Skip to Content Skip to Search Skip to Left Navigation U.S. Department of Transportation (US DOT) Logo Research and Innovative Technology Administration (RITA) Logo National Transportation Library
  ABOUT RITA | CONTACT US | PRESS ROOM | CAREERS | SITE MAP
 


Evaluation of MOBILE Vehicle Emission Model



Click HERE for graphic.


                                          NOTICE

This document is disseminated under the sponsorship of the Department
of Transportation in the interest of information exchange.  The
United States Government assumes no liability for its contents or use
thereof.



                                          NOTICE

The United States Government does not endorse products or
manufacturers.  Trade or manufacturers' names appear herein solely
because they are considered essential to the objective of this
report.


Evaluation of MOBILE
Vehicle Emission Model




prepared for:

J.A. Volpe National Transportation Systems Center
U.S. Department of Transportation

under subcontract to:

Jack Faucett Associates
Bethesda, Maryland 20217
Prime Contract  # DTRS-57-89-D-00089



June 1994




prepared by:

Sierra Research, Inc.
1801 J Street
Sacramento, California 95814
(916) 444-6666


Click HERE for graphic.

                                          PREFACE


With the passage of the Clean Air Act Amendments (CAAA) of 1990,
renewed efforts to properly account for emissions from on-road motor
vehicles were initiated.  To that end, the U.S. Environmental
Protection Agency's (EPA's) motor vehicle emission factor model,
MOBILE4, has subsequently been revised three times (i.e., MOBILE4.1,
MOBILE5, and MOBILE5a) to improve the predictive capability of the
model and to incorporate the effects of CAAA directives aimed at
reducing emissions from on-road motor vehicles.  Because of the speed
at which EPA had to make revisions to the model, documentation of
those changes was often limited.  In addition, the CAAA and the 1991
Intermodal Surface Transportation Efficiency Act (ISTEA) call for a
more active role from the U.S. Department of Transportation (DOT) in
reviewing local area transportation plans and making conformity
determinations.  EPA's MOBILE model is the analytical tool with which
those determinations are to be made.  Thus, this report was prepared
to provide DOT a better understanding of the structure and operation
of MOBILE and to document the changes that have occurred among the
model revisions since the release of MOBILE4.

Evaluation of MOBILE Vehicle Emission Model was structured so that
individuals with little background related to motor vehicle emissions
modeling could familiarize themselves with the basic parameters that
make up the MOBILE models.  For that reason, Sections 2 and 3, which
provide a summary of MOBILE components and describe motor vehicle
emission modes (e.g., exhaust versus evaporative emissions), can be
skipped by those possessing prior knowledge of motor vehicle emission
modeling without a loss of continuity.  The remainder of the report
addresses the following topics:

      the basic MOBILE modeling approach is presented with a
       discussion of registration distributions and travel fractions;

      exhaust emissions are discussed, including base emission rates,
       speed corrections, temperature corrections, operating mode
       corrections, and idle emissions;

      the approach to calculate evaporative emission rates, which is
       quite different from the exhaust methodology, is reviewed;

      the methodology to determine the impact of inspection and
       maintenance on both exhaust and evaporative rates is presented;

      the methodologies used by EPA to account for CAAA requirements
       (e.g., Tier 1 standards) are reviewed;

      the contribution of model-year groups to the fleet-average
       emission rate is presented; and

      effects of model changes on fleet-average emission factor
       estimates are presented for the MOBILE4, MOBILE4.1, and MOBILE5a
       revisions.

                                            -v-


Click HERE for graphic.

                                     TABLE OF CONTENTS

Section                                                                              Page

EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xv

1.   INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

       1.1 Summary of MOBILE Components . . . . . . . . . . . . . . . . . . . . . . . . 2
       1.2 User Inputs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
       1.3 Model Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
       1.4 Organization of the Report . . . . . . . . . . . . . . . . . . . . . . . . . 8


2.   DEFINITION OF MOTOR VEHICLE EMISSION MODES . . . . . . . . . . . . . . . . . . . . 9

       2.1 Exhaust Emissions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
       2.2 Evaporative Emissions. . . . . . . . . . . . . . . . . . . . . . . . . . . .10


3.   STANDARD MOBILE MODELING APPROACH. . . . . . . . . . . . . . . . . . . . . . . . .13

       3.1 Registration Distribution and Travel Fraction. . . . . . . . . . . . . . . .13


4.   EXHAUST EMISSION METHODOLOGIES . . . . . . . . . . . . . . . . . . . . . . . . . .17

       4.1   Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
       4.2   Data Used for Developing Base Emission Rates . . . . . . . . . . . . . . .19
       4.3   Tech IV Emission Factors Model . . . . . . . . . . . . . . . . . . . . . .21
       4.4   Final Base Emission Rate Equations . . . . . . . . . . . . . . . . . . . .28
       4.5   Correction Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
       4.6   Idle Emission Rates. . . . . . . . . . . . . . . . . . . . . . . . . . . .49
       4.7   Heavy-Duty Vehicles. . . . . . . . . . . . . . . . . . . . . . . . . . . .55


5.   EVAPORATIVE EMISSION METHODOLOGIES . . . . . . . . . . . . . . . . . . . . . . . .63

       5.1   Emission Standards and Test Procedures . . . . . . . . . . . . . . . . . .63
       5.2   Evaporative Emissions Data . . . . . . . . . . . . . . . . . . . . . . . .65
       5.3   Pressure/Purge Failure Rates . . . . . . . . . . . . . . . . . . . . . . .67
       5.4   Hot Soak and Diurnal Emissions Modeling. . . . . . . . . . . . . . . . . .68
       5.5   Running Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75
       5.6   Resting Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80
       5.7   Refueling Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80
       5.8   Crankcase Emissions. . . . . . . . . . . . . . . . . . . . . . . . . . . .80


TABLE OF CONTENTS (continued)

Section

       6.     MODELING OF I/M PROGRAMS. . . . . . . . . . . . . . . . . . . . . . . . .81

              6.1 Exhaust Emission I/M Modeling . . . . . . . . . . . . . . . . . . . .82
              6.2 Evaporative Emissions I/M Modeling. . . . . . . . . . . . . . . . . .87
              6.3 EPA's I/M Performance Standards . . . . . . . . . . . . . . . . . . .90

       7.     CLEAN AIR ACT REQUIREMENTS. . . . . . . . . . . . . . . . . . . . . . . .93

              7.1   Heavy-Duty Truck Emission Standards . . . . . . . . . . . . . . . .93
              7.2   Cold CO Emission Standards. . . . . . . . . . . . . . . . . . . . .93
              7.3   Revised Evaporative Test Procedure. . . . . . . . . . . . . . . . .94
              7.4   Tier I Emission Standards . . . . . . . . . . . . . . . . . . . . .95
              7.5   Reformulated and Oxygenated Gasoline. . . . . . . . . . . . . . . .95
              7.6   California Low-Emission Vehicle Program . . . . . . . . . . . . . .98


       8.     EMISSIONS VERSUS VEHICLE AGE. . . . . . . . . . . . . . . . . . . . . . 101

              8.1 Analytical Approach . . . . . . . . . . . . . . . . . . . . . . . . 101
              8.2 MOBILE4.1 Results . . . . . . . . . . . . . . . . . . . . . . . . . 103
              8.3 MOBILE5a Results. . . . . . . . . . . . . . . . . . . . . . . . . . 106
              8.4 Comparison of MOBILE4.1 and MOBILE5a. . . . . . . . . . . . . . . . 109


       9.     EFFECT OF MODEL CHANGES ON EMISSION FACTOR ESTIMATES. . . . . . . . . . 111

              9.1   Baseline Emissions. . . . . . . . . . . . . . . . . . . . . . . . 111
              9.2   Operating Mode Fractions. . . . . . . . . . . . . . . . . . . . . 115
              9.3   Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
              9.4   Speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
              9.5   Fuel Effects. . . . . . . . . . . . . . . . . . . . . . . . . . . 121
              9.6   Inspection and Maintenance. . . . . . . . . . . . . . . . . . . . 127


APPENDIX A: EMISSION CONTROL TECHNOLOGIES AND COMPARISON OF
       BASE EMISSION RATES AT 50,000 AND 100,000 MILES
       FOR THE LDGT1 AND LDGT2 VEHICLE CLASSES. . . . . . . . . . . . . . . . . . . . 135

APPENDIX B: SPEED-TIME PROFILES OF CYCLES USED IN SPEED
       CORRECTION FACTOR DEVELOPMENT. . . . . . . . . . . . . . . . . . . . . . . . . 149

APPENDIX C: COMPARISON OF RUNNING EXHAUST AND IDLE EMISSION
       RATES COMPUTED WITH MOBILE4.1 AND MOBILE5a . . . . . . . . . . . . . . . . . . 155

APPENDIX D: CONTRIBUTION OF MODEL YEAR GROUPS TO THE
       FLEET-AVERAGE EMISSION RATES COMPUTED WITH
       MOBILE4.1 AND MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

REFERENCES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203


                                          -viii-

                                      LIST OF FIGURES

Figure                                                                               Page

