Environmental Protection Agency
Region IX
100 California Street
San Francisco, California 94111
October 30, 1973
Technical Support Document for the
Metropolitan Los Angeles Intrastate
Air Quality Control Region
Transportation Control Plan Final Promulgation
Published in
November 12, 1973 Federal Register
-------
I. Introduction:
This document is in support of the EPA promulgated
California Transportation Control Plan for the Los Angeles
Metropolitan Air Quality Control Region (AQCR), signed on
October 30, 1973 by Acting Administrator John Quarles, and
published in the November 12, 1973 Federal Register.
The determination of the maximum amount of allowable
emissions consistent with the attainment of the National
Ambient Air Quality Standards, and the emission reduction
strategies needed to reduce emissions to the maximum allow-
able levels, are outlined and discussed in the following two
sections. More detailed information on the control strate-
gies and technical details involved in the plan are found
in the appendices.
II. Determination of Allowable Emission to Meet Federal
Ambient Air Quality Standards:
The National Ambient Air Quality Standard for photochemical
oxidant has been exceeded in this AQCR. The photochemical
oxidant control strategy discussed in this report involves
the control of high reactive hydrocarbon and other reactive
organic gases (RHC). Where possible, the RHC are defined by
the most recent EPA guidance on organic gas reactivity.
Using the EPA definition of RHC, only the following five
hydrocarbons are considered as low or non-reactive: methane,
ethane, propane, benzene, and acetylene. The EPA definition
of RHC was applied to all mobile exhaust emission sources,
and to gasoline vapor emissions from stationary as well as
mobile sources.
Due to the lack of a detailed breakdown of the organic gas
emissions from stationary sources, the Los Angeles County
Air Pollution Control District (APCD) Rule 66 chemical defi-
nition of reactivity was used for the remaining stationary
sources. Application of the Rule 66 chemical definition of
reactivity means that only those organic gases listed under
Section 1., 2., 3. of paragraph K. of Rule 66 (e.g., toluene,
aromatic compounds with eight or more carbon atoms, olefinic
hydrocarbons), are inventoried as RHC and considered for
control. It is expected that in the future, a more detailed
- 1 -
-------
stationary organic gas emission inventory will be available,
which will allow for a completely consistent definition of
RHC to be made for both mobile and stationary emission
sources of organic gases. An observation that can be made
is that the RHC definition inconsistency between the mobile
and stationary source emissions in the present RHC inventory
results in the stationary source emissions (except stationary
gasoline vapor emissions) being artificially low in relation
to mobile source emissions.
The nitrogen dioxide (NC>2) ambient air quality standard has
been exceeded in the AQCR, with a 1970 maximum yearly arith-
metic mean of .094 parts per million (p.p.m.), versus the
standard of .05 p.p.m. The control of RHC for meeting the
photochemical oxidant standard is the critical factor how-
ever, and the implementation of strategies required to meet
the oxidant standard will also result in adequate NOX
emission reductions for the attainment of the NO2 standard.
The carbon monoxide (CO) standard has been exceeded, with a
high 8-hour reading of 41 p.p.m. occurring in 1970, versus
the standard of 9 p.p.m. The control of RHC for meeting the
photochemical oxidant standard is still the critical factor,
and the control strategies required to meet the oxidant
standard should result in more than adequate CO emission
control to meet the CO standard.
The photochemical oxidant 1-hour standard is .08 p.p.m. The
critical yearly high 1-hour photochemical oxidant reading of
.62 p.p.m., occurred in the AQCR at Riverside in 1970. The
stationary RHC emissions in the 1970 base year are estimated
to be 255 tons/day, and the mobile RHC emissions are esti-
mated to be 1346 tons/day.
As a result of recommendations received at EPA public hear-
ings and technical meetings, EPA did a statistical analysis
of ambient photochemical oxidant data. The statistical
analysis entitled "Methodology for Determining the Base Year
Oxidant Level," is found in Appendix B. The results of this
study validate the use of the observed AQCR yearly high
1-hour 1970 oxidant reading of .62 p.p.m. for control
strategy planning purposes.
- 2 -
-------
Mr. Ed Schuck of EPA developed "upper limit" curves for
daily 6-9 a.m. average non-methane hydrocarbon concen-
trations versus the daily maximum 1-hour oxidant reading
over a period of time at several locations in this AQCR.
This "upper limit" relationship and others, are detailed in
the Schuck Papetti paper in Appendix D, entitled "Exami-
nations of the Photochemical Air Pollution Problem in the
Southern California Area." From this relationship it is
inferred that a 93% reduction in the critical 1970 ambient
hydrocarbon concentration would have to occur in order to
allow for the attainment of the oxidant air quality standard
(i.e., lower the .62 p.p.m. reading to .08 p.p.m.). It is
assumed that a 93% reduction in RHC emissions in the AQCR
would result in a 93% reduction of ambient hydrocarbon con-
centrations. The maximum allowable RHC emission rate in
1970 and any future year is then calculated as follows,
based on the 1970 base year emission rate, and the Schuck
"upper limit" relationship:
.07(255 + 1326) = 112 tons RHC/day
III. Control Strategy Outline:
The EPA rules and regulations that are to affect the major-
ity of the emission reductions outlined in this and the
previous section, are identified in section "I. Intro-
duction" of this document.
- 3 -
-------
Summary of Impact of
Transportation Control Regulations
In the Los Angeles AQCR in 1977
Emissions and
Reductions of
Emission Source and Control Measures RHC tons/day
Stationary source emissions I/ 236
without EPA control strategy
Expected reductions
1. Vapor recovery at gasoline stations -132
2. Dry cleaning, paint and degreasing -32
solvent controls
Stationary emissions remaining 72
Mobile source emissions without 2/ 613
EPA control strategy
Expected reductions
1. Reductions from only EPA-promulgated -61
VMT* control strategies, assuming a
conservative 14% VMT reduction 3_/
2. Catalyst retrofit, and mandatory -103
inspection and maintenance
3. Motorcycle limitations -24
4. VMT reductions and evaporative -384
emission reductions necessary from
additional control strategies to be
implemented in 1977
Mobile emissions remaining 41
Total emission remaining 113
Total emission allowable 112
*VMT is an abbreviation for "vehicle miles traveled."
-------
I/ Stationary Source
Emission Breakdown:
Petroleum Production 6
Petroleum Marketing 132
Organic Solvents 95
Miscellaneous 3
236
2/ Mobile Source
Emission Breakdown:
Petroleum Marketing 20*
Ships & RR 6
Aircraft 25
Motorcycles 45
Heavy Duty Vehicle
(HDV) Diesel 24
HDV Gasoline 56
Light Duty Vehicle
(LDV) Gasoline 437
613
*This is the amount of petroleum marketing emissions
remaining after gasoline vapor stationary controls are imple-
mented, and therefore can only be reduced by VMT reduction
measures.
3_/ Using optimistic assumptions and estimates for both EPA
and local VMT reduction measures, a total reduction of 43%
VMT, or 187 tons/day could occur. A discussion of the basis
or rationale for the VMT reductions is found in Appendix C
"California VMT Reduction Summary."
Technical reports, control tactic information (including such
details as the emission control reduction factors and the pop-
ulation fraction affected by the tactics), and other data and
information needed to calculate or understand the emission
inventory in the preceding table, are outlined or referenced
in the following appendices.
-------
APPENDIX A
Data and References Used In
Emission Inventory Calculations
A. Stationary, Aircraft, Ship and Railroad Emissions
1) Stationary emissions and ship and railroad emissions
are based on the California Air Implementation Plan
emission inventory. The aircraft emission inventory
was calculated by EPA, Washington, D.C. Headquarters
staff. The base year emissions are as follows:
1970 RHC
Emission Category Emissions (tons/day)
Petroleum Refining 5.0
Petroleum Marketing 137.0*
Solvent Users 110.8**
Agriculture, Incineration, 2.4***
Combustion
Aircraft 38.0*
Ships & Railroads 5.4*
Considering EPA reactivity factor, see Section F.
**The surface coating segment of the "solvent users"
category (49.2 tons RHC/day in 1970) is estimated by
EPA in 1977 to be half of that which would be pro-
jected from the growth projections in Section D.,
based on the increased effectiveness and incentives
of the present Rule 66.
***Based on an updated Ventura County emission
inventory in which agriculture RHC emissions are
reduced.
Future or projected emissions, not considering pro-
posed or additional controls, are obtained by apply-
ing the appropriate growth factors (see Section D.)
to the 1970 base year inventory just discussed. An
exception to this is the aircraft inventory, which
is estimated by EPA to be reduced to 25 tons/day in
1977.
- 1 -
-------
B. Emission Factors For Vehicles
1) Light and heavy duty vehicle (LDV & HDV) gasoline,
HDV diesel, and motorcycle emission factors (includ-
ing deterioration factors where applicable), were
obtained from the following document:
Compilation of Air Pollutant Emission Factors
(AP-42) 1973 Edition
Available from:
EPA Office Technical Information & Publications,
Office of Air Programs, Research Triangle Park,
N. C. 27711
The emission deterioration factors in the EPA AP-42
publication are presented as a function of vehicle
age. This analysis, however, relates the deterio-
ration factors to accumulated mileage. The accumu-
lated mileages that are associated with the vehicle
ages in AP-42, are as follows:
AP-42 Accumulated
Vehicle Age Mileage
1 17,500
2 33,600
3 46,800
4 58,200
5 69,900
6 79,900
7 90,200
8 98,800
9+ 109,700+
The emission factors presented in AP-42, are listed
for various model years in terms of grams of pollut-
ant emitted per mile travelled by the vehicle.
C. Vehicle Population, Age Distribution, and Mileage Data
1) Population data obtained from California Air Resources
Board:
- 2 -
-------
a) 1970:
Statewide gasoline LDV population - 10,560,870
Statewide gasoline HDV population - 277,120
Statewide diesel HDV population - 84,500
Statewide motorcycle population - 568,000
Los Angeles AQCR % of statewide population is
49.72%.
b) 1972:
Statewide gasoline LDV population - 11,331,900
Statewide gasoline HDV population - 296,300
Statewide diesel HDV population - 94,800
Los Angeles AQCR % of statewide population is
49.52%.
These populations are used as the most recent
base year data from which to project future
year populations. Motorcycle population pro-
jections are made from the 1970 base year popu-
lation.
C) 1977:
Los Angeles AQCR % of statewide population is
49.18%. (This factor is used only for project-
ing the AQCR 1977 motorcycle population, see
Section D.2.)
2) Vehicle Age Distributions
-------
a) 1970 (July):
LDV** Gasoline & Diesel**
Vehicle Age*(Yr) % of Population HDV % of Population
3/8 8.0 6.7
1 1/4 11.1 9.7
21/4 9.5 7.8
3 1/4 8.4 6.4
4 1/4 9.2 7.4
5 1/4 9.5 8.0
61/4 8.5 7.4
71/4 7.4 6.3
81/4 6.3 5.2
91/4 4.4 3.9
10 1/4 4.1 4.2
11 1/4 2.9 3.7
12 1/4 1.6 2.3
13 1/4 2.0 2.9
1,4 1/4 1.7 3.1
15 1/4 1.5 2.8
15 1/4+ 3.8 12.3
*The 3/8 year old vehicles are 1970 models, the 1/4
year old models are 1969 models, etc.
**Based on State of California Air Resources Board
Department of Motor Vehicle data.
-------
b) Post-1972 (July):
LDV** HDV**
Vehicle Age*(Yr) % of Population % of Population
3/8 7.9 7.2
1 1/4 9.9 8.9
2 1/4 9.5 8.0
31/4 9.2 7.5
4 1/4 8.9 7.1
5 1/4 8.5 6.9
61/4 8.2 6.8
7 1/4 7.8 6.6
81/4 6.7 5.9
9 1/4 5.4 4.9
10/1/4 4.2 4.0
11 1/4 2.9 3.4
12 1/4 2.2 3.0
13 1/4 1.7 2.7
14 1/4 1.5 2.5
15 1/4 1.4 2.4
15 1/4+ 4.4 12.5
*The 3/8 year old vehicles are the current year
models in the base year or the strategy year, the
1 1/4 year old vehicles are prior year models,
etc.
**Based on State of California Air Resources Board
and Department of Motor Vehicle data.
- 5 -
-------
3) Vehicle VMT/yr rate, §£ of July:
Vehicle Age
LDV*
VMT/yr
Gasoline HDV**
VMT/yr
Diesel HDV**
VMT/yr
1
2
6
7
3/8
1/4
1/4
3 1/4
4 1/4
5 1/4
1/4
1/4
8 1/4
9 1/4
10 1/4
11 1/4
12 1/4
13 1/4
14 1/4
15 1/4
15 1/4+
20,000***
16,300****
13,500
10,500
9,700
8,200
7,200
6,770
6,350
5,920
5,490
5,070
4,640
4,640
4,640
4,640
4,640
28,000***
21,100****
17,950
17,950
13,960
13,960
11,000
11,000
8,420
8,420
4,270
4,270
4,270
4,270
4,270
4,270
4,270
128,
96,
81,
81,
63,
63,
50,
50,
38,
38,
19,
19,
19,
19,
19,
19,
19,
000***
ooo****
600
600
600
600
200
200
400
400
440
440
440
440
440
440
440
Motorcycle VMT/yr = 4000*****
*Based primarily on California State vehicle age vs.
mileage study.
**Based on U.S. Department of Commerce study "U.S.
Truck and Inventory Study - 1967."
***The accumulated mileage of a 3/8 yr. old vehicle is
determined by multiplying this number by 3/8.
****The accumulated mileage of a 1 1/4 yr. old vehicle
is determined by multiplying this number by 1 1/4.
*****Per EPA 1973 edition of "Compilation of Air
Pollution Emission Factors" (AP-42).
The accumulated mileage for vehicles older than 1 1/4
years, is determined by adding the accumulated mileage of
a 1 1/4 year-old vehicle (see **** above) to the VMT/yr
values found in the preceding table, for each vehicle age
after 1 1/4, up to and including the vehicle age of
interest. This is illustrated by the following example.
- 6 -
-------
Calculate accumulated mileage for 4 1/4 year old
LDV:
Mileage = 16,300 x 1 1/4 + 13,500 + 10,500 + 9,700
= 54,100
The accumulated mileage is used for determining vehicle
emission deterioration factors (see Section B.).
D. Growth Projections
1) EPA stationary emissions and mobile source popu-
lation growth projections (except motorcycles) are
as follows for the Los Angeles AQCR;
1970-75 growth factor = 1.065
1970-77 growth factor = 1.104
1972-75 growth factor = 1.039
1972-77 growth factor = 1.078
The above factors are California Air Implementation
Plan growth projections, based on a California
Department of Finance, Population Research Unit
Report, "Provisional Projections of California
Counties to 2000" dated September 15, 1971.
