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

-------

-------