United States Air and Radiation EPA420-P-98-008
Environmental Protection May 1998
Agency
Description and
Documentation for
Interim Vehicle Clean
Screening Credit Utility
DRAFT REPORT
-------
May 7,1998
Transmittal of the Draft "Description and Documentation for Interim Vehicle Clean
Screening Credit Utility"
EPA is today releasing draft guidance on the interim use of clean screening methodologies in
state Inspection and Maintenance (I/M) programs. The purpose of the guidance is to share
technical information with states and other interested parties, and to advise them of certain types
of I/M changes EPA is inclined to approve, should states choose to incorporate them into existing
or planned I/M programs. The draft guidance is in the form of two documents and a computer
program. One document addresses concepts and documents the derivation of the credit loss
estimates. The other document provides detailed steps for use of the FORTRAN utility program
which is executed before executing MOBILESb.
The clean screening concepts described in the draft guidance are potentially of considerable
importance to those involved in I/M planning and implementation. In general, clean screening is
aimed at making I/M programs more cost effective by focusing inspections on cars more likely
than others to be high emitters in need of repair. Clean screening can be used to reduce the
number of cars required to be inspected each year-with less than a proportional loss of emission
reduction benefit—which can lead to a more cost-effective I/M program.
Please note that all credit estimates in this draft guidance are not final, and may change as a result
of new data or insights brought to EPA's attention during this review and comment process.
Credit loss estimates for vehicle emitter profiling will change, as EPA continues to make
refinements to this part its analysis. Even once finalized, EPA may from time to time revise the
guidance to reflect the best available data and understanding of clean screening options.
The draft guidance outlines three broad concepts which could be used to excuse likely low
emitters from traditional I/M testing: remote sensing, vehicle emitter profiling, and model year
exemptions. The guidance provides brief background and summary information on each of the
three concepts, comments on the practical application of clean screening in an operating or
planned I/M program, analyses of the probable credit ramifications associated with each
approach based on data sets currently available to EPA, instructions on how to calculate rough
estimates of the credit loss each methodology would likely cause in a specific I/M program, and a
list of references EPA used to prepare the draft guidance.
In preparing this draft guidance, EPA considered a variety of sources and data sets currently
available, which are listed in the "References" section of the guidance. The majority of
information available comes from analytical studies, pilot programs, and modeling exercises
aimed at predicting the immediate credit effect states could expect to experience when choosing
to implement clean screening in an I/M program. While two of the concepts outlined in the
guidance (RSD and model year exemption) are not new, they have not yet been used as a clean
screen in any I/M program. Consequently, there is no large body of real-world application data.
Likewise, utilizing vehicle emitter profiling to identify clean vehicles is fairly new concept as a
-------
clean screening option, rather than a tried and true approach. In addition, EPA has had to develop
a modeling approach for using information on immediate credit effects to predict credit effects in
future years when the mix and condition of cars on the road will be different. For these reasons,
EPA is particularly interested in receiving comments on the following:
• additional data sets and analysis not referenced in the draft documents
the analytical approaches used in determining the modeling protocol for both RSD and
vehicle emitter profiling, and any alternatives which may be useful
• options on how to keep an a vehicle emitter profile up to date for a given area without
compromising the technical foundation of the profile
how EPA and states should approach the ongoing evaluation of clean screening programs
in operation
EPA is soliciting comment on the draft guidance for 60 days, until July 11, 1998. Comments can
be submitted electronically, by mail, or by FAX, at the following locations:
Electronically, to Joe Somers at: somers.joseph@epa.gov
By mail: Joe Somers By FAX: 734/214-4821
U.S. EPA
Assessment and Modeling Division
2000 Traverwood Drive
Ann Arbor, MI 48105
Phone: 734/214-4321
Additionally, EPA will make every effort to include a discussion of the clean screen guidance at
any EPA/stakeholder meetings which occur during the comment period. Once EPA has reviewed
the comments and incorporated them appropriately, we will finalize the guidance for release and
immediate use. We expect this to take place in the fall of 1998.
-------
TABLE OF CONTENTS
Section Title Page Number
1.0 Summary and Concept Descriptions 1
1.1 Introduction and Overview 1
1.2 Summary of 1996 Interim Guidance on Remote Sensing
Devices 2
1.3 Newer Concepts for Clean Screening to Reduce Vehicle
Inspection Volume 4
1.3.1 Remote Sensing Clean Screening 4
1.3.2 Vehicle Emissions Profiling Clean Screening 5
1.3.3 Model Year Exemption Clean Screening 6
1.4 Calculating Effects on I/M Credit 6
1.4.1 RSD Clean Screening 6
1.4.2 LEP Clean Screening 7
1.4.3 Model Year Exemptions 8
1.5 Choosing Among the Clean Screening Options 8
1.6 Program Requirements for Use of These Credits 10
1.6.1 RSD Clean Screening 10
1.6.2 Vehicle Emitter Profiling Clean Screening 13
2.0 Detail on Remote Sensing Device, Vehicle Emitter
Profiling, and Model Year Exemptions 14
2.1 Detail on Remote Sensing Device Option for Clean
Screening 15
2.2 Vehicle Emissions Profiling 21
2.2.1 General Aspects 21
-------
TABLE OF CONTENTS
Section Title Page Number
2.2.2 Credit Losses Estimation 22
2.3 Model Year Exemptions 35
3.0 Input/Output Features of the Clean Screening Utility 35
3.1 Input/Output Features of 36
3.2 Input/Output Features for Lower Emitter Profiling 38
References 39
-------
1.0 Summary and Concept Descriptions
The purpose of this document is to provide interim guidance on use of clean screening to
excuse the low emitting fraction of vehicles from some inspection maintenance (I/M) testing that
would otherwise be required. There are several clean screening options states can use for this
purpose.
This guidance is intended to assist states in developing whatever clean screening program
they believe is appropriate as part of their I/M programs and in modifying their state
implementation plans (SIPs) accordingly.
This guidance document will not be an exhaustive summary of all that is known about
clean screening. Readers are urged to obtain and read all the listed references, many of which
have been posted on the world wide web as described in the body of the document.
This document is presently in draft form. With respect to remote sensing device clean
screening, the estimates and methods included here are those EPA intends to incorporate into the
final version, unless public and peer review surfaces superior information or persuades EPA of a
better analytical and modeling approach. With respect to low emitter profile clean screening,
EPA is planning on doing further analysis to replace the estimates given here with ones that are
more disaggregated by model year. In addition, EPA is particularly open to reviewer comments
on alternative analytical approaches, as the modeling issues are not straightforward. However, the
model year exemption approach is much more straightforward, and can be performed now using
MOBILES.
This guidance, now and when finalized, is based on the best information available at the
time of its preparation. It is subject to change as more information becomes available. Readers
and users should be aware of the possibility of revisions.
1.1 Introduction and Overview
In September 1996, EPA released a document entitled "User Guide and Description For
Interim Remote Sensing Program Credit Utility" [1]. That document contained guidance for use
of remote sensing device (RSD) technology as part of a vehicle inspection maintenance (I/M)
program. The guidance dealt primarily with the use of RSD to identify vehicles as probable high
emitters, for purposes of requiring unscheduled I/M testing at regular inspection stations. The
guidance only briefly mentioned the concept of using RSD to excuse probable low emitters from
scheduled inspections — the clean screening concept. It did promise more guidance later on clean
screening which is why EPA is releasing this draft version now. EPA and others have been
evaluating the following three different methodologies for excusing vehicles from inspection:
1) Remote sensing device (RSD) clean screening;
-------
2) Vehicle emitter profile (VEP) including high emitter profile (HEP) and low
emitter profile (LEP); and
3) Model year exemptions.
This document contains additional, interim guidance on how these three methodologies can be
used. It does not revise the September 1996 EPA guidance on RSD for identifying high emitting
vehicles, but a revision of that earlier guidance is planned for a later date.
The present guidance for clean screening is interim at present for several reasons. The
data on which the guidance is based are limited in nature. For example, the emission results for
RSD are from only a single study although generally corroborated by another published study.
Information on emission results with the low emitter profile are especially limited. While one
reference provides an emissions estimate for the current in-use fleet, no information is available
on emission losses in distant years when vehicles from the 1990 and newer model years have
aged considerably. In addition, the best ways to use clean screening and the emission credit
implications of using it will certainly become clearer with the benefit of state experience actually
using clean screening concepts in I/M programs. At present, there has been no real world
experience, just pilot programs. When and if states decide to adopt clean screening and obtain
full scale program data, this interim guidance will be further revised to reflect states' experience
and data.
This document constitutes EPA guidance and is not a final action or policy. It is intended
to share technical information with states on certain types of I/M program options. It is not
EPA's intention to limit states to considering only the types of changes described here, although
a state interested in other types of changes may need to develop technical information regarding
any other types of changes. Final EPA action on a State Implementation Plan (SIP) revision
changing an I/M program to incorporate clean screening will be made with formal notice and
opportunity for comment. All information available at that time will be considered.
1.2 Summary of 1996 Interim Guidance on Remote Sensing Devices
Remote sensing devices can be used to measure instantaneous exhaust hydrocarbon (HC),
carbon monoxide (CO), and nitrogen oxide (NOx) emissions when vehicles pass through the
RSD measurement beam while operating on the road. The RSD system is set up at various
locations, such as along roadsides or at expressway entrance/exit ramps, to measure these
emissions as vehicles pass. The system retains a video license plate recording of each vehicle
which enables the state or local government to trace the vehicle registration and notify the owner
of the need for further action such as immediate I/M testing.
Additional equipment is used in conjunction with the remote sensing device
instrumentation to measure instantaneous vehicle speed and acceleration. Vehicle acceleration is
especially important since high acceleration rates (and the resultant high load) can lead to
-------
enrichment events and high instantaneous emissions that may not be representative of a vehicle's
overall emissions. Deceleration is also important since it can result in unrepresentative (possibly
low) instantaneous emissions. These additional measurements can be used to help decrease false
passes and failures associated with RSD.
Since RSD measures only a vehicle's instantaneous emissions, which are frequently not
indicative of emissions over a wide range of operating conditions, a vehicle with high RSD
emissions should receive a confirmatory I/M test to determine if the vehicle has high emissions
or not. RSD can be useful in pinpointing vehicles with high emissions that should undergo full
I/M tests in advance of the next scheduled I/M test. Used in this fashion, RSD will result in
additional HC, CO, and NOx credits for a state's I/M program because of the early identification
and repair of high emitting vehicles. The 1996 guidance considered the use of RSD only as a
means of identifying high emitters, and thus focused on the additional benefit a state could
receive through using RSD in this way.
