United States Air and Radiation EPA420-R-02-002
Environmental Protection January 2002
Agency M6.EXH.001
&EPA Determination of
Running Emissions as a
Function of Mileage for
1981-1993 Model Year
Light-Duty Cars and
Trucks
$5b Printed on Recycled
Paper
-------
EPA420-R-02-002
January 2002
of Running as a Function of
for
Light-Duty Trucks
WI6.EXH.001
Phil Enns
Ed Glover
Penny Carey
Assessment and Modeling Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
This technical report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using data that are currently available.
The purpose in the release of such reports is to facilitate the exchange of
technical information and to inform the public of technical developments which
may form the basis for a final EPA decision, position, or regulatory action.
-------
1.0 INTRODUCTION
The MOBILE6 emissions inventory model will allocate vehicle exhaust emissions
between engine start (start emissions) and travel (running emissions). This split allows the
separate characterization of start and running emissions for correction factors such as fuel
effects and ambient temperature. It also enables a more precise weighting of these two aspects
of exhaust emissions for particular situations such as morning commute, parking lot and freeway
driving. This document describes the methodology used to calculate the in-use deterioration of
running emissions and presents estimates for model year 1981-1993 light-duty cars and trucks
proposed for use in MOBILE6. The deterioration of start emissions is addressed in a separate
document.1
Section 2 describes the Federal Test Procedure (FTP) data sources and the model year
and technology groups used. Section 3 presents the methodology for calculating running
emissions from FTP bag data. It contains a basic overview of the FTP, defines all of the
applicable emission terms, and provides the calculations for determining the base unit of engine
running emissions. Section 4 describes models and results for the in-use deterioration of
running emissions as a function of mileage. Section 5 reports on high emitter correction factors
which are applied to the deterioration estimates. Section 6 displays the final results in tabular
form.
2.0 FTP DATA SOURCES USED
FTP datasets were used to determine in-use deterioration. The FTP-based emission
estimates were then adjusted by applying high emitter correction factors derived using Ohio
EVI240 data. This section describes the FTP data sources used. Three FTP data sources were
used: (1) the test results from the EPA laboratory in Ann Arbor, Michigan; (2) data received
from the American Automobile Manufacturers Association (AAMA) based on testing conducted
in Michigan and Arizona; and (3) American Petroleum Institute (API) data collected in Arizona.
Model years range from 1981 through 1993, and vehicles include both cars and trucks. Table
1 gives a breakdown for the light duty vehicle sample by vehicle type, model year, and
technology for the three datasets combined.
Most of the 1990 and later model year vehicle data were supplied by AAMA, while most
of the pre-1990 data came from EPA laboratory testing. The API sample is a relatively small
sample (99 cars and trucks). Its chief appeal is that the vehicles have generally higher mileage
readings (all over 100,000 miles) than the rest of the sample. There is a general trend from
carbureted and open loop technologies in early model years to fuel injection in more recent
Clover, E. and P. Carey, "Determination of Start Emissions as a Function of Mileage and
Soak Time for 1981-1993 Model Year Light-Duty Vehicles," Report No. M6.STE.003,
October, 1998.
-2-
-------
years. Port fuel injected vehicles dominate in 1990 and later model years. Although not
explicitly shown in the table, new catalyst technology was phased slowly into the fleet starting
in the mid 1980's.
For this analysis, cars and trucks were each classified into the following model
year/technology groups:
MY Group / Technology Type - Cars
1988-93 Port Fuel Injection (PFI)
1988-93 Throttle Body Injection (TBI)
1983-87 Fuel Injection (PFI plus TBI)
1986-93 Closed Loop Carbureted/Open Loop
1983-85 CL Carb/Open Loop
1981-82 FI (PFI plus TBI)
1981-82 CL Carb/Open Loop
MY Group / Technology Type - Trucks
1988-93 Port Fuel Injection (PFI)
1988-93 Throttle Body Injection (TBI)
1981-87 FI (PFI plus TBI)
1984-93 Closed Loop Carbureted/Open Loop
1981-83 CL Carb/Open Loop
These groupings were selected on the basis of changes in emission standards or the
development/refinement of new fuel metering or catalyst technologies. Because of the relatively
large amount of 1988-93 fuel inj ected data, this category was split into PFI technology and TBI
technology for both cars and trucks. This produces separate deterioration functions based on
these fuel delivery technologies and allows the modeling of the future penetration of PFI
technology into the in-use fleet.
3.0DETERMINATION OF RUNNING LA4 EMISSIONS
3.1 Overview of the Federal Test Procedure (FTP)
The Federal Test Procedure (FTP) is a test cycle which is used to certify new vehicles
to emission performance standards.2 The FTP consists of a cold start segment (Bag 1), a hot
stabilized segment (Bag 2), and a hot start segment (Bag 3). Initially, the vehicle is stored for
240 CFRPart 86, Subpart B, Section 86.144
-3-
-------
a minimum of 12 hours before testing to simulate a 12 hour overnight soak period. It is then
driven over the cold start segment, which lasts 505 seconds over a length of 3.59 miles, and the
emissions collected as Bag 1. Bag 2 emissions are then immediately collected from the hot
stabilized segment, which lasts 867 seconds over a length of 3.91 miles. After a 10 minute soak,
the 505 seconds of the start segment is repeated and the emissions are collected as Bag 3.
The FTP composite emission rate is a weighted combination of the three measured bags
designed to represent two trips. The first trip is a cold start after a 12 hour soak, and the other
is a hot start after a 10 minute soak. Each trip is a "LA4" cycle, which is a combination of the
505 cycle (either Bag 1 or Bag 3) and the Bag 2 cycle. In a typical FTP test, the Bag 2 is only
measured once and the results are used for both trips. Since the 505 cycle is 3.59 miles long and
the Bag 2 cycle is 3.91 miles long, each LA4 trip is 7.5 miles long. Based on findings about
driving activity from the original FTP study, the cold start trip is weighted 43% and the hot start
trip weighted 57%. Hence the fraction of vehicle miles traveled (VMT) in Bag 1 (containing
the cold start) is:
FTP Bag 1 VMT Weighting = 43%*(3.59 miles / 7.5 miles) = 0.206
Similarly, since 57% of trips involve a hot start, the VMT weighting for Bag 3 (containing the
hot start) is:
FTP Bag 3 VMT Weighting = 57%*(3.59 miles / 7.5 miles) = 0.273
The remaining VMT represents stabilized driving (Bag 2). Since it is used for both the cold
start and hot start trips, its VMT weighting is computed from both:
FTP Bag 2 VMT Weighting = (43% + 57%)*(3.91 miles / 7.5 miles) = 0.521
Thus, the standard VMT weighting of the bags reported in grams per mile (g/mi) for the full
FTP is:
FTP = (Bag 1*0.206) + (Bag 2*0.521) + (Bag 3*0.273)
where the fractions represent the proportion of vehicle miles traveled within the three modes
during the FTP trip in grams per mile.
