United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-80-026
October 1980
Air
Methodology To Conduct
Air Quality Assessments
of National Mobile Source
Emission Control
Strategies
Final Report
-------
EPA-450/4-80-026
Methodology To Conduct Air Quality
Assessments of National Mobile
Source Emission Control Strategies
Final Report
Energy and Environmental Analysis, Inc.
1111 North 19th Street
Arlington, Virginia 22209
Contract No. 68-02-3371
EPA Project Officer: Warren P.G. Freas
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
October 1980
-------
This report was furnished to the Envfronmental Protection Agency by
Energy and Envfronmental Analysis, Incorporated, 1111 North 19th Street,
Arlington, Virginia 22209, in fulfillment of Contract No. 68-02-3371.
The contents of this report are reproduced herein as received from
Energy and Environmental Analysis. The opinions, findings and conclusions
expressed are those of the author and not necessarily those of the
Environmental Protection Agency,
Publication No, EPAT<450/4-80^026
ii
-------
TABLE OF CONTENTS
Page
1 INTRODUCTION ... .1-1
1.1 Purpose of this Report 1-1
1 2 Past Applications of Rollback and EKMA 1-2
1.3 Report Organization 1-4
2. APPLICABILITY OF ROLLBACK AND EKMA .2-1
2.1 Assumptions and Limitations 2-1
2.1 1 Rollback Assumptions ... . 2-1
2.1.2 Rollback Limitations 2-4
213 EKMA Assumptions . . . 2-5
2.1.4 EKMA Limitations 2-6
2.1 5 Applications of EKMA to Ozone Strategies .... 2-8
2.2 Use in Nationwide Studies 2-11
2.2.1 Carbon Monoxide .... 2-11
2.2.1.1 Applicability of Rollback 2-12
2 2.1.2 Comparison of Results Using Rollback
and the Indirect Source Guidelines. . . 2-14
2.2.2 Ozone 2-15
2.2.3 Nitrogen Dioxide 2-16
3. SUGGESTED METHODOLOGIES FOR COMPILING MODELING DATA. . . . 3-1
3.1 Emission Inventories 3-1
3.1.1 Ozone. . . 3-1
3.1.2 Carbon Monoxide 3-9
3.1.3 Nitrogen Dioxide 3-10
3.2 Design Values 3-14
3.2.1 Carbon Monoxide. . . 3-17
3.2.2 'Ozone 3-17
3.2.3 Nitrogen Dioxide 3-18
-------
TABLE OF CONTENTS (Continued)
Page
3.3 Growth and Retirement Rates 3-20
3.3.1 Stationary Sources 3-20
3.3.2 Mobile Sources 3-23
3.4 Control Effectiveness 3-30
3.4.1 Mobile Sources 3-30
3.4.2 Stationary Sources 3-32
3.5 Source Contribution Factors 3-40
3.5.1 Ozone 3-41
3.5.2 Nitrogen Dioxide 3-42
3.5.3 Carbon Monoxide 3-42
3.6 Transportation Control Factors 3-51
4. STRATEGY EVALUATION CRITERIA 4-1
4.1 Percentage Change in Pollutant Concentration 4-1
4.2 Number of Exceedances of the NAAQS 4-2
4.3 Number of Regions with Ambient Levels Greater than
the Standard 4-4
4.4 Other Strategy Evaluation Criteria 4-5
4.5 Conclusion 4-5
5. SUMMARY AND CONCLUSIONS 5-1
5.1 Applicability of Rollback and EKMA 5-1
5.2 Suggested Methodologies for Compiling Modeling Data . . 5-3
5.2.1 Emission Inventories 5-3
5.2.2 Design Values 5-4
5.2.3 Growth and Retirement Rates 5-4
5.2.3.1 Stationary Sources 5-4
5.2.3.2 Mobile Sources 5-5
-------
TABLE OF CONTENTS (Continued)
Page
5.2.4 Control Effectiveness 5-5
5.2.5 Source Contribution Factors 5-6
5.2,6 Transportation Control Factors 5-7
5.3 Strategy Evaluation Criteria 5-7
-------
1. INTRODUCTION
1.1 PURPOSE OF THIS REPORT
The purpose of this report is to (1) review and evaluate current proce-
dures , and suggest new procedures when warranted for applying modified
rollback modeling assumptions and the Empirical Kinetic Modeling Approach
(EKMA) to national strategy assessments, and (2) to compile national
data bases for CO, 0~, and NO- for use in future nationwide mobile
source studies. Both modified rollback and EKMA are simple models which
do not require extensive data bases. As such, they are most useful for
analyzing the impact of an emission control strategy on air quality in
nationwide studies where a number of alternative control strategies must
be analyzed in a great many areas. While these models may not estimate
absolute concentrations in each modeling region with great accuracy,
they are useful for predicting the relative impacts of a number of
alternative emission control scenarios on a nationwide basis.
The rollback model is based on the assumption that a change in pollutant
emissions will result in a proportional change in the pollutant's ambient
concentration. In its simplest form, rollback assumes that the concen-
tration of a pollutant at any point is equal to its background concentration
plus some linear function of the total emission rate of that pollutant
in the area which impacts the concentration at that point. However, the
version of the rollback equation discussed in this report provides for
different assumptions to be included for each of a number of source
categories which comprise the total emission rate of a pollutant. For
instance, growth in new sources, control efficiencies, and a typical
source-receptor relationship can be specified for each source category
of interest. The source-receptor relationship is especially important,
because it is what differentiates rollback from other more sophisticated
air quality models.
1-1
-------
EKMA is a procedure for using an isopleth diagram to estimate the effec-
tiveness of volatile organic compounds (VOC) and/or NO controls on
A
ozone concentrations. The isopleths depict the sensitivity of maximum
hourly afternoon ozone to changes in non-methane organic compounds
(NMOC) and oxides of nitrogen (NO ) under meteorological conditions
A
conducive to ozone formation. The EKMA curves have been generated using
a chemical kinetic model validated and calibrated with smog chamber data
(U.S. EPA, 1977).
1.2 PAST APPLICATIONS OF ROLLBACK AND EKMA
One of the first uses of rollback was to set goals for the emission
reductions needed from automobiles in order to meet the National Ambient
Air Quality Standards on a nationwide scale (Earth, 1970). Since that
time, projection analyses using rollback.and EKMA have been performed to
evaluate the relative air quality impacts of revisions to the automotive
emission standards, inspection and maintenance programs for autos,
alternative truck regulations, alternative motorcycle emission control
strategies, and high altitude emission standards. The results of these
analyses have been used in environmental impact statements for EPA
mobile source regulations and to supply Congress and the Administration
with an assessment of the air quality impacts of alternative automotive
emission standards. The most recent important applications of rollback
and EKMA to nationwide studies have been the EPA regulatory analyses
conducted in association with the review of ambient standards for 0- and
CO (U.S. EPA, 1979 and EEA, Inc. 1979).
Past mobile source air quality assessments have been performed using a
sample of Air Quality Control Regions (AQCR's) that have ambient monitoring
data which indicates either present NAAQS violations or the potential
for future NAAQS violations for CO, N02, or 0 . The monitoring data, or
design values, used to choose the AQCR's with the highest pollutant
levels, are used as the base year air quality concentrations in the
1-2
-------
rollback equation. Data on current emission levels are usually taken
from the U.S. EPA National Emissions Data System (NEDS) or from emission
inventories prepared by State or local air pollution control agencies.
NEDS is typically used because of the consistency of data collection
among regions. These data are grouped into source categories which have
similar growth patterns and anticipated control potential. Assumed
growth rates and control efficiencies for each source category are then
applied to estimate future year emissions. The relationships between
emissions and pollutant concentrations typically observed at design
value monitors for each source category are taken into account in the
rollback formulation so that sources that have tall stacks or are usually
located far from urban area monitors are assumed to have less impact on
design value monitors than mobile or other ubiquitous sources. For
instance, in the recent CO regulatory analysis, point source emitters
were found, through modeling, to generally have no impact on monitored
CO levels in urban areas, due to monitor placement (EEA, Inc., 1979).
Assuming that the relationship between air quality and emissions observed
in the base year remains the same in future years, projected emissions
are used to estimate ambient pollutant levels for the year of interest.
Indicators such as average percentage change in concentrations, number
of cities with projected ambient levels above the NAAQS, and the total
number of violations are then used to characterize how air pollutant
levels are expected to change. Typically, alternative emission control
options are compared using these indicators.
The EKMA isopleth curves were incorporated into the nationwide analysis
procedure in 1977. These curves depict the sensitivity of maximum
hourly ozone to changes in non-methane organic compounds (NMOC) and
oxides of nitrogen (NO ) under conditions assumed to be conducive to
X
ozone formation. Past EKMA analyses performed on a nationwide basis
1-3
-------
assumed that the NMOC/NO ratio is 9.5:1 in all areas, and that the
X
effects of natural background ozone concentrations on maximum afternoon
ozone levels for an urban area are offset by diminishing levels of
transported ozone into the areas as controls are implemented upwind. It
has also been assumed in the nationwide study applications of EKMA that
NO levels remain constant in future years.
x
The information presented in this report draws most heavily from the
recent regulatory impact analyses for the National Ambient Air Quality
Standards (NAAQS) for CO, 0_, and NO.. The CO rollback analysis was
performed by SRI International for mobile and stationary area sources
(Siddiqee, Patterson, and Dermant, 1979). The CO point source analysis
using the model PTMAX was performed by EEA, Inc. as well as the cost and
economic analysis for all significant .CO sources (EEA, Inc., 1979). The
0» regulatory analysis used both rollback and EKMA to forecast which
areas would have 0~ levels higher than various alternative ambient
standards (U.S. EPA, 1979). The NO regulatory work is still in progress.
1.3 REPORT ORGANIZATION
This report is organized in four sections in addition to the Introduction.
Section 2 discusses the applicability of using rollback and EKMA in
nationwide studies. The assumptions and limitations of the models are
presented and their use in nationwide studies is discussed separately
for CO, O.j, and NO^. Section 3 presents suggested methodologies for
compiling modeling data for use in rollback and EKMA. Section 4 critiques
the evaluation criteria typically used to compare different model runs.
A summary of the entire report is presented in Section 5.
1-4
-------
REFERENCES FOR SECTION 1
Barth, D. S. 1970. "Federal Motor Vehicle Emission Goals for CO, HC,
and NO Based on Desired Air Quality Levels". J. Air Poll. Control
Assoc. Volume 20.
Energy and Environmental Analysis, Inc. December 28, 1979. Regulatory
Impact Analysis for the National Ambient Air Quality Standards for
Carbon Monoxide. Arlington, VA.
Siddiqee, W., Patterson, R. and Dermant, A. December 1979. Methodologies
to Conduct Regulatory Impact Analysis of Ambient Air Quality Standards for
Carbon Monoxide. (EPA-450/5-80-006). SRI International, Menlo Park, CA.
(Prepared for U.S. EPA, RTP, NC).
U.S. Environmental Protection Agency. February 1979. Cost and Economic
Impact Assessment for Alternative Levels of the National Ambient Air
Quality Standards for Ozone. (EPA-450/5-79-002). Research Triangle
Park, NC.
U.S. EPA. November 1977. Uses, Limitations and Technical Basis of Procedures
for Quantifying Relationships Between Photochemical Oxidants and Precursors.
(EPA-450/2-77-021a) Research Triangle Park, NC.
1-5
-------
2. APPLICABILITY OF ROLLBACK AND EKMA
Section 2.1 presents a discussion of the assumptions and limitations of
rollback and EKMA and how these models compare with alternative air
quality modeling approaches. Section 2.2 discusses the use and applica-
bility of rollback or EKMA in nationwide studies of CO, 0., and NO,,
concentrations. Each pollutant is discussed separately in Section 2.2.
2.1 ASSUMPTIONS AND LIMITATIONS
2.1.1 Rollback Assumptions
Air quality dispersion models are mathematical constructs that attempt
to represent, with varying degrees of sophistication, the relationship
between pollutant emissions and the resulting concentration of that
pollutant in the ambient air. If only nonreactive pollutants are considered,
i.e., those that do not undergo rapid chemical transformation to some
other species, then the relationship between emissions and air quality
concentrations can be represented very simply as:
X = 2 R.Q. + B (2-1)
o iii
where:
X = Base year air quality concentration. It should ideally
be a measured concentration which is representative of
the air quality in the region of interest during the base
year. This "design value" is usually chosen to be con-
sistent with the form of the air quality standard for the
pollutant being modeled.
Q = Base year emission inventory for source category i. The
inventory should be accurate and complete for all signif-
icant sources of the pollutant being modeled. It is
usually expressed in tons/year.
2-1
-------
B = Ambient background concentration in an area (defined as
the sum of the contributions from natural emission sources
within the study area and the anthropogenic and natural
sources outside the study area affecting concentrations
in the study area) given in the same units as X .
R. = Function that relates the emission rate (Q ) to the
pollutant concentration (X ).
For example, if the Gaussian or bivariate normal equation is assumed to
be representative of the behavior of the pollutant after it is emitted,
the R. takes on the familiar form given in the traditional Gaussian
i
models. As such, R. is characterized as a function of source-receptor
distance, atmospheric stability, windspeed and the effective height of
release (stack height plus buoyant rise). An important feature of the
formulation, that will be employed later in this discussion, is the
linear dependence of the pollutant concentration on the pollutant emission
rate, i.e., for any given set of meteorology and source-receptor configurations
R. is a constant which does not depend on changes in emissions.
When the data are not available to evaluate the function R , or are im-
practical to assemble, such as in the case with an analysis of nation-
wide scope, then a simplifying assumption is usually made that reduces
R. to an arbitrary constant (K). If this is done, Equation (2-1) reduces
to the form:
XQ = K 2 Q + B (2-2)
where, as before, X and Q are the base year air quality and emission
rate (for category i) respectively, and B is the background concentra-
tion. An alternate form would be to set R = kS.. Then,
XQ = k SQ^ + B (2-3)
where the factor S is introduced because it is not desirable to totally
discard the effect of height of release and distance on the rate of
dispersion and consequently on concentration estimates.
2-2
-------
Air quality projections for some calendar year j can be made with suit-
able adjustments to the emission rates for each source category ± to
account for growth (G.,), change in emission factors (F..) resulting
from imposition of controls, dispersion characteristics (S.) and the
implementation of transportation control measures (T ) . The future
year projection for year j then becomes:
x - k + B (2-4)
where :
X. = Projected air quality concentration for calendar year j.
G . = Growth factor for source category i in year j . This is a
*" measure of the expected annual growth in the activity
level for each source category.
F . . = Emission factor ratio for source category i in year j .
^ An emission factor ratio is a ratio of the emission
factor of an average source in some future year to the
emission factor of an average source in the base year.
It is indicative of the amount of control on a source
category.
S . = Source contribution factor for stationary source category
i. This factor accounts for the relative effect of
emission height and distance from the source to the
receptor on ground level air quality. An elevated source
would be expected to contribute less to ground level air
quality than a ground level source under most meteorolog-
ical conditions.
T . - Transportation control factor, if applicable, for mobile
source category i in year j . This factor is used only to
account for measures designed to reduce vehicle miles
traveled in an urban area.
It has also been assumed that future background concentrations are the
same as in the base year and the proportionality constant (k) is unchanged,
The assumption on background invariance is not necessary and a different
background estimate could be used.
2-3
-------
The projected air quality can be estimated simply by solving Equation
(2-3) for the unknown constant k and substituting that result into
Equation (2-4). The resulting expression for "modified" rollback then
bee omes :
<2-5)
It is clear from Equations (2-4) and (2-5), that once the functional
form of R is chosen (in this case kS.) and evaluated, all strategy
evaluations for nonreactive pollutants become rollback analyses. Projected
air quality estimates are made by adjusting emission rates appropriately
to reflect the desired emission revisions. Note also, that the rollback
formulation is inherently designed to estimate changes in air quality
concentrations based on current air quality and assumed background
levels resulting from emission changes. As such it can only be used in
the relative rather than the absolute sense.