The fraction of the vehicle fleet measured by RSD (coverage) is a critical factor in
determining its benefit. The 1996 EPA guidance gives three methodologies for a state to
determine how much of its fleet is covered by RSD. The first option is to specify the number of
RSD readings annually from which the RSD utility will calculate a fleet coverage based on the
coverage obtained by the California Bureau of Automotive Repairs in its 1995 RSD
demonstration program in Sacramento [2]. The second is for a state to specify what percent of the
fleet by model year it will have measured by RSD considering local traffic patterns; the state
must then assure adequate resources are devoted to obtaining this coverage. The third is for a
state to commit to use RSD to find a specific number of failures on the confirmatory I/M test. A
state using this option must be sure RSD will, in fact, find the number of I/M failures to which
the state commits. The guidance also has an option for use of RSD by itself for an area without a
regular I/M program in which only the vehicles that fail RSD would be sent for a regular I/M test
and the vehicles that pass RSD would not be tested. Areas not needing to implement basic or
enhanced I/M may find this option of interest, especially since it may generate additional SIP
credits.
An important point to note regarding the 1996 EPA guidance is that it was based only on
RSD measurements of CO. The CO measurements were used to determine which vehicles were
high emitters and needed a more highly supervised and/or annual instead of biennial I/M test.
The excess exhaust HC, CO, and NOx emissions for the vehicles identified as high CO emitters
by RSD were based on EVI240 test results obtained from the same vehicles the same day and
were used to determine the credits given in the 1996 guidance.
However, the newer studies on which this clean screening guidance document is based
utilized cutpoints for all of HC, CO, and NOx. EVI240 measurements were made at the time of
the regularly scheduled I/M inspections, later in time than the RSD measurements, as they would
be in a real clean screening program. As such, these studies provide more realistic data than was
available for the preparation for the earlier RSD guidance.
-------
4
1.3 Newer Concepts for Clean Screening to Reduce Vehicle Inspection Volume
This draft guidance focuses on three concepts for clean screening: 1) RSD, 2) vehicle
emitter profiling, and 3) model year exemption. Since the 1996 EPA guidance on remote sensing,
two organizations have conducted projects quantifying credit for clean screening by obtaining
excess emission values from EVI240 data. This work [3,4] on the use of RSD for clean screening
has been done by Remote Sensing Technologies, Inc. (RTSi) and Applied Analysis, a contractor
and subcontractor for the Colorado Department of Public Health and Environment. Additionally,
a recent report prepared for the California Inspection and Maintenance Review Committee [5]
discusses the use of RSD. Radian International LLC and de la Torre Klausmeier Consulting
conducted a study on the use of vehicle emitter profiling for clean screening using data from
Arizona's I/M program.[5] The third option for clean screening is model year exemption of
newer model year vehicles from I/M testing which states can already use as a clean screening
tool. Credit values for model year exemption can be calculated by use of MOBILES as discussed
later in Section 4.0.
1.3.1 Remote Sensing Clean Screening
The 1996 EPA remote sensing guidance also acknowledged that clean screening of
vehicles is another option on which guidance would be provided later. In this concept, roadside
remote sensing is used to excuse some vehicles passing remote sensing from their next annual or
biennial I/M test. This approach has obvious owner convenience and cost advantages; more
specifically, clean screening can reduce the workload for I/M stations, reduce vehicle owner
costs, and increase I/M program acceptance. After skipping one scheduled test, vehicles would
have to get the next one, unless they again pass a remote sensing test. In practical application,
vehicles would have to pass two different remote sensing tests, at different times or places, and
within a limited timeframe (as discussed in Section 1.6.1) in order to be excused from each I/M
test.
Use of RSD for clean screening in general will reduce the SIP credit for I/M because
some vehicles which are high emitting on the day they are excused from I/M testing will have
appeared clean with RSD. The analyses performed to date indicate that with 80 percent fleet
coverage by RSD, using HC and CO RSD cutpoints that excuse about 50 percent of vehicles
from I/M testing results in credit losses for exhaust HC, CO, and NOx of roughly 7 percent, 6
percent, and 22 percent respectively. These numbers are from Figure IV-10 of a recent report to
the Colorado Department of Public Health and the Environment. [4] The loss of NOx credit can
be reduced to about 6 percent by the addition of an RSD NOx cutpoint which will also reduce the
HC and CO loss, but fewer vehicles will be excused from testing.
Local pilot studies with RSD clean screening will add confidence to estimates of the
credit losses it will create. Data from properly conducted local studies should be used in
preference to credit loss estimates presented in this guidance.
-------
There is no convincing evidence that RSD helps to identify low versus high emitting
vehicles with respect to evaporative HC emissions. However, older vehicles have a much higher
incidence of evaporative problems than newer vehicles, and RSD clean screening tends to fail
most older vehicles for exhaust emissions. Therefore, the loss of evaporative HC credits from
excusing a large portion of the fleet from evaporative and tailpipe I/M testing is much less than
the portion of the fleet excused. However, the loss of evaporative emissions HC credits may be
as or more significant than the loss of tailpipe credit. This loss only applies to I/M program that
perform evaporative system checks in their scheduled inspections.
1.3.2 Vehicle Emissions Profiling Clean Screening
Vehicle emissions profiling (VEP) is a method of ranking vehicles for the likelihood that
they need I/M testing and vehicle repairs. [5,6,7] It uses statistics on the historic failure rate of
vehicles of very similar design. For example, failure rates of vehicle models in the Arizona
EVI240 program can be used to predict whether a certain group of nearly homogeneous vehicles
in another state will have a high or low failure rate. This information can be used in either or
both of two ways. In low emitter profiling (LEP), the vehicle ranking would be used as a clean
screening tool to determine what vehicles to excuse from I/M testing. In high emitter profiling
HEP), the ranking is used to require some type of special test regime for the vehicles
most likely to need repair as judged from previous I/M records for an individual vehicle and/or
the composite records for a particular model. This additional testing could involve: 1) special
testing stations with more state supervision or higher qualification standards, or 2) annual
instead of biennial testing. It is possible to add other information into the ranking process, for
example, remote sensing readings. Generally, extra information (such as RSD or previous I/M
records on that particular vehicle) can improve the ranking's accuracy, and thereby reduce the
credit loss from screening out a given number of vehicles from their next regular I/M test.
In general, recent studies indicate that, for the fleet now in operation, vehicle profiling has
roughly the same emission credit ramifications as RSD, depending on the specific profiling used
by the state. Low emitter profiling carries a credit loss, due to the fact that some vehicles will be
incorrectly clean screened out of traditional I/M testing. (On the other hand, high emitter
profiling requires some vehicles to be repaired earlier or better than they would with a traditional
I/M testing thus achieving additional emission reductions and, therefore, resulting in additional
emission reduction credits.) Unlike remote sensing device readings, vehicle emissions profiling
can, in principle, target vehicles likely to fail evaporative emission tests as well as exhaust
emission tests, if the necessary database on evaporative emissions is available. For this guidance,
though, we are assuming the vehicle emissions profiling does nothing to target vehicles with high
evaporative emissions.
This document does not provide guidance on the use of HEP to increase I/M credit by
requiring special I/M testing, or the use of HEP in the absence of a periodic I/M program.
Instead, it gives guidance on use of LEP for clean screening.
-------
1.3.3 Model Year Exemption Clean Screening
This approach excuses vehicles from I/M tests until they reach a certain age, on the
premise that virtually all vehicles are clean when sold and most remain clean for at least several
years. Many states now wait until a vehicle is four or five years old before the first required
inspection, with a small loss in I/M credit for the years the vehicle is not tested. Some other states
give the owner the option of paying a fee in lieu of inspection during these years. It is the state's
choice whether to assess a program fee on new vehicles while they are exempt from testing.
Again, the effect of model year exemption varies, given the fleet mix (i.e., vehicle age
distribution) in a specific area. EPA has suggested that states consider exempting vehicles from
testing until they are at least four years old. Overall, the loss of credit for exempting the four
newest model years is small, and EPA believes most states can find ways to compensate for it if
needed to keep the SIP approvable.
1.4 Calculating Effect on I/M Credits
The effect on I/M credit using any of these three approaches for clean screening can be
modeled using MOBILESa and MOBILESb with the revised utility that can be obtained from the
remote sensing section of the EPA home page for vehicle and engine emission modeling
software at the following site:
http://www. epa.gov/OMSWWW/models. htm
This site also has the MOBILE model and the 1996 RSD utility discussed in Section 1.2.
An explanation of the type of information that has been used to develop the credit utility
for clean screening is given in Section 2, specifically, Sections 2.1, 2.2, 2.3, and in Section 3.0.
1.4.1 RSD Clean Screening
To use the utility to estimate the I/M credit effect of RSD clean screening, the user must
provide information on what portion of the fleet will be measured by the RSD testing program,
what cutpoints the RSD program will use to declare a car low emitting, and some information on
the regular I/M program. The utility then accesses stored default effects estimates to calculate the
overall fleet effect. The utility also allows the user to override the default estimates, to
accommodate any better estimates developed from future research or pilot programs or more
extended analysis of the existing databases. Section 5 more fully explains what information the
user interested in RSD clean screening must supply to use the new clean screening utility. There
is also a more detailed Clean Screening Program User Guide For Interim Remote Sensing
Program Credit Utility [8] that describes how it is used.
-------
1.4.2 LEP Clean Screening
Making the calculations to show the credit effect of a Low Emitter Profile clean screening
program will be more complicated than for RSD clean screening. The utility that accompanies
this guidance is programmed to be able to perform the mechanics of the calculation. It does so by
applying estimates of LEP clean screening credit losses by model year to MOBILE output.
However, the utility has not been programmed with any default or alternative values for the
credit losses by model year. The user must provide values for those losses by model year. The
final version of this guidance document will provide values that can be used, but even once
available in the final version they should be used with more caution than is the case for the
default estimates of RSD clean screening losses. This draft version also provides values, but at
this point they are not values that EPA believes satisfactory for immediate use because they are
based on a more aggregate analysis of model year groups than EPA intends to use for the final
document. The purpose in providing these draft values is to make the methodology and proposed
derivations for estimating credit losses with LEP clean screening easier to understand and
comment upon during the review process.
EPA has reviewed published analyses by one consultant team of a particular, Arizona-
based LEP clean screening method. [6,10] EPA found no apparent defects in this particular
approach to clean screening, and believes it is a good place for other areas to start in their own
thinking. The consultant team provided certain information on the emission reduction impacts of
this approach to LEP clean screening. However, EPA believes that LEP clean screening is
significant and complicated enough that any state seriously interested in using it ought to do an
individualized study, develop an individualized approach including a long-term plan for keeping
the profile current as the fleet evolves, and estimate credit losses for that specific approach. EPA
is available to meet with states to plan how they can approach these tasks.
While EPA favors the eventual development of locale-specific LEP plans and more
elaborately and carefully developed estimates of LEP clean screening effects, we do want to
provide states with a way of making at least ball park estimates of the credit effects they might
expect. Therefore, we have extended the consultant team's analysis to derive credit loss estimates
for use with the utility. In this draft document, these estimates are for four model year ranges,
rather than being fully model year specific. The coarseness of this level of aggregation is less
than satisfying. However, EPA believes the estimates in this draft are adequate for the immediate
purpose of obtaining peer and state review of the methodology itself. EPA is not sure that the
use of these draft estimates will correctly predict the credit loss. EPA does intend to obtain and
include a finer disaggregation of model years for the final guidance document, which should
improve the certainty in predictions. EPA urges states and interested other parties to undertake
research and experimentation in this area. Depending on resources and the level of state interest,
EPA may undertake new work itself. New work and publications may lead to revisions in the
credit estimates to be presented in this document.