3.2 Overview of the Hot Running 505 and Its Use
The FTP testing method outlined above does not allow the precise separation of start and
running emissions, since Bags 1 and 3 contain both start and running emissions. Bag 2 of the
FTP does not contain an engine start; however, the driving cycle used in the second bag is
significantly different from the cycle used for Bags 1 and 3. Thus, to estimate the amount of
-4-
-------
FTP emissions that can be allocated to engine start and running emissions, the concept of the
Hot Running 505 (HR505) must be introduced.
The HR505 refers to emissions measured from a driving test performed on the 505-
second cycle of FTP Bags 1 and 3 without an engine start.3 Appending the FIR505 cycle to a
standard three-bag FTP produces values that can be used to estimate the portions of Bags 1 and
3 attributable to start emissions following a 12 hour soak and start emissions following a 10
minute soak, respectively.
Since the HR505 has not historically been included in FTP test programs, a method of
estimating the FIR505 from FTP bag data was developed using data from a special test program.
Briefly, HR505 emissions were measured in a sample of 77 cars and trucks tested under EPA
contract. The results from this sample were used to develop a correlation between the FIR505
and FTP bag data. This correlation was then used to estimate HR505 results for the larger FTP
dataset used in this analysis.
3.3 Basic Running LA4 Emission Rate
The LA4 refers to a cycle comprised of the 505-second driving cycle used for Bags 1 and
3 of the FTP and the 867-second cycle of Bag 2. Running LA4 emissions are defined as
emissions from this 1372-second cycle with no engine start. For the MOBILE6 separation of
start and running emissions, the running LA4 represents the running portion. For a given three-
bag FTP, running LA4 emissions can be estimated using a VMT-weighted combination of the
FIR505 and the Bag 2 emissions (stabilized operation). This estimate contains all of the driving
behavior in the LA4 cycle, without engine starts. Mathematically, it is given by:
Running LA4
Emissions = (HR505*(0.206+0.273)) + (Bag 2 * 0.521)
(grams/mile)
where 0.206, 0.273, and 0.521 are the VMT weightings for Bags 1, 3, and 2, respectively.
Like the FTP, running LA4 emissions are measured in units of grams per mile. This estimate
is proposed for use in MOBILE6 as the basic exhaust emission rate from which all other running
exhaust emission estimates are derived.
Using the methods described in this section, all emissions measured using the FTP and
reported by bag can be allocated to start or running emissions before analysis. Average running
LA4 and FTP emission rate estimates for each model year are shown in Table 2 for the light-
3Brzezinski, D. and P. Enns, "The Determination of Hot Running Emissions from FTP
Bag Emissions", Report No. M6.STE.002, December, 1997.
-5-
-------
duty cars and trucks in the EPA-industry sample used in this study.
4.0 FTP-BASED MODELS OF RUNNING LA4 DETERIORATION
WITH MILEAGE
This section describes the methodology EPA used to estimate the deterioration of
running emissions. Deterioration of running emissions as a function of mileage was
examined using a number of linear and nonlinear models. The goal was to develop a
description of deterioration that is consistent with both the available test data and with
engineering judgment of past and likely future technologies.
In particular, for the model year/technology sub-fleets identified above, adequate
data are often absent in some part of the useful lifetime mileage range. Such data gaps raised
concerns when trying to fit a single functional form to a given data set, as it usually was
found that no simple description of deterioration adequately describes the full range. For
example, a fitted least squares regression often tends to overestimate emissions at low
mileage.
A number of linear and nonlinear models of deterioration were examined. The
chosen models represent a balance of simplicity and engineering judgment. They take the
general form of expressing emissions as a piecewise linear function of mileage. At low
mileage, emissions are assumed to equal the mean level estimated from those vehicles in the
dataset with less than 20,000 miles of accumulated driving. This level applies for mileages
ranging from zero up to the mean mileage for those vehicles. This approach was thought to
give the best prediction, since the vehicles tested at low mileage should not be subject to any
recruitment bias influence. The 20,000 mile cutoff is somewhat arbitrary, and was
developed in coordination with the FACA In-Use Deterioration Workgroup.
At higher mileage, emissions are modeled to deteriorate linearly. While nonlinear
models were investigated, they did not provide significant improvement over simpler linear
forms. Two linear functions are used in the final models: (1) a least squares regression
using all the data that is constrained to pass through the low mileage sample means; and (2)
a least squares regression using all the data, with no constraints. The unconstrained linear
model was chosen as the best representation of the data at higher mileages. The constrained
line was chosen to provide a transition, when needed, between the low mileage mean and
the high mileage unconstrained regression. The connection between these two lines is made
based on their relative positions for a given technology/model year class. With a few
exceptions, the following steps describe the calculation of this piecewise linear function.
1. For each of the model year/technology groups listed in Section 2.0, the mean CO,
HC, and NOX emissions, and the mean mileage were computed for all vehicles with
an odometer reading of less than 20,000 miles. This value is used to model the
group's low mileage emission level from zero miles up to the mileage determined
-6-
-------
in step 2. An exception occurs when the mean emissions for the entire sample is less
than the mean of the low mileage sub-sample, a case that is discussed below.
2. For each group, an (unconstrained) regression line was estimated for emissions
versus mileage.
a. If this line has positive slope and its intercept is less than the low mileage
mean emissions from (1) above, it defines estimated emissions beginning
at the mileage where it intersects the low mileage mean.
b. If the (unconstrained) regression has positive slope but the intercept is
greater than the low mileage mean, the constrained line defines emissions
from the mean of the low mileage sub-sample to the mileage at which it
intersects the unconstrained line. Beyond that mileage, emissions are
estimated by the unconstrained line. Thus, the constrained line links the
other lines for an intermediate range of mileages.
c. If the unconstrained regression has negative slope or the mean of the full
sample is less than low mileage mean, emissions for all mileages are set
equal to the mean emissions for the full sample. This assures that
negative deterioration cannot occur.
While these rules do not encompass all possible scenarios, they do cover all
situations arising with the FTP data on which this analysis is based. The majority of cases
are covered by option 2(a), giving a simple two-piece function. The three-piece function of
2(b) applies to several situations, usually with only a small slope change from the
constrained to unconstrained line. Finally, the simple horizontal deterioration line of option
2(c) is needed for the CO fits of the 1988-93 TBI cars and the NOx fits of the 1981-83
carbureted trucks. The underlying numerical estimates are listed in Table 3.
For the FTP data set, these rules appear to produce reasonable emission proj ections
in most cases. The two cases in which the full sample mean is less than the low mileage
mean are caused by a few low mileage outliers.