1      Fleet-Average HC Emission Estimates
              MOBILE4/MOBILE4.1/MOBILE5a. . . . . . . . . . . . . . . . . . . . . . xviii
2      Fleet-Average CO Emission Estimates
              MOBILE4/MOBILE4.1/MOBILE5a. . . . . . . . . . . . . . . . . . . . . . . xix
3      Fleet-Average NOx Emission Estimates
              MOBILE4/MOBILE4.1/MOBILE5a. . . . . . . . . . . . . . . . . . . . . . . xix
4      1990 LDGV Hydrocarbon Base Emission Rates
              MOBILE4/MOBILE4.1/MOBILE5 . . . . . . . . . . . . . . . . . . . . . . . .xx

5      1990 LDGV Carbon Monoxide Base Emission Rates
              MOBILE4/MOBILE4.1/MOBILE5 . . . . . . . . . . . . . . . . . . . . . . . xxi
6      1990 LDGV Oxides of Nitrogen Base Emission Rates
              MOBILE4/MOBILE4.1/MOBILE5 . . . . . . . . . . . . . . . . . . . . . . . xxi
7      Effect of Speed on Year 2000 NOx Emissions
              MOBILE4/MOBILE4.1/MOBILE5a. . . . . . . . . . . . . . . . . . . . . . .xxii
8      Comparison of MOBILE4.1 and MOBILE5a Fleet-Average Idle
       CO Emission Rates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
9      Sample Output File from MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . 7
10     LDV Registration Distribution. . . . . . . . . . . . . . . . . . . . . . . . . .15
11     LDGV Mileage Accumulation Rates. . . . . . . . . . . . . . . . . . . . . . . . .15
12     LDGV VMT Distribution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
13     Urban Dynamometer Driving Cycle. . . . . . . . . . . . . . . . . . . . . . . . .19
14     Overview of the Tech IV Model :. . . . . . . . . . . . . . . . . . . . . . . . .23
15     Tech IV Incidence of Emitter Regimes
              1983+ Closed-Loop Carbureted Vehicles . . . . . . . . . . . . . . . . . .26
16     Tech IV Incidence of Emitter Regimes
              1983+ Closed-Loop Fuel-Injected Vehicles. . . . . . . . . . . . . . . . .26
              17    Contribution of Regimes to HC Emission Factors
              1983+ Closed-Loop Carbureted Vehicles . . . . . . . . . . . . . . . . . .27
              18    Contribution of Regimes to HC Emission Factors
              1983+ Closed Loop Fuel-Injected Vehicles. . . . . . . . . . . . . . . . .27
              19   MOBILE4 HC Emission Factor, 1990 LDGV. . . . . . . . . . . . . . . .29
              20    Comparison of MOBILE5 HC Base Emission Rates
              by Model Year (LDGV). . . . . . . . . . . . . . . . . . . . . . . . . . .30
              21    Comparison of MOBILE5 Oxides of Nitrogen
              Base Emission Rates (LDGV). . . . . . . . . . . . . . . . . . . . . . . .30
              22    1990 LDGV Hydrocarbon Base Emission Rates
              MOBILE4/MOBILE4.1/MOBILE5 . . . . . . . . . . . . . . . . . . . . . . . .33
              23    1990 LDGV Carbon Monoxide Base Emission Rates
              MOBILE4/MOBILE4.1/MOBILE5 . . . . . . . . . . . . . . . . . . . . . . . .33
              24    1990 LDGV Oxides of Nitrogen Base Emission Rates
              MOBILE4/MOBILE4.1/MOBILE5 . . . . . . . . . . . . . . . . . . . . . . . .34
       25     LDGV  HC Emission Rates by Model Year at 50,000 miles . . . . . . . . . .35
       26     LDGV  HC Emission Rates by Model Year at 100,000 miles. . . . . . . . . .35
       27     LDGV  CO Emission Rates by Model Year at 50,000 miles . . . . . . . . . .36
       28     LDGV  CO Emission Rates by Model Year at 100,000 miles. . . . . . . . . .36
       29     LDGV  NOx Emission Rates by Model Year at 50,000 miles. . . . . . . . . .37
       30     LDGV  NOx Emission Rates by Model Year at 100,000 miles . . . . . . . . .37
       31     MOBILE5 HC Bag Fractions. . . . . . . . . . . . . . . . . . . . . . . . .38
       32     MOBILE5 CO Bag Fractions. . . . . . . . . . . . . . . . . . . . . . . . .39
       33     MOBILE5 NOx Bag Fractions . . . . . . . . . . . . . . . . . . . . . . . .39


                                           -ix-

                                LIST OF FIGURES (continued)

Figures                                                                              Page

34     Bag 1 HC Temperature Correction Factors by Fuel Delivery 
       System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41
35     HC Temperature Correction Factors by Operating Mode. . . . . . . . . . . . . . .42
36     CO Temperature Correction Factors by Operating Mode. . . . . . . . . . . . . . .42
37     NOx Temperature Correction Factors by Operating Mode . . . . . . . . . . . . . .43
38     MOBILE5 Low Temperature HC RVP Correction Factors. . . . . . . . . . . . . . . .44
39     MOBILE5 Low Temperature CO RVP Correction Factors. . . . . . . . . . . . . . . .44
40     MOBILE5 High Temperature HC 
       RVP/Temperature Correction Factors . . . . . . . . . . . . . . . . . . . . . . .45
41     MOBILE5 High Temperature CO 
       RVP/Temperature Correction Factors . . . . . . . . . . . . . . . . . . . . . . .45
42     Hydrocarbon Speed Correction Factors MOBILE4/MOBILE4.1/MOBILE5
43     Carbon Monoxide Speed Correction Factors
       MOBILE4/MOBILE4.1/MOBILE5. . . . . . . . . . . . . . . . . . . . . . . . . . . .47
44     Oxides of Nitrogen Speed Correction Factors
       MOBILE4/MOBILE4.1/MOBILE5. . . . . . . . . . . . . . . . . . . . . . . . . . . .48
45     MOBILE4.1 LDGV Idle CO Emission Rates by Temperature and
       Operating Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52
46     Comparison of MOBILE4.1 Idle and Running Exhaust CO Rates at
       50,000 Miles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52
47     Comparison of MOBILE5a Idle and Running Exhaust CO Rates at
       50,000 Miles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54
48     MOBILE4.1 vs MOBILE5a Idle Exhaust CO Rates at 50,000 Miles. . . . . . . . . . .54
49     Comparison of MOBILE4.1 and MOBILE5a Fleet-Average Idle CO
       Emission Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55
50     Heavy-Duty Engine Test Procedure . . . . . . . . . . . . . . . . . . . . . . . .56
51     Heavy-Duty Vehicle Speed Correction Factors, Hydrocarbons. . . . . . . . . . . .59
52     Heavy-Duty Vehicle Speed Correction Factors, Carbon Monoxide . . . . . . . . . .59
53     Heavy-Duty Vehicle Speed Correction Factors, Oxides of
       Nitrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
54     Heavy-Duty Gasoline Vehicle Temperature Correction Factors . . . . . . . . . . .60
55     LDGV Baseline Pressure/Purge Failure Rates MOBILE4.1 Versus
       MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67
56     MOBILE5 Hot Soak Emissions vs RVP Multi-Point, Fuel-Injected
       "Passing" Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71
57     MOBILE5 Hot Soak Temperature Correction Factors Multi-Point,
       Fuel-Injected Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71
58     Fractional Occurrence of Diurnal Episodes Used in MOBILE5a . . . . . . . . . . .73
59     Travel Parameters Used in MOBILE5a to Develop LDGV g/mi
       Evaporative Emission Rates . . . . . . . . . . . . . . . . . . . . . . . . . . .74
60     Effect of Temperature and RVP on MOBILE4 Running Loss Emission
       Estimates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75
61     MOBILE4 Model Year Specific Running Loss Emission Estimates for
       Non-Tampered LDGV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76
62     Comparison of Tampered and Non-Tampered MOBILE4 Running Loss
       Emission Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .77

                                            -x-

LIST OF FIGURES (continued)

Figure                                                                               Page