-------
Meanwhile, it is EPA's intention once this guidance is final that should a state wish to
adopt the specific type of LEP clean screening that was developed and analyzed as described in
the Radian International LLC/de la Torre Klausmeier Consulting analysis [10] and submit a SIP
revision before any better information is developed, and if other requirements of this guidance
are met, EPA would propose approval of it based on the credit estimates that can be made with
the final guidance. However, EPA would expect the SIP to contain a commitment to make
changes to or abandon the clean screening approach if future program evaluations demonstrate
that more credits were being lost than had been anticipated.
1.4.3 Model Year Exemptions
The credit effect of model year exemptions in specific I/M areas can be estimated using
existing features of MOBILESb. States with a simple system of scheduling inspections on the
anniversary of the new vehicle sales date should have no difficulty analyzing scenarios for
themselves. States with more complex scheduling may need to consult EPA. For example, an
inspection schedule tied to driver birth dates requires some special manipulation of MOBILES
inputs and outputs.
1.5 Choosing Among Clean Screening Options
EPA has no recommendations on which (if any) of the three clean screening options a
state should use as part of its I/M program. The state must examine its particular I/M program
and decide whether one of the clean screening options should be incorporated into its I/M
program. The options have varying program complexities and costs which a state must carefully
consider versus the benefit in cost and size of the I/M program associated with excusing a
specific fraction of vehicles using clean screening. With RSD clean screening, the most
important consideration will likely be the task of establishing and operating the RSD testing
program. With LEP clean screening, the most important consideration will be how to keep
acquiring data to keep the screening profile current and accurate for the local area. In all three
types, businesses currently engaged in inspections may face investment recovery issues if fewer
vehicles need a regular I/M test.
The three approaches may also differ in the perceived equity of vehicle owners'
contribution to cleaner air. Model year exemptions can be perceived to favor high income owners
who typically purchase new vehicles frequently. LEP clean screening may convey a "guilt by
association" implication to some. RSD clean screening comes closest to making the exemption
decision depend only on a vehicle's own emission performance, although some may perceive
caprice or bias in the location of the RSD testing sites.
While EPA itself has no specific privacy misgivings with RSD, privacy concerns have
been raised by others in the past when discussions of RSD or similar technologies have arisen. At
the very least, states considering RSD clean screening should carefully plan their outreach to the
public and press, before misunderstandings can occur.
-------
All three approaches allow a state to choose how many vehicles are excused from
inspection, subject to the requirement that not too much credit be lost overall. EPA's own very
general assessment is that the three approaches to clean screening do not appear at this time to
differ a great deal from one another in terms of how much credit is lost per vehicle that is
excused from inspection. The three options do differ somewhat in this regard, and states should
use the utility and MOBILESb to investigate these differences for their local situation. However,
it is critical to note that the credits in the utility are based on data that are available as of now.
These data are limited, and more limited for LEP clean screening than for RSD clean screening.
If a state chooses to implement a clean screening option, the state must continually evaluate the
program. For example, a state should have a mechanism such as a random I/M emission test
sampling of the fleet as a whole to assess the actual emissions of the vehicles clean screened
versus those that get sent for full I/M testing.
In addition, as more data become available, the emission credit losses given in the clean
screening utility will be revised. States that have already submitted SIPs to adopt clean screening
would not necessarily be expected to revise and resubmit them after such a revision in the EPA
guidance.
A state may choose to target high emitters either with RSD as discussed in the September
1996 EPA Interim Guidance [1] or with the high emitter profiling model (although this guidance
does not have data to provide specific emission benefits for the high emitter profiling option).
States wishing to use RSD to simultaneously identify high emitters to send them to special I/M
testing and low emitters to excuse them from their next scheduled I/M test should contact EPA,
since EPA has no generic guidance for this case. In addition, states wishing to further explore
using high emitter profiling by itself or in combination with low emitter profiling should contact
EPA.
The three types of clean screening will have different indirect benefits with respect to
general program assessment and management. The continually operating remote sensing testing
that is an inherent part of RSD clean screening can provide information on the frequent presence
in a nonattainment area of high emitting vehicles registered elsewhere. These may be vehicles
which have illegally evaded the I/M program, or which could be brought into the I/M program
through a rule change. The steady flow of remote sensing data can also allow at least a qualitative
indication of the trend in fleet emissions, which may be useful for clean air planning and I/M
program evaluation and improvement. The other two methods of clean screening do not produce
the same type of data flow. Model year exemptions obviously result in no data on the exempted
vehicles. LEP clean screening does require a random sample of vehicles to be tested on a regular
basis, but with more opportunity for the owner to influence whether the vehicle actually appears
for testing at all and in its typical emissions condition.
As a general note, a draft OMS reassessment of in-use deterioration of cars and light-duty
trucks will begin wide review shortly. This reassessment is part of the process for creating
MOBILE6, which will replace MOBILES in mid-1999. To the extent that MOBILE6, which will
-------
10
use this reassessment, alters I/M credits in general, the I/M credit loss from the three clean
screening options (remote sensing, low emitter profiling, and model year exemption) will be
affected. For example, exemptions of vehicles more than four years old may have less credit loss
than would be calculated with the current MOBILESb. The latest status and reports [11,12,13,
14] related to the in-use deterioration assessment are made available on the home page of the
MOBILE6 emissions model. The home page address follows:
http://www. epa.gov/OMSWWW/m6. htm
1.6 Program Requirements for Use of These Credits
States wishing to use RSD, vehicle emitter profiling, or model year exemption for clean
screening as part of their I/M program must adjust their I/M credits to reflect these new program
elements. Use of model year exemptions (as discussed in Section 2.3) is generally relatively
simple and adjustments to the I/M SIP and any SIP portions depending on I/M reductions can be
handled by using appropriate inputs for MOBILES. The following sections specifically cover
RSD and vehicle emitter profiling for clean screening. Clean screening programs not designed to
meet the requirements given below are not necessarily disallowed, but will require individual
discussion with EPA; the state has the burden of establishing and supporting credits for the
alternative programs.
1.6.1 RSD Clean Screening
In general, states adopting remote sensing device clean screening must either offset the
credit loss on a one to one basis or show that their I/M SIP and any other previously approved
SIP portions depending on emission reductions from I/M (e.g., the 15 percent rate of progress
plan or attainment demonstration) remain approvable with respect to required emission
reductions even with the credit loss caused by RSD clean screening. States may find this
adjustment possible, or may be able to achieve sufficient compensating emission reductions from
other control programs. Demonstrating such approvability involves modeling for all future years
in which emission reduction requirements are defined such as the rate of progress showing
needed for 2002. This modeling can be accomplished by using the utility program that
accompanies this guidance.
States should be advised that the remote sensing clean screening credits depend on
vehicle fleet mix in a way that causes the percentage loss (but not necessarily the tons loss) in
I/M credits due to clean screening to be larger in the future than at present. Therefore, it may be
more difficult to show that the SIP revision for clean screening is acceptable with respect to 2005
than for 1999, for example. The reason for this increase in credit loss over time is that available
information based on current RSD equipment and techniques shows that the percentage credit
loss is larger for 1990 and newer model year vehicles than for earlier vehicles. However,
refinements in remote sensing may occur that offset this trend; for example, improved test
accuracy that allows for more stringent remote sensing cutpoints without causing false failures
-------
11
may compensate for this increase. However, at this time, a state should not depend on
hypothetical improvements. A state may depend on tighter RSD cutpoints for 1990 and new
vehicles than for older cars, as a way to reduce the credit loss in future years.
Remote sensing clean screening credits are provided in the utility for two cases
distinguished by the type of regular I/M inspection performed on an annual or biennial basis: 1)
EVI240 testing with the EPA-recommended phase-in cutpoints, and 2) EVI240 testing with the
stricter EPA-recommended final cutpoints. [4] I/M credit losses in percentage terms are larger
with the final cutpoints. Programs scheduled to adopt the final cutpoints must show that their SIP
remains acceptable when analyzed on that basis. Programs not using the EVI240 tests must select
between the two available credit cases, or must develop and support customized credit
adjustments for their specific tests. Generally, states with other test types and cutpoints should
select the EVI240 cutpoint case that is closest to the state's actual program in terms of the failure
rate of the fleet. For example, a state with an Acceleration Simulation Mode (ASM) I/M test
should estimate the local failure rate with that test and its cutpoints, and with the EVI240 with
each of its cutpoints, and select whichever EVI240 case is closer in overall fleet failure rate. This
comparison does not necessarily have to be done using a local vehicle sample. EPA will even
consider demonstrations based on use of one sample to determine the EVI240 failure rate and
another sample to determine the ASM failure rate.
Two valid RSD tests with all three pollutants on each test are required to make a clean
screening determination. If a valid reading for one pollutant is not obtained, it must be assumed
that the vehicle failed for that pollutant and is not yet excused from I/M testing. However, if a
state decides not to use NOx cutpoints, valid NOx data are not required. Tests must be at
different sites or on different days. Two RSD units in close proximity do not count as two
separate tests. The two most recent tests meeting these conditions must be used. (Two clean RSD
tests followed by a dirty RSD test showing high emissions do not permit a vehicle to be excused
from scheduled testing.) It is acceptable for a state to set a reasonable cutoff date ahead of each
vehicle's scheduled I/M test (generally a month), and to stop looking for more recent RSD data
after that date. This cutoff allows a state time to match RSD tests, to identify which vehicles are
excused, and to notify vehicle owners ahead of the scheduled I/M test. The older test of the two
separate RSD tests, though, may not be more than 12 months prior to the scheduled I/M test. This
time period is roughly that used in the Colorado Department of Public Health and Environment
Studies. Further research may alter EPA guidance on the question of this time period.
As discussed earlier in Section 1.2 and later in Section 2.1, remote sensing sites must be
selected using good engineering practice in terms of traffic flow, road grade, acceleration, speed,
and other appropriate items. While there are no specific EPA requirements in this guidance, EPA
reserves the right to reject unreasonable site selection practices. Generally, however, it is EPA's
conclusion from available research that poor site selection can result in high readings on vehicles
which deserve to be excused from testing. The falsely high readings can prevent having two
clean RSD tests on a vehicle, or can come after the two clean tests and, thereby, preclude the
vehicle from being excused from its I/M test. Since poor site selection generally does not usually
-------
12
cause a low remote sensing reading on a high emitting vehicle, it does not cause a significant loss
of emission credit, but does partially defeat the very purpose of RSD clean screening which is to
excuse vehicles from I/M testing. In addition to good site selection, as mentioned in Section 1.2,
EPA encourages the use of speed and acceleration limits to define a valid test, as these can also
help achieve the goal of excusing clean vehicles from regular I/M testing. Sites that avoid
vehicles still in cold start mode are also recommended.
