EPA-600/5-76-010
September 1976
A COMPUTER SIMULATION MODEL
FOR ANALYZING MOBILE SOURCE
AIR POLLUTION CONTROL STRATEGIES
by
MATHTECH, Inc.
P. O. Box 2392
Princeton, New Jersey 08540
68-01-2952
Project Officer
John Jaksch
Criteria and Assessment Branch
Corvallis Environmental Research Laboratory
CorvaHis, Oregon 97330
CORVALLIS ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
Si,
S. t:5\
-------
DISCLAIMER
This report has been reviewed by the Corvallis Environmental
Research Laboratory, U. S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U. S. Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
11
-------
CONTENTS
List of Figures
List of Tables
List of Abbreviations and Symbols
I
II
III
IV
V
VI
VII
VIII
IX
X
Appendix A
Appendix B
Appendix C
Appendix D
Appendix E
Appendix F
Appendix G
Appendix H
Introduction
Summary
Conclusions
Recommendations
Overview of the MATHAIR Model
Application of the Model to Four Urban Regions
Changes in Air Quality Using Alternative
Air Pollution Control Strategies
Estimates of Benefits and Costs for Alternative
Strategies in the Four Study Regions
Sensitivity Analysis
References
Metric Conversion
Description of MATHAIR
Forecasting Aggregate Demand with a
Disaggregated Transportation Model
Dispersion Properties of Mobile Source and
Stationary Source Emissions
Transportation Costs and Benefits Computing
Consumers' Surplus
Input Data
Validation of Joint Probabilities for Work Trips
Predicted by the Transportation Module
MATHAIR User Documentation
IV
v
vii
1
2
4
7
8
18
33
46
61
70
73
74
87
90
95
102
163
169
111
-------
LIST OF FIGURES
Numb e ]
1
2
3
4
5
6
7
8
9
A flow diagram of MATHAIR
A flowchart of the Transportation Module
The Tri-State study region (New York, New Jersey,
Connecticut)
Chicago study region
Example of five increasingly stringent air pollution
control strategies
Strategies selected for simulation with MATHAIR . .
Net benefits of selected strategies in the four study
regions
Page
8
10
19
20
21
22
34
40
52
10 Costs of selected strategies as a function of 1975
automobile stock in the four study regions .... 53
11 Costs of selected strategies as a function of
number of units of stationary sources of
pollution in the four study regions 54
IV
-------
LIST OF TABLES
Number Page
1 Pollutant Concentrations for the Four Study Regions
Assuming No Air Pollution Controls 26
2 Summary of National Ambient Air Quality Standards . . 27
3 Damages Due to Air Pollution in Los Angeles
Assuming No Air Pollution Controls 29
4 Damages Due to Air Pollution in New York
Assuming No Air Pollution Controls 30
5 Damages Due to Air Pollution in Washington, D. C. ,
Assuming No Air Pollution Controls 31
6 Damages Due to Air Pollution in Chicago
Assuming No Air Pollution Controls 32
7 Federal Exhaust Emission Standards and Control
Levels for Light Duty Vehicles 36
8 Air Quality in Los Angeles for Three Selected
Strategies 41
9 Air Quality in New York for Three Selected
Strategies 43
10 Air Quality in Washington, D. C. , for Three
Selected Strategies 43
11 Air Quality in Chicago for Three Selected Strategies . . 44
12 Benefits and Costs for Selected Strategies in
Los Angeles 47
13 Benefits and Costs for Selected Strategies in
New York 48
14 Benefits and Costs for Selected Strategies in
Washington, D. C 49
15 Benefits and Costs for Selected Strategies in Chicago . 51
v
-------
LIST OF TABLES (continued)
Number Page
16 Selected Automobile Use Statistics for the
Study Regions 55
17 Selected Transportation Control Simulation
Outputs for the Four Study Regions 57
18 Selected Transportation Control Simulation
Outputs for New York and Chicago
Disaggregated by Trip Purpose 58
19 Work Trip Mode Choice for the Four Study Regions . . 59
20 Percentage of Automobiles of Given Age Still in
Stock 62
21 Benefits and Costs for Selected Strategies in the
Four Study Regions Assuming an Accelerated
Scrappage Rate 63
22 Air Quality in Los Angeles for Selected Strategies
Using "Worst Case" Automobile Pollutant
Emission Characteristics 66
23 Benefits and Costs for Selected Strategies in
Los Angeles Using "Worst Case" Assumption ... 67
24 Benefits and Costs of Three Transportation Measures
in Los Angeles and New York 69
VI
-------
LIST OF ABBREVIATIONS AND SYMBOLS
C-C-C or C-C
CO
CS
C-S-C or C-S
HC
I/M
MATHAIR
NO
x
O-D
OX
Pb
ppm
SO
SI
S2
S3
S4 (S4. 1, S4.2, S4. 3)
* See Figure 8 for details.
refers to round-trips whose origin and
destination are both in the center of the
city.
carbon monoxide.
consumers' surplus.
refers to round-trips whose origin is the
center of the city and whose destination is
outside the center.
hydrocarbons.
inspection and maintenance of automobiles.
MATHTECH's computer simulation model for
evaluating air pollution control strategies.
nitrogen oxides.
origin and destination.
oxidants. MATHAIR predicts three different
series: OX1, OX2, and OX3.
lead.
parts (of pollutant) per million parts (of air).
baseline or zero control strategy.
*
modest control strategy .
-if
•T»
moderate control strategy .
>'<
moderate control strategy .
*
strict control strategy with three variants .
vn
-------
LIST OF ABBREVIATIONS AND SYMBOLS (continued)
S-C-S or SC
SO
x
S-S-S or S-S
VMT
VSAD
refers to round-trips whose origin is
outside the center of the city and whose
destination is the center of the city.
sulfur oxides.
refers to round-trips whose origin and
destination are both outside the center of
the city.
vehicle miles of travel.
vacuum spark advance disconnect. A
retrofit pollution control device for
automobiles.
Vlll
-------
SECTION I
INTRODUCTION
The 1970 Amendments to the Clean Air Act required the
establishment of specific national ambient air quality standards and a
specific timetable for compliance with these standards. It also mandated
specific, and substantial, reductions in automobile emissions, again by
specified dates.
While this legislation attests to strong political support for air
pollution control, its passage has not ended the debate over air quality
standards. As is evidenced by the recently established more lenient
interim air quality standards, it is clear that there will be continuing
controversy over air pollution standards, and possible continuing changes
in these standards.
Whatever the reason for this continued interest, it is useful to be
able to evaluate alternative air pollution control strategies, in terms of
both their impact on air quality and their dollar benefits and costs.
Because the computations required to perform reasonably accurate
evaluations are sufficiently complicated, a computer simulation model
is needed for this purpose.
MATHTECH has developed a powerful and flexible computer
simulation model, MATHAIR, that can be used to estimate the impact
of strategies for controlling mobile source air pollution in terms of
both the effect on air quality and the associated benefits and costs. To
demonstrate how the model can be used, we applied MATHAIR to
evaluating some selected air pollution control strategies in four urban
areas.
-------
SECTION II
SUMMARY
THE MATHAIR MODEL
The MATHAIR model enables the user to perform experiments
evaluating the impact of air pollution control strategies in different
geographic regions. For each strategy devised, MATHAIR calculates
both the effect on air quality and the dollar benefits and costs associated
•with that strategy for a user-specified geographic region. These calcula-
tions are made relative to a baseline, or zero-control, strategy.
The effect of a pollution control strategy on air quality is
calculated by first projecting the emission sources in a region and
the level of emissions associated with each source. The impact of
the control strategy on the level of emissions from each source is then
calculated. Finally, an air quality model relates the changes in
emission levels to changes in ambient air quality. Our attention is
focused on controlling emissions of carbon monoxide (C)), hydro-
carbons (HC), and oxides of nitrogen (NO ). The principal emission
source considered is automobiles; controfmeasures considered affect
air quality both by reducing emissions holding automobile use constant, '
and by reducing automobile use.
As regards the calculation of benefits and costs, the total cost
for a strategy is taken to be the additional cost over that associated
with the baseline strategy. Thus costs do not, for example, include the
purchase price of automobiles but do include the extra capital and
operating costs of pollution control devices.
In addition to direct costs, we also calculate the implied cost to
travelers of reduced (or enhanced) mobility. The notion here is that
certain strategies may reduce the number of automobile trips, and an
imputed value of these foregone trips should be counted as a strategy
cost.
To calculate the benefits of improved air quality resulting from
air pollution control strategies, the damage due to air pollution is
quantified in monetary terms for each control strategy. The benefit of
a strategy is then calculated by subtracting those damages from the
damages associated with the baseline, zero-control, strategy.
-------
EXPERIMENTS PERFORMED WITH MATHAIR
In order to demonstrate the use of MATHAIR, four large urban
regions were selected for the simulation experiments reported here: Los
Angeles, New York, Washington, D. C. , and Chicago. These regions
provide a spectrum of the air quality problems found in large urban areas.
The air quality and damage results for the four regions, assuming
zero controls, are reported in Section VI.
The simulation strategies analyzed here include (1) measures
requiring the installation of pollution control devices directly on the
source of pollution and (2) inducements to change patterns of transporta-
tion behavior. Examples of the first type include mandated installation
of catalytic converters on automobiles built after 1975 and required carbon
adsorbers for hydrocarbons on all dry-cleaning plants. Examples of the
second type include increasing the cost per mile of automobile travel by
imposing a tax on gasoline and decreasing bus waiting times by operating
additional buses.
Control strategies representing modest, moderate, and strict
degrees of air pollution control were devised for our example experiments
and are discussed in Section VII.
The impacts of the strategies on ambient pollution concentrations,
presented in Section VII, are mixed and thus difficult to summarize.
Even modest controls effect significant air quality improvements. How-
ever, Federal standards for some pollutants cannot be met even with our
most stringent strategies. The benefits and costs of alternative strategies
in the four study regions are presented in Section VIII. A moderate level
of control proved optimal, in the sense of maximizing net benefits, in
Los Angeles and Chicago. Very modest controls were best in New York
while no controls was the optimal choice in Washington.
These results are, of course, dependent on the input data used.
A limited sensitivity analysis, reported in Section IX, showed that
certain inputs had a marked effect on model outcomes.
-------
SECTION III
CONCLUSIONS
We believe that MATHAIR strategy simulations are a practical way
to do policy analysis. MATHAIR provides a powerful, flexible framework
for testing the impact of air pollution control.
Based on the numerical results of the experiments we have con-
ducted with MATHAIR, a certain number of conclusions can be drawn.
GENERAL CONCLUSIONS
1. The first conclusion is suggested both by a priori
considerations and by the numerical results: it is
economically inefficient to impose the same pro-
gram of controls in different regions. Different
regions have different compositions of total
emissions, different meteorological carrying
capacities, and different transportation systems;
and a flexible national policy that encourages
exploitation of these differences could yield sub-
stantial economic benefits.
2. A second conclusion that emerges from our work
is that the costs associated with changes to less
preferred modes of transportation and reduction
in trip frequency can be substantial and should not
be omitted from a cost calculation. To be sure,
there may be social benefits associated with these
changes -- reduced congestion, less noise pollu-
tion, fewer accidents, and conservation of scarce
fossil fuels --as well as the air pollution reduction
benefits that we account for in our calculations.
Nonetheless, our results indicate that these
other benefits may have to be large to offset the
marginal costs of adding transportation controls
to an existing program of moderate controls.
3. MATHAIR air quality predictions show that oxidant
standards cannot be met by 1985 in any study region
even with our most stringent strategy. Our strictest
measures, including transportation controls, are
needed to meet the carbon monoxide standards.
-------
4. Inspection and maintenance of automobiles
appears to be beneficial in all regions.
5. Strategies requiring hard-ware control devices
for automobiles are most successful in regions
where the average automobile is driven
intensively.
6. Transportation c ontrols discouraging use of the
automobile are most successful, in conjunction
with other measures, in regions with the greatest
existing availability of common transportation.
7. A program consisting of transportation controls
only appears to be beneficial though of limited
impact.
SENSITIVITY ANALYSIS
We performed a limited sensitivity analysis of MATHAIR to
determine which inputs have a critical effect on strategy outcomes. The
first experiment, assuming an accelerated rate of removing automobiles
from the stock, had a minimal effect on policy implications.
The second experiment simulated a set of "worst case" assump-
tions for Los Angeles: increased control device costs, low baseline
dollar damages due to air pollution, and low estimates of the pollutant
reduction efficiencies of automobile control devices. The combined impact
of these assumptions was a dramatic decrease in net benefits.
The final sensitivity experiment simulated three versions of the
strictest control strategy, each containing a different transporation
control measure. The marginal cost of the strategy exceeded the marginal
benefit in all cases.
SOME QUALIFICATIONS
Our results and conclusions are based on an amalgam of models,
published data, and assumptions. To the greatest extent possible, we
have relied upon inputs and methods that EPA has accepted for purposes
of preparing implementation plans and transportation control plans.
Nonetheless, it has been necessary to make many independent assump-
tions. For example, we have assumed that benefits of pollution control
vary linearly with reductions in ambient concentrations of pollutants.
Our results would be quite a bit different if we had assumed that benefits
vary as (say) the square or the cube of concentration reductions. Many
strategies that fail to produce positive net benefits under the linear benefits
assumption would pass under nonlinear assumptions.
A portion of the data we have used to test the model was estimated
by us. Even the data we have from public sources are subject to con-
-------
siderable uncertaintly. For example, recent studies suggest that the
emissions control equipment is considerably less effective than is reflected
in the EPA data -we have used in our calculations. * Other things being
equal, this factor tends to cause our calculations to overstate the net
benefits of vehicular emissions control strategies.
In consequence, it must be concluded that the estimated costs
and benefits associated with alternative mobile source pollution control
strategies are very sensitive to the assumptions one makes. It -will be
possible to give more precise estimates of benefits and costs only as the
uncertainty about inputs is reduced by further research on the basic
scientific and technical aspects of air pollution control and by improving
and updating data estimates.
U. S. Environmental Protection Agency, Compilation of Air Pollutant
Emission Factors, AP42 (April 1973), and Supplement'No. 2 (September
1973) vs. the preliminary edition ©f Supplement No. 5 (April 1975).
-------
SECTION IV
RECOMMENDATIONS
In the course of developing MATHAIR, we used it to evaluate
dozens of air pollution control strategies and found it a powerful and
flexible analytic framework for policy analysis. We have only begun to
explore the kinds of strategies that can be tested with MATHAIR, and
-while our results demonstrate the feasibility and potential usefulness of
an integrated model for this purpose, we are also aware of the limitations
of the approach and the weaknesses in our data base. Our recommendations
are itemized below:
A model like MATHAIR should be widely used
to evaluate likely policy impacts in urban
areas prior to the adoption of control measures
at the state, regional, or Federal levels.
More recent and more carefully estimated data
are needed for use with MATHAIR than the data
used for the test runs. The first priority for
further work using the MATHAIR model is to
improve the data base.
A particular data need is for improving para-
meter estimates for the Transportation Module
of MATHAIR.
Module-by-module validation of MATHAIR
predictions is needed.
More extensive analysis of MATHAIR's
sensitivity to changes in the data should be
performed.
As more research is done on damage functions,
this information should be incorporated into the
Benefits Module of MATHAIR.
-------
SECTION V
OVERVIEW OF THE MATHAIR MODEL
INTRODUCTION
This section outlines MATHAIR, the model developed by
MATHTECH to estimate the costs and benefits of alternative strategies
for improving air quality. This section is intended to give the reader
an overview of the model; a more extensive discussion of MATHAIR is
presented in Appendix B.
A flow diagram of MATHAIR is shown in Figure 1.
Transportation
Module
(forecasta vehicle
mites traveled by
mode)
Automobile Stock
Module
(calculates age com-
position of stock of
juitomgbil_e»)
Emissions Module
(forecasts emissiona,
subject to installation
of control device*)
Cost Module
Air Quality Module
(foreraata ambient
pollutant concen-
trations)
Benefits Module
(forecasts reduction
in pollution damage)
Beneftti
Figure 1. A flow diagram of MATHAIR.
MATHAIR consists of the following six modules:
1. Automobile Stock Module, which computes
the composition of the automobile stock.
-------
2. Transportation Module, which forecasts vehicle
miles of travel for each mode of transportation.
3. Emissions Module, which calculates pollution
emissions for mobile sources and stationary
sources subject to the installation of control
devices.
4. Air Quality Module, which calculates the
ambient concentrations of the pollutants.
5. Benefits Module, which calculates dollar losses
due to pollution damage and the benefits (i. e. ,
reduction in losses) with respect to a baseline
case.
6. Cost Module, which calculates the cost of
implementing the control strategy.
All region-dependent information is contained in the input data so that
no adjustments to MATHAIR are required to run the model for different
urban areas. Input files can, of course, be changed or updated.
AUTOMOBILE STOCK MODULE
The inputs for this module are historical new car registrations,
predicted annual growth rate for new car registrations, and a schedule
of scrappage rates which specifies the percent of automobiles of each
age (one year old, two years old, etc. ) that are removed from the automobile
stock. This information is used to calculate the number of automobiles
of each age on the road in each simulation year.
TRANSPORTATION MODULE
A flow diagram of the Transportation Module is given in
Figure 2. It predicts trips by mode (automobile, bus, or rail), purpose
(work, shop, other), and spatial pattern (center-center, suburb-suburb,
center-suburb, suburb-center).
inputs:
For each type of trip, the module requires the following
average time and out-of-pocket costs
number of "potential" trips
average trip distance
Out-of-pocket costs are the costs incurred for an additional trip. For
automobile trips this includes the cost of gasoline, parking, and tolls
but does not, for example, include payments for the automobile or the
cost of insurance.
-------
JO
_ u «
J13.2
our*
a
I
». 2
•j a
.O w >
c o. a
C •** t.
•o
o
c
o
1 1
«
V
m*
Z?
a
.So
°£
*^ ^
j; o
c «
0 >
o2
ftH
,
r!
'
.e-§
»H U
CO
C
ti
f-t
E-"
0)
ni
u
0).
O O
u
5.2-0
O I. O
HHO
-------
• average vehicle occupancy
• base year vehicle miles traveled
These inputs are used to compute vehicle miles traveled by each
mode. Based upon time and out-of-pocket costs associated with each kind
of trip, the model predicts mode-choice probabilities and trip frequency
probabilities whose product is the joint probability of taking a trip by a
given mode between a given origin and destination. The joint probability
is then multiplied by the number of potential trips between that origin
and destination to produce an estimate of passenger trips, by mode.
Then trip distance and vehicle occupancy data are used to calculate
vehicle miles of travel (VMT). Region-wide base year VMT predictions
are calibrated to base year data.
The user has the option of bypassing the transportation model
and reading in base year VMT and an annual growth rate.
Alternative Structures Considered
Our principal requirement for the transportation model was that
demand be sensitive to the policy variables: the time and out-of-pocket
costs of taking a trip. This requirement ruled out the use of Urban
Transportation Planning models, like the travel demand submodel of
TASSIM. * In these models the number of trips demanded is deter-
mined by socio-economic factors in the zone of origin, not by the level
of service. A model like TASSIM cannot evaluate the effect on demand
of changed trip costs.
We also considered using the transportation model developed by
RAND for the San Diego Clean Air Project and subsequently recalibrated
for Los Angeles. ** We decided against this for several reasons. For
example, no user documentation had been prepared for the model.
Another drawback was that no spatial disaggregation (i. e. , between city
center and the outlying area) is possible in the RAND model since trips
are characterized only by their origin and their distance.
Ingram, G. K. , and G. R. Fauth, TASSIM; A Transportation and
Air Shed Simulation Model, U.S. Department of Transportation
(May 1974).
Goeller, B.F. et al, San Diego Clean Air Project, The RAND
Corporation, R-1362-SD, (December 1973). Also, Mikolowsky, W. T.
et al, The Regional Impacts of Near-Term Transporation Alternatives:
A Case Study of Los Angeles, The RAND Corporation, R-1524-SCAG
(June 1974).
11
-------
It should be noted that the MATHAIR Transportation Module is
a demand model, with the number of miles traveled for each mode
assumed equal to the amount demanded; no allowance is made for supply
constraints. In the case of automobile travel, this limitation presents
no difficulty, since none of the policies considered here increases
automobile use. Therefore, the existing stock of automobiles, highways,
etc. , should be adequate to meet demand for predicted automobile travel.
As regards buses, it is implicitly assumed in the present version of
MATHAIR that the supply of buses will expand to meet increased demand
for bus travel. This assumption is less realistic but does enable us to
investigate what would happen if there were no restrictions on the availability
of mass transportation.
Finally, it should be noted that while the Transportation Module
has the structure of a disaggregated behavioral model, * we have used
aggregated input data with a disaggregated model thus introducing some
bias into our predictions. The nature of the bias is discussed in
Appendix C.
EMISSIONS MODULE
The Emissions Module predicts total annual emissions from the
stock of automobiles and other mobile sources and from stationary
sources. Predictions are made for the so-called "mobile source"
pollutants carbon monoxide (CO), hydrocarbons (HC), oxides of nitrogen
(NOX), sulfur oxides (SOX) ** and lead particulates (Pb). *** The module
also simulates the effects of installing air pollution control devices.
Automobile Emissions
The emissions of an individual automobile are assumed to be
proportional to miles traveled (with an adjustment for speed). The
Emissions Module calculates the number of miles driven by automobiles
of each age for each simulation year. This calculation requires three
inputs: the age composition of the stock of automobiles (provided by the
Automobile Stock Module), automobile vehicle miles traveled (provided
by the Transportation Module), and the percent of all automobile miles
driven by automobiles of each age (a user-provided input to the Emis-
sions Module).
Disaggregated behavioral demand models are described in detail
in Domencich, T.A. and D. McFadden, A Disaggregated Behavioral
Model of Urban Travel Demand, Cambridge, Mass. , Charles River
Associates, Inc. (March 1972).
All oxides of sulfur are treated as a single pollutant in MATHAIR
since there is not yet sufficient quantitative information to
separately model the quantities and effects of SO? and SOo.
The lead content of fuel is a user-specified variable.
12
-------
The required input data include automobile emissions, in grams
per mile traveled, for each of the mobile source pollutants. These
emission factors are available for new automobiles (of different vintages)
driven at a standard speed. "Deterioration factors," which specify
the rate at which emissions increase as automobiles age, are then used
to compute emissions for old automobiles. Finally, estimated average
automobile speeds and speed correction factors are needed to adjust
emission factors for the difference between standard speeds and the
speed assumed in the simulation.
A final input required by the automobile section of the Emissions
Module are estimated efficiencies of various pollution control devices. *
The installation of control devices is simulated as a percent reduction in
automobile emissions.
Other Mobile Sources
Bus emissions are calculated as the product of a bus emissions
factor and annual miles traveled by bus. An inventory of emissions
from other mobile sources (e.g. , motorcycles) may also be provided
as input at the user's option. In the current version of MATHAIR, no
air pollution control devices can be installed on mobile sources other
than automobiles.
Stationary Sources
An inventory of stationary air pollution sources must be input
to the module. The number of "average-sized" units and emissions
of each pollutant for an average unit must be provided for each type
of source. Emissions are calculated as the product of the number of
units and emissions per unit. The installation of control devices is
simulated as a percent reduction in predicted emissions.
AIR QUALITY MODULE
We selected a simple linear rollback model for predicting air
quality. The formulation is based on the assumption that ambient
pollutant concentrations are proportional to total emissions. Although
this approach fails to exploit information about spatial and temporal
diffusion of pollution, the rollback model seems an adequate approxi-
mation for assessing broad classes of strategies based on inter-regional
comparisons. The inputs to the module are rollback coefficients, which
are exogenously estimated for each pollutant using data on pollution
emissions and concentrations for some base year, and estimates of
emissions for each simulation year (provided by the Emissions Module).
* Up to one hundred control devices can be defined without altering the
computer program.
13
-------
Air quality, expressed in terms of pollutants on a parts per million
parts of air basis, * is calculated as the product of the rollback
coefficient and predicted emissions.
It is possible to modify the simple rollback model to allow for
different coefficients for mobile and stationary source emissions. **
If the user selects this option, predicted air quality is a sum with two
components: the product of the mobile source coefficient with mobile
source emissions and the product of the stationary source coefficient
with stationary source emissions.
BENEFITS MODULE
The Benefits Module calculates the monetary value of reduced
damage and damage avoidance costs from air pollution control
measures. In the simulations reported here, three damage categories
are used: health, vegetation, and materials. However, the user may
specify up to eight damage categories.
The module predicts the dollar value of air pollution damage, for
a given pollutant and damage category, by evaluating damage functions
of the following form:
§(E-T)y if E > T
0 otherwise
where D = dollar value of damages
E = air quality
S = calibration constant
y = user-specified exponent
T = air quality threshold below -which no damage
is assumed to occur.
The Benefits Module calculates the calibration coefficient in the follow-
ing manner: the user specifies values of monetary damages, D , and
pollutant level, Eo , for a base year, as well as estimates of y and T. ***
c This unit is known as ppm. Lead particulate is measured in
;'* A discussion of why different coefficients might be appropriate is
provided in Appendix D.
:** If the latter two parmeters are not specified, default values of
y = 1 and T = 0 are used.
14
-------
One calibration coefficient 8 is then computed for each damage category
for each pollutant according to the formula
5 = Do/(Eo - T^.
The estimated damages are summed over all damage categories
for each pollutant, and the discounted present value is calculated. The
benefit attributable to an air pollution control strategy is the reduction
in pollution damage with respect to the damage associated with zero
controls. This reduction can be estimated as the difference between the
dollar value of the resulting pollution damage and the dollar value of
pollution damage in the case of no controls. The two dollar values are
calculated by MATHAIR in separate simulations, and the user must
perform the subtraction manually. For ease of exposition, however, we
discuss the Benefits Module as if it performed the subtraction and
calculated for each strategy the dollar value of benefits.
Finally, the user has the option of replacing the Benefits Module
with his own FORTRAN subroutine.
COST MODULE
The purpose of the Cost Module * is to calculate the total costs
associated with implementing a strategy. These include the costs
associated with air pollution control devices and the costs of changes in
travelers' mobility.
To calculate the costs associated with air pollution control
devices, the Cost Module requires as input unit costs and expected
economic lifetimes for pollution control devices and the number of each
device installed (from the Emissions Module). Unit costs include the
initial cost of the device and annual maintenance and operating costs.
The module calculates the present discounted value of the annualized
initial cost of automobile and stationary source control devices. It also
calculates the present discounted value of annual operating and mainten-
ance costs and the cost of changed gasoline consumption due to automobile
control devices.
A second type of cost is incurred when a traveler abandons his
preferred mode of transportation and either switches to a less convenient
mode or foregoes the trip altogether. Analogously, a benefit (negative
cost) should be counted when a strategy induces a trip that would not
otherwise have been taken. Using the number and prices of different
types of trips (obtained from the Transportation Module), the Cost
Module calculates the cost to travelers of reduced (or enhanced) mobility
Cost calculations are distributed throughout MATHAIR. It is, however,
convenient to think of a single module that performs all the cost calcu-
lations.
15
-------
due to transportation control measures. * Calculation of these costs
is a unique feature of our analysis. ##
Certain costs attributable to an air pollution control strategy are
in fact changes in cost that can only be calculated with respect to a
baseline. This is the case, for example, in calculating the cost of
reduced mobility. To calculate these costs, the user must do some manual
processing of MATHAIR outputs, i. e. , subtract the quantities and costs of
one MATHAIR output from those of another. ***
As with the Benefits Module, we discuss the Cost Module as if it
performed the subtraction and calculated the total costs incurred by a
control strategy.
MATHAIR OUTPUTS
The output of a MATHAIR simulation contains the following
categories of information:
• Predicted emissions (tons/day) of CO, HC,
NOX, SOX and Pb from automobiles, from
all mobile sources, and from all stationary
sources, respectively.
• Predicted ambient concentrations (ppm) of ,
CO, NOX, oxidants, and SOX and of Pb (/zg/m ).
• Discounted costs and benefits for the strategy
which has been simulated.
These outputs are generated for each year in the simulation horizon. A
simulation output also contains the following information:
• Discounted present value of benefits and costs
over the simulation horizon.
• Detailed intermediate output of the Transportation
Module produced at the user's option. This
-f These costs are described by the economist's paradigm of consumers'
surplus which is discussed in Appendix E.
** The RAND model [Goeller, B.F. et al, San Diego Clean Air Project,
The RAND Corporation, R-1362-SD, (December 1973) and Mikolowsky,
W. T. et al, The Regional Impacts of Near-Term Transportation Alter-
tives: A Case Study of Los Angeles, The RAND Corporation, R-1524-
SCAG (June 1974)] caclulates the number of foregone trips, but it does
not associate a dollar value with them. Nor does it analyze enhanced
mobility (i. e. , induced trips).
*##The calculations are described in Appendix E.
16
-------
output includes the number of passenger trips,
passenger miles, and vehicle miles for each
type of trip as well as mode-split and trip
frequency probabilities.
The benefit-cost framework of MATHAIR provides a single
common denominator for all outcomes: dollars. In such a framework,
the results of a strategy simulation can be represented by the dollar
value of all of the costs involved and the dollar value of the benefits due
to damage reduction. Net benefits can be calculated as one criterion
for analyzing strategies. The marginal benefits and marginal costs of
going from one strategy to a stricter one, can also be calculated. *
The RAND case study of Los Angeles [Mikolowsky, W.T. et al, The
Regional Impacts of Near-Term Transportation Alternatives; A
Case Study of Los Angeles, The RAND Corporation, R-1524-SCAG
(June 1974)], for example, reports a vector of impacts for each
strategy analyzed in the study. While such outputs are of course
useful, the dollar common denominator has the advantage of pro-
viding the decision-maker with a simple strategy characterization.
This is especially important when a large number of strategies are
being compared.
17
-------
SECTION VI
APPLICATION OF THE MODEL TO FOUR URBAN REGIONS
DESCRIPTION OF THE STUDY REGIONS AND DATA REQUIREMENTS
Description of the Study Regions
In order to test MATHAIR we have prepared input data for the
urban regions containing the cities of Los Angeles, New York, Washington,
D. C., and Chicago. These four cities represent a spectrum of the air
pollution problems facing urban areas in the United States.
In each case we selected a study region that was larger than the
city or the SMSA so that we could simulate travel within the urban area.
The particular region boundaries were generally selected because
pollutant emission and concentration data were available for those boundaries.
Maps of the study regions are shown in Figures 3 through 6. In all cases
we have divided the area into two parts which, for convenience, we call
the center and the suburbs.
Data Needs
Six categories of data, corresponding to the six MATHAIR modules,
are required for a MATHAIR simulation,, The data that were used to
test the model are presented in Appendix F. The six modules are listed
below:
Automobile Stock Module
Transportation Module
Emissions Module
Air Quality Module
Benefits Module
Cost Module
In a number of cases, more recent data are now available than
the data that were used to test the model. This is because work on this
study began late in 1974. If additional MATHAIR simulations are run
for the study regions, the data used as input to the parameter estimations
and to the model should be updated.
18
-------
Figure 3. Los Angeles study region.
Source: Prediction of the Effects of Transportation Controls on Air Quality
in Major Metropolitan Areas, U. S. Environmental Protection
Agency, APTD-1363, November 1972, pp. 3-6.
Note: The entire area shown is the study region. The cities proper of
Los Angeles and Long Beach, approximately enclosed in dashed
lines, are the region's "center" for purposes of the Transportation
Module.
19
-------
Figure 4. The Tri-State study region (New York, New Jersey, Connecticut).
Source: Vehicle Miles of Travel on Major Roadways; 1970, Tri-State
Regional Planning Commission, November 1973, p. 2.
Note: The entire area shown is the study region. New York City,
approximately enclosed in dashed lines, is the regions's
"center" for purposes of the Transportation Module.
20
-------
t
/
Figure 5. Washington, D. C., study region.
Source: Prediction of the Effects of Transportation Controls on Air Quality
in Major Metropolitan Areas, U. S. Environmental Protection
Agency, APTD-1363, November 1972, pp. 4-5.
Note: The entire area shown is the study region. The District of
Columbia, enclosed in dashed lines, is the region's "center" for
purposes of the Transportation Module.
21
-------
Figure 6. Chicago Study Region
Source: Prediction of the Effects of Transportation Controls on Air
Quality in Major Metropolitan Areas, U. S. Environmental
Protection Agency, APTD-1363, November 1972, pp. 3-7.
Note: The entire area shown is the study region. The city proper
of Chicago, approximately enclosed in dashed lines, is the
region's "center" for purposes of the Transportation Module.
22
-------
In many cases the data needed as input to MATHAIR were available
for the center only, for a metropolitan area of different boundaries than
our study area (usually the SMSA), or could not be located at all. In these
cases, -we estimated the corresponding data for the study area.
Transportation Module--
Estimating a disaggregate behavioral demand model like our
Transportation Module requires survey-type information about travel
alternatives and decisions. It -was not possible within the scope of this
project for us to estimate equations for the four study regions.