63     MOBILE5 Running Loss Emission Rates for LDGV Passing
       the Pressure/Purge Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . .78
64     MOBILE5 Running Loss Emission Rates for LDGV Failing
       the Pressure/Purge Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . .79
65     Comparison of MOBILE5 Running Loss Rates for LDGV
       Passing and Failing the P/P Test . . . . . . . . . . . . . . . . . . . . . . . .79
66     MOBILE5/Tech5 Incidence of Emitter Groups as a
       Function of Vehicle Age. . . . . . . . . . . . . . . . . . . . . . . . . . . . .84
67     MOBILE/Tech5 HC Exhaust Emission Rate by Emitter Category
       as a Function of Vehicle Age . . . . . . . . . . . . . . . . . . . . . . . . . .84
68     MOBILE/Tech5 Contribution of Emitter Categories
       to Baseline HC Emission Rate . . . . . . . . . . . . . . . . . . . . . . . . . .85
69     Tech5 HC Identification Rates. . . . . . . . . . . . . . . . . . . . . . . . . .86
70     Tech5 HC After-Repair Emission Rates as a Function
       of Before-Repair Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86
71     MOBILE/Tech5 Contribution of Emitter Categories
       to After-I/M HC Emission Rate. . . . . . . . . . . . . . . . . . . . . . . . . .87
72     Tech5 HC Emission Rates
       Before and After an IM240 Inspection . . . . . . . . . . . . . . . . . . . . . .88
73     MOBILE5a Pressure/Purge Failure Rates With and Without an
       Annual P/P Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89
74     MOBILE5 CO Exhaust Emission Benefits from Oxygenated Fuels . . . . . . . . . . .97
75     MOBILE5 HC Exhaust Emission Benefits from Oxygenated Fuels . . . . . . . . . . .97
76     CO Exhaust Benefits for 2.7% Oxygen Content
       MOBILE4.1 vs. MOBILE5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98
77     MOBILE5 LEV Program HC Emission Rates. . . . . . . . . . . . . . . . . . . . . .99
78     Sample MOBILE5a Model-Year Output. . . . . . . . . . . . . . . . . . . . . . . 102
79     MOBILE4.1 LDGV Exhaust HC Emissions by Model Year Group. . . . . . . . . . . . 104
80     MOBILE4.1 LDGV CO Emissions by Model Year Group. . . . . . . . . . . . . . . . 104
81     Model Year Contribution to Fleet-Average LDGV Exhaust HC
       Emission Rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
82     Model Year Contribution to Fleet-Average LDGV CO
       Emission Rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
83     MOBILE5a LDGV Exhaust HC Emissions by Model Year Group . . . . . . . . . . . . 107
84     MOBILE5a LDGV CO Emissions by Model Year Group . . . . . . . . . . . . . . . . 107
85     Model Year Contribution to Fleet-Average LDGV Exhaust HC
       Emission Rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
86     Model Year Contribution to Fleet-Average LDGV CO
       Emission Rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
87     Fleet-Average HC Emission Estimates
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 113
88     Fleet-Average CO Emission Estimates
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 113
89     Fleet-Average NOx Emission Estimates
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 114
90     Detailed HC Breakdown - 1995 and 2005
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 115
91     Comparison of Year 2000 HC by Operating Mode
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 116
92     Comparison of Year 2000 CO by Operating Mode
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 116


                                           -xi-

                                LIST OF FIGURES (continued)

Figure                                                                               Page

93     Comparison of Year 2000 NOx by Operating Mode
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 117
94     Effect of Temperature on Year 2000 Exhaust HC
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 118
95     Effect of Temperature on Year 2000 CO Emissions
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 118
96     Effect of Temperature on Year 2000 NOx Emissions
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 119
97     Effect of Temperature on Year 2000 Diurnal HC
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 119
98     Effect of Temperature on Year 2000 Hot Soak HC
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 120
99     Effect of Temperature on Year 2000 Running Loss HC
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 121
100    Effect of Speed on Year 2000 Exhaust HC
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 122
101    Effect of Speed on Year 2000 CO Emissions
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 122
102    Effect of Speed on Year 2000 NOx Emissions
       MOBILE4/MOBILE4.1/MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . 123
103    Effect of RVP on MOBILE5a Nonexhaust HC
       Emission Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
104    Effect of RVP on MOBILE5a Exhaust HC Emission Estimates. . . . . . . . . . . . 125
105    Effect of RVP on MOBILE5a CO Emission Estimates. . . . . . . . . . . . . . . . 125
106    Effect of Oxygenated Fuels on MOBILE5a
       CO Emission Estimates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
107    Effect of Reformulated Gas Regulations on
       MOBILE5a HC Emission Estimates . . . . . . . . . . . . . . . . . . . . . . . . 127
108    MOBILE4 LDGV Exhaust HC Emissions Before and After I/M . . . . . . . . . . . . 128
109    MOBILE4.1 LDGV Exhaust HC Emissions Before and
       After I/M. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
110    MOBILE5a LDGV Exhaust HC Emissions Before and
       After I/M. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
111    MOBILE4 LDGV CO Emissions Before and After I/M . . . . . . . . . . . . . . . . 130
112    MOBILE4.1 LDGV CO Emissions Before and After I/M . . . . . . . . . . . . . . . 130
113    MOBILE5a LDGV CO Emissions Before and After I/M. . . . . . . . . . . . . . . . 131
114    MOBILE5a LDGV NOx Emissions Before and After I/M . . . . . . . . . . . . . . . 131
115    Effect of Pressure/Purge Tests on MOBILE5a
       LDGV Evaporative Emission Estimates. . . . . . . . . . . . . . . . . . . . . . 132
116    Effect of Pressure/Purge Tests on MOBILE5a
       LDGV Running Loss Emission Estimates . . . . . . . . . . . . . . . . . . . . . 133
117    LDTG1 Exhaust Standards Summary and Predominant Emission
       Control Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
118    LDGT1 HC Emission Rates by Model Year at 50,000 Miles. . . . . . . . . . . . . 139
119    LDGT1 HC Emission Rates by Model Year at 100,000 Miles . . . . . . . . . . . . 139
120    LDGTI CO Emission Rates by Model Year at 50,000 Miles. . . . . . . . . . . . . 140
121    LDGT1 CO Emission Rates by Model Year at 100,000 Miles . . . . . . . . . . . . 140
122    LDGT1 NOx Emission Rates by Model Year at 50,000 Miles . . . . . . . . . . . . 141
123    LDGT1 NOx Emission Rates by Model Year at 100,000 Miles. . . . . . . . . . . . 141
124    LDTG2  Exhaust Standards Summary and Predominant Emission
       Control Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144


                                           -xii-

                                LIST OF FIGURES (continued)

Figure                                                                               Page

125    LDGT2 HC Emission Rates by Model Year at 50,000 Miles. . . . . . . . . . . . . 145
126    LDGT2 HC Emission Rates by Model Year at 100,000 Miles . . . . . . . . . . . . 145
127    LDGT2 CO Emission Rates by Model Year at 50,000 Miles. . . . . . . . . . . . . 146
128    LDGT2 CO Emission Rates by Model Year at 100,000 Miles . . . . . . . . . . . . 146
129    LDGT2 NOx Emission Rates by Model Year at 50,000 Miles . . . . . . . . . . . . 147
130    LDGT2 NOx Emission Rates by Model Year at 100,000 Miles. . . . . . . . . . . . 147
131    EPA Low Speed Cycle (2.5 mph). . . . . . . . . . . . . . . . . . . . . . . . . 150
132    EPA Low Speed Cycle (3.6 mph). . . . . . . . . . . . . . . . . . . . . . . . . 150
133    EPA Low Speed Cycle (4.0 mph). . . . . . . . . . . . . . . . . . . . . . . . . 151
134    New York City Cycle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
135    Speed Correction Cycle 12. . . . . . . . . . . . . . . . . . . . . . . . . . . 152
136    Speed Correction Cycle 36. . . . . . . . . . . . . . . . . . . . . . . . . . . 152
137    Highway Fuel Economy Test. . . . . . . . . . . . . . . . . . . . . . . . . . . 153
138    CARB High Speed Cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
139    Comparison of MOBILE4.1 Idle and Running Exhaust CO
       Rates at 50,000 Miles (LDGV, 75F, FTP Bag Splits) . . . . . . . . . . . . . . 158
140    Comparison of MOBILE4.1 Idle and Running Exhaust HC
       Rates at 50,000 Miles (LDGV, 75F, FTP Bag Splits) . . . . . . . . . . . . . . 158
141    Comparison of MOBILE4.1 Idle and Running Exhaust NOx
       Rates at 50,000 Miles (LDGV, 75F, FTP Bag Splits) . . . . . . . . . . . . . . 159
142    Comparison of MOBILE4.1 Idle and Running Exhaust CO
       Rates at 50,000 Miles (LDGT1, 75F, FTP Bag Splits). . . . . . . . . . . . . . 162
143    Comparison of MOBILE4.1 Idle and Running Exhaust HC
       Rates at 50,000 Miles (LDGT1, 75F, FTP Bag Splits). . . . . . . . . . . . . . 162
144    Comparison of MOBILE4.1 Idle and Running Exhaust NOx
       Rates at 50,000 Miles (LDGT1, 75F, FTP Bag Splits). . . . . . . . . . . . . . 163
145    Comparison of MOBILE4.1 Idle and Running Exhaust CO
       Rates at 50,000 Miles (LDGT2, 75F, FTP Bag Splits). . . . . . . . . . . . . . 166
146    Comparison of MOBILE4.1 Idle and Running Exhaust HG
       Rates at 50,000 Miles (LDGT2, 75F, FTP Bag Splits). . . . . . . . . . . . . . 166
147    Comparison of MOBILE4.1 Idle and Running Exhaust NOx
       Rates at 50,000 Miles (LDGT2, 75F, FTP Bag Splits). . . . . . . . . . . . . . 167
148    Comparison of MOBILE5a Idle and Running Exhaust CO
       Rates at 50,000 Miles (LDGV, 75F, FTP Bag Splits) . . . . . . . . . . . . . . 170
149    Comparison of MOBILE5a Idle and Running Exhaust HC
       Rates at 50,000 Miles (LDGV, 75F, FTP Bag Splits) . . . . . . . . . . . . . . 170
150    Comparison of MOBILE5a Idle and Running Exhaust NOx
       Rates at 50,000 Miles (LDGV, 75F, FTP Bag Splits) . . . . . . . . . . . . . . 171
151    Comparison of MOBILE5a Idle and Running Exhaust CO
       Rates at 50,000 Miles (LDGT1, 75F, FTP Bag Splits). . . . . . . . . . . . . . 174
152    Comparison of MOBILE5a Idle and Running Exhaust HC
       Rates at 50,000 Miles (LDGT1, 75F, FTP Bag Splits). . . . . . . . . . . . . . 174
153    Comparison of MOBILE5a Idle and Running Exhaust NOx
       Rates at 50,000 Miles (LDGT1, 75F, FTP Bag Splits). . . . . . . . . . . . . . 175
154    Comparison of MOBILE5a Idle and Running Exhaust CO
       Rates at 50,000 Miles (LDGT2, 75F, FTP Bag Splits). . . . . . . . . . . . . . 178
155    Comparison of MOBILE5a Idle and Running Exhaust HC
       Rates at 50,000 Miles (LDGT2, 75F, FTP Bag Splits). . . . . . . . . . . . . . 178
156    Comparison of MOBILE5a Idle and Running Exhaust NOx
       Rates at 50,000 Miles (LDGT2, 75F, FTP Bag Splits). . . . . . . . . . . . . . 179