2.1.2 Rollback Limitations
The limitations of simple rollback have previously been described (deNevers
and Morris, 1975) but will be reiterated here for completeness along
with additional limitations specific to air quality assessments of
national mobile source emissions strategies. Obviously, rollback is an
attractive methodology to employ because of its computational and resource
economy. The most significant limitations are:
1. Lack of model validation.
2. The assumed linear relationship between emissions and air
quality. Deposition, agglomeration and chemical reaction
could invalidate the linear assumption.
2-4
-------
3. Estimates can only be made at existing monitoring sites. If
the existing monitoring data have not detected the "true"
maximum concentrations, then projected air quality estimates
may be understated.
4. The growth factor used assumes all other parameters remain
unchanged, i.e., all sources grow equally and their spatial
distribution remains constant.
5. Simple and "modified" rollback assume that the meteorological
conditions occurring when the design value was measured are
the same in the projection year.
6. Very simplistic treatment of source-receptor relationships.
7. Simple rollback assumes all sources emit at the same height
and "modified" rollback assumes sources within certain cate-
gories emit at the same height.
2.1.3 EKMA Assumptions
While the rollback model has been widely used as a method for estimating
changes in air quality levels due to changes in CO emissions and N0?
area source emissions (see Section 2.2.3), the relationship between
ozone concentrations and hydrocarbon emissions is very complex and
renders the linear relationship invalid. (This is discussed further in
Section 2.1.5). Hence, the EPA is currently recommending a different
modeling methodology called the EKMA. Because EKMA has been described
in detail (EPA, 1977) only a brief discussion follows.
EKMA utilizes a set of ozone isopleths which depict maximum afternoon
concentrations of ozone downwind from a city as a function of initial
(i.e., morning) concentrations of NMOC and NO , NMOC and NO emissions
xx
occurring later in the day, meteorological conditions, reactivity of the
precursor mix, and concentrations of ozone and precursors transported
from upwind areas.
2-5
-------
EKMA is best used to determine the sensitivity of maximum hourly ozone
concentrations observed within or downwind of a city to changes in
ambient levels of NMOC and oxides of nitrogen (NO ) precursors. EKMA is
A
most suitable for addressing questions like, "How much reduction in
local prevailing precursor levels would be needed to attain the 0.12 ppm
standard for ozone" or, "What reduction in ozone levels is likely to
accompany a specified reduction in precursor levels?" The method is not
suitable for estimating the impact of strategies which result in substan-
tial changes in source configurations or for questions about what the
impact of controlling a single or small group of sources would be. As
such, EKMA can be used in nationwide studies for the same scenarios as
rollback.
There are two variations of EKMA. The first involves the use of city-
specific ozone isopleths. The second utilizes a standard set of ozone
isopleths in which fixed assumptions have been made about sunlight
intensity, atmospheric dilution rate, reactivity and diurnal emission
patterns. Because of their ease of use in a nationwide study, the
standard set of ozone isopleths was used in the ozone regulatory analysis
(EPA, 1979). The application of EKMA used in the ozone NAAQS regulatory
analysis assumed the same NMOC/NO ratio for all cities. Based on an
A
examination of data from a number of monitoring sites a median ratio of
9.5:1 was chosen. In addition, it was assumed that NO emissions remain
a
constant between the base year and the projection year. The EKMA standard
isopleth curves are illustrated in Figure 2-1. A line depicting the
9.5:1 NMOC/NO ratio is shown for reference.
A
2.1.4 EKMA Limitations
For several reasons, isopleth diagrams are most properly interpreted
when used in a relative rather than an absolute sense. First, the
absolute position of the isopleths depends upon a number of underlying
assumptions concerning meteorological conditions and emission patterns.
2-6
-------
Unless the meteorological conditions and emission patterns corresponding
to those assumed in deriving the isopleths are similar to those occurring
in the city of interest on days with high ozone concentrations, there is
no reason to expect the absolute position of the isopleths to be correct.
The relative positions of the isopleths, however, should be less sensi-
tive to these differences. Sensitivity tests have indicated that pre-
dicted control requirements for emissions of volatile organic compounds
(VOC) are not very sensitive to changing dilution rates, solar intensity,
diurnal emission patterns and changes in reactivity when the isopleths
are applied in a relative sense (Dodge, 1977). One advantage to beginning
with the observed design value of CL as input to the EKMA is that this
parameter inherently reflects the local prevailing meteorology on that
day. In this respect, the standard isopleth approach is area-specific.
It is crucial, of course, that the 0 monitoring data have been collected
at sites likely to record the maximum 0_ concentration in the region.
A second reason for applying the isopleths in a relative sense is that
they reflect the behavior of one fixed mixture of automotive exhaust.
Relative positions of the isopleths are less sensitive to the mix of
ozone precursors assumed. Third, simulations in which the kinetics
model was used to derive the standard EKMA isopleths did not include
allowance for any injection of precursors after 8 a.m. LDT. A sensitivity
test conducted with the model indicated that the maximum ozone concentra-
tion formed on the first day of the simulation is increased by post-8
a.m. precursor emissions (Dodge, 1977). However, this same sensitivity
study indicated that estimated control requirements (using the isopleths
in a relative manner) were generally insensitive to the post-8 a.m.
emissions. It is important to point out that in the sensitivity studies,
post-8 a.m. emissions were reduced in proportion to reductions in initial
NMOC concentrations. Thus, in using the standard isopleths, proportional
reductions in all emissions (both 6-8, a.m. and post-8 a.m.) must be
assumed in order for the curves to be applied properly. No distinction
2-7
-------
FIGURE 2-1
-^1
ftt
ISOPLETH CURVES FROM SMOG CHAMBER EXPERIMENT
Oa ,
.00 .12 .16 .20 .24 .20
.32 .36
0,0 1.0 1.2
MM! 1C ppm C
1.4
i.a
i.a
20
-------
can be made between 6-9 a.m. emissions and other emissions with the
standard isopleths. Fourth, it should be kept in mind that the isopleths
have been primarily validated against smog chamber data. Smog chambers
represent a simplification of the urban atmosphere in that several
contaminants which may have a potential impact on maximum CL may have
been excluded from either the chamber or the model or both. Finally, a
limitation in this method, as well as in linear rollback or any other
empirical technique, is that there will always be some uncertainty about
whether the maximum ozone concentration is measured. On any given day
it is unlikely that the ozone monitor is at the precise location experiencing
the maximum ozone concentration.
In the Los Angeles basin, standard EKMA tends to underestimate changes
in ozone concentrations observed between the mid-1960's and the mid-1970's
(Trijonis and Hunsaker, 1978). Studies are currently underway to
compare both city-specific and standard versions of EKMA with more
recent trends observed in the Los Angeles basin. Results of the 1978
Trijonis and Hunsaker study, as well as the more recent studies, indicate
that the agreement between EKMA and Los Angeles trends is sensitive to
the NMOC/NO ratio assumed to prevail in the mid-I960's.
A
Applying EKMA to a nationwide study by assuming the same NMOC/NO ratio
A
in all AQCR's will produce some errors. Therefore, area specific ratios
should be used whenever available. However, if the variance about the
default 9.5:1 ratio is not large, then indicators like average percentage
change in 0. concentration and number of areas above the ambient standard
should be meaningful for comparing control strategies.
2.1.5 Applications of EKMA to Ozone Strategies
For the rollback relationship to hold for ozone, there must be a single
proportionality constant between VOC emissions and maximum ozone.
Re-examining Figure 2-1 shows that this is not always the case. Even if
it were assumed that the isopleths in Figure 2-1 are not absolutely
2-8
-------
correct, the proportionality "constant" between NMOC and 0_ depends on
(1) the prevailing level of NO and (2) the changes in NO accompanying
X A.
hydrocarbon control strategies. For example, for 0^ = .28 ppm, ac-
cording to Figure 2-1, the proportionality "constant" can be anything
from 0.36 to 0.14. Proceeding from the 0« = .28 isopleth down to the
0 = .08 ppm isopleth at a constant NMOC/NO ratio of 9.5:1, one finds
proportionality constants between NMOC and 0. varying between .26 and
.50.
As noted previously, control estimates are usually based on the scenario
where transported ozone (from upwind areas) is diminished concurrently
with implementation of local emission control programs. Ideally, this
reduction would result in ozone transported to downwind cities being
reduced to levels approaching natural background. Because transport and
natural background affect maximum ozone via a similar mechanism (i.e.,
from aloft), to be consistent, natural background should be ignored if
transport is ignored. For ozone design values in the range 0.15 to
0.16, the scenario in which transport is actually reduced, ignoring both
transport and natural background results in calculated control requirements
that are within ±5% of those in which reasonable estimates of both
transport and natural background are considered. Hence, this approach
may be considered adequate given the general uncertainty of the levels
of present and future transport in many areas and the long time frame
often considered in regulatory analyses.
An assumption in rollback is that the amount of VOC emission controls
needed to attain the ozone standard is independent of the prevailing
NMOC/NOx ratio. However, smog chamber experiments suggest that the
lower the ratio, the more effective the VOC reduction is in reducing the
2-9
-------
TABLE 2-1
NATIONWIDg EMISSION ESTIMATES, 1977
(10 metric tons/year)
Source Category
NO
VOC
CO
Transportation
Highway Vehicles
Non-highway vehicles
Stationary fuel combustion
Electric utilities
Industrial
Residential, commercial, & institutional
Industrial processes
Chemicals
Petroleum refining
Metals
Mineral products
Oil & gas production & marketing
Industrial organic solvent use
Other processes
Solid waste
Miscellaneous
Forest wildfires & managed burning
Agricultural burning
Coal refuse burning
Structural fires
Miscellaneous organic solvent use
TOTAL
9.2
6.7
2.5
13.0
7.1
5.0
0.9
0.7
0.2
0.4
0
0.1
0
0
0
0.1
0
0.
0
0
0
0
23.1
11.5
9.9
1.6
1.5
0.1
1.3
0.1
10.1
2.7
1.1
0
0
3.1
2.7
0.3
0.7
1
1
4
0
0
0
0
3.7
28.3
85.7
77.2
8.5
1.2
0.3
0.6
0.3
8.3
2.8
2.4
2.0
0
0
0
1.1
2.6
4.9
4.3
0.5
0
0.1
0
102.7
NOTE: A zero indicates emissions of less than 50,000 metric tons/year.
SOURCE: U.S. EPA, National Air Quality, Monitoring, and Emissions Trends
Report, 1977 (EPA-450/2-78-052), Research Triangle Park, N.C.,
December, 1978.
2-10
-------
maximum ozone formed. Thus, at very low NMOC/NOx ratios, rollback may
underestimate the effectiveness of VOC controls. Conversely, at high
ratios, estimates obtained with rollback may be overly optimistic. With
the prevailing ambient conditions which appear to be typical in many
cities, the net effect of rollback assumptions is to estimate that less
control is needed to attain the ozone standard than would be implied by
the EKMA. Under the ambient conditions which apparently prevail in most
cities, rollback is likely to differ most from predictions of kinetics
models during the period in which initial control increments are exercised
(i.e, when the NMOC/NO ratio is still relatively high). Thus, for
X
example, if VOC emission reductions of 30 percent were implemented in a
city experiencing moderate (9.5:1) NMOC/NOx ratios, the corresponding
improvement in maximum ozone concentrations may be considerably less
than the 30 percent predicted with rollback. The risk in relying on
rollback is that when the corresponding 30 percent reduction in ozone
levels failed to materialize, questons may arise about whether con-
trolling volatile organic compounds will ever be effective in reducing
ambient ozone. Many of these questions would arise because a model
which is not based directly on cause-effect relationships between ozone
and its precursors is used. Therefore, use of rollback for estimating
VOC reductions needed to attain the ozone standard is not recommended.
2.2 USE IN NATIONWIDE STUDIES
2.2.1 Carbon Monoxide
Concentrations of CO vary with time, season, and geographic location.
These variations often follow predictable trends. In most cities, CO
levels peak at 7 to 9 a.m., 4 to 7 p.m., and 10 p.m. to midnight. The
first two peaks arise from automobile traffic coupled with meteorological
conditions. The midnight peak can be primarily attributed to calm wind
conditions which result in a reduced dispersion of CO emissions.
2-11
-------
The highest CO levels tend to occur in the fall and winter. In the fall
and winter seasons, the tendency toward colder ambient temperatures
results in increased production of CO emissions from cars, in addition
to CO emitted from other fuel burning sources. Also, the more stable
atmospheric conditions and low wind speeds which occur during winter and
fall result in decreased dispersion of CO emissions and contribute in
substantial part to the occurrence of high ground level CO concentrations
Carbon monoxide is commonly understood to be a mobile source problem
with wide variations in concentrations within an urban area. As can be
seen from Table 2-1, transportation sources emitted more than 80 percent
of the anthropogenic CO emissions in the U.S. in 1977. Some cities show
that mobile sources emit even a higher proportion of CO than 80 percent.
For instance, the Denver SIP estimates that 93 percent of all CO emis-
sions in 1978 were from motor vehicles in that region.
2.2.1.1 Applicability of Rollback
A recent study performed by SRI International (Shelar, Ludwig, and
Shigeishi, 1979) for EPA's Office of Mobile Source Air Pollution Control
studied CO source/receptor relationships in four urban areas. These
areas were San Jose, Seattle, Phoenix, and Chicago. One of the purposes
of this study was to identify the contribution of CO from local versus
regional sources at sites where violations of the 8-hour CO standard are
likely. The site chosen for study in San Jose was near a congested
suburban intersection in a commercial area. The Seattle and Chicago
sites were in heavily congested downtown areas. The site studied in
Phoenix was downtown near government office buildings where parking lots
filled and emptied during rush hours in the morning and evening. All of
the sites were chosen because they were considered likely to have viola-
tions of the CO standard during the one week monitoring period. Both
"local" and "background or regional" monitors were used at each site.
2-12
-------
The results of the monitoring in these four cities illustrate two situa-
tions that can produce CO 8-hour standard violations. The first situa-
tion is where high traffic density and congested conditions produce
excessively high CO emissions which cause high ambient levels. This
situation is characterized by diurnal CO patterns which closely follow
traffic level variations. The second situation is where an inversion
forms to limit vertical mixing, and emissions during the time the inver-
sion is in place build up to produce peak CO levels. CO violations
during this time can be widespread in an urban area. However, there is
no evidence to 'suggest that CO is transported from any distance during
the inversion. If this is true, "local" emissions are responsible for
high observed CO levels.
Because CO has been labeled as a "local" problem with variations in con-
centrations thought to be due to nearby traffic densities, little work
has been done to characterize how CO concentrations can reach peak
levels during periods when emissions are low. Therefore, there is
considerable uncertainty in identifying which sources or types of sources
are contributing to high nighttime levels of CO. For an application of
rollback to be valid, it is important that the mix of sources contributing
to the problem be accurately known. This issue is discussed further in
Section 3.
In the CO Regulatory Analysis, the rollback equation was applied to
estimate the number of counties that would have CO levels exceeding
alternative ambient standards for CO in future years. Because of the
different nature of mobile and stationary source CO emission problems,
and the location of the existing monitoring network, it was concluded
that recorded violations in CO nonattainment areas are due primarily to
mobile sources and localized area sources. Therefore, only mobile
source and other area source emissions were included in the emission
inventories used in the rollback analysis. An analysis of large stationary
2-13
-------
(point) sources was conducted and it indicated that stationary source
emissions had negligible effects on existing urban monitor sites. Thus,
the analysis assumed for existing problem areas that all reductions
would come from controlling mobile and other area source emissions. The
regulatory analysis assumes that 100 percent of all mobile source emissions
in a county and 20 percent of all area source emissions affect concentrations
at design value monitors. However, the mix of sources used in the CO
regulatory analysis is not applicable for cases where high CO levels are
measured during nighttime or early morning hours.