The alternative we adopted was to use parameters estimated for
Pittsburgh by other investigators.* We recalibrated the constant terms so
that the Transportation Module's base year predictions for the study regions
would fit observed data. The nature of the model offers some justification
for the method we adopted since behavior of individuals faced with the
same price and service characteristics of transportation alternatives can
be assumed to be relatively independent of where the individual lives.
We undertook a validation of some of the output series of the
transportation model against independent data series. The two series
are generally acceptably close. Comparisons are presented in Appendix G.
Benefits Module--
The MATHAIR Benefits Module requires for each study region
base year air quality measurements for each pollutant and the dollar
value of each damage category by pollutant for the same base year. These
data are used to calibrate the damage functions.
To our knowledge, estimates of the dollar value of pollution
damage have been made only on a nationwide basis, and independent
estimates are not in close agreement. Therefore, two problems had
to be faced:
• select nationwide damage estimates, and
• disaggregate these estimates into damage estimates
for specified urban regions.
The national pollution damage estimates used to test the model
are the estimates of Barrett and Waddell, extended by Babcock and Nagda. **
* A Disaggregated Behavioral Model of Urban Travel Demand, Charles
River Associates (March 1972).
:<# All reported in Justus, C. G. , J. R. Williams, and J. D. Clement,
Economic Costs of Air Pollution Damage. STAR, Inc. (May 1973).
The 1971 Barrett and Waddell estimates (using 1968 data) were extended
by Babcock and Nagda in a 1973 study to include the effects on health
and property of the mobile source pollutants.
23
-------
Considerably lower estimates were made by Justus et al. * We have used
these lower estimates in one of the scenarios simulated with MATHAIR in
order to test the sensitivity of the model.
The methodology we have used for disaggregating national
damage estimates to urban area estimates of damage by pollutant and
damage category was developed by Justus et al. ** They calculated
a "multiplier" for each of 65 SMSA's for five pollutants and seven
damage effects. The product of the multiplier and the national loss
estimate for all damage due to air pollution, is an estimate of the loss
in one SMSA for one category of damage by one pollutant. The multipliers
are determined by three factors:
• a severity factor (Spe) -which is a measure of the
"toxicity" of pollutant p for damage effect e,
• an air quality exceedance factor (Amp) which
measures the extent to which the air quality standard
for pollutant p was exceeded in SMSAm in the
base year, and
• an exposure factor (Eme) -which measures the level
of exposure in SMSAm to possible damage from
effect e.
Total damage is assumed to be exhaustively distributed among
the 65 SMSA's. The multiplier for an SMSA is the product of the three
factors, normalized so that the sum of the multipliers is 1.00.
For each of our study regions we used the Justus multipliers
adjusted for the difference in boundaries between the SMSA's and our
study regions.*** In the test runs we have made, only three damage
categories are considered: nealth, vegetation, and materials. Other
categories of damage, like noise and congestion, which are also affected by
the control strategies are not considered because they are even more
difficult to quantify. The omission of the damages associated with these
categories means, other things being equal, that the damages (and the
benefits) calculated by MATHAIR should be regarded as lower bounds.
ESTIMATES OF AIR QUALITY AND MONETARY VALUE OF DAMAGES
ASSUMING NO CONTROLS
We now report MATHAIR forecasts of air quality and dollar value
of damages in the four study regions for the period 1975-1984 assuming
no air pollution controls.
* Ibid..
** Ibid.
The study region multiplier was calculated as the sum of the multipliers
for all SMSA's included in the study region. Where an SMSA was
partly included, judgmental decisions were made.
24
-------
Air Quality
The air quality predictions are shown in Table 1. The scenario
assumes no air pollution controls; that is, post-1967 automobiles in
New York, Washington, D. C. , and Chicago have 1967 vintage characteristics
and post-1965 Los Angeles automobiles have 1965 characteristics. * The
existing level of control of stationary sources is assumed.
Table 1 shows concentrations of CO, OX, and NO to be increasing
over time. This increase is due solely to increased demand for trips
as the population grows over the simulation horizon since stationary
source emissions are assumed to remain constant. Concentrations of
SO are also increasing, but very slowly, since mobile sources emit
refatively small quantities of this pollutant. Lead concentrations are seen
to be decreasing because we have assumed a gradual phasing out of lead
from gasoline according to the schedule shown in Table F-24 of
Appendix F.
The national ambient air quality standards are shown in Table 2
so that we can compare the MATHAIR outputs with the standards. **
The standards for sulfur oxides are met in all four regions over the entire
simulation horizon. This unrealistic prediction is due in part to the choice
of a zero growth rate for stationary source emissions over the simulation
horizon. In Washington, D.C., the standards for nitrogen oxides
are also met. The standards for carbon monoxide and for oxidants are
greatly exceeded in all areas.
Damages
MATHAIR forecasts of the dollar value of damages due to air
pollution (1975 present value in 1973 dollars) for the four study regions are
shown in lables 3 through 6.
The loss due to sulfur oxides is shown to be zero for Los Angeles
and for Washington, D. C. The explanation of this anomaly is that, in
* In California, post-1965 automobiles were subject to controls. In
the rest of the nation, post-1967 automobiles were the first to be
controlled.
** The averaging times for measured pollutant concentrations used in
calibrating rollback parameters for MATHAIR are shown below. The
relevant standard is shown in the third column.
Pollutant Averaging Time Standard
CO 8 hour 9 ppm
NO annual arithmetic mean 0. 05 ppm
Oxidants 1 hour 0. 08 ppm
SOX annual arithmetic mean 0. 03 ppm
Pb 24 hour
25
-------
TABLE 1.
POLLUTANT CONCENTRATIONS FOR THE FOUR STUDY REGIONS
ASSUMING NO AIR POLLUTION CONTROLS 3
Region
Los Angeles
'New York
Washington, D. C.
Chicago
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Carbon
Monoxide
(CO)
54. 7388
55.7445
56.7804
57.8472
58.9462
60.0781
61.2439
62. 4447
63.6815
64.9555
30.6820
30.8631
31.0460
31.2307
31.4173
31.6057
31.7960
31.9382
32. 1824
32.3784
42.5291
43. 1936
43.8781
44.5830
45. 3091
46.0570
46.8273
47.6208
48.4380
49.2798
34.2627
34.7956
35. 3444
35.9098
36.4920
37.0918
37.7095
38. 3458
39.0012
39.6762
OX1
0.5096
0.5169
0.5246
0.5326
0.5409
0. 5495
0.5585
0.5677
0.5773
0.5872
0.2558
0.2569
0.2580
0.2592
0.2604
0.2617
0.2630
0.2643
0.2657
0.2671
0. 3497
0.3557
0.3618
0.3681
0. 3746
0.3812
0.3879
0. 3946
0.4015
0.4086
0.2464
0.2491
0.2519
0.2549
0.2579
0.2610
0.2642
0.2675
0.2710
0.2745
Oxidants M
OX2 1 OX3
0.4720
0.4786
0.4855
0.4926
0. 5000
0.5077
0. 5156
0.5238
0. 5323
0.5410
0.2197
0.2205
0.2213
0.2221
0.2229
0.2238
0.2247
0.2256
0.2265
0. 2274
0. 3403
0. 3454
0. 3506
0. 3560
0. 3615
0. 3671
0. 3729
0. 3787
0. 3847
0. 3907
0.2357
0.2382
0.2408
0.2434
0.2462
0.2490
0.2519
0.2549
0.2580
0.2611
0.4483
0.4544
0.4607
0.4672
0, 4740
0.4811
0.4883
0.4958
0. 5036
0.5116
0.2140
0.2147
0.2151
0.2161
0.2169
0.2176
0.2184
8.2192
0.2200
0.2208
0. 3348
0. 3393
0. 3440
0. 3488
0. 3538
0. 3588
0. 3639
0. 3692
0. 3745
0. 3800
0.2246
0.2268
0.2291
0.2315
0. 2340
0.2365
0. 2392
0. 2419
Nitrogen
Dioxide a
(NOX)
0.0637
0.0645
0.0653
0.0662
0.0671
0.0680
0.0689
0.0699
0.0709
0.0720
0.0797
0.0798
0. 0800
0.0802
0.0803
0.0805
0.0807
0.0809
0.0811
0.0812
0.0086
0.0087
0.0088
0.0089
0.0090
0.0091
0.0092
0.0093
0. 0094
0.0095
0.0916
0.0974
0.0932
0.0941
0.0950
0.0959
0.0968
0. 0978
0.2446 ! 0.0983
0.2475
0.0998
Sulfur
Oxides
(SOX)
0.0158
0.0158
0.0158
0.0159
0.0159
0.0159
0.0159
0.0160
0.0160
0.0161
0. 0109
0.0109
0.0109
0.010°
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0155
0.0155
0.0155
0.0155
0.0155
0.0155
0. 0155
0.0156
0.0156
0.0156
0.0217
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
Lead c/
(Pb)
31.97
22.02
13.88
10. 13
7.14
4.93
3.39
2.28
1.54
1.02
6.69
4.71
3.05
2.31
1.70
1.20
0.82
0.55
0.36
0.24
21.04
14.60
9.42
7. 18
5.30
3.80
2.61
1.74
1. 18
0.79
9.54
6.75
4.42
3. 39
2.51
l.SO
1. 24
0. 84
0.57
0. 38
Source: MATHAIR outputs.
a/ All units are reported in parts per million (ppml except lead particulate which is reported in
by Three alternative prediction equations were used for oxidant concentrations. . The equations are listed
below:
OX1 =
0X2 = y(HCs +
OX3 -
NO
XS
where a, ft, Y, and 8 are calibration constants
HC - = hydrocarbons
NO , - nitrogen oxides
Subscripts: S - Stationary source and M = mobile source
OX1 is the estimate used by the air quality module.
c/ Lead concentrations arc seen to be decreasing because a achedule for phasing out the lead content
gasoline was assumed.
26
-------
TABLE 2. SUMMARY OF NATIONAL AMBIENT AIR QUALITY
STANDARDS a/
Pollutant
Particulate
matter
Sulfur oxides
Carbon
monoxide
Nitrogen
dioxide
Photochemical
oxidants
Averaging
Time
Annual (Geometric
mean) , .
24-hour-7
Annual (Arith-
metic mean)
24-hour b/
W
3-hour-^7
8-hour -/
1 -hour —
Annual (Arith-
metic mean)
1 -hour —
Primary
Standards
75 Mg/m3
3
260 ng/m
80 ptg/m3
(0. 03 ppm)
365 Mg/m3
(0. 14 ppm)
-
3
10 mg/m
(9 ppm)
3
40 mg/m
(35 ppm)
100 /ig/m
(0. 05 ppm)
3
160 fig/rci
(0. 08 ppm)
Secondary
Standards
3
60 /xg/m
3
150 ng/m
-
•3
1300 Mg/m
(0. 5 ppm)
(Same as
primary)
(Same as
primary)
(Same as
primary)
Source: Air Quality Data - 1972 Annual Statistics. U. S. Environmental
Protection Agency (March 1974), p. iv.
a/ The air quality standards and a description of the reference methods
were published on April 30, 1971, in 42 CFR 410, recodified to 40 CRF 50
on November 25, 1972,
b_/ Not to be exceeded more than once per year.
27
-------
the Justus study, these two cities are assigned air quality exceedance
factors of zero for the base year for SOX. Given the method of calculation,
they will therefore always show a zero loss! These are obviously
underestimates of the actual loss, and the allocation method furthermore
forces the other cities to show compensating overestimated portions of
the national estimates. While this is a serious problem with the procedure
for disaggregating damages, it has a minimal effect on the benefit-cost
analysis we will perform in Section VIII. *
Benefits are defined as the difference between damage loss under a
given scenario and damage loss in the baseline case, and the difference
forecasted in our test runs will always be zero. However, this is not
a serious error. Since both quantities are in fact small positive numbers,
the actual difference will be close to the predicted value of zero.
28
-------
TABLE 3. DAMAGES DUE TO AIR POLLUTION IN LOS ANGELES
ASSUMING NO AIR POLLUTION CONTROLS (1975
PRESENT VALUE IN MILLIONS OF 1973 DOLLARS)
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Damage
Category
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Carbon
Monoxide
(CO)
23.5
0.0
61.8
22.2
0. 0
59.5
21.0
0. 0
57.2
19.8
0.0
55.0
18.7
0.0
53.0
17.7
0. 0
51.0
16.7
0.0
49.1
15. 8
0.0
47.3
15.0
0.0
45.6
14.7
0.0
43.9
Oxidants
(OX)
152.5
35.9
486.6
143.5
33. 3
467.2
134.9
30.9
446.8
127. 1
28.6
428.6
119.7
26.6
411. 1
112. 8
24.8
394.5
106.4
22.9
378.7
100.2
21. 3
363.6
94.5
19.9
349.2
89.1
18.5
335.5
Nitrogen
Oxides
(NOX)
80. 3
6.9
111.5
75.5
6.4
106.6
70.9
5.9
102.0
66.6
5.5
97.7
62. 6
5.1
93.5
58.8
4.6
89.5
55.4
4. 3
85.7
52.0
4.0
82. 1
48.9
3.8
78. 6
46.0
3.4
75.4
Sulfur
Oxides
(SOX)
0.0
0.0
0. 0
0. 0
0.0
0. 0
0.0
0.0
0.0
0. 0
0.0
0. 0
0.0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0.0
0.0
0.0
Lead
(Pb)
0.2
0.0
0. 3
0.1
0.0
0.2
0.0
0. 0
0. 1
0. 0
0. 0
0.0
0. 0
0. 0
0. 0
0. 0
0.0
0. 0
0. 0
0.0
0. 0
0. 0
0.0
0. 0
0. 0
0. 0
0. 0
0.0
0. 0
0.0
Source: MATHAIR outputs.
29
-------
TABLE 4. DAMAGES DUE TO AIR POLLUTION IN NEW YORK
ASSUMING NO AIR POLLUTION CONTROLS (1975
PRESENT VALUE IN MILLIONS OF 1973 DOLLARS)
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Damage
Category
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Carbon
Monoxide
(CO)
2.5
0.0
87.6
2.3
0.0
82.6
2.2
0.0
78.0
1.9
0.0
73.6
1.8
0.0
69.5
1.7
0.0
65.5
1.5
0.0
61.8
1.4
0.0
58.4
1.3
0.0
55. 1
1.2
0.0
52.0
Oxidants
(OX)
21. 3
0.8
570.2
19.7
0. 8
537.2
18.2
0. 6
506.2
16.9
0. 6
477. 1
15. 6
0.5
449.8
14.4
0.5
424. 0
13.4
0.4
399.8
12.4
0.4
377. 0
11.4
0.4
355.5
10.6
0.3
335.3
Nitrogen
Oxides
(NOX)
15. 6
1. 1
364. 2
14.4
1. 1
342. 5
13.2
1.0
322.0
12, 3
0.9
302. 8
11. 3
0.8
284.7
10.4
0. 8
267.6
9.6
0. 6
251.6
8.9
0.6
236. 6
8.2
0.5
222.5
7.5
0.5
209.2
Sulfur
Oxides
(SOX)
204.4
32.8
2386.6
188. 3
30.2
2240.2
173.4
27.7
2102.3
159.7
25.5
1972.0
147. 1
23.5
1850. 3
135. 5
21.5
1736. 1
124.8
19.82
1629.5
114.9
18.2
1529.3
105.9
16.7
1434.6
97.6
15.4
1346. 3
Lead
(Pb)
0.2
0.0
5.7
0. 1
0. 0
3.8
0. 0
0. 0
2. 3
0.0
0. 0
1.6
0.0
0.0
1. 1
0. 0
0. 0
0. 8
0. 0
0.0
0.4
0. 0
0.0
0.3
0. 0
0. 0
0.2
0.0
0.0
0. 1
Source: MATHAIR outputs.
30
-------
TABLE 5. DAMAGES DUE TO AIR POLLUTION IN WASHINGTON,
D. C. ASSUMING NO AIR POLLUTION CONTROLS (1975
PRESENT VALUE IN MILLIONS OF 1973 DOLLARS)
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Damage
Category
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Carbon
Monoxide
(CO)
0.0
0.0
23.8
0.5
0.0
23.0
0.4
0.0
22.3
0.4
0.0
21.5
0.4
0.0
20.8
0.4
0.0
20. 1
0.3
0.0
19.4
0.3
0.0
18.8
0.3
0.0
18. 3
0.3
0.0
17.7
Oxidants
(OX)
1.7
1.0
91.2
1.6
1.0
88.3
1.5
0.9
85. 5
1.4
0.9
82.8
1.4
0.8
80.2
1. 3
0. 8
77.7
1.2
0. 6
75.2
1.2
0.6
72. 8
1. 1
0.5
70.5
1.0
0.5
68.4
Nitrogen
Oxides
(NOX)
0.4
0.3
10.4
0.4
0.3
10.0
0.4
0.2
9.6
0.4
0.2
9.3
0.3
0.2
8.9
0.3
0.2
8. 6
0.3
0.2
8.3
0.3
0.2
8.0
0.3
0.1
7.6
0.2
0. 1
7.3
Sulfur
Oxides
(SOX)
0.0
0.0
0.0
0.0
0.0
0. 0
0. 0
0.0
0.0
0. 0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0. 0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Lead
(Pb)
0.0
0.0
0.4
0. 0
0.0
0.2
0. 0
0. 0
0. 1
0. 0
0. 0
0.1
0.0
0.0
0. 0
0.0
0. 0
0. 0
0.0
0. 0
0.0
0.0
0.0
0. 0
0.0
0.0
0. 0
0.0
0. 0
0.0
Source: MATHAIR outputs.
31
-------
TABLE 6. DAMAGES DUE TO AIR POLLUTION IN CHICAGO
ASSUMING NO AIR POLLUTION CONTROLS (1975
PRESENT VALUE IN MILLIONS OF 1973 DOLLARS)
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Damage
Category
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Health
Vegetation
Materials
Carbon
Monoxide
(CO)
1.2
0.0
24. 3
1.2
0.0
23.4
1. 1
0.0
22.4
1.0
0.0
20.9
1.0
0.0
20.5
0.9
0.0
19.7
0.9
0.0
18. 3
0. 8
0.0
18. 1
0.8
0.0
17.3
0.6
0.0
16.6
Oxidants
(OX)
16.6
3.4
442.6
15.5
3. 1
421.9
14.4
2.9
402. 3
13.5
2.7
383. 6
12.5
2.5
365.9
11.6
2.3
354.5
10.9
2.2
333.2
10. 1
1.9
318.0
9.5
1.8
303.6
8.8
1.7
289.9
Nitrogen
Oxides
(SOX)
46.2
10.2
531. 1
42.9
9.4
505. 1
39.8
8.7
480.6
37.0
8.0
457.2
34.5
7.4
435. 1
32. 0
6.8
414. 1
29.8
6.2
396.3
27.7
5.8
375.3
25.7
5.4
357.5
24.0
5.0
340.5
Sulfur
Oxides
(SOX)
139.8
83. 1
1034.5
128. 8
76.0
975.5
118.7
69.5
920.1
109.3
63. 5
867.7
100.7
58. 1
818.4
92.7
53.1
771.9
85.5
48.6
728.1
78.7
44. 3
686.7
72.6
40.5
647.7
66.9
37.0
610.9
Lead
(Pb)
0.0
0. 0
0.3
0. 0
0.0
0.2
0. 0
0. 0
0. 1
0. 0
0. 0
0. 0
0.0
0.0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0.0
0. 0
0. 0
0. 0
0.0
0.0
Source: MATHAIR outputs.
32
-------
SECTION VII
CHANGES IN AIR QUALITY USING ALTERNATIVE
AIR POLLUTION CONTROL STRATEGIES
In this section we use the MATHAIR model to calculate the
impact of alternative pollution control strategies on air quality. We
begin by describing how the particular strategies were chosen and then
examine their possible impact on air quality in Los Angeles, New York,
Washington, D. C. , and Chicago.
AIR POLLUTION CONTROL STRATEGIES CONSIDERED
The Principle of Strategy Selection
The basic principle for selecting strategies was to develop air
pollution control strategies involving progressively more stringent
controls. Starting with a zero strategy, SO, involving no controls,
more stringent control is achieved through either additional control
measures or stricter application of existing controls. Strategies
ordered by amount and degree of control are assumed to result in
progressively improved air quality. While there is no single physical
unit for measuring air quality (or pollution reduction) since several
pollutants are involved, the improvements in air quality are measured
in MATHAIR by the estimated dollar value of damage reduction.
The strategies considered here consist of combinations of
measures for controlling emissions of carbon monoxide, hydrocarbons,
and nitrogen oxides. We have not included measures for controlling
emissions of lead or sulfur oxides in the strategies. Neither Federal
nor State transportation control plans have required controls of these
pollutants. Furthermore, mobile sources account for less than 5
percent of sulfur oxide emissions. Control measures for these
pollutants could, however, be simulated with MATHAIR.
The five types of control measures of which our strategies are
composed are the following:
hardware devices for new automobiles
retrofit devices for existing automobiles
33
-------
inspection and maintenance (I/M) for new
and existing automobiles
transportation controls for reducing the
use of the automobile
hardware devices for stationary sources
of pollution
Figure 7 shows an example, in matrix form, of four increasingly
stringent strategies and the zero control strategy. The columns of the
matrix correspond to the five types of control measures and the rows
to the five strategies.
Strategy 0 (SO)
Strategy 1 (SI)
Strategy Z (SZ)
Strategy 3 (S3)
Strategy 4 (S4)
Control Devices
For New
Automobiles
-
low
moderate
moderate
high
Retrofit
Control
Devices
-
-
-
-
required
Inspection
and
Maintenance
-
-
-
for all
automobiles
for all
automobiles
Transportation
Controls
-
-
-
-
required
Stationary
Source
Controls
-
low
moderate
moderate
moderate
Matrix columns correspond to control measures and rows correspond to strategies. Entries describe
the level of control.
Figure 7. Example of five increasingly stringent air
pollution control strategies.
The first row of the matrix corresponds to SO. Since zero
controls are required, there are no entries in this row. The first
strategy, SI , involves a low level of control for each of two types of
measures. These might take the form of minor hardware adjustments
on all new automobiles built after a certain date and the installation of an
assortment of suitable control devices on various types of stationary
sources of pollution. The low level of control in the latter case might be
reflected by installing controls on only a small percentage of each type
of plant. *
It is convenient to designate an individual unit of stationary source
pollution as a "plant. " This term must be understood in a very
general sense. Airplane ground operations, for example, are
a "stationary" source of pollution and a plant, in this case, is a
single airport.
34
-------
The second strategy, S2 , requires an increased level of control
for the same two measures. A moderate level of control through
devices for new automobiles might take the form of more extensive
adjustment, including additional devices, and an earlier starting date.
In the case of stationary source controls, the percentage of plants fitted
with control devices could be increased.
The next row of the matrix in Figure 7 corresponds to the third
strategy, S3. S3 adds I/M for all automobiles to the requirements of
S2.
The final strategy, S4 , requires two new kinds of control and
increases the severity of a third measure. A higher level of control
through new automobile devices could be achieved by requiring addi-
tional devices not specified in the preceding strategy. In addition,
certain retrofit control devices are required and a reduction in the use
of automobiles is achieved through transportation controls.
Control Measures for Mobile Sources
Hardware Devices for New Automobiles--
Automobile manufacturers have been required to comply with new
automobile emission standards since 1967 (California since 1965).
Emission reductions have been achieved through engine adjustments and
installation of pollution control devices. The zero control strategy, SO,
assumes that these adjustments and installation of devices were not
made, and that automobiles built after 1967 (1965 in California) have
1967 (1965) model emission characteristics. *
We will consider the following three measures for inclusion in
our control strategies:
• All automobiles built before 1974 are assumed
to meet the applicable Federal (or California)
new automobile interim emission standards (shown in
Table 7) and 1974 standards are required for 1975
and later model automobiles.
• All new automobiles built before 1975 are
assumed to meet the applicable Federal (or
California) new automobile interim emission
standards, and the 1975 interim standards are
required for 1976 and later model new auto-
mobiles.
In the case of the stationary sources of pollution, however, SO
assumes the existing level of control.
35
-------
|8
S|i
QLC HH
ri n
« W H
<; > H
9^£
*£3
£«3
££E
o £ w
1-1 O s.
W M •>
J3 ^ to
£? to
2 K n
H O &
FH ^0
w w .
5riffl
3 W j
X H §
W H} o
1-1 1-3 NO*
< O re*
Cp Pn
H H p
W O 5
h O id
'
r-
H
^*
j^
pH
to
41
•o
1-1
X
O
C
4)
O
4J
•H
55
c
O 4-1
•r4 C
4J <\l
O O
3 (^
•O 41
4) O.
OS
41
»— 1
*H
^
CO
O
Monoxide
c
o
o
01
c
o
.0
u
cd
u
O
•«
a
•*
B
Reductio
percent
t
Gram/Mile
O 4J
•H B
4J CD
U O
3 r.
i CU
S
OJ rH
•O CX
0 S
» w
| ^ »^ ^ ff)
t •& *O CO ^t *O
|H «-+° IJ 1-1 )H IH
*o 41 ccf1™^ o rt 3 o cd
4) T3 CO) AJTJ^H 4J*O
rH O O^O 3CIH 3B
t— 1 fij -H h 4-ICOCU 4-1 CO
O 4JCd « 4-1 4J CO 4J
rJ
r-l N,.
1 CO in vO IN.
41 IN. IN. fN (N.
PL, r-< rH r-4 rH
•
XNi
*T^
U
1
to
> 41
O rJ --H
N^ IH nj
a -"
CO N^ {H
2s 0
3 2 P
•0 60 5
4» C
O o O
O * ^
U CM .S
^^ -__j V**
*J "c C
CO co M
4) "
H ^
4> rt
« -S £
00 •-" tJ
* x ™
So h
C •"
rf | .2
S « 'S
^5 a
41 Ij rQ
IS CO j
4J ° ^
o o 3 G
« 2
I ^ =s
•>
O
r- H
t— 1
• H
rH
N?
CO
0)
oo
oi
u->
ty)
(U
oj
• H
J_>
*s
tn
CO
C
O
0)
C 4) "TJ
3rH < O
iH J->
jO 4 g J*
32 i-i 5
oo U ^
CD CUE
5 * "go:
B ° *» a
o on o
•• O *rH
TJ CO T S •"
4i -a U u
UP (U
3 cd ci) i v
S "B ^2
fl) a) M "
S ™ r f^
CO
U cd
C8 1-4 (U
B • cj
w U a JL
r-4 0 CD "
4) >H t) P
> iH i-l O
4) rH X r/-)
r-l Cd O
U
B B
o s a>
•r4 iH 60
09 MO
ca 4) VJ
iH 4J U
a c i-i
M MB
^^ ^N,
CO J3
N~X *"^
36
-------
All new automobiles built before 1975 are
assumed to meet the applicable Federal (or
California) new automobile interim emission
standards, and 1975 interim standards are
required for 1976 and later model new auto-
mobiles. In addition 1978 standards for
carbon monoxide, hydrocarbons, and nitrogen
oxides must be attained by all post-1977 auto-
mobiles (as with a three-way catalytic con-
verter).
Retrofit Devices for In-Use Automobiles--
Certain kinds of pollution control hardware devices can be
retrofitted onto existing automobiles. Three devices, which are likely
to be widely used for this purpose, are listed below with brief descrip-
tions of their mode of control:
Vacuum Spark Advance Disconnect (VSAD) --
The idle air-fuel mixture is made leaner, idle
speed is increased and vacuum timing advance
is made inoperative during normal engine
operation. VSAD can be installed in pre-1968
automobiles.
Air Bleed to Intake Manifold -- The air-fuel
ratio is increased by means of an air valve to
the intake system. This device is useful for
pre-1972 automobiles.
Oxidizing Catalytic Converter -- A catalytic
converter is installed in the exhaust system of
some pre-1974 automobiles to oxidize CO and
HC.
I/M for New and In-Use Automobiles--
I/M programs are intended for the diagnosis and subsequent
correction of excessive emissions. It is generally believed that a
transportation program requiring hard-ware control devices on new and/or
in-use automobiles should be accompanied by mandatory I/M to get full
advantage of the control device. I/M is also useful as a stand-alone
measure. *
Measures for Reducing Automobile Travel--
Automobile travel can be reduced through directly restraining
the use of automobiles and through improving the level of service of mass
transportation. Both types of measures encourage a shift from automobile
* A maintenance program ean improve the gasoline consumption character-
istics of automobiles. We will have more to say about this subject in
Section Vlll.
37
-------
travel to mass transportation.
Restraints on the use of automobiles include physical barriers,
like exclusion from certain roads or effective gasoline rationing, and
increased costs due to taxes or fees on gasoline, parking, or road use.
Besides improving air quality, these measures may have other effects
like reducing noise, congestion, and energy consumption.
The level of service of mass transportation (bus and rapid transit)
can be improved by shortening headways, * extending areas covered,
reducing fares, and in, in the case of buses, establishing exclusive bus
lanes.
For the purpose of specifying strategies, we will be interested
in the direct effects of such measures on travel time and out-of-pocket
costs. From this view, the above measures can be restated as follows:
increase in per/trip or per/mile cost of
automobile travel
increase in automobile transit time
decrease in fares for mass transportation
decrease in waiting or transit times for mass
transportation
Control Measures for Stationary Sources of Air Pollution
The major stationary source pollutants are sulfur oxides
and particulate matter. However, it is possible to identify a number
of stationary sources of the mobile source pollutants, CO, HC, and
NOX , and to specify suitable controls. Examples of control measures
would include the folio-wing:
Hydrocarbon emissions during gasoline
distribution and marketing can be controlled
by installing hardware devices to collect
the vapors.
Tank roof designs can incorporate controls
for hydrocarbon emissions from petroleum
storage tanks.
Thermal or catalytic incineration can con-
trol industrial evaporation of hydrocarbons
from industries like printing or surface
coating.
Aircraft ground operations emissions of
carbon monoxide and hydrocarbons can be
controlled by modifying the operating
characteristics of the turbine or piston engine.
Headway is the scheduled time between two vehicles traveling the
same route.
38
-------
Nitrogen dioxide emissions from coal-, oil-,
and gas-fired utility boilers can be controlled
by combustion modifications like water injec-
tion or low excess air.
Strategies Selected
Four increasingly stringent strategies, SI through S4 , using
the control measures discussed earlier, are presented in Figure 8.
There are no entries in Figure 8 for the zero control strategy
SO. The first strategy, SI , represents a modest level of control. SI
requires some new automobile control devices and stationary source
controls on specified percentages of different types of plant. Control
device emission reduction characteristics typically assume an otherwise
uncontrolled source. * The percentages of installations to be controlled by
the strategies were arrived at after making rough estimates of the existing
levels of control of different types of plant.
The next two strategies, S2 and S3 , can be considered moderate or
"most likely" scenarios. S2 requires more of the same kinds of controls as
SI. An additional new automobile device is required and a larger
percent of stationary installations is controlled. The third strategy
adds I/M for all automobiles to the requirements of S2.
We have specified three variants of S4, our most stringent
strategy. All three require a more effective device on post-1977 auto-
mobiles than does S3 as well as adding retrofit devices for the in-use auto-
mobile stock and transportation controls. Transportation controls for
reducing the use of the automobile may take the form of increasing the cost
of using the automobile or decreasing the cost of mass transportation. It
is generally believed that the latter type of control must be accompanied
by some form of automobile use restraint in order to be effective in
reducing the use of automobile. To test this hypothesis, we have devised
three transportation control measures (whose impacts are compared in
Section IX). Strategy S4. 1 increases the cost per mile of automobile
travel by half: this measure might, for example, be implemented by
increasing the gasoline tax. S4. 2 decreases bus fare by one-third.
Finally, S4. 3 combines both of these transportation control measures.
One example is the case of incremental improvements: Combustion
modifications on an uncontrolled utility boiler will be more effective
than the same adjustments to a boiler that has already undergone
other combustion modifications. Another type of example is the case
where control devices are already in place. If 50 percent of existing
petroleum storage tanks are equipped with roof designs that limit
hydrocarbon evaporation, only the remaining 50 percent is subject to
this type of control.
39
-------
a
—
"3 "°
3 i-
!J
o
^ •
1 K
3
in
e
_o
*
0«
£.•3
HO
2
Inspection and
Maintenance (J
*o
g
U
l|
et G
•i •
1|
c a
SQ
I'D
^ t.
Eg
zu
C
I
3
0
fO — N fM M (M
£
lo 1
r; -* x c »
3 * "5 <•' *•
— •» "C 2 c "s
I * — "5 ** o — *
5> *3 ** 3 P* ° *•*
_*
j|. |ri
^ x 'i j z ** "
^E±T g.J*S
^•*^^ *o2£^
• 3 • 1 §Is«
^-•^ ;3"'
w
* """ """""
g g
•B 3 >.
S J S ?
C C ^ ±
« v "S . . -c
to u J U Q
1.1 l|ll3
Iffilllll
_
Jj • *O
•0 J JJ „ , g.