Remote sensing units must be designed, programmed, maintained, calibrated, and quality
assured in keeping with good engineering practice. EPA does not certify remote sensing
equipment or RSD contractors, but reserves the right to disapprove a SIP that does not provide
for reasonable data quality. EPA invites consultation with states if they are considering
innovative approaches that lack a performance record. Data recording and storage must allow
any given remote sensing record being used to excuse a vehicle from regular I/M testing to be
traced to a specific date, site, and testing unit. License plate decoding and matching a particular
vehicle to its owner must be accurate, so that the right vehicles are excused from I/M testing.
Relevant data used to excuse a vehicle from its next I/M test must be stored for a reasonable
period. The SIP must address and demonstrate that this matching will occur accurately. The
California Bureau of Auto Repair has prepared material relevant to quality control in RSD
testing.
Remote sensing sites must be selected and rotated to achieve broad vehicle coverage. The
studies done to date have used multiple RSD sites [2, 3,4]. While there are no specific
requirements established in this guidance, EPA reserves the right to accept or reject site selection
plans. The goal is that owners of clean vehicles should not be able to voluntarily present their
vehicles for remote sensing while owners of dirty vehicles (or vehicles that have not received
recent maintenance) avoid remote sensing. For example, a site selection system that facilitates
vehicles obtaining remote sensing measurements soon after passing one scheduled I/M test and
then avoiding remote sensing for the following 11-12 months is not acceptable.
A state can announce remote sensing sites or use one or more of its sites consistently, so
that vehicle owners who believe their vehicles are clean can choose to get tested. This approach
can achieve increased vehicle coverage, and thereby allow more vehicles that deserve to be
excused to actually be excused. These announced or repeated sites must be on-road sites, rather
than special lanes set up off road, unless the special lanes are demonstrated to give RSD readings
representative of on-road sites selected with good engineering practice. If vehicle owners are
given an opportunity to present their vehicles for remote sensing intentionally, the program must
have safeguards to make it difficult for owners to defeat the program by altering their vehicles
just for the two test days, or by temporarily switching license plates. One approach is to also
operate sites likely to capture a true cross section of the fleet. EPA does not rule out other
approaches, but details need to be discussed with EPA to be sure that the system over all will
have an effectiveness as good as the system that produced the data upon which these credits have
been based. The idea is to try to preclude vehicle owners from adjusting their vehicles just for the
remote sensing test ("clean for only a day") and make readjustments afterwards.
-------
13
A random sample of the vehicles which would normally be excused from I/M testing
based on their remote sensing tests must instead be required to get their regularly scheduled I/M
test. This sample is needed to provide data to assess and ensure on a continuing basis that the
remote sensing system is correctly identifying high and low emitting vehicles, to identify remote
sensing sites that need to be reconsidered, etc. To assure these vehicles are truly random and not
specially altered for the I/M testing, owners of these vehicles may not be informed of their
special status. The size of the random sample is not set by this guidance, but needs EPA approval
on a case by case basis. These data should be used to evaluate the credit losses with RSD clean
screening and, as needed, to adjust them to better reflect the local situation in a state. The SIP
must contain details on how this random sample will be obtained and used to evaluate RSD clean
screening.
States may conduct local pilot studies to improve, update, or customize any of the
parameters that determine the I/M credit adjustments. For example, failure rates by age of
vehicles may be different than estimated in the default credits. These failure rates are important
in determining evaporative emissions adjustments.
States that have committed in their SIPs to conduct evaporative system purge or pressure
testing but are awaiting EPA guidance on practicable methods for such testing may adopt clean
screening for tailpipe emissions. If these states have an approved SIP which assumed credit for
an evaporative emission test, the state may adopt RSD clean screening for the tailpipe test with
no adjustment in the evaporative emission credits used in the SIP until the EPA guidance on
evaporative methods is provided and triggers a request for the state to fulfill the SIP commitment
for evaporative inspections. States which are currently conducting or which, in the future, begin
to conduct the type of evaporative inspections specified in their SIP must account for the effect of
clean screening on evaporative emission credits using the provisions of this new RSD clean
screening utility.
1.6.2 Vehicle Emitter Profiling Clean Screening
As with remote sensing clean screening, states adopting low emitter profiling for clean
screening must either offset the credit loss on a one-to-one basis or show that their I/M SIP and
any other previously approved SIP portions depending on I/M (e.g., the 15 percent rate of
progress plan or attainment demonstration) remain approvable with respect to required emission
reductions even with the credit loss from use of the low emitter profiling. States may be able to
achieve needed compensating emission reductions from other control programs. With RSD clean
screening, all future years in which emission reduction requirements are defined must be
modeled in order to determine the approvability of an LEP clean screening plan.
There appears to be a model year dependence for low emitter profiling that also shows
increasing percentage credit loss for later model year vehicles with RSD. If this is true, then it
will be somewhat harder to show SIP approval for a 2005 evaluation date than for the evaluation
date of 1999, for example. Actual events may differ from this apparent trend, but until a showing
-------
14
otherwise, a SIP revision must calculate credits accordingly. States can decrease the LEP clean
screening exemption rates to limit the percent credit loss in future calendar years. EPA suspects
that the credit losses that appear in this draft version of the guidance may overstate the credit loss
in future years, so users should at present consider this an area of some uncertainty.
Since EPA intends the estimates of LEP clean screening credit loss in this guidance to be
more illustrative than definitive for an extended period, and since these estimates rely on a single
study with limited if any peer review so far, these estimates have not been made default values in
the clean screening utility for low emitter profiling. States will have to enter the values given in
Section 2.2 by hand. The final utility issued after peer review and public comment may be
modified to incorporate these values as defaults.
An LEP clean screening State Implementation Plan (SIP) must state how and when the
initial profile for credit losses will be updated and revised as more data become available. As
with RSD clean screening, a random sample of vehicles excused from I/M testing based on the
low emitter profiling must get their regular I/M test. Testing on this random sample of vehicles
will provide data for evaluating whether the profiling is correctly identifying high and low
emitting vehicles. If it is large enough, it will also provide information to help update the profile
algorithm or list.
Evaporative emissions credits should be handled as discussed in Section 1.6.1.
2.0 Detail on Remote Sensing Device, Vehicle Emitter Profiling, and Model Year
Exemptions
There are three possibilities for use of clean screening in conjunction with an I/M
program to excuse low emitting vehicles from I/M testing:
1. Use of RSD to excuse low emitters;
2. Use of low emitter profiling to excuse low emitters; and
3. Use of model year exemption for recent model year vehicles.
Each of these options is discussed in detail in the following sections.
The guidance on RSD clean screening was developed primarily on the basis of two
studies by the Colorado Department of Public Health and Environment [3,4] in Greeley and
Denver. The guidance on low emitter profiling for clean screening was developed from work by
Radian International LLC and de la Torre Klausmeier Consulting.[5,6,7,8,10]
The RSD clean screening guidance covers eight specific scenarios with four different
RSD cutpoints for HC, CO, and (as applicable) NOx with either of two sets of I/M program
-------
15
standards (interim and final IM 240). EPA is open to states using other outpoints but expects
those states to propose and support credits developed in a manner consistent with or superior to
the approach used for the eight scenarios. The software with the RSD clean screening utility can
accommodate customized credits.
The credit impact information for the low emitter profile is limited to a single approach to
LEP clean screening, that presented by Radian International LLC and de la Torre Klausmeier
Consulting, with some conceptual variation to accommodate the evolution of the fleet mix
through the future. The starting point is a list in which every "engine family" is ranked by the
frequency with which vehicles belonging to it have failed the EVI240 test.
The credit loss for model year exemption can be calculated using MOBILES.
2.1 Detail on Remote Sensing Device Options for Clean Screening
Use of RSD for clean screening results in specific vehicles being excused from their next
annual or biennial I/M test because they pass the remote sensing cutpoints. Obviously, this has
owner convenience and cost advantages. Use of RSD results in some I/M credit loss since some
high emitting vehicles may appear clean with RSD and will be excused from the confirmatory
I/M testing and repair for that I/M cycle.
There are some steps states can take to improve the tradeoff between the convenience and
emission reduction loss. As mentioned in Section 1.6.1, vehicles should be required to pass two
different RSD tests at different times or places which is the basis for the numbers in the EPA
clean screening utility. Careful RSD siting practices and good quality control and monitoring can
give clean vehicles the greatest chance of passing while minimizing the passing of dirty vehicles.
The credits given in this guidance assume all these features are implemented. Even with these
features, RSD is not highly accurate in predicting the emission levels of vehicles with only
moderately high emissions. To avoid allowing moderately high emitting vehicles to escape a high
quality periodic I/M test, RSD cutpoints would be set to produce a lower pass percentage than the
normal I/M program cutpoints. Typically, about 50 percent or fewer of vehicles from the fleet as
a whole tested by RSD would be considered passes and would be excused from the regular
periodic I/M test requirement. The other 50 percent would be required to take their I/M test,
which some vehicles will fail and some will pass.
Because I/M programs are designed to address repairable emissions problems and such
problems are strongly correlated with vehicle age, it is not surprising that RSD clean screening
will excuse many more younger vehicles than older ones. In the Colorado study, 24 percent of
vehicles 15 to 18 years old were exempt compared to 63 percent of vehicles younger than 8 yeas,
when an overall 50 percent of the fleet was excused. In concept, a state could decide to impose
some constraint on the exemption pattern that would tend to occur naturally. For example, the
state could decide that in every model year, the engine families representing the cleanest 10
-------
16
percent of the vehicles of that age will be exempt. Conversely, a state could decide that among
vehicles that have reached some age, none will be exempt.
The RSD pass rate and the loss of credit due to errors in passing high emitter vehicles in
the current vehicle fleet varies with model group (or age cohort) as shown in the Colorado
Department of Public Health and Environment reports, specifically Figures IV-10 and IV-13 of
reference. [4] These figures as well as other figures used directly from the different references are
given in Appendix A. One of the trends observable is that the percent I/M credit retained varies
with the model year grouping of the vehicles. For example, the exhaust emission HC I/M credit
retained with RSD cutpoints of 200 ppm HC, 0.5 percent CO, and 1500 ppm NOx is 97 percent,
95 percent, and 90 percent for 1982-85, 1986-89, and 1990 and later model year vehicles
respectively.
For this interim guidance, the assumption is made that the variation in the loss of credit in
percentage terms is a function of technology differences among the model year groups rather than
specific vehicle age. In other words, a cohort of 1990 and newer vehicles will have one value for
the percentage of emission credits retained regardless of its age. While MOBILES shows
increasing emissions of a given vehicle model year with increasing age due to deterioration, the
same percentage of emission credit retained is applied to the increased emissions. Thus, in the
distant years beyond 1998 when most of the fleet consists of 1990 and newer model year
vehicles, the excess emission credit retained will be the smaller percentage shown for these
vehicles rather than the larger percentage for the earlier model year vehicles. This assumption is
one that should be examined and revised, as appropriate, when additional RSD clean screening
data become available. EPA does not rule out that a higher credit retention might naturally occur
on these vehicles as they age. EPA is open to evidence to that effect which could be used to
revise or depart from this guidance. Additionally, it seems reasonable to suppose that even more
stringent RSD cutpoints than those given in this guidance would increase the credit retention rate
and may be practical to implement, especially if RSD technology continues to improve. EPA is
open to additional data, but for now believes states should not presume greater credits than
estimated by the utility.