2.2.1.2 Comparison of Results Using Rollback and the Indirect Source
Guidelines
In past uses of rollback for analyzing the impact of nationwide control
programs on CO concentrations, it has been assumed that although CO is
largely thought to be a "hot-spot" problem, that if the county or AQCR
inventory used in the modeling accurately reflects the source contri-
butions at the design value monitor, that rollback was appropriate for
estimating future year CO concentrations. By comparing the technique
for estimating the impact of changing localized emissions on nearby CO
concentrations described in the Indirect Source Guidelines (U.S. EPA,
1978) with the rollback equation, the validity of this assumption can be
tested. The difference in the procedure used to estimate future year CO
concentrations using the Indirect Source Guidelines versus using rollback
is only in the way the design value is chosen. The Indirect Source
Guidelines use the HIWAY model along with traffic characteristics to
estimate local CO contributions. (Background CO should be added to
obtain a total concentration.) Rollback uses measured CO values in the
urban area. Both methods use variations in emission rates due to changes
in emission factors, vehicle miles traveled, and vehicle speed to estimate
future year concentrations. Therefore, both methods of estimating
future concentrations assume the basic relationship that emission changes
result in proportional changes in air quality.
2-14
-------
In any city where high CO concentrations are a localized phenomenon,
rollback can be used instead of the Indirect Source Guidelines to esti-
mate future year CO concentrations if (1) the monitor measuring the
design value is in a location representative of the highest CO in the
city, if (2) the emission sources affecting the monitor are well charac-
terized, and if (3) the same meteorological conditions are assumed to
prevail in the future. Then, the only difference between rollback and
ISG predicted future CO concentrations is that monitored versus estimated
concentrations, respectively, are used as design values.
2.2.2 Ozone
Ozone, a secondary pollutant, results from the reaction of non-methane
organic compounds (NMOC) and oxides of nitrogen (NO ). Ambient ozone
* a
concentrations occur as the result of two processes. One is a physical
process involving dispersion and transport of the precursor emissions.
The second is a chemical process involving reaction of the precursors
under the stimulus of sunlight. Obviously, then, meteorological conditions
are an important factor in determining ambient ozone levels. Atmospheric
dilution and transport affect ambient pollutant levels and the geographic
relationship between source areas and corresponding ozone problem areas.
Solar radiation and ambient temperature affect the chemical processes.
Because meteorological conditions influence ozone formation to such a
great extent, they tend to obscure both the absolute and relative effects
of the emission-related factors to degrees that vary by geographic
location and season. However, because nationwide analyses are the
principal application for the modeling procedures described in this
document, the average predicted changes in concentrations should not be
unduly biased by yearly meteorological variations in a few cities.
In previous nationwide studies it has been informative to use both EKMA
and rollback to produce separate estimates of expected ozone concentration
2-15
-------
re'ductions in future years. However, EKMA is regarded as a more theoretically
sound approach for estimating the effects of VOC reduction on ozone
concentrations, and the use of rollback is no longer recommended.
2.2.3 Nitrogen Dioxide
The formation of NO- in the atmosphere is a highly complex process. NO,
formed primarily as a result of fuel combustion, can be converted to NO-
through oxidation with atmospheric oxygen, ozone or various organic
compounds. Only about five percent of the NO in the combustion products
is in the form of NO-. Production of NO- by oxidation with atmospheric
oxygen is immediate for NO concentrations greater than about 100 ppm and
significant 0^ concentrations. Such high concentrations of NO only
occur very near the point of exhaust, however. As the NO is diluted to
concentrations below 100 ppm, the rate of conversion by oxygen decreases
to a point where at 1 ppm, the direct reaction with 0- becomes unimportant.
For the most part, the NO to NO- conversion through this mechanism is
limited to roughly ten percent of ambient NO concentrations.
A
Oxidizing agents such as ozone produce a rapid conversion of NO to NO-.
The degree of conversion, under certain conditions, is directly propor-
tional to the level of ozone present. In the absence of photochemical
activity, N0_ formation due to 0- is on a one-to-one basis. In the
presence of ultra-violet light, an equilibrium set of reactions between
NO- and ozone result. On days with high solar radiation, NO- levels are
set by these competing reactions. The presence of organic pollutants
(e.g., NMOC) may, in some cases, shift this equilibrium. In spite of
these very complex chemical processes, some success in applying rollback
in NO control strategies has been achieved, as will be shown in the
X
following discussions.
Measured violations of the current annual average NAAQS for NO- are
generally believed to be caused by area, especially mobile, sources.
2-16
-------
Therefore, only area source emitters of NO are typically included in
X
applications of the rollback model to NO- regulatory analyses. This
approach is supported by recent studies on the relative impact of mobile
and stationary sources on high NO- concentrations. One study analyzed
NO , CO, and SO- data from the Welfare Island monitoring site in New
York City (Iverach, 1978). It assumed that ambient SO can be used as a
surrogate for stationary source emissions and ambient CO can be used as
a surrogate for mobile source emissions. Therefore, the concentrations
of CO and SO during periods of high NO- should show which source type
predominates. 'The conclusion of the analysis was that on annual basis
for the area around the Welfare Island, NY site, NO emitted from mobile
and other area sources is more likely to result in high measured NO
than NO emitted from point sources.
x
Other studies also tend to suggest that mobile and area sources of NO
are the primary contributors to the levels occuring at ambient monitoring
sites. First, it has been shown (Chang, Norbeck, and Weinstock, 1980) ,
using LA Basin data, that high one-hour N0? resulted mainly from vehicular
sources. An examination of the NO /CO and SO-/CO ratios shows that a
X £»
significant impact of elevated NO sources on high N09 occurs rarely.
X t»
In fact, a stationary (both ground level and elevated) source contribution
of greater than 10 percent was rare at the LA Basin sites examined. The
only exception to this would be a case where stationary sources with low
stacks were near a monitor.
Chang, Norbeck, and Weinstock also concluded that rollback calculations
that assume high hourly N02 concentrations are proportional to the total
tonnage of NO emissions from all sources will greatly underestimate the
X
improvement in N02 air quality to be expected from a reduction in vehicle
N0x emissions in the LA Basin. Therefore, source contribution factors
must be used to discount the effect of emissions from stationary sources
on observed N02 levels. See Section 3.5 for a further discussion of
source contribution factors.
2-17
-------
A recent study by SRI International (Martinez and Nitz, 1979) examined
high (greater than 0.20 ppm) hourly NCL levels at 48 monitoring sites in
California during 1975-77. Approximately 1800 site-days when N02 exceeded
0.20 ppm were reviewed. It was found that mobile source emitters predo-
minated at all stations. Point-source effects linked to high hourly N0_
were infrequent.
In addition, Martinez and Nitz found that high hourly NO- levels in
Southern California coincide with increased emissions from stationary
area sources of NO . The diurnal variations of the high N00 levels at
x 2
the California monitors were associated with the traffic cycle. There-
fore, some combination of mobile and stationary area sources contributes
to peak NO- levels.
As noted above, these studies tend to support the contention that high
measured NO- concentrations are attributable to mobile or other area
source impacts and, thus point source emissions can be neglected. Therefore,
although ambient concentrations of N0_ are created by a mix of emissions
from point and area sources, it is suggested that for regulatory analyses
only area sources be modeled using rollback. This is a reasonable approach,
particularly if, as appears true from the available evidence, most
monitors in urban areas are sited to reflect contributions primarily
from area sources of NO emissions.
A
2-18
-------
REFERENCES FOR SECTION 2
Chang, T.Y., Norbeck, J.M., and Weinstock, B. February 1980. "NO Air
Quality Precursor Relationship: An Ambient Air Quality Evaluation in
the Los Angeles Basin". J. Air Poll. Control Assoc. Volume 30, No. 2.
deNevers, N. and Morris, J.R. 1975. "Rollback Modeling: Basic and
Modified," J. Air Poll. Control Assoc^ Volume 25, No. 9.
Dodge, M.C. June 1977. Effect of Selected Parameters on Predictions of
a Photochemical Model. EPA-600/3-77-048.U.S. EPA.Research Triangle
Park, N.C.
Energy and Environmental Analysis, Inc. December 28, 1979a. Regulatory
Impact Analysis for the National Ambient Air Quality Standards for
Carbon Monoxide. Arlington, VA.
Iverach, D. August 1978. "A Technique for Estimating the Relative
Impact of Mobile and Stationary Sources on N02 Concentrations". J. Air
Poll. Control Assoc. Volume 28, No. 8.
Martinez, J.R. and Nitz, K.C. August 1979. Analysis of High NO,, Concentrations
in California, 1975-1977 EPA-450/4-79-034a. SRI International. Menlo Park, CA.
(Prepared for U.S. EPA, RTP, NC).
Shelar, E., Ludwig, F.L., and Shigeishi, H. May 1979. Analysis of
Pollutant and Meteorological Data Collected in the Vicinity of Carbon
Monoxide "Hot Spots". Draft. SRI International, Menlo Park, CA.
(Prepared for U.S. EPA, Ann Arbor, MI.)
Trijonis, J., Peng, T., McRae, G. and Lees, L. January 1976. "Emissions
and Air Quality Trends in the South Coast Air Basin," EQL Memo No. 16
Environmental Quality Laboratory, California Institute of Technology,
Pasadena, CA.
Trijonis, J. and Hunsaker, D. February 1978. Verification of the
Isopleth Method for Relating Photochemical Oxidant to Precursors.
EPA-600/3-78-019. Published by U.S. EPA, Research Triangle Park, NC.
U.S. Environmental Protection Agency. November 1977. Uses, Limitations
and Technical Basis of Procedures for Quantifying Relationships Between
Photochemical Oxidants and Precursors. EPA-450/2-77-021a. Research
Triangle Park, NC.
2-19
-------
U.S. EPA. September 1978. Guidelines for Air Quality Maintenance
Planning and Analysis Volume 9 (Revised): Evaluating Indirect Sources,
EPA-450/ 4-78-001 (OAQPS No. 1.2-028R). Research Triangle Park, NC.
U.S. EPA. February 1979. Cost and Economic Impact Assessment for
Alternative Levels of the National Ambient Air Standards for Ozone.
EPA-450/5-79-002. Research Triangle Park, NC.
2-20
-------
3. SUGGESTED METHODOLOGIES FOR COMPILING MODELING DATA
This section discusses potential methodologies for compiling modeling
data for use in a nationwide analysis of CO, 0-, or NO- concentrations.
Procedures for estimating emission inventories, design values, growth
rates, retirement rates, control effectiveness, source contributions,
and transportation controls will be presented.
3.1 EMISSION INVENTORIES
The first step in compiling modeling data for use in applying rollback
or EKMA is to collect an emission inventory for each region of interest.
For each pollutant to be modeled, there are two key questions that must
be answered before inventorying can begin. They are:
What size area should be inventoried?
Within this area, which source categories have a significant
effect on monitored concentrations?
The first question can be answered by determining how large a source area
typically affects design value concentrations of the pollutant being
modeled. The design value and the emission inventory should be within a
common boundary. Emission patterns and ambient air quality data are
reviewed for each pollutant for a sample of urban areas to answer this
question. To answer the second question, ambient data and modeling
studies are examined separately for CO, 0_, and N0_.
3 £»
3.1.1 Ozone
In order to determine an appropriate source area for ozone, studies in a
number of cities were reviewed. The first such study examined here was
performed during the summer of 1977 in Tulsa, Oklahoma. (Eaton et al.,
3-1
-------
1979) Ozone was monitored at 10 sites; NO/NO at eight sites; non-
2v
methane hydrocarbons at four sites; wind speed and direction at four
sites; and solar radiation at one site. An airborne program measured
ozone, NO/NO , temperature and other variables for seven days during the
X
study period. This data base is useful in determining the spatial
distribution of ozone and the relationship between major emission sources
and observed maximum concentrations.
On three of the four days that the aircraft measured a distinct ozone
plume, the maximum ozone concentration was measured approximately 48 km
(30 mi) north of the center of Tulsa. (Aircraft monitoring flights were
only made on days with consistent southerly winds). On the remaining
day, the maximum ozone concentration was measured approximately 32 km
(20 mi) downwind. It is interesting that the urban ozone plume did not
experience a significant amount of spread in traveling the 64 km (40 mi)
to the location of the aircraft final horizontal traverse. The average plume
spread on all horizontal traverses was approximately 34 km (21 mi). The
city of Tulsa is approximately 26 km (16 mi) wide in the east to west
direction. Therefore, the ozone plume from Tulsa was only slightly
wider than Tulsa and did not widen appreciably in traveling the 64 km
downwind.
The morning ozone precursor concentrations at the downwind sites were
generally very low compared to those measured in the city. Therefore,
ozone synthesis from local sources at the downwind sites probably repre-
sents only a small contribution to the maximum ozone concentrations
measured at those sites. This means that a large percentage of the
ozone measured at those sites is from precursors which originated in the
Tulsa metropolitan area.
In addition, during the study, three sites were considered to be simul-
taneously under urban influence on 39 days. The maximum ozone con-
centration in the network was measured at Vera (34 km downwind) on 49
3-2
-------
percent of those days, at Ochelata (48 km downwind) on 32 percent of
those days, and at Sperry (14 km downwind) on 8 percent of those days.
This points out the widespread nature of high ozone levels, and the
necessity of modeling a large enough area to include both the monitor
with the highest 0_ levels and all sources affecting readings at that
monitor.
Project Da Vinci II involved ambient monitoring aloft of ozone and
precursors near and downwind of St. Louis (Decker et al., 1977). During
this study, ozone concentrations measured aloft above the radiation
3
inversion remained in the range 230 to 290 pg/m from 1700 on June 8
until the flight was ended the next day. These measurements strongly
suggest overnight transport of ozone aloft with minimum dilution and/or
destruction, (i.e., <20 percent). In addition, the occurrence of high
3
ozone concentrations (>250 (Jg/ffl ) on tne morning of June 9, 1976, in a
rural area in southwest Indiana was attributed to long range transport
of ozone from St. Louis. These measurements suggest that a large source
area should be modeled for ozone analyses.
Ozone data collected during the St. Louis RAPS study are also useful for
determining the appropriate source area for ozone. Table 3-1 shows
monthly maximum ozone values for July, August, and September for 24
sites in and near St. Louis. They are classified by their distance and
directional location with respect to the St. Louis Central Business
District (CBD) . The outermost sites are as much as 50 km from the CBD.
A map of these sites is shown in Figure 3-1. The monthly maxima in
Table 3-1 show no discernable pattern. At least one value greater than
the 0.12 ppm hourly standard was measured during these three months at
every site. Therefore, it is likely that a large source area contri-
butes to the design value monitor in St. Louis. At the least, it can be
concluded that a large area (AQCR) must be inventoried to ensure in-
clusion of all of the significant emissions affecting each site.
3-3
-------
TABLE 3-1
MONTHLY MAXIMUM HOURLY OZONE FOR RAPS SITES IN ST. LOUIS (1976)
CONCENTRATIONS IN PPM
Distance from
Site Clusters
1.
