED %
OXIDANT CHANGE,
1965 to 1974
7:1 RATIO 12:1 RATIO
Basinwide Isopleths, Azusa
95th Percentile
Basinwide Isopleths, Azusa
2nd Maximum
Basinwide Isopleths, Basinwide
2nd Maximum
5-Hour Isopleths, DOLA
95th Percentile
5-Hour Isopleths, Anaheim
95th Percentile
7-Hour Isopleths, Azusa
95th Percentile
9-Hour Isopleths, San
Bernardino 95th Percentile
-18%
-21%
-18%
-28%
-29%
-18%
+ 9%
-15%
-15%
-14%
-32%
-12%*
-12%
0%
- 9%
- 8%*
- 8%
-14%*
+ 5%*
*
+ 1%
+ 6%
significant difference based on potential errors in actual
oxidant trends and in estimates of precursor trends
The ambient data for NMHC and NOX (Chapter 3) gave us fairly good con-
fidence that the median atmospheric 6-9 AM ratio was 12:1 (or slightly high-
er) in 1965. However, a 12:1 atmospheric ratio may not be equivalent to a
101
-------
12:1 ratio in the isopleth model (which is a mathematical model using
propylene and n-butane, calibrated with smog-chamber results using auto
exhaust). It is possible that a given level of ambient NMHC in Los Angeles
is equivalent to a lower level of NMHC in the isopleth model. This would be
the case if the atmospheric NMHC were of lower reactivity (per ppmc) than
the isopleth NMHC. Thus, an ambient ratio of 12:1 may possibly be equi-
valent to a ratio of 7:1 in the isopleth model.
To investigate this factor further, the reactivity of atmospheric NMHC
in Los Angeles should be compared to the reactivity of the isopleth NMHC mix.
Such a reactivity analysis should consider both the number of moles per ppmc
and the oxidant producing potential per mole of the hydrocarbons.
Post 9 AM Emissions
The existing versions of the EKMA isopleths relate ozone to initial
NOX and NMHC (assumed equivalent to 6-9 AM NOX and NMHC), neglecting em-
issions after 9 AM. It is expected that NOX emitted after 9 AM would act
more as an ozone inhibitor than initial (6-9 AM) NOX. If post 9 AM emis-
sions were added to the model, the isopleths in the upper left-hand corner
of Figure 1 (or Figure 25) should bend more to the right because of greater
ozone inhibition from NOX. The critical ratio in the isopleth model (the
top of the ozone "hill" or the ratio at which hydrocarbon control becomes
effective) might also become larger. These effects would tend to reduce
the discrepancies between the isopleth predictions and historical oxidant
trends in Los Angeles.
The addition of post 9 AM emissions should be most important for short
irradiation times; the most significant changes should occur in the 5-hour
isopleths. This is encouraging because the greatest discrepancies between
actual and predicted values have been found in the cases involving short
irradiation times, i.e. DOLA and Anaheim.
EPA is presently investigating the possibility of adding post 9 AM
emissions to the isopleth model (Personal communication with Edwin Meyer,
EPA Office of Air Quality Planning and Standards, Durham, N.C., November
1977). When these new isopleths become available, the verification studies
should be repeated. We would expect the results of the verification tests
to improve.
102
-------
Meteorology
Another reason for the disagreement could be meteorological bias in
the actual oxidant trends. The greatest discrepancies occur in 1973-1975,
when actual oxidant is, in most cases, substantially lower than predicted
oxidant. This may, in part, be due to meteorology; it has been previously
noted [21, 22] that pollution potential in Los Angeles appeared to be lower
in 1973-1975 than in 1964-1966.
It would be useful in future work to normalize the actual oxidant trends
for meteorological variance. This would provide a more appropriate test
of the isopleth method. Normalization for meteorology should also decrease
the error bounds on the actual oxidant trends, resulting in a more finely-
tuned validation study. Zeldin and Meisel [23] have recently completed a
guideline document for EPA on meteorological normalization of air quality
trends; in the future they may be applying their method to the Los Angeles
oxidant data.