2) Motorcycle population growth projections for the
entire state;
Growth factors are determined using the ratio
of estimated statewide motorcycle population
projections in the California Department of
Motor Vehicle Report No. 31, March 1970. The
motorcycle growth rate factors derived from
Report No. 31 are as follows:
1970-75 growth factor =1.46
1970-80 growth factor =1.91
- 7 -
-------
E. Strategy Assumptions & Reduction Factors
1) State and local programs in effect or committed:
Population
Base (Vehicle Percentage Percentage
Model Years Population Population RHC CO NOx
or Sources) Base Afftd. Base Afftd. Reduction Reduction Reduction
Program Affected in '70 by ' 77 Factor Factor Factor
1
oo
1
NOx retrofit
control
NOx retrofit
control
Crankcase (PCV)
retrofit
control
Rule 66 incen-
tives &
1955-65 LDV
Exhaust
1966-70 LDV
Exhaust
1955-62 LDV
Crankcase
Surface
coating
0%
0%
93%
0%
67%
100%
approx .
100%
approx .
100%
0.25
0.12
1.00
0.50
0.09
0.10
0.00
0.00
0.23
0.48
0.00
0.00
restrictions
-------
2) Proposed EPA programs:
Population
Base (Vehicle Percentage
Model Years Population
Program
or Sources)
Affected
Percentage
Population
RHC
CO
NOx
Base Afftd. Base Afftd. Reduction Reduction Reduction
in '70
by '77
Factor
Factor
Factor
Annual Inspec.
Maintenance
Oxidizing
Catalyst
Retrofit
Oxidizing
Catalyst
Retrofit
Dry Cleaning
Solvent
Control
New motorcycle
emission
standards,
1976 & later
models
Degreasing
Solvent
Control
All LDV 0%
Exhaust
1971-74 LDV 0%
Exhaust
1966-70 LDV 0%
Exhaust
All RHC Dry 0%
Cleaning
Sources
All motor- 0%
cycle
Emissions**
All RHC 0%
Solvent
Sources
100%
75%
20%
100%
100%**
100%
0.15
0.58*
0.58*
0.90
0.28
0.90
0.12
0.50
0.50
0.00
0.28
0.00
0.00
0.00
0.00
0.00
0.00
0.00
*This factor accounts for a hydrocarbon reduction factor of 0.5 and a lowering of
the exhaust reactivity factor from 0.77 to 0.64. (See Section F.)
**While only new 1976 and later model years are affected, the entire population is
included here because the reduction factors are derived on the basis of the total
population.
-------
o
I
Program
Population
Base (Vehicle Percentage
Model Years Population
or Sources Base Afftd.
Affected in '70
Percentage
Population
Base Afftd.
by '77
RHC CO NOx
Reduction Reduction Reduction
Factor Factor Factor
Petroleum All Petroleum 0% 100%
Marketing Marketing approx .
Controls
Parking sur- All gasoline 0% 100%
charge and LDV and
review, bus & petrol,
carpool pri- marketing
ority treat- emissions
ment, & employ-
ees transit
incentives*
Total gasoline All gasoline 0% 100%
ban vapor and
combustion
emissions
Total Diesel All HDV 0% 100%
Fuel Ban Diesel
*See Appendix C, "California VMT Reduction Summary,
other VMT reduction measures.
The June 8, 1973 Federal Register discusses various
0.87 0.00 0.00
0.14 0.14 0.14
1.00 1.00 1.00
1.00 1.00 1.00
" for a discussion of these and
mobile Source Control programs
or tactics, and outlines the reduction factors associated with the tactics.
-------
F. Hydrocarbon Reactivity Factors
Per recent EPA guidelines, the following factors indicate
the weight fraction of hydrocarbons that are considered
to be reactive (i.e., do not contain unreactive hydrocar-
bons which are methane, ethane, propane, benzene, acety-
lene) :
Emission Source Weight Fraction of RHC
Gasoline LDV exhaust 0.77
Gasoline LDV exhaust after 0.64
catalyst treatment*
Gasoline HDV exhaust 0.79
Diesel HDV exhaust 0.99
2-stroke motorcycle exhaust 0.96
4-stroke motorcycle exhaust 0.86
Piston & turbine aircraft 0.90
exhaust
Gasoline vapor 0.93
*This factor is to be applied to the exhaust of all 1975
and later LDV models, and to those pre-1975 LDV models
that are to have retrofit catalyst devices installed.
Crankcase emissions are estimated to consist of equal
amounts of uncombusted gasoline vapor and combustion
vapor.
A rule 66 chemical definition of reactivity is used for
the remaining emission sources.
- 11 -
-------
APPENDIX B
Methodology for Determining the Base Year Oxidant Level
INTRODUCTION
This paper discusses a method for selecting the maxi-
mum values used in the calculation of emission reduction
requirements.
The methodology described in this paper is neither new
nor original. Dr. R. I. Larsen, Meteorology Laboratory,
NERC, Research Triangle Park, outlined such a technique in
1967 and has published numerous papers since that time ex-
plaining the use of his model in the establishment of
standards and in relating air quality measurements to such
standards (Reference 1, 2, and 3).
The rationale for selecting this method is outlined
and some of the advantages and shortcomings are covered.
A comparison of actual measured values with model calculations
is provided.
BACKGROUND
The development of a control strategy to achieve a
National Ambient Air Quality Standard is frequently based
on the premise that the concentration of a man-made
-------
- 2 -
pollutant In the ambient air Is linearly related to the rate at which
the pollutant is emitted in the atmosphere.
This assumption permits the use of a simple proportional (or
rollback) model to determine emission reduction requirements. Such
a model states that:
QOO) icurrent air quality) - (air quality standard) reQulred reductlon
U00; (current air quality) - (background) required^reauction
Current air quality is defined as the maximum measured concentration.
The development of the transportation control strategies did not
rely totally upon the rollback model. A non-linear relationship between
oxidant levels and hydrocarbon emissions developed by Schuck (See
Appendix B) was also employed. In some areas data was not available
for the verification of such a non-linear model and the simple proportional-
relationship had to be applied.
Regardless of which of these models was used, the selection of an
appropriate maximum concentration was a critical factor in determining
the emission reduction requirements.
-------
- 3 -
There are several methods that can be used to
determine the maximum value needed for these"roll-back"
calculations. Among such methods are:
a. Diffusion modeling
b. Selection of a maximum value from a base year
c. Choosing the highest value over a number of years
d. Determining a maximum value by statistical analysis
Diffusion modeling, where validated models can be ap-
plied, probably represents the best method for determining
both the concentration and the location of high pollutant
levels. Unfortunately, a model with the required accuracy
is not yet available for determining specific oxidant
concentrations.
The selection of a value from a base year, where the
year is usually selected as the year of the latest emission
inventory, has the advantage of being most closely related
to the emission data. It also provides a convenient base
for comparing data at different locations. However, high
concentrations of oxidant occur under certain, as yet not
fully quantified, meteorological conditions and different
sets of these conditions may apply to the production of high
levels at different locations. Since meteorological para-
meters do not necessarily follow an annual cycle, the adverse
-------
- 4 -
conditions producing high levels may not always occur every
year at any given location. The data indicate that maximum
levels at a particular monitoring station may vary from year
to year by as much as a factor of two. High values within a
given region do not always occur at the same site and maxi-
mum concentrations selected from all stations within a region
may also vary considerably, although not usually by as much
as they do at a single location.
Extreme values can occur either because of unusual
meteorological conditions or because sane abnormal periods
would not necessarily be expected to occur every year but
perhaps only once in 5 or 10 years. Thus, the selection of
such an extreme value could require overly stringent control
measures. Conversely, abnormally low values could also be
selected if the data record is short.
A statistical analysis of data collected over a period
of years tends to smooth out the variations due to the
meteorology and to local anomolies. Such an analysis can
also provide a prediction with a specified probability of
occurrence and the extreme or outlying values can be weighed.
This paper compares the results obtained by applying a
particular statistical method to the calculation of maximum
oxidant levels with the actual measured maximum concentrations
at selected stations from data collected over the past three
years.
-------
- 5 -
THE ANALYSIS
Selection of Technique:
The objective of the analysis was to find an oxidant level (con-
centration) that represented the highest level expected to be achieved
with a frequency of one hour per year. The rationale for this objective
is the National Ambient Air Quality Standards for oxidant: 160 ug/m3
(0.08 ppm) - maximum 1 hour concentration not to be exceeded more than
once per year.
Although there are a number of statistical methods that could be
applied, a technique described in the Office of Air Programs publica-
tion No. AP-89, "A Mathematical Model for Relating Air Quality Measure-
ments to Air Quality Standards" by R. I. Larsen, November, 1971, seemed
to best fit the objective. This model is based on the assumption that
the air quality data fit a log-normal distribution. There is some
disagreement about whether or not this is an appropriate assumption.
For example, Mitchiner & Brewer (5) have suggested the use of a
'double-exponential' distribution. This is a widely known extreme
value technique. Their analysis, however, was limited to data collected
in three summer months and used only the maximum daily hour data. A
report by Mosher, Fisher, and Brunelle (6) indicates peak oxidant con-
centrations of 0.50 ppm or greater have occurred in Los Angeles County
in all months of the year except January and February. The selection of
only certain months could, therefore, tend to bias the results. Addi-
tionally, extreme value techniques seem most applicable to the selection
of an absolute maximum concentration and not necessarily to the
-------
- 6 -
concentration expected to occur once per year. However, a comparison
of the values calculated by the Mtchiner-Brewer method indicate that
they do not differ greatly from Larsen's method, at least at the one
station covered in their analysis, even though a different data set was
used.
Larsen (7) analyzed all oxidant data for all California stations
for the period 1963-1967 and presented the cumulative frequency distri-
butions and a calculated maximum concentration for each station. The
tables in that publication were used in conjunction with later available
measured data to determine the location (or areas) of the highest con-
centrations. Stations within those areas were then selected for further
analysis. An attempt was made to obtain a three-year period of record
for each station. It was felt that the period should be comparable to
the latest emission inventory data available (in most cases this was 1970
data) and also should contain a sufficiently long period to help over-
come the problem of meteorological variability. A period of 5 to 10
years would have been desirable, but because of the changing patterns of
emissions and changing vehicular emission factors, it was felt that a
period longer than three years would tend to introduce more emission
variability than the meteorological variability that would be factored
out. Data for 1972 were not available so the period January, 1969, through
December, 1971, was selected. Unfortunately, there were many gaps in the
record and data was not available for some of the desired stations.
-------
- 7 -
Fourteen stations were finally selected for analysis and cum-
mulative frequency distributions for the selected stations for the
three-year period were then obtained. The data were analyzed according
to Larsen (4). A sample of the frequency distribution used is shown in
Figure 1.
CALCULATION OF MAXIMUM CONCENTRATIONS
The frequency distribution as given in Figure 1 is plotted on a
logarithmic probability graph as indicated in Figure 2. If the data
were perfectly log-normally distributed, all points on the graph would
be on a straight line. As can be seen in Figure 2, this is not the
case. However, the points in the frequency ranges from 10% to .01$ do
appear to closely approximate a straight line. Since these are the
frequencies of most concern when considering very high values, only
those points are considered. To find the value that would be expected
once a year, Larsen (4) suggested using the .01 and the .10 frequency
points and extrapolating the line connecting these points to the
desired once per year frequency point. This was done for each location
for which frequency distributions were available. The extrapolation can
be done either graphically or mathematically. The mathematical method
is as follows:
-------
- 8 -
The desired frequency using this log-normal distribution is
obtained from
f = r-0.4 (100$?)
n
where: r = the rank of the desired concentration if all
the concentrations were ordered from one
through the number of possible samples within
a selected time period
f = the frequency of occurrence in percent
n = total number of samples
FOR EXAMPLE: To find the frequency corresponding to the highest
one-hour average in a year, all of the 8760 one-hour averages in a year
would be listed in order from 1 (the highest) to 8?60 (the lowest). The
rank order, r, then is equal to 1, n, or the total number of samples, is
8760, and
f = 1-0.4 (100?) = 0.00685$?
8760
Next, the extrapolation of the data to this desired frequency, using
the two known concentration vs . frequency points , is as follows :
The equation of a straight line passing through two known points
and
this can be rearranged so that
' y = Yl + yg- yl
-------
- 9 -
In this case the x and y without subscripts are the intercepts of
the unknown point on this line.
Where we are using the log-normal distribution, the y intercepts are
logarithms and the x intercepts are in terms of standard deviations from
the median. In a normal distribution each frequency can be located as a
distance (standard deviations) from the center of the profile (median).
If the y intercepts are logarithms, then the equation for the
straight line becomes:
In y = In y
(x2-x-L)
The concentration at an unknown point 'x,yr is then equal to the
anti- logarithm of
In y2
(In y, + _ 3T (x-x-]_) )
(x2-x1)
or to put it in another form: -^ y
concentration at y = exp [ In y + ^1 (x-x-j_) ]
(x2-x1)
where 'exp' indicates that 'e', the base of natural logarithms, is
raised to the power in the brackets, 'e1 is approximately equal to
2.71828.
Following Larsen's suggestion (8), the two known points at the .01
and the .10 percentile levels are used to define the straight line we
wish to extend. Prom a statistical table, such as is given in
Reference 3 on Page 30 , the x intercepts at these percentile points
-------
- 10 -
can be determined. In the case of a log-normal profile, the .01
percentlle point is 3.72 standard deviations from the median; the
.10 percentile point is 3.09 standard deviations; and the unknown point
at .00685% is 3.81 standard deviations from the median. The y intercepts
are the concentrations at each of these percentile points. These x and y
values are then substituted into the above straight line equation and
the unknown concentration at the .00685% frequency is determined.
To illustrate the procedure, the data from Figure 1 have been
rep lotted on Figure 3, and the points that are used below have been
labeled.
frequency concentration standard deviations
f(%) y(ppm) _ x _
.00685 to be determined(y) 3.8l (x)
.01 .27 (y-,) 3.72 (x, )
.10 .23 (yp 3.09
Substituting these values into the straight line equation:
-
concentration at y = exp [ In. 27 + m (.27} (3.81-3.72) ]
(3.09-3.72)
= exp [ -1.30933 + (-0. 1603*0 (0.09) ]
(-0.63)
= exp [ -1.286424]
concentration at y = 0.28 ppm
From the example, a concentration of 0.28 ppm would then be the
highest concentration expected to be reached (or exceeded) once each
year.
These maximum concentrations were calculated for each of the
selected stations within each Air Quality Control Region. The results
are listed in Table 1.
-------
- 11 -
TABLE I. Hourly average concentrations for selected frequencies of
occurence.
LOCATION
South Coast AQCR
Riverside
Azusa
Pasadena
San Diego AQCR
San Diego (8th & E)
El Cajon
Sacramento Valley AQCR
Creekside
Chico
San Joaquin Valley AQCR
Fresno (So. Cedar)
San Francisco Bay AQCR
Livermore
San Leandro
Fremont
Percent of time given concentration
equaled or exceeded
0.10%
0.34
0.42
0.39
0.16
0.27
0.18
0.14
0.0136
0.56
0.51
0.51
0.23
0.30
0.24
0.15
0. 00685$ ( Annual Maximum)
0.60
0.52
0.53
0.24
0.30
0.25
0.15
0.20
0.24
0.19
0.22
0.25
0.32
0.27
0.27
0.26
0.33
0.28
0.28
Data used in this Table were hourly averages for the period of
January, 1969, to December, 1971.