West of CBD
125
Center City
50-60
(km)
July
0.145
August
0.168
September
0.106
2. Western Edge of CBD
119
120
121
Central Business District
101
102
105
106
107
111
112
113
Downwind Edge of CBD
103
104
110
114
118
Downwind 10 km
109
115
116
117
Outer Periphery
122
123
124
10-20
10-20
20-30
0
<0
SO
SO
SO
SO
SO
10-20
SO
50
SO
10-20
10-20
10-20
20-30
20-30
10-20
50-60
30-40
40-50
10-20
10-20
20-30
.126
.155
.126
.160
.176
.187
.107
.146
.109
0
SO
SO
SO
SO
SO
SO
10-20
.151
.107
.122
.155
.154
.107
.154
.179
.145
.149
.180
.169
.152
.134
.134
.158
.092
.249
.098
.110
.090
.108
.188
.108
SO
SO
SO
10-20
10-20
.158
.112
.141
.223
.144
.202
.152
.146
.177
.143
.248
.087
.120
.132
.154
10-20
20-30
20-30
10-20
.159
.169
.167
.137
.180
.192
.180
.166
.120
.132
.124
.107
50-60
30-40
40-50
.148
.141
.160
.190
.010
.111
.126
.109
.153
3-4
-------
FIGURE 3-1
RAMS MONITORING SITES IN ST. LOUIS
£RSEYVILU
©122
RAPS CENTRAL
« RAMS STATIONS
E7HALTO
EDWARDSVILLE
-------
Most, if not all, of the available evidence suggests that VOC inven-
tories to be used in a nationwide rollback or EKMA modeling approach
should be compiled on an AQCR basis. The only evidence presented here
that suggests otherwise is the situation in Los Angeles, where it is
suggested that a county inventory covers the source area affecting
maximum ozone (Trijonis and Hunsaker, 1978). In order to estimate the
error involved in using an AQCR inventory where a county inventory is
appropriate, a sample of regions was used to compare the distribution of
sources in a county versus an AQCR. Data for this comparison were taken
from the National Emissions Data System (NEDS). Chicago, Washington,
Los Angeles, Philadelphia, Baltimore, Detroit, Pittsburgh, and Houston
were the cities used in the comparison. The results of this comparison
are shown in Table 3-2. The only area with a significant difference
between the distribution of sources is Washington, DC. The percentage
of mobile source emissions in the District of Columbia is much lower
than it is in the remainder of the AQCR. Because of the way NEDS calcu-
lates mobile source emissions, the miles traveled in DC by vehicles
registered and fueled in the neighboring States of Maryland and Virginia
are probably not included in the DC totals. Therefore, it is likely
that in actuality, the distributions of emissions in the District of
Columbia alone and in the entire AQCR are approximately equal.
Because the source distributions are roughly the same in counties and in
the AQCR's that encompass those counties, if AQCR inventories are used
in all areas, when in some cases a county inventory is appropriate, the
resulting errors will be minor.
It is recommended that VOC inventories for use in nationwide analyses be
compiled using the source categories listed in Table 3-3. These cate-
gories are essentially the same as those recommended by EPA for use in
preparing VOC emission inventories (U.S. EPA, 1977). Therefore, they
correspond closely to the inventories included in the SIP's. National
3-6
-------
TABLE 3-2
DISTRIBUTION OF VOC SOURCES FOR COUNTY VERSUS AQCR INVENTORIES FROM NEDS
PERCENTAGE OF ANNUAL EMISSIONS IN EACH SOURCE CATEGORY
Source Categories
AQCR 216 Harris AQCR 197 Allegheny AQCR 123 Wayne
Houston County SW Penn County Detroit County
AQCR 115 Baltimore
Baltimore County
1. Petroleum Refineries
2. Storage, Transportation and
Marketing of Petroleum
Products
3. Industrial Processes
4. Organic Solvent Evaporation
5. Stationary Combustion Sources
6. Mobile Sources
Source Categories
1
2
3
4
5
6
. Petroleum Refineries
. Storage, Transportation and
Marketing of Petroleum
Products
. Industrial Processes
. Organic Solvent Evaporation
. Stationary Combustion Sources
, Mobile Sources
.07
.18
.31
.19
.01
.24
AQCR 047
National
Capital
0
.05
0
.37
.01
.57
.03 0
.24 .05
.19 .09
.23 .41
.02 .01
.29 .44
AQCR 067
D.C. Chicago
0 .01
0 .10
0 .03
.64 .48
.01 .02
.35 .37
0 .01 .02 .01
.04 .05 .05 .04
.12 .02 .03 .04
.37 .54 .50 .45
.01 .02 .02 .02
.47 .36 .38 .44
AQCR 045 Philadel-
Cook Philadel- phia AQCR 024
County phia County Los Angeles
0 .03 .01 .05
.04 .19 .04 .06
.01 .04 .01 0
.54 .37 .42 .45
.01 .01 0 .04
.40 .36 .51 .39
.03
.04
0
.50
.01
.42
Los Angeles
County
.06
.07
0
.49
.04
.34
-------
TABLE 3-3
VOLATILE ORGANIC COMPOUND EMISSION INVENTORY
SOURCE CATEGORIES
Stationary Sources
1. Petroleum Refineries
2. Storage, Transportation and Marketing of Petroleum Products
3. Industrial Processes
4. Industrial Surface Coating
5. Other Solvent Use
6. Other Miscellaneous Sources
Mobile Sources
7. Light-Duty Gasoline Vehicles
8. Light-Duty Diesel Vehicles
9. Light-Duty Trucks
10. Heavy-Duty Gasoline Trucks
11. Heavy-Duty Diesel Trucks
12. Other Transportation Sources
3-8
-------
Emissions Data System (NEDS) data can also be easily structured into
these source categories. Light-duty diesel numbers are not readily
available at this time from NEDS, however.
3.1.2 Carbon Monoxide
CO concentration patterns across a city relate primarily to nearby
traffic levels and meteorological conditions. The most important
meteorological variables are wind speed, atmospheric stability, and
mixing height. CO concentrations usually vary greatly within an urban
area, so it is important to characterize only those emissions near the
design value monitor. However, two meteorological conditions, occurring
simultaneously or separately, can occur to produce widespread high CO
levels. One is thermal circulation, such as sea-land breezes, lake-land
breezes, drainage winds, or mountain-valley winds, that is caused by
differential heating of topographic features. These winds generally
flow in one direction during the day, then in the opposite direction at
night. As a result, an urban area can experience "blowback" of CO
emissions emitted during the day resulting in higher CO concentrations
at night. The second meteorological condition that can lead to wide-
spread high CO is inhibited vertical mixing caused by high atmospheric
stability or presence of a shallow mixed layer. An important example of
limited vertical mixing occurs as the result of a "surface" or "radiative"
inversion. Such conditions usually occur at night, with light winds and clear
skies. The surface inversion usually persists for hours and, because it
typifies stable atmospheric conditions, it can lead to high and rela-
tively widespread CO concentrations.
Presence of a shallow mixed layer may often occur as the result of a
subsidence inversion, caused by the adiabatic warming of a descending
atmospheric layer. The subsidence inversion usually persists on the
order of days and tends to contribute to high urban background CO concen-
trations.
3-9
-------
Because CO problems seem to be predominantly related to microscale
emission levels, and occasionally more widespread emissions (in the
presence of adverse meteorological conditions), the recommended inventory
area is a county. While in many cases, nonattainment areas for CO are
smaller than a county (sometimes they include only a few blocks in a
downtown area) for practical purposes a county inventory is the most
reasonable approach for a nationwide study. This is true because the
NEDS summaries are available only down to the county level.
Table 3-4 lists the source categories suggested for compiling county
emission inventories for use in a nationwide analysis. These categories
are recommended based on the assumption that stationary sources of CO do
not significantly affect concentrations measured at design value monitors,
and the source categories typically used in NEDS are sufficient (see
Section 3.5 for a more complete discussion of source contributions).
3.1.3 Nitrogen Dioxide
Observed concentrations of NO- result primarily from chemical reactions
in the ambient air involving NO and other pollutants. The principal
mechanisms leading to high NO- levels are photochemical synthesis and
the reaction of NO with ozone (titration). Photochemical synthesis
occurs when NO and NMOC react in the presence of sunlight to form NO .
Site locations which primarily experience high NO- concentrations due to
photochemical synthesis are usually within 1-4 hours travel time of
where heavy NO and VOC emissions occur. The titration reaction of NO
X
with ozone to form N02 occurs very rapidly in the atmosphere, much
faster than the photochemically driven synthesis.
3-10
-------
TABLE 3-4
CARBON MONOXIDE EMISSION
INVENTORY SOURCE CATEGORIES
Mobile Sources
1. Light-Duty Gasoline Vehicles
2. Light-Duty Diesel Vehicles
3. Light-Duty Trucks
4. Heavy-Duty Gasoline Trucks
5. Heavy Duty Diesel Trucks
6. Other
Off-Highway
Rail
Aircraft
Vessels
Stationary Sources
7. External Combustion
Residential
Industrial
Commercial-Institutional
8. Solid Waste Disposal
9. Misc el laneous
-------
FIGURE 3-2
1977 N02 CONCENTRATIONS IN THE LOS ANGELES AIR QUALITY CONTROL REGION
CO
I
_AQC|l_Baindlli|
A
B-Ami
C-Burtunk...
.109
..121
IM
D-Cimlllo 43
E-GuUNea M1
MONITORING SITE AND READINGS
F Li Habn 1W1 K Los Angles Ct) 105
G Lenxu 122 \_ Lymnod 120
H-LontBudi Ul M-Nnilull M
I -Los Antfln IK H-P>»d«* 163
J Los Annl«, 1* 0-Poun 134
P-Redtandi (0
fl KuKdon M'
R-SwBetui4iiu Si
S-Smll Balkan ft'
T-HMttiM I3«
1977 ARITHMETIC MEAN
1 MuswedbySiltaunColoiinUiclMiod
2 Hun fcmtd Iroa l«ss than 75toMoljl
possible obunatigns
HUES
-------
Therefore, a monitor with high N02 readings produced by a reaction with
ozone is most likely near significant NO sources.
Because the highest annual average NCL concentrations will be measured
at less than four hours travel time, and oftentimes immediately downwind,
of significant NO sources, it seems appropriate to compile NO inventories
X X
for nationwide studies on a county basis. This conclusion can be judged
further by examining spatial variation patterns of ambient N0? data.
Few areas have enough N02 monitors to allow spatial variation patterns
to be determined. The Los Angeles Basin and the St. Louis area do have
sufficient data; Los Angeles through regular monitoring and St. Louis
from the RAPS study. Figure 3-2 shows isopleths of annual average NO
in the Los Angeles Basin from 1977 data. Based on the Los Angeles data,
for an analysis of annual mean NO., it is probably appropriate to use a
county inventory.
During 1976, 18 individual stations monitored hourly NO- concentrations
in St. Louis as part of the RAPS study. The site locations are shown in
Figure 3-1. Site 101 is in the city center. If the other sites are
grouped with respect to their distance from site 101, conclusions can be
made about the spatial variation in N02 concentrations in the vicinity
of St. Louis.
Site Distances from Number of Annual Average
the City Center (km) Sites N02 (ppm)
0 1 .028
0-4 3 .027
4-10 5 .027
10-20 7 .015
20-30 2 .012
3-13
-------
The annual averages in the above table are averages of the 1976 annual
means at all of the sites in each group. While the annual averages
within 10 km of the city center are relatively constant at .027-.028
ppm, there is a considerable drop in the annual average at sites more
than 10 km from site 101. Therefore, it seems reasonable to assume that
a county rather than an AQCR-wide emission inventory is appropriate for
characterizing the sources contributing to the design value concen-
trations in St. Louis. This is consistent with the findings in Los
Angeles.
NO emission inventory source categories recommended for use in nation-
wide studies are shown in Table 3-5. Point sources are aggregated into
one category because it was found that point sources do not significantly
impact NO. concentrations at the present set of continuous NO monitors.
They can be used in emission projections, however. Source contributions are
discussed further in Section 3.5.
3.2 DESIGN VALUES
The design value is defined as the concentration that must be reduced to
the level of the standard for an area to be in attainment. In using
ambient monitoring data in national strategy assessments, design values
are usually chosen to correspond to the form of the standard for each
pollutant. In addition, design values can be chosen from a time period
consistent with the emission inventory. Therefore, recommended techniques
for determining design values are presented separately for CO, 0-, and
N02. The EPA Storage and Retrieval of Aerometric Data (SAROAD) System
is the source of monitoring data typically used to determine design
values for use in nationwide studies. SAROAD is simply a compilation of
ambient air quality data from air monitoring operations of State, local,
and Federal networks.
3-14
-------
TABLE 3-5
OXIDES OF NITROGEN EMISSION
INVENTORY SOURCE CATEGORIES
Mobile Sources
1. Light-Duty Gasoline Vehicles
2. Light-Duty Diesel Vehicles
3. Light-Duty Gasoline Trucks
4. Heavy-Duty Gasoline Trucks
5. Heavy-Duty Diesel Trucks
6. Other Mobile Sources
Off-Highway
Aircraft
Vessels
Locomotives (Rail)
Fuel Combustion (Area Sources)
7. Residential Oil/Gas
8. Comm/Inst. Coal
9. Comm/Inst. Oil/Gas
10. Industrial Coal
11. Industrial Oil/Gas
12. Other
3-15
-------
TABLE 3-5 (Continued)
Solid Waste
13. S.W. Disposal
Point Sources
14. Point Sources
3-16
-------
3.2.1 Carbon Monoxide
There are currently two NAAQS for carbon monoxide; an eight-hour standard
of nine ppm and a one-hour standard of 35 ppm. As a result of the
review and revision of the health criteria, EPA has recently proposed to
lower the one-hour standard to 25 ppm (45 FR 55066). For the most part,
the eight-hour standard is more stringent; therefore, all of the design
values used in this study are for an eight-hour averaging time.
Under the present form of the standard, design values are determined by
choosing the second highest monitor reading (non-overlapping) in any
year. Thus, the highest second high in the three year period encompassing
the emission inventory base year should be chosen as the design value.
While this technique for choosing a design value is acceptable given the
current deterministic form of the standard, a change to a statistical
standard has been proposed. A daily interpretation of exceedances of
the CO NAAQS has been proposed which is consistent with the interpreta-
tion of exceedances for the ozone NAAQS.
3.2.2 Ozone
The ozone NAAQS contains the phrase "expected number of days per calendar
year." This differs from the previous NAAQS for photochemical oxidants
which specifies a concentration "not to be exceeded more than once per
year." Although the new standard form appears to be complicated, the
basic principle is relatively straightforward. In general, the average
number of days per year above the level of the standard must be less
than or equal to one. The simplest method for determining compliance
with the standard is to record the number of exceedances at each site
each year and then average this number over the past three years to
determine if it is less than or equal to one. Adjustments can be made
for incomplete monitoring data and year to year variations in emissions
and meteorology. (U.S. EPA, 1979a).
3-17
-------
Because the choice of a design value is primarily influenced by the few
highest values in a data set, it is possible to construct a simple table
look-up procedure to determine a design value (U.S. EPA, 1979a). To use
this tabular approach, it is only necessary to know the total number of
daily values and a few of the highest values. For example, if there are
1,017 daily values, then the ranks of the lower and upper bounds obtained
from Table 3-6 are 3 and 2. This means that an appropriate design value
would be between the third-highest and second-highest observed values.
In using this table, the upper bound should be used as the design value.
For data sets that are 75 percent complete, but still have less than 365
days of data, the maximum observed concentration may be used as a tenta-
tive design value as long as the data set was 75 percent complete during
the peak times of the year.
For a complete discussion of this table look-up procedure and other
alternative methods for estimating ozone design values, refer to the
Guideline For the Interpretation Of Ozone Air Quality Standards (U.S.
EPA, 1979a).
3.2.3 Nitrogen Dioxide
Annual average NO design values can be simply determined by choosing
the highest annual average from the past three years. However, care
should be taken to ensure that only valid annual averages are considered
in choosing a design value. SAROAD uses the criterion for calculating
annual means from continuous observations with sampling intervals of
less than 24 hours that data must reflect a minimum of 75 percent of the
total number of possible observations for the applicable year. Criteria
for noncontinuous observations with sampling intervals of 24 hours or
greater are as follows:
3-18
-------
TABLE 3-6
TABULAR ESTIMATION OF DESIGN VALUE
Number of Daily
Values
Rank of Upper
Bound
Rank of Lower
Bound
Data Point Used
for
Design Value
365 to 729
730 to 1094
1095 to 1459
1460 to 1824
1825 to 2189
1
2
3
4
5
2
3
4
5
6
highest value
second highest
third highest
fourth highest
fifth highest
SOURCE: U.S. EPA. January 1979. Guideline for the Interpretation
of Ozone Air Quality Standards.EPA-450/4-79-003.OAQPS
No, 1.2-108. Research Triangle Park, NC.
3-19
-------
Data representing quarterly periods must reflect a minimum of
five observations for the applicable quarter. If there are no
measurements in one of the three months of the quarter, each
remaining month in that quarter must have no less than two
observations reported.