                                          -xiii-

                                      LIST OF TABLES

Table                                                                                Page

1    Required and Optional Inputs for the MOBILE Models . . . . . . . . . . . . . . . . 4
2      Emission Regime Cutpoints for 1983
       and Subsequent Model Years . . . . . . . . . . . . . . . . . . . . . . . . . . .24
3      Tech IV Model Hydrocarbon Emission Rates
       of Regimes for 1983+ Model Years . . . . . . . . . . . . . . . . . . . . . . . .24
4      Technology Distribution by Model Year Used
       in MOBILE4 (Passenger Cars, LDGV). . . . . . . . . . . . . . . . . . . . . . . .28
5    MOBILE5 LDGV Base Emission Rate Equations. . . . . . . . . . . . . . . . . . . . .31
6      LDGV Exhaust Standards Summary and
       Predominant Emission Control Technology. . . . . . . . . . . . . . . . . . . . .32
7      Light-Duty Vehicle Evaporative Emission Standards. . . . . . . . . . . . . . . .64
8      1981+ LDGV Non-Tampered Emission Equations Used in MOBILE4 . . . . . . . . . . .68
9      Tampered Vehicles Running Losses . . . . . . . . . . . . . . . . . . . . . . . .77
10     MOBILE5a Parameters Used to Model the Performance Standard
       Targets for Basic and Enhanced I/M Programs. . . . . . . . . . . . . . . . . . .91
11     Basic and Enhanced I/M Performance Standards Based on
       MOBILE5a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92
12     Comparison of Model Year Contribution to the Fleet-Average
       LDGV Emission Rate for Calendar Year 1995. . . . . . . . . . . . . . . . . . . 109
13     Applicability of MOBILE Model Algorithms
       and Options by Vehicle Class . . . . . . . . . . . . . . . . . . . . . . . . . 112
14     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE4.1; No I/M; LDGV, LDGT1, LDGT2
       Calendar Years 1985, 1990, and 1995. . . . . . . . . . . . . . . . . . . . . . 184
15     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE4.1; No I/M; LDGV, LDGT1, LDGT2
       Calendar Years 2000, 2005, and 2010. . . . . . . . . . . . . . . . . . . . . . 185
16     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE4.1; Basic I/M; LDGV, LDGTI, LDGT2
       Calendar Years 1985, 1990, and 1995. . . . . . . . . . . . . . . . . . . . . . 188
17     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE4.1; Basic I/M; LDGV, LDGT1, LDGT2
       Calendar Years 2000, 2005, and 2010. . . . . . . . . . . . . . . . . . . . . . 189
18     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE5a; No I/M; LDGV, LDGT1, LDGT2
       Calendar Years 1985, 1990, and 1995. . . . . . . . . . . . . . . . . . . . . . 192
19     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE5a; No I/M; LDGV, LDGTI, LDGT2
       Calendar Years 2000, 2005, and 2010. . . . . . . . . . . . . . . . . . . . . . 193
20     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE5a; Basic I/M; LDGV, LDGT1, LDGT2
       Calendar Years 1985, 1990, and 1995. . . . . . . . . . . . . . . . . . . . . . 196
21     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE5a; Basic I/M; LDGV, LDGTI, LDGT2
       Calendar Years 2000, 2005, and 2010. . . . . . . . . . . . . . . . . . . . . . 197
22     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE5a; Enhanced I/M; LDGV, LDGT1, LDGT2
       Calendar Years 1985, 1990, and 1995. . . . . . . . . . . . . . . . . . . . . . 200
23     Model-Year Group Contribution to the Fleet-Average Emission Rate
       MOBILE5a; Enhanced I/M; LDGV, LDGT1, LDGT2
       Calendar Years 2000, 2005, and 2010. . . . . . . . . . . . . . . . . . . . . . 201

                                           -xiv-

                                     EXECUTIVE SUMMARY


With the passage of the Clean Air Act Amendments (CAAA) of 1990,
renewed effort to properly account for the emissions characteristics
of on-road motor vehicles was initiated.  As part of this effort, the
U.S. Environmental Protection Agency's (EPA's) motor vehicle emission
factor model, MOBILE4, has subsequently been revised twice (i.e.,
MOBILE4.1 and MOBILE5) to improve the.predictive capability of the
model and to incorporate other CAAA directives aimed at reducing
emissions from onroad motor vehicles.  In addition, the 1990 CAAA and
the Intermodal Surface Transportation Efficiency Act (ISTEA) of 1991
call for a more active role from the U.S. Department of
Transportation (DOT) in reviewing local area transportation plans and
making determinations regarding whether these plans are consistent
with, and conform to, the State Implementation Plan.

Title I of the CAAA outlines the requirements for conformity
determinations.  Specifically, air pollutant emissions occurring as
a result of changes to an area's transportation network cannot:

             cause or contribute to new violations of the national
              ambient air quality standards,

             increase the frequency or severity of violations, or

             delay attainment of the standards or any required interim
              emission reductions.


All federally funded or approved transportation projects must meet
these so-called "conformity requirements," and the DOT, in
conjunction with local metropolitan planning organizations, is
responsible for making such a determination.  EPA's MOBILE model is
the analytical tool by which the emissions impacts are estimated and
the conformity determination is made.


OVERVIEW

Because the MOBILE model must be used in developing the on-road motor
vehicle emissions estimates required to make conformity
determinations, DOT requires a better understanding of its structure
and operation.  To that end, Sierra Research (Sierra) was retained to
document and evaluate the changes among the MOBILE4, MOBILE4.1, and
MOBILE5 versions of the model.  This consisted of reviewing
succeeding variations to the model to identify changes to existing
components (e.g., speed corrections, temperature corrections,
deterioration rates, etc.), as well as to evaluate the addition of
new components (e.g., modeling of oxygenated

                                           -xv-

fuels, resting loss emissions, I/M program modifications, etc.). In
addition, the impacts of model changes on emission factor estimates
were quantified under a variety of vehicle operating conditions
(e.g., speed, temperature, hot/cold start mix, I/M, etc.).

The approach that Sierra took in performing this evaluation consisted
of first identifying and defining the primary components of MOBILE4. 
This provided a basis of comparison from which succeeding versions of
the model were compared.  The information sources utilized by Sierra
primarily consisted of the FORTRAN source code and the materials
distributed at the workshops held by EPA prior to the release of each
model version. (Three workshops were held for MOBILE4; MOBILE4.1
received a single workshop; and two workshops were conducted for
MOBILE5.) For cases in which additional detail was needed, EPA's
Office of Mobile Sources provided valuable assistance.

DIFFERENCES AMONG MOBILE REVISIONS

MOBILE4, MOBILE4.1, and MOBILE5 are based on the same programming
structure, but new data and algorithms have been added in succeeding
versions to better represent emissions from in-use motor vehicles. 
Generally, revisions to the model have been incorporated either (1)
to account for compilation of additional test data, or (2) to add new
features and options.  The summary below describes some of the more
substantive changes that have occurred among the three MOBILE
releases investigated in this study.

MOBILE4 - This version of the model was released in February 1989,
and updated the previous MOBILE3 model.  Some of the added data and
features included the following.

       New Data

             Basic emission rates were updated.

             Tampering and misfueling rates were updated.

             Temperature and speed correction factors were updated.

       New Features

             The evaporative emissions component of the model was
              completely revised to account for minimum temperature,
              maximum temperature, fuel weathering, and RVP.  Further,
              running loss emissions were included for the first time.

             Higher deterioration rates for light-duty vehicles were
              assumed for mileages above 50,000.


MOBILE4.1 - Section 130 of the 1990 CAAA directed EPA to review and
revise, if necessary, the methods used to estimate emissions of
carbon monoxide, volatile organic compounds, and oxides of nitrogen. 


                                           -xvi-

As a result of this directive, MOBILE4.1 was released in November
1991.  Some of the more significant changes to the model are listed
below.

       New Data

             Basic emission rates were updated.

             Tampering rates were updated.