Strictly speaking, it is appropriate to use the data from the RSD study in Colorado in the
way this guidance does use it only if vehicle owners do not often repair their cars in anticipation
of the initial EVI240 test. If a car in Colorado passed RSP despite having an emissions defect, or
prior to developing one, but the owner pre-emptively obtained an effective repair prior to the
EVI240 test, that car would not play any role in the calculation of the credit retention/loss factors
in this report. In actuality, if such a car were exempted based on its RSD tests the owner might
not get the repair, and the I/M program benefit would be less. EPA invites reviewers of this draft
guidance to comment or offer specific data relevant to this uncertainty.
Vehicles cannot be clean screened out of regular I/M testing if they are unsuccessfully
tested by RSD. Getting two RSD readings on a large fraction of the vehicle fleet within a 12-
month period is not simple, and requires multiple testing units moving among many sites. The
-------
17
1995 California RSD pilot program in Sacramento demonstrated a 47 percent vehicle coverage
for a single RSD reading with 26 percent of the fleet being measured more than once.[l] The
more recent Colorado Department of Public Health and Environment work in Greeley, though,
shows a 72 percent coverage of the vehicle fleet. However, the fraction of the vehicle fleet
receiving two or more RSD readings was only 45 percent.[3, 16] Fleet coverage above 50 percent
should be possible, but will require more resources unless owners begin to seek out the RSD test.
The rough estimates for credit impact given later are only for purposes of illustration.
Since achieving a certain coverage can be a complicated issue, states need to be careful in
making estimates on coverage to assure that the coverage assumed can be obtained with
reasonable resources. States need to consider the number of practical remote sensing locations
(and the associated traffic flows), the number of RSD devices available, the staffing levels to
operate these devices, how many hours they will be operated, and the quality of the readings
obtained. The selection and achievement of a coverage rate is the responsibility of the state; EPA
will approve a state estimate or commitment used based on the weight of evidence for it.
The clean screening utility allows a state two options for fleet coverage. The first option
is commitment to a level of effort in which a state commits to a specific number of vehicles with
valid RSD tests (i.e., two valid readings of HC/CO and NOx if a NOx cutpoint is used on each
vehicle, valid recording of license plate, ability to contact vehicle owners for any required I/M
test) in a given time period. The RSD clean screening utility will internally convert the number of
RSD tests to fleet coverage assuming the vehicles to be eligible for clean screening must receive
two or more valid RSD tests. The second option is commitment to a specific fleet coverage (e.g.,
a certain percentage of the fleet will receive two or more valid RSD tests). The coverage is
specified by a separate fraction for each age of vehicle from new vehicles to those that are 25
years old.
Some additional information from Applied Analysis [15] shows that, at a given overall
fleet coverage with RSD, the RSD coverage is greater for the newer vehicles and lower for older
vehicles. For example, the RSD coverage for the 1990-97 model year vehicles is greater than the
overall average coverage for the entire vehicle fleet in the Denver study. Correspondingly, the
coverage for the pre-1990 vehicles in the Denver RSD study is less than that of the overall fleet
average. This phenomenon is expected due to the different use pattern of newer vehicles (which
have a greater annual "vehicle miles traveled" per vehicle than older vehicles) and, thus, are
more likely to be seen by RSD units along roadways.
The first option for RSD coverage can be used to demonstrate what individual model
year coverage and overall fleet coverages would be achieved by a given level of testing.
Calculations were done for level of effort programs that achieve 30 percent, 50 percent, and 70
percent overall fleet coverage. The individual model year coverages for these three fleet
coverages are given in the User Guide [9] which also explains how the individual model year
coverages need to be entered for the second option. Using cutpoints of 200 ppm HC, 0.5 percent
CO, and no NOx cutpoint, the percent of the fleet excused from I/M testing is 16 percent, 27
-------
18
percent, and 37 percent for the 30 percent, 50 percent, and 70 percent fleet coverage. If a NOx
cutpoint is also used, these percentages are lower.
The loss of I/M credit associated with RSD falsely passing dirty vehicles depends on the
cutpoints chosen. This EPA interim guidance provides credits for cutpoints of only 200 ppm HC
and 0.5 percent CO; other HC/CO cutpoints are not included since no analysis was done for other
HC/CO cutpoints in the Colorado Department of Public Health and Environment study.
Additionally, the utility provides an analysis of I/M credit impacts without any NOx cutpoint or
NOx cutpoints of 1,000, 1,500, or 2,000 ppm. Choosing a NOx cutpoint (especially more
stringent ones such as 1,000 ppm versus 2,000 ppm) reduces the fraction of vehicles that will
pass clean screen but, also, reduces the I/M credit loss since there will be few vehicles falsely
passing RSD.
The Colorado work [4] is the basis for the EPA interim credits for clean screening. The
published report shows the percent of excess emission credit retained for the following three
groupings of model years:
1. 1990 and later model year vehicles;
2. 1986-89 model year vehicles; and
3. 1982-85 model year vehicles.
The EPA clean screening RSD utility uses these three model year groupings, and also assumes
that the pre-1982 model year vehicles have the same I/M credit retained as the 1982-85 model
years. Excess emissions are defined as those emissions over either the looser phase-in or stricter
final EVI240 cutpoints. Figures IV-2 and IV-3 of the Colorado Department of Public Health and
Environment report lists the phase-in EVI240 cutpoints and the final EVI240 cutpoints for light-
duty vehicles (LDVs) (passenger cars) and light-duty trucks (LDTs) (both LDT-1 and LDT-2
categories).
A total of 594 vehicles with two RSD readings for all three pollutants and matched
EVI240 values were in the Colorado Department of Public Health and Environment study. The
percent or overall excess emissions retained among these 594 vehicles (taken directly from
Figures IV-9 and IV-12 of the Colorado report) with clean screening found in this study are given
in Tables 1 and 2 for both the phase-in and final EPA EVI240 cutpoints, for the four possible sets
of RSD cutpoints. Since the phase-in cutpoints are significantly higher than the final cutpoints.
the percent of excess emissions retained is higher when using the phase-in EVI240 cutpoints
rather than the stricter final cutpoints. This point is very important for states to consider as they
determine what clean screening programs are appropriate for their use when modeling future
years in which the SIP commits to final EVI240 standards.
-------
19
Table 1
RSD Clean Screening Effectiveness on a Sample of Vehicles in Denver, Assuming the
Regular I/M Program Uses Interim EPA IM240 Standards
Clean
Screening
Cutpoints
HC 200 ppm
CO 0.5%
NOx - none
HC 200 ppm
CO 0.5%
NOx 2,000
ppm
HC 200 ppm
CO 0.5%
NOx 1,500
ppm
HC 200 ppm
CO 0.5%
NOx 1,000
ppm
Vehicles
Tested
594
594
594
594
% Vehicles
Passing
Clean
Screen
51%
40%
37%
29%
HC
(% excess
exhaust
emission
credit
retained)
98%
98%
99%
99%
CO
(% excess
emission
credit
retained)
93%
93%
100%
100%
NOx
(% excess
emission
credit
retained)
77%
88%
89%
93%
-------
20
Table 2
RSD Clean Screening Effectiveness on a Sample of Vehicles in Denver, Assuming the
Regular I/M Program Uses Final EPA IM240 Standards
Clean
Screening
Cutpoints
HC 200 ppm
CO 0.5%
NOx - none
HC 200 ppm
CO 0.5%
NOx 2,000
ppm
HC 200 ppm
CO 0.5%
NOx 1,500
ppm
HC 200 ppm
CO 0.5%
NOx 1,000
ppm
Vehicles
Tested
594
594
594
594
% Vehicles
Passing
Clean
Screen
51%
40%
37%
29%
HC
% excess
exhaust
emission
credit
retained
91%
94%
95%
96%
CO
% excess
emission
credit
retained
93%
95%
99%
99%
NOx
% excess
emission
credit
retained
72%
85%
88%
93%
The excess emissions retention values in Tables 1 and 2 are based on all 594 vehicles.
The model year distribution that occurred in the 594-vehicle sample may not be representative of
the particular model year distribution found in any given state's fleet. It surely does not represent
the distribution that will exist in future years. However, the utility treats the several model year
groupings as distinct, and MOBILES will recombine them according to the mix implied by the
state's vehicle registration distribution input. More details, including the more detailed credit
retention values for each of the three model year categories, for each scenario of EVI240
standards, are shown in Figures IV-10 and IV-13 of the Colorado Department of Health and
Environment Report. [4]
-------
21
A critical note is that the credits should be revised as additional data beyond these 594
vehicle become available. In particular, as states obtain additional data relating RSD and I/M
values and from the random full EVI240 tests being done, these credits should be updated. In
addition, results ideally should be obtained for individual model years rather than continuing to
use model year groupings such as the three model year groups shown above. The use of
individual model year results should lessen any large discontinuities when going from one model
year group to another. The 594-vehicle sample itself is in EPA's opinion too small to allow
reliable estimates for individual model years.
In determining the total credit loss for HC, the exhaust component mentioned above must
be considered along with the evaporative HC as discussed in Section 1.6.1. Sections 1.3.1 and
1.6.1 discussed evaporative emissions and RSD clean screening. The clean screening utility
assumes that RSD clean screening operates randomly with respect to evaporative emissions
malfunctions. If 50 percent of vehicles of a given age pass the RSD cutpoints, 50 percent of the
potential evaporative benefit of I/M is lost. The clean screening utility is programmed to make
these calculations on a model year by model year basis and to give a composite result.
2.2 Vehicle Emissions Profiling
This section covers a general conceptual overview of how vehicle emissions profiling
would be used and the changes in I/M credits associated with it.
2.2.1 General Aspects
Vehicle emitter profiling (VEP) is a concept which centers upon ranking vehicles for the
likelihood that they would fail the I/M test and need emission repairs, using statistics on the
historic failure rate of other vehicles of very similar design [14]. Those who have been active in
this concept to date have considered vehicles to be "of very similar design" if they match at least
on all of model year, make, model, engine size, and fuel metering system (e.g., carbureted vs.
fuel injection). Those that do match are said to be in the same "engine family," a term borrowed
from EPA's regulatory program but used here in not precisely the same way. The ranked list and
the failure probability for each engine family is called the profile. Once ranked, the engine
families with the lowest probability of failing the I/M test are categorically excused from testing,
even though some individual vehicles in those engine families would fail if tested. The state can
decide how far to go down the list when exempting engine families, but the credit loss per car
excused increases rapidly beyond the point at which one-half of the fleet is excused.
Because I/M programs are designed to address repairable emissions problems and such
problems are strongly correlated with vehicle age, it is not surprising that a typical clean
screening profile will excuse many more younger cars than older. In the example in the
references, virtually all vehicles aged one or two years old were excused, but only 10 percent of
10-year-old cars, when overall 50 percent of the fleet was excused. In concept, a state could
decide to impose some constraint on the exemption pattern that would tend to occur naturally.