"st^j z"13 u £"« ^
* | •- s S-s • « ?. g
^J«_S ""fE-J fS J
I°"S g|s5 |^ ?
ZE£^«; 2.-S ,» <
•
iC »*>i/\ m «/i *«
g 1
• 2 F »
1 1 lo 1
i £• E- °*
1.1^ |T|||
1 JHt|s:i
M.oU^P4,q »O<
"3
•8 .
ill
in
fit
"3 — "°
• S * •* • 1:
, "S-S S "^ * ^ -w-^
3|"§J^-i"° ~*-c 5
^•P*'^— O^"*>" "S*'^- CT-
°^ " *J "S " "X"??'^ °"'o "
I'i'ilfl?! #i« i
-e-"< s"-" srs *
IA
^ «*•> I/N in tn m
c o
D O
•0 B K
* * ** M
*• ** ** a
c c 5 i
& » |u "g
£• i1 E - .S x
•** •** * "*• * -S ?
llsllisll
;S -ZS SS -Si?
o* ^ ** Us"l2!!
2 .S — W y— OC^SjJ'-'
•i a) S >• "^ *O •i.e^'7
£,— -7 <>• v i; £,-• i »-!
gigs 5£_ !l|S*;r
is^s Js* ?J^ssv
o a
• 0 • 0 •
5
'o •
Iji
^•1
111
- •» i .
°i gi Ul
5r "5 ^ e •* •"
§2 g g2|
— tf> *• S
o7|| ||gj
>- E< EoS~
-1 |a ^:
ill. -Es ii!^ "*^
•5^3° 3^*^ 7-^'c"^.p'3
» 5-^^ E-IS " S *S1 2° *
— £ «« £ '3££JI **-«•?< »».**«
,,-j*.,™ o E "C °S— £**•
Sill slSs slsd ^,si.
«
rt
c
o
•H
-»->
R)
,—I
d
8
JH
O
0)
-l->
O
0)
CO
CO
0)
• H
bC
CD
oo
(D
h
S
bo
•H
fa
40
-------
EFFECTS OF CONTROL STRATEGIES ON AIR QUALITY FOR THE FOUR
STUDY REGIONS
In the second part of this section we report the MATHAIR air quality
predictions for the period 1975 to 1984 in the four urban areas for the
pollution control strategies discussed earlier. Results for strategy
S2 are not presented for reasons which will be discussed in Section VIII.
Futhermore, results for only the first variant of S4 (S4. 1) are presented
in this section.
Air quality outcomes for the zero control strategy, SO, are
shown in Table 8 and discussed in Section VI. Tables 8 through 11
show the air quality predictions for the four study regions for strategies
SI, S3, and S4. 1 and also indicate the air quality standards for each
pollutant.
Air quality predictions for Los Angeles are shown in Table 8.
TABLE 8 AIR QUALITY IN LOS ANGELES FOR
THREE SELECTED STRATEGIES (PPM)
Strategy
Strategy 1
Strategy 3
Strategy 4. 1
Standards
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Carbon
Monoxide
(CO)
27.0579
27. 0336
23.4120
22. 2543
21. 5320
21.0851
20.8940
20.8681
20. 9625
21. 1525
22. 2574
18.7369
15. 5773
13. 0419
1 1. 0070
9. 4865
8. 4625
7. 7119
7. 2409
6. 9338
13.4732
11.7851
10.2726
8. 8677
7. 7150
6. 7976
6. 1389
5. 6250
5. 2746
5. 0362
9.0
Oxidants (OX)
OX1
0.2661
0. 2537
0.2444
0.2386
0.2353
0.2360
0.2378
0.2607
0. 2820
0. 3017
0.2268
0.2024
0. 1816
0. 1655
0. 1528
0. 1458
0. 1417
0. 1595
0.1768
0. 1930
0. 1814
0. 1651
0. 1533
0. 1426
0.1341
0.1294
0. 1264
0. 1404
0.1536
0. 1661
OX2
0. 2989
0. 2876
0.2791
0. 2738
0.2711
0.2716
0. 2733
0. 2895
0. 3047
0. 3192
0. 2641
0.2387
0. 2164
0. 1986
0. 1850
0.1764
0. 1714
0. 1818
0. 1925
0. 2031
0.2212
0. 2043
0. 1895
0. 1740
0. 1616
0. 1529
0. 1468
0. 1528
0. 1593
0. 1662
OX 3
0. 3113
0. 3001
0. 2917
0.2865
0.2839
0.2844
0.2861
0.3006
0. 3140
0. 3266
0.2772
0. 2510
0. 2279
0.2094
0. 1953
0. 1862
0. 1809
0. 1900
0. 1989
0.2074
0.2337
0.2160
0.2005
0. 1889
0. 1705
0.1608
0. 1539
0.1578
0. 1618
0.1659
.08
Nitrogen
Oxides
(N0x)
0.0588
0. 0573
0. 0562
0.0555
0. 0553
0.0553
0.0556
0. 0560
0. 0565
0.0571
0. 0547
0. 0503
0.0462
0. 0428
0. 0403
0. 0384
0.0373
0.0365
0.0361
0. 0360
0. 0486
0.0454
0. 0423
0. 0383
0.0350
0.0323
0.0308
0.0287
0. 0275
0.0268
.05
Sulfur
Oxides
(sox)
0.0158
0.0158
0.0158
0.0159
0.0159
0.0159
0.0159
0.0160
0.0160
0.0161
0.0158
0.0158
0. 0158
0. 0159
0.0159
0.0159
0.0159
0.0160
0.0160
0.0161
0.0156
0.0156
0.0157
0.0157
0.0157
0.0158
0.0158
0.0158
0.0158
0.0159
.03
Source: MATHAIR output
41
-------
One of the impacts of strategy SI is a significant reduction in carbon
monoxide (CO), and oxidant (OX) concentrations with respect to the zero
control strategy outcome. Concentrations of the third major mobile
source pollutant, nitrogen oxides (NOX), are also reduced. Table 8 shows
an initial decline in concentrations of these three pollutants over time as
the controls are introduced. Eventually, all three levels begin to
increase (starting in 1983 for CO, 1980 for OX, and 1980 or 1981 for NOX).
The increasing use of the automobile dominates further emission
reductions, once the controls have had their principal impact. SOX is
not subject to controls and, therefore, concentrations are the same as
they were in the case of the zero strategy, SO. The standards for CO,
HC, and NOX are not met even by 1984.
Concentrations of CO, HC, and NOX are further reduced under
strategy S3. The CO and NOX controls are sufficiently effective in the
later years of the simulation horizon that reduction in concentrations is
monotonic. The oxidant levels, however, begin to increase in 1982.
Under this scenario, the CO standard is met by 1981 and the NOX
standard by 1977.
All pollutant concentrations are further reduced under strategy
S4. 1. The CO standard is met by 1978. The NOX standard is satisfied
for each simulation year. The oxidant standard cannot be met, and
concentrations begin to increase after 1981. Even though there have
been no direct controls of SOX, concentrations are somewhat lower
than in the case of the other strategies since the use of automobiles
has been reduced.
Overall, controls in Los Angeles are effective in improving air
quality. However, the MATHAIR simulation results confirm the well-known
oxidant problem.
Predicted pollutant concentrations for New York are shown in
Table 9. As in the case of Los Angeles, CO and OX concentrations are
significantly reduced and NOX concentrations are somewhat attenuated
by the SI scenario. Oxidant levels begin to increase after 1980 for the
reason cited in the analysis of Los Angeles outcomes. The CO, HC,
and NOX standards are not satisfied in any simulation year.
Concentrations of CO, HC, and NOX are somewhat further
reduced by S3. The analysis of SI outcomes holds equally for S3. With
S4. 1 the CO standard is met in 1975 and the NOX standard is satisfied
by 1982. While the oxidant standard cannot be met even with our strictest
control strategy, it is nearly achieved. SOX concentrations are slightly
reduced due to restricted use of the automobile. Overall, we can
conclude that controls are effective in improving air quality in New York.
Table 10 shows predicted pollutant concentrations for Washington,
D. C. As in the cases of Los Angeles and New York, CO and OX levels
are significantly reduced by the first control strategy. However, CO
and OX concentrations begin to increase in 1983 and 1980, respectively,
and neither the CO nor the HC standard can be met. While NOX concentra-
42
-------
TABLE 9. AIR QUALITY IN NEW YORK FOR THREE
SELECTED STRATEGIES (PPM)
Strategy
Strategy 1
Strategy 3
Strategy 4. 1
Standards
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Carbon
Monoxide
(CO)
16.9442
15.6198
14.5587
13.7236
13. 1359
12.7420
12.4871
12. 3279
12.3420
12.Z074
15.3224
13.8792
12.8021
12.0563
11.6207
11.4196
11. 3675
11. 3933
11.4561
11.5382
8. 1096
7.7558
7. 5471
7.2366
6.9804
6. 7555
6. 5498
6.3495
6.1497
5.9684
9 A
. \J
Oxidants (OX)
0X1
0.1485
0. 1406
0. 1340
0.1294
0.1262
0.1250
0.1308
0. 1364
0.1416
0.1483
0.1271
0.1176
0.1107
0.1063
0.1089
0.1036
0.1105
0.1171
0. 1230
0. 1304
0.0909
0.0872
0. 0846
0.0823
0.0806
0.0797
0.0824
0.0848
0.0869
0.0898
OX 2
0.1569
0. 1513
0. 1468
0. 1433
0. 1409
0.1397
0.1425
0. 1453
0.1479
0.1515
0. 1396
0.1331
0. 1281
0. 1247
0.1226
0. 1218
0.1251
0.1284
0. 1314
0. 1353
0. 1084
0. 1058
0. 1039
0. 1010
0. 0986
0. 0967
0. 0971
0. 0975
0. 0978
0.0988
OX 3
0. 1574
0. 1518
0. 1472
0. 1437
0. 1412
0. 1400
0. 1429
0. 1457
0. 1484
0.1519
0. 1400
0. 1333
0. 1281
0. 1246
0. 1223
0. 1216
0. 1252
0. 1287
0.1318
0. 1357
0. 1081
0.1054
0. 1033
0. 1004
0.0980
0.0963
0.0969
0.0975
0.0980
0.0990
. 08
Nitrogen
Oxides
(Ncy
0. 0742
0.0730
0.0719
0.0710
0.0702
0. 0697
0. 0694
0. 0693
0. 0692
0. 0692
0. 0685
0. 0672
0. 0660
0. 0649
0. 0641
0. 0635
0. 0632
0. 0630
0. 0629
0. 0628
0. 0572
0. 0567
0. 0561
0. 0545
0. 0531
0. 0518
0. 0507
0. 0499
0. 0491
0. 0486
. 05
Sulfur
Oxides
(sox)
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0109
0.0108
0.0108
0.0108
0.0108
0.0108
0.0108
0.0108
0.0108
0.0108
0.0108
. 03
Source: MATHAIR output
TABLE 10. AIR QUALITY IN WASHINGTON, D. C. , FOR THREE
SELECTED STRATEGIES (PPM)
Strategy
Strategy 1
Strategy 3
Strategy 4. 1
Standard*
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Carbon
Monoxide
(CO)
20.0693
19.2152
17.9463
17. 0237
16.4456
16. 1495
16.0240
16.0151
16.0956
16.2396
18.6065
16.6259
16. 3071
14.5138
14. 1748
14.1822
14. 3904
14.6975
15.0212
15.3611
11.4460
10.7654
10.4634
9.9176
9. 5247
9. 1663
8.8652
8. 5180
8. 126o
7.7333
9. 0
Oxidants (OX)
OX1
0. 1706
0. 1584
0. 1498
0. 1443
0. 1415
0. 1419
0. 1554
0. 1683
0. 1807
0. 1962
0. 1540
0. 1390
0. 1295
0.2345
0. 1231
0. 1254
0. 1411
0. 1560
0. 1700
0. 1868
0. 1172
0. 1088
0. 1042
0. 0999
0.0975
0.0969
0. 1065
0. 1149
9. 1223
0. 1318
OX2
0.2159
0.2053
0. 1974
0. 1922
0. 1893
0. 1891
0. 1983
0.2073
0.2162
0. 2274
0.2022
0. 1894
0. 1806
0. 1755
0. 1733
0. 1742
0. 1846
0. 1949
0. 2047
0.2166
0. 1667
0. 1594
0. 1547
9. 1463
0. 1396
0. 1348
0. 1402
0. 1402
0. 1426
0. 1469
0X3
0.2296
0.2136
0.2103
0. 204S
0.2018
0.2015
0.2109
0.2199
0.2286
0.2391
0.2154
0.2019
0. 1924
0. 1869
0. 1845
0. 1855
0.1966
0.2070
0.2196
0.2278
0. 1773
0. 1693
0. 1641
0. 1554
0. 1485
0. 1436
0.1487
0. 1486
0. 1501
0. 1531
.08
Nitrogen
Oxides
,NOx)
0.0083
0.0081
0. 0079
0. 0078
0.0077
0.0077
0.0077
0. 0077
0.0078
0. 0078
0.0081
0.0079
0.0077
0.0075
0. 0074
0.0074
0.0074
0.0074
0. 0074
0.0075
0. 0072
0. 0071
0. 0070
0.0065
0.0061
0. 0057
0.0054
0. 0052
0.0050
0. 0048
.05
Sulfur
Oxides
(sox)
0.0155
0.0155
0.0155
0.0155
0. 0155
0.0155
0.0155
0. 0156
0.0156
0.0156
0.0155
0.0155
0.0155
0.0155
0.0155
0.0155
0.0155
0.0156
0.0156
0.0156
0.0154
0.0154
0. 0154
0.0154
0.0155
0. 0155
0. 0155
0.0155
0.0155
0.0155
.03
Source: MATHAIR output
43
-------
tions are slighly reduced from the levels predicted by MATHAIR for the
zero control strategy, the standard for NOX is met even with no
controls.
Concentrations of CO, HC, and NOX are somewhat further
reduced by S3. The analysis of SI outcomes holds equally for S3.
All concentration levels are further reduced by S4. 1. CO concen-
trations decrease monotonically over time, and the standard can be met
by 1981. SOX emissions decline slighly, as in the case of the other study
areas, due to diminished automobile usage. The oxidant standard still
cannot be met.
Overall, the controls are effective in improving air quality.
Finally, we turn to Table 11 which shows the MATHAIR predictions
of pollutant concentrations in Chicago. As in the other study regions, SI
has the effect of significantly decreasing CO and OX concentrations and of
slightly reducing the ambient level of NO . However, levels of these three
pollutants begin to increase in 1983, 198?and 1982 respectively. The
standards cannot be met for any of these three pollutants.
TABLE 11.
AIR QUALITY IN CHICAGO FOR THREE SELECTED
STRATEGIES (PPM)
Strategy
Strategy 1
Strategy 3
Strategy 4. 1
Standards
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
198Z
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Carbon
Monoxide
(CO)
18. 3919
17. 1505
16. 1662
15.4514
14.9687
14.6845
14. 5446
14. 5036
14. 5415
14.6372
16.6288
15.2169
14.2058
13. 5473
13.2171
13. 1374
13.2192
13. 3840
13. 5889
13.8163
9.4046
9.0501
8.8739
8. 5566
8.2965
6. 0656
7.6505
7. 6334
7. 4085
7. 2013
9.0
Oxidants (OX)
OX1
0. 1502
0. 1442
0. 1396
0. 1365
0. 1346
0. 1342
0. 1405
0. 1467
0. 1526
0. 1601
0. 1330
0. 1256
0. 1205
0. 1174
0. 1160
0. 1165
0. 1238
0. 1308
0. 1374
0. 1455
0. 1049
0. 1018
0.0999
0. 0980
0.0967
0.0961
0.0993
0. 1022
0. 1047
0. 1083
OX2
0. 1655
0. 1601
0. 1558
0. 1528
0.-1508
0. 1503
0. 1547
0. 1592
0. 1635
0. 1692
0. 1508
0. 1443
0. 1395
0. 1364
0. 1347
0. 1346
0. 1396
0. 1446
0. 1494
0. 1554
0. 1202
0. 1174
0. 1155
0. 1123
0. 1094
0. 1073
0. 1082
0. 1092
0. 1100
0. 1118
OX3
0. 1719
0. 1668
0. 1626
0. 1596
0. 1577
0. 1571
0. 1606
0. 1643
0. 1679
0. 1725
0. 1583
0. 1520
0. 1473
0. 1441
0. 1423
0. 1421
0. 1462
0. 1504
0. 1544
0. 1593
0.1266
0.1239
0.1220
0. 1132
0. 1147
0. 1121
0. 1119
0.1119
0. 1118
0.1125
.08
Nitrogen
Oxides
(N0x)
0.0880
0. 0863
0.0848
0.0835
0.0827
0.0822
0. 0821
0.0823
0.0823
0.0832
0.0843
0.0823
0.0806
0.0791
0.0781
0.0775
0. 0773
0.0773
0.0776
0.0780
0.0683
0.0675
0.0667
0.0637
0.0609
0.0585
0.0565
0.0546
0.0534
0.0523
.05
Sulfur
Oxides
(sox)
0.0217
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0217
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0218
0.0216
0.0216
0.0216
0.0216
0.0216
0.0216
0.0217
0.0217
0.0217
0.0217
.03
Source: MATHAIR output
44
-------
While concentrations are somewhat further reduced by S3, the
analysis of SI outcomes holds equally for S3.
The impact of S. 4 is to further reduce the levels of all pollutants.
CO concentrations are seen to decrease monotonically over time and the
standard can be met by 1977. SOX levels are slightly reduced since
automobile use has diminished. Neither the OX standard nor the NOX
standard can be met. However, our overall evaluation for Chicago, as
for all the other study regions, is that a program of even a modest
level of pollution controls can significantly reduce ambient concentra-
tions of the mobile source pollutants.
In this section we have discussed the impact of three control
strategies on air quality. In the next section we analyze the benefits
and costs of alternative strategies for the four study regions.
45
-------
SECTION VIII
ESTIMATES OF BENEFITS AND COSTS
FOR ALTERNATIVE STRATEGIES IN THE FOUR STUDY REGIONS
REVIEW OF BENEFIT AND COST COMPONENTS
Strategy Benefits and Costs for the Study Regions
The individual components of costs and benefits have already been
discussed in Section V. To briefly review, however, the cost of an air
pollution control strategy is the sum of;
• automobile and stationary source control
device costs
• operating and maintenance costs for automobile
and stationary source control devices, including
the cost of increased (or decreased) gasoline
consumption due to automobile control devices.
• the value to travelers of reduced (or enhanced)
mobility.
The benefits of a strategy are the dollar value of reduced air pollution damage
and damage avoidance costs for all damage categories and for all pollutants.
LOS ANGELES
We now turn to an analysis of the benefits and costs of selected
strategies in Los Angeles, shown in Table 12. The dollar value of damages
for the zero control strategy, SO, is given on the table because it provides
the baseline for calculating the benefits attributable to the control strategies
(SI, SZ, S3, S4. 1), which have already been defined. *
One strategy, S2, involves higher costs and lower benefits than S3,
and thus turns out not to be cost effective.''"'* The difference between the two
strategies is that S3 adds mandatory inspection and maintenance (I/M) for
As we noted earlier, benefits must be tallied by the model user by
subtracting the damages attributable to a strategy from the damages
of the zero control strategy.
A similar result was found for the other study regions.
46
-------
TABLE 12. BENEFITS AND COSTS FOR SELECTED STRATEGIES
IN LOS ANGELES (1975 PRESENT VALUE IN BILLIONS
OF 1973 DOLLARS)
Strategy
SO
SI
S2
S3
S4. 1
SO with trans-
portation
controls of S4. 1
Damages
7.8
4.2
3.03
2.97
2.5
6.5
Benefits
3.6
4.76
4.83
5.3
1. 3
Costs
1.9
2.6
2.4
3.6
0.7
Net Benefits
(Benefits-
Costs)
1.7
2.2
2. 5
1. 7
0.6
Source: MATHAIR output.
automobiles. An analysis of MATHAIR output shows that the predicted
saving to motorists from improved gasoline mileage is greater than the
cost of I/M; since I/M also reduces pollutant emissions, the net effect
is that S3 dominates S2. One wonders, however, why motorists do not
voluntarily have their automobiles inspected and maintained in light of
the expected cost saving. One possible explanation is that the data on
gasoline consumption savings and inspection and maintenance cost are in
error.
Table 12 also shows that strategy S3 maximizes net Benefits and
thus is the optimal choice among the strategies considered. * The final
row of the table reports the results of a "stand-alone transportation controls'
strategy that will be discussed below.
NEW YORK
The benefits and costs of selected air pollution control strategies**
for New York are shown in Table 13. For strategies SI, S3, and S4, the
benefits are smaller and the costs are greater than for the corresponding
strategies in Los Angeles. Furthermore, net benefits in New York are
negative in all three cases.
It, however, a policy decision were made to require S4. 1 because it comes
nearer to meeting the air quality standards, net benefits would still be
positive.
Results of S2 will not be reported for the remaining study regions since it
is not cost-effective.
47
-------
TABLE 13.
BENEFITS AND COSTS FOR SE'LECTED STRATEGIES
IN NEW YORK (1975 PRESENT VALUE IN BILLIONS
OF 1973 DOLLARS)
Strategy
SO
SI
S3
S4. 1
SO with trans-
portation
controls of S4. 1
Damages
28. 1
25.2
24. 6
23. 3
27.9
Benefits
2.9
3.6
4.9
2. 3
Costs
3.3
4.1
6.6
1.7
Net Benefits
(Benefits-
Costs)
-0.4
-0.5
-1.7
0.7
Source: MATHAIR output.
Although a detailed comparison of the different strategy outcomes
between regions would lie outside the scope of this study, an approximate
evaluation of the underlying causes can be made. The benefit of a strategy
that controls a single pollutant is the product of two terms: the dollar
value of damage per unit of the pollutant and the reduction in the number of
units of the pollutant emitted both by mobile and stationary sources.
Generalizing to the case of several pollutants, -we developed some
rough estimates of the two terms which show that the dollar value of damage
per unit of pollutant is more than twice as great in Los Angeles as in New
York* while the number of reduced units of pollutants in New York is only
about fifty percent greater than in Los Angeles. '*''* The product of the two
terms is thus more than one-third greater in Los Angeles, accounting for
the greater dollar benefit in that region.
Costs are essentially proportional to the number of automobiles and
the number of stationary sources since these are the units on which control
devices are installed. Control costs are almost twice as great in New York
as in Los Angeles because there are about fifty percent more automobiles''"'"'^
and more than twice the number of stationary sources''"'"'"'' in New York.
Estimated from base year data in Tables F-35 and F-37 of Appendix F.
Estimated using MATHAIR outputs.
Table F-l of Appendix F.
Judging from total emissions shown in Table F-28 of Appendix F.
48
-------
In addition to explaining these results, we were prompted by the
poor apparent performance of the control strategies in New York to investi-
gate whether an even lower level of control would be beneficial. We tested
an additional strategy involving stand-alone transportation controls; that is,
the zero control strategy with the transportation controls of S4. 1. The net
benefits, seen in Table 13, are small but positive.
appears optimal in New York.
This lenient* strategy
The impact of stand-alone transportation controls in New York
motivated us to test the impact of the strategy in Los Angeles; the results
are shown in Table 12. While net benefits are also positive in Los Angeles,
a greater level of control is optimal in that region.
WASHINGTON, D. C.
Benefits and costs for SI, S3, and S4. 1 for Washington, D. C. , are
shown in Table 14. As in New York, net benefits are negative for the three
strategies, implying that the optimal decision is to require no controls. ''"'*
TABLE 14.
BENEFITS AND COSTS FOR SELECTED
STRATEGIES IN WASHINGTON, D. C.
(1975 PRESENT VALUE IN BILLIONS OF
1973 DOLLARS)
Strategy
SO
SI
S3
S4. 1
Damages
1. 1
0.5
0.5
0. 3
Benefits
0.6
0.7
0.8
Costs
0.75
0.78
1.2
Net Benefits
( Benefits -
Costs)
-0. 14
-0. 13
-0.45
Source: MATHAIR output.
In this context lenient means both low cost and low benefit.
We did not simulate the stand-alone transportation controls strategy in
Washington, D. C. It is possible that such a lenient strategy would be
beneficial.
49
-------
Strategy benefits and costs are considerably lower in Washington, D. C.
than in the other study regions. While the dollar value of damage per unit
of pollutant is roughly comparable to that in New York, the initial level of
pollution and thus the amount of reduced emissions both for mobile and
stationary sources, is by far the lowest of all study regions. For this
reason the benefits realized by control strategies are low. Similarly, be-
cause costs are proportional to the number of automobiles and of stationary
sources, the small number of both of these sources of pollution in
Washington, D. C. explains the low strategy costs.
We next sought to explain why the marginal impact of SI was
considerably greater than that of S3 by inspecting each control measure of
the two strategies. In addition to moderate stationary source control, SI
requires that all the mobile source controls mandated between 1967 and
1974 be implemented, and that future automobiles have the characteristics
of 1974 models. In contrast, S3 adds to SI only two controls in addition to
stricter stationary measures; I/M on all automobiles and the catalytic
converter on post-1974 automobiles. Therefore in the case of SI the benefits
and costs of stationary source controls form only a small proportion of
total benefits and costs and the impact of the mobile source measures pre-
vails. However the impact of the stationary source control measure
prevails in the case of S3.
While the proportion of automobiles in Washington, D. C. , relative
to the other regions is small, the proportion of stationary sources is even
smaller. For this reason, the benefits and costs of S3 are particularly
low. While none of our strategies isolates the stationary source controls,
our results corroborate the a priori judgment that these measures have
the smallest impact in areas with a small inventory of stationary sources
of pollution.
CHICAGO
Benefits and costs of Si, S3, and S4. 1 for Chicago are shown in
Table 15. While net benefits of all three strategies are positive and of
similar magnitude, the optimal strategy of the three is S3.
We performed a rough analysis of the strategy costs and benefits,
as we did for New York and Washington, D. C. Since the number of auto-
mobiles and stationary sources of pollution in Chicago is about half that in
New York, this analysis suggests that costs in Chicago should be about half
those in New York, as indeed a comparison of Table 15 and Table 13 shows
them to be.
Comparison of the two tables also shows benefits in Chicago to be
about eighty percent of those in New York. Since base year emission levels
in Chicago are less than eighty percent of those in New York, it must be
that the dollar value of damage per unit of pollutant is higher in Chicago.
Rough estimates are not sufficiently detailed to permit direct verification,
but it is probably a greater dollar value of damage per unit of pollutant
in Chicago that accounts for the positive net benefits in Chicago.
50
-------
TABLE 15. BENEFITS AND COSTS FOR SELECTED
STRATEGIES IN CHICAGO (1975 PRESENT
VALUE IN BILLIONS OF 1973 DOLLARS)
Strategy
SO
SI
S3
S4. 1
Damages
18. 3
16.0
15.5
14.1
Benefits
2.3
2.8
4.2
Costs
1.9
2.3
3.7
Net Benefits
(Benefits-
Costs)
0.46
0. 50
0.48
Source: MATHAIR output.
SOME GENERAL IMPLICATIONS
The net benefits of strategies Si, S3, and S4. 2 for the four study
regions are assembled in Figure 9. Net benefits of all strategies are
positive in Los Angeles and Chicago and negative in New York and Wash-
ington, D. C. and net benefits are highest in Los Angeles and lowest in
New York. While care must be taken in generalizing the results of
simulations in only four regions, we feel that some useful statements can
be made about the reasons underlying these results.
Strategies Involving Pollution Control Devices
Strategies Si, S3, and S4. 1 require pollution control devices for
automobiles and stationary sources. The costs of the strategies are
plotted in Figure 10 as a function of the size of the 1975 automobile
stock and in Figure 11 as a function of the number of units of stationary
source pollution. *
For each type of stationary source (e.g., airport ground operations),
the number of units is equal to total emissions of that type of source
in a given region divided by the emissions of a "typical" unit (e.g. ,
ground operations at an average size airport). For purposes of
Figure 11, units are totaled over all types of stationary sources.
51
-------
o
o
u
s
u
w
3
o
z
CO
O
o
2 »
§ Jj
Sq
o
H
o
BC
CO
w
z
X
fl
o
•l-<
<3JD
(U
T)
d
4->
CO
?H
S
O
m
(U
£3
•H
0]
0)
•H
be
U
(1)
.^J
(U
W
<4H
O
CO
-M
•H
m
CO
fl
0)
fn
J3
bJD
52
-------
.fl «
O O
8-
53
in "
-i W
tn
*H •
rt
O VH
U o
>-s
60 S
(U O
JL -H
n) Z!
ID
?
n)
to
co
tt)
B O
S "S
U S
53
-------
to "*H rt m o
4) ° C (U •Jj
Hs*B2
H -H TI 3 —*
3 C JS O °
2 £> W CO Pn
M)
0)
tm
4)
rt
ui
_nj
«^
O
o
s
o
o
o
CO
o
o
LCI
o
o
o
PO
o
o
m
o
o
o
o
o
rt
o
O
O
<
U
s
U
W
U
O
D
U
Q
O
U
I
<4H
O
•a
1
co
cj
o
0)
fH
CO
(U
• H
DO
cu
O
CO
O d
0) O
U »
CD [>s
^ ^
o eg
CO O
4-J .H
03
d
be
O
U
rt —
54
-------
It is hardly surprising that costs for all strategies increase with
the number both of automobiles and of stationary sources since costs are
incurred on a per unit basis. However, while the total (and marginal)
costs of Si and S3 in all regions are in line with the size of the automobile
stock, the total (and marginal) benefits and the ratios of marginal benefits
to marginal costs for both strategies SI and S3"* are highest in Los Angeles.
One reason, that has already been discussed, is that the dollar value of
damage per unit of pollution is higher in Los Angeles than in the other study
regions. A second reason lies in the intensity of automobile use. To
illustrate this notion, suppose that there are the same number of automobiles
in two different regions, A and B, but that the average automobile in A is
driven twice as many miles and thus emits twice the quantity of pollutants
as an average automobile in B. Now let us assume that an air pollution
control device that reduces emissions by 10% is installed on all automobiles
in both regions. Since there are the same number of automobiles in both
regions, air pollution control device costs will be the same in A as in B.
However, for the given percentage decrease in emissions, the reduction of
units of pollutant emitted in A will be twice as large as in B. If all other
things are equal, in particular, the dollar value of damage per unit of
pollutant, the larger reduction in emissions will translate into a larger
dollar amount of benefits. This means that the dollar benefit produced by
mobile source control devices per dollar of cost will be greater in a region
where automobiles are driven intensively than they are in other regions.
Data on the average daily miles driven by an automobile in Los
Angeles, New York, Washington, D. C. , and Chicago, given in Table 16,
indicate that automobiles are driven most intensively in Los Angeles and
least intensively in Washington, D. C. This ranking is consistent with the
ranking of marginal benefits and with the ranking of the dollar benefit per
dollar of cost.
TABLE 16. SELECTED AUTOMOBILE USE STATISTICS
FOR THE STUDY REGIONS
Los Angeles
New York
Washington, D. C.
Chicago
1975 Daily Auto
Miles Traveled
159
200
36.5
113
1975 Automobile
Stock
4,680,750
6,777, 545
1, 694,500
3, 831, 175
Average Daily
Miles Per Auto
34.0
29.5
21.5
29.5
Source: All numbers are calculated from data in Appendix F.
a millions of miles.
Strategy S4. 1, which involves transportation controls as well as hardware
devices, is discussed in the next part of this section.
55
-------
While it is true that strategies SI and S3 include moderate stationary
source controls as well as automobile controls, -we conclude that one
component of the large net benefits seen in Los Angeles for strategies
requiring automobile pollution control devices is the relative intensity of
automobile use in that region and that automobile control devices are more
likely to be beneficial in high-pollution regions like Los Angeles -where
the automobile is used intensively than in areas with low pollution and low
intensity automobile use. This does not mean that all pollution control
devices are beneficial. For example, we calculated the benefits and costs
for Los Angeles of a new strategy which adds the three -way catalytic
converter''" to the control measures of S3. The marginal cost was about
twice the marginal benefit, indicating that the usefulness of a device depends
upon its cost and control characteristics and upon the level of control that
has already been attained.
Strategies Involving Transportation Control Measures
The fourth strategy adds three controls to S3: the three-way catalytic
converter for new post-1978 autos, several retrofit devices for in-use autos,
and a transportation control measure (50 percent increase in per mile auto
cost). The marginal costs of this strategy are greater than the marginal
benefits in all four study regions. However, both the marginal benefits and
the ratio of marginal benefits to marginal costs are highest in Chicago and
New York.
The data on changes in transportation patterns of the four regions,
given in Table 17, suggest that the key to these results lies in the different
natures of the transportation systems. The increased cost of automobile
transportation leads to a 25-30 percent decline in automobile passenger
trips in Chicago and New York, well above the 12-18 percent decline in Los
Angeles and Washington, D. C. Furthermore a much smaller fraction of
these foregone automobile trips are replaced by bus trips in Chicago (27
percent) and New York (11 percent) than in Los Angeles (41 percent) and
Washington, D. C. (50 percent). Thus there is very little pollution offset
from increased bus pollution in the first two cities.