             Speed correction factors were updated, particularly for
              speeds above 48 mph.

             Registration data were expanded from 20 to 25 model years.

             The evaporative emissions data base was substantially
              enhanced with data stratified according to whether EPA's
              functional evaporative pressure/purge test was passed.

       New Features

             Modeling the effects of oxygenated fuels on CO emissions
              was included.

             Reporting of hydrocarbon results was expanded (i.e., THC,
              NMHC, VOC, TOG, or NMOG).

             Evaporative resting losses were included.

             Modeling of Tier I CO and cold-temperature CO emission
              standards was included.

             Modeling the effects of a transient, chassis-based I/M
              program was included (i.e., IM240).

             Modeling the effects of a functional evaporative test
              (i.e., pressure/purge) was included.


MOBILE5 - Initially released in December 1992, the MOBILE5 model
included the capability to model the motor vehicle requirements
called for in the 1990 CAAA. (In March 1993, MOBILE5a was released. 
This version was essentially the same as MOBILE5, incorporating some
minor corrections to the model.) The primary features associated with
MOBILE5 include the following:

       New Data

             Basic emission rates were updated, but they were based on
              IM240 data.  This represented a significant departure from
              EPA's historical method of developing base emission rates
              from surveillance program vehicles.

             Additional evaporative test data were used to update
              emission rates for diurnal, hot soak, resting, and running
              losses.


                                          -xvii-

       New Features

             Modeling of "industry average" and reformulated gasoline
              was included.

             Tier I and heavy-duty NOx emission standards were taken
              into account.

             Modeling of Low-Emission Vehicles (LEVs) was included.

             Modeling the effects of the new evaporative standards and
              test procedures was included.

COMPARISON OF MODEL OUTPUT

In general, the hydrocarbon and carbon monoxide model output from the
MOBILE4 and MOBILE4.1 versions is reasonably similar for equivalent
model conditions of speed, temperature, etc., but the NOx results
differ.  All emissions estimates from the MOBILE5 release are
substantially different compared to MOBILE4 and MOBILE4.1. This is
seen graphically in Figures 1, 2, and 3 for hydrocarbons, carbon
monoxide, and NOx, respectively. (The model runs upon which these
figures were based assumed no inspection and maintenance, an average
speed of 19.6 mph, 75F temperature, and operating mode splits based
on the Federal Test Procedure.)


Click HERE for graphic.

                                          -xviii-

Click HERE for graphic.


Click HERE for graphic.

Although many model changes account for the differences in fleet-
average emission rates observed in Figures 1 through 3, one of the
primary factors is revisions made to the base emission rate
equations.  These are shown for light-duty gasoline vehicles in
Figures 4 to 6 for HC, CO, and NOx, respectively.  These figures
indicate that a much higher base emission rate is being predicted in
MOBILE5 compared to the previous versions (this change is primarily
the result of EPA's use of IM240 data to represent the in-use fleet). 
It is interesting to note, however, that the MOBILE5 emissions
estimates are reduced substantially when accounting for the impact of
an enhanced I/M program containing chassis based emission testing and
EPA's recommended functional evaporative system check. (This point is
expanded upon in Section 10.)



Click HERE for graphic.

                                           -xx-

Click HERE for graphic.


Click HERE for graphic.


                                           -xxi-

Another change occurring in the MOBILE5 model that may have
implications for conformity analyses is the shape of the fleet-
average Nox emission rate as a function of vehicle speed.  This is
illustrated in Figure 7. The MOBILES emission rate, shown by the top
line in the figure, is predicted to Increase as speed increases above
20 mph.  Historically, emissions have been assumed to decrease
throughout this range until higher speeds were obtained (e.g., refer
to the MOBILE4.1 results in Figure 7).  However, the MOBILE5 results
indicate that traffic control programs that increase average vehicle
speed (e.g., from 35 mph to 40 mph) will also increase NOx emissions.



Click HERE for graphic.

Finally, idle emission estimates have also changed as the MOBILE
models have been revised.  Although MOBILE5 and MOBILE5a were
released with the idle emissions algorithm disabled, EPA has
subsequently issued guidance on how to convert MOBILE5a gram per mile
output into gram per hour idle emission rates. (That methodology is
reviewed in Section 5.6 of this report.) A comparison of MOBILE4.1
and MOBILE5a fleet-average idle CO emission rates (at 75F and under
a no-I/M scenario) is presented in Figure 8, which indicates a
significant increase in idle CO emissions when using MOBILE5a,
particularly beyond 1995.  That increase is attributable to the same
model revisions that caused increases in the running exhaust emission
rates in MOBILE5a, e.g., increased base emission rates, corrections
to account for in-use fuel effects, and increased mileage
accumulation rates.


                                          -xxii-


Click HERE for graphic.



Although many changes have occurred among the model revisions, there
are several areas that have remained relatively constant.  For
example, temperature correction factors have not been revised since
MOBILE4, and modeling of heavy-duty vehicle exhaust emissions has
also not been significantly revised since MOBILE4.


                                          -xxiii-

1. INTRODUCTION


The estimation of emissions from on-road motor vehicles is important
from a number of viewpoints.  First, these rates are used to develop
regional emission inventories which give an indication of progress
made toward meeting (or maintaining compliance with) ambient air
quality standards.  Second, they are important in determining where
efforts should be placed to control air pollution in a community. 
Finally, on-road motor vehicle emission estimates are used to
determine if regional transportation plans and projects are
consistent with, and conform to, the State Implementation Plan.

Because of the need to accurately predict motor vehicle emission
rates, the U.S. Environmental Protection Agency (EPA) expends
considerable resources on data collection programs that help quantify
the rate at which pollutants are emitted by individual categories of
motor vehicles under a variety of operating conditions.  The data
collected through these efforts have been used to develop and
continually update a computer model that estimates emission rates
from the fleet of motor vehicles in a given community.  That model,
termed "MOBILE", accounts for the effects of a variety of local
parameters (such as vehicle mix, speed, temperature, control
programs, etc.) on the mass of pollutants emitted from on-road motor
vehicles .

EPA has released several versions of MOBILE over the past few years:

             MOBILE4 - released February 1989,

             MOBILE4.1 - released November 1991,

             MOBILE5 - released December 1992, and

             MOBILE5a -  released March 1993.

The model has generally been updated to incorporate new data that
better represent emissions from in-use vehicles; however, the
MOBILE4.1 and MOBILE5 versions also addressed motor vehicle
requirements stemming from the Clean Air Act Amendments of 1990.
(MOBILE5a, a variant of MOBILE5, was prepared to correct errors that
were found in MOBILE5 subsequent to its release.  In this report,
much of the analysis was based on MOBILE5.  However, in cases where
MOBILE5 and MOBILE5a differ, an effort was made to present the
MOBILE5a results.)


                                            -1-

1.1 SUMMARY OF MOBILE COMPONENTS*

The primary components of the MOBILE emission factor models include
the base emission factors, the effect of local conditions (e.g.,
temperature and vehicle speed), characterization of the vehicle
fleet, the impact of fuel characteristics, and the effect of
inspection and maintenance programs.  None of these factors are 
static: technology is continually evolving, leading to changing in-
use emission performance, while changes in economic conditions can
lead to changes in vehicle sales and travel patterns.  As alluded to
above, EPA expends considerable effort to quantify and stay current
with the influence of all of these factors on motor vehicle emission
levels.  The key factors in the MOBILE models are discussed below.

Emission Factors - Also known as "base emission rates," these factors
are developed from test measurements of in-use vehicles at various
odometer readings.  The emission factors are represented by two
components: a zero-mile level (or intercept) and a deterioration rate
(or slope).  The zero-mile level represents the new-vehicle emission
rate, while the deterioration rate depicts emission control system
deterioration that takes place as the vehicle ages.

Test Conditions - Standardized test procedures have been developed
(e.g., the Federal Test Procedure, or FTP) to measure emission rates
from motor vehicles.  These procedures include specific driving
cycles (i.e., speed versus time profiles), temperatures, vehicle
load, and starting conditions.  Although the test procedures were
developed from data intended to represent average urban driving
conditions, they do not necessarily match those that vehicles
experience in each community.  Therefore, EPA has developed
correction factors to account for differences between the test
procedures and actual operating conditions.

Fleet Characteristics - The base emission rates represent the average
emission level of each model year in the vehicle fleet for each
vehicle class (e.g., light-duty vehicles versus heavy-duty vehicles). 
A fleet average emission rate incorporating the contribution of all
model years and vehicle classes is the output from the MOBILE model. 
The age distribution, the rate of mileage accumulation, and the mix
of travel experienced by the vehicle classes considered in MOBILE can
all influence the contribution of each vehicle class to the fleet
average emission rate.  Although the MOBILE models utilize national
average data as default values for these parameters, local data can
be input by the user to tailor a run for a specific community and
provide a more accurate estimate of emissions.

Fuel Characteristics - Emission test measurements are normally
conducted on a standardized test fuel known as Indolene.  The
characteristics of this fuel are well defined and ensure that test
results are repeatable.  However, in-use fuels are generally much
different than Indolene, and


______________________________

* Much of Section 2.1 is patterned after a similar discussion
contained in EPA'S, "Procedures for Inventory Preparation.  Volume
IV: Mobile Sources," dated July 1989.