Source Area Definition
Another potential explanation for the disagreement is that the source
areas have not been properly defined. Perhaps the precursor changes of con-
sequence are the precursor changes in the sub-areas of greatest emission
density (which have low growth rates) rather than the precursor changes
throughout the entire upwind area.
To assess the effect of redefining source areas, we repeated the
analysis for the Azusa second maximum using the precursor trends for the
DOLA source region (a high-density/low-growth sub-area of the Azusa source
region). The results of this analysis are presented in Figures 54 to 57.
There is some improvement in the verification study for the 12:1 ratio
(compare Figure 55 to Figure 31), but predicted oxidant still exceeds
actual oxidant in 1974. Overall, the predicted values for a 12:1 ratio
using the DOLA source area (Figure 55) resemble the predicted values for a
7:1 ratio using the Azusa source area (Figure 30).
103
-------
NMHC/NOX = 7:1
60-
50 —
40 —
Q.
a.
Actual
Oxidant
Trend
Predicted
Oxidant
Trend
T
I
i
30 —
o
o
§
hi
X
o
20 —
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
10—
^ I r ' I ' ' I ' ' I '
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 54. Oxidant Trends in the Second Maximum for
Azusa, Predicted for 7:1 Ratio vs. Actual,
Predicted Values Based on DOLA Source Region
104
-------
NMHC/NOX =12:1
50 _
40
Predicted
"T Oxidant
Of Trend
Actual
Oxidant
Trend
Q.
EX
30 _
UJ
O
20_
X
O
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 55. Oxidant Trends in the Second Maximum for
Azusa, Predicted for 12:1 Ratio vs. Actual,
Predicted Values Based on DOLA Source Region
105
-------
60 _
NMHC/NOV =23:1
A
50_
40_
O-
Q-
O
7 Predicted
Oxidant
Trend
Actual
Oxidant
Trend
30_
UJ
o
o
o
X
O
20_
A: Statistical error in ambient
oxidant trends
B: Error in precursor trends
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 56. Oxidant Trends in the Second Maximum for
Azusa, Predicted for 23:1 Ratio vs. Actual,
Predicted Values Based on DOLA Source Region
106
-------
60-,
NMHC/NOX = 12:1 with range of 7:1 to 23:1
50-
40-
E
Q.
Q-
30_
O
X
O
20-
Predicted
Oxidant
Trend
Actual
Oxidant
Trend
Statistical error in ambient
oxidant trends
Range in NMHC/NOX ratio and
error in precursor trends
10
—r -j i -i i i i | . r- -] i
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
YEAR
Figure 57. Summary of Oxidant Trends in the Second
Maximum for Azusa, Predicted vs. Actual,
Predicted Values Based on DOLA Source Region
107
-------
Spatial Coverage of Oxidant Monitoring Stations
Our analyses with the basinwide EKMA isopleths rely essentially on
oxidant trend data for the Azusa monitoring site. Oxidant trends at Azusa
may not be representative of the trends in the basinwide oxidant maximum;
in particular, the historical oxidant decrease at Azusa may have been
greater than the decrease in the basinwide oxidant maximum. This is
plausible because the location of the basinwide oxidant maximum has been
shifting eastward, downwind of Azusa, as reductions in the NMHC/NO ratio
* x
have retarded the photochemical reaction rates. The oxidant maximum at
Azusa may have decreased relative to the basinwide maximum because of this
spatial shift.
That we may have overestimated the historical decrease in the basin-
wide oxidant maximum because of the limited spatial coverage of the mon-
itoring stations could explain some of the discrepancies in the verification
tests using the basinwide isopleths. This, however, would not explain the
even greater discrepancies found in the tests using isopleths for fixed
irradiation times.
Inappropriate definition of source areas and limited spatial coverage
of oxidant monitoring stations are possible factors contributing to the ob-
served discrepancies. However, it is our opinion that the three most likely
explanations for the disagreements are (1) non-equivalency between atmos-
pheric NMHC and chamber NMHC, (2) omission of post 9 AM emissions in the
isopleth model, and (3) meteorological bias in the actual oxidant trends.
*
Maximal oxidant in the Los Angeles basin presently tends to occur near
Upland, approximately 20 miles downwind of Azusa. As explained in Chapter
2, we did not include Upland in our trend analysis because only three years
of data were avaliable for that location.