However, Chicago and New York differ substantially in the allocation
of the remaining foregone trips. In Chicago 35 percent of the foregone trips
are switched to rail, so that only 38 percent of the reduced automobile trips
represent cancelled trips. On the other hand, in New York only 12 percent
of the foregone trips are switched to rail, meaning that 77 percent of the
reduced automobile trips represent cancelled trips. The differences are
even more striking when the changes in trips are disaggregated by trip
purpose and mode as shown in Table 18 for Chicago and New York.
The fraction of foregone automobile trips represented by foregone
work trips is about the same for Chicago (15%) and New York (19%). In
both cities, about forty five percent of the work trips formerly made by
automobile are now made by rail. The real difference lies in the foregone
This device is relatively expensive and effective. Its characteristics are
shown in Table F-41 in Appendix F.
56
-------
W
D<;
n
^ 2
E-H Q
*3
o w
^
l j y}
H W
< rt
i~
M k
1_]
, "~i
H-^ »— |
O ^3
On — *
t-< CA
§2
UG
o*
^8
rt f"1
o 5
w H
^
On """ i
HQW?
Q k W
t-C
O t-» ^
wo<
w fe Q
r-
W
PQ
f5
i
—]-—•
f" J3 tu ct) +ICM
" O ,£! rt "^ — •
•« P ° J2I
o g *; -o ~^
fa IS 5
CD 0)
•« £ i! ~
0 o x H
C "*"* ?
0 3 OT _
•rt <; „ co in
~ M g
£ S)H£
-o »
'a g0
rt w o
c M .
tX r^ ' '
(U ^ ^. rj<
01 -tn GO —
^ co m a>
rH H H t/5
CO ^
g .H* ^
® H u
-S s|
a) aot!o °°
03 q ' '
nj qj O
** CO **
^ m flj
g ^ 3
l-H A Q
1
° S
3 M) pS ?
< C ° ? co
rti '^ oo "n v
" — 01 v u Jj
"S e 'C i ^ ^
CD
1)
-71 c
S o
'3 ° '-°
ry* 2 *"*•
Op *~H
T^ O **^
rt 3 ™
CQ
S
O
rt
t— (
a
t~*
-------
TABLE 18.
SELECTED TRANSPORTATION CONTROL SIMULATION
OUTPUTS FOR NEW YORK AND CHICAGO DISAGGREGATED
BY TRIP PURPOSE (FIGURES ARE ON A DAILY BASIS)
New York
Work Trips
Shopping Trips
Other Trips
Total
Chicago
Work Trips
Shopping Trips
Other Trips
Total
Reduction in
Automobile
Passenger
Trips
750, 000
1, 390, 000
1,800, 000
3, 940, 000
357, 000
513, 000
1,479, 000
2, 349, 000
Increase
In Bus
Passenger
Trips
422, 000
0
0
422,000
195, 000
126, 000
318, 000
639, 000
Increase
In Rail
Passenger
Trips
330, 000
90, 000
50, 000
470, 000
162, 000
124, 000
527, 000
812, 000
Cancelled
Trips *
0
1, 300,000
1, 750,000
3, 048, 000
0
263,000
634, 000
898,000
Source: MATHAIR outputs.
* Figures may not add exactly to totals shown in the last column of this table or in
Table 17 due to rounding.
58
-------
automobile trips for other purposes. In Chicago about 55 percent of these
trips are replaced by bus or rail trips while the percent in New York is
negligible. We can only speculate about the source of these differences.
One possible explanation is that substitution possibilities between automobiles
and mass transportation exist in Chicago while in New York the automobile
is reserved for situations when substitution by mass transit is difficult.
A contributing factor to the relatively high marginal benefits of
S4. 1 in New York and Chicago (or, more exactly, to the relatively low
benefits in Los Angeles) is the time of day of most transfers from automo-
bile to mass transportation. In these two regions (and in Washington, D. C. ),
mass transportation is heavily used for work-trips (shown in Table 19) at
peak hour occupancy rates. By contract, bus transportation is little used
for work trips in Los Angeles, thus reducing the average bus occupancy
rate''~ below that of the other regions. Transfer of a given number of
passengers to bus therefore requires more buses, involving more pollution,
than in the other regions.
TABLE 19. WORK TRIP BY MODE CHOICE FOR THE
FOUR STUDY REGIONS (1970 data)
New York
Chicago
Washington, D. C.
Los Angeles
% of Employed
Central City
Residents Using:
Auto
30
60
57
90
Transit
70
40
43
10
c'j of Employed Non-
Central City
Residents Using;
Auto
82
88
91
97
Transit
18
12
9
3
Source- Calculated from AIP-MVMA, Urban Transportation
Factbook (1974), p. 1-22, assuming that all work trips
use either automobile or mass transit.
MATHAIR uses a peak bus-occupancy rate for work trips and an off-peak
rate for shopping and other trips. These two rates remain constant even
if demand for bus trips increases.
59
-------
In sum, our analysis suggests that automobile use restrictions may
be relatively successful in regions -with well-developed networks of mass
transportation for two reasons: first, travellers are more likely to forego
the trip by automobile. Second, they are less likely to switch to bus, the
second most polluting means of transportation.
60
-------
SECTION IX
SENSITIVITY ANALYSIS
INTRODUCTION
An important step in the development of any simulation model is to
test the sensitivity of the model to changes in the variables and parameters.
Sensitivity results indicate the stability of the model's calculations -with
respect to changes in the data, identifying those data series for which small
variations may have a significant impact on the simulation outcomes. In
this section we report the results of a limited sensitivity analysis of the
MATHAIR model.
In the first case, strategies SI, S3, and S4. 1 are
simulated for all four study regions using an
accelerated rate of removal of automobiles from
the stock.
Strategies SI, S3, and S4. 1 are simulated for Los
Angeles making some "worst case" assumptions
about several variables.
Two alternative transportation control measures
are defined, and strategies S4. 2 and S4. 3 are
constructed by substituting these new measures
for the transportation controls of S4. 1. Outcomes
for the three strategies are compared for Los
Angeles and New York.
ACCELERATED RATE OF REMOVAL FROM AUTOMOBILE STOCK
Nature of the Change
An input to MATHAIR 's Automobile Stock Module is the automobile
scrappage rate. Strategies SI, S3 and S4. 1 were simulated for all four
regions using a dramatically accelerated automobile scrappage rate.
Specifically, the proportion of automobiles i years old still on the road,
Pi, was assumed to be given by
P =
61
-------
where P, is the proportion still on the road after k years in the standard
case. The impact of this substitution is indicated by Table 20, which gives
the values of P. and Pi . The effects of this change in scrappage rate are;
reduced size of automobile stock
reduced proportion of older automobiles in stock
increased number of miles driven per automobile.
TABLE 20. PERCENTAGE OF AUTOMOBILES OF GIVEN
AGE STILL IN THE STOCK
Age of
Automobile
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Standard Case
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
100. 0
99.0
98.0
96.0
92.0
85. 0
74.0
60. 0
44. 0
32. 0
23.0
17.0
12.0
9.0
6.0
4. 0
3. 0
2.0
1.0
0.5
Assuming Accelerated
Scrappage Rate
(P/)
100. 0
99.0
97.0
93. 1
85.7
72. 8
53. 9
32.3
14. 2
4.6
1. 1
0.2
0.0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
0. 0
62
-------
Impacts on Benefits and Costs
The benefits and costs for strategies SI, S3, S4. 1,
calculated with the accelerated scrappage rate schedule, are shown in
Table 21 with net benefits for the standard case in parentheses. ''*
Accelerating the scrappage rate results in somewhat lower costs, higher
benefits and thus higher net benefits than in the standard case. Lower
costs clearly result from the reduced size of the automobile stock and,
in particular, a smaller number of automobiles with pollution control
devices. Because the average automobile is driven more miles than in
TABLE 21.
BENEFITS AND COSTS FOR SELECTED
STRATEGIES IN THE FOUR STUDY
REGIONS ASSUMING AN ACCELERATED
SCRAPPAGE RATE (1975 PRESENT VALUE
IN BILLIONS OF 1973 DOLLARS)
Los Angeles
Strategy 0
Strategy 1
Strategy 3
Strategy 4. 1
New York
Strategy 0
Strategy 1
Strategy 3
Strategy 4. 1
Washington, D. C,
Strategy 0
Strategy 1
Strategy 3
Strategy 4. 1
Chicago
Strategy 0
Strategy 1
Strategy 3
Strategy 4. 1
Costs
1. 7
2.3
3.4
3.0
3. 9
6.3
0.68
0. 73
1. 1
1.7
2.2
3.7
Damages
7.8
3.9
2. 7
2.3
28. 1
25.0
24. 3
23. 1
1. 1
0. 5
0.4
0.3
18 3
15.8
15.2
14. 0
Benefits
3.8
5. 1
5. 5
3. 1
3. 8
5. 0
0. 6
0. 7
0. 8
2. 5
3. 0
4. 3
•Jf
Net Benefits'"
2.1 (1.7)
2.9 (2.5)
2.1 (1.7)
0.2 (- .4)
-0. 1 (- .5)
-1.3 (-1.7)
-0. 04 (-. 14)
-0.04 (-.13)
-0.3 (-.5)
0.81 (.46)
0.83 (.50)
0.6 (.48)
Numbers in parentheses are the net benefits in the standard case from
Tables 12 to 15,
Corresponding benefits and costs for the standard case are reported in
Tables 12 through 15.
63
-------
the standard case, each automobile causes more pollution. Therefore,
a given percentage decrease in automobile emissions from a control device
will mean more units of pollution reduction than in the standard case.
Since the benefits of a strategy equal the product of the dollar damage per
unit of pollutant and the number of units of pollutant reduced by the strategy,
benefits are higher in the case of the accelerated scrappage rate.""
The optimal strategies in Los Angeles, Chicago and Washington remain
unchanged (S3 for Los Angeles and Chicago and no controls for Washington).
In New York, the net benefits for SI are now positive and this strategy is
preferable to one involving zero controls.
We conclude that small variations in the scrappage rate -will have
only a small impact on strategy outcomes since even the steeply accelerated
scrappage rate we have simulated produced a relatively minor impact.
A "WORST CASE" CALCULATION FOR LOS ANGELES
Nature of the Change
Several of data series input to MATHAIR are subject to considerable
uncertainty. This is evidenced by official "up-dates" which often represent
significant changes in the series and by widely different estimates obtained
by different researchers. As an example of how the model's results can be
sensitive to data uncertainty, we conducted simulations for Los Angeles
based on a set of "worst case" assumptions. These assumptions, designed
to reduce benefits and increase cost, are as follows;
10% increase in costs of emission control devices
lower base year damage estimates
decreased automobile control drive efficiencies
The meaning of the first is straightforward. The other two assumptions
require explanation. The base year damage estimates used in the standard
case simulations are based upon estimates of Barrett and Waddell extended
by Babcock and Nagda. ''"'" However, Justus et al. propose a range of
nationwide estimates that are considerably lower.v>'"v For our "worst
case" simulations, we have used Justus's "Lower bound" estimates, which
are less than twenty percent of the standard case estimates. Low estimates
of pollution damage in the base year mean a low damage cost per unit of
pollution and hence a low benefit per unit of pollutant reduced by a control
strategy.
This argument was also made for Los Angeles in Section VIII.
Reported in C. G. Justus, J.R. Williams, and J. D. Clement,
Economic Costs of Air Pollution Damage, prepared for Southern
Services, Inc., Birmingham, Alabama (May 1973).
Justus et al. , ibid. Both sets of estimates are shown in Tables F-35
and F-36 of Appendix F.
64
-------
The automobile emission characteristics used in the standard case
are from the EPA report Compilation of Air Pollution Emission Factors -
AP42 (April 1973), as updated (Supplement No. 2, September 1973). An
unedited preliminary edition of Supplement No. 5 has recently been made
available (April 16, 1975). In this document, the approach to calculating
mobile source pollutants has been modified in several ways. Most
significantly, measured and estimated new car emissions for recent
vintages have been increased considerably (over AP42 estimates) and, at
least for California, reported new car emissions no longer show a mono-
tonic decrease for more recent model years. * If these new estimates are
correct, it means that automobile control devices employed since 1967
(1965 in California) have not been as effective as previously claimed.
We have not incorporated the changes in Supplement No. 5 into our
standard case both because of the tentative nature of the data and because
the level of effort required would be prohibitive for a project near
completion. However, for purposes of the "worst case" sensitivity analysis,
we have developed a set of emission characteristic estimates based on
Supplement No. 5. *>;< While these estimates are monotonic (decreasing with
more recent vintage year), they do reflect the higher emission levels'1"'"0 of
the more recent document. Emissions under the zero control strategy
will be the same as in the standard case, but other strategy outcomes
should be affected in two ways: air quality will deteriorate relative to
the standard case, and benefits should decrease. The latter consequence
results from a smaller number of units of pollutant reduction.
The scenario incorporating these three kinds of data changes is a
"worst case" from the point of view of air pollution control because each
change contributes to a deterioration of air quality, an increase in strategy
costs and/or a reduction of strategy benefits.
Impacts on Air Quality
MATHAIR air quality predictions for the "worst case" are shown in
Table 22 along with the air quality standards. Pre-control air quality,
corresponding to the zero control strategy, is the same as in the standard
case. Air quality under strategies Si, S3, and S4. 1 can be compared
against the standard case results for Los Angeles shown in Table 8.
For example, new car CO emissions for California are reported in
Supplement No. 5 as follows (based on 1975 Federal Test Procedure):
1971 42. 3 gr/mi (measured, p. 33).
1972 46. 7 gr/mi (measured, p. 35).
1973 37.0 gr/mi (estimated, p. D10).
'< >'<
These estimates and the standard case automobile emission estimates
are shown in Tables F-21 and F-22 of Appendix F.
>!' -f
For 1975 models driven in California, for example, Supplement No. 5
estimates of emissions per mile of CO, HC, and NO are more than
75% higher than corresponding estimates in Supplement No. 2.
65
-------
TABLE 22.
AIR QUALITY IN LOS ANGELES FOR SELECTED
STRATEGIES USING "WORST CASE" AUTOMOBILE
POLLUTANT EMISSIONS CHARACTERISTICS (PPM)
Strategy
Zero
Control
Strategy
Strategy 1
Strategy 3
Strategy 4. 1
Year
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Standards
Carbon
Monoxide
(CO)
54. 7388
55. 7445
56. 7804
57. 8472
58. 9462
60. 0681
61. 2439
62.4447
63. 6815
64. 9555
38. 8726
37. 5759
36. 6050
36. 0123
35. 7826
35. 7810
35. 9872
36. 3410
36. 8083
37. 3665
32. 1118
27.4362
22. 9741
19. 1371
16. 0006
13. 5364
11. 8848
10. 6463
9.8661
9. 3433
18. 8779
16. 6658
14. 5358
12. 3258
10.4562
8. 8932
7.7781
6. 9082
6. 3176
5. 9131
9.0
Oxidants
(OX)
0. 5096
0. 5169
0. 5246
0. 5326
0. 5409
0. 5495
0. 5585
0. 5677
0. 5773
0. 5872
0. 2900
0. 2761
0. 2660
0. 2597
0. 2560
0. 2543
0. 2564
0. 2797
0. 3015
0. 3218
0. 2477
0. 2205
0. 1974
0. 1791
0. 1647
0. 1539
0. 1491
0. 1665
0. 1835
0. 1998
0. 1944
0. 1776
0. 1635
0. 1508
0. 1407
0. 1328
0. 1291
0. 1425
0. 1555
0. 1679
0.4720
0.4786
0.4855
0.4926
0. 5000
0. 5077
0. 5156
0. 5238
0. 5323
0. 5410
0. 3285
0. 3188
0. 3188
0. 3078
0. 3063
0. 3062
0. 3091
0. 3263
0. 3427
0. 3582
0. 2908
0. 2671
0. 2465
0. 2300
0. 2173
0. 2079
0. 2037
0. 2149
0. 2266
0. 2380
0. 2397
0. 2246
0. 2115
0. 1942
0. 1795
0. 1672
0. 1592
0. 1637
0. 1691
0. 1753
0. 4483
0. 4544
0. 4607
0. 4672
0.4740
0. 4811
0. 4883
0.4958
0. 5036
0. 5116
0. 3428
0. 3340
0. 3276
0. 3240
0. 3228
0. 3230
0. 3260
0. 3425
0. 3579
0. 3722
0. 3056
0. 2819
0. 2609
0. 2439
0. 2307
0. 2207
0. 2162
0. 2280
0. 2396
0. 2507
0. 2535
0. 2380
0. 2244
0. 2060
0. 1903
0. 1771
0. 1684
0. 1713
0. 1746
0. 1785
. 08
Nitrogen
Oxide
(NOX )
.X.
0. 0637
0. 0645
0.0653
0. 0662
0. 0671
0. 0680
0. 0689
0.0699
0. 0709
0. 0720
0. 0654
0. 0652
0. 0651
0. 0653
0.0657
0. 0662
0. 0669
0. 0677
0.0686
0. 0695
0.0609
0. 0582
0.0557
0.0536
0. 0522
0.0511
0. 0506
0. 0504
0. 0505
0. 0508
0. 0534
0. 0515
0. 0497
0. 0454
0. 0416
0. 0381
0. 0355
0. 0333
0. 0317
0. 0306
. 05
Sulfur
Oxides
(SOX )
!X
0. 0158
0. 0158
0. 0158
0. 0159
0. 0159
0. 0159
0. 0159
0. 0160
0. 0160
0. 0160
0. 0158
0. 0158
0. 0158
0. 0159
0. 0159
0. 0159
0. 0159
0. 0160
0. 0160
0. 0161
0. 0158
0. 0158
0. 0158
0. 0159
0. 0159
0. 0159
0. 0159
0. 0160
0. 0160
0. 0161
0. 0156
0. 0156
0. 0157
0. 0157
0. 0157
0. 0158
0. 0158
0. 0158
0. 0158
0. 0159
. 03
Source: MAT HAIR outputs
66
-------
While there is considerable pollution reduction with respect to the
zero controls levels, the "worst case" results for Si show no further
improvement in air quality over the 1975 to 1984 simulation horizon. After
a small initial decline, concentrations of all pollutants increase and, except
for carbon monoxide, eventually exceed the 1975 levels. The reduction in
pollution due to devices required by SI is apparently offset by pollution
increases from the growing use of automobiles with "worst case" emission
characteristics.
None of the air quality standards can be met by SI in either the
standard case or the "worst case. " Strategy S3, by contrast, effects sig-
nificant pollutant concentration reductions in both cases. Still, the
standards for CO, OX, and NOX cannot be met in the "worst case" while
only the OX standard is not met in the standard case. Strategy S4. 1 also
results in a significant reduction in pollutant concentrations in both cases.
While concentrations of all pollutants are higher than in the standard case,
in both cases all standards, except that for oxidants, can be met by 1984.
The "worst case" assumes reduced efficiencies for all automobile
pollution control devices specified in strategy Si. The efficiencies of
devices for new automobiles built after 1975 and of retrofit devices for
old automobiles specified by strategies S3 and S4. 1 are not reduced by
the "worst case" assumptions. Hence, the latter two strategies are more
effective in improving air quality.
Impacts on Strategy Benefits and Costs
Table 23 shows the benefits and costs of SI, S3, and S4. 1 using
"worst case" assumptions in Los Angeles (with net benefits under the
standard case given in parentheses). As anticipated, control costs for all
strategies are 10% higher in the "worst case. " "Worst case" benefits are
considerably lower than standard case benefits for all strategies, reflecting
both the decreased value per unit of pollutant reduction and the smaller
number of units reduced.
TABLE 23.
BENEFITS AND COSTS FOR SELECTED
STRATEGIES IN LOS ANGELES USING
"WORST CASE"* ASSUMPTIONS (1975
PRESENT VALUE IN BILLIONS OF 1973
DOLLARS)
Strategy 0
Strategy 1
Strategy 3
Strategy 4. 1
Costs
2. 1
2.6
4. 0
Damages
1. 0
0.6
0.4
0. 3
Benefits
0.4
0.6
0. 7
Net Benefits**
-1.7 (1.7)
-2.0 (2.5)
-3.3 (1.7)
• 10% control device cost increases
• Justus lower bound damage estimates
• Emissions adapted from Supplement No. 5 to AP 42.
**Numbers in parentheses are standard case results from
Table 12.
67
-------
While all strategy benefits are greater than the corresponding costs
in the standard case, all "worst case" strategy costs are greater than the
corresponding benefits. The optimal strategy is to require no controls, and
this remains true even if costs are not increased by ten percent.
We conclude that if the California standard case efficiencies of
automobile control devices and the base year estimates of damage due to
air pollution are seriously overestimated, the benefits due to strategies
like those -we have considered are equal to only a fraction of their cost.
THREE TRANSPORTATION CONTROL, MEASURES IN LOS ANGELES AND
NEW YORK
Nature of the Measures
Our most stringent strategy, S4. 1, includes a transportation
control measure involving a fifty percent increase in the cost per mile of
using the automobile. In order to test the sensitivity of MATHAIR to
different transportation controls, we have devised two other versions of
S4, S4. 2 and S4. 3, involving the folio-wing transportation controls:
S4. 2 Bus fares decreased by one-third through a
government subsidy
S4. 3 Cost per mile of automobile use increased by
half and bus fares decreased by one-third.
Impact on Benefits and Costs
The benefits and costs of strategies S4. 1, S4.2, and S4. 3 for Los
Angeles and New York are shown in Table 24. The benefits attributable
to each strategy, in both regions, are about the same (five billion dollars).
While the costs in New York are also comparable for the three strategies,
S4. 3 is considerably more costly than the other strategies in Los Angeles.
In Los Angeles, the net benefits of S4. 1 and S4. 2 are positive while
the costs of S4. 3 are slightly higher than the benefits. The optimal choice
among the three strategies is S4. 2, but the optimal choice among all
strategies simulated is still S3.
In New York, the net benefits are negative for all three strategies.
However, as in Los Angeles, the best (or least undesirable) choice of the
three is S4. 2 and the most costly is S4. 3.
It is often said that effective transportation controls should include
a constraint on the use of automobiles. While our results show decreased
bus costs (S4. 2) to be more beneficial than increased automobile costs (S4. 1),
this may be due to the magnitude of cost changes selected for our simulations.
However, the least beneficial of the three measures is the simultaneous
change in both costs indicating that more transportation controls are not
necessarily better than fewer.
68
-------
TABLE 24. BENEFITS AND COSTS OF THREE TRANSPORTATION
MEASURES IN LOS ANGELES AND NEW YORK (1975
PRESENT VALUE IN BILLIONS OF 1973 DOLLARS)
Los Angeles
Strategy 0
Strategy 4. 1
Strategy 4. 2
Strategy 4. 3
New York
Strategy 0
Strategy 4. 1
Strategy 4. 2
Strategy 4. 3
Costs
3.6
3. 1
5. 2
6.6
5. 3
6. 8
Damages
7.8
2. 5
2.4
2.6
28. 1
23. 3
23. 7
23. 1
Benefits
5. 3
5. 1
5. 2
4.9
4.4
5. 0
Net Benefits
1. 71
1.95
-0. 02
-1. 77
-0. 87
-1. 80
69
-------
SECTION X
REFERENCES
Environmental Protection Agency
Aircraft Emissions; Impact on Air Quality and Feasibility of Control,
(undated, assumed 197Z).
Air Quality Data, 1972 Annual Statistics, EPA-450/a-74-001 (March 1974).
Barrett, L. B. and N. L. Nagda, "The Cost of Air Pollution Damage: A
Status Report," (1971), cited in Justus, C.G. , J.R. Williams,
and J.D. Clement, Economic Costs of Air Pollution Damage,
STAR, Inc. (May 1973).
Compilation of Air Pollution Emission Factors, AP-42 (April 1973),
Supplement No. 2 (September 1973) and Supplement No. 5
(April 1975).
The Cost of Clean Air, Annual Report of the Administrator Environmental
Protection Agency to the Congress of the United States (April 1974).
Crenshaw, J. and A. Basala, "Analysis of Control Strategies to Attain
the National Ambient Air Quality Standard for Nitrogen Dioxide, "
presented to the Washington Operation Research Council's Third
Cost-Effectiveness Symposium, National Bureau of Standards (1974).
Heavy Duty Vehicle Driving Pattern and Use Survey - New York City,
prepared by Wilbur Smith and Associates, APTD-1523 (May 1973).
Hydrocarbon Pollutant Systems Study, EPA Contract No. 71-12, MSA
Research Corporation, Vol. 1 (1972).
Horowitz, J. , "Transportation Controls, " prepared for publication in
Environmental Science and Technology (1974).
Horowitz, J. and S. Kuhrtz, Transportation Controls to Reduce Automobile
Use and Improve Air Quality in Cities, EPA-400/ 11-74-002 (1974).
1975 County Emissions Summary Reports, National Emissions Data. System.
Prediction of the Effects of Transportation Controls on Air Quality in
Major Metropolitan Areas, APTD-1363 (November 1972).
70
-------
Stationary Internal Combustion Engines in the U.S. (1973).
Systems and Costs to Control Hydrocarbon Emissions From Stationary
Sources, Emission Standards and Engineering Division (August 1973).
Technical Support Document for the Metropolitan Los Angeles Intrastate
Air Quality Control Region Transportation Control Final
Promulgation, Region IX (October 30, 1973).
Transportation Control Strategy Development for the Metropolitan Los
Angeles Area, APTD-1372 (December 1972).
Transportation Control Strategy Development for the New York Metropolitan
Area, APTD-1371 (December 1972).
Waddell, J. , Economic Damages of Air Pollution (1974).
Other References
Auto Facts and Figures, Motor Vehicle Manufacturers Association, (1972
and 1973 editions).
Babcock, L. R. and N. L. Nagda, "Cost Effectiveness of Emission Control,"
Journal of the Air Pollution Control Association, 23, 173-179 (1973)
cited in Justus, C.G., J.R. Williams, and J. D. Clement, Economic
Costs of Air Pollution Damage, STAR, Inc. (May 1973).
Bostich, T. A. and H. J. Greenbalgh, Relationships of Passenger Car Age
and Other Factors to Mileage Driven, Washington, D. C. : U.S.
Department of Commerce, Bureau of Public Roads (1967).
Control Techniques for Nitrogen Oxide Emissions from Stationary Sources,
U.S. Department of Health, Education and Welfare, (March 1970).
Domencich, T. A. and D. McFadden, A Disaggregated Behavioral Model
of Urban Travel Demand. Cambridge, Mass., Charles River
Associates, Inc. (March 1972).
Federal Register, Vol. 7, No. 36, "Proposed Rule-Making: Regulation of
Fuels and Fuel Additives, " (1972).
Goeller, B. F. et al. , San Diego Clean Air Project, The RAND Corporation,
R-1362-SD (December 1973).
Ingram, G. K. , and G. R. Fauth, TASSIM: A Transportation and Air Shed
Simulation Model, U.S. Department of Transportation (May 1974).
Justus, C. G. , J.R. Williams, and J. D. Clement, Economic Costs of Air
Pollution Damage, STAR, Inc. (May 1973).
McFadden, D. and F. Reid, Aggregate Travel Demand Forecasting From
Disaggregated Behavioral Models (February 1974, revised).
71
-------
Mikolowsky, W. T. et al. , The Regional Impacts of Near-Term Transportation
Alternatives: A Case Study of Los Angeles, the RAND Corporation,
R-1524-SCAG (June 1974).
Moran, John B. , Lead in Gasoline: Impact of Removal on Current and
Future Automotive Emissions, for Air Pollution Control Association
(1974).
Motor Truck Facts, Motor Vehicle Manufacturers Association, (1973).
Report by the Committee on Motor Vehicle Emissions, National Academy
of Sciences, (November 1974).
Status of Regional Public Transportation Service and Use, Tri-State
Regional Planning Commission (May 1974).
Stoddard, L. and P. Downing, The Economics of Air Pollution Control
for Used Cars, Riverside: University of California, working
paper No. 15.
A Survey of Average Driving Patterns in Six Urban Areas of the United
States, System Development Corporation, (1971).
Ter Haar et al. , "Composition, Size, and Control of Automotive Exhaust
Particulates, " APCA Journal, Vol. 22, No. 1 (January 1972).
Turner, D. B. , Workbook of Atmospheric Dispersion Estimates, U.S.
Department of Health, Education and Welfare, (March 1970).
Urban Transportation Factbook. American Institute of Planners - Mctor
Vehicle Manufacturers Association, (1974).
Vehicle Miles of Travel on Major Roadways: 1970, Tri-State Regional
Planning Commission, (November 1973).
72
-------
APPENDIX A
METRIC CONVERSIONS
Conversion of parts per million (ppm) micrograms per cubic meter
) for five pollutants.
Carbon Monoxide
Hydrocarbons
Oxidants
Nitrogen oxides
Sulfur oxides
1 ppm. in //g/m'
1140
667
2000
2000
2600
Conversion of grams per mile (gr/mi) to grams per kilometer
(gr/km).
1 gr/mi = 1.6 gr/km
Conversion of tons (T) to kilograms (kg).
IT = 907 kg
73
-------
APPENDIX B
DESCRIPTION OF MATHAIR
In this appendix we provide the equations for the following six
MATHAIR modules:
Automobile Stock Module
Transportation Module
Emissions Module
Air Quality Module
Benefits Module
Cost Module
MATHAIR user documentation is given in Appendix H.
AUTOMOBILE STOCK MODULE
The following four equations define the structure of this module:
j
(A.I) VEH, . = REG, ... x II JUNK; (for base year only)
v ' byr,j byr-j+1 1=1
(A. 2) VEH L x (1 + GROW)(y"byr); (for new cars only)
(A. 3) VEH . = VEH , . . x JUNK. (general case)
YJ y-1 J-1 J
y £ byr, j ± 1
where VEH . = number of vehicles that will be of age j at the
y^ end of year y
REG = number of new car registrations in year y (for
y past years y < byr)
Note; The first year of the simulation horizon, byr, is 1975 in our test
runs and the old car stock includes vintages early as 1955. Both
of these years are variables in MATHAIR. For example, the user
could specify the simulation horizon to begin in byr = 1980 .
74
-------
JUNK. = 1 - scrappage rate of vehicles of age j
GROW = presumed annual growth rate of new car registrations
after base year
TRANSPORTATION MODULE
Model Type and Capabilities
MATHEMATICAL transportation demand module is a disaggregated
behavioral model based on the work of Domencich and McFadden. The
module provides the following outputs:
Mode-choice probabilities
Trip-frequency probabilities
Joint travel demand probabilities
Number and composite price of auto, bus, and
rail passenger trips
Auto, bus, and rail passenger miles
Auto, bus, and rail vehicle miles of travel (VMT)
All outputs are calculated for three travel modes, three trip
purposes, and four spatial trip patterns.
Module Logic
The decision tree in Figure B-l is the individual's travel decision
process assumed by the Transportation Module. It corresponds to the
following sequence of choices:
• The purpose of the trip is identified: work, shop,
or other purpose.
• Auto, bus, or rail is chosen by the decision-maker
after evaluating the relative cost and performance of
each alternative.
• The probability of making or not making the trip
(trip frequency) is determined as a function of the trip
purpose and of the specific cost of travel by the
preferred mode.
• Trip time of day, while not explicitly identified in the
model, is a function of the purpose of the trip. Work
trips are assumed to take place during peak-hour
(high vehicle occupancy). Shopping and other-purpose
trips are assumed to be undertaken at other times of
the day.
In the model of Domencich and McFadden, trip frequency is determined
before mode choice.
A flow chart of the Transportation Module used in MATHAIR is
shown in Figure B-2.
75
-------
1
a
0)
to
rt
(1)
-------
V.
o
to
c
rt -c.
r, 2
'N
V
u
c
c.5
'C *
HQ
^-/
0
u
u
1
1
en
^i
^5i
U
.cL S
Hfc
3
T)
O
O
• H
-tJ
.2
o
P.
2
rt
EH
0>
O
^
o
V
to
-------
Definition of Variables and Major Equations in Transportation Module
AC-' = Automobile parking charges and operating costs (in
dollars).
AIVJ = Automobile in-vehicle time (in minutes) by purpose.
AO -1 = Average vehicle occupancy for travel by i mode
for j"1 purpose.
ATD •* = Average trip distance by i mode for j purpose.
AVOTR = Average value of time riding.
AVOTW = Average value of time waiting and walking.
FARE •* = The fare by i mode for j purpose.
Pi •* = Probability an individual will choose the automobile
for a trip of purpose j rather than the bus.
PT -1 - Probability an individual will choose the automobile
for a trip of purpose j rather than the train.
H = Probability an individual will choose the bus for a
trip of purpose j rather than the automobile mode
(HBJ = 1 - HA'J).
HF = Probability an individual will make a trip.
IT! • *
H " = Probability an individual will choose the train for a trip
of purpose j rather than the automobile (H^j = 1 - H-^-'J)
NT •' = Number of trips by i mode for j purpose.