-------
22
For example, the state could decide that in every model year, the engine families representing the
cleanest 10 percent of the vehicles of that age will be exempt. Or, a state could decide that among
cars that have reached some age, none will be exempt.
Vehicle emissions profiling requires a very large and broad database of historical
information on in-use vehicle emissions to be successful. This database needs to be continually
updated with more recent information since the profile will change in future calendar years. It is
preferable that the new data come from within the area of the I/M program doing the profiling, so
that it reflects emissions of vehicles as actually affected by the local I/M program and repair
industry. It is essential that within an "engine family," the sample be a true random cross section
of vehicles operating in the I/M area.
Experts consulted by EPA/OMS believe that the database needs to be on the order of,
roughly, one million vehicle records for a 1- or 2-year period; EPA has not confirmed this. With
a database of this size, even engine families with low market share will have enough records to
allow failure rates to be estimated with reasonable accuracy. With a smaller database as might be
obtained with random I/M testing of only a small portion of the fleet, vehicle models with too
few vehicles present in the data would have to be excluded from any possibility of exemption
since there is no way to know with any type of certainty that such vehicles would have low
emissions. However, states can concentrate the random testing on vehicle models needed to
provide updates to the profile. For example, instead of testing many more vehicles of a popular
model than needed to update the profile, fewer numbers of these vehicles could be tested. The
state could then concentrate testing on vehicle models where there would be too few on a random
basis to update the model. Also, states may want to investigate increasing sample sizes in the
portion of the fleet for which errors would cause the greatest credit loss.
More complicated forms of LEP clean screening would use information other than engine
family and that family's I/M failure rate to make the exemption decision, in particular
information about an individual vehicle such as data from any recent RSD encounters or data
from the vehicle's own last I/M test. Conceivably, information about the vehicle owner or the
vehicle's place of registration could play a role. Assuming the underlying data is good and kept
up to date, in principle any extra complication has at least the potential to make the screening
program more efficient, in terms of reducing the credit loss per vehicle excused from testing.
Some of the references discuss some of these concepts, but this guidance does not address them
further.
By definition, vehicle emitter profiling covers 100 percent of the vehicles for a given
model year, so "coverage" is much simpler than for RSD clean screening.
2.2.2 Credit Losses Estimation
The credit loss estimates in this guidance have been developed by extending an available
analysis by Radian International LLC and de la Torre Klausmeier Consulting. They analyzed
-------
23
IM240 data from the Arizona I/M program to develop a LEP clean screening profile. IM240
failure rates were determined for each engine family. The list ranks engine families from lowest
to highest for most recent failure rate on the IM240 test. In actual use in an I/M program, the
managing agency would use the vehicle identification number (VIN) of each car in the fleet, in
conjunction with Radian International LLC's VIN Decoder or its equivalent, to match the vehicle
to its engine family and thereby determine its place on this exemption ranking.
To help evaluate the validity of its approach to creating the ranked list and the usefulness
of an Arizona-based list across different state programs, Radian International LLC applied the
list developed on the Arizona data to the Colorado fleet. The performance of the model in
identifying high, very high, and super emitters is almost the same for both the Arizona and
Colorado fleets as judged by comparing model predictions to the actual EVI240 data in each state.
For both states, the cleanest 50 percent of the fleet contain less than 10 percent of the EVI240
failures.
Table 3 below gives the emission impacts of clean screening for the Arizona fleet now in
operation using the low emitter profile for the full range overall fleet exemption rates. Excess
exhaust emissions are defined as those above the final EVI240 cutpoints. The model predicts that
excusing 50 percent of the current fleet from I/M testing results in losses of exhaust I/M credits
for HC, CO, and NOx of 5.5 percent, 5.7 percent, and 6.8 percent respectively. This table is
taken directly from an analysis done by de la Torre Klausmeier Consulting.[14]
-------
24
Table 3
Emission Impacts on the Current Arizona Fleet of
Targeting with the Low Emitter Profile
Vehicle Fraction
Excused
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
% Excess Emissions in Excused Fraction
Exhaust HC
0.00%
0.23%
0.65%
1.45%
2.93%
5.46%
10.98%
24.90%
37.72%
57.69%
100%
CO
0.00%
0.30%
0.80%
1.75%
3.58%
5.71%
10.28%
23.21%
35.10%
53.08%
100%
NOx
0.00%
0.26%
0.73%
1.57%
3.41%
6.79%
12.32%
25.90%
38.98%
62.30%
100%
-------
25
The values in Table 3 can give a general sense of the potential for LEP clean screening to
reduce I/M test volume with relatively minor loss of I/M exhaust emission reduction. (To this
credit loss must be added some loss in evaporative HC credits, which generally will be larger in
percentage terms than the loss in exhaust HC credits. Instructions for calculating evaporative HC
credit losses are at the end of this subsection.) However, these are overall fleet results for
Arizona, in a recent period. Caution is in order before conclusions are reached for other times
and places. These overall fleet averages obscure any differences among model year groups in
how well low emitter profiling works. Such differences would affect the credit impacts in other
areas with a different mix of model year groups, and even areas with a different I/M program
design since I/M program type can affect how much each model year contributes to overall
program benefit. To help EPA determine whether such differences do exist and, if so, to allow
EPA to develop a flexible method for estimating credit losses in other areas and in future years,
Radian International LLC and De la Torre Klausmeier Consulting provided EPA some
previously unpublished analysis of the data, which is summarized and further analyzed below.
This analysis is in the list of references and is available on the web [14].
The data analysis supplied to EPA showed that the Vehicle Emitter Profile (VEP) was
based on the following number of vehicles in different model year categories:
1982-85 model year 73,052 vehicles
1986-89 model year 132,359 vehicles
1990+ model year 190,007 vehicles.
Information was also submitted showing how many vehicles of each model year category
were included in the different emitter categories. Emitter categories were defined using the EPA
classification of normal, high, very high, and super emitters for HC/CO and normal and high
emitters for NOx. The EVI240 emission limits for these categories as defined by EPA (CARB
definitions are slightly different) for HC and CO are shown below:
Normal emitters:
Less than two times the FTP HC standard for HC
Less than three times CO standard for CO
High emitters:
Between two and four times the FTP exhaust HC standard for HC
Between two and four times the FTP CO standard for CO
Very high emitters:
Over four times the FTP HC standard, but below 10 g/mile for HC
Over four times the FTP CO standard, but below 150 g/mile for CO
-------
26
Super emitters:
Over lOg/mileforHC
Over 150g/mileforCO.
For NOx, these categories are defined based on IM240 NOx emissions as follows:
Normal emitters:
Under 2 g/mile NOx
High emitters:
Over 2 g/mile NOx.
Table 4 below (taken from information provided by Radian International LLC and de la
Torre Klausmeier Consulting) shows the distribution of the almost 400,000 vehicles used in the
low emitter configuration of the vehicle emitter profile in 10 percent increments starting with
vehicles in engine families deemed least likely to fail the I/M test and thus most deserving of
being exempted.
For example, when 50 percent of the overall fleet is exempted, 162,120 vehicles in the
1990 and new model year group are exempted, and these are 85 percent of all 190,007 vehicles
which were 1990 or newer. In contrast, only 27 percent of cars in the 1986 to 1989 age group and
essentially zero percent of vehicles in the 1982 to 1985 group would be among the engine
families most deserving of being exempt.
Table 4 also provides information on the exemption fate of the vehicles that had "very
high" HC and/or CO and those that had "high" NOx in the EVI240 test. While there are additional
categories of emitters, these two categories account for most of the vehicles and most of the
excess emissions of their respective pollutants, and are useful for understanding the patterns that
develop with LEP clean screening. Ideally, few of these would be exempted even when lots of
vehicles in a model year group are being exempted. For example, when 50% of the fleet is
exempted, only 2,922 of the very high HC/CO cars of 1986 to 1989 vintage are exempted, which
is only 11 percent of all 25,645 cars in that age group that had very high HC/CO. This compares
well with the 27 percent exemption rate for all vehicles in this age group.
Further comparisons among columns and rows of Table 4 show that the model year
groups do display differences in how they are affected by clean screening. Their exemption rates
for any given fleet exemption rate differ, as do the exemption rates for their very high HC/CO
emitters and their high NOx emitters. Furthermore, even when different rows (i.e., different
scenarios for overall fleet exemption rate) are chosen to try to equalize the exemption rates
among the model year groups, the exemption rates among their high emitters are still different.
Radian also supplied figures for the exemption rates of the other emitter categories: high HC/CO
and super HC/CO.
-------
27
Table 4
Characteristics and Number of Arizona Vehicles Used in Low Emitter Profile
Percent of
overall fleet
exempted as
probable low
emitters
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Number of
vehicles
exempted as
probable low
emitters
39,542
79,082
118,826
158,168
197,708
237,252
276,794
316,336
355,878
395,418
1990+ Model Vehicles
Number of
vehicles
exempted
as probable
low
emitters
39,117
78,097
114,486
137,917
162,120
178,690
181,577
188,084
190,003
190,007
Number of
exempted
vehicles
which have
very high
HC/CO
497
1,206
2,403
3,371
4,929
7,322
7,737
8,758
9,160
9,160
Number of
exempted
vehicles
which have
high NOx
122
493
1,179
1,849
2,981
3,882
4,076
4,901
5,086
5,086
1986-89 Model Year Vehicles
Number of
vehicles
exempted
as probable
low
emitters
425
987
4,340
20,251
35,561
57,957
83,128
110,665
128,144
132,359
Number of
exempted
vehicles
which have
very high
HC/CO
41
93
289
1,444
2,922
6,306
12,263
18,936
23,854
25,645
Number of
exempted
vehicles
which have
high NOx
44
51
174
1,049
2,490
5,371
9,858
14,854
19,450
20,422
1982-85 Model Year Vehicles
Number of
vehicles
exempted
as probable
low
emitters
0
0
0
0
27
605
12,089
17,587
37,731
73,052
Number of
exempted
vehicles
which have
very high
HC/CO
0
0
0
0
4
183
7,334
10,439
23,776
51,805
Number of
exempted
vehicles
which have
high NOx
0
0
0
0
2
82
2,797
4,338
11,111
25,060
-------
28
One striking observation from Table 4 is that, as expected, the bulk of vehicles excused from I/M
testing by the low emitter profile are 1990 and newer model years. With 50 percent of the fleet
exempted, 85 percent of these newer vehicles are exempted. As the fleet turns over, however,
maintaining an overall 50 percent exemption rate will require that a smaller fraction of these newer cars
be exempted, since the older vehicles with higher exemption rates will gradually disappear from the
fleet. EPA's task in developing a method for estimating credit losses was to account for this type of
program evolution. A related question concerns how the aging process will affect the profiling
characteristics of any of the age groups shown in the table. At present an 85 percent exemption rate
among 1990 and newer vehicles (which occurs at a 50 percent exemption rate for the overall fleet in the
table above) corresponds to a 54 percent exemption among the very high HC/CO vehicles in that group.