Data representing annual periods must reflect four quarters of
observation that have satisfied the quarterly criteria.
3.3 GROWTH AND RETIREMENT RATES
3.3.1 Stationary Sources
Changes in stationary source activity levels are accounted for in the
rollback model by a combination of growth and retirement rates. This is
done because regulations affecting new sources differ from those af-
fecting existing sources. Therefore, different control assumptions are
identified for each. Growth rates and controls are applied to estimate
new source emissions. Retirement rates are applied to estimate how
emissions from existing sources will decrease.
The most recent industry-by-industry projections of growth applicable to
an air pollution analysis are those performed by the Bureau of Economic
Analysis (BEA, 1977). They are shown in Table 3-7. These growth rates
were calculated based on earnings and, therefore, are assumed to repre-
sent net growth for an industry. In other words, retirement of existing
sources is taken into account. Retirement rates for existing sources
are shown in Table 3-8. These estimates of plant retirement rates were
developed for EEA by Data Resources, Inc. (U.S. DOE, 1979) in support of
EEA's ISTUM model development.
The equation which should be used to incorporate the data in Tables 3-7
and 3-8 is as follows:
3-20
-------
TABLE 3-7
STANDARD INDUSTRIAL CLASSIFICATION (SIC) ANNUAL GROWTH RATES*
Industry SIC
Agriculture 1-7
Forestry and Fisheries 8-9
Metal Mining 10
Coal Mining 11-12
Oil and Gas Extraction 13
Mining and Quarrying 14
Contract Construction 15-17
Food and Kindred Products 20
Textile Mill Products 22
Apparel and Other Textiles 23
Lumber and Furniture 24-25
Paper and Allied Products 26
Printing and Publishing 27
Chemicals and Allied Products 28
Petroleum Refining 29
Primary Metal Industries 33
Fabricated Metal Products 34
Machinery, Except Electrical 35
Electrical Machinery 36
Transportation Equipment 37
Miscellaneous Manufacturing Industries 39
Railroad Transportation 40
Motor Freight Transportation 42
Transportation Services 47
Communication 48
Electric, Gas and Sanitary Services 49
Average Annual
Growth (1977-2000)
0.6
1.7
1.3
3.1
2.2
1.5
3.6
1.2
2.4
2.0
2.3
2.7
2.3
3.1
1.9
2.4
2.5
2.6
2.9
1.5
3.3
0.2
3.2
2.9
4.9
3.5
Population Growth Rate
0.8
SOURCE: Bureau of Economic Analysis, 1977.
* Based on projected earnings by industry, calculated in 1967 dollars.
3-21
-------
TABLE 3-8
INDUSTRIAL RETIREMENT RATES
Industry
Agricultural Production
Agricultural Services
Forestry
Fishing, Hunting, and Trapping
Metal Mining
Anthracite Mining
Bituminious Coal and Lignite Mining
Oil and Gas Extraction
Mining and Quarrying
Building Construction
Construction Other than Buildings
Construction - Special Trade
Food and Kindred Products
Tobacco
Textile Mill Products
Apparel
Lumber and Wood Products
Furniture and Fixtures
Paper and Allied Products
Printing and Publishing
Chemicals and Allied Products
Petroleum Refining
Rubber and Miscellaneous Plastics
Leather and Leather Products
Stone, Clay, Glass and Concrete
Primary Metal Industries
Fabricated Metal Products
Machinery, Except Electrical
Electrical Machinery
Transportation Equipment
Miscellaneous Instruments
Miscellaneous Manufacturing Industries
Railroad Transportation
Interurban Transit
Motor Freight Transportation
U.S. Postal Service
Water Transportation
Air Transportation
Pipe Lines, Except Natural Gas
Transportation Services
Communication
Electric, Gas and Sanitary Services
General Government, Except Finance
Justice, Public Order, and Safety
Public Finance and Taxation
Average Annual
Retirement Rates
01
07
08
09
10
11
12
13
14
15
16
17
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
91
92
93
26
26
26
26
26
26
26
26
26
26
26
26
4.56
3.35
20
18
3,
3,
6.37
,68
.13
4.92
5.07
4.48
2.97
,09
,93
,97
3.23
4.10
4.61
4.09
4.83
4.41
4.26
4.26
4.26
4.26
4.26
4.26
4.26
4.26
4.26
4.
4.
4.
26
26
26
4.26
SOURCE: U.S. Department of Energy, 1979
-------
QQ = QQ {[(1 + G.)t - 1] FQ + (1 - Ri)t Fe + [1 -(1 - R^] Fn} (3-1)
where:
Q = emissions in projection year
Q = emissions in base year
R. = retirement rate
i
F = emission factor ratio for existing sources
G. = growth rate
F = emission factor ratio for new sources
n
The first term in the equation represents new source growth and controls,
the second term accounts for retirement and controls for existing sources,
and the third term accounts for replacement source controls. It should
be noted that the G. term in equation (3-1) represents net growth. If a
total growth rate (G.') is used, the first term in equation (3-1) should
be changed to [(1 + G.' - R.)1" - 1] F . (Estimates for F and F are discussed
6 lv i i' J n en
in Section 3.4.)
In order to facilitate use of the growth and retirement rates in Tables 3-7
and 3-8 in a nationwide analysis, Tables 3-9 and 3-10 are included.
They present the data in Tables 3-7 and 3-8 for 15 different stationary
source category groupings commonly used in nationwide analyses. Annual
percentage ranges in growth and retirement along with the applicable SIC
category are shown in the tables.
3.3.2 Mobile Sources
One of the most difficult and important variables to project for future
years is growth in vehicle miles traveled (VMT) which is used as a
surrogate for mobile source emissions growth. Growth in VMT is difficult
to predict for two reasons. One is the uncertainty in how changes in
fuel prices will .affect demand for gasoline and travel. The second is
3-23
-------
TABLE 3-9
STATIONARY SOURCE ACTIVITY LEVEL GROWTH
Source Categories
1. External fuel combustion
2. Internal fuel combustion
3. Chemical manufacturing
4. Food/agriculture industry
5. Primary metal products
6. Mineral products
7- Petroleum industry
8. Wood products
9. Industrial solvent evaporation
10. Other industrial processes
11. Solid waste disposal
12. Petroleum storage & transport
13. Gasoline service stations
14. Area source solvent evaporation
15. Mi sc ellaneous
1975-1995
Annual
Percentage
Change
3.5
3.5
3.1
0.6
2.4
1.3
1.9
2.3
3.3
3.3
0
1.9
1.9
0.8
0
As sumed
SIC
49
49
28
20
33
10
29
24
39
39
-
29
29
population
SOURCE: BEA, 1977.
3-24
-------
TABLE 3-10
STATIONARY SOURCE RETIREMENT RATES
1. External fuel combustion
2. Internal fuel combustion
3. Chemical manufacturing
4. Food/agriculture industry
5. Primary metal products
6. Mineral products
7. Petroleum industry
8. Wood products
9. Industrial solvent evaporation
10. Other industrial processes
11. Solid waste disposal
12. Petroleum storage and transport
13. Gasoline service stations
14. Area source solvent evapoation
15. Miscellaneous
Annual
Percentage
Change
4.3%
4.2
5.1
4.6
5.0
4.9
4.5
6.4
4.4
4.4
0
4.5
4.5
0
0
Assumed
SIC
49
49
28
20
33
10
29
24
39
39
-
29
29
-
^
SOURCE: U.S. DOE, 1979.
3-25
-------
identifying how VMT will change in the areas that specifically affect
high pollutant levels. For instance, growth in VMT in central business
districts may vary from that in outlying areas. VMT is an important
variable to estimate accurately because the rollback or EKMA predictions
are sensitive to the value assumed, especially if projections are made
more than 10 years into the future.
On a nationwide basis, VMT has historically grown at a rate of slightly
over four percent per year (Congress, 1979). Immediately prior to the
1973-1974 oil embargo, annual growth reached nearly five percent. After
a rapid drop during the embargo, VMT gradually increased again to a four
percent annual rate. However, recent events suggest that fuel prices
will continue to rise, and VMT may not increase as rapidly as historical
trends indicate. Gasoline consumption fell by five percent in 1979 and
is down eight to nine percent so far this year (1980) . (With recent
fuel economy gains, changes in gasoline consumption do not track directly
into VMT changes). Until the elasticity of gasoline demand with respect
to price can be quantified, it is only speculation how VMT will change
in future years.
The California Transportation Department determined that weekend traffic
on State highways declined by 11 percent last December (1979) from a
year earlier. However, weekday traffic increased by four percent. It
is important to make this differentiation when performing an air pollu-
tion analysis, because the traffic affecting the periods of highest
concentration are of interest. Assuming that most high concentrations
are observed on a week day, growth in travel during that period should
be used.
Although one response to rising gasoline prices is reducing the number
of trips, work trips are not likely to be reduced as much as trips for
other purposes. These are the trips that contribute most to high pollutant
levels. Therefore, in estimating future VMT growth rates, it is important
not to use total VMT changes when weekday travel is of interest.
3-26
-------
The second major issue regarding VMT growth estimates is that only the
VMT specifically affecting the design value concentrations should be
considered in growth projections. This means that SMSA growth rates are
not always applicable. If the nonattainment area is the central business
district, then expected growth in VMT in that area alone should be
included in the projections. In addition, if a specific time period of
the day is of interest, e.g., 6-9 a.m., then VMT changes for that period
should be accounted for as much as possible.
Despite the fact that in most analyses the same growth rate is used for
all mobile source categories, it is appropriate to use a separate growth
rate for each vehicle type if possible. Therefore, the nationwide
average growth rates presented in Table 3-11 are specific to each vehicle
type. The 1977 market share for diesels is negligible when compared
with 1990. Therefore, total VMT estimates are shown for both years.
In cases where nationwide average growth rates do not apply and it is
necessary to estimate how VMT growth on urban roads (especially within
the CBD) is likely to change in future years, it is useful to consider
how roadways are designed. Rather than build a road to handle the
maximum volume likely to travel on it during a year, traffic engineers
either measure or estimate the hourly distribution of traffic volume
during a year, pick an hourly traffic volume along that distribution,
and build the roadway with the chosen capacity. This is done because it
is uneconomical to provide for the extreme hourly volumes of traffic
that may occur but a few times during a year. Based on the relationship
between the peak hourly flows and the annual average daily traffic, it
has been determined that the ratio of benefit to expenditure is near its
maximum at the thirtieth highest hourly traffic volume (Wohl and Martin,
3-27
-------
TABLE 3-11
MOBILE SOURCE GROWTH RATES*
A, Highway Vehicles
1. Gasoline
B
a. Light-Duty Vehicles
b. Light-Duty Trucks I
c. Light-Duty Trucks II
d. Heavy-Duty Trucks
e. Motorcycles
2. Diesel
a. Light-Duty Vehicles
b. Light-Duty Trucks I
c. Light-Duty Trucks II
d. Heavy-Duty Trucks
Off-Highway Vehicles
1. Gasoline
2. Diesel
C. Locomotives
D, Aircraft
E. Vessels
Annual
Percentage
Change
0.5%
2.3
6.2
-2.0
2.5
5.0
2.3
2.3
0.2
2.7
2.7
1977
VMTq
(x 10*)
1990
VMTq
(x 10*)
1.23
0
0
201.79
32.19
33.65
Light-duty vehicle and light-duty truck growth are estimated from
data presented in McNutt, Dulla, and Lax, 1979. Heavy-duty truck
growth rates are taken from U.S. EPA, 1979b. Growth for motorcycles
is assumed to be 2.5%, which is FHWA's latest estimate of expected
growth in VMT for the vehicle fleet as a whole. For off-highway
vehicles, growth rates are estimated from projections of growth in
earnings by industry (BEA, 1977).
3-28
-------
1967). Hence, for most cases, the thirtieth highest hourly volume for
the year is generally a reliable criterion of the needed capacity for
which it is most practical to design.
This design criterion, if commonly used, has important implications for
analyzing future VMT growth. If peak hour traffic is of interest, and
roadway capacity is typically reached an average of 30 times per year,
then even if VMT grows during off-peak periods, mobile source contribu-
tions to peak pollutant levels are unlikely to change. However, this
conclusion is dependent on the averaging time of the standard being
examined.
A question that has been raised a number of times with respect to nation-
wide studies is whether area specific growth rates should be used for
mobile sources. In order to test the sensitivity of rollback results to
the use of area specific growth rates, a test case was run using the
data from the CO regulatory analysis. The results are shown in the
table below:
Average Percentage Changes in CO Concentrations
(1976 Base Year)
Projection Years
Growth
Scenarios 1982 1985 1987 1990 1999
No Growth
1% Growth
3% Growth
Area Specific
Growth -17 -33 -40 -46 -40
The above table shows that rollback predictions are sensitive to the
assumed growth rate, but that area specific growth rates are roughly
-28
-25
-16
-46
-41
-32
-54
-49
-39
-61
-56
-45
-65
-58
-39
3-29
-------
equivalent to using a three percent VMT growth assumption for the whole
country. The indicators of "number of cities above the standard" and
"the total number of expected violations" are also essentially equiva-
lent for the three percent and the area specific growth assumptions.
The area specific growth rates are based on DOT-FHWA projections from
1975-1990. These projections were all submitted by the individual
States and may be derived from differing assumptions.
In summary, it is recommended that the point estimates of mobile source
growth in Table 3-11 be used in analyses where nationwide average VMT
growth rates are needed. Because there is considerable uncertainty in
predicting how VMT will change in future years, upper and lower bound
growth rates should be included in the analysis in addition to the point
estimate. In adjusting the Table 3-11 growth rates to run sensitivity
analyses, it is probably appropriate to assume that growth increases or
decreases proportionately for all vehicle categories.
3.4 CONTROL EFFECTIVENESS
3.4.1 Mobile Sources
The effectiveness of future controls for mobile sources can be estimated
using EPA's mobile source emission factor program (MOBILE2). This
program accounts for both expected future emission levels by model year
by vehicle type and the rate of retirement of existing vehicles from the
present fleet. For use in a nationwide study, MOBILE2 emission factors
for the base year and the projection year are calculated and used to
determine an emission factor ratio. This emission factor ratio is
simply expressed as:
Emission Factor Ratio = Projection Year Emission Factor
Base Year Emission Factor
This emission factor ratio is calculated separately for each motor
vehicle type.
3-30
-------
Despite slight differences in emission standards for California and high
altitude areas, the technologies applied to reduce mobile source emis-
sions are the same throughout the country. California and high altitude
emission factors should be used where appropriate, however, using emission
factor ratios computed from 49-state emission factors in a nationwide
analysis will yield nearly the same results as using California specific
emission factors.
The MOBILE2 emission factors program is used to account for controls
mandated by the Federal Motor Vehicle Emission Control Program (FMVECP).
If an estimate of the potential effectiveness of an inspection and
maintenance (I/M) program is needed, the MOBILE2 program (U.S. EPA,
1978) can be used to generate this estimate. Because the emission
factors including I/M credits are calculated within the MOBILE2 program,
the effect of an I/M program is accounted for in rollback by changing
the emission factor ratios.
I/M programs are required by the Clean Air Act Amendments (1977) if
SIP's for ozone and/or CO cannot demonstrate attainment of the NAAQS by
1982. The original timetable for implementation of I/M programs was as
follows (U.S. EPA, 1978):
June 30, 1979 Legal Authority for I/M Programs Required
December 31, 1981 Decentralized I/M Programs in Place
December 31, 1982 Centralized I/M Programs in Place
Although not all states met the June 30, 1979 deadline for enabling
legislation, as of September 1, 1980, only two of the 29 states required
to have I/M legislation did not have it. Taking into consideration the
3-31
-------
fact that several areas already have operating I/M programs, it is
reasonable to assume that on a nationwide average, all areas projected
to not attain the CO and/or the 0_ standard by December 31, 1982 will
have an I/M program operating beginning in 1982.