5.0 HIGH EMITTER CORRECTION FACTORS
Since the estimates of running emissions deterioration are based on FTP tests
obtained from public vehicle recruitment programs, there is some concern that low vehicle
recruitment acceptance rates (typically less than 25%) in these programs may introduce
recruitment bias. Whether such bias results in overestimation or underestimation of the true
emissions deterioration is a matter of debate. This section addresses this issue, describes the
methodology for adjusting emission factor estimates to account for bias, and presents the
-7-
-------
results.
Most of the 1990 and later model year vehicle data for this analysis were supplied
by the domestic automobile manufacturers (the AAMA dataset). The manufacturers have
expressed the opinion in FACA meetings and MOBILE6 workshops that owners of vehicles
experiencing problems would be more likely to respond to the manufacturers' recruitment
efforts, especially considering that repairs were included as an incentive to participate. The
AAMA dataset is also composed of vehicles tested when they were roughly 2-3 years of age,
when gross emitters should be few in number and any recruitment bias influence should be
minimal.
Most of the pre-1990 data were collected by EPA; the average age of the vehicles
was roughly 3-5 years at the time of testing. In this case, tampered vehicles or vehicles with
problems should be greater in number, but owners may be more reluctant to participate in
a program run by a regulatory agency, resulting in an underestimation of high emitters. The
California Air Resources Board (CARB) has tested this hypothesis by comparing estimates
from its CALEVIFAC emissions inventory model, which are based on surveillance programs
similar to those run by EPA, with emissions obtained from a California Pilot Project fleet
with a high (60%) vehicle capture rate. In general, the comparison showed that the modeled
estimates tend to underestimate emissions in older model year vehicles and slightly
overestimate the emissions of newer vehicles. CARB developed high emitter adjustment
factors (HECFs) for use in its EMFAC model to account for these discrepancies.
EPA developed high emitter correction factors using EVI240 data collected in
Dayton, Ohio during 1996-97. Like other inspection and maintenance (I/M) data, these form
a large sample of vehicles within their geographical region, and are considerably less subj ect
to sources of bias found in non-mandatory programs. The data and their translation to
running LA4 estimates are described in more detail in a separate document4.
Because of problems with the Ohio data odometer readings, the data were condensed
to their mean running LA4 values by age, which then were associated with the
corresponding region-specific mileage accumulations obtained from 1995 Nationwide
Personal Transportation Survey (NPTS) data. After smoothing these values in the manner
required for use in MOBILE6, these points were graphed with the emission rates fitted from
the FTP data as described in Section 4. For each pollutant and within each model
year/technology group, the difference between the Ohio mean and FTP-based fit was
computed. These values were regressed through the origin against mileage. (The line was
forced through the origin so that at zero miles the difference is zero.) Finally, the fitted
differences were added to the fitted FTP-based values to obtain corrected values.
4Enns, P., E. Glover, P. Carey and M. Sklar, "Analysis of Emissions Deterioration
Using Ohio and Wisconsin EVI240 Data," Report Number M6.EXH.002, October, 1998.
-------
In a few model year/technology groups, the Ohio adjustment is negative and, when
applied to the deterioration line, causes negative deterioration. For these cases, deterioration
is held equal to zero up to the mileage at which the adjusted emissions exceed the low
mileage constant level.
Figures 1 and 2 illustrate how the adjusted and original values compare for each
model year/technology group as a function of mileage for the car sample. Ninety-five
percent confidence bands for the unadjusted lines are drawn to help judge the impact of the
corrections. If the adjusted values fall inside these bands, it suggests that the Ohio EVI240
data agrees fairly closely with the FTP data, i.e., bias is not a large problem. Otherwise, the
recruitment bias is more serious. The graphs show varying levels of disagreement between
the two data sources. In these graphs, the mileage interval for a given set of lines
corresponds to the average mileages assigned in the NPTS survey to the model years for that
group of vehicles. For example, the 1990 to 1993 cars range in mileage from about 45,000
to 70,000. Thus, the graphs show line fits for those vehicles in that interval.
Figures 3 to 9 present emission estimates for each model year/technology group as
a function of mileage both with and without the high emitter correction factors for cars and
trucks. For MOBILE6, deterioration estimates with the high emitter corrections will be used.
6.0 RESULTS
Results for each vehicle type/model year/technology group are presented in Tables
3 and 4. Included are the slopes and intercepts of the constrained and unconstrained
regression lines, low mileage emissions and mileage intervals for each line segment. Table
3, described in Section 4, gives the unadjusted slopes. Applying the adjustment factors
effectively changes the line segment slopes. The high emitter correction factors and the
corresponding adjusted slopes are displayed in Table 4.
Shown below is a sample calculation of running emissions. It illustrates how the
model coefficients given in Table 4 are used.
Example: Calculate HC running emissions for a 1985 model year Fl-equipped car with:
a) 15,000 miles, b) 75,000 miles, and c) 125,000 miles.
From Table 4:
a) At 15,000 miles, mileage
-------
b) At 75,000 miles, first cornersecond corner, therefore:
Running (g/mile) = ZML + (First Slope * First Corner) +
(Second Slope) * (Second Corner - First Corner) +
(Third Slope) * (Mileage - Second Corner)
= 0.1479 + (0.0000*18.89) + (0.0078)*(81.38-18.89) +
(0.0059)*(125-81.38)
= 0.8927 g/mile
-10-
-------
7.0 RESPONSE TO COMMENTS FROM PEER AND STAKEHOLDER
REVIEW
(1) "MOBILE6 Determination of Running Emissions as a Function of Mileage for 1981
- 1993 Model Year Light-Duty Cars and Trucks", and (2) "Analysis of Emissions
Deterioration Using Ohio and Wisconsin IM240 Data"
Comment #37 (AAMA) June 5.1997
Comment: "EPA did not describe, however, how they intend to estimate
deterioration for current and future technology vehicles, particularly Tier I and LEV-
type vehicles with extended durability (100K) and onboard diagnostic controls."
EPA's Response: EPA 's estimates of future technology vehicle emission factors
can be found in documents M6.EXH. 00 7 and M6.EXH. 009. The first document
discusses the HC andNOx emission factors and the second paper discusses the CO
emission factors.
Comment: "AAMA differs significantly with the EPA on its approach to estimate
in-use exhaust emissions and deterioration. EPA should not base in-use emission
rates used in MOBILE6 on I/M240 results, for the following reasons:
1. EPA has not demonstrated sufficient correlation between the I/M240 and
FTP.
2. EPA has little control over vehicle test fuel, preconditioning, or temperature
at which I/M240 tests are conducted. The methods used by EPA in
MOBILES to correct I/M240 results for temperature and fuel effects are
questionable.