-------
- 12 -
It should be noted that these calculated concentrations are not
necessarily the highest values to be expected. It is quite possible
that this value could be nearly twice as high on an unusually "smoggy"
day. Based on this analysis, however, such very "smoggy" days would
not normally occur every year.
COMPARISON WITH MEASURED MAXIMA
The calculated maximum values were compared with the actual maxi-
mum values that have been reported within each of the Air Quality Control
Regions since 1969. These values are shown in Table 2. In all cases
the calculated maximum concentration is within .03 ppm of the actual
measured maximum, even though an additional year of measured data was
considered and the high value for the region may have been reported at
a station other than one included in the calculations.
TABLE 2. Comparison of measured and calculated highest hour average
oxidant concentrations calculated.
AQCR
South Coast
San Joaquin
San Diego
S.P. Bay Are
Sacramento
Calculate
Maximum
.60
.26
.30
.33
.25
d Measured
Station Maximum
Riverside
Rr'esno
El Cajon
Livermore
Creekside
.62
.24
.32
.36
.28
Station
Riverside
Modesto
Escondido
San Leandro
Creekside
Year
1970
1972
1972
1971
1972
-------
- 13 -
The number of occurrences of concentrations in excess of the cal-
culated maximum within each Air Quality Control Region was also tabulated.
For comparison, the daily maximum hourly averages from 1969-1972 were
used. The calculated maximum was equaled or exceeded three times in the
San Francisco Bay Area, once in 1969 and at two separate locations on
the same day in 1971. In the South Coast Basin the calculated concentra-
tion was exceeded once. In the San Diego Area twice, once each in 1971
and in 1972 in Sacramento once and once in the San Joaquin Air Quality
Control Region.
Again, it should be noted that the calculated value represents a
level that is expected to be reached or exceeded once per year and that
the analysis does not attempt to predict the highest possible concentra-
tion. Thus, the occurrence of a concentration greater than the predicted
value tends to verify the procedure if no other concentration measured
during the year was equal to or greater than the calculated maximum.
EVALUATION OF METHOD
The fact that the calculated values are close to the actual measured
concentrations and that the values have been reached or exceeded only
once in a given year, would tend to indicate that reliability of Larsen's
technique. There are, however, some obvious shortcomings to the analysis
presented here. A full three-year period of record was not available from
all of the air monitoring stations within each basin, nor from each of
-------
the stations listed in Table 1. The shorter the period of record that
is available, the less reliable are the calculated values. To improve
the reliability, additional data should be analyzed and a larger sanple
from each Air Quality Control Region should be selected.
Also, it was assumed that the stations selected represented the highest
concentrations within the given Air Quality Control Region. This is not
necessarily a valid assumption. Although only limited data is available,
newly established monitoring sites appear to be recording higher values
than some of the listed stations. For example, data from Escondido was
used to develop the strategy in the San Diego Air Quality Control Region.
The station was established in mid-1972 and the .32 ppm oxidant measured
there represents the highest concentration within the San Diego metro-
politan area in recent years. Agencies are usually continually expanding
their networks to include new areas of high concentrations, and additional
analyses should be performed as new data become available.
The calculations were based on the data measured during the years
1969-1971. They reflect only the emissions during that period of time.
Assuming no changes in emission patterns or emission controls at the
source, these values could be used to predict future air quality. However,
none of the areas considered are static with respect to growth, or to
the numbers and ages of motor vehicles in operation, or even with respect
to the numbers of and outputs from stationary sources. Some care should
be exercised in attempting to relate the concentrations to emissions in
areas of rapid growth.
-------
-15-
The oxidant data do not exactly fit a long-normal distribution and
the degree of fit varies at different locations. Thus, use of this
method may result in more reliable results in some areas than in others.
Also, the calculated maximum is quite sensitive to the selection of the
percentile points used in the calculations especially where the log-normal
fit is poor. Larsen (4) has suggested the use of the concentrations at
the .01 and .1 percent frequencies as being most representative of the
distribution of the higher concentrations. In some instances it appears
that the point at the .01 percentile fits the overall log-normal
distribution least well. The problem is particularly evident when a
short period of record is used. In most of the data examined, use of
the .1 and the 1 or the 10 percentile points would result in higher
maximum, levels than when the .01 percentile is included. This would
indicate that for some reason, probably meteorological, the maximum
possible values are not achieved. In other cases the .01 percentile
value seems too high. A study of the individual days could perhaps
provide an answer to the reasons why some of the high values seem out
of line.
The calculation of the maximum value is quite simple, but it does
require the preparation of cumulative frequency distributions. These
distributions are best processed by computer because of the large amounts
of data required. Once they are available, several other analyses can be
performed (see Larsen, 3). Additionally, a comparison of the different
yearly and three-year distributions suggests a possible method for trend
analysis.
-------
- 16 -
SUMMARY
Air quality data for a number of California air monitoring stations
were reviewed and analyzed according to a method suggested by Larsen.
The objective of the analysis was to determine a maximum oxidant concen-
tration for certain Mr Quality Control Regions that could be related to
the available emission data and used to determine the emission reductions
needed to achieve the National Ambient Air Quality Standards.
Because of the variability of concentrations from year to year, at
least a three-year period of record would appear to be required for
analysis. This limits the selection of maximum concentrations to these
stations where data have been collected over that long a period and
could eliminate areas where higher concentrations are possible.
Although values obtained in this analysis compare favorably with
measured conentrations, other statistical approaches may provide equally
meaningful solutions and should be compared with this method. The method
outlined in this paper, however, is relatively simple and well documented
and is applicable to all pollutants. The use of some statistical approach
is certainly less arbitrary than the selection of one particular measured
concentration.
ACKNOWLEDGEMENTS
The author is indebted to Dr. R. I. Larsen of the Meteorology
Laboratory for his assistance and Mr. Don Worley of the Data Systems
Division for providing the necessary frequency distributions.
-------
- 17 -
REFERENCES
(1) LARSEN, R. I. (1967) "Determining Reduced Emission Goals Needed to
Achieve Air Quality Goals - A Hypothetical Case," YAPCA 17, pp. 823-829
(2) LARSEN, R. I. (1969) "A New Mathematical Model of Air Pollutant
Concentration, Averaging Time and Frequency," YAPCA 19, pp. 2^-30
(3) LARSEN, R. I. (197D "A Mathematical Model for Relating Air Quality
Measurements to Air Quality Standards," Publication AP-89, U. S.
Environmental Protection Agency, Research Triangle Park, NC 27711
(4) LARSEN, R. I. (1973) "An Air Quality Data Analysis System for Interrelating
Effects Standards, and Needed Source Reductions," presented at WMO-WHO
Technical Conference on Observation and Measurement of Atmosphere Pollution,
Helsinki, Finland, July 30 to August 4, 1973
(5) MITCHINER, T. L., and BREWER, J. W., "A Comment on the Method Used by
EPA to Calculate Required Reductions in Emissions," University of California
at Davis, unpublished
(6) MOSHER, T. C., FISHER, E. L., and BRUNELLE, M. F. (1972) "Ozone Alerts in
Los Angeles County," Air Pollution Control District, County of Los Angeles, CA
(7) LARSEN, R. I. (1971) "Air Pollution Concentrations as a Function of
Averaging Time and Frequency
(8) LARSEN, R. I. (personal communication, 1973)
-------
CONCENTRAT
OX10ANT
ION V$ -AVKAGlNG T1M6 AND FREQUENCY FOR
(PPMI: LOS ANGELES, S. SAN PEDRO ST.
JANUARY
It 1966TO
Q
PERCENT
WERAGING
TIME
5 MIN
lO
i5
40
1 HOUR
3
3 6-9 AM
8
i?.
1 DAY
2
ft
7
i.4
1 MONTH
2
?
6
I YEAR
2
3-
6
MEAN
000.
000.
000.
000.
0.03
0.03
0.01
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
000.
MAX
000.
000.
000.
000.
0.33
0.27
0.05
0.21
0.12
0.10
0.09
0.07
0.06
0.06
0.05
0.05
0.04
0.03
0.03
0.03
"O."03
000.
MIN
000.
000.
000.
000.
0.01
0.01
0.01
0.01
0.01
0.01
0.0.1
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.03
0. 03
000.
0
.001
PERCENT
000.
000.
000.
000.
069.
072.
073.
073.
073.
074.
074.
075.
075.
075.
075."
075.
"075."
075.
075."
050.
"100."
000.
0.00
0.00
0.00
0.00
0.33
0.27
0.05
0.21
0.12
0.10
0.09
0.07
0.06
0.06
0.05
0.05
0.04
0.03
0.03"
0.03
0.03
0.00
DEC.
STATI
ON 001
31, 1971
OF TIME CONCENTRATION IS
EQUALED OR EXCEEDED
99.9
0.01
0.00
0.00
0.00
0.00
0.27
0.27
0.05
0.21
0.12
0.10
0.09
0.07
0.06
0.06
0.05
0.05
0.04
0.03
0.03
0.03
0.03
0.00
0.1
0.00
0.00
0.00
0.00
0.23
0.21
0.04
0.18
0.10
O.C9
0.09
0.07
0.06
0.06
"0 . 05
0.05
0.04
0.03
"0 i 03
0.03
'0.03"
0.00
1
0.00
0.00
0.00
0.00
0.16
0.15
0.03
0.13
0.08
0.07
0.07
0.06
0.06
0.06
0.05
0.05
0.04
0.03
0.03
0.03
0.03
0.00
10
0.00
0.00
0.00
0.00
0.07
0.07
0.02
0.07
0.05
0.05
0.05
0.05
"0.04
0.04
"0.04"
0.04
0.04
0.03
0.03
0.03
0.03
0.00
20
0.00
0.00
0.00
0.00
0.04
0.04
0.01
0.04
0.04
0.04
0.04
0.04
0.04
0.04
"0.04
0.04
0.04
0.03
0.03
0.03
0.03
0.00
30
0.00
0.00
0.00
0.00
0.03
0.03
0.01
0.03
0.03
0.04
0.04
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.00
40
"0.00
0.00
0.00
0.00
0.02
0.02
0.01
0.02
0.03
0.03
0.03
0.03
0.03
0.03
0.03-
0.03
0.03
0.03
0.03
0.03
'0.03"
0.00
50
0.00
0.00
0.00
0.00
0.01
0.01
0.01
0.02
0.02
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.00
60
0.00
0.00
0.00
0.00
0.01
0.01
0.01
0.01
'0.02
0.02
0.02
0.02
0.02
0.02
0.03"
0.03
0.02
0.03
0.03
0.03
0.03
0.00
70
0.00'
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.03
0,03
0.00
80
0.00
0.00
"0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.03
0.03
0.00
90
0.00'
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.01
0.02
0.02
0.02
0.03
0.03
0.00
99
0.00
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.03
0.03
0.00
99.999
99.99
0.00
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01"
0.01
0.02
0.02
0.02
0.03
0.03
0.00
0.00
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.03
0.03
0.00
- ---
99.999
0.00
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.03
0.03
0.00
- - - -
0.00
0.00
0.00
0.00
0.01
0.01
0.01
0.01
0.01
0.01
0.01
C.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.03
0.03
0.00
Figure 1. Cumulative frequency distribution of oxidant concentrations for Los Angeles, Downtown station.
-------
-V
.v
.00
•05
0.0=) 0.1 as. c.5
'j.c IV.G ?.:.;:
iiM^irifrwA
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION IX
-------
1.0
5-81 V2. W
^?fi
0.8
0.7
O.fc
£
O.I
'' '
I N ">• (?•-
»• -'
1.78
f-, •
-'; i i N f I • .*- A ; i1-.
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION IX
-------
Appendix C
CALIFORNIA VMT REDUCTION SUMMARY:
Presentation of Analytic Basis
for
VMT Reduction Estimates
-------
SUMMARY
Purpose
It is the purpose of this paper to summarize recent
studies relating various transportation control measures such
as car pooling, bus lanes, gasoline sales limitations, etc.,
to VMT reductions. In particular, the paper includes those
measures proposed for California Transportation Control Plan
(TCP) .
Concentrating on the measures proposed by EPA and local
AQCR task forces, the paper shows the range of VMT reductions
that can be expected from the TCP measures individually and
combined in a complementary strategy for each AQCR. Deter-
minations of the effective range of VMT reductions achieveable
are based on application of various transportation modal split
models developed for the purpose of predicting commuter trans-
portation patterns.
Most of the effects of measures promulgated by EPA are
predictable within a range of certainty. Results of pilot
studies and public reaction attitude surveys have been used
as inputs to the data base. Measures that can not be assessed
quantitively at this time or that have delayed effects and are
unable to effect air quality fast enough for the Clean Air Act
standards are not included in EPA promulgation. However, many
of these measures have future short and long-term potential.
EPA encourages local proposal and implementation of these
measures.
Although the socio-economic impact of VMT reduction
measures is important, the scope of this analysis is limited
to the technical effects. Many of the studies done for EPA
have addressed the socio-economic questions.
I. INTRODUCTION
THE ROLE OF VMT MEASURES
In the majority of air quality control regions requiring
additional controls, the combined impact of stricter controls
on stationary sources and the establishment of an inspection
-------
-2-
and maintenance system will not provide emission reductions
adequate to achieve the air quality standards by 1975. Conse-
quently, EPA has promulgated a variety of measures to reduce
vehicle miles traveled in these regions. In several urban
areas a shift from present reliance on automobiles occupied by
one or two persons to a greater reliance on other forms of
transit is essential to the achievement of the air quality
standards. Significant reductions in vehicle miles traveled
can also be accomplished within a limited time span.
The States have had practically no experience with trans-
portation control measures as a means of dealing with air quality
problems and the success of particular VMT reduction measures
is difficult to predict. However, recent developments involving
bus lanes, mass transit improvements, carpool programs, bikeways,
and other innovations indicate that many VMT reduction measures
are available and feasible. Furthermore, attitude surveys show
that the public in many of our urban areas recognizes the need
to place less emphasis on the automobile for urban mobility and
is already encouraging the implementation of steps to develop
alternative forms of transit.
Some of the regulations being promulgated will have signif-
icant effects on the future development of urban transportation
in the major cities of this country. A clear implication of
these air plans is that future augmentation of mass transit must
focus not only on the center city streets but also on urban/sub-
urban routes. It is expected that the regulations will lead
not only to substantial reductions in air pollution, noise,
congestion, and energy comsumption, but to the development of
more mass/rapid transit to serve the growing urban and suburban
regions of the nation. The need, desirability, and feasibility
of reducing urban auto use are no longer issues. The problem
is determining the degree to which VMT reductions can be
reasonably implemented within the limited time frames.
The amount of VMT reduction that can be considered "rea-
sonably available" varies greatly according to a city's indi-
vidual characteristics and the ability of other modes of trans-
portation to absorb the demand that would be created by a
significant VMT reduction. A measure cannot be considered
"reasonably available" if putting it into effect would cause
severe economic and social disruption. Although some reduc-
tion in personal travel could certainly be absorbed without
disruption, to achieve a significant VMT reduction, the bulk
of the travel displaced from single-passenger automobiles
must be absorbed by such other modes of transportation as
carpools, walking, bicycling, or public transit.