If changes in mobile source emission standards need to be evaluated,
MOBILE2 can again be used to compute calendar year emission factors to
account for these revisions. To change the emission standards for a
vehicle type, new values for the zero-mileage emission rate and the
emission factor deterioration rate per 10,000 miles must be input to
MOBILE2. Changes in emission standards can and should be evaluated
separately for each vehicle type. Again, emission factor ratios are
used to estimate the impact of mobile source controls.
3.4.2 Stationary Sources
Estimating the effectiveness of stationary source controls for use in a
nationwide analysis is not as straightforward as it is for mobile source
controls because there are many more source types to deal with. Sta-
tionary source categories used in nationwide analyses are usually chosen
so that source types with common growth rates and control efficiencies
are grouped together. Therefore, the control efficiency used in the
rollback or EKMA model is a weighted average based on control effi-
ciencies for each source type. This process is further complicated by
the fact that control efficiencies within each source type differ based
on the type of fuel burned.
Of the three pollutants being discussed in this report, ozone is the
only one that has been identified as having a significant stationary
source influence at monitors measuring design values. Therefore, this
section concentrates on providing control efficiency estimates for VOC
sources. In nonattainment areas, the Clean Air Act Amendments require
new and modified stationary sources to meet the lowest achievable emission
3-32
-------
reduction (LAER) before a source can be located in an area or modified.
Since EPA has not determined what constitutes LAER for stationary VOC
sources, it should be assumed that new and modified sources will have to
achieve a minimum level of control equal to reasonably available control
technology (RACT).
For areas that fail to attain the standard with mobile source and new
and modified stationary source control, existing sources will also be
required to install RACT. RACT will vary among industries and may well
vary among sources within an industry. RACT is defined as the lowest
emission limit that a particular source is capable of meeting by applying
control technology that is reasonably available considering technological
and economic feasibility. Since economic feasibility is basically
source specific, RACT may vary among sources in an industry. Since
national assessments generally only consider broad categories of sources
and do not consider the feasibility of RACT on individual sources, RACT
is broadly defined here as technology available for application in
categories of sources which will lead to adequate levels of control
based solely on technical considerations. In actual practice, the
reasonableness of technology for particular sources will be determined
by State and local agencies based on technical guidance issued by EPA
and depending upon economic and energy feasibility.
EPA has published guideline documents for VOC sources which assess the
technology available for these sources. Based on these documents,
estimates were made of the efficiency of RACT for major VOC sources.
These estimates are shown in Table 3-12. As of January 1978, Control
Techniques Guideline (CTG) documents had been issued for 15 stationary
source categories. Legally enforceable RACT regulations must be sub-
mitted by the states on an annual basis in the January of the year
following the publication of the CTG (Clean Air Act, 1977). Therefore,
3-33
-------
TABLE 3-12
ASSUMED EFFICIENCY IDENTIFIED OF RACT FOR
VOC STATIONARY SOURCE CATEGORIES
Efficiency of Assumed
Source Category and Sources Included* RACT (%) SIC
1. Chemical Manufacturing 28
Organic Chemical Manufacturing Industry
- Process Emissions 90%
- Fugitive Emissions 80%
- Storage and Loading Emissions 90%
- Secondary Emissions 75%
Pharmaceutical Industry 95%
Paint Manufacture 90%
Rubber Industry 75%
Weighted Average of RACT** 82%
Current Emissions Affected by RACT*** 79%
Emission Reduction in Source
Category Achievable with RACT 65%
2. Petroleum Industry 29
Gas and Crude Oil Production 90%
Petroleum Refining
- Vacuum Jets 100%
- Waste Water Separators 95%
- Miscellaneous Sources 91%
- Process Unit Turnaround 98%
Natural Gas and Gasoline Plants 96%
Weighted Average of RACT** 95%
Current Emissions Affected by RACT*** 95%
Emission Reduction in Source
Category Achievable with RACT 90%
3. Industrial Solvent Evaporation
Auto and Light Duty Truck Manufacturing 80%
Flatwood Products 80%
3-34
-------
TABLE 3-12 (Continued)
ASSUMED EFFICIENCY IDENTIFIED OF RACT FOR
VOC STATIONARY SOURCE CATEGORIES
Efficiency of Assumed
Source Category and Sources Included* RACT (%) SIC
Paper Coating 80%
Fabric Coating 80%
Wire Coating 90%
Can Coating 80%
Metal Furniture 85%
Industrial Machinery 80%
Commercial Machinery 80%
Coil Coating 85%
Fabricated Metal Products 80%
Large Appliances 85%
Small Appliances 80%
Weighted Average of RACT** 80%
Current Emissions Affected by RACT*** 75%
Emission Reduction in Source
Category Achievable with RACT 60%
4. Area Source Solvent Evaporation
Dry Cleaning 65%
Vapor Degreasing 55%
Cold Cleaning 50%
Graphic Arts 80%
Adhesives 80%
Cutback Asphalt Paving 100%
Weighted Average 75%
Current Emissions Affected by RACT 39%
Emission Reduction in Source
Category Achievable with RACT 30%
3-35
-------
TABLE 3-12 (Continued)
ASSUMED EFFICIENCY IDENTIFIED OF RACT FOR
VOC STATIONARY SOURCE CATEGORIES
Source Category and Sources Included*
Efficiency of
RACT (%)
Assumed
SIC
5. Petroleum Storage and Transport
Gasoline and Crude Oil Storage
Gasoline Bulk Terminals
Gasoline Bulk Plants
Weighted Average
Current Emissions Affected by RACT
Emission Reduction in Source
Category Achievable with RACT
29
75%
95%
76%
80%
100%
80%
6. Gasoline Service Stations
Storage
Refueling
Weighted Average of RACT**
Current Emissions Affected by RACT***
Emission Reduction in Source
Category Achievable with RACT
29
90%
90%
90%
83%
75%
7. Fuel Combustion****
8. Other Industrial Processes****
49
39
9. Solid Waste****
10. Miscellaneous****
SOURCES: U.S. EPA, 1979
Peterson and Sakaida, 1978
3-36
-------
TABLE 3-12 (Continued)
ASSUMED EFFICIENCY IDENTIFIED OF RACT FOR
VOC STATIONARY SOURCE CATEGORIES
* Sources included are those for which screening studies or guideline
documents are being prepared. Many solvent evaporation sources
and other industrial processes are not included because the sources
have not been identified or control levels have not been defined.
** Represents weighted average of RACT for sources listed based on
current national emissions for each category.
*** Represents the proportion of current national emissions for which
RACT can be applied. Excludes emissions from source categories
for which RACT control levels have not been identified and the
residual emissions from sources which have already controlled to
the RACT level.
**** RACT has not been identified for these source categories.
3-37
-------
it is reasonable to assume that unless projections are being made for a
year prior to 1982, that the RACT efficiencies in Table 3-12 can be
used.
If the emission inventory in each region is grouped by the source cate-
gories identified in Section 3.1, the average emission reduction that
could be achieved with the application of identified RACT must be deter-
mined for these categories. To do this, it is first necessary to derive
the weighted average of the efficiency of RACT for the sources in each
category for which RACT has been estimated. This weighted average takes
into account each sub-category's relative contribution to current nation-
wide emissions, and the efficiency of RACT for each source. Therefore,
the total emission reduction in each general source category is determined
by multiplying the weighted efficiency of RACT for applicable sources by
the percentage of emissions in the general source category which would
be affected by RACT. The emissions affected by RACT include only those
sources for which RACT has been identified and which have not already
controlled emissions to the RACT level.
As Table 3-12 indicates, emissions from existing VOC sources can be
reduced by 65 percent from the chemical manufacturing industry (indus-
trial processes), 90 percent from the petroleum industry, 80 percent
from petroleum storage and transport, 75 percent from gasoline service
stations, and 60 percent from industrial surface coating operations.
However, controls on identified area source solvent evaporation sources
will reduce emissions from the total category by only 30 percent, since
over half of the emission inventory results from sources which have not
been identified and for which control levels have not been estimated.
Table 3-13 presents stationary source control efficiencies for CO and
NO^. While they will not be commonly needed for nationwide studies
where design values are measured at urban monitors, they are included
3-38
-------
TABLE 3-13
STATIONARY SOURCE CONTROL EFFICIENCIES
1. External fuel combustion
2. Internal fuel combustion
3. Chemical manufacturing
4. Food/agricultural industry
5. Primary metal products
6. Mineral products
7. Petroleum industry
8, Wood products
9. Industrial solvent evaporation
10. Other industrial processes
11. Solid waste disposal
12. Petroleum storage & transport
13. Gasoline service stations
14. Area source solvent evaporation
15. Miscellaneous
* Source: PEDCo Inc., 1979.
** Source: U.S. EPA, 1979a.
CO*
RACT
90%
90
RACT
25%
30
NO **
NSPS
33%
30
LAER
90%
60-90
90
99
99
25
30-60 90
3-39
-------
because they might be useful in emission projections or special studies
where stationary source influences are likely. CO control efficiencies
in Table 3-13 are assumed to represent RACT. A wider variety of control
options are available for NO , so three levels of control are identified
A
for that pollutant in Table 3-13.
While the CO control efficiencies for RACT shown in Table 3-13 are
currently available control technologies, the likelihood of their being
applied to meet the current CO ambient standards is low. In most cases,
CO SIP's concentrate on controlling mobile soures.
For NO , there are currently very few nonattainment areas. Therefore,
A
RACT controls will only be applied to areas with annual averages greater
3
than 100 pg/m . New Source Performance Standard NO controls for utility
A
boilers were originally proposed in 1971 and should affect all boilers
larger than 73 MW which have begun operation since 1976. This assumes a
five year lag time between NSPS proposal date and boiler start-up date.
For coal-fired utility boilers, a 0.6 Ib/million Btu NSPS NO level was
A
proposed in 1978. Assuming the five year lag time, boilers coming on-line
after 1983 should be meeting the 0.6 Ib/million Btu standard.
The NO control efficiencies listed in Table 3-13 for internal fuel
X
combustion and petroleum industry sources under NSPS represent levels
that new sources can attain with present control technology rather than
proposed NSPS levels. Therefore, in a nationwide analysis, it should not
be assumed that new sources in these categories will control to the NSPS
levels shown in Table 3-13 unless NSPS's for NO are proposed.
3.5 SOURCE CONTRIBUTION FACTORS
As noted in the formulation of modified rollback (Equation 2-5), a
source contribution factor (S±) must be specified for each source category,
i.e., mobile, stationary area and stationary point sources. This factor
3-40
-------
is to account, in a very simplistic manner, for the effect of emission
height and source-monitor separation on the contribution of that source
category to the measured ambient concentrations. A factor of 1.00
implies that emissions are at or near ground level and near the monitor,
and that they will have the greatest impact on air quality. Factors
less than 1.00 imply that pollutants are emitted above ground level or
at a considerable distance from the monitor, and because of a better
opportunity for dilution, the emissions have a lesser impact on air
quality than nearby ground level emissions. In a national analysis,
these factors must represent some universal weighting of emissions for a
specific category averaged over all sites. If sufficient data and
resources were available, then the use of a more sophisticated disper-
sion model would provide these weighting factors directly. Of course,
such detailed analyses are not usually within the scope of a nationwide
analysis. Therefore, estimates of the source contribution factors must
be based on extrapolations from previous site specific studies and
engineering j udgement.
3.5.1 Ozone
All sources of volatile organic compounds (ozone precursors) are generally
considered to have a source contribution factor of 1.0 regardless of
their distance from the design value monitor or of the height of emission.
This assumption is made because ozone is a secondary pollutant requiring
some time for formation in the ambient air, and is generally characterized
as leading to a pervasive, as opposed to a "hot spot," air quality
problem. The complexity of the chemical transformations (NMOC/NO to
A
ozone), and the large geographic extent of ozone nonattainment areas,
suggests that all VOC emissions should be considered collectively. In
addition, height of release is not considered because ozone levels are
assumed to be more dependent on total atmospheric loadings than on
initial dilution, at the point of emission. Recall that ozone is formed
only after photochemical reactions with NMOC and NO have taken place,
A
which is generally after the plume has been well mixed in the atmosphere
and its initial height of release has become inconsequential.
3-41
-------
3.5.2 Nitrogen Dioxide
Mobile and stationary area NO sources have traditionally been assigned
A
source contribution factors of 1.0. The rationale for this selection
follows from similar emission heights of these source categories and the
annual averaging time imposed by the current NAAQS. The effect of the
latter point is to "wash out" the effects of the spatial distribution of
these sources. Therefore, no support for using a source contribution
factor other than 1.0 is evident, provided that an annual average NAAQS
is under consideration. Conversely, it is unclear what impact stationary
point sources, which have been shown (see e.g., Chang, et al., 1980a) to
make significant air quality contributions on a short term basis, have
on NO- monitors in the existing network. However, none of the existing
sites appears to have any significant impact from stationary point
sources; they are dominated by mobile and stationary area sources. It
follows then, that unless new information becomes available or new
monitor sites are selected that are dominated by stationary point sources,
that a source contribution factor of 0.0 should be used for point sources.
3.5.3 Carbon Monoxide
The selection process for source contribution factors for CO follows,
very closely, that of NO . Again, the CO regulatory analysis has shown
X
that, while stationary point sources can be shown to contribute to
significant ambient concentrations, none of the CO monitors that cur-
rently indicate nonattainment are significantly impacted by stationary
point sources. Hence, a weighting factor of 0.0 is assumed for stationary
point sources. Note that the above discussion may not apply to all
monitor sites nor to new sites.
3-42
-------
Because of the availability of the SRI work mentioned in Section 2.2A.I,
an assessment of the relative impact of mobile and stationary CO area
sources can be made. Recall that SRI examined, in detail, monitors in
four areas: San Jose, Seattle, Phoenix, and Chicago.
Monitoring data in San Jose show a correlation between traffic peaks and
CO levels. This is illustrated in Figure 3-3 for one of the monitors
closest to the intersection being studied. There is a sharp increase in
CO concentration starting at 7 a.m. Concentrations drop rapidly after.
9 a.m. and remain relatively constant through mid-day until they increase
again around 3 p.m. Peak evening values occur at 8 p.m. The two sites
most frequently upwind of the nearby roadways show the least evidence of
the diurnal cycles. This is evidence that local contributions to CO are
important at the downwind sites. Figure 3-4 shows the diurnal pattern of
"local" CO contributions averaged across all monitors at the San Jose
intersection. The local contribution is determined by subtracting the
background concentration for that hour from the maximum observed concen-
trations. The local contribution to ambient CO levels during periods
when the CO concentration was above 9 ppm (the 8-hour CO NAAQS) ranged
from 62 to 98 percent and averaged 80 percent.
In Seattle, samplers were deployed on a oneway street in the Central
Business District (CBD). Because the street was one-way, the bimodal
traffic patterns observed in San Jose were not evident in Seattle. As
can be seen from Figure 3-5, CO concentrations rise steadily through the
day and peak between 5 and 7 p.m. Apparently, the street where the
monitors are located is more heavily traveled outbound during the evening
rush hour period than it is during the morning. Figure 3-6 shows the
estimated percentage of locally generated CO at the Seattle site. As
in San Jose, the maximum local contribution is determined by subtracting
a background concentration for that hour from the maximum observed
concentration. Although the local CO contribution usually ranges between
3-43
-------
FIGURE 3-3
DIURNAL VARIATIONS IN CO CONCENTRATIONS OBSERVED BY
HOUR OF THE DAY AT A SAN JOSE, CA ROADWAY MONITOR*
MAXIMUM AND MEAN ONE-HOUR AVERAGE MEASURED DECEMBER 8-15,1977
CO (ppm)
30
20-
15-
10-
5-
MAXIMUM
10
I
15
20
25
HOUR OF DAY
Hourly values are averages of measurements during the one week monitoring period.
