EPA's Proposal for MOBILE6
           Basic Exhaust Emissions and Deterioration

                         March 14, 1997


Abstract

     In MOBILE5, the base exhaust emission rates are expressed in
terms of composite (three bag) Federal Test Procedure (FTP)
emissions in grams per mile.  These base emissions were then
adjusted to reflect user specified trip speeds, operating mode,
ambient conditions and fuels.  In MOBILE6, EPA proposes to
separate the emissions due to engine start from the "running"
emissions.


Summary of the Proposal

     EPA proposes to use only the emissions from bag 3 (the FTP
505 driving cycle, including a hot engine start) of the FTP,
after removing the effect of engine start from bag 3 (a "running"
505), instead of using the entire FTP cycle for the base exhaust
running emissions.  Please refer to the document, "EPA's Proposal
for MOBILE6: Exhaust Emissions Due to Engine Start," for more
information about how start emissions will be extracted from the
FTP bag emissions.  Using the running 505 as the base exhaust
running emissions allows the use of existing correction factors
data based on the FTP bag emissions.


Supporting Data

     EPA has tested thousands of vehicles in Hammond, Indiana and
Phoenix, Arizona since 1991 using the IM240 transient test
procedure.  Several hundred of these vehicles were recruited for
additional FTP and other testing.  This data is proposed to be
used to develop the correlation between running 505 emissions and
the IM240 test results.

     EPA will obtain test results from thousands of vehicles
tested on the IM240 transient test procedure as part of the
vehicle inspection program from the State of  Ohio.  Some of the
data collected will be from areas which have never previously had
a mandatory vehicle inspection program.  EPA proposes to use
these initial test results to represent the "non-I/M" base
exhaust running emission rates.

     If we are unable to use Ohio IM240 data due to
unavailability or too many complications with the data sample,
then we will use IM240 data from an I/M area, which poses other
problems (see section on "Outstanding Issues").


Methodology

     EPA's proposed effort to create base exhaust running
emissions requires a process for predicting FTP Bag 3 values from
IM240 scores. Proposals for analyses approaches for this work are
documented in two reports prepared under contract for EPA:
"Methodology for Estimating Basic Emission Rates for Use in the
MOBILE Emission Factor Model" (E.H.Pechan & Associates, Inc.);
and "Development of a Methodology for Estimating Basic Emission
Rates for Use in the MOBILE Emission Factor Model" (Sierra
Research, Inc.).

     While the basic form of this relation will be estimated from
the Hammond and Phoenix data, applying that function to a larger
set of IM240 tests must take into account a number of factors,
including:
     
     1.   Fundamental differences between normal and high
               emitting vehicles (refer to the document "EPA's
               Proposal for MOBILE6: Definition of High Emitter
               Category").
     2.   Differences among technology groups (refer to the
               document "EPA's Proposal for MOBILE6: Emission Control
               Technology Distribution")
     3.   Fuel and temperature effects, as reflected in:
     
          a.   Geographical location of test
          b.   Test date (seasonal effect - see Sierra, p.6)
          c.   Fuel type (Sierra, p.16)
          d.   RVP (Sierra, p.17)
          e.   Oxygenate level
     
     4.   Usage adjustments - effect of mileage and age
               accumulation (Pechan, p. 8)
     
     Outstanding Issues

     A number of issues remain to be resolved.
     
     5.   Treatment of odometer outliers in IM240 data sources.
     6.   In areas subject to I/M, large effort required to back-
              out non-I/M deterioration rates.
     7.   Ohio IM240 dataset: 
          a.   Not all vehicles received full IM240 tests
          b.   Many emissions given in total grams per test over
                    variable time frame - unable to determine gram per
                    mile value
          c.   Potential problems identifying vehicle technology,
                    e.g., fuel metering, catalyst, EGR
     8.   Hammond, Indiana  and Phoenix, Arizona IM240 datasets:
          a.   Vehicles have been through several cycles of I/M
                    program.
          b.   Small sample size (20,000 vs. approx. 1,000,000 in
                    Ohio dataset).