-------
-3-
The significant expansion of public transit facilities
that can be accomplished by 1975 depends on the upgrading and
expansion of bus service. Much can be done in this regard.
Scheduling and service can be improved. Individual lanes of
freeways and other major roads can be set aside for the exclu-
sive use of buses. Significant numbers of new buses can be
purchasds and put into service by 1975; according to the
Department of Transportation figures, 2,500 transit buses were
sold in this country in 1972, but there is considerable poten-
tial for expansion of the transit industry's production by
two or three-fold. Foreign sources of supply could provide
additional resources.
The Environmental Protection Agency is working with the
Department of Transportation to assure increased Federal sup-
port for short-term augmentation of mass transit capacity and
appropriate modifications of highway facilities to permit
increased utilization of mass transit.
In addition to public transit, part of the transportation
demand created by VMT reductions can be absorbed by carpools.
Private automobiles, which are designed to carry four to six
persons, carry an average of 1.1 to 1.4 persons per trip for
work trips in major urban areas, and thus represent the largest
unused pool of transportation capacity presently available.
The measures mentioned above are primarily concerned with
providing an alternative to low-occupancy use of private auto-
mobiles. Although measures such as buying more buses and
improving bus service, providing for carpool programs, building
bicycle paths, and (possibly in the long run) building new
rapid transit systems increase the availability and attrac-
tiveness of alternative transit forms, VMT reductions will not
necessarily be achieved unless disincentive restrictions are
placed on the use of automobiles.
The applicability of both measures—incentives such as bus
lanes that increase the attractiveness of alternative transit
forms and disincentives such as parking limitations that dis-
courage the low-occupancy use of private automobiles—varies
according the the conditions in the individual urban area.
For example, bus lanes are a more appropriate strategy in
Washington, D.C. than certain other areas. Similarly,
parking restrictions are more applicable to a major center
like Boston than to a small city with few transit alterna-
tives like Fairbanks, Alaska.
After consideration of the already available transit
alternatives, the city's local conditions, and the applica-
bility of various incentive and disincentive measures, the
-------
-4-
EPA has determined that varying degrees of VMT reduction
are feasible in particular areas. The Agency believes a 3
to 10 percent VMT reduction can be achieved in some of the
regions by 1975. Since the Clean Air Act specified that
all reasonable available measures be instituted before any
time extension is granted, the Administrator is taking into
consideration all VMT-related measures presently being
implemented by a municipality and augmenting those measures
with methods that are available, applicable, and adoptable
in the individual area by 1975.
Through studies and the public hearing process, the
Agency has also determined that it may be unrealistic to
expect reductions in auto use greater than 10 to 20 percent
by 1977. Generally, reductions beyond 10 to 20 percent
would require a special and, in most cases, unreasonable
effort unless driving is to be cut without a corresponding
increase in mass transit. Achievement of even the levels
provided for in these plans will require a strong commitment
by local areas to implement strict disincentive programs,
improve mass transit, and make carpooling or other programs
work.
EPA has promulgated a number of measures designed both
to increase the attractiveness of alternative forms of transit
and to discourage the low-occupancy use of automobiles. The
measures include: regulatory fees for mass transit augmenta-
tion, bus/carpool lanes, carpool matching systems, and carpool
programs stressing preferential treatment. Local task forces
have proposed additional measures, applicable to their par-
ticular regions, that will be implemented along with EPA
strategies. These measures include: traffic flow improve-
ments, ramp metering, fringe parking for park and ride, dial-
a-ride service, bicycle lanes and facilities, reduced transit
fares, four day work weeks, and taxation and pricing measures.
State, Local, and Federal Implementation of_ Control Measures
In order to preserve the intent of the Clean Air Act
that pollution problems be dealt with primarily at the local
level, the Agency is requiring that State and local govern-
ments take action wherever possible and will involve the
Federal Government only in the direct implementation of some
programs. State and/or locally enforced, Federal promulgated
requirements are: retrofit programs; parking supply and sur-
charge; bus and carpool lanes; inspection and maintenance;
and stationary source controls. Federally operated programs
will be: motorcycle controls, gasoline limitation, and a
bus/carpool incentive regulation directed at major employers.
-------
-5-
DESCRIPTION OF EPA PROMULGATED VMT REDUCTION MEASURES
Bus Lanes
Bus priority treatment consists of allocating highway
facilities preferentially to buses for the purpose of improving
the quality of bus service. Methods of effecting bus priority
treatment in the transportation plans include reserved lanes
for buses (and/or carpools), preferential access for buses at
metered freeway ramps, and certain traffic engineering improve-
ments. The forms of bus lanes set forth in either the plans
proposed or approved by the States or promulgated by EPA in-
clude normal bus flow lanes, and contra-flow lanes. In Cali-
fornia, the Department of Transportation suggested that only
certain freeways or major roads be dedicated to the bus lane
concept and EPA agreement with this suggestion is reflected in
these promulgations. The method of selecting the lanes has
been changed from one based on the number of lanes in the road
to one looking to the establishment of a coherent network of
such lanes along transportation corridors. In some regions,
pilot programs will be conducted to discover the best way to
implement a full-scale program. In some cases, measures such
as the conversion of entire streets to bus and carpool use may
prove preferable to limited lanes.
The use of bus (and/or carpool) lanes has been observed
to increase mass transit freeway speeds by a factor of two or
more. Through the elimination of congestion problems, bus
service dependability is increased as late arrivals are sig-
nificantly reduced. Furthermore, bus ridership will increase,
and the fares may eventually be reduced. Because of these
factors, the regulations set forth for bus lanes are expected
to be a positive inducement to increased bus patronage. The
timetables for implementation of bus lanes will vary according
to regional situations.
Carpool Systems
Experience to date with carpool programs suggests that
policies to encourage carpooling might double auto occupancy
rates for downtown peak period work trips. If a 10 to 50
percent increase in auto occupancy is adopted as a realistic
range of possible effects, the net effect of carpool policies
on total urban area auto use might be a 5 to 10 percent re-
duction .
EPA is promulgating measures that provide computerized
carpool matching programs and preferential carpool treatment
-------
-6-
programs. The matching program provides for the formation
of carpools and the preferential treatment programs provide
incentives such as free parking to encourage carpools. Under
the measures included in some plans, disincentives such as
parking space reduction or paid, rather than free parking,
are included to discourage single occupancy on commuter trips.
In all Regions EPA is requiring the establishment of
carpool matching systems to enable persons with similiar daily
travel patterns to make contact with each other and arrange
carpools. In some regions, pilot programs will be established
prior to establishment of the system throughout the region.
Such a measure is necessary if the restraints on individual
vehicle use contained in this plan are to have the desired
effect of reducing VMT.
The EPA Regional Office in San Francisco has contacted
various Federal agencies in order to facilitate the implemen-
tation of the pilot programs called for in the regulation.
The Regional Office has experienced initial success in its
first contacts, and this effort is continuing. A detailed
guide for the operation of a bus/carpool matching program,
along with a discussion of a number of successful programs
in operation in many areas of the country is found in a U. S.
Department of Transportation Federal Highway Administration
Publication "Carpool and Buspool Matching Guide (Second
Edition)", May 1973. This report discusses the considera-
tions involved in a successful program such as public infor-
mation, incentives, data processing, and a continuing updating
of the service, and is an excellent guide and reference for
conducting such a program.
The EPA believes that this approach to reducing vehicle
miles traveled is an excellent short-term strategy. It
involves a minimum investment and deserves the active promo-
tion and support of government and industry.
Employer Provisions for Mass Transit Priority Incentives
As was pointed out in the public hearings by some of
those testifying, employer-paid privileges for employees tend
to encourage employees to drive to their place of employment
rather than use carpools or mass transit. The promulgated
regulation therefore, provides for employer-paid mass transit
fares and special parking privileges to those who travel in a
carpool. It also requires that individuals who drive may not
be provided with free unlimited parking, but must pay the
prevailing surrounding parking rate.
-------
-7-
This regulation will be implemented in stages, the first
stage applicable to employers of 700 or more employees, and
the second to employers of 70 or more.
The purpose of this regulation is to effect sizeable
reductions in VMT caused by commuting, with appears to be the
mode of travel most easily diverted to mass transit and car-
pools.
Control of Existing Parking Spaces; Surcharge on Parking
The proposal that spaces in public parking facilities
be reduced by 20 percent drew almost universal adverse comment
during the rulemaking proceedings. At the same time, the use
of regulatory fees to discourage pollution-causing activities
was widely supported. In particular the use of fees to control
parking was mentioned.
EPA also believes that the use of such parking fees has
much to recommend as a matter of policy. Accordingly, EPA is
not promulgating a reduction in publicly owned parking spaces
and is instead promulgating a regulatory fee to increase the
price of parking in, and so discourage traffic to, selected
trip attraction centers in the three most heavily polluted
AQCR's. The regulation's coverage will be increased in
three phases. At least 50 percent of the revenues will be
used for mass transit.
Several of the plans call for the imposition of regulatory
fees on parking. In earlier Notices of Proposed Rulemaking,
the Agency expressed some doubt about its authority to impose
such fees. That legal question has been extensively reexamined,
and EPA has now concluded that such a step is authorized by the
statute. The transportation control measures promulgated by
EPA will require a significant change in the driving habits of
the American people. The use of fees can help to bring that
change about with a minimum of social disruption of the wide
latitude they leave to individual choice. Those whose needs or
preferences are strongly in favor of using the single-passenger
automobile may continue to do so, although at a somewhat higher
cost; those who can easily adapt to the use of other modes of
transportation have a financial incentive to do so. Many public
comments supported the adoption of such fees. In addition, the
enforcement of such fees will be less difficult than some other
measures. Finally, such fees will be used to support mass
transit. Expansion of mass transit is essential if the disin-
centives to automobile use imposed by transportation control
-------
-8-
plans are to have the desired effect. Such a use of the proceeds
will also greatly mitigate the potentially regressive nature of
such fees.
In requiring the States and EPA to impose transportation
controls where they are needed to meet air quality standards,
the Congress imposed a regulatory task whose difficulty and
complexity are virtually unparalleled. The legislative history
shows that Congress fully recognized the magnitude of the
problem. At the same time, the statute's description of the
exact types of measures that may be imposed is extremely broad
and general. In the face of this broad language, the Adminis-
trator concluded that the Congress intended him to impose the
method of control that he determined was best able to achieve
the purposes of the statute.
Parking Management Program
The proposal for review of new commercial parking facil-
ities has been modified, from an earlier proposal, to allow
a wider range of variables to be considered. In essence, the
regulation promulgated today would forbid the construction of
any facility that could be expected to lead to a VMT increase
unless either (1) the application retired from service or
caused the retirement from service of an equal number of
spaces elsewhere in the AQCR, or (2) the applicant could show
in a separate hearing devoted only to that question that the
impact of the proposed facility on VMT, and thus air quality,
would be insignificant.
The promulgated regulations will require that the appro-
priate local government submit to the Administrator a plan
outlining the locally planned expansion of parking facilities
for the next five years. If a submittal is not made that
shows to the satisfaction of the EPA that such planned parking
expansion does not conflict with the California State Imple-
mentation Plan, the EPA will review each proposed new parking
facility individually. Such review by either the State or
EPA will be consistent with the previously discussed complex
sources regulations to be promulgated shortly.
Motorcycle Controls
In the July 16, 1973, proposal, regulations were included
that would have restricted 2-stroke motorcycle operations during
the "smog season" in California. This action was taken due to
the very high pollution potential of the 2-stroke motorcycle.
The average 2-stoke motorcycle emits approximately 31 times as
much exhaust hydrocarbons per mile as a new California 1975
-------
-9-
automobile will emit. Consequently, prevention of increases
in the number of motorcycles was proposed to prevent counter-
productive shifts from automobiles to motorcycle as a result
of other elements of the control strategy. The Agency has
evaluated the feasibility of establishing emission standards
for new motorcycles and is currently evaluating the availability
of motorcycle emission control technology for existing motor-
cycles to reduce emissions.
Based upon testimony presented by motorcycle manufacturers,
testimony presented by motorcycle trade associations, and an
independent analysis by the Environmental Protection Agency, it
appears that significant reductions in the emissions from new
motorcycles can be achieved.
Accordingly, the EPA is no longer requiring an uncondi-
tional ban on motorcycle operations. Instead, the ban regula-
tion has been rewritten to provide that it will not go into
effect in the event that nationally applicable Federal regula-
tions are promulgated that require at least a 50 percent reduc-
tion of 2-stroke motorcycle emissions by 1976 and conformity
with the 1976 automobile standards by 1979. Comparable emission
reductions will be required of 4-stroke motorcycles.
Vehicle Free Zones - Not Promulgated
Traffic free zones are primarily promulgated to control
local carbon monoxide problems. The zones are necessarily
restricted in size (approximately ten blocks or less) in order
to provide foot access. Consequently, the zones can be put
into effect by 1975 since no additional transit facilities
are required. Although increasing the size of the vehicle
free zone tends to increase the potential air quality im-
provements , such action also increases the problems of access,
circulation, and peripheral congestion and pollution.
Selective Vehicle Use Prohibitions - Not Promulgated
In several regions, EPA proposed a regulation under wliich
the vehicle population would have been divided into five cate-
gories . Each category of vehicles would have been required to
display prominently a tag of distinctive color; on one day of
each working week vehicles marked with one such color would
have been forbidden to operate.
Testimony at all the public hearings indicated that
measures of this type so far proposed would be unenforceable
because of their severeness and arbitrary nature. In addition,
the number of additional enforcement personnel necessary to
-------
-10-
implement such a program would then have been so great as to
preclude the reasonable availability of this measure. Were
they to be implemented, many very workable methods of evading
the requirements would doubtless be devised.
Gasoline Supply Limitations
The proposed transportation controls included measures
to limit the gasoline supply in certain areas in order to
reduce vehicle miles traveled. The measure included two types
of regulations: (1) a gasoline supply lid that would become
effective in 1974 to limit the quantity of gasoline sold to
fiscal 1973 levels; and (2) a regulation to be implemented on
May 31, 1977, to reduce an area's gasoline supply, and thus
VMT, to the extent necessary to achieve the standards.
The gasoline supply requirement has been dropped as a
primary measure. The Act requires that all "reasonably avail-
able" measures must be implemented by May 31, 1975, before
granting an extension. Based upon the comments received at
the public hearings on this measure and the Agency's evalua-
tion of the feasibility of implementing and administering
successful gasoline supply limitations, the Administrator has
determined that a gasoline supply lid cannot be considered
"reasonably available." The possibilities of evasion, the
likelihood of noncompliance, and the difficulty of enforce-
ment are too great to make this measure practicable at this
time.
The gasoline supply reduction regulation to be implemented
on May 31, 1977, however, has been retained in several plans.
As was noted above, the Clean Air Act required air quality
standards to be achieved by 1977 without regard to cost or
social disorganization that may result as a by-product of
achievement. If gasoline supply limitations are needed to
achieve the standard, the "reasonableness" criteria is not a
determining factor. Accordingly, the Administrator was obli-
gated to use gasoline supply limitations as a final resort
measure in certain areas with severe air pollution problems.