3. The use of the I/M240 results along with dubious correlation equations and
correction factors for fuels and temperature very likely result in EPA arriving
at emission rates for current and future vehicles that are significantly
different than the California Air Resources Board's (ARB) emission rates for
exactly the same vehicles. If this is the case, then EPA must explain why
these emission rates are so different than ARB's emission rates. If EPA is
convinced that the I/M240 must be used, then EPA should convince ARB to
estimate emission rates in the same way, so that some consistency in
emission rates on identical vehicles is achieved."
-11-
-------
EPA's Response:
1. Emission results from the IM240 exhaust emission test procedure were not
used directly to estimate the emissions of vehicles using the FTP. The
IM240 was usedprimarily to assure that recruitment bias in the FTP testing
did not affect the overall emission deterioration estimate. Adjustments were
made to the IM240 results in order to approximate FTP emission levels,
however, these emission levels are never used to estimate FTP emission
levels for MOB1LE6 directly.
2. The adjustments to IM240 results for preconditioning, fuel and temperature
effects are applied only as an attempt to reduce, not eliminate, the effect of
these parameters. EPA feels that the application of these adjustments
improved the credibility of the overall analysis.
3. EPA does feel that the emission estimates for similar future vehicles in
California and in the federal fleet should be comparable. However, since
the specific rules related to new vehicle certification federally and in
California are not identical, there is room for disagreement on the emission
impact of the new rules. EPA and California are sharing data and methods
so that EPA is confident that, once all factors have been considered by both
groups, that the differences between EPA 's and California's estimates of
emissions from future vehicles will be negligible.
Comment:: "EPA should not base FTP emission rates on fast-pass I/M240 data, for
the same reasons as above. AAMA believes there will be a weak correlation
between fast-pass data and full I/M 240 data, and a weak correlation between full
I/M240 data and the FTP. FTP values developed from fast-pass data and these two
weak correlations will be subject to a high degree of error."
EPA's Response: EPA admits that the relationship between the fast pass IM240
test and the FTP test is not well characterized in the work and may possess a low
level of correlation. This would be particularly true for high emitting vehicles or
vehicles which possess intermittent emission problems. Fortunately, in the case of
high emitting vehicles (Ohio failures), no fast pass results were used since all
failures automatically received a full IM240 test.
Comment: "The automakers submitted extensive in-use FTP data to the EPA a year
ago which show that emission rates of 1990-1994 vehicles are significantly less than
MOBILES estimates, furthermore the data do not show any accelerated deterioration
after 50,000 miles as EPA now assumes. The emission rates developed from this
-12-
-------
data were similar to ARB's emission rates for the same vehicles. Data were
submitted on over 2,000 cars and 900 light duty trucks. This data was not mentioned
by EPA in the Workshop. This data significantly adds to the EPA data on the same
model years, and should give EPA a much larger database to analyze for MOBILE6.
AAMA urges that the MOBILE6 emission rates should be based on all available
FTP testing of vehicles by various sources, including EPA. ARE, and the Industry."
EPA 's Response: The EPA acknowledges the receipt of these data and their use in
characterizing in-use emission behavior. The FTP test data provide by these
various sources became the backbone of the MOBILE6 emission factors. It was
used almost exclusively to characterize the normal and high emitter emission levels.
It was also the basis for the average emission calculation presented in reports
M6.EXH.001 andM6.STE.002.
Comment: "100K and Useful Life Effects: For purposes of certification, under
Section 206 (Clean Air Act), emission standards were established for useful life
extended to 10 years or 100.000 miles. EPA needs to explicitly incorporate the effect
of this added requirement in the MOBILE model as this was not done in
MOBILES."
EPA's Response: The MOBILE6 1981-1993 model year emission factors were
empirically developed from the FTP and Ohio datasets, and no assumptions were
added to account for extended useful life. However, the 1994 and later vehicle
modeling was less empirical in nature, and contains factors that account for the
extension of the useful life from 50,000 miles to 100,000 miles.
Comment #56 (J.F. KOWALCZYK. State of Oregon) Dec. 19.1997
Comment:: "It is recommended that EPA go through the precise calculations to
make the adjustment to mobile 6 based on Ohio I/M data and 1995 NPTS data as
outlined above. Additionally, EPA should use the 1995 NPTS mileage
factors in the final mobile 6 model which may necessitate speeding up
the Acurex work. The 1995 NPTS data is clearly more up-to-date and most
likely more accurate than the 1990 data. Regional specific NPTS mileage
data should be used as there appears to be significant differences
between regions.
I also recommend that CARB's revised EMFAC model, which according to
Mark Carlock should be available in January, be analyzed against
mobile 6 and that full harmonization between the two models be sought
or, at the very least, a believable explanation of the difference be
-13-
-------
provided."
EPA's Response: The final analysis byAcurex usedinMOBILE6 included the 1995
Nationwide Personal Transportation Survey (NPTS) database. As suggested, the
regional NPTS results were used for estimating mileage accumulations in the
analysis of the Ohio I/M data.
We have attempted to coordinate with California as much as possible when
constructing MOB1LE6, including sharing of data and analysis. We have met
repeatedly with California Air Resource staff during the development ofMOBlLE6
to share ideas. Some aspects of the California fleet, however, are unique to
California and will not be identical in the federal (non-California) fleet.
Comment #58 (Data Analysis team, In-Use Deterioration and Modeling Workgroups)
January 14.1998
Comment: EPA should consider a broad range of mass emissions databases,
including those reviewed by the Team and those which EPA has said it is still
seeking to obtain and/or analyze (e.g., more complete Ohio EVI240 data, California
Pilot I/M Program data). EPA should report to the Team at a later date its proposal
for the role of each data base in revising MOBILE, groupings of model year,
technology, model year/technology, emitter categories, etc. The team wishes to
review and discuss EPA's draft revision of the in-use deterioration estimates in the
MOBILE model.
EPA Response: EPA agrees that the data summarized in the workgroup report are
enough to warrant a serious reassessment of the MOBlLESa emissions. EPA has
considered the data during that reassessment. EPA has presented interim analyses
to the Workgroup for review, including the model year/technology groupings
chosen, and the role of each data base in revising MOBILE. The documentation for
the proposed in-use deterioration estimates will be provided to the Workgroup for
review.
Comment: EPA should acknowledge the overestimation that results when
predicting future evaporative emissions based on current experience and existing
technology.
EPA Response: EPA agrees that cars meeting the new evaporative test procedures
and standards should be modeled as being lower emitting in-use than previous cars.
This was in fact the case in MOBILES, but the size of the difference is worthy of
reconsideration. The reconsideration ought to start with an understanding of the
causes of high evaporative emissions in the older cars, and then apply an
-14-
-------
understanding of how changes in design and materials would affect those causes.
Comment: EPA should take an active role in obtaining high quality data from state
IM240 programs.