P" = Joint possibility of taking a trip by i mode for j
purpose.
P1-1 = The price of travel for the average trip by i mode
and j purpose.
TSS •* = Transit station to station time (in minutes) plus wait
time by mode and purpose.
TW1-' = Transit walk time (in minutes) for the i mode and
the j"1 purpose.
UT-' = Potential number of trips for j purpose.
VMT1-' = Vehicle miles traveled by i mode for j purpose.
78
-------
Modal Split- -
(A. 4) log
(A. 5) log
Automobile-Bus Modal Split
HAj HAJ
1-HAJ
= log
H
Bj
= a + a
B.
«3(ACJ - FARE J)
Automobile-Train Modal Split
H
A'j
= log — =, - = a +a TWTj + a (AIVj - TSSTj)
Tj ° l 2
H
-
+ a3(ACJ - FARE J
Normalized Modal Split Probabilities
(A. 6)
HAJ+HBJ+HTJ
J = triP Purpose
/ A .71 i
( ) g
Trip Frequency
. HF
r , r Aij
^° l
i = Trip mode
j = Trip purpose
(A. 8)
(A. 9)
(A. io)
Mode Specific Cost Equations
Auto Trip Cost
PAj'
x AVOTRJ+ ACJ'
Bus Trip Cost
pBi = (
(
TWBJ x AVOTW + (TSSBJ" x AVOTR) + FAREBJ
Rail Trip Cost
) + (
AVOTW + TSSTJ' x AVOTRJ+ FARET;i
79
-------
Joint Probabilities--
(A.inp1' = (Hij)x(Wij) i = trip mode
^ ' v ' j = trip purpose
where H ^ is assumed equal to one for work trips.
Transportation Module Outputs --
• Number of Trips Per Day
(A. 12) NTj = UTJ x P^ i = triP mode
j = trip purpose
• Passenger Miles Traveled Per Day
mode
(A. i 3) PMT = NT x
v • ' Jrxvij- j = trip purpose
• Vehicle Miles Traveled Per Day
(A. 14) VMTj, =
EMISSIONS MODULE
Automobile Emissions
The following equations are used to calculate automobile emissions.
First it is necessary to calculate the number of miles driven by automobiles
of each age:
(A. 15) AMILE . = VMT . x VEH . x MILE. / >. VEH x MILE
yj yi yj J/ La ya a-J
where AMILE . = average daily mileage driven by automobiles
^ of age j in year y
VMT = automobile vehicle miles traveled/day (VMT
yL is bus vehicle miles traveled (VMT)) y
MILE. = average daily miles presently driven
J by an automobile of age j (exogenous)
/ ^ = summation is over all ages in stock
a
80
-------
Automobile VMT, needed in the above equation, may be provided exogenously
or else generated as an endogenous variable by the Transportation Module.
Emissions are then calculated as follows:
(A. 16) Ei>poll - ^ CCpoll X Cpoll, vint X Dpoll, vint, age X SCspeed, poll
age
x AMILE. x Q
i, age I 907,185
CC
Poll
3 • ) + P •, •
1, i.age' d, i, age
(l-Rd, poll, age)
= rr (i-p, . x R_ )
d d, i, age d, poll, age'
where E.
i, poll
CC
poll
poll, vint
poll, vint, age
, ,,
speed, poll
emissions in tons /day of pollutant poll
in year i
improvement factor due to devices
emissions in grams/mi, of pollutant
poll, from average unmodified automobile
of vintage vint
deterioration factor ( 1. 0) as automobile ages
speed correction factor (corrects for
difference between actual speed and
speed at which standard emission
measurements were made)
AMILE.
i, age
average daily mileage in year i by
automobile of given age.
1
907, 185
conversion from grams to tons
P, .
d, i, age
proportion of automobiles of given age
(vintage) on which device d is installed in
year i
81
-------
R
d.poll, age
Stationary Source Emissions
proportion of emission of pollutant poll
held back when device d is installed
in automobile of given age.
The following equations describe the calculation of stationary source
emissions:
(A. 18) S. „ = 7 SE .
v ' i, poll L-J s, i,
poll
s e S
(A. 19) SE . „ = SS .. x
s, i, poll s, poll
- (P , x SDEV, ..)
v sd d, poll'
where S.
i, poll
SE . ..
s, i, poll
SS
s, poll
TT
d
total stationary source emissions in
year i for pollutant poll in tons/day
set of all stationary sources
total daily emissions of pollutant poll in
year i from source s
emissions of pollutant poll from source
s in tons/day
product over all devices installed
AIR QUALITY MODULE
The following equations describe the Air Quality Module:
(A'20) ci,co
s
i,co
(A.21) C.
a2Si,NO
i, NO
x
Ci,OX
Si,HC
(A. 23) C.
ij S°x ~ °4 S i, SO + /3. M. qn
x 4i» ^o
(A. 24) C.
Si,Pb
82
-------
where C. ^ = concentration of pollutant poll in year i
,. = daily emissions of pollutant poll in year i
9 " -f •** *"\rvi o-f-raifir\Tlo i»Tr o /-MI -i*/•» o a
M. = daily emissions of pollutant poll in year i
' P from mobile sources
a. - city-specific stationary source rollback coefficients
•j*
ft. = city-specific mobile source rollback coefficients
BENEFITS MODULE
MATHAIR's module for calculating the economic costs of air
pollution damage can be bypassed if the user wishes to provide his own
benefits module.
The MATHAIR Benefits Module allows the user to specify up to
eight benefits categories (i = 1, . . . , 8) for each of the five pollutants
(j = 1, . . . , 5) or forty benefit/pollutant categories. For each, the following
information is required;
(i) dollar value of damages plus damage avoidance costs
in some base year (D..)
(ii) level of pollutant concentration producing baseline
damage and damage avoidance costs (C.)
(iii) rate at which damage (cost) will grow if pollutant
level remains constant (r.)
(iv) exponent for pollution concentration in damage
function ( y..)
ij
(v) threshold below which no damage occurs (T,.)
The damage function is assumed to be of the following form, where
the coefficients §.. are computed by MATHAIR from the baseline conditions;
(A.Z51D = 5..(C. -T..)^
*ctj and /3j are calibrated (outside of MATHAIR) from base year
estimated emissions and measured concentrations and then input to
MATHAIR. Values for a and ft are discussed in Appendix D.
83
-------
If the concentration of pollutant j is E. in year byr + k, then
damages in that year (discounted to year byr) -dill be equal to
WV x(1 + v " E-
J
otherwise
When a strategy is simulated, the Benefits Module computes the
resulting damage and damage avoidance costs. Benefits due to a strategy
can be calculated by subtracting associated costs from those produced in
the baseline case.
COST MODULE
Automobile Device Costs
MATHAIR requires both input data and strategy information
concerning air pollution control devices for automobiles. The input data
describe emission reduction characteristics, expected device life, and
device costs. Strategy information describes which devices should be
installed, on which automobiles, and in -what years.
The annual cost of increased fuel consumption and increased main-
tenance and the installed cost of the device are calculated by the Cost
Module as follows:
(A. 26) COST. , . = DCOST, , . x VEH.. x P. .. (l+R)1"byr
v ' i,fuel k, fuel ij kij
where byr = base year
i = a year between the year of installation
and the year of removal
COST. . = the incremental fuel cost in year i of
lj u having installed device k in autos of
age j
DCOST, . , = the annual unit increase in fuel cost for
k'fuel device k
VEH.. = the number of autos of age j in year i
P = the proportion of autos of age j in year
J i on which device k has been installed
R = the discount rate
84
-------
The calculations are analagous for the maintenance costs associated -with
a control device. *
If the device is installed in year i and either it is removed before
the simulation terminates, or the device has been on for the whole of its
economic life when the simulation terminates, the entire cost is attributed
to the simulation period. If neither of these conditions is met, only part
of the cost is attributed to the simulation period. In either case, the
annualized cost in year i is calculated as follows;
(A. 27)
where t = the last year simulated
A. = the annualized device cost
k
such that DCOST, , = A, + A. + . . . + A
k, dev k k_ k
1+R (L-l)
(1+R)
where L, = the maximum assumed device life or the number
of years of use before the device is removed,
whichever is less
DCOST, , = the purchase price of the k device
and the other variables are as defined above.
Stationary Source Device Costs
MATHAIR requires both input data and strategy information concerning
air pollution control of stationary sources. Input data include an inventory
of the stationary sources in each study region and device descriptions
(emission reduction characteristics, device lifetime, device costs). Strategy
information describes which devices should be installed, on which sources,
and in which years.
The cost equations are analagous to those already described for
automobile devices. Annual costs should reflect the economic value, if any,
of the recovered pollutant.
Changes in the cost of fuel, due to the effects of control devices on gasoline
consumption, are thus calculated as an annual cost per automobile although
they are incurred on a per mile basis. Strategies restricting the use of the
automobile do not affect the size of the stock but do reduce the number of
miles driven. In these cases, the gasoline costs are currently adjusted
manually after the model is run.
85
-------
Other Costs
The cost of reduced (or enhanced) traveler mobility is manually
calculated by the model user, using the method described in Appendix E.
The current version of MATHAIR also calculates the costs of new buses
required to meet any increase in demand for bus trips. Since we assume
fares will cover these costs, they have been manually netted out of the costs
presented in this report.
86
-------
APPENDIX C
FORECASTING AGGREGATE DEMAND WITH A DISAGGREGATED
TRANSPORTATION MODEL*
The disaggregated behavioral transportation model is estimated
with observations of individual travel behavior and is intended to provide
a basis for forecasting individual transportation demand probabilities.
This type of model can also be used to predict aggregate demand.
The probability of the jth individual's demanding an auto trip
rather than a bus trip, for example, can be expressed as
X a X '
i i i
P. = f(XJ, , ... Xi) = — ==, r
where the parameters (aj , i = 1 , . . . , n) are assumed to be constant
among individuals, but the probabilities (Pj) and differential level of
service attributes (X^ , i = 1 , . . . , n) are assumed to vary among
individuals. If these individual functions are added over all N persons
in the population and then both members of the equation are divided by
N, we get
N N
xn>
where P is the average probability for the population. In the very
special case where f is linear, then
= f(x1
Based on the discussion in D. McFadden and F. Reid, Aggregate Travel
Demand Forecasting from Disaggregated Behavioral Models, presented
at Fifty-fourth Annual Meeting of the Transportation Research Board,
Washington, D. C. (January 1975).
87
-------
and f is easily aggregated, with the average probability being equal to the
same function f of the average of the variables. Analytic aggregation
may be feasible even if f is not linear.
The function f specified above is derived from assumptions
about the individual's utility function. In the disaggregate behavioral
model described by Domencich and McFadden, the utility function is
assumed to contain a stochastic element to reflect the individual's
idiosyncracies in taste or unobserved variations in the observed attribute
vector X = JX^ , . . . , Xn \ . Assuming that the non-stochastic part of the
utility function is linear in the attributes, McFadden and Reid have
investigated the special case where the stochastic elements in the individual
utility function upon which the probability expressions are based is
standard normal (the probit model). In this case, a simple closed formula
can be obtained for the aggregate probability (using matrix notation):
P. = (a'XJ)
J
where P. is the probability of response (i. e. , choice of particular
J mode) by jth individual.
<£> is cumulative normal density.
a is the vector of estimated weights
X^ is the vector of differential impedance between modes
for the jth individual
X is assumed normal with mean X and covariance matrix A .
If we let a = a'Aa, then McFadden and Reid show that
p = x
V i + o-
In other words, the aggregate prediction obtained from using aggregate
data in the disaggregated model should be attenuated by a factor reflecting
the degree of heterogeneity of the variables facing the individuals. If the
area is homogeneous, then cr^ = 0 and the disaggregate model can be
used without correction factor for aggregate prediction.
The model used in this study is not a probit model, but a conditional
logit model, estimated by McFadden, in which the stochastic elements
are assumed to have a Weibull distribution. While McFadden makes this
assumption for computational facility within the disaggregate framework,
it does not lead to a simple closed formula for aggregate prediction.
88
-------
In this study we have implicitly assumed that zones are
homogeneous by making aggregate predictions with the disaggregated
model. According to McFadden and Reid, the nature of the bias with any
sigmoid-shaped choice function (and thus with any cumulative density
function) is toward the extreme possible values: too low for values below
0. 5, too high high above 0. 5 -- that is, the assumption of homogeneity
for a zone that is in fact heterogeneous means that demand elasticities
will be over-estimated.
Since -we recalibrate the constant terms (in all choice functions)
to ensure proper base-year predictions, the bias in our model enters
when we simulate policy change without calculating (or estimating)
accompanying change in the covariance matrix. As McFadden and Reid
state:
We note that specifying the precise effect of a policy
change on A [the covariance matrix A] may be
challenging; various approximations may become
necessary, including the extreme approximation . . .
that A is always zero and is therefore unchanged by
policy.
This difficulty is present even in the case of the probit model where an
analytic expression exists for the aggregate function. Given our use of
the logit model and the absence of such an analytic expression, we are
obliged to accept the error introduced by this approximation --an
over-estimation of responsiveness to policy change.
89
-------
APPENDIX D
DISPERSION PROPERTIES OF MOBILE SOURCE
AND STATIONARY SOURCE EMISSIONS
This somewhat technical appendix draws on material from
D. Brace Turner's Workbook of Atmospheric Dispersion Estimates. *
The purpose of the appendix is to display the results of some calculations
comparing the dispersion properties of non-reactive air pollutants from
stationary and from mobile sources. The results make it possible to
compare the cost-effectiveness of pollution control strategies for the two
kinds of sources. If we assume that the ambient concentration of a
pollutant is linear in emissions and is equal to a times mobile source
emissions plus /3 times stationary source emissions, then we can show
that the ratio a//3 ranges over several orders of magnitude. The value
of the ratio depends on such variables as climatic stability, wind
direction, and distance from the source. If a single value must be
chosen, we propose 100 based on preliminary calculations reported in
this appendix.
Mobile and stationary sources differ in two essential respects:
height of emission and spatial aggregation. Mobile sources have an
emission height of at most a few meters; stationary sources have a
height of up to 200-300 meters. Furthermore, a given quantity of
pollutant emitted from a single stationary source is obviously more
concentrated than the same total quantity emitted from several mobile
sources. We will now examine the consequences of these differences.
CONSEQUENCES OF STACK HEIGHT
The height of emissions (H) is the height of the plume centerline
when it becomes essentially level. In -what follows, we will assume that
H = 10 meters for mobile source and H = 100 meters for stationary
sources. We assume that the plume has a Gaussian distribution in both
the horizontal and vertical planes and that there is no deposition or
reaction at the surface. We originate a coordinate system at the base
of the stack and are interested in calculating concentration (g/mr) at
(x, y, z,) where the x-axis is parallel to the wind direction. The system
is illustrated in Figure D-l.
*D. Bruce Turner, Workbook of Atmospheric Dispersion Estimates, U. S.
Department of Health, Education and Welfare (1969).
90
-------
(x,-y,Z)
(x,-y,0)
Figure D-l.
Coordinate system showing Gaussian
distributions in the horizontal and
vertical.
Source: D. Turner, op. cit.
When temperature decreases with height at a rate greater than
5.4 per 1,000 feet (1°C per 100 meters), the atmosphere is in
unstable equilibrium (and vertical motions are enhanced). When tempera-
ture decreases at a lower rate or increases with height (inversion),
vertical motions are damped or reduced and equilibrium is stable.
Stability categories can be designated on the basis of surface wind speed
and amount of incoming solar radiation (day) or cloud cover (night)
illustrated in Figure D-2. The categories range from A (unstable)
through F (stable), and D is considered the neutral class.
91
-------
<00r
TEMPERATURE, 'C
WIND STEED, n/uc
Figure D-2.
Examples of variation of temperature and -wind
speed with height (after Smith, 1963).
Source: D. Turner, op. cit.
GROUND LEVEL CONCENTRATION DOWNWIND
In the Workbook of Atmospheric Dispersion Estimates, Turner
plots relative ground-level (i. e. , z = 0), down-wind (i. e. , y = 0" concen-
trations times windspeed ( Yu/Q, where X is concentration, u is
windspeed, Q is emission) against distance (i. e. , for x = . 1 to 100
kilometers), at various heights of emission (H) and various limits to
vertical mixing, for each stability class. We will fix the downward
distance at one kilometer. Given that Xu/Q = Y, then X = YQ/u. Com-
paring Y/u (or simply Y) for H = 10 meters and for H = 100 meters,
under otherwise identical conditions, we find the following values on a
neutral day:
H = 10
Y
-4
H = 100
-10
-6
Given the assumptions we have made, the multiplier for mobile sources
(a) is about 100 times greater than for stationary sources (/3)> The
ratio ( y - d//3) approaches infinity on stable days and approaches one
on unstable days (at a distance of one kilometer).
92
-------
An average value for y would appear to be 100 considering
emission height differences only. Because of the significant differences
in actual patterns of spatial distribution, we will inspect y when y £ 0
(i. e. , not directly downwind) and when z 4 0 (i. e. , at points higher than
ground Level).
CONCENTRATIONS ABOVE GROUND LEVEL
When z >0 (i.e., above ground level), the following values of
Y = Yu/Q pertain for emission heights of H = 10 and H = 100 meters:
z = 0 meters
10 meters
50 meters
100 meters
H = 10
-io"4
-lo-4
-.5 x IO"4
-io-6
H = 100
-10'6
-io-6
~.2 x IO"4
-7 x IO"4
These calculations show that the ground level emission
characteristics are reasonable estimates for at least the first thirty
feet of the atmosphere.
GROUND LEVEL CONCENTRATIONS OTHER THAN DOWNWIND
When y/0 (i.e., not directly downwind), Xu/Q must be
multiplied by a factor of K = e"1/2 (y/cry)2 , where ay is the standard
deviation of plume concentration distribution in the horizontal direction.
We find the following values for K for a stability class D day when
y = 1 kilometer;
K
x = 1 kilometer
1.2 x 10
-34
x =10 kilometers
.36
Thus, concentrations other than downwind are extremely
sensitive to the location of the source of emissions.
These observations suggest the importance of the location of a
fixed source with respect to the prevailing wind directions of a given
urban area. Let us represent the urban area as a circle and the
emission sources as in Figure D-3.
93
-------
Figure D-3. Three possible distributions of emission sources.
In (a) the emissions are uniformly distributed on the surface of the area,
as they might be idealized to represent mobile sources. In (b) the emis-
sions are uniformly distributed around the perimeter, and in (c) around
part of the perimeter. Both (b) and (c) are representative of stationary
sources of air pollution. If total emissions are the same in all three
cases, average concentration above the area will, in general, not be the
same in all three cases depending, in particular, upon wind direction.
Furthermore, the average differences in concentrations may span many
orders of magnitude.
94
-------
APPENDIX E
TRANSPORTATION COSTS AND BENEFITS:
COMPUTING CONSUMERS' SURPLUS
The purpose of computing the consumers' surplus (CS) is to measure
one aspect of the cost (or benefits) of a change in transportation policy: the
cost to consumers of a strategy in terms of changed travel costs including
the value of foregone (or induced) trips. We shall consider CS to be the
maximum consumers are willing to pay for a decrease in the price (or the
minimum they would accept as compensation for an increase in the price) of
a given mode of travel.
In aggregating from the individual consumer to all consumers in the
market, we are making the implicit assumption that the marginal social
utility of income is the same for all consumers (i. e. , that an additional
dollar of income contributes an equal amount to social welfare regardless
of who receives it). Furthermore, our model includes auto, bus and rail
trips and no other commodities. This amounts to assuming that price changes
in other sectors of the economy have an insignificant effect on the consumers'
surplus we calculate.
CONCEPT OF CONSUMERS' SURPLUS
Let us illustrate the concept with an example. Suppose the cost of
other travel modes remains unchanged and the cost of a bus trip is reduced
from Pn to P, as illustrated in Figure E-l.
1
.Demand Curve For Trips
Number of
Tripi
Figure E-l. Consumers' surplus.
95
-------
The demand curve in Figure E-l is the Locus of points (here
assumed Linear) relating the quantity of trips taken (horizontal axis) to
given trip prices (vertical axis). Initially QQ trips are taken at unit
price PQ . When price is decreased to PI , bus trip consumption is
increased to Q-, .
The downward slope of the demand curve has two interpretations:
(a) it reflects the fact that fewer people travel as trips become more
expensive; and (b) it indicates that the value of a trip varies over
members of the population; some are -willing to pay more than the trip
price (up to P*), while others are willing to pay only the trip price and
no more. Moreover, as the difference between value (points on the
demand curve) and prices (price lines PQ and Pj , respectively)
diminishes, more trips are demanded, up to a final demand level
and Qj, respectively).
Corresponding to interpretation (b), the triangular area between
the demand curve and each price line measures the aggregate value of
trips, above and beyond the money outlay required to purchase the number
of trips demanded. The difference between these two areas is the CS
shaded on Figure E-l.
Paralleling perspective (a) above, as the price of travel changes,
inducing more or fewer trips, the area between demand curve and the
price line will change, producing CS.
The shaded area represents the most consumers will pay to have
the reduction in price. This consists of two parts: a rectangular portion
corresponding to CS on the original QQ trips plus a triangular portion
corresponding to CS on the Qj - QQ additional trips induced by the price
change. (In the case of a price increase, the latter corresponds to a loss in
CS due to foregone trips. )
THE CONSUMERS' SURPLUS MODEL
The CS model provides a methodology for calculating the value of
CS for specified strategy changes as described in Figure E-i: Within the
Transportation Module, demand curve parameters for all trip modes,
patterns and purposes are accessible. These curves are constructed and
the area bounded by each curve and the relevant price line is computed.
Trip Demand
Curve
-3
Integration
Under the
Demand Curve
-?
Summation
Over All Zones
And Purpose*
_}
Din count
Rate
Sum ma b on
Over All
M-..
Present
Discounted Value
"? of Consumer
Surplus
Figure E-2. Flow chart of consumers' surplus model.
96
-------
For each mode, the measured CS is summed over trip zonal
patterns and over trip purposes. The summed modal CS values are
further combined to yield total CS which is discounted to yield the present
value of all benefits accruing in the future.
FORM OF THE DEMAND CURVE
The trip demand curves in the Transportation Module have the
following form:
Q.., = UT.. x W.., x HF..,
ijk ij ijk ijk
where i = zone.
j = purpose.
k = mode.
Q = quantity of trips demanded.
UT = number of potential travelers.
W = mode choice probability.
HF = trip frequency probability.
The term "price, " as used in the transportation demand functions,
includes both time and out-of-pocket costs incurred in securing trans-
portation services. The time and money cost variables are multiplied
by coefficients which estimate their relative impact on the total composition
of trip prices perceived by travelers.
Cross-price terms are expressions containing prices for two or
more transportation modes. These are essential to the initial determination
of mode split. The mode-split probabilities (Wj^) represent the
substitutability between competing modes. Thus, cross-price terms enter
into the final demand curve for each mode indirectly in the form of the
mode-split probabilities.
Own-price expressions are equations containing only the time
and money costs specific to the particular mode under study. Cross-
price is affected by change in own-price.
The relationship between quantity and own-price is expressed
by decomposing the members of the right-hand side of the equation which
depend on price:
97
-------
HF..
1) log
- HF...
ijk
/ • \
(S{ x P )
V X 1J*/
• •
= SJ + S x P
HF.., = e
Pijk
1 + e
x P..
2) W.., is also affected by price by a relationship we will
denote W.., = W.., (P.., ) .
ijk ijkv ijk'
The final demand function can then be written
Q.-T = UT.. x W.-T (P.., ) x e
ijk ij ijk v ijk'
x P..
CS in the case of price decrease is evaluated by the following
integral:
CS. .,
UT.. x W.., (P) x HF.... (P) dP
ij ijk ' ijk x
P..
This function is not readily integrable by analytic means, but we can
approximate the change in CS by assuming that the demand function
is a linear function of price over the relevant range. Graphically the
change in CS for an increase in price of automobile trips and for a decrease
in price of bus trips is approximated as shown in Figure E-3.
Price /\
Price
Loss In
Consumers' Surplus
0 Q. Q0 Number of
1 Trips
a. Increase in Price of
Auto Trips
Gain in
Consumers' Surplus
Q, Number of
Trips
b. Decrease in Price of
Bus Trips
Figure E-3. Approximation of consumers' surplus
98
-------
In both cases the broken line is the Linear approximation to the demand
curve for prices between PQ and P^ .
TWO OR MORE SIMULTANEOUS PRICE CHANGES
When the prices of two (or more) substitutes change simulteneously,
the accepted methodology is to consider the price changes one at a time
sequentially (in any order) and to sum the corresponding consumers' surpluses.
If we assume a simultaneous increase in the cost of automobile trips and decrease
in the cost of bus trips, we will sum:
(i) The (loss of) CS due to the increase in auto-
mobile trip cost given the original bus trip cost,
and
(ii) the CS due to the decrease in bus trip
cost given the new automobile trip cost.
Furthermore, this will be the same as the sum of:
(iii) The CS due to the decrease in bus trip
cost given the original automobile trip cost, and
(iv) the (loss of) CS due to the increase in automobile
trip cost given the new bus trip cost.
This equality -will hold provided the income effect is small, where income
effect is the product of the following two terms: proportion of income spent
on transportation and income elasticity of transportation demand. The
approximation to CS in the case of two simultaneous price changes
(i. e. , increase in price of automobile trips and simultaneous decrease in
price of bus trips) is calculated as follows.
First we calculate the approximation of the two quantities
indicated as (i) and (ii) above. The first quantity, (i), is calculated
as in the case of a single price change: increase in the price of an automobile
trip. The second quantity, (ii), however, has to be calculated with
respect to the new demand curve for bus trips which shifts outward (from
D()D0 to D}D} in Figure E-4) as a result of the increased price of
automobile trips. The figure indicates the quantity of bus trips originally
demanded (Qn), the quantity demanded at the original price of a bug
trip after the increase in automobile price (Q1), and the quantity
demanded after the decrease in bus price (Q]^). The resulting gain
in consumer surplus to bus trip takers is the triangle indicated on
Figure E-4.
99
-------
D,
PA D,
\
Q
Q Q.
Figure E-4. Gain in consumers' surplus corresponding to (ii).
The analagous calculation is made in the second case indicated as (iii)
and (iv). The loss in automobile consumers' surplus corresponding to (iv)
is the triangle indicated in Figure E-5.
P. - P
Figure E-5. Loss in consumers' surplus corresponding to (iv).
100
-------
The linear approximation to the shifted demand curve is not shown in
Figures E-4 and E-5 to facilitate the exposition.
Under some measures to reduce automobile attractiveness or
increase the attractiveness of mass transportation, the cost of reduced
mobility or benefit of enhanced mobility is less than the full decrement or
increment to CS. For example, if automobile costs are increased via a
tax, the public treasury will collect additional taxes in the amount of the
tax rate times the amount of taxed activity remaining after the imposition
of the tax. These collected taxes represent a transfer, and should be
netted out of calculated CS loss. In like manner, reduction of bus fares
will require an increase in the operating subsidy required for bus service.
This also represents a transfer, and should be netted out of any calculated
increase in CS stemming from a reduction in bus fares. Figure E-6 shows
the same demand curves as Figure E-3 with transfer payments indicated.
Price
Gasoline Tax
Price
Qn Number
U of Trips
Bus
/Subsidy
Q. Number
of Trips
Automobile Trips
Bus Trips
Figure E-6. Transfer payments.
CS used in the calculations presented in this report is the average
of (i) plus (ii) and (iii) plus (iv) net of transfer payments.
101
-------
APPENDIX F
INPUT DATA
This appendix describes the input data required for the MATHAIR
model. Where appropriate, data are presented for each of the four
regions (Los Angeles, New York, Washington and Chicago) used for policy
simulation studies.
AUTOMOBILE STOCK MODULE INPUTS
The Automobile Stock Module takes three inputs: new automobile
registrations for 1955-75, shown in Table F-l; scrappage rates, given
in Table F-Z; and annual growth rate for new automobile registrations,
assumed to be 2 percent.
TRANSPORTATION MODULE INPUTS
A flowchart of the Transportation Module is shown in Figure F-l.
Module inputs, including parameter estimates, represented on the flow-
chart have been numbered on this figure and are presented in this appendix.
Trip Costs, Times, and Distances (Boxes 1-5)
We have assumed that the length of any trip for a given purpose is
independent of the mode choice, and have used estimates based on
automobile distances. Average bus speeds (and rail transit speeds for
New York and Chicago) have been obtained from a variety of sources
including detailed city-specific published reports such as the Metropolitan
Washington Council of Governments report on Existing Transportation
Systems in the Washington Metropolitan Area, less detailed general
reports such as the MVMA Urban Transportation Factbook, and in some
cases from conversations with transit company officials. Speeds, waiting
times and walking times were estimated. Fares •were derived from published
schedules and conversations with transit company officials. Table F-3
shows a sample calculation of automobile parking costs; Table F-4 shows
typical automobile round-trip distances, speeds and in-vehicle times. The
following ten tables (Tables F-5 through F-14) show the Transportation
Module inputs corresponding to boxes 1 through 5 of Figure F-l for the
four study regions. ''"
Value of time, a cost component not shown here, is derived from the
mode-split parameters and is presented in the next section.
102
-------
COMPOSITION OF AUTO STOCK
TABLE F-l. NEW AUTOMOBILE REGISTRATIONS, 1955-1975
(in thousands)
Year
1975
1974
1973
1972
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
I960
1959
1958
1957
1956
1955
Los Angeles
520
510
500
513
451
476
561
562
601
855
1170
601
541
481
391
397
649
643
517
841
589
New York
780
770
770
730
690
720
760
830
820
1160
1470
820
738
656
533
541
886
877
705
1148
803
Washington,
B.C.
260
250
216
165
162
169
164
178
205
186
200
205
185
164
133
135
221
219
176
287
201
Chicago
470
460
457
464
403
382
409
420
445
584
703
445
401
356
289
294
481
476
383
623
436
Source; Estimated using Motor Vehicle Manufacturers Association
of the U. S. , Inc. , Automobile Facts and Figures (1972).
Note: The growth rate in new automobile registrations for 1975 through
1984 was assumed to be 2% for all urban areas.
103
-------
TABLE F-2. AUTOMOBILE SCRAPPAGE RATE
Age of Automobile
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Percent of Automobiles
of Given Age
Still in Stock
100.0
99.0
98.0
96.0
92.0
85. 0
74.0
60.0
44.0
32. 0
23. 0
17.0
12.0
9.0
6.0
4.0
3.0
2.0
1.0
0. 5
Source: Estimated using Motor Vehicle Manufacturers
Association 1972 Automobile Facts and Figures,
p. 30.
104
-------
r4
0
0
U •
•3 «>
" -3
JL
Passenger
Miles
HQ
Number o
Trips
(Traveler
T
1
•
V
>r4
.0
rt
So
££
i
-2
,%H
T)
O
C
o
(0
a
rt
!H
H
a)
1 1
Q
V
w
w
"3
•o .t;-g
^ "ft. **
2W0.
.2-»
I* OJ
HS
i -i
o rs
^•s
.B-S
h h
HP.
o
I
h
0)
3
M
& V
^ w
f>^
>• 2
0 C
a o
.2-li
1* to O*
LH fr* WJ
T
5.5-0
'T
* •
L' - , , . .. ,.._
J
V17
105
-------
By way of example, Tables F-3 and F-4 show how some of the
variables were calculated for New York.
TABLE F-3. SAMPLE CALCULATION OF AUTOMOBILE
PARKING COSTS FOR NEW YORK
WORK TRIPS
SHOPPING TRIPS
OTHER TRIPS
Origin/
Destination
C-C-C
S-S-S
C-S-C
S-C-S
Hr». x C x PPP = Total
8 0.30 50 $1.20
8 0.10 25 0.20
8 0.10 25 0.20
8 0.30 50 1.20
Hr«. xC x PPP = Total
1 0.75 100 $0.75
1.5 0.20 50 0.15
2 0. 20 50 0. 20
2 0.75 65 0.975
Hr». x C x PPP = Total
1 .75 100 $0.75
1 .20 100 0.20
1 .20 100 0.20
1 .75 100 0.75
Source; Estimated
Notes;
Hrs.
C
PPP
C-C-C
S-S-S
C-S-C
S-C-S
average length at time parked
average cost per hour for parking
percentage of total parkers using paid spaces (as opposed to
employer-provided and/or on-street free parking)
trips within center
trips within suburb
round-trips from center to suburb
round-trips from suburb to center
106
-------
Source
V
"u
2 <
>J
S
SN
** C
*«
e
u
|
itination
9 1
Q*°
3*
"C
0
Cl
o
0
1
£u " ^ § >•
i "> S "O 3 M 2
•gjJs S 2 »d .!:
5***1 >s tl O ® "^
£5 g- S ^ uU ?
Isjj £ . If; |
*^£'° en P- *3 «! ** **
S * I g '5 2 =j M £
•0-g.S I* »^.S g
o p T3 »* 7; o. o c ?