3-44
-------
FIGURE 3-4
DIURNAL VARIATIONS IN CO CONCENTRATIONS DUE TO
LOCALLY GENERATED EMISSIONS AT A SAN JOSE, CA INTERSECTION
PERCENTAGE OF OBSERVED 8-HOUR (Running) AVERAGE MEASURED DECEMBER 8-15,1977
AVERAGE PERCENTAGE OF
LOCALLY GENERATED CO
100
80-
60-
40-
20-
10
i
IS
20
25
HOUR OF DAY
Hourly values are averages of measurements during the one week monitoring period.
3-45
-------
FIGURE 3-5
DIURNAL VARIATIONS IN CO CONCENTRATIONS DUE TO LOCALLY
GENERATED EMISSIONS IN DOWNTOWN SEATTLE, WA *
PERCENT OF OBSERVED ONE-HOUR AVERAGE MEASURED JANUARY 11-18, 1978
AVERAGE PERCENTAGE OF
LOCALLY GENERATED CO
100
80-
50-
40-
20-
10
15
I
20
25
HOUR OF DAY
* Hourly values are averages of measurements during the one week monitoring period.
3-46
-------
FIGURE 3-6
HOURLY VARIATIONS IN CO CONCENTRATIONS DUE TO LOCALLY
GENERATED EMISSIONS IN DOWNTOWN SEATTLE
PERCENTAGE OF OBSERVED 8-HOUR (Running) AVERAGE MEASURED JANUARY 11-18,1978
PERCENTAGE OF LOCALLY
GENERATED CO
100
80-
60-
40-
20-
20 40 60 80 100 120 140 160 180 200
NUMBER OF HOURS
3-47
-------
60 and 80 percent, at some hours in Seattle it is much lower. The
"regional" contribution at times is as high as 70 percent. Therefore,
the CO problem at these Seattle sites cannot be considered strictly
"local" in nature.
CO concentration patterns in Phoenix were much different than those in
the previous two cities. Although the study site in Phoenix was near a
complex of government office buildings in the downtown area with dense
traffic likely from 7 to 9 a.m. and 4 to 6 p.m., the highest CO concen-
trations were observed between 10 p.m. and 2 a.m.; a period with low
traffic volume. Because high CO was observed at numerous monitors
during this period, it was theorized that a recirculated air mass caused
these readings. However, based on discussions with the meteorologist
for the State of Arizona Bureau of Air Quality Control, SRI's conclusions
about how these high levels occur may be somewhat misleading. In the
winter months when high CO concentrations are most likely to be observed
in Phoenix, the following pattern is predominant:
1. High CO emissions during 6 to 9 a.m. traffic peak.
2. From 10 a.m. on, upslope (westerly) winds predominate. There is
intense surface heating which causes instability and promotes good
vertical mixing. Therefore, no CO buildup is likely during the
day.
3. High CO emissions during 4 to 6 p.m. traffic peak.
4. At approximately 6 p.m. an inversion forms. Winds continue from
the west.
5. Sometime between 10 p.m. and midnight the wind direction changes,
and downslope winds bring post 6 p.m. emissions back into the urban
area causing high CO readings.
6. The inversion breaks up by 10 a.m. the next day.
Given this weather pattern, then, locally generated CO emissions
end up producing the high observed values. The difference is that the
3-48
-------
highest CO emission periods (morning and evening rush hours) are not the
contributors to the highest ambient CO. Because CO emitted before
6 p.m. is not trapped by the inversion, it is unlikely to be carried
back into the city late at night. Therefore, 6 p.m. to midnight emissions
are the most directly related to high CO.
In a city with this type of situation, rollback is only applicable if
the mix of sources contributing to the problem is the same as that in
the emission inventory being used. If an air mass leaves and then
returns to the downtown area, it is likely to be well mixed (within the
inversion layer). Therefore, it is more likely in a city like Phoenix
that non-mobile area sources or point sources can contribute to CO
concentrations at the design value monitor if there is any transport.
The monitor locations in downtown Chicago were representative of a
"street canyon" situation, with nearby buildings all taller than ten
stories. Figure 3-7 shows that while the local CO contribution is
typically between 60 to 80 percent, it can be as low as 20 percent.
Diurnal variations in CO concentrations were more evident at the "local"
monitoring sites than at the "background" sites. Therefore, it can be
concluded that to some extent these monitors reflect "local" changes in
emissions. Because there were only two eight-hour CO standard violations
during the Chicago study period, it is difficult to reach conclusions
based on such a small sample. However, it should be noted that the
"local" contribution to these two violations averaged 86 percent.
The monitoring data collected in Chicago are less representative of
typical conditions in that city than the data from the other three
cities. The CO concentrations observed in Chicago during the SRI study
period were much lower than normal for the same period in previous years
3-49
-------
FIGURE 3-7
HOURLY VARIATIONS IN CO CONCENTRATIONS DUE TO
LOCALLY GENERATED EMISSIONS AT A CHICAGO, IL SITE
PERCENTAGE OF OBSERVED 8-HR (Running) AVERAGE MEASURED FEBRUARY 24-MARCH 3,1978
PERCENTAGE OF LOCALLY
GENERATED CO
100-
60-
40-
20-
20
40
60
I
80
i
100
120
140
I
160
I
180 200
NUMBER OF HOURS
3-50
-------
as measured at a nearby SAROAD site. Violations of the 8-hour standard
occurred approximately one percent of the time at the nearest SAROAD
site in the first quarter of 1975. Concentrations at SRI's mobile
laboratory did not even exceed 3 ppm one percent of the time during the
test week. This difference is at least partially attributable to adverse
weather conditions (snow) which reduced traffic during the study.
Because stationary CO area sources constitute major contributors to CO
levels, they pose a perplexing problem. Based on emission height, they
should probably be weighted the same as mobile sources. However, because
of the short term averaging time of the NAAQS, averaging cannot be
easily invoked (as with N0?) to resolve spatial variability of sources.
Area sources of CO will not, in general, be as heavily concentrated
around the nonattaining monitors as mobile sources. Therefore, a weighting
factor less than 1.0, but larger than 0.0 should be used. Because
ratios of mobile to area source densities are not available around each
site, some judgement concerning the level of the factor must be made.
Traditionally, (see e.g., Chang, et al., 1980b) the area source contri-
bution factor has been chosen to be 0.2. However, the value selected
for this factor is usually not critical because stationary area souce CO
emissions consitute only about 7% of the national total mobile and
stationary area source inventory (EPA, 1980). Thus, an area source
contribution factor of 0.20 can be used for national air quality assessments.
3.6 TRANSPORTATION CONTROL FACTORS
These factors are used to show the extent to which transportation control
measures are expected to reduce mobile source emissions. They should
only be used to account for measures designed to reduce VMT. Emission
reductions due to the Federal Motor Vehicle Emission Control Program
(FMVECP) and inspection and maintenance are accounted for in the mobile
source control factors.
3-51
-------
Transportation control measures (TCM's) are designed to change current
travel patterns via either pricing or service incentives, or both. They
are usually projects which involve a small investment of capital and
attempt to induce immediate changes in travel behavior. Major capital
projects like new freeways and subway systems which affect long-term
travel are not TCM's. Through TCM's, an automobile driver may be
induced to:
Eliminate a trip
Reduce the length of a trip
"Double up" by carpooling
Use transit.
Determining potential credits for TCM's for use in a nationwide analysis
is difficult because there may be considerable variation in the trans-
portation control plans among regions. Therefore, TCM's in one region
may be designed to reduce emissions by only a few percent, while another
region with a more extensive set of TCM's may expect their program to
reduce emissions by 10 percent or more. Because of this variability, it
is recommended that any projection analysis be run using more than one
assumption about TCM credits to test the sensitivity of the results to
changes in this variable. It is unlikely that TCM's will reduce mobile
source emissions more than 10 percent, however.
Because transportation control plans are being continually changed
through the SIP process, future nationwide analyses should take these
changes into account. To the extent that transportation control plans
are the same in different regions, a point estimate of TCM effectiveness
can be made. Otherwise, a range of estimates should be used.
Determinations of credits for TCM's should be based on the most recent guidance
from EPA's Office of Transportation and Land Use Policy (OTLUP).
3-52
-------
REFERENCES FOR SECTION 3
Bureau of Economic Analysis. October 1977. "Population, Personal
Income, and Earnings by State - Projections to 2000." U.S. Department
of Commerce. Washington, B.C.
Chang, T.Y., Norbeck, J.M., and Weinstock, B. February 1980a, "NO^
Air Quality-Precursor Relationship: An Ambient Air Quality Evaluation in
the Los Angeles Basin," J. Air Poll. Control Assoc. 30:157.
Chang, T.Y., Norbeck, J.M., and Weinstock, B. January 1980b, "Urban-
Center CO Air Quality Projections," J. Air Poll. Control Assoc. 30:1022.
Congress of the United States. February 1979. Changes in the Future
Use and Characteristics of the Automobile Transportation System. Volume II:
Technical Report. Office of Technology Assessment. Washington, B.C.
Decker, C.E., Worth, J.J.B., Ripperton, L.A., and Bach, W.D. January
1977. Ambient Monitoring Aloft of Ozone and Precursors Near and Down-
wind of St. Louis. EPA-450/3-77-009. Research Triangle Institute, RTP,
NC. (Prepared for U.S. EPA, RTP, NC.)
Eaton, W.C., Decker, C.E., Tommerdahl, J.B., and Dimmock, F.E. April
1979. Study of the Nature of Ozone, Oxides of Nitrogen, and Nonmethane
Hydrocarbons in Tulsa, Oklahoma. Volume II. Data Analysis and Inter-
pretation. EPA-450/4-79-008c. Research Triangle Institute, Research
Triangle Park, NC. (Prepared for U.S. EPA, RTP, NC.)
McNutt, B., Dulla, R., and Lax, D. March 1979. "Factors Influencing
Automotive Fuel Demand." SAE Technical Paper Series. No. 790226.
Congress and Exposition. Detroit, MI.
PEDCo Environmental, Inc. February 1979. Control Techniques and Costs
for Carbon Monoxide Emissions. Interim Report No. 1. Cincinnati, OH.
Peterson, P.R. and Sakaida, R.R. December 1978. Summary of Group I Control
Technique Guideline Documents for Control of Volatile Organic Emissions from
Existing Stationary Sources. EPA-450/3-78-120. Pacific Environmental
Services, Santa Monica, CA. (Prepared for U.S. EPA, Research Triangle Park,
NC)
Trijonis, J. and Hunsaker, D. February 1978. Verification of the
Isopleth Method for Relating Photochemical Oxidant to Precursors.
EPA-600/3-78-019. Technology Service Corporation. Santa Monica, CA.
(Prepared for U.S. EPA, RTP, NC.)
J-53
-------
U.S. Department of Energy. October 1979. Industrial Sector Technology
Use Model, Industrial Energy Use in the United States. 1979-2000.
Volume I -Primary Model Documentation. DOE/FE/2344-1. Washington, D.C.
U.S. EPA. December 1977. Procedures for the Preparation of Emission
Inventories for Volatile Organic Compounds, Volume I. EPA-450/2-77-028.
RTP, NC.
U.S. EPA. January 1979a. Guideline for the Interpretation of Ozone Air
Quality Standards. EPA-450/4-79-003. OAQPS No. 1.2-108. RTP, NC.
U.S. EPA. December 1979b. Regulatory Analysis and Environmental Impact
of Final Emission Regulations for 1984 and Later Model Year Heavy Duty
Engines. Office of Mobile Source Air Pollution Control. Ann Arbor, MI.
U.S. EPA. March 1980. 1977 National Emissions Report. EPA-450/4-80-005.
RTP, NC.
U.S. EPA. January 1978a. Control Techniques for Nitrogen Oxides Emmisions
from Stationary Sources-Second Edition. EPA-450/1-78-001. Research Triangle
Park, NC. (Prepared by Acurex Corporation, Mountain View, CA) .
U.S. EPA. August 1978b. User's Guide to MOBILE1; Mobile Source Emissions
Model. EPA-400/9-78-007. Washington, D.C.
U.S. EPA. 1980. User's Guide to MOBILE2; Mobile Source Emissions Model.
(In press) Ann Arbor, MI.
Wohl, M. and Martin, B.V. 1967. Traffic System Analysis for Engineers
and Planners. McGraw-Hill Book Company. New York, NY.
3-54
-------
4. STRATEGY EVALUATION CRITERIA
Indicators like percentage change in concentrations, number of exceedances
of the NAAQS, and number of regions experiencing NAAQS violations are
typically used in nationwide studies to compare the effectiveness of
alternative motor vehicle emission standards or other control programs.
The usefulness of each of these indicators in predicting differences
among control scenarios is discussed in this section. In addition, some
less frequently used indicators such as concentration frequency distri-
butions, populations at risk, and number of unhealthful days will be
presented and critiqued. The indicators discussed in this section are
important because they are the criteria by which alternative scenarios
are judged.
4.1 PERCENTAGE CHANGE IN POLLUTANT CONCENTRATION
Calculating the percentage change in air quality between the base year
and the projection year is accomplished using equation (4-1):
Projection Year A.Q. - Base Year A.Q..Ul_n<,. _ . _, // T\
J Base Ye;r A.Q. *-100% = Percentage Change (4-1)
This computation is made for each area in the nationwide sample and then
the specific area values are averaged to determine a nationwide average.
This measure essentially provides a weighted average of the effective-
ness of emission controls. Emissions are weighted by the source contribution
factors. The average percentage change in air quality equals the average
percentage change in emissions if the source contribution factors are
all equal to one, and there is no background.
4-1
-------
This indicator is useful if the purpose of the analysis is to compare
the effectiveness of different control strategies like different auto
emission standards. Then, knowing the weighted average effectiveness of
the controls vis-a-vis other source types is important. It is a good
indicator to use in a nationwide analysis where a sample of regions is
chosen to be representative of the entire U.S. However, for a regulatory
analysis where attaining/not attaining a standard is of most interest,
its usefulness is limited.
The significance of a percentage decrease in pollutant concentration
depends on the base year ambient levels and their magnitude in relation
to the NAAQS. For example, a two percent reduction in ambient levels
may be sufficient to bring a region that is already close to the stan-
dard into compliance. However, in a region with pollutant levels much
higher than the standard, two percent would be a totally inadequate
reduction.
4.2 NUMBER OF EXCEEDANCES OF THE NAAQS
More than one method can be used to project the number of exceedances of
the NAAQS. The basic procedure is to choose a frequency distribution
(e.g., exponential) which has been shown to be representative of the
pollutant in question and use this distribution to estimate the number
of violations from the projected concentration for a given sampling
time. The most straightforward manner to apply this technique is to use
a single distribution for all areas. The more complex alternative is to
project a separate distribution for each region.
An example where a distribution was used to predict the number of NAAQS
violations is in a past analysis of alternative automotive emission
standards. (U.S. EPA, 1979).
4-2
-------
In this analysis, violations of the ambient CL standard were estimated using
an exponential distribution for CL concentrations. The exponential
distribution had the form:
F(C) = 1 - exp (-C/C) (4-2)
where:
F(C) = the fraction of values less than or equal to the
concentration C
C = the mean of the observed concentrations
Hourly 0_ data were used to determine the design value and the mean for
the base year. These values were then projected to future years using
the rollback model, and used in equation (4-2) to estimate the number of
hours that the standards would be violated.
Other researchers have confirmed the use of the exponential distribution
as a good descriptor of the upper tail of most ambient concentration
distributions (Breiman, Gins and Stone, 1978) (Curran and Frank, 1975).
These references are particularly useful because they concentrate on
predicting peak air pollution concentrations. In addition, it should be
noted that the one-parameter exponential distribution is readily adapted
to the rollback formulation where only the peak concentration value is
available.
Because this measure compares ambient values to the NAAQS, it provides
some information on the potential health risk of exposures to high
pollutant levels. Presumably, the greater the number of violations, the
greater the health risk. However, there are some uncertainties involved
in estimating the number of ambient standard violations. One is that
the distribution selected as representative of the data in the base year
is accurate. Another is whether the distribution of air quality values
in the base year will be the same in future years. Because control
4-3
-------
strategies may be focused on the highest recorded level, the distribu-
tions may typically change over time. However, if the sample of cities
used in an analysis is large enough, and the errors in estimating the
number of exceedances are randomly distributed between cities, the
projected nationwide number of violations should be a valuable indicator
for comparing alternative control options.