EPA Response: EPA has obtained state IM240 data from Colorado, Arizona,
Wisconsin, and Ohio. Effort in 1998 is limited due to staff time and resource
constraints; however, we remain interested in encouraging and perhaps financially
supporting better preconditioning, more full IM240 tests, and more careful
recording of vehicle information, on a sizable sample of vehicles.
Comment: An Auto/Oil hot soak pilot study has been conducted. In addition, CRC
has conducted a real time diurnal study that will measure 24 hour diurnal emissions
from 151 vehicles. EPA is also conducting a diurnal emission study. The results of
these analyses should be reviewed when available to provide insight into evaporative
emissions deterioration.
EPA Response: EPA plans to use these data along with the data collected under
EPA sponsorship.
Comment #83 (Rick Barrett. CDPHE) July 28.1999
Comment: The 60-day comment period is not sufficient to complete the review of
MOBILE 6 documents.
EPA Response: We agree that due to the simultaneous posting of large amounts of
MOB1LE6 documents that a 60-day comment period may not be sufficient time to
complete a thorough review of the MOBILE6 documents. As such, we have accepted
comments for several months following the original 60-day deadlines.
Comment: "there is likely still an under prediction in both the frequency and
projected emissions levels of high emitters in the proposed Mobile 6 estimates."
EPA Response:
EPA does not believe that the emission factors and high emitter frequency contained
inMOB!LE6 seriously under or over predict (within a reasonable statistical range)
the emission results that have been collected from vehicles or will likely be collected
from typical vehicle emission testing. This conclusion is backed up by considerable
laboratory and in-use IM240 testings, and by recent repeat testing of vehicles
(multiple tests over a long period of time).
-15-
-------
However, it doesn't necessarily mean that the emission factors based on
dynamometer tests exactly match either individual vehicle behavior or even fleet
behavior in the 'real world'. Vehicle emission testing has confirmed that individual
vehicle emission behavior is highly variable. Thus, any individual vehicle or small
sample of vehicles may have results which are considerably different that these
averages. Also, the 'real world' may contain factors that greatly affect emissions
which have not been accounted for in the MOBILE6 modeling due to testing and
experimentation resource limits.
Comment: The EPA has several documents related to the Tier 0 and Tier 1
vehicles, as well as I/M effects, which show the high emitting vehicles have a
constant average emission rate with respect to age or mileage. CDPHE staff believes
this assumption does not make sense. As a result of an audit of Colorado's I/M
program conducted last year by Environ, vehicles identified as failures, and then
repaired, were deteriorating at a faster rate than the remainder of the fleet. In
addition an analysis of Colorado's data conducted by Peter McClintock has shown
that high emitters average emissions do increase with age. CDPHE staff believes
EPA should reevaluate these assumptions based on I/M data where the emissions
performance of the same vehicles can be tracked through multiple inspection cycles.
EPA Response: The assumption of zero deterioration of High emitters was made
because absolutely no correlation could be found in the data (FTP and a limited
sample of Ohio I/M data) that indicated that it is a function of mileage or age. Thus,
the assumption of a constant emission level with respect to mileage or age is the
most prudent assumption to make. The engineering rationale for this assumption
is that the more serious emission related problems are not necessarily more likely
to occur at high mileage than at low mileage. However, the large observed
variance of the high emitter level does suggest that a variety of emission related
problems occur, and that they can occur at any mileage (few problem occur under
20,000 miles, though). The rising High emitter fraction with mileage also implies
that the frequency of High emitter appearance increases with mileage.
Comment: The CDPHE staff feels that EPA has made reasonable assumptions
regarding the effectiveness of OBD systems and motorist response to OBD MIL
indications.
EPA Response: Thanks..
Comment #85 (Joel Schwartz, California Inspection & Maintenance Review
Committee) August 3.1999
-16-
-------
Comment: Data biases should be examined and corrected: Some cars never get
tested. Some motorists prepare for the test by setting up their car to pass the test
without making substantive repairs. And some motorists make substantive repairs
to their cars in preparation for the test. Either (1) validate the EVI240 test-lane data
against data collected in random roadside pullovers or by remote sensing, or (2) use
roadside or RSD data to generate the emission factors for the model. Base RSD
credits on more representative RSD data and encourage the use of the CRC best
practices for collection of high-quality RSD data.
EPA Response:
(Also see responses #1 and #20 of section 7 ofM6.IM.001)
The remote sensing device (RSD) has the potential to collect massive amounts ofin-
use data that could potentially be used for the development of emission factor
models or the confirmation of emission factor models. Unfortunately, at the time of
data collection and analysis, high quality RSD data in large quantities were not
available for use. Also, the quality and applicability of use of RSD data is
frequently dependent on the actions of individual operators and the specific test
locations. Improper siting of the RSD units can lead to results which are not
representative of overall vehicle operation (100% ramp siting for example), or do
not measure a random sample of the vehicle population.
Comment: Base model assumptions on real-world data from vehicle emissions and
human behavior studies, and not on incorrect pre-conceptions about the way people,
vehicles, or programs ought to behave. EPA should provide justification for
"Engineer Assumptions". I/M "Saw-Tooth" does not represent real I/M programs.
Human behavior should explicitly included in the model. Real-world data should
drive evaporative emissions calculation. RSD benefits are contingent on incorrect
assumptions about effectiveness of scheduled I/M programs
EPA Response:
(Also see Response #20 of section 7 ofM6.IM.001)
The RSD benefits have been removed from the MOBILE6 model.
Comment: MOBILE should not be used for regulatory purposes until it has been
appropriately validated. Include multiple internal validation checks so users can
ensure that MOBILE's output jibes with real-world measurements, both overall and
in detail. Make real-world measurements, rather than MOBILE predictions, the final
arbiter of the emissions reduction credit attributed to an I/M program.
-17-
-------
EPA Response: The process for using a model such as the MOBILE model series
for regulatory purposes rather than conducting a multitude of elaborate and
expensive studies of 'real world' emissions has been established through Federal
regulation over several decades. Until this process has been changed, the need for
MOBILE6 will continue.
Comment #87 (Alison Pollack, Till Stoeckenius and Cuong Tran, Environ) August 18,
1999
Comment: Comments from Peter McClintock of Applied Analysis (7/4/1999) and
Tom Wenzel of Lawrence Berkeley National Laboratory.
EPA Response: Are these comments answered or addressed? Comments from Peter
McClintock of Applied Analysis (7/4/1999 conversion of IM240 to running LA4
and 6/25/1999 memo to Phil Lorang regarding change vehicle registration patterns)
Comment: Fast Pass time is incorrectly included in the regression during
conversion of fast pass to full EVI240 using Wisconsin data. Fast pass time (variable
name FSEC in EPA regressions) is assumed to be linearly related to the log of fast
pass emissions.