^ u c, 5S -**-Qu
Data for typical one-way trips were
The average peak CBD speed of 21.
atudy was felt to be too high to rcflc
was adjusted downward ZSfo to ) 7 m
Data from typical one-way trips tak<
doubled to derive suburb round-trip
Data for typical long-dititance work
trips. Average speed increased \ 0(
on morninycom.-nuteout of CRD and .
Data for typical long-distance (type
to derive round trip data.
tf) »-4 l*> O"*
9* »* eo in
O -O CO O
t» -J r^ •»
*+ M fO fO
(4 (MOO
— -.•»•»
" w 0 co
U Irt W L)
1 i • i
U W 0 to
X J< J< X
U h »« k
O O O O
a
25
1 4-0 **
5 5 ** w*
0 O, 0 • ?" *S
•o -r ** »-i
^ *M •**
•o ** - o u
^ rul"* V
•i C C o C
M '** V V
«] O ^ J* O
•g " i 2 ^
ti Q > <"
O.CQ * t« >.
"(JO >" J
O ** "U *rt
•s « ? M L!
Data from typical shop trip used wil
ward 15% to reflect downtown conge
also felt to be too long and were ass
SDC Survey
Average C-S-C work trip speed inc
conditions.
Average S-C-S work trip speed inc:
peak conditions.
^ • •
*A fO fl
_ o- t- r-
_J V o^ »n
0 0 o 0
N f d d
N co
o- G- £• S-
o o o o
•
a
V 0.
°°'C
i* ^j
V M
> C
* 'S.
' "
3 U
n >
f) O
« A
Composite roundtrip of 20. 6 miles
equal to off-peak speeds assumed a
*& >o »n N
oo' o^ o' r-
« a- r- r-
*~ ^r o^ in
(>a «M en
-------
TABLE F-5. AUTO TRAVEL DATA, LOS ANGELES
(Boxes 1-5 of Figure F-l)
Trip Type
Work
C-C-C
S-S-S
C-S-C
S-C-S
Shop
C-C-C
S-S-S
C-S-C
S-C-S
Other
C-C-C
S-S-S
C-S-C
S-C-S
Round
Trip
Distance
(miles)
11.6
11.6
30.6
30.6
5.2
5.2
15.0
15.0
20. 6 =
20.6
30.0
30. 0
Average
Speed
(mph)
20. 5
22. 7
30.4
27. 0
21. 7
21. 7
30. 7
30.7
22. 3
22. 3
30.7
30. 7
Total
In-Vehicle
Time
(minutes)
33. 9
30. 6
59.4
67. 9
14.4
14.4
29. 3
29. 3
55.4
55.4
58. 6
58.6
Fixed Cost
Per Trip
$ . 05
0
0
. 05
$ 0
0
0
0
$ 0
0
0
0
Variable
Cost
Per Mile
$ 0. 04
0. 04
0. 04
0.04
$ 0. 04
0. 04
0. 04
0. 04
$ 0. 04
0. 04
0. 04
0. 04
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
108
-------
W
O
O
H
<
ffl
a
CQ
„
-s
H "^
B
pq
Its
rt
0)
S
1 1 •!-<
H
> to ~
10 *• is"
p ~
a
a
•H
O O O O
vo *^o in in
• • • •
i— < i— i
•ee-
o o in m
i— i i — i i — t i — i
vo oo [ — t^-
Tp in v£> in
vO CO O vD
in ^o ^0 ^P
•-H •— i ro CM
00 00 O O
CO CO CO CO
•— I i— 1 CO CO
U co U «
0 ^ co 6
g O co O to
^
O O 0 0
vD vD 00 00
• • • •
•W5-
o o in in
r*~ ^0 ^o *^o
i — i i — i i — ii — i
co oo m in
• — i i — i i — i i— i
co -^ o o
• • • •
r — vo co to
i— I I-H CO CO
CO O O O
r— 1 I— 1 CO CO
o o o o
T}< •rp in in
i — i »— i
U rn 0 CO
6 to '
£X /-, ' ,\ '
o U co O to
to
o o o o
oo oo m m
• • • •
r— t i-H
•w-
o o to in
* * » >
i — i i — i i — i i — i
co oo m m
i — i i — i i— i i — i
vO 00 O O
co co o o
v£) vO vO ^O
co o o o
I-H •— I CO CO
**Q vO CO CO
O O O O
CM CNJ CO CO
U „ U to
^ ^ co w O
% O to U to
O
en
O
OJ
o
fn
0)
0)
U
a a
o o
en
•H
f-t
^
^_)
i
O
f-t
U co
to U
U to
a;
o
in
a.
O
•§
en
tn
a.
to
U CO
109
-------
TABLE F-7. AUTO TRAVEL DATA, NEW YORK
(Boxes 1-5 of Figure F-l)
Trip Type
Work
c-c-c
S-S-S
c-s-c
S-C-S
Shop
C-C-C
S-S-S
c-s-c
S-C-S
Other
C-C-C
S-S-S
C-S-C
S-C-S
Round
Trip
Distance
(miles)
11.2
11.2
43.0
43.0
2.0
4.0
20.0
20.0
20.6
20.6
40.0
40.0
Average
Speed
(mph)
17. 0
21. 6
37. 8
34. 0
21. 1
24. 9
39.7
35.7
21. 1
24. 9
39. 7
39. 7
Total
In -Vehicle
Time
(minutes)
39.5
31. 1
68. 3
75. 9
5.7
9.6
30.2
33. 6
58. 6
49. 6
60. 5
67.2
Fixed Cost
Per Trip
$1.20
. 20
.20
1. 20
$ . 75
. 15
.20
. 975
$ . 75
.20
. 20
. 75
Variable
Cost
Per Mile
$ 0. 04
0. 04
0. 04
0.04
$ 0. 04
0. 04
0. 04
0. 04
$ 0. 04
0. 04
0. 04
0. 04
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
110
-------
O
>H
^
H
fcj -
co h
^ ««
< o
PH m
"
D «
I-1 (i)
ffi x
o
. CQ
oo *"
3
CQ
g fn Ot
J
o ^
nt
a
rrt "
0 '^
" co £-
0) ^
T3 O 02
O r^ •+•> "^
r i /-^
^j r~^ [/} c,
OH .^t C
P ~
H
a
•H
H
o o o o
r-- r-- in m
• • • •
i — i i — i
O O 0 0
[^ sD vO t^
i-H I — 1 1 — 1 I — 1
o o m in
in vo in in
r^ oo i-i I-H
• • • •
i— H ^J* OO 00
m *^ ^^ cr^
ro m oo co
>— 1 i— 1 00 IN]
oo oo o o
i-H i— ( CO CO
0 ^ CO (J
g 0 co U co
o o o o
h- r- o o
• * • »
*
0 O 0 0
t^ vo NO r^
i— 1 i— 1 i-H i— 1
0 O 0 0
1 — 1 1 — 1 1 — 1 1 — 1
in o o o
oo m o o
^* sD CD CD
i— l i— I CO CO
O O O O
CNJ -^ o o
oo oo
U ^ co u
o U co U co
o o o o
r- t~- r-- r-
• • • •
i— < i— 1
-W-
o o o o
• • • •
t^- vO vD t*-
o o o o
,_( ,-H rH i-H
CM CO O O
oo r-- o o
oo r-- oo oo
^ \o o o
•-H r-H OO CO
xO vO ^^ ^D
0 CD CD O
, y} U CO
',11
., U Jj co o
S ' ' ' •
^ U co U co
O
rt
a)
u
O CO
a a
o o
VH ««
en en
CX DJ
.H -H
h IH
-4-> •(->
I I
O co
I I
co o
O co
(1) H
-(-> 3
C o
-------
O
w
»
<
S
EH be
r-l .H
H-
•J
to
*
O
W
J
PQ
in
0}
s
ni H
H rt
0>
IH.-S
1) -^
-digs
%'SI
0)
nJ (U r£
rH 0) ^
l#i
o> _
T3 O tn
i •£ rt ^
Q —
^2
o o o o
?»
I
O
rH
I I
co u
U co
rH
0)
(U
U
en
a
U
U
tn
d
•rH
^
0-
• H
rH
co
i
co
U co
112
-------
TABLE F-10. AUTO TRAVEL DATA, WASHINGTON, D. C.
(Boxes 1-5 of Figure F-l)
Trip Type
Work
C-C-C
S-S-S
C-S-C
s-c-s
Shop
C-C-C
S-S-S
C-S-C
S-C-S
Other
C-C-C
S-S-S
C-S-C
S-C-S
Round
Trip
Distance
(miles)
12.2
12. 2
28.4
28.4
5.6
5.6
12. 0
12. 0
20. 6
20. 6
25. 0
25. 0
Average
Speed
(mph)
20. 0
23. 2
33. 8
29. 0
22. 7
22. 7
36. 0
36. 0
24. 0
24. 0
36. 0
36. 0
Total
In-Vehicle
Time
(minutes)
36.6
31. 6
50. 9
58. 7
14. 8
14. 8
20. 0
20. 0
51. 5
50. 5
41. 6
41. 6
Fixed Cost
Per Trip
$ 1. 00
0. 20
0. 20
1. 00
$ 0. 20
0. 10
0. 15
0. 30
$ 0. 10
0. 05
0. 075
0. 15
Variable
Cost
Per Mile
$ 0. 04
0. 04
0. 04
0. 04
$ 0. 04
0. 04
0. 04
0. 04
$ 0. 04
0. 04
0. 04
0. 04
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
113
-------
O
p
o
H
O
J~ J.
EH bO
<} >H
D MH
H O
i—i
CO "?
I -
PQ
W
J
PQ
<
H
T3 a,
° £ h
CD
SH
rP r^
rt
T3 "
[-1 O . 'rH
""* i r i
3 rH
O r .M
<"£
£
O ^
t— i cn
O k^ r ^
r^ p> ||— 1 «H
a ~
CD
rd CD ^
rH CD B/
CD a H
> CO •—
__^
i-g-g «
PH ^ .H 6
b ~
CD
a
H
a
rH
H
O O O O
00 00 00 00
<»
o o in in
r^ vd MD vo
i — i i — t i — i i — i
v^} oo r — r^
o o oo oo
• • • •
CO 00 O 00
xD in v£) xD
vD O O O
r— i "^h oo in
•— i i— i 00 00
oo oo ^f -^
00 OO 00 00
i-H r-H 00 OO
y co U co
V w °? u
^ 6 co O co
o
?5
o o o o
00 00 O O
•ee-
o o m m
l> vO vO vD
l — 1 rH i — 1 i — 1
co oo in in
i — i i — i r— t i — i
o o oo oo
• * • •
r- r-H oo oo
oo oo oo oo
CO sO O O
oo m co co
i— l i— l CO CO
vO SO O O
in m oo oo
i— i i— i
^ CO ^ W
U co w 0
CL 1 '
o U co U co
JH
CO
0 O 0 O
oo oo r^ r^
•W5-
o o m m
r^ vo so ^
i — i i — i i — i i — i
co oo in in
1 — 1 I— 1 l— 1 f-H
in oo in in
• • • •
o CT^ in m
rH
CO s£> O O
oo m co co
^H rH CO CO
vO vD O O
o o in in
00 OO 00 00
'iii
Ui
r A CO t }
H ' i ' i
% U co U co
*TJ
O
•§
rH
CD
4->
a
CD
u
o
rH
rH
H->
I
O
rH
O
w
U
CP
O
rH
V
U
O
O
rH
Vl
en w
O. P-
rH
H->
I
O
rH
CO
a
en en
a a
• H -rH
rH rH
CO
U co
114
-------
TABLE F-12. AUTO TRAVEL DATA, CHICAGO
(Boxes 1-5 of Figure F-l)
Trip Type
Work
c-c-c
s-s-s
c-s-c
s-c-s
Shop
C-C-C
S-S-S
c-s-c
s-c-s
Other
C-C-C
S-S-S
C-S-C
S-C-S
Round
Trip
Distance
(miles)
13. 8
13. 8
33.0
33. 0
4.0
4.0
15.0
15.0
20.6
20. 6
30. 0
30. 0
Average
Speed
(mph)
18. 6
20. 6
29. 3
27. 0
22. 2
22. 2
33. 0
33. 0
22. 2
22. 2
33. 0
33. 0
Total
In-Vehicle
Time
(minutes)
44. 5
40. 2
67.6
73. 3
10. 8
10. 8
27. 3
27. 3
56. 2
56. 2
54. 6
54. 6
Fixed Cost
Per Trip
$ 1. 00
. 20
.20
1. 00
$ . 15
. 10
. 10
.20
$ . 10
. 05
.05
. 10
Variable
Cost
Per Mile
$ 0. 04
0. 04
0. 04
0. 04
$ 0. 04
0. 04
0. 04
0. 04
$ 0. 04
0. 04
0. 04
0. 04
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
115
-------
O
o
o
U
•v
H
Q
H
co
<
PQ
<
H
bO
in
en
4)
X
pq
173 a <"
*Hf"
0)
s
r— 1 .H
O ^
£— I p*M
CU
O r, 4->
Cd
Total
In- Vehicle
Time
(minutes)
sit
§Sl5
^ H m C
CC ..-i C
Q ^
CD
O O C5 O
00 00 CO CO
U CO U W
1 • I I
O w CO (J
£ U 10 U co
6
•§
en
(U
U
^H
-M
I
0)
U
W
a
• i-l
JH
-j_>
I
O CO
CO (J
U co
01
4->
a
D
O
c
H!
C.
tn
0.
co
U CO
116
-------
O
O
<
O
I-H
E
U
<
<
Q
CO
g,
m
o
CQ
w
^
CQ
^ 0
3 'M rt
ti
B
I — 1 *rH
O L/
H <-H
n)
- J
PH" ^ '"^
r-H ^
r_, .;;
-------
Parameter Estimates for Mode Split and Trip Frequency Equations and the
Value of Time (Boxes 6-8)
The mode split equations (Box 6 in Figure F-l) have the following form:
log -^— = a + a x TW + a x (AIV - TSS) + a x (AC - FARE)
J. ™ JTi \J J. LJ J
where H = probability an individual will choose the auto mode
rather than a transit mode
TW = transit walk time
AIV = automobile in-vehicle time
TSS = transit station-to-station time plus wait time
AC = automobile costs for parking, tolls
FARE = transit fare
The values used for the parameters (a~, a , a_, and a ) are shown in
Table F-15. The constant term (an) has been recalibrated for each study
region.
We can now use the mode split equation parameters to calculate the
value of time and the total cost of a transit trip. The total cost of a trip
(Box 8 of Figure F-l) is calculated by adding three cost components; the
cost of waiting time, the cost of in-vehicle time, and the out-of-pocket
costs. Data on out-of-pocket costs were presented in the last section. We
can derive an estimate for the value of waiting and in-vehicle time from the
Domencich and McFadden mode-split coefficients.
The mode split equation for shopping trips has the following form;
log -j~- = -6.77 + . 374TW - . 0654 (AIV-TSS) -4.11 (AC-FARE)
where the variables are as defined above.
Let y = log Y~J3~ • Then
dFARE - dy / o-y _ . 374
o-TW dTW o-FARE ~ -4.11 ~ V '
This means that the transit fare would have to be decreased 9£ to compensate
for a one-minute increase in walk time. Therefore the value of transit walk
time can be estimated at 9£/minute. Analagously, the value of riding time
can be estimated at
dFARE _ .0654 _ n9
dTSS - ^47TT ~ ~'UZ
or 2£/minute.
118
-------
TABLE F-15. MODE SPLIT PARAMETERS
(Box 6 in Figure F-l)
Mode Split Equation Recalibrated Constant Terms (a,,)--
c-c-c
s-s-s
c-s-c
s-c-s
Los Angeles
Work Shopping
Trip Trip
-.05 -3.98
.98 -3.37
-.05 -3.98
.98 -3.37
New York
Work Shopping
Trip Trip
- .36 -5.87
1.38 -4.08
-2.49 -5.77
.73 -4.28
Washington, D. C.
Work Shopping
Trip Trip
-1.45 -7.50
- . 98 -5. 25
- .45 -6. 00
.43 -6.60
Chicago
Work Shopping
Trip Trip
- .28 -5. 18
- .65 -4. 33
. 32 -5. 18
1.95 -4.33
Note: We used the same constant for "other" trips as for shopping trips. In
their study of Pittsburgh, Domencich and McFadden estimated the con-
stant term for Pittsburgh at -4. 77 for work trips and -6. 77 for shopping
trips.
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
Mode Split Equation Parameters--
Work trips
Shopping trips
al
. 114
. 374
a2
-.0411
-.0654
a3
-2.24
-4. 11
Note; Estimated by Domencich and McFadden from a survey of
trip behavior in Pittsburgh by the Southwestern Pennsylvania
Regional Planning Commission. Their equations included an
additional variable. We have used the same coefficients for
"other" trips as for shopping trips.
Source: A Disaggregated Behavioral Model of Urban Travel Demand,
Charles River Associates (1972).
119
-------
The cost of a transit trip has the following three components:
(TW x .09) + (TSS x .02) + FARE
Turning now to the calculation of trip frequencies, the trip-frequency
equation (Box 7 in Figure F-l) has the following form:
HF
where HF = probability an individual will take a trip
P = trip price
It is assumed that HF = 1 for work trips.
The values used for the parameters (b« and b,) are shown in
Table F-16.
Potential Trips (Boxes 9-H)
Three inputs are needed for the calculation of potential trips: city
center and suburb population and growth factors (Boxes 9 and 10 of Figure
F-l) and a matrix of potential trip factors (Box 1 1 of Figure F-l).
We have used the number of potential trip-takers as a proxy for
the number of potential trips (round trips). For work trips -we can safely
assume that the number of trips equals the number of travelers: that is,
that each worker makes an average of one work trip a day. The potential
travelers for a work trip wholly within the central city would be, for
example, that fraction of total central city population who work in the
central city. The matrix below gives the distribution of work trips by
origin and destination for New York in 1970:
120
-------
TABLE F-16. TRIP FREQUENCY PARAMETERS
(Box 7 of Figure F-l)
Trip Frequency Equation Recalibrated Constant Terms (bn)--
Auto
Bus
Rail
Los Angeles
Shopping Other
Trips Trips
2.8 4.1
6.7 9.8
New York
Shopping Other
Trips Trips
3. 7 4.7
0.0 0. 0
3.1 4.6
Washington, D. C.
Shopping Other
Trips Trips
3.8 6.4
5.2 7.4
Chicago
Shopping Other
Trips Trips
2.9 5.1
6.0 7.00
5.6 7.4
Note.- Domencich and McFadden estimated the constant for Pittsburgh
shopping trips at 3. 9.
Trip Frequency Equation Parameters--
bl
Shopping Trips
-1.72
Note: We have used the same price coefficient for "other" trips as for
shopping trips.
Source; A Disaggregated Behavioral Model of Urban Travel Demand.
Charles River Associates (1972).
121
-------
WORK TRIPS - NEW YORK CITY
(1000's of Trips)
Living C
In g
Total
Working In;
C S
2549 99
335 890
2885 989
Total
2648
1225
3874
Population
7346
3341
10778
Source: Urban Transportation Factbook, American Institute of
Planners and Motor Vehicle Manufacturers Association (1974).
Note; C - city center; S - suburb
Using the above data, we calculate the base year (1970) proportion
of the population -who are workers, for each origin-destination pair:
Working in;
C S
Living
In
C
S
. 347
. 100
. 013
.266
The product of these ratios with current population is the number of
total potential work trips (travelers) between a given origin and destination
(O-D). For New York, for example, the work-trip matrix of potential trip
factors and how it is used in the calculations of the transportation module
are shown below:
Total Potential Work
Trips (Travelers) With
O-D Combination:
Work Trip
Matrix Population
"c-c-c"
S-S-S
C-S-C
_S-C-S
_
". 347 0. 0 "
0.0 .266
.013 0.0
0.0 .100
X
rPop
Pop
's
(4x1) (4x2) (2x1)
where Pop and Pop refer to central city and suburb population,,
pectively, and the numbers in parentheses are the matrix dimensions,
res
122
-------
It is more problematic to determine the number of "potential"
trips for shopping and "other" purposes. Whereas a worker takes one
daily •work trip and a non-worker does not, no similar distinction can
be made between potential shoppers and non-shoppers. We have assumed
that the number of trips is proportional to the number of households. We
have also assumed that in any given year the average daily number of
shopping trips per household will not be greater than one, although the
average number of trips taken by those households actually taking trips
on a given day may be greater than one.
As in the case of work trips, potential trip factors are calculated
for shopping and "other" trips. Here a base year ratio (also 1970) of
households to total population in the origin is multiplied by current pop-
ulation. This yields current number of households in the origin, our
proxy for (number of) potential trips.
The inputs used in the numerical calculations reported in this study
are shown in Table F-17.
Vehicle Occupancy Rates (Box 12)
The final input needed by the Transportation Module, vehicle
occupancy rates {Box 12 of Figure F-l), are used to translate passenger
miles into vehicle miles. These data are shown in Table F-l 8.
Calibration Data
To calibrate the Transportation Module, data on actual vehicle
miles traveled for a base year (1975) were obtained. These data are
shown in Table F-l 9.
EMISSIONS MODULE INPUTS
The Emissions Module requires five kinds of input;
a) Automobile mileage
b) Emission factors
c) Deterioration factors
d) Automobile speeds
e) Control device efficiencies
Automobile JVTi.1 e_age
The distributions ot" annual automobile miles traveled by each age
automobile are shown in Table F-20. These distributions and forecasted
vehicle miles of travel (provided by the Transportation Module) are used
to calculate the number of miles driven by each age of automobile.
123
-------
TABLE F-17. POTENTIAL TRIP FACTORS, POPULATION,
AND POPULATION GROWTH
(Boxes 9, 10, 11 of Figure F-l)
Los Angeles
Center: Population 3. 2 million
Growth rate 0%
Suburb: Population 6. 4 million
Growth rate 3%
New York
Center: Population 7. 9 million
Growth rate 0%
Suburb: Population 12. 3 million
Growth rate 1%
Washington, D. C.
Center: Population . 77 million
Growth rate 0%
Suburb: Population . 99 million
Growth rate 3%
Chicago
Center: Population 3. 3 million
Growth rate 0%
Suburb: Population 3. 7 million
Growth rate 3%
Work Trip Matrix
of Potential Trip
Factors*
".267 0
0 .237
. 094 0
0 . 109
. 347 0
0 .266
. 013 0
_ 0 . 100_
".285 0
0 .260
.069 0
0 . 130_
".297 0
0 .274
. 062 0
0 .098 _
Shopping/ Other Trip
Matrix of
Potential Trip Factors*
".375 0
0 . 313
. 375 0
0 . 313
. 3597 0
0 . 301
.3597 0
0 .301_
~. 337 0
0 .257
. 337 0
0 .257_
".300 0
0 . 336
. 300 0
0 . 336 _
The matrices contain the potential trip factors. Matrix rows represent
origin-destination pairs in the following order:
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
Matrix columns represent the center and suburb, respectively. The way
a matrix is used in calculations is illustrated in the text.
124
-------
TABLE F-18. VEHICLE OCCUPANCY INPUT DATA
(Box 12 of Figure F-l)
New York
Chicago
Los Angeles
Washington, D. C.
AUTO
Work
1.20
1.23
1. 1
1.3
Shop/
Other
1.15
1. 18
1.4
1.2
BUS
Work
34
40
30
42
Shop/
Other
17
18
14
21
RAIL
,_ Work
70
44
—
—
Shop/
Other
15
20
—
—
Source; estimated.
125
-------
TABLE F-19.
DAILY VEHICLE MILE TRAVELED (VMT) FOR
BASE YEAR CALIBRATION
Region
New York (Tri-State
Region)
Auto
Bus
Rail
Chicago Region
Auto
Bus
Rail
Washington, D. C.
Region
Auto
Bus
Los Angeles Region
Auto
Bus
1970
VMT /Day*
do6)
84
.269
.14*
33.1
.0897
144.0
.217
Annual
Gr owth
Rate++ (%)
1
1
2
2
2
2
Geographical
Factor+++
1.3
1.3
1975
VMT /Day
do6)
200
.6
1.2
113
.365
.6**
36.5
.099
159.0
.24
* Rapid transit only
** Estimated
Sources; + Urban Transportation Factbook, American Institute of Planners
and Motor Vehicle Manufacturers Association, (1974), Tables
122, 143.
++ Based on 1960-1970 population growth as reported in Factbook
p. 1-9. Assumed growth rates are approximately two-thirds
those for 1960-1970.
+++ Where 1970 figures are for the city alone, the geographical
factor is the ratio of the metropolitan area population to city
population. (Factbook, p. 11-21 for Chicago.) Resulting 1975
figures are estimates for the region.
New York (Tri-State Region) bus and rail figures are projected
to 1975 from Table II and Chart II of the Status of Regional
Public Transportation Service and Use - 1972 (May 1974).
The automobile figure is projected from Vehicle Miles of Travel
on Major Roadways, 1970 (November 1973). Both documents
are published by the Tri-State Regional Planning Commission.
[New York rail VMT assumes ten cars to a train. ]
126
-------
TABLE F-20.
DISTRIBUTION OF TOTAL, ANNUAL MILES
BY AGE OF AUTOMOBILE
Age of
Automobile
(in years)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
>'£
Total
Percent of Total Miles Traveled by Automobile of Given Age
Series A
9.75
8.86
7. 69
6. 80
6. 24
5. 74
5.41
4.91
4.63
4. 35
4. 13
3.92
3.73
3. 54
3.36
3.20
3.04
2.88
2. 74
2.60
2.47
100.00
Series B
8.67
7.88
7.23
6.31
6. 17
5.71
5. 65
5.32
4. 79
4.60
4.37
4. 15
3. 94
3. 74
3. 56
3. 38
3.21
3. 05
2. 90
2. 75
2.61
100. 00
Series C
12.42
10. 88
10. 28
8.23
7. 20
6. 00
4. 54
4. 28
3. 77
3. 60
3.42
3. 25
3. 09
2. 93
2. 78
2. 64
2. 51
2. 39
2. 27
2. 15
2. 05
100. 00
Source: Series A was obtained from a study of air quality programs in the
Chicago metropolitan area conducted by G. S. Tolley of the University
of Chicago.
Series B is the national average reported in T. A. Bostich and H. J.
Greenbalgh, "Relationship of Passenger Car Age and Other Factors
to Mileage Driven, " U. S. Department of Commerce, Bureau of
Public Roads. Washington, D. C. , 1967.
Series C is a Los Angeles basin-specific distribution from "The
Economics of Air Pollution Control for Used Cars," by Lytton
Stoddard and Paul Downing, University of California, Riverside,
Working Paper No. 15.
We used Series A for Chicago and New York, Series B for Washington,
D. C. , and Series C for Los Angeles.
;'<
Totals may not equal exactly 100% due to rounding. It is assumed that
vehicles older than 21 years are driven a negligible amount of miles
annually.
127
-------
Emission Factors
Emissions data for automobiles, for other mobile sources, and for
stationary sources are shown in Tables F-21 through F-29- Table F-21
shows EPA estimates of new automobile emissions of carbon monoxide,
exhaust hydrocarbons, and nitrogen oxides in grams per mile for cars made
before 1975 in compliance with EPA standards.'1* A high emission-rate
scenario was developed for new cars in Los Angeles based on measurements
from a more recent EPA study. These estimates are presented in Table
F-22 and are used to test the sensitivity of MATHAIR to automobile emission
assumptions. Table F-23 shows other automobile emissions (non-exhaust
hydrocarbons and sulfur oxides). Table F-24 describes the scenario for the
lead content of automobile fuel used in our simulations. Bus emission factors
are shown in Table F-25. The other categories of emissions are calculated
on a tons per day basis. Tables F-26 to F-29 show these emissions data.
Deterioration Factors
The emission performance of automobiles deteriorates with age.
Deterioration factors are shown in Tables F-30 and F-31.
Automobile Speeds
The emission characteristics discussed above are for automobiles
driven at 20 miles per hour. Assumed average automobile speeds for the
four study regions are shown in Table F-32 and speed correction curves
are shown in Figure F-2.
Control Devices
The final type of input required is the efficiency of control devices
in reducing emissions. These data are presented along with the device
costs in the final section of this appendix entitled "Cost Module Inputs. "
AIR QUALITY MODULE INPUTS
The Air Quality Module requires user-specified rollback parameters.
These are shown in Table F-33. The data used to calculate these parameters
are shown in Table F-34.
California emissions are shown separately because emission control
standards were implemented there on a different schedule from the
Federal emission standards.
128
-------
to
O
i—i
CO
! ' < d
I C -0-ig
i -t- *
i ,s - IF- S 60
: - w ~
lOthe:
S-Q^
i
cti
• iH
0
o
1—I
rt
u
d) en ,-
60 (1)
a
o
cci
o
O
ffi
vO xO vD CO IT) c-l 00 00 CO CO CM
oooooo
I Jj « O
! «3 o —
ti
<1) en
X
r—
oJ
:U
C
O
O
o
K-
oo o sor
cr*
oo
OJ
in
-------
TABLE F-22. CALIFORNIA NEW CAR EXHAUST EMISSIONS:
HIGH ESTIMATES
(grams /mile)
Year
Pre 1966
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
CO
87
70
65
61
56
52
47
42
37
37
5.4
HC
8.8
7.9
7.1
6.2
5.3
4.5
3.6
3.2
3.2
3.2
.6
NO
X
3.6
3.6
3.6
4. 3
5.5
5. 1
3. 8
3.8
3. 1
3. 1
2.0
Source: These figures are based on those reported in Preliminary Edition
of Supplement 5 to Compilation of Air Pollutant Emission Factors,
AP-42, EPA (April 1975). We have rendered the CO and HC figures
monotonic since the non-monotonicity of the corresponding
measurements in Supplement 5 have yet to be verified.
130
-------
TABLE F-23. OTHER AUTOMOBILE EMISSIONS
a •
"1^2f I1 •* us^ Hydrocarbon Emissions
Pre 1971
Diurna"! loss (grams per vehicle per day)
Running it 1 Hot Soak Loss (grams per mile)
23. 6
2. 1
1972 and later
16, 3
1.45
b. Sul_fT"- ''K'4'1'-; F
- J3 gram.-:. per mile
Source;
a. Calt>p^/i Corpv-i alii-1,, 5 • K'notive Exhaust Emissions Surveillance,
A Summary, " L-P/i Report AP'T'D 1544 (May, 1973). Data is derived
from a test of 31 vehicles in T.. ?, Angeles.
b- c 2 EIHil] ^li* ^ Pi A 'Jl-I' °n'i " '-'•'• ^.9 n Factors, AP-42, EPA, April
1 ''573 and Supple me-jt -, , jepteaibor 1973.
TABLE F-24. SCENARIO FOR LEAD CONTENT OF AUTOMOBILE FUEL
n In-mand for Lead-free Oapoiine by Percent oi Automobiles of each Model Year'
(percent)
1971 T
1972 25
1973 35
1974 60
1975 85
1976 UO
b. Lead Content of (Leaded) Gasoline
(grams per gallon)
1975
1976
1977
1. 70
1. 50
1. 25
Source;
a. John B. Moran, Lead in Gasoline; Impact of Removal on Current and
Future Automotive Emissions, for Air Pollution Control Association,
Denver, "Coles re do, June 9-13, 1974, Table 12, Scenario II or III
(estimates by J.sormt-r and Moore Associates),
b- Fe_deral Register ^ol 7, No, ';',, February 23, ] 972, "Proposed
Rule-Making; Regulation of Fuels and Fuel Additives. "
131
-------
TABLE F-25. BUS EMISSION FACTORS
(grams per mile)
Carbon Monoxide
Hydrocarbons
Nitrogen Oxides
Sulfur Oxides
20.4
3.4
24. 0
2.4
Source: Compilation of Air Pollutant Emission
Factors, AP-47, EPA, April 1973 and
Supplement No. 2, September 1973.
TABLE F-26. HEAVY DUTY VEHICLE EMISSIONS
(tons/day)
Los Angeles
New York
Washington, D. C.
Chicago
HEAVY DUTY GASOLINE
CO
711.2
2467. 0
183.0
946.9
HC
863.0
297. 0
21. 5
115.0
NOX
47. 7
166. 0
12. 0
63.6
HEAVY DUTY DIESEL
CO
33. 1
38. 0
2. 9
15.3
HC
5. 5
5. 0
0. 5
2.6
NOX
55. 0
64. 0
5. 0
25. 5
Sources: Los Angeles; data on total Los Angeles basin gas truck registrations
in Appendix N of the Environmental Protection Agency's report APTD-1372
was multiplied by the average daily mileage (31) reported in MVMA's 1973 Motor
Truck Facts to estimate VMT. VMT estimates were then used with emissions
factors to arrive at total basin emissions. It was assumed (following data
reported by an R. L. Polk Co. survey for New York City) that gas heavy duty
vehicles are 90 percent of total truck population, giving an estimate of regional
heavy duty diesel trucks. This estimate was then applied as above to get heavy
duty diesel emissions. New York: an estimate of 3, 757, 000 daily VMT in the
five counties area reported in Heavy Duty Vehicle Driving Pattern and Use Survey--
New York City, (prepared for the Environmental Protection Agency by Wilbur Smith
and Associates, May, 1973 - APTD-1523) was used (with 90-10 gas/diesel
split) with emissions factors to estimate center city heavy duty emissions.