Most of these areas required reductions in vehicle miles,
traveled far in excess of 10 to 20 percent. In some regions,
however, the required VMT reductions may well be accomplished
through the specified VMT reduction measures. Gasoline supply
limitations were required in these areas only to assure the
attainment of the standards by 1977. If a review of air
quality data and VMT reduction monitoring information prior
to 1977 indicates that the gasoline reduction measure is not
required, supply limitations will not be implemented.
-------
-11-
ADDITIONAL VMT REDUCTION MEASURES - LOCAL PROPOSALS
State/Local Task Forces
State/local task forces have been formed in all AQCRs
(except the Southeast Desert) covered by this promulgation to
develop alternatives to the EPA-proposed control measures,
with the goal of developing draft plans by mid-October 1973.
Meetings were held between the task force and EPA representa-
tives to discuss potential alternatives for inclusion in the
EPA control plan promulgated for each AQCR. Although the
EPA promulgations are not wholly comprised of recommendations
of the task forces, EPA hopes that they will more properly
reflect reasonable and locally acceptable measures to improve
air quality in each AQCR. EPA also hopes that the recommended
alternative plans being developed by the task forces later
this fall will be approvable by EPA and will allow EPA to
rescind its regulations.
The membership of the task forces follows:
Los Angeles: District VII Cal/Trans, California Air
Resources Board, City and County of Los Angeles,
California Highway Patrol, Southern California
Association of Governments, Los Angeles County Air
Pollution Control District, Southern California Rapid
Transit District, South Coast Air Basin Coordinating
Council, the League of California Cities.
San Francisco: District IV Cal/Trans, California Air
Resources Board, San Francisco Bay Area Metropolitan
Transportation Commission, Association of Bay Area
Governments, and the Bay Area Pollution Control
District.
San Diego: District XI Cal/Trans, California Air
Resources Board, San Diego Comprehensive Planning
Organization (CPO), San Diego County Office of
Environmental Management, San Diego County Pollution
Control District, City of San Diego, the San Diego
Unified Port District, and San Diego Rapid Transit
District.
San Joaquin Valley: Cal/Trans, California Air Resources
Board, County, City, Regional, Fresno Community Council,
-------
-12-
local transit officials, and county, city, and govern-
mental bodies including the Fresno County Air Pollution
Control District.
Sacramento Valley: Cal/Trans, California Air Resources
Board, Sacramento Regional Area Planning Commission,
county, city and regional governmental bodies, and
the Sacramento Regional Transit District.
Four-Day Work Week Schedule
The four-day week would reduce VMT generated in work
commute travel. Like staggered work hours, this would be a
useful measure if there were a localized, temporal problem in
employment concentration areas. However, indications are that
increased recreational and other non-work travel will fully
replace if not exceed the reductions in VMT resulting from
decreased work commuting. Thus, this measure does not respond
well to hydrocarbon emission problems.
Staggered Work Hours
Changes in work schedule by staggering work hours have
been proposed as a control measure in some cities as they
tend to produce some flow improvements by reducing commute
period traffic congestion. This measure, however, would
produce only marginal reduction in emissions.
Staggered work hours do not decrease total daily VMT but
simply spreads the time of VMT generation. Such a strategy is
most applicable when the problem is a short duration, localized
concentration of pollutant, which results from temporal con-
centration of traffic flow. High concentrations of carbon
monoxide are most typical of this type of problem. Staggered
work hours, however, also tend to reduce the potential for car
pooling, a measure which relates well to hydrocarbon emission
reduction, since it tends to directly reduce VMT.
Traffic Flow Improvements
Measures to achieve emission reductions through improved
traffic flow fall into two categories: construction of new
major traffic facilities (freeways, expressways and major
arterial linkages); and operational improvements to existing
streets and highways. The emission reductions are brought
about by increases in vehicle speeds, reduced idling, and a
general shortening of trip times.
-------
-13-
Major facility construction normally enables significant
increases in vehicle travel speed in the corridors affected
but also tends to activate latent travel demand. In the long
run this reinforces auto dependence and increases vehicle miles
traveled. Over the short-range time frame of primary concern
in this study, the air quality impacts of new traffic facil-
ities can be assumed positive.
Operational improvements to existing streets and highways
cover a broad range of programs. These include freeway improve-
ments such as ramp metering and removal of bottlenecks; and
surface street improvements such as area wide signal system
integration, intersection channelization, minor widening of
streets and intersection approaches, institution of one-way
street systems, and the like. Because they do not produce
dramatic shifts in accessibility, operation improvements gener-
ally do not lead to activation of latent travel demand and their
near-term impact on emissions and air quality is assessed as
positive but their specific contribution to areawide emission
reduction is small and difficult to quantify. At best, the
planned operational improvements can be expected to accommodate
an ever increasing amount of travel without decrease in the
level of service.
Ramp Metering
Ramp metering is used to optimize the efficiency of traf-
fic movement in a freeway corridor. Metering also has potential
utility for shutting down the freeway for episode control,
and as a means to provide preferential entry for vehicles
that have a higher utilization (car pools, buses).
Mass Transit Improvements
Since personal travel requirements cannot be diminished,
some form of transportation alternatives must be provided if
vehicle use is to be reduced, particularly if vehicle re-
straints are implemented. One form for these alternatives
is public transit.
Improvements to public transit systems include both
extensions and/or upgrading of bus systems and provision of
rapid transit on separate rights-of-way. In conventional
bus operation, improvements include level of service (area
of coverage, headway, etc.) betterment and amenity promotions
(air conditioning, bus stop shelters, etc.). Most of the
urban areas already have or are in the process of setting up
transit districts to expand public transit service. These
-------
-14-
could result in significant patronage increases, but it is
unlikely that such improvements would induce major shifts of
choice riders from auto to transit without a system of con-
current disincentives for single occupancy use of the auto-
mobile.
Fringe Parking, Dial-A-Ride, Jitneys
Fringe park and ride facilities could allow suburban
commuters to park their cars on the peripheries of urban areas
and then take either mass transit or carpools into the central
business district. To have significant impact on lowering
emissions, local meterological conditions in a specific geo-
graphical area would have to be suitable. For instance, if
such a facility were in a basin, the amount of pollution
reduced would be diminished. Dial-a-ride and jitney service
could serve as feeders to either mass transit or carpool
rendezvous points, in addition to serving as a primary means
of transit.
Network of_ Bicycle Paths and Facilities
Greater use of bicycles could be encouraged through
designation and protection of bicycle lanes and incorporating
bike/pedestrian paths in new developments. A recent EPA
study suggested that increased use of bicycles in urban com-
muting could reduce auto vehicle miles traveled by as much
as three percent in some areas particularly amenable to
bicycle travel.
Imposition of restraints on auto usage, particularly
measures like gas pricing and rationing, could be expected
to encourage bicycle use. The greatest increase in bicycle
ridership would probable occur for children in getting to
school, recreation, etc., as parents pre-empt the car for
more essential functions. Work trips by bicycle could be
encouraged by providing exclusive bike lanes in city
streets and a carefully laid out bicycle grid system.
Free or Lower Transit Fares
Lowering the fare is one of the more effective means
of improving the competitive position of transit vis a vis
the automobile, particularly for intercity travel. However,
very little travel diversion from auto could be expected from
short trips, or from longer trips where time is valued over
price (see TRW and DOT travel demand elasticities models).
-------
-15-
Accompanied by auto use time penalty strategies or mass
transit time savings measures, lower fares could play a more
important role in reducing VMT.
Tax Disincentives
A "pollution" tax could be charged in direct ratio to
the emission rate and mileage of each motor vehicle or to
increase the tax on gasoline (consumption varies directly to
mileage). Schemes to reduce vehicle mileage through gasoline
pricing may be very effective if prices are set high enough.
If imposed indiscriminately on all segments of society, the
largest impact is felt by limited income groups.
Various taxes on automobiles have been proposed. Low
fees are not effective in reducing VMT and high fees are
extremely regressive. Substantial registration fees on second
or third family autos might provide reductions in VMT and still
avoid some of the more regressive elements of this type of
taxation.
Tolls
The imposition of tolls on freeways is a potential method
of regulating road use. It is possible, however, that a high
percentage of those priced off the freeways by tolls may drive
on surface streets rather than shifting to car pools or transit.
This could produce increased emissions as a result of reduced
travel speed and idling on surface streets. Tolls also tend
to be regressive since many of those priced off the roads will
be low income persons.
Gasoline Limitation
Recent increasing fuel demands and the predicted fuel
shortages may cause some gas rationing in the near future,
and therefore result in a VMT reduction. However, this
would be a by-product, and not a TCP stragegy. Administrator
Train recently announced a decision not to use gasoline
limitation as a TCP strategy, if possible. Studies have been
done evaluating the impact of schemes to raise the price of
gasoline and thus reduce consumption. The use of gasoline is
inelastic in lower price ranges and more elastic with higher
prices with increased effects if accompanied by a range -of
other VMT reduction incentives. Experience in European
countries shows that even when a gallon of gasoline is priced
as high as $1.50 a gallon, VMT rates continue to grow. TRW
[Ref. 1] estimates that it would take two years to evaluate
-------
-16-
the effects of raising the price of gasoline. However, a gas
tax accompanied by other pricing penalties on private use of
the automobile - parking surcharge, tolls, high registration
fees on second and third cars, emission taxes - would help
decrease VMT as well as contribute funds to mass transit.
II. ANALYTIC BASIS OF VMT REDUCTION ESTIMATES
To evaluate the relative effectiveness of various VMT
reduction strategies, it is necessary to have analytical
methodology which can predict the potential transit rider-
ship of average automobile occupancy rates for a set of
critical variables. Unfortunately, it is very difficult
to'quantitate the factors people use to rate the attrac-
tiveness of car pools relative to driving alone or riding
in a bus. Many of the measures EPA and the local task
forces have designed to reduce VMT are untried. EPA has
gathered the available data and studies that analyze the
range of VMT reductions possible from various transportation
controls. Analysis has included surveys to measure public
attitude and anticipated behavior, pilot studies conducted
in specific areas to gain better understanding of carpool
modal splits, and modeling to estimate the proportion of
total trips between two geographical locations that will be
made via mass transit (see Appendix 1 for listing of studies).
A Federal Department of Transportation draft report indi-
cated some important relationships between increased use of
mass transit and total auto travel.1 The following quotes
from that study illustrate the difficulties in achieving
shifts away from personal use of the car by only providing
improved transit incentives:
"(a) The price elasticity of demand for transit work
trips is only -0.19. This means that a one percent
change in price will only result in .19 percent change
in demand for transit work trips. Therefore, a 20 per-
cent decrease in transit fares would only increase
transit ridership by 3.8 percent."
"(b) The price elasticity or demand for transit shopping
trips is only -.323. Therefore, a 20 percent reduction
in price will result in an increase in shopping trip
ridership of 6.5 percent."
•'•/Alternative Transportation Investments and/or Controls for
Reduction of Air Pollution in Major Metropolitan Areas,
U. S. Department of Transportation, 1972, page 10.
-------
-17-
"(c) The time elasticity of demand for transit work
trips is -.709. Therefore, a 20 percent decrease in
travel time will increase the work trip ridership by 14
percent. (Note that time elasticity of demand is about
3.7 times the price elasticity.)"
"(d) The time elasticity of demand for transit shopping
trips is -.593. Therefore, a 20 percent reduction in
transit time will increase ridership by about 12 percent.
(Note that time elasticity of demand for shopping is
about twice as responsive to changes in travel time as
it is to changes in fare: -.593 compared to -.323.)"
In addition, the cross elasticity for mass transit may be
low without accompanying disincentives and penalties on per-
sonal use of the automobile.
EPA has funded studies to assess the possible impact of
a range of incentives and disincentives to achieve VMT reduc-
tions in five of the California AQCR's. TRW has conducted
studies for the Los Angeles Basin, San Francisco Bay Area,
San Joaquin and Sacramento Valley. A joint IREM/Rand project
[Ref. 12, 14] studied San Diego. Each region's transportation
patterns can be applied to a model that can reasonably esti-
mate the proportion of total trips between two geographical
locations that will be made via alternate means other than
private automobile when different incentives or disincentives
are applied (Appendix 2 displays the various models and cal-
culations used to arrive at the VMT reduction estimates).
The model that has been applied to most of the California
transportation studies was developed by Alan M. Voorhees and
Associates (San Francisco).
The Voorhees Model
1. A "marginal utility" function is calculated for a
typical trip between the two sub-areas between which projec-
tions are being made. "Marginal utility" is defined as a
measure of the advantages of the private automobile over
public transit. Therefore, higher marginal utilities will
result in lower levels of transit ridership and vice versa.
2. Diversion curves are developed empirically for each
regional area, which specifically relate transit ridership to
marginal utility.
-------
-18-
3. The diversion curves are used to predict the expected
transit ridership for any calculated marginal utility. It is
important to remember that the utility curves were derived
empirically, and hence subject to uncertainties. It would be
a mistake to view the model as an absolute predictor. Since
the model has inherent uncertainties, any values calculated
have the same uncertainties plus the additional uncertainties
associated with various assumptions made.
There are a number of variables which can be used to
measure the relative advantage the private auto has over
public transit. Voorhees related marginal utility to nine
variables, which fall into two classes - time and money.
Within the group of nine variables used to formulate the
marginal utility function, certain variables are more sus-
ceptible to change than others. Variables which are sus-
ceptible to change in the direction of decreasing marginal
utility are the ones which hold most potential for decreasing
VMT. Appendix 2 includes tables of the nine parameters and
some implications associated with changing their marginal
utility.
Theoretically, some or all of the nine parameters can be
modified in hopes of affecting increases in transit ridership.
Application of the model invariably shows that demand for pub-
lic transit will increase if the marginal utility of the auto-
mobile is decreased.
4. Transit ridership estimates are extrapolated for
different marginal utilities. They reflect three levels of
optimism for what transit ridership might become under rather
ideal conditions. The patronage level of forecast used
depends on the transportation characteristics of the region
in which the model is applied.
5. Within the defined patronage level of a region, the
Voorhees curves are drawn for three different levels of income.
Various income groups will exhibit different responses to, or
perceive differing marginal utility changes to a uniform change
in actual conditions. This is explained by the differing
values placed on time and money within each economic class.
Low income drivers will probably divert to public transit more
rapidly than middle income drivers in the event a substantially
increased cost penalty is associated with driving. However,
since more low income people ride transit now, the percentage
increase in transit ridership within each income level could
be about the same. Appendix 2 shows the expected changes in
transit ridership for various parameters as a function of
three income levels.
-------
-19-
For each region in which the model is used data is col-
lected and typical values assumed for each parameter needed
to simulate conditions experienced during a commute type trip.
Using the assumed values and the variables curves can be gen-
erated for estimating the percent transit ridership as a
function of each of the variables. Marginal utility is deter-
mined by holding all but one of the variables constant as one
is allowed to change.
Conclusions From Applied Model•
In comparison to transit variables, controls aimed at
penalizing the automobile, are more effective. Theoretically,
if the penalty placed on the auto is severe enough, high
transit ridership and therefore, large VMT reductions are
possible. However, the measures necessary to achieve large
VMT reductions have to be very drastic.