EPA Response: It would seem logical that the fast pass time would be a significant
variable in the correlation between fast pass IM240 and the Full IM240. It's
exclusion would clearly be a mistake. Only Phil Enns can definitively answer this
question. However, it is my understanding that he double checked all of his
regression equations, and found and corrected any errors that may have been
present.
Comment: Use of the logarithmic transformation is inappropriate and reduces the
effects of the high emitters, leading to underprediction of emissions.
EPA Response: Because of the highly skewed nature of vehicle emission data, a
logarithmic transformation was necessary in-order to perform standard least-
squares regression analysis. In order to overcome any possible under-prediction
of emissions, an additive and positive logarithmic transformation factor was added
to the regression equation. It has the effect of increasing the average emissions and
restoring the effect of high emitters. It is a valid and fairly commonly used
statistical transformation.
-18-
-------
Comment: We recommend that EPA review the proposed approach for estimating
MOBILE6 deterioration rates, taking into account the comments received, and revise
the data bases and modeling methods accordingly. In addition, we recommend that
EPA perform sensitivity and uncertainty analyses to guide further data collection
efforts. In addition, as stated above, we recommend that EPA make all data sets
readily publicly available, and also that full regression statistics be provided for any
regression models using in the development of these and other MOBILE6 emission
factors.
EPA Response: Throughout this process EPA has reviewed all comments and
revised our methodology when appropriate. Allnon confidential or non proprietary
data sets will be made public for subsequent analysis, and most regression models
will also be available as part of the report.
Comment #89 (Robert Slott. Consultant) August 23.1999
Comment: "... large data sets of random full EVI240 measurements exist in a
number of states. These data sets should be used to get a first approximation of
deterioration rates in each of these states." "Analysis of remote sensing
measurements, corrected for vehicle specific power, should be developed to give
more realistic on-road deterioration."
EPA Response: Data from IM240 testing done in Ohio and Wisconsin was used in
MOB1LE6 to adjust the deterioration of the national average fleet estimates in
MOBILE6. Time and resources were not available to do a more extensive review
of the available 1M240 testing results from other States. There was not a sufficient
consensus on how to interpret remote sensing measurements in order to use them
for estimating on-road deterioration for MOB1LE6. The appropriate use of remote
sensing will be considered as apart of future EPA model development.
-19-
-------
Rgum 1: RUNNING LM: FTP-BASED MOBILES and OHIO IM240 ADJUSTMENTS
1961-03 FUEL INJECTION CARS
HC (g/ml)
3.0
«•
2.0
1.6-
1.0-
as
0.0-
CO (a/ml)
120
in
(XIOOO)
ao
10
10
6
O-
; ociooo)
NOX (0/mO
1.6
ao
«clOOO)
88—98 MOBLEB HHH 83—87 MOBLEO
88-88 ADJUSTED A** 83-87 ADJUSTED
LOWER 90* CL —• LOWER 00% CL
UPPER 95% CL ~~ * UPPER 98% CL
120
*** 81-82 MOBILES
*** 81-82 ADJUSTED
— • LOWER 90% CL
— • UPPER 98% CL
-20-
-------
Figure 2: RUNNING LM: FTP-BASED MOBILES and OHIO IM240 ADJUSTMENTS
1981-88 CARBURETED CARS
HC (g/mi)
100
CO (g/fnl)
120
•MO
•MO
MLE8 (XIOOO)
1O-
10O
NOK (gtrti)
2.0 i
•tf-
1.0
OB-
001
i OcMOO)
' LOWER flMt CL
1 UPPER MKCL
— • LOWER 9BKCL
— • UPPER 9BK CL
-21-
-------
HC (g/ml)
Figure 3: FTP-BASED MOBILES PROJECTIONS and OHIO IM240 ADJUSTMENTS
RUNNING LA4. 1868-88 PFI CARS
0.0
200
-22-
-------
HC (g/ml)
1.0 d
Figure 4: FTP-BASED MOBILES PROJECTIONS and OHIO IM240 ADJUSTMENTS
RUNNING LA4. 1868-83 TBI CARS
O4
0.0
200
-23-
-------
HC (g/ml)
Figure 6: FTP-BASED MOBILES PROJECTIONS and OHIO IM240 ADJUSTMENTS
RUNNING LA4, 1866-83 GARB CARS
0.0
200
-24-
-------
HC (g/ml)
Figure 6: FTP-BASED MOBILES PROJECTIONS and OHIO IM240 ADJUSTMENTS
RUNNING LA4, 1863-87 F\ CARS
NOX
-------
HC (g/mi)
Figure 7: FTP-BASED MOBILES PROJECTIONS and OHIO IM24O ADJUSTMENTS
RUNNING LA4, 1863-86 GARB CARS
0.0
200
-26-
-------
HC (g/mi)
Figure 8: FTP-BASED MOBILES PROJECTIONS and OHIO IM240 ADJUSTMENTS
RUNNING LA4, 1861-82 H CARS
1-
200
-27-
-------
HC (g/mi)
Figure 9: FTP-BASED MOBILES PROJECTIONS and OHIO IM240 ADJUSTMENTS
RUNNING LA4. 1981-82 GARB CARS
1-
200
-28-
-------
Table 1
Distribution of Vehicles by Model Year and Technology for the Combined FTP Dataset
81
82
83
84
85
86
87
88
89
90
91
92
93
CARS
TECHNOLOGY
CARB
657
71
57
30
74
34
15
22
OPLP
367
71
63
5
24
7
PFI
29
8
62
35
66
92
106
113
103
250
426
347
366
TBI
15
74
127
46
56
60
76
69
38
160
91
57
29
SUB
TOTAL
1,068
224
309
116
220
193
200
197
163
410
517
404
395
TRUCKS
r
TECHNOLOGY
CARB
3
22
30
9
OPLP
124
45
8
26
33
14
PFI
6
41
4
144
92
93
TBI
13
23
6
144
141
92
90
SUB
TOTAL
124
45
11
49
82
87
10
145
285
184
183
TOTAL
1,192
269
320
165
302
280
210
197
163
555
802
588
578
M6.EXH.001 -29-
-------
TOTAL 977 538 2,003 898 4,416 64 250 382 509 1,205 5,621
M6.EXH.001 -30-
-------
Table 2
Mean Running LA4 and FTP Emission Levels by Model Year for Light-Duty Cars and Trucks
for the Combined FTP Dataset
CARS TRUCKS
|HC_RUN | HCFTP | CO_RUN | COFTP | NOX_RUN | NOFTP | HC_RUN | HCFTP | CO_RUN | COFTP
NOXRUN NOFTP
IMYR
81 | 0.42l| 0.706| 6.489| 9.667| 0.795| 0.897| 0.759| 1.275| 10.876
18.1581 1.6621 1.7521
-1- -1-
82 0.588
83 0.230
84 0.533
85 0.355
-t-
0.789
0.431
0.756
0.533
-1-
5.394
16. 77^
2.760
13.22.