Suburb emissions were estimated by multiplying center emissions by the ratio
of suburb/center total vehicle VMT reported in the Environmental Protection
Agency report APTD-1 371. Chicago: total truck registrations for Chicago
reported in MVMA's 1972 Auto Facts and Figures were used in a similar manner
as discussed above for Los Angeles. Registrations were multiplied by average
daily miles to get VMT which were in turn multiplied by emissions factors to
get total emissions. Washington, D. C. data was not available for truck
registrations or VMT. Judgmental estimates were made based on Chicago
emissions data.
132
-------
TABLE F-27. MOTORCYCLE EMISSIONS
(tons/day)
Los Angeles
Other Cities
CENTRAL CITY
CO
34.6
HC
9.6
NOT
NO
X
0. 3
ESTI.M
SUBURB
CO
70.4
ATED
HC
19.4
NO
X
0.6
Source: Basin-wide population (107,500 2-stroke, 175,000 4-stroke) was
multiplied by average miles driven per day (11) to get motorcycle
vehicle miles traveled (VMT). Total VMT then was used to estimate
total area emissions using Environmental Protection Agency report
AP-42 emissions factors. A city-suburb split of 33/67 (based on
population) was used to separate total basin emissions regionally.
133
-------
TABLE F-28. STATIONARY SOURCE EMISSIONS
(tons/day)
Los Angeles
New York
Washington, D. C.
Chicago
CO
175
369
33
657
HC
444
1,041
24
837
NO
X
245
1, 557
81
374
SO
X
228
1, 035
142
725
Source: Estimated from 1975 County Emission Summary Reports
provided by the Data Processing Section, NEDS, EPA. Los
Angeles region emissions assumed to be 43 percent (by pop-
ulation) of Los Angeles County emissions. New York data are
estimated from 1974 Emissions Inventory Summary of New
York City Department of Air Resources. New York region
emissions assumed to be 300 percent of New York City emissions.
134
-------
H
W
U
O
CO
O T3
i—( —^
EH co
P-i O
O
U
w
^
w
oo
9
CQ
X
o
o
QD 7 "\
«t u
o E
U
o
u
. o
Q
00
m r- oo o
i — i
r- vo
o
r-
00
^
i — i
oo in
vO
-H vO r-
o vO m
00
r~t CO
1-1 in
^H 00
oo vo i— i
m .— i r~-
i — i
vO
1—1
G
o
_A ^
(U ^H
^H ] *
CU CO
G 3
CU *&
O d
n
£ &
•H S
-S S
i 0
U h
CU +>
— i CU
W OH
o r- o — i
m ^H oo <— i
00 ,-1
m o
Tf CO
i— i
CO
4)
G "
0 d
rt M ^
0 S >s
OH ZJ h
rc) T3 n5
> G G
W at 0
ft) cu +j 45
'h G «« W
Xi -rt fCl
co -^ h h
•o I ^ I
5 0 < 0
-*
t-
CO
r-
CO
00
[^
in
vD
i — i
00
00
CO
CO
r-
tn
in
i — i
_,
o
i — i
0
CO
in
00
^
^
in
r-
3
"o
w ^ z
«
a
.
-P
.
cu
O
!H
d
o
w
co
3
O
0)
G
rM JS
h _<
o cu
>H O
•^ w
!> -i-l
!> ^
CU i<
^;r
. «H
OB O
G -M
O G
•s s
.s I
r1 3)
CU
rt
o
Uco'
co G
cu o
—' -H
cu co
ao co
w
O
T -i-
I— I .r
(U
U GO
X
-H
d
•n
G
o cu
Ct)
Us
0) 01 -
_, CQ "H
§ « §
•H co •£
00
O CU
'5o>
cy o
CO
CU
135
-------
w o
o
CO
PQ
Vehicle Age, years
Al
oo
-
-0
in
*
CO
CM
i — i
O
Pollutant
and Model
Year
0
o
t — 1
o
o
o
0
1— 1
o
o
1 — 1
o
o
1 — 1
o
o
r-H
0
o
1 — 1
o
o
. — 1
o
o
1 — I
o
o
1 — (
Carbon
monoxide
Pre 1968
CM CM NO ON
t*- oo m-sO
i — i i — i i — 1 1 — t
r— 0s t — i co
co m ONO
NO i*- in m
i — t i — i • — 1 1 — i
00 i — i t^- 0s
m Is- "41 ^
i — i i — i i — 1 1 — i
CO OO "^ *\f*
1 — 1 1 — 1 1 — 1 1 — I
t^- CO OCO
I — 1 I — 1 r— 1 i — 1
r-H O^ OONO
-Nt< m coco
i — i i — i i — 1 1 — i
m co CMO
co in co co
r-H I — 1 I — 1 r-H
•Nf< CM 00-Nf
CM ^f I-H O
r-H i — 1 I — ( r-H
O O OO
o o oo
_ ,-H i-lr-H
SO
o
rH
oo 0s o ^ in
NQ vO t*~ ^^ ["""
ON ON QN r-H QN
i — 1 i — 1 i — t i — t
O m r-H
O CO CO
i — 1 r— 1 i — 1
O CM CT
O CO CM
i — i i — t i — i
o o oo
0 CO CM
o oo in
O CM CM
I — 1 I — 1 r-H
O NO CO
O CM CM
O CO r-H
O CM CM
i — i i — i i — i
O r-H 00
0 CM r-H
r-H r-H I-H
O OO NO
O r— 1 r-H
i — ( i — 1 i — 1
O CM O
O r-H i-H
O O O
O O O
1 — 1 , — I 1 — 1
Exhaust
hydrocarbons
Pre 1968
1968
1969
NO CO
CM vQ
r-H r-H
•Np NO
CM in
• — i i — i
CM O
CM m
o co
CM Th
1 — 1 1 — 1
r-H CO
m o
1-1 CM
i — i i — i
CO CM
r-i CM
i — i i — t
O CO
I-H r-H
i — I r-H
m o
O 0
o o
o o
i — 1 i — !
GO
O
rH
o ^ m
f^ O |N_
r-H r-H
O
O
1 1
O
o
1 — 1
o
0
1 — I
o
o
1 — 1
0
o
,-H
o
o
1 — 1
o
o
1 — 1
0
o
1 — t
o
o
o
o
Nitrogen oxides
Pre 1973
NO in
CM •*
i — i i — i
m in
CM •*
i — i i — i
^ in
CM
CO
•a
CO
rH
O
-4->
U
rd
fl
o
•rH
CO
CO
H
rH
•H
CO
G
o
£
o
U
(U
co
o
MH
CM
U
G
(1)
bo
G
o
(U
Pl-K
CO
d,
0)
o
rH
d
o
CO
136
-------
'
H
Qi
II
W0
H w
w • oo
vQ 00
un oo
OiTI.-itn.-i
o co oo co oo
sO
oo
o oo o r- o
o oo ir> NO r-
in
o m o oo 0s
o oo oo oo oo
00
O .— i N£> OO IT)
o oo -^ NO r~-
OO
LO o
O 00 CO O 00 OO
o oo .-H oo oo oo
00
O vO sO
0 00 OJ
o tn -^t1
o oo oo
O ^ — 1
O 00 00
O ON O 00 i—i t*~ 00 O O t^- 00 LO O
o oo m oo oo •NfLO o 0s m NO oo r--
O OO CO in NO ^ Is- O 00 r-H 00 ON! I-H
• •••• •• ••••• •
o m 0s r^ oo ooo or—^00.—< in
o oo oo •"* vo •*}< NO o CM •—i oo CM >-<
i-H i—I i—I .—I i-H i—t i-H i-H r-H i—I i-H i—* r-H
o-N^coi—icr^ ooo oinoor-noo co
OOOOO^m cOin O 00 •—! 00 >-H I-H
• ••«• *t ••••• •
Or-nooinoo oo^f OOOOOONO o
OONl—ICOin COCO OOOr-lr-Hr-H ,-H
.—li—Ir-Hi—li—I i—I.—I i—li—\ t—li—li—I .—I
O.—ii—(00"^ '—i.—i Or—fO.—i'—i O
ooooo oo ooooo o
ooooo oo ooooo o
r-H r—1 i—I i—I r-1 r—Ir—I r-Hr-
n)
co
C
-t->
O
a
co
cc
I
rt
o
OH
CO
C!
O
•H
a
o
U
co
oo
43
o
Pj^
CO
OJ
o
o
CO
137
-------
TABLE F-32. ASSUMED AVERAGE AUTO SPEEDS
Region
Los Angeles
New York
Washington, D. C.
Chicago
Average Auto Speed
(miles per hour)
22
20
22
20
Source: Based on A Survey of Average Driving Patterns in
Six Urban Areas in the United States, Systems
Development Corporation (1973).
TABLE F-33. ROLL-BACK PARAMETERS FOR
AIR QUALITY MODULE
CO
ox
NO
X
SO
X
Pb
Los Angeles
.00139
. 0000603
.0000306
. 00001026
.576
New York
. 00275
. 000095
. 000099
. 0000293
1.49
Washington, D. C.
. 01226
. 000622
. 000345
.000105
10. 60
Chicago
. 00363
.000194
.0000634
.0000627
3.66
Source: Calculated from data in Table F-34.
138
-------
H
W
<
Bi
<
IX
X,
u
<
ffl
o
a!
D
U
u
OH
g
tn
W
«
+*
JS
U
M
flj
(X
(fl
4)
T3
O X
K 0
M
CO
—
•sg
o ~~
i Oxides
Ox>
x
S 2
B0~
O
2
01
"E
'x
o
§6
2 y
c
o
XI
id
U
» 1 "| I
id S fi M
2 o
CO
c —
O ca
>•••- c
.71 ^ O
id F—J
Q~£
«
e ' ~~ •+•
s t, £ (I, •p
£ c jx* g
"y L> rt C
45 o G
^ t6 "£
m
C ^
^ >m o
Q 'S £l
w
1 -— -
Sc -S1 h
4) _P
*C* o c
f* n
^> . . 4-»
en
K-2^
co C
>> CO O
S 'g H
id C —
Q W
Maximum
Concentra-
tion (ppm)
(annual ar-
ith. mean) +
c
•-J m C
«.5^
Q gb.
W
H g? +
!!!•;
'i C § oo
CO
c
^.s"-
^ OT S
rt m °
C * —
^
W
oj
4)
•^ o o .r-
O 00 00 O
sO CM CO OO
O ^.O CJ^ ro
t~- co ro O
^^ "'•I* ^O
o co m ro
o o o o
• • • •
o in oo -^
in m ^ in
CM CO sD CO
o in co in
O CO ** O^
CN] O^ "^f1 CO
m tM -H
C^ CO P— CO
C O 0 O
O "*j* CO O
in vo CM co
^ CM
~< :J"o o^ o
-sf CM CM fM
O O CT* O
O ro ^ ro
co co ro \O
-— • ^ CM C
l—i •—* f-H
0)
T3*
U
V
0) rt c 4)
t!f!
3™ u id 2
^ ^ 0
l
.
S
u
fc ^^
CM ™
(J
•rH C
Tf U w
ro ^
t <° H
_,
cc
c
t,
It
3 *§ cu
i rrt rt ^*
0- H O
<*» )H
CO CO i -5 c
P Q £ 2
H H H «
(X (X -5
< < t^ o
co"
rt
C
id
4^
"o
O
0)
2
>H
o
'id
2
-S
Air Quality
ntrols on
0
O
o
«H
"S
o
a
01
d
id
"3
CO
O
CU
«*H
u
0
a
o
•H
4-*
O
'•3
CU
(X
CO
id
c
CTJ
• ?H
"3
a
o
cu
2
V.
o
re)
2
c
4-t
nj
3
a
ntrols on
o
O
.2
"rt
o
a.
CO
c
^
H
'o
4->
CJ
flj
S3
M
cu
(U
2
u
JS
I
£
o
o
>•
Q
01
13
4-*
C/J
o
CJ
0
O
a
o
.t-4
"rd
4J
t-l
O
ft
CO
C
id
|H
H
•rH
M
CU
OH
T_4
O
'c
o
0
4-)
"3
3
n
rastate Air (
4-»
a
CO
CU
cu
CU
<
CO
o
c
id
•rH
*o
a
o
£
2
1)
*C
"^
J-i
o
?
cu
3
CJ
o
Q
'J
o
a
a
w
CN;
^
ro
•vO
CO
— (
1
D
H
(X
<
CO
rt
^
^<
C
iti
4-i
"o
o.
o
CU
a 2
ro
^
O
ro
^
CU
X!
O
U
O
a
c
o
id
C
^
id
3
a
i.
c
o
CO
Agency Regi
tion Control
rotection
IX
rt
"c
£
/-<
aviroi
W
CO
CU
"a
~*i?
•a
CU
C
c"
FH
4.J
id
tu
3
Prom
id ,i^ o
rtl
4->
h
O
a
CO
c
(d
H
o
CO
o
41
V-4
V4
W
cu
**
o
c
o
4^
u
'•S
CU
(X
from
-a
01
rt
•H
CO
Wi
T3
S
o
00 •<*
» r~
§5
•5 *
£w
'c "id
£ fi
£<
o
cu o^
fl -.
o
cu Q
.a
id
CO £
h O
t~ ™
id
i Id
r- „
B
co u
C CJ
2 c
CO tj
•c-
lo'
id CO
X
O
C 0)
O 0
S (X
< w
id
41
co
id
E
c
o
3
a
a
41
V
CJ
C ^
1)
CO CO
m
t
d
e
41 «
h'^
S ^
S^
£ S
I g
cu id
o v
C *
O *4-t
CJ O
» o
r-H CO
•B T3
£3 cu
S3
(U S
t/1 O
raging
3 4J
bo co
S 2
139
-------
o
n)
fl
O
o
U
TJ
0)
-------
BENEFITS MODULE INPUTS
The Benefits Module requires base year damages and air quality
inputs. We have taken 1%8 as the base year. Damage estimates are
shown in Tables F-35 and F-36. Corresponding base year emissions are
presented in Table F-37. *
COST MODULE INPUTS
In this section we present the control device costs for automobiles
(Tables F-38 through F-41) and stationary sources (Tables F-42 through
F-55) that are used in our simulations. All costs are presented in 1973
dollars.
The tables also indicate the efficiency of the devices. ''"'~ We found
it convenient to report device costs and efficiencies in one table, even
though the costs are input to the Cost Module and the device efficiencies
are input to the Emissions Module.
Air quality is calculated by applying the rollback coefficients to emissions.
We realized after the study was completed that, in the standard case strategy
simulations, 1970 emissions were used with 1968 damage estimates. The 1970
emissions are about 5-10% lower than 1968 emissions (due to intervening con-
trols) and therefore produce an error of a 5-10% overestimate of damages per
unit of pollutant (and of benefits per unit reduced by air pollution controls).
Only costs are reported (in Tables F-38 and F-39) for bringing automobiles
up to 1975 standards. Accompanying pollution reduction efficiency is
reflected in emission estimates.
141
-------
TABLE F-35. BASE YEAR DAMAGE ESTIMATES BY POLLUTANT
AND DAMAGE CATEGORY
(millions of 1973 dollars, inflated from
1968 dollars using Wholesale Price Index)
Los Angeles
Health
Materials
Vegetation
New York
Health
Materials
Vegetation
Washington, D.C.
Health
Materials
Vegetation
Chicago
Health
Materials
Vegetation
Carbon
Monoxide
(CO)
8. 95
23. 11
--
1.38
48.25
--
.30
13.43
--
. 89
17. 58
--
Oxidants
(OX)
85.36
267.41
20.34
14.39
377. 71
. 52
1.08
55. 63
.64
10.22
266. 27
2. 11
Nitrogen
Ox ide s
(NOX)
53.66
73.05
13. 90
17. 03
390.47
1.22
.36
7. 75
. 23
39. 12
439. 59
8. 72
Sulfur
Oxides
(SOX)
--
--
--
278. 21
3189. 60
44. 74
--
--
--
139. 81
1010. 78
83. 88
*
Lead
(PB)
1. 17
2.22
. 31
. 24
5. 68
. 01
. 01
.44
. 01
. 16
2. 27
. 04
Source: Computed from air quality exceedance factors, exposure
factors, and severity factors and total U.S. damages from
Barrett and Waddell (1971) extended by Babcock and Nagda
(1973). all reported in Justus et al., Economic Costs
of Air Pollution Damage, STAR, Inc. , (May 1973).
Pb estimated as 1. 4 percent of particulates (33 percent of trans-
portation sources and 1 percent of industrial process losses),
based on Table 20 of Waddell, Economic Damages of Air Pollution,
EPA (1974)and Ter Haar et al., "Composition, Size and Control
of Automotive Exhaust Particulates, " APCA Journal, Vol. 22, No. 1,
p. 43, (January 1972).
142
-------
TABLE F-36.
BASE YEAR LOW DAMAGE ESTIMATES BY POLLUTANT
AND DAMAGE CATEGORY
(millions of 1973 dollars, inflated from
1970 dollars Using Wholesale Price Index)
Los Angeles
Health
Materials
Vegetation
New York
Health
Materials
Vegetation
Washington, D. C.
Health
Materials
Vegetation
Chicago
Health
Materials
Vegetation
Carbon
Monoxide
(CO)
1.02
2. 63
__
. 16
5. 50
--
.04
1. 53
--
. 10
2. 00
Oxidants
(OX)
12. 02
37. 64
2. 86
2.03
53. 16
. 07
. 15
7. 83
.09
1. 44
37.48
.30
Nitrogen
Oxides
(NOX)
4. 48
6. 10
1. 16
1.42
32. 62
. 10
.03
.65
. 02
3. 27
36. 72
. 73
Sulfur
Oxides
(SO)
--
--
--
10.75
123.20
1. 73
--
--
--
5.40
39.04
3. 24
Lead
(Pb)
, 17
. 33
. 05
. 03
.84
--
--
.07
--
. 02
.34
. 01
Source: Computed from air quality exceedance factors, exposure factors,
severity factors, and lower bound of total U. S. damages reported
in Justus et at, Economic Costs of Air Pollution Damage, STAR,
Inc. (May 1973). " "
>!<
Pb estimated as 1. 4 percent of particulates (33 percent of transportation
sources and 1 percent of industrial process losses), based on Table 20 of
Waddell, Economic Damages of Air Pollution, EPA (1974) and Ter Haar,
et al. , "Composition, Size and Control of Automotive Exhaust Particulates,
APCA Journal, Vol. 22, No. 1, p. 43, (January 1972).
143
-------
TABLE F-37.
BASE YEAR 1970 MOBILE AND STATIONARY
SOURCE EMISSIONS
(tons per
Los Angeles
CO
HC
NOX
sox
Pb
New York
CO
HC
NOx
SOX
Pb
Washington, D. C.
CO
HC
NOX
SOX
Pb
Chicago
CO
HC
NOX
SOX
Pb
Mobile
5709
922
470
28
57
11778
2106
952
37
12
2014
323
146
1.4
2.4
8633
1301
483
27
47
Stationary
148
578
214
300
782
794
1928
1432
19
36
49
142
683
317
311
725
Source: Estimated from pages 3 - 10, 4 - 10, 3 - 28, 3 - 66 of
Prediction of the Effects of Transportation Controls on
Air Quality in Major Metropolitan Areas, APTD-1363
(November 1972).
SOX estimates from 1975 County Emission Summary
Reports Provided by the Data Processing Section,
NADB, EPA.
Pb estimated from 1975 MATHAIR predictions adjusted to
1970 in proportion to average change in other mobile source
pollutants.
144
-------
TABLE F-38. PER VEHICLE ESTIMATES FOR POST-1967 CONTROL
POLICIES EXCLUDING CALIFORNIA (1973 dollars)
Automobile
Vintage
1968
1969
1970
1971
1972
1973
1974
1975b
Equipment Cost
$ 5.40
5. 40
34. 50
34. 50
34. 50
82. 50
82. 50
177. 50
267. 50
Annual Fuel Cost
$15
18
16
26
25
34
31
10
20C
Annual Maintenance
Cost
$16. 00
16. 00
16. 00
16. 00
16. 00
16. 00
16. 00
13. 00
20. 00
a. Percentage loss converted to dollars assuming 10,000 miles/year, 13. 1
miles/gallon, $. 52/gallon.
b. The first figure is without catalytic convertor; the second includes its
cost.
c. Assumed.
Source: The Cost of Clean Air Annual Report of the Administrator Environmental
Protection Agency to the Congress of the United States ( April 1974)
Tables 111-4,111-6, III-7. All costs are EPA estimates either stated
or assumed to be 1973 dollars. For example, dollar figures cited
in the Report from a 1972 EPA publication but not revised by EPA
are assumed to be 1973 estimates.
145
-------
TABLE F-39. PER VEHICLE ESTIMATES FOR POST-1965 CONTROL
IN CALIFORNIA ( 1973 dollars)
Automobile
Vintage
1966*
1967*
1968*
1969
1970
1971
1972
1973
1974
1975b
Equipment Cost
$ 4.00
1. 10
.30
5.40
34.50
34. 50
34. 50
82. 50
82. 50
177. 50
267. 50
Annual Fuel Cost
$11. 25
3. 00
. 75
18.00
16. 00
26.00
25.00
34. 00
31. 00
10. 00
20.00°
Annual Maintenance
Cost
$12. 00
3.20
. 80
16. 00
16.00
16. 00
16. 00
16. 00
16. 00
13.00
20. 00
a. Percentage loss converted to dollars assuming 10,000 miles/year, 13. 1
miles/gallon, $. 52/gallon.
b. The first figure is without catalytic converter; the second includes its cost.
c. Assumed
* Estimated
Source: The Cost of Clean Air Annual Report of the Administrator Environ-
mental Protection Agency to the Congress of the United States (April
1974), Tables III-4, III-6, III-7. All costs are EPA estimates either
stated or assumed to be 1973 dollars. For example, dollar figures
cited in the document from a 1972 EPA publication but not revised
by EPA in 1973 are assumed to be 1973 estimates.
146
-------
•Q «
d w
rd co
•M
*J C
o a
fl
CO
rd
-M
rd
U
CO
co
__i
PQ
• H
^
P
^c
CO
>
CO
ri 0 o oo
U i-H i-H
i__l
CO CO
to to
rd rd
O 0
O -^ ,-,
^ °
1 ' c o CM
OO I-H B1- g^ .o
vQ r^ o o o ,
o o m m ^
^ -H oj
ti-t UH i — t
O O -6O-
o in
CM r-
03 — 'to' —
J-l CQ ^ CO
Cu W cti
O (ti fi fd
U CJ
OT ^ oo ^
2n ^ ^ ^
1 00 I 00
CM g) v£) CO vD
r^ JH o^ LJ cj^ c^
1 1
CO £^~ ^- ^-^- O
to 00 O •— I in (M
fL| m -rt* CM CM o -w-
• H
1 — 1
rd
U
oo _d
,_, ^ o m on o
1 ^? (M CM CM
(D ^5"
M ^
ft CO
JH
ft
. d d o "Jo
^ ° ° £ o
z3 £ £ u u
•rH 0 0 3
rQ 13 d ""O "T?
rd T) T3 co co
_U CO D CE[ Zj
"Q X -2
£ o u o 2
< U ffi S 5
^ CM O~
•60- -1- rH
•te-
i — i
-H
rd
.,_(
§ ^5
.2 ^
•a 'bio
• i-(
CM "W>
CO
•efl- 2;
oo
^^
r- i o
O -6O-
<» +i
in £?• o
OO >-H
-6r> 1 1
•6O-
CO
O
U
co
U to
d o
rd _
C 4^
CO — •
3 ^
(^ ^
0)
13 «
d —i
d co
< fc,
a
o
• r-4
-4->
cd
!H
(U
NI I
Ǥ
ft
•. »
NI CO
s
OS
to Cd
rd -
» ^
(i) O
n H
rd rt
-^ co
d co
O to
T3 ho co
+> u^ M
"rd ^ ^
O to -,
^ U -H
d jj^
'H CO O
TJ CO
co -y <;
a- co
ra rd
to O
rd d •<->
- .2 «
o
T3
U
CO
a
d
CO
CO
rd
to
rd
CO
CO
•4->
rd
T3
d
d
CO
S
d
o
o
Q
n)
H
CO
_H
o
d
o
U
d
o
to
O
P-
CO
d
rd
to
r->
•t
NI
O
E
CO
o
to
g
CO
147
-------
TABLE F-41. DESCRIPTION OF THREE-WAY CATALYTIC CONVERTOR
Proportion
CO
. 73
Emission Reduction
HC
.69
(over 1975)
NO
X
. 82
Device Cost Components:
Installed Cost: $370
Annual Additional Fuel Cost: $40
Annual Additional Maintenance: $23
Source: Estimated from Report by the Committee on Motor Vehicle
Emissions, National Academy of Science (November 1974),
Table V-4, p. 89. Proportion emission reduction is that
between 1975 and 1978 emissions assumed in the document.
148
-------
TABLE F-42. UNIT COST AND EFFECTIVENESS OF CONTROLS APPLIED
TO HYDROCARBON (HC) EMISSIONS FROM PRINTING
(1973 dollars)
Emissions Control
None
Thermal
Incinerator "
Catalytic
Incinerator °
HC
Emissions
Tons/ Year
4. 72
1.77
.59
Percent
Reduction
63
88
Change in
Emissions
Tons/Year
2.95
4. 13
Annual
Cost
$13, 500
12, 900
Capital
Cost
$31, 000
50,000
aAssumes as a "standard" a web offset litho/letter press, 5000 scfm,
emitting 5lbs/hr, 8 hrs/day, 260 days/yr.
b These control alternatives are mutually exclusive.
Source: Computations based on data in Systems and Costs to Control
Hydrocarbon Emissions from Stationary Sources, U. S. EPA
Emission Standards and Engineering Division (August 1973),
p. 16.
149
-------
TABLE F-43.
UNIT COST AND EFFECTIVENESS OF CONTROLS APPLIED
TO HYDROCARBON (HC) EMISSIONS FROM DEGREASINGa
(1973 dollars)
Emissions Control
None
Temperature controls
plus careful operation
Temperature controls
plus refrigerating
cooling coils
Carbon absorption
HC
Emissions
Thousand-
Tons/Year
31. 14
18. 87
15.48
3. 02
Percent
Reduction
39
18
90
Change in
Emission
Thousand-
Tons/Year
12. 27
3. 39
28. 12
Annual
Cost
($1,428)
( 1,070)
( 620)
Capital
Cost
$ 300
3, 000
19, 500
aAssumes as a standard open tank vapor decreases operating 40 hours/week,
29 ft^ tank, uncontrolled emission rate = . 825 Ibs/hr-ft .
b These control alternatives are mutually exclusive.
Note: Numbers in parentheses indicate cost savings due to recovery of
solvents.
Source: Computations based on data in Systems and Costs to Control
Hydrocarbon Emissions from Stationary Sources, U. S.
Environmental Pro tection Agency, Emission Standards and
Engineering Division (August 1973), pp. 5-8.
150
-------
TABLE F-44. UNIT COST AND EFFECTIVENESS OF CONTROLS APPLIED
TO HYDROCARBON (HC) EMISSIONS FROM PETROLEUM
STORAGE3 (1973 dollars)
Emissions Control
None
Pontoon floating
roof
HC
Emissions
Tons/Year
176. 8
12. 7
Pe r c e nt
Reduction
93
Change in
Emissions
Tons/ Year
164. 1
Annual
Cost
$ 0
Capital
Cost
$25, 000
aAssumes a standard storage tank at 90' diameter with a 50, 000 BBL
capacity.
Source: Computations based on data in Systems and Costs to Control
Hydrocarbon Emissions from Stationary Sources, op. cit,
p. 24 and Hydrocarbon Pollutant Systems Study, EPA contract
No. 71-12, MSA Research Corporation, vol. 1 (1972).
151
-------
TABLE F-45.
UNIT COST AND EFFECTIVENESS OF CONTROLS
APPLIED TO HYDROCARBON (HC) EMISSIONS FROM
GASOLINE MARKETINGS (1973 dollars)
Emissions Control
None
Vapor recovery at
terminal
Vapor recovery at
terminal plus recycle
vapor to truck
Plus carbon cannister
Plus refrig,
condensation"
HC
Emissions
Tons/Year
881.3
645. 6
645. 6
382.4
69.4
69.4
Percent
Reduction
27
41
82
82
Change in
Emissions
Tons/ Year
235. 8
263. 1
313. 0
313. 0
Annual
Cost
$21, 000
29, 000
390, 000
808, 000
Capital
Cost
$235,000
322,000
1, 827,000
3, 800,000
b
'Assumes standard hypothetical gasoline distribution system with the
following parameters: Terminal flow = 250 x 10^ gal/day; Number of
stations = 300; Number of tank trucks = 15 Top Loading; Normal
summer operating conditions.
These control alternatives are mutually exclusive.
Source: Computations based on data in Systems and Costs to Control
Hydrocarbon Emissions from Stationary Sources, U. S.
Environmental Protection Agency, Emission Standards and
Engineering Division, August 1973, pp. 27-29.
152
-------
TABLE F-46.
UNIT COST AND EFFECTIVENESS OF CONTROLS
APPLIED TO THE EMISSIONS FROM AIRCRAFT
TURBINE ENGINESa (1973 dollars)
Emissions Control
None
Divided Fuel Supply
Modified compressor
1 I 1 ^
air bleed
THC
Emis sions
Thousand
Tons/ Year
9.31
7. 67
6. 05
Pe r c e nt
Reduction
18
35
Change in
Emission
Thousand
Tons/ Year
1. 64
3. 26
Annualized
Cost
Thousands
of Dollars0
942
547
Assumes a standard commercial airport with 270, 000 LTO's in 1979 (based
on Los Angeles). Three engine fan jet assumed. Time in various modes
used were 70^ Taxi/Idle, 20
-------
TABLE F-47.
UNIT COST AND EFFECTIVENESS OF CONTROLS
APPLIED TO THE EMISSIONS FROM AIRCRAFT
PISTON ENGINES3- (1973 dollars^
Emissions Control
None
Air/Fuel Controlb
Air Injection
Thermal Reactor0
Catalytic Reactor0
Direct Flame
Afterburner °
Water Injection0
THC
Emissions
Thousand
Tons/Year
35. 92
17. 96
17. 96
8. 98
8. 98
3.59
8. 98
Percent
Reduction
50
50
50
50
80
50
Change in
Emissions
Thousand
Tons/ Year
17. 96
17. 96
8. 98
8. 98
14. 37
8. 98
Annualized
Cost
Thousands
of Dollars
0
452
1153
1435
1153
1051
aAssumes standard general aviation airport with 660, 000 LTO's/year
In 1979 (based on figures for Van Nuys Airport). Single engine aircraft
assumed.
These control alternatives are mutually exclusive.
° These control alternatives are mutually exclusive and are designed
to be used with air/fuel control.
Since, in terms of operations, Van Nuys represents 1. 4°/i of general
aviation activity, this figure was used to allocate estimated cost.
e Costs include CO reductions reported in next tables.
Source: Computations based on data in Aircraft Emissions: Impact on
Air Quality and Feasibility of Control. U.S. Environmental
Protection Agency, p. 14, 63, 65, 68, assumed 1972. 1972
dollars have been inflated using wholesale price indices.
154
-------
TABLE F-48. UNIT COST AND EFFECTIVENESS OF CONTROLS
APPLIED TO CARBON MONOXIDE (CO) EMISSIONS
FROM AIRCRAFT TURBINE ENGINES3 (1973 dollars)
Emissions Control
None
Divided fuel supply
Modified compressor
air bleed b
CO
Emissions
Thousand
Tons/Year
6. 87
5. 67
4.46
Percent
Reduction
17
35
Change in
Emissions
Thousand
Tons/Year
1.20
2.41
c
Annualized
Cost
Thousands
of Dollars
942
547
Assumes a standard commercial airport with 270, 000 LT^'s in 1979 (based
on Los Angeles). Three engine fan jet assumed. Time in various modes
used were 70^ Taxi/Idle, 20^ Approach and 10^ Takeoff.
These controls are mutually exclusive.
Q
Los Angeles operations represent approximately 4. If of air carrier
operations. This figure was used to allocate costs. Costs include THC
reductions reported in preceding tables. 1972 dollars have been inflated
using wholesale price indices.
Source: Computations based on data in Aircraft Emissions: Impact on
Air Quality and Feasibility of Control, U.S. Environmental
Protection Agency, p. 14, 60, 69, assumed 1972.