108
-------
REFERENCES
1. "Alternatives for Estimating the Effectiveness of State Implementation
Plans for Oxidant," Draft Report by the Air Management Technology Branch,
Monitoring and Data Analysis Division, Office of Air Quality Planning
and Standards, Environmental Protection Agency, March 1977.
2. Dimitriades, B., "Oxidant Control St^tegies. Part I. Urban Oxidant
Control Strategy Derived h>'->. ; >isting Smoq Chamber Data, " Environ. Sci.
Techno!.. 11, 80 (1977).
3. California Air Resources Board, "Ten-Year Summary of California Air
Quality Data 1963-1972," and "Three-Year Summary of California Air
Quality Data 1972-1975," Technical Services f)ivision, Sacramento,
Calif., January 1974 and November 1976.
4. Eldon, J. and J. Trijonis, "Statistical Oxidant Precursor Relationships
for the Los Angeles Region, Part i. Data Quality Review and Evaluation,"
Interim Report to the Air Resources Board under Contract NO. A5-020-87,
January 1977.
5. "California Air Quality Data," Vol. 7, #4, p. 3-5, California Air
Resources Board, Technical Services Division, Sacramento, Calif.,
October-December 1975.
6. r-ieiburget , A., IN A. Renze'ti, R, Tice, "Wind Trajectory Studies
of the Movement of Polluted Air in the Los Angeles Basin," Report #13
of the Southern California Air Pollution Foundation, April, 1956.
7. Hurst, W. C., Draft Environmental Impact Report: Paktank Pacific
Company Oil Storage Terminal, Terminal Island, Los Angeles Harbor,
Los Angeles Harbor Department, July 1975,
8. DeMarrais, G. A., G. C. Holzworth, C. R. Hosier, "Meteorological
Summaries Pertinent to Atmospheric Transport and Dispersion over
Southern California," U.S. Weather Bureau, Technical Paper #54, 1965.
9. Tiao, G. C., G. E. P. Box, W. J. Hamming, "Analysis of Los Angeles
Photochemical Smog Data; A Statistical Overview," Journal of the
Air Pollution Control Association, Vol. 25 #3. p. 260, March 1975.
10. Vaughan, L. M., A. R. Stankunas, "Field Study of Air Pollution Trans-
port in the South Coast Air Basin," Metronics Associates, Inc., pre-
pared for California Air Resources Board, Contract #ARB-658, July 1974.
11. Stevenson, R. E., " Winds Over Coastal Southern California," Bulletin
of Southern California Academy of Sciences, Vol. 59, part 2, p. 103,
1960.
109
-------
12. Poppendiek, H. F., J. 6. Edinger, M. L. Greenfield, W. J. Hamming,
L. H. McEwen, "A Report on an Atmospheric Pollution Investigation
in the Los Angeles Basin," University of California, Departments of
Engineering and Meteorology, prepared for Los Angeles Air Pollution
Control District, June 1948.
13. Pack, D. H. and J. K. Angell, "A Preliminary Study of Air Trajectories
in the Los Angeles Basin as Derived from Tetroon Flights," Monthly
Weather Review, p. 583, October-December 1963.
14. Neiburger, M., J. G. Edinger, "Summary Report on Meteorology of the
Los Angeles Basin with Particular Respect to the .'Smog' Problem,"
Report #1 to the Southern California Air Pollution Foundation,
April 1954.
15. Trijonis, J., T. Peng, G. McRae, and L. Lees, "Emissions and Air
Quality Trends in the South Coast Air Basin," EQL Memo NO. 16.
Environmental Quality Laboratory, California Institute of
Technology, Pasadena, California, January 1976.
16. Trijonis, J., G. Richard, R. Tan, R. Wada, and K. Crawford, "An
Implementation Plan for Suspended Particulate Matter in the Los
Angeles Region," TRW Environmental Services, EPA Contract No. 68-02-1384,
March 1975.
17. Trijonis, J. C., "An Economic Air Pollution Control Model--Application:
Photochemical Smog in Los Angeles County in 1975," Ph.D. Thesis,
California Institute of Technology, Pasadena, 1972.