7.622
10.63:
5.561
-1-
8.318
1 1.74(
5.073
5 1.40.
9.968
3 1.38'
6.935
-t-
0.750
3 1.73
0.677
5 1.43
0.785
7 1.4C
0.687
-t-
0.872
2
0.806
5
0.893
5
0.770
-t-
1.163
0.865
0.419
0.923
-t-
1.732
1.361
0.802
1.281
-t-
8.987
5.759
3.597
8.999
-31-
-------
+ + + + + +
+ + I
0.759| 0.926| 8.738| 10.432| 0.612| 0.713
8.789| 1.006| 1.057|
+ + + + + +
+ + |
0.456| 0.656| 7.005| 8.366| 0.698| 0.789
4.610| 0.53l| 0.605|
+ + + + + +
+ + |
0.2121 0.4061 3.3441 4.5741 0.5641 0.6681
86
0.5611 0.8231 6.248
87
0.1641 0.4011 2.959
88
+ + + + + + + + + + .
+ + |
89 I 0.1521 0.3111 2.6451 3.9111 0.5531 0.6521 .1 .1
91
93
90 | 0.109| 0.274| 2.087| 3.614| 0.400| 0.633| 0.163
. | 0.376 | . |
2.245
0.078| 0.237| 1.572| 3.145| 0.353| 0.524| 0.187| 0.800| 2.228
9.510| 0.486| 0.885|
92 | 0.094| 0.267| 2.599| 4.327| 0.322| 0.508| 0.152
.| 0.469| .|
2 .172
0.061| 0.225| 0.977| 2.55l| 0.286| 0.466| 0.137| 0.420| 1.668
5.3631 0.4591 0.8471
M6.EXH.001
-32-
-------
Table 3
Running Emission Deterioration Model Coefficients for HC (Unadjusted)
Light-Duty Cars
ModelYear/
Technology
88-93 PFI
88-93 TBI
83-87 FI
86-93 CARS
83-85 CARS
81-82FI
8 1-82 CARS
ZML Mean
Emissions
(gr/m)
0.0516
0.0843
0.1479
0.0815
0.1691
0.1240
0.2108
First
Slope
(gr/m/ 1000m)
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
First
Corner
(1000 miles)
20.03
34.39
14.10
19.83
25.24
11.29
10.18
Second
Slope
(gr/m/ 1000m)
0.0023
0.0020
0.0079
0.0019
0.0095
0.0038
0.0110
Second
Corner
(1000 miles)
N/A
N/A
81.38
N/A
N/A
70.55
N/A
Third
Slope
(gr/m/ 1000m)
N/A
N/A
0.0060
N/A
N/A
0.0037
N/A
Light-Duty Trucks
88-93 PFI
88-93 TBI
84-93 CARS
81-87 FI
81-83 CARS
0.0932
0.0783
0.2495
0.2927
0.6587
0.0000
0.0000
0.0000
0.0000
0.0000
23.40
16.24
22.03
29.38
15.99
0.0025
0.0043
0.0136
0.0136
0.0110
N/A
55.16
N/A
N/A
N/A
N/A
0.0042
N/A
N/A
N/A
Note: The first slope is zero, since it is assumed that the ZML emission rate is constant from zero miles to the first corner. For the
cases with a single corner, the second slope is determined from the unconstrained regression and the corner occurs at the mileage
where that line intersects the ZML mean emissions. For the case with two corners, the second slope was obtained using a
regression line constrained to pass through the ZML mean emissions-mileage. The third slope is for the unconstrained regression
line and applies at mileages above the second corner. (Unadjusted refers to estimates obtained using the FTP dataset only.)
M6.EXH.001
-33-
-------
Table 3 (cont.)
Running Emission Deterioration Model Coefficients for CO (Unadjusted)
Light-Duty Cars
ModelYear/
Technology
88-93 PFI
88-93 TBI
83-87 FI
86-93 CARS
83-85 CARS
81-82 FI
81-82 CARS
ZML Mean
Emissions
(gr/m)
0.7983
2.5684
2.1416
0.6910
1.0983
1.7270
2.9361
First
Slope
(gr/m/ 1000 m)
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
First
Corner
(1000 miles)
13.78
N/A
14.10
21.13
22.69
16.62
8.79
Second
Slope
(gr/m/ 1000 m)
0.0397
N/A
0.1142
0.0307
0.1739
0.0585
0.1494
Second
Corner
(1000 miles)
N/A
N/A
69.78
N/A
N/A
N/A
15.02
Third
Slope
(gr/m/ 1000 m)
N/A
N/A
0.0898
N/A
N/A
N/A
0.1459
Light-Duty Trucks
88-93 PFI
88-93 TBI
84-93 CARS
81-87FI
81-83 CARS
0.9017
1.1439
1.5384
5.2337
9.0704
0.0000
0.0000
0.0000
0.0000
0.0000
16.80
17.54
19.30
55.03
18.86
0.0357
0.0491
0.1986
0.0644
0.0635
58.68
N/A
N/A
N/A
N/A
0.0297
N/A
N/A
N/A
N/A
Note: The first slope is zero, since it is assumed that the ZML emission rate is constant from zero miles to the first corner. For the
cases with a single corner, the second slope is determined from the unconstrained regression and the corner occurs at the mileage
where that line intersects the ZML mean emissions. For the case with two corners, the second slope was obtained using a
regression line constrained to pass through the ZML mean emissions-mileage. The third slope is for the unconstrained regression
line and applies at mileages above the second corner. (Unadjusted refers to estimates obtained using the FTP dataset only.)
M6.EXH.001
-34-
-------
Table 3 (cont.)
Running Emission Deterioration Model Coefficients for NOx (Unadjusted)
Light-Duty Cars
ModelYear/
Technology
88-93 PFI
88-93 TBI
83-87 FI
86-93 CARS
83-85 CARS
81-82FI
8 1-82 CARS
ZML Mean
Emissions
(gr/m)
0.2582
0.2931
0.5976
0.5522
0.5614
0.6370
0.6121
First
Slope
(gr/m/ 1000 m)
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
First
Corner
(1000 miles)
18.58
21.55
34.25
26.12
12.52
16.36
8.79
Second
Slope
(gr/m/ 1000m)
0.0048
0.0047
0.0042
0.0023
0.0059
0.0129
0.0063
Second
Corner
(1000 miles)
N/A
N/A
N/A
N/A
N/A
N/A
17.00
Third
Slope
(gr/m/ 1000m)
N/A
N/A
N/A
N/A
N/A
N/A
0.0060
Light-Duty Trucks
88-93 PFI
88-93 TBI
84-93 CARS
81-87 FI
81-83 CARS
0.3782
0.3346
1.3234
0.5388
1.6660
0.0000
0.0000
0.0000
0.0000
0.0000
21.20
16.24
22.20
21.43
N/A
0.0044
0.0040
0.0040
0.0084
N/A
N/A
55.16
N/A
N/A
N/A
N/A
0.0032
N/A
N/A
N/A
Note: The first slope is zero, since it is assumed that the ZML emission rate is constant from zero miles to the first corner. For the
cases with a single corner, the second slope is determined from the unconstrained regression and the corner occurs at the mileage
where that line intersects the ZML mean emissions. For the case with two corners, the second slope was obtained using a
regression line constrained to pass through the ZML mean emissions-mileage. The third slope is for the unconstrained regression
line and applies at mileages above the second corner. (Unadjusted refers to estimates obtained using the FTP dataset only.)