                                            -2-

differences in fuel volatility and other fuel parameters (e.g.,
oxygenate content) influence both evaporative and exhaust emission
rates.  All MOBILE versions reviewed in this work require fuel
volatility as an input.  Additionally, MOBILE4.1 provided an option
to model the effects of oxygenated fuels on carbon monoxide
estimates, and MOBILE5 includes the ability to model the impact of
reformulated and oxygenated gasolines on hydrocarbon and oxides of
nitrogen emissions as well.

Emission Control Programs - The model-year-specific emission factor
equations are based on test data that do not include the effects of
local emission control program (i.e., inspection and maintenance
(I/M) and anti-tampering programs).  These programs are intended to
reduce emissions from in-use vehicles, and differences in program
design (e.g., annual versus biennial testing) can have a significant
impact on their effectiveness.  Thus, MOBILE contains provisions for
identifying the specific parameters applicable to the program being
modeled.

1.2 USER INPUTS

As alluded to above, there are a number of required and optional user
inputs to the MOBILE models that allow the user to account for
regional differences in travel parameters, ambient conditions,
enforcement programs, etc.  These are listed in Table 1, and briefly
described in the following discussion.

Required Inputs - To run the MOBILE models, a number of local
conditions are required that describe the travel parameters, ambient
conditions, and fuel parameters.  These include the following:

             Volatility class - Although not required in MOBILE4.1, this
              parameter is required in the MOBILE4 and MOBILE5 versions. 
              It was used in MOBILE4 as a surrogate for fuel volatility
              in some calculations, 1* and MOBILE5 requires it for
              modeling the effects of reformulated gasoline.

             Temperature - Because emissions are a strong function of
              ambient temperature, minimum, maximum, and average daily
              temperatures are required.

             Reid Vapor Pressure (RVP) - The fuel volatility (measured
              as RVP in pounds per square inch) also is an important
              parameter in emissions calculations.  The RVP has a
              significant influence on evaporative emissions (higher fuel
              volatility translates into higher evaporative emissions),
              and it also impacts exhaust emission estimates.

________________________

* Superscripts denote references provided in Section 11.

                                            -3-
Click HERE for graphic.

             Region - The region (i.e., low or high altitude) must also
              be input by the user.  Because vehicles (particularly older
              vehicles with mechanically based fuel delivery systems)
              generally run rich at high altitudes (i.e., more fuel is
              introduced into the combustion chamber than can be
              completely burned by the available oxygen), hydrocarbon and
              carbon monoxide emissions are magnified.

             Calendar year - The calendar year of evaluation must be
              specified by the user.  Because of increasingly stringent
              motor vehicle emission standards, the fleet-average
              emission rate generally decreases as future years are
              specified.

             Average speed - Speed also plays an important role in
              estimating vehicle emissions.  Because emissions are
              reported in grams/mile, lower speeds (i.e., below 20 mph)
              result in higher emissions (i.e., it takes a longer time to
              cover the same distance).  At high speeds (above 48 mph in
              MOBILE4.1; above 55 mph in MOBILE5), emissions are also
              predicted to increase.

             Operating mode - As will be described in the next section,
              the condition (i.e., cold start, hot start, or stabilized)
              under which the vehicle is operating has a significant
              impact on vehicle exhaust emissions.  For example, exhaust
              hydrocarbon emissions can be several times higher while the
              vehicle is warming up compared to those under stabilized
              operation.


Optional Inputs - In addition to the above required inputs, the
MOBILE models allow for a number of optional inputs that better
describe the locality being modeled; these are also shown in Table 1.
Among the more important optional inputs are the following:

             Registration distribution - Because the age of the vehicle
              fleet is an important parameter in determining the fleet-
              average emission rate, many local-level analyses make use
              of this option.  Generally, these data are readily
              available from state Departments of Motor Vehicles.

             Inspection and maintenance (I/M) programs - If an area has
              an operating I/M program, the effects of this can, and
              should, be modeled.  The impact can be quite significant,
              particularly for transient, loaded mode programs which can
              be modeled by MOBILE5.

             Anti-tampering programs - I/M programs often include a
              visual check of emission control system components to
              assure the vehicle owner has not disabled or otherwise
              tampered with the system.  The MOBILE models allow for the
              impact of these programs to be modeled.
             Refueling emissions - Although many air pollution control
              districts consider refueling emissions (i.e., emissions
              that occur when a fuel tank is filled) to be stationary
              source emissions, the MOBILE models are capable of modeling
              this process.

                                            -5-
             oxygenated and reformulated fuels - Fuels containing
              oxygenates (e.g., ethanol) result in significant reductions
              in exhaust hydrocarbon and carbon monoxide levels. 
              Additionally, the reformulated gasoline requirements
              contained in the 1990 CAAA result in decreased hydrocarbon
              emissions.  MOBILE4.1 is capable of modeling the impact of
              oxygenated fuels on carbon monoxide emissions, while
              MOBILE5 has the capability of modeling the effects of
              oxygenated fuels and reformulated gasolines on hydrocarbon,
              carbon monoxide, and oxides of nitrogen emissions.

1.3 MODEL OUTPUT

The MOBILE output consists of exhaust hydrocarbon (HC), carbon
monoxide (CO), and oxides of nitrogen (NOx) emission rates (in
grams/mile [g/mi]) for eight separate vehicle categories:

             LDGV light-duty gasoline vehicles (i.e., passenger cars),

             LDGTL light-duty gasoline trucks (under 6000 lbs. gross
              vehicle weight),

             LDGT2 - light-duty gasoline trucks (6000 to 8500 lbs. gross
              vehicle weight),

             HDGV - heavy-duty gasoline vehicles (over 8500 lbs. gross
              vehicle weight),

             LDDV - light-duty Diesel vehicles (i.e., passenger cars),

             LDDT - light-duty Diesel trucks (under 8500 lbs. gross
              vehicle weight),

             HDDV - heavy-duty Diesel vehicles (over 8500 lbs. gross
              vehicle weight), and

             MC - motorcycles.


In addition, evaporative HC emissions are reported as g/mi or
grams/event (e.g., grams per hot soak).

A sample output file from a MOBILE5a run for the year 2000 is given
in Figure 9. For this run, two different biennial I/M programs were
specified: a 2500 rpm/idle test for 1968 to 1985 model year LDGV,
LDGT1, and LDGT2; and an IM240 test (i.e., a chassis-based transient
test) for 1986 and subsequent model year LDGV, LDGTI, and LDGT2.  In
addition, a functional check of the evaporative emission control
system was specified, as was an ATP.  The emission results are
reported for each vehicle type separately, and a composite rate for
all vehicles is also computed by the model and included in the
output.  Note that the "Evaporat HC" value reported in g/mi includes
hot soak, diurnal, and crankcase emissions that have been converted
to a g/mi basis.


                                            -6-

Click HERE for graphic.

1.4 ORGANIZATION OF THE REPORT

Following this introduction, Section 3 contains an overview of motor
vehicle emission modes (e.g., cold start, hot start, etc.) which
introduces the reader to the processes that lead to emissions from
motor vehicles.  This background information serves as a basis from
which modeling of these modes can be discussed.  Section 4 is a
summary of the standard modeling approach used by the MOBILE models. 
Information on fleet make-up is introduced in this section.  Exhaust
emissions modeling is detailed in Section 5. This includes a
discussion of standard test procedures as well as the development of
base emission rates and various correction factors.  Considerable
time is spent in comparing how these parameters have changed among
revisions to the model.  Evaporative emissions are treated in Section
6, which details the modeling methodologies used in the MOBILE models
and makes comparisons among MOBILE revisions.  Because of its
increasing importance as an air pollution control.strategy, modeling
of I/M programs is discussed separately in Section 7. The main focus
of this discussion is on the methodology utilized in MOBILE5 to model
the impact of I/M.  The Clean Air Act Amendments of 1990 contained a
number of features related to the control of on-road motor vehicles
(e.g., more stringent emissions standards, reformulated gasoline). 
The methodologies by which these requirements are modeled are
discussed in Section 8. Section 9 contains a summary of emissions
versus vehicle age which provides insight regarding the fraction of
emissions attributable to older versus newer vehicles.  That fraction
has ramifications in terms of determining where efforts should be
placed to control emissions from motor vehicles (i.e., should the
focus be on new vehicle standards or in-use control programs). 
Finally, Section 10 summarizes how the changes incorporated in
MOBILE4, MOBILE4.1, and MOBILE5 resulted in differing emissions
estimates for on-road motor vehicles. (MOBILE5a was used in making
these comparisons.)


                                            -8-

2. DEFINITION OF MOTOR VEHICLE EMISSION MODES


Motor vehicle emissions consist of a large number of chemical species
that primarily result from combustion within the engine and from fuel
evaporation at various locations throughout the fuel delivery and
storage system.  Three particular emission components are modeled by
the MOBILE models: hydrocarbons (HC), carbon monoxide (CO), and
oxides of nitrogen (NOx).  These are the emission components that
result in the two major nonattainment pollutants related to mobile
sources: ozone and carbon monoxide.