4.3 NUMBER OF REGIONS WITH AMBIENT LEVELS GREATER THAN THE STANDARD
This is a straightforward number to calculate, but an important measure
to use in comparing control strategies. It is especially important in
determining the potential control costs to achieve a standard and may be
used as a surrogate for exposure. In the recent regulatory analyses,
this indicator was extremely important because a change in the ambient
standards directly affects costs to industries and consumers in areas
which must impose controls to meet those standards. The steps in the
regulatory analyses were:
Use base year monitoring data to determine design values
Project design values to the future year of interest
Compare projection year design values to ambient standards
For areas with values greater than the standard, impose
controls to bring them into compliance
Calculate the cost of meeting the ambient standard.
Therefore, estimating whether a region will or will not comply with an
ambient standard is crucial in evaluating needed nationwide control
strategies and their associated costs. While this is a useful measure
in a nationwide analysis, the use of rollback or the standard EKMA
isopleths to predict attainment or nonattainment by any single area is
inappropriate. Control strategies in specific areas may be far different
than those assumed as a nationwide average.
4-4
-------
4.4 OTHER STRATEGY EVALUATION CRITERIA
The approach used for evaluating the effectiveness of expected emission
reductions in the next 10-20 years by the Council on Environmental Quality
(CEQ, 1976) in its annual reports is somewhat different than that typically
used in projection analyses. Instead of using a design value to judge
attainment or nonattainment of a standard, the CEQ projects the distribution
of concentrations. This distribution shows both absolute concentration
intervals and the number of days in a year with maximum concentrations
in those intervals. This technique assumes a 1:1 relationship between
emissions and air quality reductions. Essentially, expected reductions
in emissions are accounted for by shifting the distribution downward
according to the percentage reduction in emissions assumed.
Another CEQ indicator for measuring air quality trends is the number of
unhealthful days. The 1979 CEQ Annual Report uses the Pollutant Standards
Index (PSI) for each city to determine the number of days with unhealthful
air. In this CEQ report, they use PSI data from 1975-1977 to rank SMSA's
by the number of "unhealthful" and "very unhealthful" days. For each
county with monitoring data, CEQ multiplies the county population by the
number of days on which the levels of each of the criteria pollutants
exceeded the primary NAAQS. This yields the number of person-days of
exposure in the county for the year.
The chief advantage of these techniques is that they present more than
just the expected change in the design value for a region. However,
when considering them for use in a nationwide analysis, this advantage
is outweighed by the amount of effort to assemble the data for each
region.
4.5 CONCLUSION
The current version of the modified rollback computer program used by
EPA uses three summary statistics as indicators of the nationwide dif-
4-5
-------
ferences in control options. They are the projected average percentage
change in air quality, the number of cities with concentrations above
the standard, and the total number of violations. Each one of these
indicators is useful for comparing control options, but some indicators
should receive more emphasis than others depending on the type of analy-
sis being performed. For example, if different motor vehicle emission
standards are being compared, the average percentage change in air
quality is probably the most useful. Alternatively, if a regulatory
analysis is being performed, the number of cities above the standard is
the most important indicator because costs for controls are only incurred
in areas that violate the standard. In\ both cases, the total number of
violations indicator provides useful additional information about the
extent of high concentrations in the projection year.
In the CEQ reports, the indicators provide some additional information
when compared with the current rollback output, but with the exception
of including county population estimates to the data base, they are only
marginally useful and would be extremely time consumptive to include in
a rollback format.
4-6
-------
REFERENCES FOR SECTION 4
Breiman, L., Gins, F. and Stone, C. November 1978. Statistical Analysis
and Interpretation of Peak Pollution Measurements. Technology Service
Corporation, Santa Monica, CA (Prepared for U.S. EPA, Research Triangle
Park, N.C.)
Council on Environmental Quality- September 1976. Environmental
Quality - 1976. The Seventh Annual Report of the Council on Environ-
mental Quality. Washington, D.C.
Council on Environmental Quality. December 1979. Environmental Quality -
1979. The Tenth Annual Report of the Council on Environmental Quality.
Washington, D.C.
Curran, T.C. and Frank, N.H. 1975. "Assessing the Validity of the
Lognormal Model when Predicting Maximum Air Pollution Concentrations"
Paper No. 75-51.3, 68th Annual Meeting of the Air Pollution Control
Association, Boston, MA.
U.S. EPA. November 1979. "Data Assumptions and Methodology for Assessing
the Air Quality Impact of Proposed Emission Standards for Heavy-Duty
Vehicles." OAQPS, RTP, NC.
4-7
-------
5. SUMMARY AND CONCLUSIONS
5.1 APPLICABILITY OF ROLLBACK AND EKMA
This study recommends methodologies appropriate for evaluating the
nationwide impact of emission control strategies on ambient levels of
CO, NO and 0_. These methodologies are not recommended for SIP analyses
*» o
of specific urban areas. They are particularly appropriate for evaluating
mobile source control strategies. Therefore, they apply to the mobile
source related pollutants. These methods use a sample of regions with
ambient data indicating a standards violation or the potential to exceed
the standard in the analysis. While emissions and air quality data
specific to each area are used, other assumptions about control effec-
tiveness, growth in new sources, retirement of existing sources, and
source contributions are made on a nationwide average basis.
CO and NO air quality projections are performed using the Modified
Rollback Model (de Nevers and Morris, 1975) to relate pollutant emis-
sions and ambient concentrations. This model assumes that the ambient
concentration of a pollutant is a linear function of pollutant emissions
within a specified geographical area. Therefore, a change in CO or NO
X
emissions is assumed to result in a proportional change in CO or N02
concentrations.
Ozone air quality projections are made using the Empirical Kinetic
Modeling Approach (EKMA) (U.S. EPA, 1977). EKMA utilizes a set of ozone
isopleths which depict maximum afternoon concentrations of ozone down-
wind from a city as a function of initial (i.e., morning) concentrations
of NMOC and NO , VOC and NO emissions occurring later in the day.
X X
meteorological conditions, reactivity of the precursor mix, and con-
centrations of ozone and precusors transported from upwind areas.
5-1
-------
However, in the version of EKMA typically used in nationwide studies,
NO levels are assumed to remain constant. Therefore, only reductions
x
in VOC emissions are assumed to affect 0 concentrations. The NMOC/NO
«J X
ratio assumed for each area affects the percentage reduction in 0
attributable to a given reduction in VOC emissions. Therefore, area
specific ratios should be used when available. A default value of 9.5:1
is used in areas with no monitoring data.
When comparing models like rollback and EKMA. with complex dispersion
models, the most important difference is in determining the source
contributions. The extent to which a reduction in emissions will
produce a reduced pollutant concentration for a given source cate-
gory depends on source-receptor distance, atmospheric stability, wind
speed and effective stack height. The source contribution factors used
in a rollback application are determined from monitoring and modeling
data. A source contribution factor is determined separately for each
source type which contributes significantly to measured concentrations.
Then, once the source contribution factors are chosen, all strategy
evaluations become rollback analyses.
The main advantage of using rollback is its computational and resource
economy. There are several limitations associated with rollback how-
ever. The most significant are:
1. Lack of model validation
2. The assumed linear relationship between emissions and air
quality. Deposition, agglomeration and chemical reaction
could invalidate the linear assumption.
3. The current base year concentration may not be known
based on coverage of the monitoring network.
4. The growth factor used assumes all other parameters
remain unchanged, i.e., all sources grow equally and
their spatial distribution remains constant.
5-2
-------
5. Simple and "modified" rollback assume that the meteoro-
logical conditions occurring when the design value was
measured are the same in the projection year.
6. Very simplistic treatment of source-receptor relationships.
7- Simple rollback assumes all sources emit at the same
height and "modified" rollback assumes sources within
certain categories emit at the same height.
8. Estimates can only be made for existing monitoring sites.
It should also be noted that the rollback formulation is inherently
designed to estimate changes in air quality concentrations based on
current air quality and assumed background levels resulting from emis-
sion changes. Therefore, it should be used to examine relative dif-
ferences between control strategies rather than absolute levels. Likewise,
EKMA isopleths are most properly interpreted when used in a relative
rather than an absolute sense.
5.2 SUGGESTED METHODOLOGIES FOR COMPILING MODELING DATA
5.2.1 Emission Inventories
The National Emissions Data System (NEDS) is recommended for use in
compiling emission inventories for nationwide studies. NEDS is the only
source which uses consistent inventorying procedures and compiles data
in a common set of source categories. NEDS inventories are typically
provided on a county or AQCR basis. After study of the characteristics
of the three mobile source related pollutants, it is recommended that
county inventories be used for CO and NCL analyses, and AQCR inventories
be used for 0 analyses. Although CO is predominately a microscale
problem, county inventories are the most detailed included in NEDS, and
it would be impossible to find a consistent inventory area smaller than
a county for which information would be available on a nationwide basis.
5-3
-------
The source category groupings recommended for use in a nationwide study
are shown in Tables 3-3, 3-4, and 3-5.
5.2.2 Design Values
The design value is defined as the concentration that must be reduced to
the level of the standard for an area to be in attainment. In using
ambient monitoring data in national strategy assessments, design values
are usually chosen to correspond to the form of the standard for each
pollutant and for a time period consistent with the emission inventory.
Design values for the three pollutants considered in this report are
usually chosen as follows:
Carbon Monoxide -
The highest second high eight-hour average in the three year
period encompassing the emission inventory base year.
\
Ozone -
The one-hour average measured in the three year period encom-
passing the emission inventory base year which is expected to
be exceeded an average of once per year (See Table 3-6).
Nitrogen Dioxide -
The highest annual average measured in the three year period
encompassing the emission inventory base year.
5.2.3 Growth and Retirement Rates
5.2.3.1 Stationary Sources
Future changes in industrial activity can be accounted for by a combina-
tion of growth and retirement rates. Growth and retirement are differen-
tiated because regulations affecting new sources differ from those
affecting existing sources. Therefore, different control assumptions
are identified for each. Growth rates and controls are applied to
estimate new source emissions. Retirement rates are applied to estimate
how emissions from existing sources will decrease.
5-4
-------
When applying the growth and retirement rates listed in Section 3, it
should be noted that these are national averages based on broad station-
ary source categories, which are judged to be sufficient for a national
scale analysis. They may not be representative of a particular urban
area.
5.2.3.2 Mobile Sources
Predicted growth in vehicle miles traveled (VMT) serves as a surrogate
for estimating the true growth rate in mobile source emissions. However,
growth in VMT is difficult to predict for two reasons. One is the
uncertainty in how changes in fuel prices will affect demand for gasoline
and travel. The second is identifying how VMT will change in the areas
that specifically affect high pollutant levels. Therefore, because of
the considerable uncertainty in predicting VMT over a 10-20 year period,
it is advisable to use a range of possible growth rates to judge the
sensitivity of the modeling results to changes in this factor.
It should also be noted that the growth rates in Section 3 may differ
from the area specific VMT growth rates used in state and local analyses.
5.2.4 Control Effectiveness
The effectiveness of future controls for mobile sources can be estimated
using EPA's mobile source emission factor program (MOBILE2). This
program accounts for both expected future emission levels by model year
by vehicle type and the rate of retirement of existing vehicles from the
present fleet. MOBILE2 can also be used to evaluate the effectiveness
of I/M programs.
Estimating the effectiveness of stationary source controls for use in a
nationwide analysis is not as straightforward as it is for mobile source
controls because there are many more source types to deal with. Sta-
tionary source categories used in nationwide analyses are usually chosen
5-5
-------
so that source types with common growth rates and control efficiencies
are grouped together. Therefore, the control efficiency used in the
rollback or EKMA model is a weighted average based on control effi-
ciencies for each source type.
Estimates of control efficiencies for stationary sources of VOC detailed
in Section 3 are taken from the Control Techniques Guideline (CTG)
documents published to date. As new CTG documents are published, the
numbers in Table 3-12 can be updated. Control efficiencies for NO and
CO stationary sources can be updated through new editions of their
respective Control Techniques Documents (CTD's).
5.2.5 Source Contribution Factors
Source contribution factors are an attempt to account for the effect of
stack height and source-monitor separation on the contribution of each
source category to measured ambient concentrations. A factor of 1.0
implies that emissions from that source category are typically at or
near ground level and near the monitor measuring the design value.
Factors less than 1.0 imply that pollutants are emitted above ground
level or at a considerable distance from the monitor, and because of a
better opportunity for dilution, the emissions have a lesser impact on
air quality than nearby ground level emissions.
All ozone precursors are generally considered to have a source contri-
bution factor of 1.0. This assumption is made because ozone is a secondarily
formed pollutant and is generally characterized as leading to pervasive,
as opposed to "hot spot", air quality problems. For N0_, point sources
are not significant contributors to annual average concentrations measured
at urban monitors. Therefore, they are assigned a source contribution
factor of zero. Mobile and other area source emissions are assumed to
contribute 100 percent to ambient N0_. Similarly for CO, it is assumed
that point sources have zero contribution at design value monitors. A
5-6
-------
source contribution factor of 1.0 is recommended for mobile sources, and
a 0.2 factor is recommended for stationary area sources. The 0.2 source
contribution factor is based largely on the typical distance of CO area
sources from urban CO monitors in comparison with mobile sources.
5.2.6 Transportation Control Factors
These factors are used to show the extent to which Transportation Control
Measures (TCM's) are expected to reduce mobile source emissions. They
should only be used to account for measures designed to reduce VMT.
Emission reductions due to the Federal Motor Vehicle Control Program
(FMVCP) and I/M are accounted for in the mobile source control factors.
Determinations of credits for TCM's should be based on the most guidance
from EPA's Office of Transportation and Land Use Policy (OTLUP). Because
there may be considerable variation in the transportation control plans
among regions, it is advisable to model a range of TCM credits rather
than one point estimate.
5.3 STRATEGY EVALUATION CRITERIA
The current version of the modified rollback computer program used by
EPA uses three summary statistics as indicators of the nationwide differences
in control options. They are the projected change in air quality, the
number of cities with concentrations above the standard, and the total
number of violations. Each one of these indicators is useful for comparing
control options, but some indicators should receive more emphasis than
others depending on the type of analysis being performed. See Section 4
for more details about the interpretation of these factors.
5-7
-------
REFERENCES FOR SECTION 5
deNevers, N. and Morris, J.R. 1975. "Rollback Modeling:
Modified", J. Air Poll. Control Assoc. Volume 25, No. 9.
Basic and
U.S. Environmental Protection Agency. November 1977. Uses, Limitations
and Technical Basis of Procedures for Quantifying Relationships Between
Photochemical Oxidants and Precursors. EPA-450/2-77-021a. Research
Triangle Park, NC.
5-8
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA 450/4-80-026
|3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Methodology To Conduct Air Quality Assessments of
National Mobile Source Emission Control Strateqies
5. REPORT DATE
October, 1980
6. PERFORMING ORGANIZATION CODE
]7. AUTHOR(S)
James H. Wilson, Jr. ,
Mark A. Scruggs
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Enerqv and Environmental Analvsis, Inc.
Ill North 19th Street
Arlington, Virainia 22209
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3371
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Anency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer - Warren P. Freas
16. ABSTRACT
This report describes a methodology for conducting air quality
assessments of national mobile source emission control strategies
using the Modified Rollback Model and standard EKMA isopleths. Both
modified rollback and EKMA are simple models which do not require
extensive data bases. As such, they are most useful for estimating
the impact of an emission control strategy on air quality in nationwide
studies where a number of alternative control strategies must be
analyzed in a great many areas. Recommended methodologies and data
assumptions are consistent with those used in the regulatory impact
analyses for alternative national air standards for ozone and carbon
monoxide.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Mobile Sources
Air Oualitv Analvses
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
100
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
------- |