155
-------
TABLE B-49. UNIT COST AND EFFECTIVENESS OF CONTROLS
APPLIED TO CARBON MONOXIDE (CO) EMISSIONS FROM AIRCRAFT
PISTON ENGINESa (1973 dollars)
Emissions Control
None
Air/Fuel Control
A- T • t. b
Air Injection
Thermal Reactor
Catalytic Reactor
Direct Flame
Afterburner
Water Injection
CO
Emissions
Thousand
Tons/Year
40. 12
20. 06
4. 01
4. 01
4. 01
4.01
4. 01
Percent
Reduction
50
75
75
75
75
75
Change in
Emissions
Thousand
Tons/Year
20. 06
16. 05
16. 05
16. 05
16. 05
16. 05
Annualized
Cost
Thousands
of Dollars
0
452
1153
1435
1153
1051
Assumes a standard general aviation airport with 660, 000 LTO's/year
in 1979 (based on figures for Van Nuys Airport). Single engine
aircraft assumed.
These control alternatives are mutually exclusive.
Since, in terms of operations, Van Nuys represents l.4°/t of general
aviation activity, this figure was used to allocate estimated cost. Costs
include THC reductions reported in preceding tables. 1972 dollars have
been inflated using wholesale price indices.
Source: Computations based on data in Aircraft Emissions: Impact on
Air Quality and Feasibility of Control, U. S. Environmental
Protection Agency, p. 14, 63, 65, 68, assumed 1972.
156
-------
TABLE F-50. UNIT COST AND EFFECTIVENESS OF CONTROLS
APPLIED TO NITROGEN OXIDE (NOX) EMISSIONS FROM GAS FUELED
UTILITY BOILERS a (1973 dollars)
Emissions Control
None
Low Excess Air
(LEA)
Two-stage ,
combustion
Fuel gas
recirculation
Water Injection
NOX
Emissions
Thousand
Tons/ Year
48. 08
32. 21
4. 81
9. 62
43. 27
Percent
Reduction
33
85
70
10
Change in
Emissions
Thousand
Tons/ Year
15. 87
27.40
22. 59
4.81
Annualized
Cost
Thousands
of Dollars
(116)
0
246
175
Assumes a standard 1000 megawatt boiler used 6, 120 hrs/yr.
These controls are mutually exclusive and are designed to be used
with LEA.
Note: Numbers in parentheses represent savings.
Source: Computations based on data in Control Techniques for NOy
From Stationary Sources, Public Health Service, p. 5 - 3
(1970). Costs have been inflated to 1973 dollars using
wholesale price indices.
157
-------
TABLE F-51. UNIT COST AND EFFECTIVENESS OF CONTROLS APPLIED
TO NITROGEN OXIDE (NOX) EMISSIONS FROM OIL-FUELED
UTILITY BOILERa (1973 dollars)
Emissions Control
None
Low Excess Air (LEA
Two-stage
combustion
Fuel Gas
recirculation
Water Injection
NOX
Emissions
Thousand
Ions/ Year
27. 22
18. 23
7.35
8. 17
24.49
Percent
Reduction
33
60
55
10
Change in
Emissions
Thousand
Tons/ Year
8. 99
10. 88
10. 06
2. 73
Annualized
Cost
Thousands
of Dollars
(363)
0
246
218
aAssumes a standard 1000 megawatt boiler used 6120 hrs/yr.
These controls are mutually exclusive and are designed to be used with
LEA.
Note: Numbers in parentheses represent savings.
Source: Computations based on data in Control Techniques for NOy
from Stationary Sources, Public Health Service, p. 5 - 3
(1970). Costs have been inflated to 1973 dollars using whole-
sale price indices.
158
-------
TABLE F-52. UNIT COST AND EFFECTIVENESS OF CONTROLS APPLIED
TO NITROGEN OXIDE (NOX) EMISSIONS FROM COAL-FIRED
UTILITY BOILERSa (1973 dollars)
Emissions Control
None
Low Excess Air (LEA)
Two-stage ,
combustion
Fuel gas ,
recirculation
Water Injection
Emissions
Thousand
Tons/ Year
27. 22
20.41
10. 89
12. 25
24.49
Percent
Reduction
25
47
40
10
Change in
Emissions
Thousand
Tons/ Year
6. 81
9.52
8. 16
2. 73
Annualized
Cost
Thousands
of Dollars
(96)
364
246
174
e\
Assumes a standard 1000 megawatt boiler used 6120 hrs/yr.
These controls are mutually exclusive and are designed to be used
with LEA.
Note: Numbers in parentheses represent cost savings.
Source: Computations based on data in Control Techniques for NOy
from Stationary Sources, Public Health Service, p. 5 - 3
(1970). Costs have been inflated to 1973 dollars using
wholesale price indices.
159
-------
TABLE F-53. UNIT COST AND EFFECTIVENESS OF CONTROLS APPLIED
TO NITROGEN OXIDE '(NOX) EMISSIONS FROM COAL-FIRED
INDUSTRIAL BOILERa '1973 dollars)
Emissions Control
None
Low Excess Air
(LEA)
Fuel gas
recirculation °
NOX
Emissions
Tons/ Year
418. 85
314. 14
279. 23
Percent
Reduction
25
11
Change in
Emissions
Tons/Year
104. 71
34. 91
Annual
Cost
$ 130
9, 689
Capital
Cost
$141,000
aAssumes a standard boiler producing 190, 000 Ibs of steam per hour.
Can be used with LEA.
Source: Computations based on data in Analysis of Control Strategies to
Attain National Ambient Air Quality Standard for NO2, J. Crenshaw
and A. Basala (1974).
160
-------
TABLE F-54.
UNIT COST AND EFFECTIVENESS OF CONTROLS APPLIED
TO NITROGEN OXIDE (NOX^ EMISSIONS FROM STATIONARY
INTERNAL COMBUSTION ENGINESa (1973 dollars)
Emissions
Control^
None
Mode
Mode
Mode
Mode
b
Mode
Mode
Mode
Mode
X
Emissions
Tons/ Year
49. 34
26.02
20. 77
28. 29
30. 88
34.44
17.05
41.31
47. 50
Percent
Reduction
47
58
43
37
30
65
16
4
Change in
Emissions
Tons/Year
23.32
28.56
21. 05
18.46
14. 90
32.39
8.03
1. 84
Annualizedc
Cost
$ 466. 50
585. 75
2980. 50
3079. 00
2555. 50
8412. 85
2161. 50
746.40
Assumes a standard spark-fired gas engine, 1080 bhp and 300 rpm
operated 250 days 12 hrs/day.
These are mutually exclusive controls. The various modes involve
timing, manifold temperature, etc. and are described in the source
document.
c
Annualized cost is operating cost.
Source: Computations based on data in Stationary Internal Combustion
Engines in the U.S. , Environmental Protection Agency, pi6~0,
(1973). Costs have been inflated to 1973 dollars using wholesale
price indices.
161
-------
TABLE F-55. UNIT COST AND EFFECTIVENESS OF CONTROLS
APPLIED TO NITROGEN OXIDE (NOX) EMISSIONS
FROM NITRIC ACID MANUFACTURED1 (1973 dollars)
Emissions Control
None
Catalytic Combustion
(Natural Gas)b
Catalytic Combustion
(75% Natural Gas &
25% H2)b
Catalytic Combustion
(Hydrogen) b
NOx
Emissions
Tons/Year
2163
204. 1
40. 82
30. 62
Percent
Reduction
91
98
99
Change
in
Emissions
Tons/ Year
1958. 9
2122. 18
2132. 38
Annual
Cost
Thousands
of Dollars
13. 94
24.40
55. 77
Capital
Cost
$43,931
43,931
43,931
Assumes a standard model plant producing 326. 6 T/Day of Nitric Acid oper
operating 250 Days/Yr.
These controls are mutually exclusive.
Source: Computations based on data in Control Techniques for NO
from Stationary Sources, Public Health Service
"h
es j.ui INV^X
, p. 7-11,
(1970). Costs have been inflated to 1973 dollars using
wholesale price indices.
162
-------
APPENDIX G
VALIDATION OF JOINT PROBABILITIES FOR WORK TRIPS
PREDICTED BY THE TRANSPORTATION MODULE
While total base year vehicle miles traveled (VMT) by mode is
exogenous, the Transportation Module also breaks down the predicted
totals by zone within the urban area and by trip purpose. In this appendix
we validate the work-trip breakdown in Table G-l against the MATHAIR
outputs reported in Tables G-2 through G-5.
TABLE G-l.
CHOICE OF WORK TRIP
TRANSPORTATION MODE
(1970)
New York
Chicago
Washington, D. C.
Los Angeles
% of Employed
Central City
Residents Using:
Auto
30
60
57
90
Transit
70
40
43
10
% of Employed Non-
Central City-
Residents Using;
Auto
82
88
91
97
Transit
18
12
9
3
Source- Calculated from AIP-MVMA, Urban Transportation
Factbook (1974), p. 1-22, assuming that all work trips
use either automobile or mass transit.
For New York, for example, Table G-3 shows that about 30 percent of
the work trips originating in the center are taken by automobile, as are about
163
-------
80 percent of those originating outside the central city. * These numbers
compare favorably with the first line of Table G-l. A similar analysis
may be made for -work trips in the other urban areas.
*For example, we calculate that in New York 80% of the trips originating
outside the center of the city are taken by automobile in the following way:
(. 942 x 3, 270, 000) + (. 430 x 1, 230, 000) /4, 500, 000 = . 802 = 80. 2%
where . 942 = joint probability for using the automobile
for work trips entirely within the suburbs
3, 270, 000 = the number of potential work trips entirely
within the suburbs
. 430 = joint probability for using the automobile
for work trips from the suburbs to the
center of the city
1, 230, 000 = the number of potential work trips from
the suburbs to the center of the city
(and 3, 270, 000 + 1, 230, 000 = 4, 500, 000)
All figures are shown in Table G-3.
164
-------
TABLE G-2. JOINT PROBABILITY PREDICTIONS FOR LOS ANGELES
Joint Probabilities For Automobile
Work
Shop
Other
C-C-C
. 945
.869
. 673
S-S-S
. 981
. 872
.675
C-S-C
. 961
.670
.499
S-C-S
. 982
.675
. 502
Joint Probabilities For Bus
Work
Shop
Other
C-C-C
. 055
. 006
. 017
S-S-S
. 019
. 003
. 014
C-S-C
.039
.013
. 010
S-C-S
. 018
. 007
. 005
Joint Probabilities For Rail
Work
Shop
Other
C-C-C
.0
.0
.0
S-S-S
. 0
. 0
. 0
C-S-C
.0
.0
. 0
S-C-S
. 0
. 0
. 0
Potential Trips in The Areas (millions)
Work
Shop
Other
C-C-C
.85
1. 20
1. 20
S-S-S
1. 61
2. 12
2. 12
C-S-C
. 30
1. 20
1. 20
S-C-S
. 74
2. 12
2. 12
C-C-C trips within center
S-S-S trips within suburb
*
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
Source: MATH AIR output
165
-------
TABLE G-3. JOINT PROBABILITY PREDICTIONS FOR NEW YORK
Joint Probabilities For Automobile
Work
Shop
Other
C-C-C
. 315
.400
. 000
S-S-S
.942
. 945
.774
C-S-C
.273
.442
. 327
S-C-S
.430
. 190
. Ill
Joint Probabilities For Bus
Work
Shop
Ocher
C-C-C
. 269
. 002
. Oil
S-S-S
. 015
. 000
. 000
C-S-C
.328
.000
. 000
S-C-S
. 235
. 000
. 000
Joint Probabilities For Rail
Work
Shop
Other
C-C-C
.416
. 118
. 105
S-S-S
.043
. 000
.000
C-S-C
.399
. 041
. 001
S-C-S
. 335
. 008
. 003
Potential Trips in The Areas (millions)
Work
Shop
Other
C-C-C
2.71
2.84
2.84
S-S-S
3.27
3. 70
3. 70
C-S-C
. 10
2.84
2.84
S-C-S
1.23
3.70
3. 70
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
Source; MATH AIR output
166
-------
TABLE G-4. JOINT PROBABILITY PREDICTION FOR WASHINGTON, D. C.
Joint Probabilities For Automobile
Work
Shop
Other
C-C-C
. 569
. 823
.875
S-S-S
. 906
. 926
. 923
C-S-C
.958
.800
.937
S-C-S
. 903
. 655
. 910
Joint Probabilities For Bus
Work
Shop
Other
C-C-C
.431
. 052
. 031
S-S-S
.094
.007
. 020
C-S-C
. 042
. 044
.011
S-C-S
. 092
. 106
. 021
Joint Probabilities For Rail
Work
Shop
Other
C-C-C
.0
.0
.0
S-S-S
. 0
.0
. 0
C-S-C
.0
.0
.0
S-C-S
.0
.0
.0
Potential Trips in The Areas (millions)
Work
Shop
Other
C-C-C
.219
. 259
. 259
S-S-S
.257
. 334
. 334
C-S-C
.053
.259
.259
S-C-S
. 129
. 254
. 254
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
Source; MATHAIR output
167
-------
TABLE G-5. JOINT PROBABILITY PREDICTIONS FOR CHICAGO
Joint Probabilities For Automobile
Work
Shop
Other
C-C-C
. 564
. 867
. 731
S-S-S
. 929
.885
.826
C-S-C
. 718
.589
. 539
S-C-S
. 665
. 584
.605
Joint Probabilities For Bus
Work
Shop
Other
C-C-C
. 117
. 007
. 002
S-S-S
.071
. 003
.008
C-S-C
. 108
.039
.055
S-C-S
. 114
. 026
. 033
Joint Probabilities For Rail
Work
Shop
Other
C-C-C
.319
. 003
. 077
S-S-S
.0
. 0
.0
C-S-C
. 174
.040
. 105
S-C-S
. 221
. 027
. 065
Potential Trips in The Areas (millions)
Work
Shop
Other
C-C-C
. 980
.990
. 990
S-S-S
1. 04
1.28
1.28
C-S-C
. 205
.990
. 990
S-C-S
. 372
1.28
1.28
C-C-C trips within center
S-S-S trips within suburb
C-S-C round-trips from center to suburb
S-C-S round-trips from suburb to center
Source; MATHAIR output
168
-------
APPENDIX H
MATHAIR USER DOCUMENTATION
All MATITAir, programs and data are stored as files on an
IBM 371-68, accessed through the NCSS time-sharing computer network.
A listing and a card deck of all programs and data are being submitted
nnder separate ^ovct to Hie Project Officer.
Four type,- -if filt", -ir.' needed for r\ simulation run:
ol commands
MATKAIR programs
;• tra t<-~ py do s r, ci pi,i on
i npnt d 'i!;.
Tn addition, MATHAiH ;., ,-ie riles Kvo -.utput data files.
\V h;) f MATHAIK onsists '-f " i v ''ur. ci i on^i I moduJes, it is stored
a ^ f'uif - 'il, i-ou(-i m s in ff,,. fVil i o'.'inp ('<)!> I'l'AN I'iles:
• A IH
• UKADKM
• TJ MILKS
» ix;
Dfi 'R the ['.CHI- fits iV'ofl'iie snb rout i lie. If j.hc usoi' wishes to provide his
own bonef'i*-^ n-"idule, IK' should |x r(-pi;ir..rl l)y -MIother subroutine of the
sn rne nn rn e .
I1 or (•;){ h ~. Li i1 >i i -• t' < >n, fi"e in[> -* i .. l.w scr^Jegy files, and two
output file -; Mpist be -.\>< < \\\"<.\ 1 he |i! s .->-if \v< lia /e used to test MATH-
AIR arf natn>->d .ind iir^-ct -tM-M in • s.1, "'• ! ' !- Kj'.,'!ie f-7 f.'how5 the
control Innguar^e fill "- vh'i h < fmj-in rtief i f l c ^ f i o^: =-. of nil the files need for
Olle S i f MM 1 '; f 1 O'i .
'/a •. ei.-, i (• ("• ,'. f i >•(! -• in ill- Mi;)ur _'ri K" s r'iusr be- set:
-------
Number
1
2a*
b
c
3
4
5
6a=:'
b
c
d
e
7a*
b
c
8
9
Los An
File Name
LA
LEM75
LEM74
LEM65
LA-DEV
LA -IN
DA ML A
A65
A74
L75
L75IM
LFULL
SNULL
SHALF
SFULL
PR
LAOUT
geles
File Type
DATA
DATA
DATA
DATA
DATA
DATA
DATA
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
DATA
DATA
New
File Name
NY
EM75
EM74
EM67
ALL-DEV
NY-IN
DAMNY
A67
A74
A75
A75IM
AFULL
SNULL
SHALF
SFULL
PR
NYOUT
York
File Type
DATA
DATA
DATA
DATA
DATA
DATA
DATA
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
DATA
DATA
Chic
File Name
CH
EM75
EM74
EM67
ALL-DEV
CH-IN
DAMCH
A67
A74
A75
A75IM
AFULL
SNULL
SHALF
SFULL
PR
CHOUT
ago
Filp Type
DATA
DATA
DATA
DATA
DATA
DATA
DATA
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
DATA
DATA
Washingto
File Name
DC
KM75
EM74
EM67
ALL-DEV
DC-IN
DAMDC
A67
A74
A75
A75IM
AFULL
SNULL
SHALF
SFULL
PR
DCOUT
n, D.C.
File Type
DATA
DATA
DATA
DATA
DATA
DATA
DATA
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
STRATEGY
DATA
DATA
In categories 2, 6, and 7, one of the set of files must be selected.
Number
Description of File Contents
a
b
c
d
e
a
b
c
General information about the respective urban areas.
Emissions information (which is the same for all areas except Los Angeles).
It is assumed that post-1975 vehicles have 1975 emission characteristics.
It is assumed that post-1974 vehicles have 1974 emission characteristics.
It is assumed that post-1967 vehicles have 1967 emission characteristics (1965 for Los
Angeles).
Descriptions of new and retrofit devices for automobiles.
Input to transportation submodel.
Input to damage submodel.
Strategy for auto pollution control.
It is assumed that post-1967 vehicles have 196? characteristics (1965 for Los Angeles).
It is assumed that post-1974 vehicles have 1974 characteristics.
It is assumed that post-1975 vehicles have 1975 characteristics.
Same as 6c; also inspection and maintenance required for all vintages.
Same as 6d; also post-1977 vehicles required to meet interim- 1 978 standards
and a set of retrofit devices are required.
Strategy for stationary source pollution control.
No controls required.
Light control.
Strict control.
Output of the transportation model.
Output file.
Figure H-l. MATHAIR input and output files
170
-------
Number
1
2
3
4
5
Los Angeles
File Name
LAI
LA 2
LAS
LA 4
LAB
File Type
EXEC
EXEC
EXEC
EXEC
EXEC
Other Regions
File Name
SI
S2
S3
S4
S5
File Type
EXEC
EXEC
EXEC
EXEC
EXEC
Number
Emission and Strategy Files Used
(See Figure H-l for descriptions)
1
2
3
4
5
uses LEM65 (EM67), A65 (A67) SNULL
uses LEM74 (EM74), A74, SHALF
uses LEM75 (EM75), L75 (A75), SFULL
uses LEM75 (EM75), L75IM (A75IM), SFULL
uses LEM75 (EM75), LFULL(AFULL), SFULL
and assumes transportation times and costs
within LA-IN (CH-IN, etc. ) may have been
altered
For example, LAI EXEC is as follows:
FILEDEF 3 DSK SNULL STRATEGY
FILEDEF 12 DSK A67 STRATEGY
FILEDEF 5 DSK LA DATA
FILEDEF 15 DSK LEM67 DATA
FILEDEF 10 DSK LA-DEV DATA
FILEDEF 7 DSK LAOUT DATA
FILEDEF 8 DSK LA-IN DATA
FILEDEF 16 DSK PR DATA RECFM FA LRECL 133
BLKSIZE 133
FILEDEF 1 DSK DAMLA DATA RECFM FA LRECL 133
BLKSIZE 133
Figure H-2. MATHAIR control language
171
-------
• ISW (in LA DATA) must have one of the following
values:
0 if the transporation model is to predict VMT.
1 if exogenous VMT is specified.
• ALPHA and BETA (in LA DATA):
ALPHA: five stationary source coefficients for
air quality model.
BETA: model source coefficients. Presently
two sets of coefficients have been calibrated
(a = /3 and a' = 100 x /81). The user should
be sure the desired values have been assigned.
• TERM (in LA DATA) - number of years for which
model is to run.
• KFLAG (in LA DATA) must have one of the
folio-wing values:
- 0 if intermediate transportation output is not
desired.
1 if PR DATA is desired.
• MFLAG (in LA-IN DATA) must have one of the
following values:
- 0 if only the transportation model is to be run.
1 the general case where the whole model is
run.
• IFLAG (in DAMIN DATA) must have one of the
following values:
0 if no mobile or stationary source control
devices are to be simulated (i. e. , base case
for damages).
- 1 if some strategy is being implemented.
The FILEDEF commands in the EXEC files must correspond
to the files for a given simulation. In particular:
FILEDEF 1Z for appropriate mobile strategy
file.
FILEDEF 3 for appropriate stationary
strategy file.
FILEDEF 15 for appropriate emissions file.
Also, the proper city name must appear in all places: for
example, CH DATA must be changed to NY DATA if a
simulation for New York is desired.
172
-------
Only two commands are necessary to execute the model:
XX, where XX is one of SI, . . . , S5, LAI, . . . , LAS
and
RUN AIR READEM NMILES DG.
The first command executes the control language file which
identifies the input and output files. The second command loads
and runs compiled versions of the four MATHAIR modules.
Output is stored as disk files. Output files may be either edited
and/or printed at the terminal or else offline printed as in the
following example for Los Angeles:
TITLE LA NULL STRATEGY (optional--generates
a cover page)
OFFLINE PRINT LAOUT DATA
MATHAIR INPUT FILE DESCRIPTIONS
In the following section, we give a detailed description of form of
the five input files for Los Angeles. Each line of each input files begins
with a keyword of ten characters or less, and the keyword is followed by
one or more numbers. MATHAIR does not read the keyword.
In the following pages, each line of each input file is described in
three ways:
• the MATHAIR format that is used to read it
• the keyword
• a. description of the inputs that follow the keyword
NAME OF FILE: LA DATA
Format Keyword Description
20A4 (no keyword) name of urban area (or other title)
10X, 12 ISW 0 if transportation model is to
predict VMT
1 if exogenous VMT is to be used
10X, 315 OYRIYRTER 3 variables: • year of oldest
vintage car in stock
to be described
• present year (base
year for model)
• term of model
173
-------
F or mat
10X, 3E10. 3
10X, 7E10. 3
10X, F10. 3
10X, 7E10.3
10X, 7F10. 3
1CXF5.2
10X, 15
10X, 15
10X, 15
10X,F5. 2
10X, F5. 2
10X, 3F5.2
10X, 3F5.2
10X, F5.2
10X,F5. 2
10X, F5.2
10X, F5. 2
Keyword
DVMT
AUTOREGS
AUTOGROW
AMILE
A JUNK
DR
IPOL
IAGE
NYR
SPEED
HCDIURN
SULF
BUSEM
LEAD1975
LEAD1976
LEAD1977
NONLEAD1975
Description
daily VMT: auto, bus, train
number of new car registrations
from present year back to year
in which oldest vintage car was
new (7 to a line)
rate at which new car registrations
will grow in the future
average number of miles a year
driven by cars of successively
greater age for all ages in stock
1 - scrappage rate of cars of
successively greater age for all
ages in stock
discount rate
number of pollutants besides sulfur
and lead (i. e. , always three in
present model)
number of years provided for aging
of a vehicle in deterioration factors
number of vintage years for
deterioration factors
average auto speed (weighted by
relative frequency of different
purpose trips)
daily diurnal HC loss per auto
SOX emissions per VMT (auto,
bus, train)
bus emissions per VMT of CO,
HC, NOX
gr. lead/gal, fuel (1975)
gr. lead/gal, fuel (1976)
gr. lead/gal, fuel (1977)
% cars requiring non-leaded
gasoline (1975)
174
-------
Format
10X, F5. 2
10X, F5. 2
10X, F5.2
10X, F5. 2
10X, 5F10.0
10X, 5F10.0
10X, 5F10.0
10X, 5F10.0
10X, 5F10.0
10X, 5F10.0
10X, 5E10.2
10X, 5E10.2
10X,2E10.2
10X, 1015
20A4
10X, 5F10.0
10X, 5F10.0
Keyword
NONLEAD1974
NONLEAD1973
NONLEAD1972
NONLEAD1971
MTC
GROWTH
MHDDF
GROWTH
MHDG
GROWTH
ALPHA
BETA
XYZ
NUMBSO
name 1
name 1
GROWTH
Description
% cars requring non-leaded
gasoline (1974)
% cars requring non-leaded
gasoline (1973)
% cars requring non-leaded
gasoline (1972)
% cars requring non-leaded
gasoline (1971)
motorcycle emissions (CO, HC,
NOX, SOX, Pb)
annual growth rate for motorcycle
emissions
heavy-duty diesel emissions
annual growth rate for heavy-duty
diesel emissions
heavy-duty gasoline emissions
annual growth rate for heavy-duty
gasoline emissions
stationary source coefficients for
air quality
mobile source coefficients
coefficients for alternative OX
models
number of installations for n
(n ^ 10) kinds of stationary sources
name of first source (at a typical
installation)
emissions of first source (CO, HC,
NOX, SOX, Pb)
annual growth rate for name 1
20A4
name
n
name of nth source
175
-------
Format
10X, 5F10.0
10X, 5F10.0
10X, 15
20A4
Keyword
name
n
GROWTH
NUMBDE
name.
10X, 5F5.2,2E10.4 SDEV,
Description
emissions of nth source
annual growth rate for name n
number of control devices to be
described
name of first control device
proportion by which it reduces the
five pollutants; installation cost,
annual maintenance
name
NUMBDE
20A4
10X, 5F5. 2.2E10.4
NAME OF FILE: LAEMIS DATA
10X, 10F5.2 DETCO1975
DETHC1975
name of last device
characteristics of last device
deterioration factors for autos of
named vintage (NYR vintage) for
given pollutant when car is new,
1 year old, . . . , up to IAGE years
old
10X, 10F5.2
10X, 10F5.2
10X, 10F5.2
10X, 4F5.2
DETNOX1965
FDETCO76+
FDETHC76+
FDETNOX76+
FDETNOX76
EMIS1975
up to IAGE years old
deterioration factors for the CO
and HC in years 1976 and later
and for NO^ in 1977 and later
X.
deterioration factors for NOX in
1976
emissions of CO, HC, NOX from
auto of given vintage and running
and hot soak (HC) loss (gr. /mi. )
10X, 4F5.2
10X, 4F5.2
EMIS1975
FEMIS1976
same for autos of given (future)
vintages
10X,4F5.2
FEMIS1985
176
-------
NAME OF FILE: LA-DEV DATA
Format
10X,A6, 8X, 14
10X, 10F7.2
10X, 10F7.2
10X, 10F7.2
10X, 3F7.2
Keyword
DEVICE
CO
HC
NOX
COST
Description
name of auto pollution control
device and vintage of auto
% reduction in CO emissions for
new cars, one-year old, up to
IAGE years old
% reduction in HC emissions
% reduction in NO emissions
X.
installed cost; annual costs in
lost fuel and in maintenance
Devices are described by consecutive sets of five cards each. The file
should terminate with a blank card (indicating that there are no more
devices).
NAME OF FILE: DAMLA DATA
10X, 12
2A4
2A4
2A4
10X, 12
10X, 5E10.4
10X, 5E10.4
10X, 5E10.4
10X, 5F10.4
10X, 5F10.4
10X, 5F10.4
10X, 5F10.4
NCAT
CAT2
CAT
neat
IF LAG
DAMHEALTH
DAMPLANT
DAMMATER
number of damage categories to
be used (48). Currently neat = 3
name of first category (health)
name of second category (vegetation)
name of last category (materials)
not used
$ value in base year of damage to
health by CO, OX, NOV, SO , Pb
-X ,2£
same for damage to vegetation
same for damage to materials
CBASEHEALT concentration in ppm (mg/m3 per Pb)
of pollutant causing base year damage
to health
CBASEPLANT
CBASEMATER
DGROWHEALT
same for vegetation
same for materials
annual growth rate of base year
damage to health assumed at
unchanged pollutant concentration
177
-------
10X, 5F10.4
10X, 5F10.4
10X, 5F10.4
10X, 5F10.4
10X, 5F10.4
10X, 5F10.4
DGROWPLANT
DGROWMATER
EXPOHEALTH
EXPOPLANT
EXPO MATER
THRESHHEAL
10X, 5F10. 4 THRESHPLAN
10X, 5F10.4 THRESHMATE
NAME OF FILE: LA-IN DATA
10X, 12 MFLAG
10X, 12
10X,2E12.4
10X, 2F7. 0
10X, 6F7. 0
10X, 9F7. 0
10X, 6F7.0
10X, 6F7.0
10X, 6F7. 0
10X, 12
KFLAG
POP
POPGRO
POPFAC
AVDIST
ABG
ABGX
DELLAM
M
same for vegetation
same for material
exponent in health damage function
(one for each pollutant)
same for vegetation
same for materials
threshold concentration below
which no health damage occurs
(one for each pollutant)
same for vegetation
same for materials
0: transportation model only is run
1: full model is run
0: PR DATA output file is
abbreviated
1: full details in PR DATA
1975 center-city and suburban
population estimates
center-city and suburban popula-
tion growth factors
factors used to generate potential
trips from center-city and suburb,
by purpose
average travel distance by pur-
pose and area for each mode
regression coefficients for
mode-split equations
regression constraints for
mode-split equations
parameters of trip frequency
equation
flag for execution control (default
M=l)
178
-------
Format
10X, 6F7. 0
10X, 6F7. 0
10X, 6F7. 0
10X, 6F7. 0
10X, 6F7. 0
10X, 6F7. 0
10X, 6F7. 0
10X, 6F7.0
10X, 6F7. 0
10X, 3F7. 0
10X, 3F7. 0
10X, 3F7. 0
10X, 2F7. 0
10X, F7. 0
10X, F7. 0
10X, F7. 0
10X, 3F7. 0
10X, 2F7. 0
Keyword
AIV
ACT
ACM
BUSFAR
TWWB
TSSB
RAIFAR
TWWR
TSSR
AOGL
AOB
A OR
AVOT1,AVOT2
NBUS
BCAP
BMAINT
COR
CSI, CSJ
Description
automobile in-vehicle time
fixed cost of auto trip (parking, etc. )
auto costs per mile (fuel and
operating)
bus fare
•waiting and walking time for bus
trip
station-to-station riding time for
bus trip
rail fare
•waiting and walking time for rail
trip
station-to-station riding time for
rail trip
auto occupancy level by trip
purpose
bus occupancy level by trip purpose
rail occupancy level by trip purpose
average value of time spent waiting
and riding
number of buses in initial year
acquisition cost of a bus
maintenance per bus-mile
correction factors for auto, bus,
rail VMT as output by transporta-
tion model
mesh descriptors for approxima-
tion of consumer surplus integral
(suggested values: CSI =5 and
CSJ - . 5)
179
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/5-76-010
3. RECIPIENT'S ACCESSION NO.
4. TITLE ANDSUBTITLE
A COMPUTER SIMULATION MODEL FOR ANALYZING
MOBILE SOURCE AIR POLLUTION CONTROL STRATEGIES
5. REPORT DATE
September 1976
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
MATHTECH, Inc.
P.O. Box 2392
Princeton, New Jersey
10. PROGRAM ELEMENT NO.
1HA094
11. CONTRACT/GRANT NO.
68-01-2952
12. SPONSORING AGENCY NAME AND ADDRESS
Corvallis Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
13. TYPE OF REPOFIT AND PERIOD COVERED
Final Report Nov.1974-Aug.1976
14. SPONSORING AGENCY CODE
EPA/ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report describes MATHAIR, a computer model that simulates the impacts of
strategies for controlling mobile source air pollutants. Vehicle miles traveled
(VMT) for different modes of ground transportation are predicted by a transportation
module. Riven VMT predictions, an inventory of stationary sources of mobile source
pollutants and certain other inputs for the urban area, pollutant emissions are
calculated using EPA-developed relations. A linear rollback air quality module then
predicts ambient pollutant concentrations. Finally strategy costs and benefits due
to damage reduction are calculated.
A set of experiments were performed for four selected urban regions using MATHAIR.
MATHAIR casts all strategy costs (including the imputed social cost of trips foregone
or induced by a strategy) and all benefits in dollar terms so that strategies can
be compared in terms of their total impact. A baseline, zero control strategy and
four strategies of increasing stringency are defined and simulated over a ten-year
horizon. One general conclusion suggested by the results of the test is that it is
economically inefficient to impose the same program of controls in different urban
regions. Another conclusion is that the control strategies are very sensitive to
certain input data.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air Pollution
Benefit/cost analysis
Economic analysis
Economic effects
Econometric models
Automotive
Transportation models
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Environmental Economics
Air Pollution Economics
Automobile Pollution
Envi ronmental Modeli ng
Benefit/cost analysis
05C
Behavioral and
Social Science/
Economics
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
189
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
180
U S GOVERNMENT PRINTING OFFICE 1976—698-166/5 REGION 10
------- |