18. Kinosian, J. and J. Paskind, "Hydrocarbon, Oxides of Nitrogen, and
Oxidant Trends in the South Coast Air Basin 1963-1972," Division of
Technical Services, California Air Resources Board, 1974.
19. Meyer, E.,W. Freas, and J. Summerhays, "Procedures for Quantifying
Relationships Between Photochemical Oxidants and Precursors," Draft
Report by Monitoring and Data Analysis Division, Office of Air Quality
Planning and Standards, Environmental Protection Agency, August 1977.
20. Zeldin, M., "Weather Adjusted Oxidant Trends for Selected Cities in
the South Coast Air Basin," Statewide Air Pollution Research Center,
University of California, Riverside, California, 1976.
21. Horie, Y., J. Trijonis, "Analysis and Interpretation of Trends in Air
Quality and Population Exposure in the Los Angeles Basin," prepared
for EPA Office of Air Quality Planning and Standards by Technology
Service Corporation under Contract No. 68-02-2318, March 1977.
110
-------
22. Southern California Air Pollution Control District, Metropolitan Zone,
"Air Quality and Meteorology - 1975 Annual Report," 1976
23. Zeldin, M. and W. Meisel, "Guideline Document for Meteorological
Adjustment of Air Quality Data," prepared by Technology Service
Corporation under Contract No. 68-02-2318 to the EPA Office of Air
Quality Planning and Standards, November 1977.
Ill
-------
APPENDIX A
TABLE OF OXIDANT TREND DATA
STATION
OXIDANT AIR
QUALITY INDEX
TRENDS IN 3-YEAR AVERAGES (pphm)
1964-66 1967-69 1970-72 1973-75
BASINWIDE ANNUAL SECOND HIGHEST
ONE HOUR CONCENTRATION 49.0 52.0 47.0
39.7
AZUSA ANNUAL SECOND HIGHEST
ONE HOUR CONCENTRATION 47.3 50.0 47.0 36.7
95TH PERCENTILE OF DAILY
MAXIMAL ONE HOUR
CONCENTRATIONS 33.3 35.0 32.0 27.0
DOWNTOWN LA
95TH PERCENTILE OF DAILY
MAXIMAL ONE HOUR
CONCENTRATIONS
25.3
22.3
18.0 18.0
ANAHEIM
t
18.9
17.0
13.0 13.5
SAN BERNARDINO
,t
21.1
20.3
21.1 23.0
t
Oxidant measurements taken at locations outside Los Angeles County have
been multiplied by 0.80 to account for differences in calibration pro-
cedures.
Two-year average (1973-74).
112
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-6QO/3-78-019
3. RECIPIENT'S ACCESSION»NO.
4. TITLE AND SUBTITLE
VERIFICATION OF THE ISOPLETH METHOD FOR RELATING
PHOTOCHEMICAL OXIDANT TO PRECURSORS
5. REPORT DATE
February 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
J. Trijom's
D. Hunsaker
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG \NIZATION NAME AND ADDRESS
Technology Service Corporation
2811 Wilshire Boulevard
Santa Monica, CA 90403
10. PROGRAM ELEMENT NO.
1AA603 AC-29 (FY-78)
11. CONTRACT/GRANT NO.
68-02-2299
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Historical trend data for oxidant concentrations in the Los Angeles region
were used to check the isopleth method that has been proposed as a replacement
for the Appendix J method for relating oxidant to non-methane hydrocarbon (NMHC)
and nitrogen oxide (NO ) precursors. Using the median 6-9 AM NMHC/NO ratio
measured during the summer as input to the isopleth model, significant discrep-
ancies were found between the isopleth predictions and actual oxidant trends.
Using a range in the NMHC/NO ratio, in particular a low value for the ratio,
much better agreement was fofind between the predicted and actual trends.
Potential explanations for the discrepancies are discussed.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATi Field/Group
*Air pollution
*0zone
*Mathematical Models
Verifying
Los Angeles
•13B
07B
12A
14B
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report;
UNCLASSIFIED
21. NO. OF PAGES
123
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
113
-------
------- |