The quantification of emissions involves determining emission rates
for two fundamentally different types of emission-producing
processes: emittance from the vehicle's exhaust system, and
evaporation from the fuel storage and delivery system.  Emissions
from each of these basic types of emission-producing processes,
exhaust and evaporative, can be further categorized, as discussed
below.

2.1 EXHAUST EMISSIONS

Cold Start - Under cold start conditions, the vehicle engine has been
turned off for some time and the catalytic converter (if the vehicle
is so equipped) is cold.  HC and CO emissions are higher when a cold
engine is first started than after the vehicle is warmed up.  This is
because catalytic emission control systems do not provide full
control until they reach operating temperature (i.e., light-off), and
a richer fuel air mixture must be provided to the cylinders under
cold operating conditions to achieve satisfactory engine performance
(e.g., startability and driveability). (EPA considers a cold start
for a catalyst-equipped vehicle to occur after the engine has been
turned off for one hour.  For non-catalyst vehicles, a four-hour
engine-off period distinguishes a cold start.) Rich mixtures are
necessary to achieve smooth combustion during warmup because gasoline
does not fully vaporize and mix with the air in a cold engine.  Extra
fuel is added to ensure that an adequate amount of fuel is vaporized
to achieve a combustible mixture.  Complete vaporization eventually
occurs in the engine cylinder as a result of the high temperatures
created by combustion.  However, the excess fuel that was needed to
ensure adequate vaporization to start the combustion process cannot
be completely burned due to a lack of sufficient oxygen in the
cylinder.  The result is that partially burned fuel and unburned fuel
are emitted in relatively high concentrations from a cold engine. 
Elevated emissions of these pollutants in this cold transient phase
occur from the time a cold engine starts until it is fully warm. 
While engine-out NOx emissions tend to be low during rich operation
of a cold



                                            -9-

engine, the lack of catalyst activity to control this pollutant
results in elevated cold start NOx emissions as well.

Hot Start - Under hot start conditions, the vehicle engine has been
turned off for such a short time that the catalyst has not had time
to cool to ambient temperature.  Thus, the warm-up period is shorter
(if present at all) than that required under cold start conditions. 
For that reason, HC and CO hot start emissions are significantly
lower than under cold start operation.  Under the standard test
procedure used by EPA, a "hot start" is a test that begins exactly 10
minutes after a fully warmed up engine has been shut off.  After only
10 minutes, no mixture enrichment is required to achieve a reliable
re-start and the catalyst is usually still above its "light-off"
temperature.

Hot Stabilized - After warmup has occurred, and the engine and
emission control systems have reached full operating temperatures,
the vehicle is considered to be in the hot stabilized mode. 
Generally, emissions are relatively low (compared to cold start
emission rates) under hot stabilized conditions.  However, emissions
are also highly dependent on vehicle speed and engine load.  Recent
analysis by Sierra2 has revealed that varying off-cycle load
conditions (i.e., under conditions not tested on the standard
automotive driving cycle used for vehicle certification purposes) can
have a dramatic impact on emissions.

Idle Emissions - Although not generally considered for inventory
purposes, idle emissions may need to be considered for certain
transportation-related analyses.  The MOBILE models report idle
emissions in terms of grams/hour, and emissions are a function of
operating mode (i.e., stabilized or cold start) and temperature. 
Because many of the changes incorporated into the MOBILE5 model
(e.g., incorporation of Tier I emission standards, reformulated
gasoline, low emission vehicles) resulted in unreliable estimates of
idle emissions, the idle subroutines have been temporarily disabled
in MOBILE5.  When MOBILE5 was released, EPA recommended continued use
of MOBILE4.1 to estimate idle emission rates.  However, EPA has
recently developed a methodology to convert MOBILE5a gram/mile
running exhaust emission rates to gram/hour idle rates. (This
approach is discussed in Section 5.6.)

2.2 EVAPORATIVE EMISSIONS

Evaporative emissions consist entirely of hydrocarbon emissions. 
These emissions can be categorized into the six groups discussed
below.

Hot Soak - When a hot engine is turned off, fuel exposed to the
engine (e.g., in carburetor float bowls or in fuel injectors) may
evaporate and escape to the atmosphere.  These so-called "hot soak"
emissions are modeled by MOBILE as grams/event, which are then
converted within the model to a g/mi basis.

Diurnal - Diurnal temperature fluctuations occurring over a 24-hour
period cause "breathing" to occur at the gasoline tank vent.  To
prevent the escape of fuel vapor, the vent is routed to a charcoal
canister where the vapor can be adsorbed and later purged into the
running

                                           -10-

engine.  These emissions are calculated by the MOBILE models in terms
of grams/event, and, as with hot soak emissions, are then converted
to a g/mi basis.

Running Losses - Running loss emissions are those resulting from
vapor generated in gasoline fuel tanks during engine operation. 
Running losses are especially a problem on vehicles that have exhaust
systems in close proximity to the gasoline tank.  Running loss
emissions occur when the vapors emitted from the tank vent exceed the
rate at which they are being purged from the canister by the engine. 
This HC emission category had long been assumed to be insignificant,
but research over the past several years has shown this assumption to
be incorrect, and running losses have been included in EPA's emission
factors models since the MOBILE4 version (published in 1989).  These
emissions are calculated by the MOBILE models in terms of g/mi.

Resting Losses - Resting losses have only recently been included in
the MOBILE models, with the MOBILE4.1 version (published in 1991)
including resting losses for the first time.  EPA considers resting
losses to be "those emissions resulting from vapors permeating parts
of the evaporative emission control system (e.g., rubber vapor
routing hoses), migrating out of the carbon canister, or evaporating
liquid fuel leaks."3 EPA also states that a portion of what are now
considered resting losses was previously included in the hot soak and
diurnal categories.  Resting losses are dependent upon temperature
and the type of carbon canister that is used in the evaporative
emission control system (i.e., open-bottom versus closed-bottom). 
These emissions are calculated as grams/hour and are then converted
to g/mi.

Refueling Losses - There are two components of refueling emissions:
vapor space displacement and spillage.  As a fuel tank is being
refueled, the incoming liquid fuel displaces gasoline vapor that has
established a pseudo-equilibrium with the fuel in the tank,
effectively "pushing' the vapor out of the tank.  Spillage simply
refers to a small amount of fuel that is assumed to drip on the
ground and subsequently evaporate into the ambient air.  Refueling
emissions are calculated in terms of grams/gallon of dispensed fuel
and are converted to a g/mi basis.

Crankcase Emissions - Although not a true "evaporative" source,
crankcase emissions are generally considered in the evaporative
emissions category.  They are the result of defective positive
crankcase ventilation (PCV) systems that allow blow-by from the
combustion process (which is normally routed to the vehicle's intake
manifold) to escape to the atmosphere.  These emissions are modeled
as grams/mile.

                                           -11-

3. STANDARD MOBILE MODELING APPROACH

Emissions from each of the processes outlined in Section 3 are
estimated by EPA's MOBILE models separately for the eight vehicle
classes included in the models (i.e., LDGV, LDGTI, LDGT2, HDGV, LDDV,
LDDT, HDDV, and MC).  The emissions estimate is performed by first
determining the emission rate of each model year making up the
vehicle class, weighting the model-year-specific emission rate by the
fractional usage experienced by that model year (i.e., travel
fraction or VMT fraction), and summing over all model years that
comprise the vehicle class.  In addition, a variety of corrections
are applied to the base emission rates to account for conditions that
are not included in the standard test cycles used to develop the base
emission rates (e.g., exhaust emission rates may be corrected for
non-standard speeds, evaporative emissions may be corrected for non-
standard temperatures, etc.).

In equational form, the calculation can be described by:

                           n
              EFi,j,k =  Tpm * (BERj,k,m * CFj,k,m...)                        [4-1]
                           m-1


where EFi,j,k       =      fleet-average emission factor for calendar year
                           i, pollutant j, and process k (e.g., exhaust,
                           evap);
       TFm          =      fractional VMT (i.e., travel fraction)
                           attributed to model year m (the sum of TFm over
                           all model years n is unity);
       BERj,k,m     =      base emission rate for pollutant j, process k,
                           and model year m;
       CFj,k,n      =      correction factor(s) (e.g., temperature, speed)
                           for pollutant J, process k, model year m, etc;

and the sum is carried out over the n model years making up the
vehicle class (e.g., 20 years for LDGV-in MOBILE4, 25 years in
MOBILE4.1 and MOBILE5).

3.1 REGISTRATION DISTRIBUTION AND TRAVEL FRACTION

The registration distribution and resulting travel fraction has an
important impact on the calendar-year-specific, fleet-average
emission factors.  This is particularly apparent for cases in which
the model year emission factors undergo significant changes from one
year to the next (e.g., when new emission standards are implemented). 
A higher proportion of the travel fraction being allocated to newer
vehicles (with presumably lower emission rates) will result in a
lower fleetaverage emission rate, and thus, lower inventory
calculations.