This indicates some tendency for very high HC/CO cars to have appeared in particular engine families,
causing them to stand out and be disqualified for exemption. As the vehicles age, will age be the
equalizer and make engine families more similar to one another? The discussion and tables that follow
walk the reader through the method devised by EPA to temporarily address these issues, and the steps
that a modeler should follow to estimate credit loss. As stated in Section 1.4.2, the goal of this effort was
to provide states with at least ball park estimates-but better than simply using Table 3 directly —
pending the development of better methods and assumptions.
The basic concept in the approach recommended in this guidance is to estimate credit losses
within each of the three model year groups, using the exemption rate within that model year group as the
measure of exemption level. Only at end of the calculation process, which will happen within the utility
and MOBILESb, would these pieces be reassembled into an overall fleet effect.
EPA's first step has been to use the information in Table 4 to estimates the credit retention within
each model year group when a certain fleetwide exemption rate is used. The results appear in Table 5
below. These numbers are based on the number of very high HC/CO and high NOx vehicles in the
excused sample fraction versus the total number in each model year group. (High and super emitters for
HC/CO were not considered. Use of only the very high emitters is an expedient approximation to the
strictly proper approach of using individual vehicle data to express the retention directly in terms of
retention of excess emissions. For the final version of this guidance, EPA's goal is to switch to the latter
approach. Meanwhile, the values in Table 5 are approximate.) Excusing 20% of the overall fleet from
I/M testing results in 1,206 of 9,160 very high HC/CO emitters for the 1990 and newer model years in
the profile being excused. Excusing this number of vehicles results in an I/M exhaust credit loss of 13.17
percent or a credit retention of 86.83 percent for this model year grouping. Again, the assumptions in
these calculations are simplistic and do not consider the actual amount of emissions contained and do
not consider that vehicles other than very high emitters for HC/CO and high emitters for NOx contribute
to I/M failures and benefits.
-------
29
Table 5
I/M Exhaust Emission Credits Retained with Low Emitter Profile for Specific Model
Year Groupings With Specific Exemption Rates for Overall Current Fleet
Percent
overall
fleet
excused
from
I/M
testing
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Vehicle Model Year Groupings
Pre-1982
1982-85
1986-89
1990+
Pollutant
HC/CO
100%
100%
100%
100%
100%
99.99%
99.65%
85.84%
79.85%
54.10%
0%
NOx
100%
100%
100%
100%
100%
99.99%
99.67%
88.84%
82.69%
55.66%
0%
HC/CO
100%
100%
100%
100%
100%
99.99%
99.65%
85.84%
79.85%
54.10%
0%
NOx
100%
100%
100%
100%
100%
99.99%
99.67%
88.84%
82.69%
55.66%
0%
HC/CO
100%
99.84%
99.64%
98.87%
94.37%
88.61%
75.41%
52.18%
26.16%
6.98%
0%
NOx
100%
99.78%
99.75%
99.15%
94.86%
87.81%
73.70%
51.73%
27.26%
4.76%
0%
HC/CO
100%
94.57%
86.83%
73.77%
63.20%
46.19%
20.07%
15.53%
4.39%
0%
0%
NOx
100%
97.60%
90.31%
76.82%
63.65%
41.39%
23.67%
19.86%
3.64%
0%
0%
Next, EPA calculated the exemption rate within each model year group that occurs with each
possible fleetwide exemption rate. Table 6 below shows in percentage terms how fleet exemption rates
in the Arizona sample corresponded to exemption rates within model year groups. For example, when
the fleet exemption rate is 30 percent, no pre-1982 or 1982-85 vehicles would be exempted. Table 5 is
derived from the "number of vehicles exempted as probable low emitters" columns of Table 4 by simple
divisions, with the 100% exempt line in Table 4 as the denominator.
-------
30
Table 6
Exemption Rates by Model Year Group
Percent of total
fleet excused
from I/M test
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Percent Within Model Year Groupings Excused From I/M Test
1990+
0%
20.59%
41.10%
60.25%
72.59%
85.32%
94.04%
95.56%
98.99%
100%
100%
1986-89
0%
0.32%
0.75%
3.28%
15.30%
26.87%
43.79%
62.80%
83.61%
96.82%
100%
1982-85
0%
0%
0%
0%
0%
0.04%
0.83%
16.55%
24.07%
51.65%
100%
Pre-1982
0%
0%
0%
0%
0%
0.04%
0.83%
16.55%
24.07%
51.65%
100%
The next step by EPA was to combine Tables 5 and 6 to express model year group-specific credit
retention as a function of model year group-specific exemption rates. This required interpolation to
obtain results at regular intervals of exemption rate. The results are shown in Table 7 below. The table
below gives these numbers for HC/CO and NOx.
-------
Table 7
I/M Exhaust Emission Credits Retained with Low Emitter Profile for Specific Model
Year Groupings Versus Specific Exemption Rates for Those Model Year Groupings
Percent
of
model
year
group
excused
from
I/M
testing
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Vehicle Model Year Groupings
Pre-1982
1982-85
1986-89
1990+
Pollutant
HC/CO
100%
92%
83%
75%
66%
56%
45%
34%
23%
12%
0%
NOx
100%
93%
86%
77%
68%
57%
46%
35%
24%
12%
0%
HC/CO
100%
92%
83%
75%
66%
56%
45%
34%
23%
12%
0%
NOx
100%
93%
86%
77%
68%
57%
46%
35%
24%
12%
0%
HC/CO
100%
96%
92%
86%
79%
68%
56%
44%
31%
17%
0%
NOx
100%
97%
92%
86%
77%
67%
55%
44%
32%
16%
0%
HC/CO
100%
98%
96%
92%
87%
80%
74%
66%
55%
34%
0%
NOx
100%
99%
98%
95%
91%
85%
77%
68%
52%
32%
0%
With these results, an LEP clean screening planner can proceed to analyze any future year. The
planner presumably will want to begin with an intended overall fleet exemption volume, since this
affects the needed capacity in the I/M network and the number of vehicle owners who can look forward
to exemption. This target exemption volume needs to be assigned among three model year groups: 1990
and newer, 1986 to 1989, and pre-1986. (EPA intends to re-divide and increase the number of groups in
the final guidance.) In a real application of clean screening, the ranking of engine families based on a
local random sample will cause this assignment. For modeling purposes now, some supposition is
required. A starting point is the Arizona case.
-------
32
The LEP clean screening planner should start by applying the group-specific percentages from
one line of Table 6 (corresponding to the intended fleetwide exemption percentage) to the expected
vehicle population in that group for the area and calendar year of interest, and determine whether the
resulting number of exempted vehicles matches the exemption volume target. If the calendar year of
interest is close to the present, the match may be close, but for future years it will be considerably off.
The planner then needs to make adjustments to achieve a match. We recommend that if the calculated
exemption volume needs to be reduced, the exemption rates for pre-1986 vehicles should be set to zero
if not already zero, that two initially exempted 1990 and newer vehicles be "de-exempted" for every
additional retention of a 1986 to 1989 car that is "de-exempted" (until no 1986 to 1989 cars remain
exempt), and the remaining adjustment be made to the 1990 and newer group. This algorithm is not
necessarily optimized, and experimentation may show some other way of balancing exemptions that has
a smaller credit loss. For future years, this process will tend to make the 1990 and newer exemption rate
about equal to the target overall fleet exemption rate, since 1990 and newer cars will be the large
majority of the projected fleet mix.
Once the exemption rate within each model year group is selected, Table 7 can be used to find
the credit retention fraction that should be input into the utility for all ages of vehicles that fall within
that model year group in the calendar year to be evaluated. If the exemption rate is not a multiple of 10
percent, it will be necessary to linearly interpolate the between the entries in Table 7.
For example, if the clean screening planner wants to calculate the effect in 2005 of a clean
screening approach that exempts 50 percent of 1990 and newer vehicles and does not exempt any older
vehicles, the credit retention factor for vehicle ages 1 through 16 should be set at 80 percent for each of
HC and CO, and at 85 percent for NOx. The credit retention factor for ages 17 through 25 should be 100
percent for all three pollutants.
EPA recognizes that the use of a single 1990 and newer model year group causes a coarseness in
the modeling that is not satisfying to the careful analyst particularly for evaluation years as far into the
future as 2005. EPA intends to solve this problem in the final version of this guidance by treating each
model year separately.
Despite the shortcoming just described in the numbers available in this draft guidance, EPA
believes the method described for the final guidance will provide good ball park estimates of credit
losses. EPA hopes reviewers of this document will exercise the utility with the numbers available in the
above tables. Doing this will make the mechanics of the process clearer, allowing more informed
comment regarding the appropriateness of the method proposed for the final guidance.
Vehicle emissions profiling can affect evaporative emissions I/M benefits just as RSD clean
screening affects evaporative emissions I/M benefits. Thus, the utility for emission profiling assumes, as
does the RSD-based clean screening utility, that clean screening will reduce the benefits of evaporative
emissions control system checks by the fraction of vehicles in that vehicle age exempted from
inspection. Again, this credit loss is calculated on a model year specific basis. For example, if 40 percent
-------
33
f vehicles of a given age are exempted by low emitter profiling, the benefits of the evaporative system
checks for that vehicle age are reduced by 40 percent.
The utility requires the exemption rate to be input by individual vehicle age in order to calculate
the evaporative credit loss. For the older three of the four model year groups in Table 6, the span of
model years is small and the exemption rates among the individual model years will not be great, and it
would not be unreasonable to use a single exemption rate for all the model years in the group. However,
this is again a degree of coarseness that may not be satisfying. For the 1990 and newer group it definitely
should not be satisfying, especially for future evaluation dates when this group will span a wide age
ranges. In the final version of this document, EPA will provide a more fully researched method of setting
exemption rate by individual model year. Meanwhile, EPA believes a user of this draft guidance can
reasonably approximate the age/model year dependence of exemption rate within the 1990 model year
group by making exemption rate a linear function of age, starting at 100 percent for the newest vehicles
within the group and decreasing for older vehicles. Some trial and error may be needed to obtain the
desired overall exemption rate, accounting for different vehicle populations within the model years.
Also, for exemption rates of less than 50 percent, it may be necessary to have zero exemptions for
several of the oldest model years in the group.
As an example, Table 8 shows one plausible pattern of exemption rate by model year in the year
2005, for a case in which the exemption rate for 1990 and newer vehicles overall is selected to be 40
percent.
-------
34
Table 8
Example of Exemption Rates in 2005 by Model Year for 1990 and Newer Vehicles
Model Year
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
Age in 2005
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Total 1990+ Vehicles
Total Exempted Vehicles
1990+ Group Exemption
Rate
Registrations
136,000
133,929
146,314
144,213
142,955
143,549
143,081
141,251
137,323
135,431
131,864
121,600
106,705
89,243
75,257
61,671
1,990,386
40%
Exemption Rate
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.1
0.1
0.0
0.0
0.0
0.0
Exempted Vehicles
136,000
120,635
117,051
100,949
85,773
71,774
57,232
42,375
27,464
13,543
13,186
12,160
0
0
0
0
798,045
In actual practice, a clean screening planner would use some sample of vehicles spanning all
model years to determine the ranking of the many engine families, and this ranking would imply an
exemption rate in each model year. However, for making estimates of future credit losses, some method
-------
35
of assigning exemption rates to each model year is needed. The method just described is reasonably
consistent with the clean screening case in the references, in which exemption rates did generally follow
a linear trend downward reaching zero exemptions at age 11 or 12.