A fairly consistent finding is that the private auto
user is more affected by time loss and inconvenience than by
monetary considerations. To achieve large VMT reduction by
economic means alone would necessitate severe economic pen-
alties for private auto use. The impact would be very in-
equitable affecting the lower classes the most.
An example from the TRW study for the L.A. Basin shows
the degree of impedence necessary in money or time to achieve
a VMT reduction response. Assuming no transit operation
improvement, the combined vehicle operating and parking cost
would have to be raised to approximately $4.00 per day ($2.00
per trip), an increase of 14 cents per mile in order to get
to the saddle-point of the curve that separates the flat por-
tion from the steeper, more responsive section. Similarly,
if varied alone, the automobile travel time would have to be
lengthened to more than 80 minutes (an average speed of
approximately eight miles per hour) or the parking terminal
time lengthened to 30 minutes to reach the same saddle point
(present commute time is 23 minutes).
Unfortunately, attempts to use time to penalize the
private auto are often counterproductive to reducing pollution.
It has been shown that cars tend to emit less hydrocarbons and
carbon monoxide at higher levels of speed than when idling or
crawling in congested traffic. Therefore, the closest alter-
native is to implement measures that achieve great time savings
for modes of travel (bus or carpool) other than the auto. Even
this incentive does not achieve the same level of possible VMT
reduction that a direct private auto time penalty would achieve,
-------
-20-
To lenghten the parking terminal time implies the need to elim-
inate present parking spaces, a measure that is considered
infeasible in light of public testimony and comment received.*
However, parking management will control the growth of avail-
able spaces and impose time penalties if there is a growth in
private car VMT.
A combination of adequately severe economic disincentives
and of improved mass transit incentives has been promulgated
for the regions in California requiring VMT reduction measures.
Estimates of the expected VMT reductions achievable have been
made based on the best available data and its application to
the Voorhees model, or, in the case of San Francisco, to a
modified BATSC/MTC** model.
* Periphery parking cannot be expected to cause significant
increases in parking terminal time, since one would be able
to take advantage of improved local transit or minibus to
get into the CBD.
**Model designed by Bay Area Transportation Study Commission
(BATSC) to estimate transportation modal choice in Bay Area.
The Metropolitan Transit Commission (MTC) updated this modal
to reflect recent conditions in the Bay Area.
-------
- 21 -
III. ESTIMATED IMPACT OF VMT REDUCTION MEASURES
A. Los Angeles Intrastate AQCR
The majority of the data derived for Southern California
was derived from the California Division of Highways' ongoing
Los Angeles Regional Transportation Study (LARTS). Although
the geographic boundaries do not coincide exactly with the
State's definition of the South Coast Air Basin, it is assumed
by TRW and by the Department of Public Works that the study
area is equivalent to the AQCR. An analysis of population figures
for the two areas show them to be very comparable. The LARTS
study provides the input data for the Voorhees model, dis-
cussed earlier.
Table A shows the VMT control measures that are to be
implemented in the Region. Measures 8-10 are locally pro-
posed measures that were not promulgated by EPA but are to
be implemented upon their adoption as part of the State plan.
All of the measures shown on the table are complementary to
the goal of mutually reinforcing incentives and disincentives
that promote the use of multi-occupancy vehicles and penalize
driver-only cars. However, the reinforcing nature of many of
the measures makes addition of their individual impacts diffi-
cult. However, estimates of ranges are possible, holding all
parameters except one constant in testing a measure's potential
impact in the Voorhees model.
This procedure was followed by TRW in estimating approxi-
mate impacts of various VMT control measures. Table A shows
the results in the column under TRW. Estimates have also
been calculated for some of the measures by other sources.
Differences in estimates for the same measure may occur
because assumptions used in applying the model can vary,
as explained in the preceding section on the Voorhees model.
A percent transit mode split required to achieve specified
total daily VMT reduction goals is given in the TRW study.
In order to achieve a 10 percent reduction in VMT the per-
cent commute transit ridership would have to be 38 percent,
or a 30 percent increase over present levels. (8 percent
now) A 66 percent ridership would be necessary to achieve
a 20 percent VMT reduction. (Estimates assume a constant
trip demand and no change in the level of car pooling).
-------
- 22 -
Bus Improvements—
Projections of increased bus patronage by the Southern Cali-
fornia Rapid Transit District (SCRTD) show low VMT reduction
potential. A Mini-bus Project and the San Bernardino Busway
Project with park and ride facilities are estimated to reduce
VMT by only .5 percent. However, the study points out that
projecting the 1977 ridership is very speculative at this
point because the attractiveness of systems of this type in
the Los Angeles travel context is unknown. Combined with
other VMT reduction measures, the potential of the bus system
may be as effective as the Shirley Highway Busway Project in
Washington D.C. (Demand increased by 50% in one year).
Park and Ride—
A park and ride express bus service along the Santa Ana
Freeway is estimated to reduce the area's VMT by .5 percent.
TRW estimates for the combined impact of the programs planned
before EPA promulgation (of the measures listed in Table A) is
about 1.3 percent VMT reduction. It is unclear whether this
reduction has been accounted for in estimates of reductions
due to bus lanes and transit improvement, or whether it is in
addition to the present estimates.
Results of the Voorhees model for L.A. data has given
more encouraging estimates of VMT reduction. Even with no
measures penalizing the auto, a reduction in transit ride
time can attract 15 percent additional transit ridership.
A free transit would attract less than 10 percent ridership.
As indicated in Section II of this report, time turns out to
be more important than money as a variable. Using very optimis-
tic assumptions in the Voorhees model which would make transit
service free and as fast as the private auto, one might be
able to achieve an overall 10 percent VMT reduction. If
more reasonable assumptions are used in the model, one
achieves only a 3 percent VMT reduction in the Basin. These
estimates are shown in the table.
-------
Parking Surcharge—
When a surcharge on parking is brought in as a variable,
holding transit riding time equal to the auto, approximately
75 cents per trip (making the combined costs equal to $1.25
per trip) is necessary to reach the elastic portion of the
curve. Elasticity occurs at about 20 percent transit demands.
A 20 percent ridership during commute periods would get a
4 percent VMT reduction in the basin. If the surcharge were
raised to $2.00 per trip (making the combined costs equal to
$2.50 per trip) a reduction of over 10 percent could be
achieved.
The impact of increased parking costs is mitigated
because only six percent of the commuters presently pay
for parking. In order for parking surcharges to be effec-
tive, a massive program will have to be undertaken to sub-
stantially increase the percentage of drivers who pay for
parking. This could possibly be achieved in conjunction
with incentives for employees measures initiated by employers.
Free parking would only be made available to car pools. This
would probably help diminish the effect of the many parking
lots that are free and in private (employer?) ownership.
Pricing schemes would have to be worked out so that penalties
don't tend to be regressive.
The Impact of Increased Car Pools--
Since the auto occupancy rate is so low for work trips (1.1) ,
significant VMT reductions can be achieved through increased car
pool activity. To get a 10 percent reduction in VMT by car pools
the commuting auto occupancy would have to increase from 1.1 to
1.7. The freeway commuting auto occupancy would have to
increase from 1.1 to 2.1. Commute travel by freeway users
accounts for less than 25 percent of daily VMT. If all such
commuters were in car pools of three and no additional VMT
were generated by car pool assembly and dispersal, the maxi-
mum VMT reduction possible would be 17 percent. This result
is not shown in Table A because the additive breakdown attrib-
utable to various car pool incentive measures is not clear.
However, in estimating a total range of VMT reduction percents,
this figure should be kept in mind. On the optimistic side of
the tally one would be able to at least say that 17 percent is
a bottom estimate of VMT reduction achievable.
-------
Exclusive Bus and Car Pool Lanes on the Freeway—
A study by Voorhees and Associates for the U.S. Department
of Transportation on reserved freeway lanes for buses and car
pools estimates that this measure would create a 3 percent
shift into buses and a 3 percent shift into car pools. With
an increased bus ridership obtained through increased bus
service on existing routes and establishment of new routes
in the corridor, shifts greater than 5 percent into buses
5 percent into car pools would not be unrealistic. It is
unclear whether the study anticipated the possibility that
the 5% increase to buses due to transit improvements would
partially come from the 3% shift to car pools, thereby not
increasing total VMT reduction but shifting the proportion
of car pools to buses. Table A reflects the predicted 6 to
10 percent range.
TRW's study arrives at a lower estimate of the bus/car
pool lane's potential effect. By generously assuming a twenty
minute travel time advantage over automobile travel maintained
by the buses or car pools in the special lanes, a 15 percent
ridership could be assumed, giving a 2.5 percent reduction in
daily VMT. If this measure achieved a 1.5 persons per vehicle
occupancy on the L.A. freeway system during commute hours, the
VMT reduction would be about 4.4 percent of the daily Basin-
wide VMT.
Impact of Local Task Force Measures—
The task force recommended a series of measures to speed
the flow of traffic and avoid the bottlenecks and stop-and-go
driving that are both polluting and wasteful of energy. These
measures include: automated and interconnected traffic sig-
nals, freeway ramp metering, expanded fringe park-and-ride
facilities, and other traffic flow improvements. The task
force estimates that these measures will total an additional
5.3 percent reduction in pollution.
Total Ranges of VMT Reductions—
Using the TRW based estimates, one can expect about a 12.5
to 30.5 percent reduction in VMT plus whatever additional percent
reductions can be calculated from parking supply management,
and non-overlap mass transit incentives for employees, etc.
Also added on to the sum would be the extra VMT reduction
achieved by local traffic flow improvements, if any, and mass
transit improvements. Substituting the DOT bus/car pool lane
estimates the sum increases to 13.5 to 34.5 percent.
-------
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
LOS ANGELES
Control Measures
Exclusive Bus/Car Pool Lanes
Bus/Car Pool Matching (not
necessarily additive with
measure 1)
Mass Transit Incentive for
Employees
Parking Supply Management
Parking Surcharge
VMT/Air Quality Improvement
Monitoring Program
Gasoline Limitations
Mass Transit Improvements
and Development
Bicycle Network & Facilities
Traffic Flow Improvements *
Fringe Parking for Park
and Ride
Dial-a-Ride
-25-
TABLE A
- VMT REDUCTION MEASURES
Estimated percent
Reduction in Daily
Vehicle Miles
Traveled - DVMT
0) >i W
0) TJ >i X
•H -3 'd *-* H
O 4J 3 RJ O
CO) CO -P -P 15
(1) O W tO
Cn to in Q &
Promul- Proposed So So KW P*
gated by by Local ^ ^ Q •£ ° ^ g ^
EPA & Cal/Trans own on << a
O<0»H W^" Q
X X 4 6-10 2.5- 2-3
4.4
X 5-6
***
X 1-2
***
X 1-2
X 4-10
X
x ****
X 3-10
X
** ***
X 5.3 0-1
**
X .3 .5-4
***
X .1-4
* Traffic flow improvements may awaken the latent demand for increased
use of the less congested facilities. Therefore, caution must be taken
in calculating the VMT reduction for this measure.
** Estimates are for the percent of pollution reduced, not VMT
*** Estimate by EPA, Region IX
**** Amount necessary to meet air quality standards in 1977.
-------
- 25 -
APPENDIX I
References
1. TRW, Transporation Control Strategy Development for the
Metropolitan Los Angeles Region, December 1972
2. TRW, Transportation Control Strategy Development for
Sacramento Valley
3. TRW, for San Joaquin
4. TRW, for San Francisco
5. TRW, Air Quality Implementation Plan Development for Central
California Regions; Summary Report July 1973
6. State of California, Department of Public Works, for Air
Resources Board, Can Vehicle Travel Be Reduced 20% in South
Air Basin? January 1973.
7. California Institute of Technology, EQL, SMOG;
A Report to the People of the South Coast Air Basin,
January 1972
8. DOT, a Computer Simulation Model for Evaluating Priority
Operations on Freeways June 1973.
9. Alan M. Voorhees and Associates, for Department of
Transportation(DOT), Summary Report:
Feasibility and Evaluation Study "oT Reserved Freeway
Lanes for Buses and Car Pools, January 1971.
10. MTC, Proposed Regional Transportation Plan and Revisions;
Improvement Proposals, June 1973.
11. Local Agencies Plan for LA Basin,
Clean Air, August 9, 1973.
-------
-26-
12. San Diego Air Pollution Control Task Force Report, Local
Transporation Control Plan for the San Diego Basin,
October 5, 1973.
13. San Diego Air Pollution Control Task Force, Evaluation of EPA's
Proposed Air Pollution Control Plan with Alternatives
14. Rand Corporation, A Policy Oriented Urban Transportation
Model; Notes on the Rand Approvals, March 9, 1973.
15. TRW, Report on the IREM/RAND Clean Air Project.
16. EPA Region IX, Technical Support Document for LA Basin,
July 31, 1573
17. Federal Register, Proposed Regulations for California Air
Quality Control Districts, Vol. 38, No. 126, Monday,
July 2, 1973
18. Federal Register, Promulgation for Transportation Control
Plans for California, Vol. 38, No. 217, Part I,
November 12, 1973
-------
- 27 -
APPENDIX 2a
VOORHEES MODEL
Marginal utility = U = 2.5 (Ta + Tw) + Tr-(2.5 At + Ar) + F-(Ao + Ap) (1)
0.251
where : Ta = transit access time in minutes (i.e. the time to walk
to the bus stop)
Tw = transit wait time in minutes
Tr = transit riding time in minutes
At = automobile parking access time in minutes (i.e., the
time necessary to find parking and walk to destination)
Ar = automobile riding time in minutes
P = transit fare in cents
Ao = automobile operating costs in cents (excluding
depreciation and insurance)
Ap = automobile parking cost in cents (averaged over the
round trip)
I = mean income of the home based zoned in cents per minute
-------
- 28 -
TABLE 7.4
Control of Marginal Utility Parameters
Marginal Utility
Parameter a
Ta
Tw
Tr
At
Ar
F
Ao
Ap
I
Potential
for Control
Low
Medium
Medium
Low
High
High
High
Medium
Very low
Example (s) of Control Aimed
at Decreasing Marginal Utility
More bus stops and/or routes
Improved frequency of service
Exclusive busways for freeway lanes
Peripheral parking, auto free zones
Ramp metering
Lowered fares
Gasoline tax, "smog" tax, tolls
Increased parking costs
Lower personal income levels
a - see Equation 1
-------
- 29 -
TABLE 7.5
ASSUMED VALUES OF MARGINAL UTILITY PARAMETERS
FOR A TYPICAL COMMUTE TRIP
Variable Assumed Value Source
Trip Length 10.5 miles Table 5.4 (EPA publication APTD-1372)
Ta 5 minutes Generally accepted value
Tw 7 minutes 15 minute headway, no transfer
Tr 48 minutes LARTS, 1971 Travel Time Study
F 38 cents 30C basic fare + 8£ for one
additional zone
Ao 50 cents 4.8C per mile - assumed by Voorhees
Ap 2.5 cents 90C per day for estimated 5.8
percent who pay See Table 7.7.
I 8.3 cents/minute $10,000 per year
At 2 minutes A. M. Voorhees & Associates
Ar 23 minutes Table 5.4 (EPA publication No.