M6.EXH.001
-35-
-------
Table 4
High Emitter Adjusted Running Emission Deterioration
Model Coefficients for HC
Light-Duty Cars
ModelYear/
Technology
88-93 PFI
88-93 TBI
83-87 FI
86-93 CARS
83-85 CARS
81-82FI
8 1-82 CARS
ZML Mean
Emissions
(gr/m)
0.0516
0.0843
0.1479
0.0815
0.1691
0.1240
0.2108
First
Slope
(gr/m/ 1000 m)
0.0013
0.0013
0.0000
0.0039
0.0003
0.0094
0.0048
First
Corner
(1000 miles)
20.03
34.39
18.89
19.83
25.24
11.29
10.18
Second
Slope
(gr/m/ 1000m)
0.0036
0.0033
0.0078
0.0058
0.0098
0.0132
0.0158
Second
Corner
(1000 miles)
N/A
N/A
81.38
N/A
N/A
70.55
N/A
Third
Slope
(gr/m/ 1000m)
N/A
N/A
0.0059
N/A
N/A
0.0131
N/A
Adjustment
Additive
(gr/m/ 1000m)
0.0013
0.0013
-0.0001
0.0039
0.0003
0.0094
0.0048
Light-Duty Trucks
88-93 PFI
88-93 TBI
84-93 CARB
81-87 FI
81-83 CARB
0.0932
0.0783
0.2495
0.2927
0.6587
0.0013
0.0013
0.0000
0.0000
0.0018
23.40
16.24
36.01
40.58
15.99
0.0038
0.0056
0.0083
0.0099
0.0127
N/A
55.16
N/A
N/A
N/A
N/A
0.0055
N/A
N/A
N/A
0.0013
0.0013
-0.0053
-0.0038
0.0018
Note: Adjusted refers to estimates obtained using the high emitter correction factors. To obtain the adjusted values, the additive
adjustments given in this table were applied to the unadjusted slopes in Table 3. Slope values of zero were assigned in cases where
the additive adjustments would have resulted in negative deterioration.
M6.EXH.001
-36-
-------
Table 4 (cont.)
High Emitter Adjusted Running Emission Deterioration
Model Coefficients for CO
Light-Duty Cars
ModelYear/
Technology
88-93 PFI
88-93 TBI
83-87 FI
86-93 CARS
83-85 CARS
81-82FI
8 1-82 CARS
ZML Mean
Emissions
(gr/m)
0.7983
2.5684
2.1416
0.6910
1.0983
1.7270
2.9361
First
Slope
(gr/m/ 1000 m)
0.0310
0.0310
0.0000
0.0727
0.0000
0.1817
0.1414
First
Corner
(1000 miles)
13.78
N/A
19.04
21.13
25.68
16.62
8.79
Second
Slope
(gr/m/ 1000m)
0.0707
N/A
0.1091
0.1034
0.1537
0.2401
0.2908
Second
Corner
(1000 miles)
N/A
N/A
69.78
N/A
N/A
N/A
15.02
Third
Slope
(gr/m/ 1000m)
N/A
N/A
0.0846
N/A
N/A
N/A
0.2873
Adjustment
Additive
(gr/m/ 1000m)
0.0310
0.0310
-0.0051
0.0727
-0.0203
0.1817
0.1414
Light-Duty Trucks
88-93 PFI
88-93 TBI
84-93 CARB
81-87 FI
81-83 CARB
0.9017
1.1439
1.5384
5.2337
9.0704
0.0326
0.0326
0.0000
0.0545
0.1040
16.80
17.54
28.90
55.03
18.86
0.0683
0.0817
0.1327
0.1190
0.1675
58.68
N/A
N/A
N/A
N/A
0.0623
N/A
N/A
N/A
N/A
0.0326
0.0326
-0.0660
0.0545
0.1040
Note: Adjusted refers to estimates obtained using the high emitter correction factors. To obtain the adjusted values, the
additive adjustments given in this table were applied to the unadjusted slopes in Table 3. Slope values of zero were assigned
in cases where the additive adjustments would have resulted in negative deterioration.
M6.EXH.001
-37-
-------
Table 4 (cont.)
High Emitter Adjusted Running Emission Deterioration
Model Coefficients for NOx
Light-Duty Cars
ModelYear/
Technology
88-93 PFI
88-93 TBI
83-87 FI
86-93 CARS
83-85 CARS
81-82 FI
8 1-82 CARS
ZML Mean
Emissions
(gr/m)
0.2582
0.2931
0.5976
0.5522
0.5614
0.6370
0.6121
First
Slope
(gr/m/ 1000 m)
0.0010
0.0010
0.0023
0.0021
0.0003
0.0000
0.0003
First
Corner
(1000 miles)
18.58
21.55
34.25
26.12
12.52
30.66
8.79
Second
Slope
(gr/m/ 1000m)
0.0058
0.0058
0.0064
0.0045
0.0062
0.0069
0.0066
Second
Corner
(1000 miles)
N/A
N/A
N/A
N/A
N/A
N/A
17.00
Third
Slope
(gr/m/ 1000m)
N/A
N/A
N/A
N/A
N/A
N/A
0.0063
Adjustment
Additive
(gr/m/ 1000m)
0.0010
0.0010
0.0023
0.0021
0.0003
-0.0060
0.0003
Light-Duty Trucks
88-93 PFI
88-93 TBI
84-93 CARS
81-87 FI
81-83 CARS
0.3782
0.3346
1.3234
0.5388
1.6660
0.0002
0.0002
0.0000
0.0000
0.0008
21.20
16.24
1754.24
32.21
N/A
0.0046
0.0042
0.0001
0.0056
N/A
N/A
55.16
N/A
N/A
N/A
N/A
0.0034
N/A
N/A
N/A
0.0002
0.0002
-0.0040
-0.0028
0.0008
Note: Adjusted refers to estimates obtained using the high emitter correction factors. To obtain the adjusted values, the
additive adjustments given in this table were applied to the unadjusted slopes in Table 3. Slope values of zero were assigned
in cases where the additive adjustments would have resulted in negative deterioration.
M6.EXH.001
-38-
------- |