2.3 Model Year Exemptions
This is the simplest of the clean screening options and requires almost no extra effort on the part
of the state. This approach exempts vehicles from I/M testing until they reach a certain age, on the
premise that virtually all vehicles are low emitting when sold and remain so for at least several years.
Many states now wait until a vehicle is four or five years old before the first I/M test, with a small loss
in credit for the years without testing. Some other states give the owner the option of paying a fee instead
of I/M testing during the exempted years. It is the state's choice whether to assess a program fee on new
vehicles during the exemption period. EPA has suggested that states consider exempting vehicles from
I/M testing until they are at least four years old.
The effect of model year exemption varies from area to area, given the fleet mix and travel
characteristics (i.e., vehicle age distribution, local vehicle miles traveled, even mix of cars/trucks).
Overall, the loss of I/M credit for excusing the four or so newest model years is small; EPA believes
most states can find ways to compensate for it if needed to keep the state implementation plan (SIP)
approvable. However, model year exemptions do result in some high emitting new vehicles being
excluded and some older model year clean vehicles being tested.
The effect of model year exemption in specific I/M areas can be estimated using existing features
of MOBILESb. States with a simple system of scheduling inspections on the anniversary of the new
vehicle sales date should have no difficulty analyzing scenarios for themselves. States with more
complex scheduling (such as requiring I/M tests during the month of the vehicle owner's birthday) may
need to consult EPA to assure the credit adjustment is correctly apportioned considering the time period
the vehicle would operate before the first scheduled inspection.
Generally, model year exemptions result in under 10 percent exhaust HC credit loss with up to
the five newest model years being exempted.
3.0 Input/Output Features of the Clean Screening Credit Utility
The Clean Screening User Guide [18] provides complete step by step details on how to use the
clean screening credit utility. This purpose of this section is to provide an overview of the information
needed for the RSD and vehicle emitter profiling options.
As mentioned in Section 2.3, determining the I/M credit loss for model year exemptions is
straight forward and can generally be easily done with MOBILES without further guidance. Thus, neither
the detailed User Guide nor this summary provides further guidance on model year exemptions.
However, if there are specific questions on how to calculate the change in I/M credits with model year
exemptions, the state should contact EPA.
-------
36
The basic formula for determining the new I/M credit with clean screening is as follows.
New Credit^ = Old Credit^ * Fm * Em,p
Where:
Old Credit = I/M credit from the periodic inspection program
F = Adjusted fraction of the inspected fleet eligible for clean screening.
E = Effectiveness of clean screening at identification of clean vehicles.
m = Quantity is a function of vehicle model year; and
p = Quantity is a function of pollutant (i.e., HC, CO, or NOx).
In the equation, the influence of the underlying I/M program is represented by the variable Old
Creditmp which is chosen from the already-released I/M credits used with the MOBILES model. The
variables Fm and Emp in the equation represent the clean screening fleet coverage and effectiveness. The
fleet coverage is determined as explained in the September 1996 user guide and summarized briefly in
Section 3.1. Only coverage options 1 and 2 are allowed for clean screening scenarios.
The clean screening utility adjusts the I/M credit to account for the I/M credit loss from high
emitting vehicles that pass the clean screening test and are exempted from further I/M testing.
3.1 Input/Output Features of the RSD Portion of the Clean Screening Utility
The first item needed as an input is fleet coverage which is the fraction of the fleet for which
there are two or more valid RSD readings. There are two options for fleet coverage as explained in
Section 2.1:
1) Commitment to a level of effort; and
2) Commitment to a specific fleet coverage.
The first is for a specific number of valid RSD tests which the RSD clean screening converts internally
to a fraction of the fleet receiving two or more valid RSD tests. The second option is simply for the user
to provide this fraction directly on a model year basis.
The next input needed is for the standard I/M credit data files. Each of these files has an
automatic default value. The record descriptions for the input files (discussed in more detail in the Users
Guide itself) follow:
-------
37
1981 and newer model year credits
1981 and newer model year retest-hybrid credits
Pre-1981 model year credits
Pre-1981 model year retest-hybrid credits
Standard pressure/purge credits
Retest-hybrid pressure/purge credits.
The utility adjusts these credit files to account for clean screening.
The age at which vehicles first become eligible for targeting by remote sensing must be given
which allows the state to apply clean screening only to vehicles older than a given age (e.g., 1-3 years
old). Using a specific age at which clean screening is allowed means that vehicles newer than that age
would not be subject to any I/M test requirements. In a way, this combines model year exemption with
clean screening. The default input is 1-year old, meaning all vehicles 1-year old and greater are eligible
for exemption from I/M testing with clean screening.
The user can also determine which model years of vehicles will be included in the clean
screening program. The user has the following model year groupings from which to select:
Clean screening for Pre-1975 model year vehicles
Clean screening for 1975-80 model year vehicles
Clean screening for 1981-85 model year vehicles
Clean screening for 1986-89 model year vehicles
Clean screening for 1990 and newer model year vehicles.
The user also has one of four RSD cutpoints from which to select as follows:
0.5%/200 ppm/1000 ppm CO/HC/NOx cutpoints
0.5%/200 ppm/1500 ppm CO/HC/NOx cutpoints
0.5%/200 ppm/2000 ppm CO/HC/NOx cutpoints
0.5%/200 ppm CO/HC cutpoints (no NOx cutpoint).
The RSD clean screening utility provides default values for clean screening effectiveness as a
function of model year and RSD cutpoints chosen. In addition, the user has an option of an alternate
RSD clean screening scenario where the user supplies the effectiveness numbers based on data that the
user would have developed such as from an RSD pilot program similar to the one run by the Colorado
Department of Public Health and Environment.
3.2 Input/Output Features for Low Emitter Profiling (LED
The input features for the low emitter profiling (LEP) are somewhat different than those for the
RSD clean screening option.
Low emitter vehicle profiling, by definition, covers 100 percent of the fleet. There are three
-------
38
default exemption rates for low emitter profiling: 30 percent, 40 percent, and 50 percent. These
percentages can be on a model year basis instead of the overall fleet. For example, these defaults can be
selected for different vehicle ages increasing the exemption rate for newer vehicles where a larger
fraction of the vehicles are low emitting. However, presently, the credit adjustments have been
developed only for a single percentage number for the overall fleet.
A major input for the clean screening utility for low emitter profiling is the effectiveness or
percent of the I/M emissions remaining for the fleet once a certain number of vehicles have been
excused from I/M testing. The effectiveness numbers are also for vehicle age going from 1 to 25 years. A
number of 1.00 means that the profile is fully effective in screening out only vehicles for that model year
that would pass the confirmatory I/M test. If, for example, 5 percent of the excess I/M 240 emissions
would be in the excused fraction for a particular model year, the effectiveness for that model year would
be 0.95.
-------
39
References
1. * "User Guide and Description For Interim Remote Sensing Program Credit Utility," U. S.
Environmental Protection Agency, Office of Air and Radiation, Office of Mobile Sources, Report EPA
420-R-96-004, September 1996.
2. "Evaluation of the California Pilot Inspection/Maintenance (I/M) Program," Draft Final
Report for California Bureau of Automotive Repair, Rob Klausmeier, Sandeep Kishan, Rick Baker,
Andrew Burnette, and Joel McFarland, de la Torre Klausmeier Consulting, Inc., and Radian
Corporation, March 31, 1995.
3. * "The Greeley Remote Sensing Pilot Program Final Report," Report for Colorado
Department of Public Health and Environment, Remote Sensing Technologies, Inc., Rob Klausmeier
and Peter McClintock, de la Torre Klausmeier Consulting, Inc., and Applied Analysis, November 4,
1997.
4.* "The Colorado Enhanced I/M Program 0.5% Sample Annual Report," Report for
Colorado Department of Public Health and Environment, Remote Sensing Technologies, Inc., Peter M.
McClintock, Applied Analysis, January 27, 1998.
5. "Recommendations for Remote Sensing Use in a Smog Check Program," Report for
California Inspection and Maintenance Review Committee, Rob Klausmeier, de la Torre Klausmeier
Consulting, Inc., and Sandeep Kishan, Radian International LLC, February 9, 1998.
6. "Exemption of Vehicles from I/M Requires in the Phoenix Area: the Arizona Clean
Screen Program," Report for Arizona Department of Environmental Quality, Radian International LLC,
December 4, 1996.
7. "Profiling Vehicle Emission with the High Emitter Profile Mode," Report of California
Bureau of Automotive Repair, Rob Klausmeier, Sandeep Kishan, Meredith Weatherby, Joel McFarland,
Tim DeFries, Dave Amlin, and Bob Benjaminson, de la Torre Klausmeier Consulting, Inc., Radian
International LLC, and California Bureau of Automotive Repair, October 30, 1997.
8. "Profiling Vehicle Emissions Using the High Emitter Profile Model," Briefing for EPA,
Sandeep Kishan and Rob Klausmeier, Radian International LLC and de la Torre Klausmeier Consulting,
Inc., January 29, 1998.
9.* "Program User Guide for Interim Vehicle Clean Screening Program Credit Utility," draft
report EPA420-P-98-007, April 1998.
10.* "Description of the High and Low Emitter Profiles Models," Memo to Joe Somers and
Phil Lorang, EPA from Rob Klausmeier, de la Torre Klausmeier Consulting Inc., February 19, 1998.
-------
40
11.** "Determination of Running Emissions as a Function of Mileage for 1981-1993 Model
Year Light-Duty Cars," EPA Report No. M6.EXH.001 (in preparation; will be posted on the web page
later this month).
12.** "Analysis of Emissions Deterioration Using Ohio and Wisconsin IM240 Data," EPA
Report No. M6.EXH.002 (complete; will be posted on the web page imminently).
13.** "The Determination of Hot Running Emissions from FTP Bag Emissions," EPA Report
No. M6.STE.002 (posted on web page 9/30/97 for 60-day stakeholder review; stakeholder review
complete on 11/29/97).
14.** "Determination of Start Emissions as a Function of Mileage and Soak Time for
1981-1993 Model Year Light-Duty Vehicles," EPA Report No. M6.STE.003 (posted on 10/10/97;
stakeholder review complete on 12/08/97).
15. "Additional Information Requested Regarding RSD Clean Screen Percent by Model
Year,: Memo from Peter McClintock of Applied Analysis to Phil Lorang and Jim Lindner of EPA,
February 23, 1998.
16. "Study Confirms Clean Screening Effective in Emissions Testing," Colorado Department
of Public Health and Environment Press Release, March 5, 1998.
* References available on EPA web page titled "Clean Vehicle Screening & Remote Sensing -
Vehicle Profiling & Remote Sensing Device (RSD) Issues" at following address:
http://www. epa.gov/omswww/rsd. htm
** References available on EPA web page titled "MOBILE6 Vehicle Emission Modeling Software"
at following address:
http://www. epa.gov/OMSWWW/m6. htm
------- |