APTD-1372)
-------
- 30 -
TABLE 7.6
MARGINAL UTILITY PARAMETER COMBINATIONS
Marginal Utility Variables
Transit fare (one way)
Transit access and waiting time
Transit riding time
Auto operating and parking
cost (one way)
Auto riding time
Auto terminal time
Symbol
F
Ta+Tw
Tr
AQ+AP
Ar
At
Potential
for Control
High
Medium
Medium
Medium
Medium
Low
Relative
Effectiveness
Low
Medium
High
High
Low
Low
NOTE: Income is considered a constant throughout, primarily because
it is virtually impossible to control.
-------
- 31 -
40
30
«_
Qi
•o
>20
V-
o»
10
j L
•Currently .38
.30
1.20
1.50
.60 .90
F - Dollars
Figure 7.3. Transit Fare (One Way) vs. Per Cent Transit Rldershlp.
Source: Developed from the Los Angeles Metropolitan Area Mode Choice Model,
r
-------
20.0
17.5
15
- 12.5
2 10
-------
- 33 -
20
18
16
I14
•o
5 12
i/i
£ 10
O)
-------
40
o.
.c
01
•o
•£ 30
c
0)
5 20
10
.50
Currently .52
1.00
1.50
2.00
2.50
AQ + A - Dollars
Figure 7.6. Automobile Operating and Parking Cost
(One Way) vs. Per Cent Transit Ridership.
Source; Developed from the Los Angeles Metropolitan Area Mode Choice Model.
-------
40
OJ
•o
30
20
c
01
O
u
Of
0.
10
10
- 35 -
Currently 23
30
50
70
90
110
Af - Minutes
r
Figure 7.7. Automobile Riding Time vs. Per Cent Transit RldersMp.
Source; Developed from the Los Angeles Metropolitan Area Mode Choice Model.?
-------
-36-
40
30
£
•o
•r™
ac.
I/I
£20
-------
Sample Calculation 09
Impact of Exclusive Bus-Carpool Lane
- 37 -
Assumptions:
1) frip length = 10.1 miles
2) 20 minute travel time
advantage for users of the
lane
3) Values of Voorhees parameters
T, = 2 minutes
d
Tw = 3 minutes
F = 46*
AQ = 48* (or 780
Ap = 2.5*
At = 2 minutes
Tr-Ar = 20 minutes
I = 8.3*/minute
U - 2.5
r - (2.5At+Ar)
Source
See Appendix G, Table G-8.
(Trip to freeway is excluded.)
San Bernardino Busway will have
18 minute advantage; Shirley Highway
Busway has 30 minute advantage.
Assumed
5 minute headway for correct bus
Basic 30* fare + 2 zones
4.8* per mile (+3* per mile "tax")
Assumed
See Reference E-5
20 minute time advantage
$10,000 per year
F - (A +A )
U = -15 (-29 with 3* per mile "tax")
Based on the above assumptions and Figure 7.2, the estimated transit
patronage 1s approximately 15 percent without the tax and 20 percent with
the tax.
-------
Sample Calculation #9 (Cont'd) - 38 -
OCB(1-FT) + OCA(FT)
% VMT Reduction = 100-100
OCB(1-F
OCA(F,
,Sample
Calculation #1)
The percentage of total commute VMT which is travelled on the freeway system
. . - - • .. • ""• •"•"'•" -
In the*Basin is approximately 17 percent (see Appendix G, Table.G-8); therefore",
OCB(1-FT) •
OCB(1-FT )-
0
* OCA(FT) "
H OCA(FT )
0 .
% VMT Reduction « ,17 100-100
Assume:
1) OCB = 50 persons per bus
2) OCA « 1.1 persons per car
3) F, = negligible on freeways presently
o
4) FT = .15 (.20 with "tax")
% VMT Reduction * 2.5 (or 3.2 with "tax")
To achieve an occupancy of 1.5
Assume:
1) During commute periods, average volume per lane of traffic
of 1500 vehicles per hour
2) Carpool is three or more people per automobile
3) Negligible carpooling on the freeway presently
4) Eight lane freeway (four in each direction)
5) Capacity of freeway lanes is 1800 vehicles per hour
6) No diversion to transit
7) Total person trips on freeway 1s constant
-------
Sample Calculation 09 (Cont'd) - 39 -
Let N » Number of cars in carpool lane (three persons per car) when '
all other lanes are at capacity and the resultant average
automobile occupancy is 1.5 persons per vehicle.
Then * (3)(1800 veh./1n-hr)(l.1 persons/veh.) + N(3 persons/veh.)
(3)(1800) + N veh./ln-hr '
•*• N = 1440 veh./ln-hr
Persons diverting to carpools = 1440 x 3 = 4320
Total number of person trips = (3)(1800)(1.1) + (1440)(3) = 11,340
% of freeway commuters shifting to car pools = n * » 38%
I I ,
Impact of Shifts in Carpools
Assume occupancy can be increased to 1.5 persons per vehicle.
From Sample Calculation #2 for commute periods periods,
% VMT Reduction = 32.4 fl- lf21
This strategy impacts 72.5 percent of the commute VMT (see Appendix G,
Table G-8).
% VMT Reduction = .725(32.4)(1-
= 23.5 (1- l
= 4.4%
-------
APPENDIX D
EXAMINATION OF THE PHOTOCHEMICAL
AIR POLLUTION PROBLEM IN THE
SOUTHERN CALIFORNIA AREA
E. A. Schuck and R. A. Papetti
I. INTRODUCTION
Many areas of the world which are highly urbanized experience
tvo types of major air pollution problems. The first is due to
direct emission into the air we breathe of a variety of components
mainly from energy producing and related activities associated with
the use of fossil fuels. Many of these exhaust components such as
carbon monoxide, sulfur dioxide, and particulate matter produce
significant deleterious health and welfare effects at the concentration
observed in urban areas. Others such as most unburned or partially
oxidized hydrocarbons, carbon dioxide, water, and nitric oxide have
no significant direct health effects at observed ambient concentrations.
Certain of the hydrocarbon products do, however, result in significant
welfare effects. A major example is the vegetation damage resulting
from the hydrocarbon automobile exhaust component, ethylene. /
j'
Some of these largely innocuous exhaust components, especially
the hydrocarbons and nitric oxide, react after emission into the
atmosphere to produce new products thus leading to the second type
of major pollution problem, termed "smog," or more specifically
-------
- 2 -
photochemical air pollution. Many of the major products of these
sunlight induced atmosphere reactions, i.e., ozone, peroxyacyl nitrates,
nitrogen dioxide, and certain aldehydes, pose a serious health problem
and in addition many of these secondary products are deleterious to
vegetation growth and some such as the peroxyacyl nitrates and certain
aldehydes produce eye irritation. To further compound the effects
these atmospheric reactions also result in the rapid oxidation of part
of the sulfur dioxide present and thus to the production of sub-micron
sulfuric acid aerosols vhich result in greatly reduced visibility as
veil as further increasing the health problem resulting from particulate
matter.
. N
There has been a tendency in recent years to sharply differentiate
cities into those whose air pollution problem is due to primary
emissions, i.e., sulfur dioxide and particulates, and those whose
problems are due to the secondary effects of photochemical air pollu-
tion. Such categorization, while correctly identifying the immediate
most pressing problem of a given urban area nevertheless also tends
to suggest that photochemical air pollution is unique to certain /
areas in the world. This concept is incorrect since chemical
processes which lead to photochemical air pollution takes place during
daylight hours wherever there are atmospheres containing hydrocarbons
and oxides of nitrogen. Indeed the chemical reactions involved are
-------
- 3 -
an essential part of the oxidative processes vhich prevent the buildup
of hydrocarbons in the earth's atmosphere. In lieu of such reactions
the overall carbon dioxide cycle would be interrupted and the air we
breathe in all probability would contain concentrations of certain
hydrocarbons which would produce deleterious health effects. Thus,
if urban areas concentrate all their efforts on the Immediate obvious
problems, such as excess concentrations of sulfur oxides or directly
emitted particulate matter, they will, after resolution of these
factors, be faced with the more insiduous problems resulting from
photochemical air pollution. Whether recognized or not the presence
and therefore detrimental effects of photochemical air pollutants
s
exist in all urban areas and will continue to be a mojor problem as
long as our energy base is dependent on the use of fossil fuels.
f
Areas in the United States where the effect of photochemical air
pollution was early recognized as a major problem are located on the
western edge of California. Because of such early recognition an
extensive monitoring network for both precursors and products has
been in use for the past 15-20 years. The most extensive of these /•
multistation networks, and the one whose records are examined here,
exists in the Southern California area surrounding Los Angeles and
known as the South Coast Air Basin (SCAB). Into the common air mass
covering this region the ten million inhabitants dump approximately
-------
- U -
20,000 tons of pollutants each day. The semi stagnant characteristics
of this air mass leads to the mild weather conditions which the
population finds desirahle "but also results in restrictive dilution
and dispersion of pollutants therefore creating a severe air pollu-
tion problem.
The magnitude and extent of the resulting health and welfare
effects is impressive and alarming. Imagine if you will an area of
one thousand square miles on a typical morning with 50 mile visibility.
Four hours later the visibility has*been reduced to less than a mile,
not by direct emission of particulates, but rather by chemical reactions
in the atmosphere which have created a submicron aerosol containing
as its principal constituent sulfuric acid. Furthermore, as the day
progresses one can visually trace this pollution cloud for hundreds
?
of miles downwind. This visibility reduction, although physcologically
depressing, is the least of the problem. This visually apparent air
mass contains many substances which are detrimental to all forms of
biological life. As an example, consider that certain types of sensitive
commercial crops can no longer be grown in or near the SCAB. Millions
/
of trees in the surrounding mountains are also being destroyed. Yet
these effects are minor compared with the human health aspects. On
as many as 250 days per year from 7 to 9 million of the local citizens
must breathe air for several hours which, along with other deleterious
-------
- 5 -
components, contains oxidants (ozone) above the level of the health
related national air quality standard (0.08 ppm for one hour).
Furthermore, between 1 and 3 million of these citizens are breathing
air which contains oxidant concentrations above the point where medical
authorities advise against engaging in physical activities (above
0.27 ppm for one hour).
Since the early 1950's many local governmental agencies and in
particular the Los Angeles County Air Pollution Control District
have pursued an active and successful program to reduce the amounts
of pollutants being dumped into this air basin. Without a doubt
the automobile is the greatest contributor to ambient concentrations
V
of carbon monoxide and to the precursors of photochemical air
pollution, i.e., hydrocarbons and oxides of nitrogen. Following
» .
California's lead additional restrictions on the amounts of hydro-
carbons and carbon monoxide in new motor vehicle exhaust were
instituted by the federal government starting in 1968, It is the
purpose of this discussion to examine the air monitoring data base
of the SCAB in order to show what changes have occurred during the /
past several years and to explore the reasons for these changes. An
auxiliary purpose in this examination is to update the relationships
between ambient early morning concentrations of precursors and
subsequent oxidant values.
-------
- 6 -
II. The Data Base and its Analysis
Since the early 1960's, at "between 8 and 20 locations, oxidants
and their precursors have been continuously measured in the South
Coast Air Basin. This is particularly true for total oxidants, nitric
oxide and nitrogen dioxide. Measures of total hydrocarbons "being more
difficult to obtain, vere until recently only measured at one or two
locations. Starting in 1971 however, continuous hydrocarbon measure-
ments have been available from eight locations. This hydrocarbon measure
contains an appreciable quantity of methane which does not participate
in the oxidant forming reactions. Direct methods of measuring non-
methane hydrocarbons are not in widespread use. In lieu thereof the
N
Division of Chemistry and Physics, EPA, has on the basis of detailed
chromatographic analysis of ambient hydrocarbons developed an
t
empirical formula for converting the total hydrocarbons measurement
to a non-methane hydrocarbon measurement.
This formula is:
Non-methane hydrocarbon (ppmC) = 0.7(THC-1.3)
vhere THC is the total hydrocarbon measurement (ppmC) and the 1.3 /
/
represents the average background methane concentration. Unless
otherwise stated the hydrocarbon values cited in this discussion
have been derived by application of this formula.
-------
- 7 -
Since a major objective is to quantitate changes in air quality
the data "base consisting of hourly average frequency distributions at
8 stations for oxidant and at 11 stations for nitrogen dioxide vere
examined. Specifically examined vere the hourly frequency distribu-
tions during the 19^3 through 1967 time period as compared with the
distributions observed during the 1968 through 1971 time period.
Essentially the earlier time period represents a pre-automotive exhaust
control period while the 1968-1971 data is a post-control period in
which substantial changes in emissions were predicted as a result of
increasing impact of automotive emission controls. The oxidant com-
parison was made to determine what effect if any the automotive
s
hydrocarbon control program has produced. The nitrogen droxide comparison
was made because the method chosen to control automotive hydrocarbons
*
also resulted in substantial increases, in nitric oxide. This" latter
comparison thus was made to determine the effect of such increases of
nitric oxide emissions on subsequent nitrogen dioxide concentrations.
III. Discussion of Results
A. Ambient Oxidant Changes. ' /
In table I are listed the changes observed in hourly average
oxidant concentrations at eight stations when comparing the noted
post automotive control period with the pre-control period. Substan-
tial reductions in hourly average oxidants have occurred at all stations
-------
Table I
Change in Hours Per Year - Average Hourly Oxidant Concentrations
Comparision of 1968 - 1971 Data With 1963 - 1967 Data
Station and Code
North Long Beach 72
Downtown Los Angeles 1
West Los Angeles 71
Reseda 74
Pomona 75
Riverside 126
Azusa 60
Bur bank 69
Average Change for
All Stations
Percent Change in Stated Concentration Ranges (pphtn)
1-8
+3
+3
+6
*
+18
+13
+6
0
+1
/
+6
8-25
-57
-42
-29
-27
-25
_-14
-3
+3
-20
25+
-67
-83
-78
-65
-28
-26
-21
-60
-39
•
-------
- 8 -
with the greatest reductions occurring in the higher concentration
range (i.e., 25 pphm and above). Two items are worthy of note-
First these reductions are much greater than could he expected from
hydrocarbon control alone since the maximum hydrocarbon reduction
from the pre-control to post-control period is 8-10$. Second the
stations with greatest upwind contributions, i.e., Riverside, Azusa,
Pomona, show the smallest reductions. This is in line with previous
determinations that in general the concentrations of any given
pollutant at a specific location is proportional to number of and
intensity of sources over which the air mass has passed. The noted
increases in the 1-8 pphm range are discussed under Section III-C.
N
B. Nitrogen Dioxide Changes
As shown in Table 2 for 11 stations hourly average nitrogen
»
dioxide above a value of 25 pphm demonstrate substantial increases
except for the Riverside and Lennox sites. Why these latter sites
exhibit a decrease is not known however, it will be noted that
this decrease is not reflected in total exposure or in the yearly
mean values. For that matter there is no ready explanation for the
very substantial increase noted in hours above 25 pphm for the Basin
as a whole (+U2$) since the projected increases in nitric oxide
emission are only about 16-20$ which is more in line with the noted
changes in basin wide total exposure and yearly mean.
-------
Table 2
Observed Changes in Selected Hourly Average
Nitrogen Dioxide Concentrations and Total
Nitrogen Dioxide Exposrue During
1963-1971 in the South Coast Air Basin
Station and Code
Anaheim 176
Riverside 126
West Los Angeles 71
Reseda 7**
Pomona 75
North Long Beach 72
Downtown Los Angeles 1
Lennex 76
Burbank 69
Azusa 60
San Bernardino 151
Average Percent Change
(All Stations)
Hours Per Year
Above 25 pphm
1963-1967
1
25
70
1*1
25
88
111*
ll*0
96
6
0
1965-1971
20
12
88
70
81
111*
131
61
251*
29
'2
Percent Change in 1968-1971
Compared to 1963-1967
Hours Above
25 pphm
+520
-52
+26
+71
+220
+30
+15
-56
+165
+380
—
+1*2
" Total Exposure
(Cone. X Time)
+1*0
+16
+33 -!
+35
+17
+17
+13
+8
+38
+26
+1*5
+23
Yearly Mean
+25
+25
+1*0
+1*0
+11*
+33
0
+17
+1*3
+20
0
+23
-------
- 9 -
It is probably more correct to say there is no precise quantifi-
able explanation for these changes rather than implying an area of
unknown factors. From a diverse and substantial data base generated
in environmental chambers it is quite apparent that the relationship
between oxidant and oxidant precursors is quite complex. As long
as ten years ago, Dr. Philip A. Leighton pointed out the danger of
changing the ambient hydrocarbon to oxides of nitrogen ratio. He
did in fact predict that reductions in hydrocarbon without an attend-
ing reduction in oxides of nitrogen would lead to some reductions in
peak nitrogen dioxide concentrations and also to an increase in
nitrogen dioxide total exposure. This effect observed at two
" x
stations in Table 2 may in fact be more general however, it is not
possible from this data base to substantiate such an effect since it
is difficult to seperate the effects due to the substantial increase
in nitric oxide emissions from those due strictly to hydrocarbon
reductions.
One positive effect of the increase in nitric oxide emissions
has been the depressant effect on oxidant concentrations. Nitric
oxide and ozone react extremely rapidly in the atmosphere to produce
nitrogen dioxide and oxygen. Thus the substantial decreases in
oxidant shown in Table I are as much if not more due to increases in
nitric oxide emissions as they are due to reductions in emissions of
0
hydrocarbons. Obviously the tradeoff is that in the process the
-------
- 10 -
ambient nitrogen dioxide concentrations have increased. While it is
highly desirable to continue taking advantage of the presence of
excess nitric oxide, the positive effects on oxidant reduction must
"be "balanced against the negative effects of increased nitrogen dioxide
since the latter also has health implications as well as resulting in
coloration of the atmosphere. Moreover, nitrogen dioxide is the
principal light absorber leading to the formation of photochemical
air pollution.
C. Frequency Distribution Changes
Of particular interest are the changes in hourly average oxidant
and nitrogen dioxide as a function of concentration range. A summary
of these changes for. the. Basin air mass is shovn in Table 3. It
vill be noted that a reversal in the sign of.change occurs for both
components at.about 5pphm. Thus the reductions in oxidants 'observed
previously lead to., increases in hours of oxidant. in the 1-5- pphm
range. The reverse is true for nitrogen dioxide concentrations. The
reason,for these reversals is that for any given, pollutant the frequency
distribution is primafilly determined/by meteorological.variables.
/
Thus in the case of nitrogen,dioxide,the. increases in hours .above.. 5
pphm have to ,be accounted for someyhere in the frequency distribution.
Many of the hours that formerly contributed to hours, in the 1 to 5
pphm range have becasue of increases, in nitric oxide emissions moved
up to higher concentration ranges. The reverse is true for the oxidant
-------
Table 3
Observed Changes in Frequency Distributions
of Hourly Average Concentrations of
Oxidant and Nitrogen Dioxide
1963 through 1971
Hourly Average
Concentration
Range (pphm)
1-2
2-3
3-U
U-5
5-6
6-8
8-10
10-15
15-20
20-25
25-30
30-UO -
UO-60
Percent Change in Hours Per Year
in 1968-1971 Compared to 1963-1967
Oxidant
+8
+9
+15
+5
-9
-6
-16
-20
-23
-27
-39
-1U
-8
Nitrogen Dioxide
-1*6
-23
-11
• -7 '
+;6
+21
+31
+U9
+63
+52
+1*1
+38
+7
1 Average Number of Hours
Per Year in 1968-1971
Oxidant
1630
952
617
U26
259
U56
279
U09
192
88
3U
16
3
Nitrogen
Dioxide
575
935
1070
1020
930
lUOO
900
1050
3U8
123
U8
26
6
(l) Oxidant data at 8 South Coast Air Basin Stations. Nitrogen dioxide data at 11 South Coast
Air Basin Stations.
-------
- 11 -
frequency distribution. Depending on weither the interest is peak
valves or dosage these frequency shifts will have a decided effect
on calculated changes. Thus if total oxidant dosage should "become
an important health factor the change in dosage over all concentration
ranges would lead to a much less dramatic reduction than that shown
in Table 3 as contrasted to the case where only peak valves above
8 pphm are important.
D. Relationship of Oxidant to Precursors
Since ambient concentrations are a complex function of the
precursor hydrocarbons and oxides of nitrogen there have been various
attempts to quantify this relationship for control purposes. The
v
simplest such approach, termed linear rollback, assumes a one to one
relationship between oxidant and precursors. Taking the highest
oxidant valve observed in an area one can calculate what percent
reduction in oxidant is required to reach a lower oxidant value such
as the national air quality standard. It is then assumed that this
same calculated reduction will apply to the precursors. Most
frequently this is applied to hydrocarbon control alone although the
implication is for equal control of both precursors. Data from
environmental chambers however, as well as observations of ambient
atmospheres consistently show that the oxidant precursor relationship
is not linear even when the hydrocarbon to oxides of nitrogen ratio
-------
- 12 -
is held constant. The available data suggests that the linear -
rollback technique tends to underestimate the degree of precursor
control required.
A second approach is to use the relationship as determined in
atmospheric simulations in environmental chamber studies. Unfortunately,
at least until very recently, the vails of these chambers have caused
fluctuations in the results and also the investigators have been forced
by limitations of measurements to operate at precursor concentrations
veil above 'those found in real atmospheres. While all such studies
are in qualitative agreement no quantitative projection can be made
from this data base. What does become clear from these environmental
chamber studies is that hydrocarbon control alone vill in all cases
lead to reductions in oxidants. These studies further shov that oxides
of nitrogen reductions can lead to both increases and decreases in
oxidants depending to a large extent on vhat hydrocarbon to oxides of
nitrogen ratio are being investigated. Investigation of early
morning ambient atmospheres indicate an existing hydrocarbon to
oxides of nitrogen ratio in the vicinity of about 10 to 1. Given
/
this fact the chamber studies suggest the following guidelines
for oxidant control in the South Coast Air Basin.
(l) Hydrocarbon control is the prime consideration for oxidant
control.
-------
- 13 -
(2) Oxidant control will be enhanced if the ratio of hydrocarbons
i
to oxides of nitrogen decreases. This can be accomplished
by (a) not controlling oxides of nitrogen or (b) controlling
oxides of nitrogen at a slower rate than hydrocarbons.
.Partial oxides of nitrogen control appears necessary in
view of the health and welfare effects of nitrogen dioxide.
(3) If the hydrocarbon to oxides of nitrogen ratio is permitted
to increase to 20 to 1 then hydrocarbon control will not be
as effective in oxidant "control. This event is not likely
to happen since for most areas only minimal control of
oxides of nitrogen is required in order to meet the national
air quality standard for nitrogen dioxide.
The third approach to delineation of the oxidant-precursor
relationship involves equating early morning precurson levels to
subsequent oxidant peaks. The method chosen by the Environmental
.Protection Agency in 1970 was to equate the highest observed oxidant
for a given day with the 6 - 9 am hydrocarbon peak. This has been
termed the upper limit concept since for any given early morning t
/
hydrocarbon valve there will exist a range of subsequent oxidant valves
from zero up to an observed maximum. This method has been severely
criticized as being non statistical however, as we will
discuss the more recent application of regression analysis has
-------
- 1U -
validated this upper limit concept.
Since 1970 the body of applicable data has increased making it
possible to redefine the upper limit curves for multiple stations
in the South Coast Air Basin. Three such curves "based on 1968-1971
data are shown in Figure I.. Since little data exists in this Basin
"below 0.5 ppmC of hydrocarbon the curves have "been extropolated to
zero. The data points indicated in the vicinity of these extropolation
are those obtained from other large U.S. cities where actual measure-
ment of non-methane hydrocarbons as well as peak oxidants were observed.
At first glance one might expect substantial differences between
stations in.projected degree of hydrocarbon reductions to reach a
given oxidant valve. In reality thevprojections from the highest
oxidant observed at each station give quite similiar projections. In
order to reach the 8 pp'hm one hour standard the three curves in
Figure I gives a valve of 93+. 3% for hydrocarbon control. Again, as
in the case of linear rollback the curves shown in Figure 1 assume
equal control of the two precursors. Thus while linear rollback tends
to underestimate the degree of control the-use of the curves in Figure
.'
1 tends to overestimate the degree of control since we know that the
hydrocarbon to oxides of nitrogen ratio will decrease thus enhancing
oxidant control.
As previously noted the 1971 data base includes 8 stations in
-------
Figure 1
70
6o
50
I
P<
aJ
•a
«>30
c
o
20
10
Riverside
Downtown Los Angeles
I
_L
1 2 3 1* 5
Non Methane Hydrocarbons (ppmC) (6-9am Average)
UPPER LIMIT ONE HOUR OXIDANT CURVES FOR
SELECTED SOUTH COAST AIR BASIN STATIONS
1968-1971 DATA
-------
- 15 -
the Basin where hydrocarbons as veil as oxidants vere measured.
This permits comparing the average 6-9 am hydrocarbon value for the
Basin (as measured "by these eight stations) with the highest oxidant
observed at any of the stations in the Basin. This Basin average
hydrocarbon value should provide a more direct link to emission
inventories since these latter are also on a Basin wide basis.
The resulting upper limit curve developed from this 1971 data base is
shown in Figure II. Using the highest oxidant observed during that
year suggests a 91$ reduction in hydrocarbons in required in order
to attain the national air quality standard of 8 pphm for one hour
once per year. .Again since the data base explores only the existing
ambient precursor ratios it must be assumed to be an overestimate of
the degree of hydrocarbon control required.
A fourth approach 'to determination of the oxidant-precursor
relationship is the use of regression analysis. The specific relation-
ship used both by EPA researchers and independent investigators at
the Chevron Research Company assumes peak daily oxidant to be .
proportional to power functions of early morning hydrocarbon and,.
/
oxides of nitrogen values. Unfortunately such assumed proportionality
cannot take into account the complex chemistry involved. The resulting
relationships correctly identify the direct or inverse relationship
to hydrocarbon but cannot agree on a direct or inverse relationship
to oxides of nitrogen. These analysis do however provide a
-------
60
t50
G
CO -H
a to
O -P
a a)
•p
a5
•d
30
s 10
CO
1 2
Non Methane Hydrocarbons (ppmC)
(Average for 8 Stations) (6-9am)
(1971 Data)
Figure 2
UPPER LIMIT OXIDANT VALUES IN 'THE SOUTH COAST AIR
BASIN AS A FUNCTION OF AVERAGE 6-9am HYDROCARBON
CONCENTRATIONS AT 8 STATIONS
-------
- 16 -
statistical check on the validity of the upper limit approach. Thus
if the Chevron or EPA regression expression for the Downtown Los
Angeles location are projected to the once per year event they
duplicate the curve shown in Figure I for that station.
-------
- 17 -
IV. Summary
Comparing a pre -automotive control period (1963-196?) with
a post-control period (1968-1971) shows that changes in automotive
emissions patterns have resulted in substantial changes in subsequent
photochemical air pollutant concentrations in the South Coast Air
Basin. Projection by the Los^Angeles County Air Pollution Control
District suggests that the post-control period is characterized by
an 8-10$ reduction in hydrocarbon emissions and a 16-20$ increase in
nitric oxide emissions when compared with the pre -control period.
. Examination of the air monitoring data from 8-11 stations in this
•
1000 square mile area shows that during this same time period the
average number of hours of oxidant concentrations above 25 pphm has
decreased by 39$ and the average number of hours of nitrogen dioxide
N
concentrations above 25 pphm has increased by ^2$. The reduction in
oxidant concentration is partially due to the decrease in hydrocarbon
t
emissions and partly due to the increase in the nitric oxide emissions.
These latter react instantaneously with oxidant (ozone) to form
oxygen and nitrogen dioxide. Thus increased nitrogen dioxide
concentrations is the resulting trade-off. for much of the observed
oxidant reductions. Nevertheless the essential point is that peak /
oxidant values are on the decrease as a direct result of control on
automotive emissions. Furthermore these reductions are occurring in
spite of a substantial population growth. To what extent we can
continue to take advantage of the depressant effect of excess nitric
-------
- 18 -
oxide emissions on peak oxidant concentrations -will be a function
of the health and welfare effects of the resulting increased nitrogen
dioxide concentrations. As of this point in time in the South Coast
Air Basin the health threat from oxidants would appear to "be greater
than that from nitrogen dioxide.
Reevaluation of the relationship between ambient oxidants
and their early morning precursors in the light of an expanded air
monitoring data base shows no essential change over previously •
developed relationships. The expanded data base does permit relating
Basin wide average hydrocarbon concentrations to observed Basin wide
peak oxidant rather than using such relationships at individual
stations. The result of such a Basin wide relationship should permit
a more direct connection to emission levels since the latter are also
on an area basis. The results of this newly developed oxidant-precursor
relationship are not significantly different than that obtained at
single stations. This is not an unexpected result since the stations
in this 1000 square mile area are highly interrelated because of gross
meteorological factors. As- an example, if the average daily peak
oxidant for all stations is known, it is possible to calculate the
.Basin wide maximum peak within a deviation of
The Basin wide oxidant-precursor relationship predicts a
need for 91% control of ambient hydrocarbons in order to reach the
national air quality standard of 8 pphm for one hour once per year.
-------
- 19 -
Use of a linear rollback technique, which is an underestimate, suggests
a hydrocarbon control value of 85$. Both methods assume equal control
of both hydrocarbon and oxides of nitrogen. This is not likely to
occur, i.e., nitric oxide will persist in the emission and thus the
depressant effect of these latter emissions will probably lead to the
need for a hydrocarbon control somewhere between 85$ and 91$- It will
be noted, however, that these projections do not take into account
expected growth in emissions. If such growth continues, a degree of
hydrocarbon control greater than 91% will eventually be required.
V. Acknowle dgment s
The authors wish to express their appreciation to D. L. Worley
and R. I. Larsen for their expert assistance in analysis of the air
monitoring data base. Additionally, the suggestions and assistance
of A. P. Altshuller and J. L. Horowitz proved invaluable to this study.
------- |