                                           -13-

The methodology to calculate travel fractions has not changed among
the three versions of the MOBILE models being evaluated in this work. 
It is based on applying an estimated annual mileage accumulation rate
by vehicle age (determined from National Purchase Diary data) to an
estimated registration distribution (from Polk data) for the number
of model years assumed to comprise the fleet (i.e., 20 model years
for MOBILE4; 25 model years for MOBILE4.1 and MOBILE5).  The travel
fraction for each model year (TFm) is calculated from:

                             REGm * MILESm
              TFm   =      --n---------------
                               (REGm *MILESm)
                            m=1



where MILESm represents the annual mileage accumulation for model year
m, REGm. represents the registration fraction for model year m, and
n is the total number of model years in the fleet.  Prior to
performing the calculation, however, the registration and mileage
accumulation data ' are modified to reflect a January 1 analysis
date. (MOBILE4 and MOBILE4.1 calculated emissions on January 1 of the
calendar year of evaluation, whereas MOBILE5 also allows the option
of a July 1 evaluation date.  However, the MOBILE5 July 1 methodology
simply interpolates between two consecutive January 1 evaluations. 
Thus, the VMT distribution is still based on a January 1 evaluation
date, regardless of whether a July evaluation is requested.)

The registration distributions used to calculate the travel fractions
for LDGV in MOBILE4, MOBILE4.1, and MOBILE5 are shown in Figure 10.
(Note that the registration distributions hard-coded in the models
are on a July 1 basis.  Thus, optional area-specific registration
distributions, which can be input by the user, must also be on a July
1 basis.) As seen, the registration distribution for light-duty
vehicles (which includes Diesels in this figure) was substantially
modified for the MOBILE4.1 version.  The uneven distribution for
MOBILE4.1 and MOBILE5 reflects EPA's choice to use data that depict
actual sales fractions through the 1980s, rather than develop a more
generic curve that assumes a steady decline in population as vehicles
age.  Such an approach provides more reliable estimates of emissions,
as long as the historical sales peaks and valleys move as the
evaluation year changes.   This is not the case with MOBILE4.1 and
MOBILE5, as the same registration distribution is used for all
evaluation years.  Thus, future-year projections made with MOBILE4.1
and MOBILE5 assume an historical registration distribution. (EPA
understands this concern and is considering an alternative approach
to project future-year registration distributions that could be used
in conjunction with the model.)

Figure 11 illustrates the mileage accumulation rates used in the
MOBILE models.  As seen, mileage accumulation by vehicle age was
assumed to increase between MOBILE4.1 and MOBILE5, and the same
relative increase (i.e., 9.7 percent) was assumed for all vehicle
ages.  This change was made to update the MOBILE4 and MOBILE4.1
mileage accumulation rates, which were based on 1984 National
Purchase Diary (NPD) data, with 1990 data.  The methodology employed
by EPA consisted of comparing data from


                                           -14-


Click HERE for graphic.


Click HERE for graphic.


the Federal Highway Administration on total vehicle miles in 1984 and
1990.4 The results indicated a 9.7% increase in miles per vehicle
between 1984 and 1990.  Thus, this factor was applied to the mileage
accumulation rates.

Finally, Figure 12 shows the calculated travel fraction for the three
MOBILE versions.  MOBILE4.1 and MOBILE5 attribute a higher fraction
of VMT to newer vehicles than does MOBILE4.  As discussed above, this
can be important in terms of estimating the benefits of new control
measures.


Click HERE for graphic.


                                           -16-

                             4. EXHAUST EMISSION METHODOLOGIES


The cornerstone of exhaust emissions estimates using the MOBILE
models is the model-year-specific base emission rates.  For light-
duty vehicles, these emission rates (in grams/mile) are determined by
testing vehicles over a standardized test cycle on a chassis
dynamometer.  Because the standard test cycle does not include all
operating conditions that can be encountered, correction factors have
been developed that adjust the base emission rates for operating mode
fractions (i.e., percent of VMT spent in cold start, hot start, and
stabilized modes), temperatures, and speeds not represented by the
test cycle.  Presented below is an overview of the test cycle used to
generate exhaust emissions data, the analytical approach taken in
developing the model-year-specific base emission rates, and a
discussion of the various corrections that are applied to account for
non-standard conditions.  Although emissions modeling of heavy-duty
vehicles contains many of the same elements as that of light-duty
vehicles, the treatment is different enough to warrant a separate
subsection.  Thus, this section concludes with a review of the
methodologies employed by the MOBILE models to estimate emissions
from heavy-duty vehicles.

4.1 BACKGROUND

Standard Test Procedure - The basic exhaust emissions data contained
in the MOBILE models are based on a standardized driving cycle called
the Urban Dynamometer Driving Schedule (UDDS), or "LA4 cycle." This
chassis dynamometer test cycle involves duplicating a speed-time
profile from an actual road route identified in the Los Angeles area
in the late 1960s and chosen to represent the typical urban area
driving pattern.  The driving cycle was then incorporated into EPA's
Federal Test Procedure (FTP), the testing process used by all motor
vehicle manufacturers to certify that their vehicles are capable of
meeting federal emission standards.

The FTP consists of three distinct segments at a standard test cell
temperature of 68 to 86F.  Because the mass emissions from each of
the three segments are collected in separate tedlar bags, the three
operating modes are often referred to in terms of "bags." A complete
FTP is comprised of:

             a cold start (or cold transient) portion ("Bag 11), which
              is the first 3.59 miles of the UDDS (505 seconds in
              length);

             a stabilized portion ("Bag 2"), which is the final 3.91
              miles of
              the UDDS (867 seconds in length); and


                                           -17-

              a hot start (or hot transient) portion ("Bag 3"), which is
              the first 3.59 miles of the UDDS and follows an engine-off
              period of 10 minutes.


The LA4, shown in Figure 13, has been the standard driving cycle for
the certification of light-duty vehicles since the 1972 model year. 
It is approximately 7.5 miles in length with an average speed of 19.6
mph.  The cycle includes 18 segments of non-zero speed activity (at
varying engine loads) separated by idle periods.  Since the 1975
model year, the initial 505 seconds of the cycle have been repeated
following a 10 minute hot soak.  It is believed that after cold/hot
start weighting factors (i.e., 43 percent of the starts are assumed
to be cold starts, 57 percent are assumed to be hot starts) are
applied to the Bag 1 and Bag 3 results, the test provides a more
accurate reflection of typical customer service than running just one
7.5 mile cycle from a cold start.5 This change was incorporated
because a significant fraction of vehicle starts do not occur with
the vehicle in a completely cold condition, and running the first 505
seconds of the LA4 with the vehicle in a warmed-up condition gives an
indication of stabilized emissions over a higher speed cycle than
represented by Bag 2. The FTP composite emission rate is calculated
from the bag-specific emissions results according to the following
formula:

              3.59 * (0. 43 *BAG1 + 0.57 *BAG3)             3.91 *BAG2
BER =         ----------------------------------  +  ----------                [5-1]
                                 7.5                              7.5
which reduces to

       BER = 0.206 *BAG1 + 0.521 *BAG2 + 0.273 *BAG3                           [5-2]

where         BER   =      composite FTP base emission rate (g/mi),
              BAG1  =      bag 1 emission rate (g/mi),
              BAG2  =      bag 2 emission rate (g/mi), and
              BAG3  =      bag 3 emission rate (g/mi).


Corrections for Nonstandard Conditions - Because the emissions test
is performed over the same standard operating conditions for all
vehicles, the MOBILE emission factor models make use of a variety of
correction factors to tailor the FTP results to the specific local
conditions being modeled.  For example, operating mode correction
factors (also referred to as bag correction factors) are applied to
the FTP composite emission rate to determine the emission rate of a
vehicle in the cold start mode, stabilized mode, or hot start mode. 
Because emission rates from motor vehicles are much higher at very
low temperatures, temperature correction factors are used to account
for this effect.  Finally, a vehicle's emission rate is also a strong
function of its average speed.  Thus, speed correction factors modify
the FTP results to account for speeds different from the FTP average
of 19.6 mph.


                                           -18-



Click HERE for graphic.



The overall emission rate from a vehicle is the product of the basic
emission rate (determined from the FTP results) and a number of
correction factors that are specific to the conditions being modeled. 
Although the actual modeling procedure is very complex, this can be
simply represented by:

              EF = BER * BCF * TCF * SCF             [5-3]


where         EF    =      emission factor (g/mi) corrected for operating
                           mode, temperature, and speed;
              BER   =      composite FTP base emission rate (g/mi);
              BCF   =      bag (or operating mode) correction factor;
              TCF   =      temperature correction factor; and
              SCF   =      speed correction factor.


Each of these parameters is described below.

4.2 DATA USED FOR DEVELOPING BASE EMISSION RATES

As outlined above, the base emission rates form the foundation upon
which all ensuing calculations are based.  The base emission rate
equation is a linear function of vehicle mileage, with emissions



                                           -19-

increasing as the vehicle ages.  Thus, the BER is described by two
components: a zero-mile (ZM) level and a deterioration rate (DR). 
For some model years, pollutants, and vehicle types, however, EPA has
determined that the deterioration rate increases beyond 50,000 miles. 
In some cases, therefore, the BER equation has two deterioration
rates, termed DR1 and DR2.

There are many factors that must be considered when developing the
base emission rate equations, but primary among these is the
necessity that the base emission rates truly represent emissions from
in-use vehicles for the vehicle class and model year being analyzed. 
Because it is often not possible to test enough vehicles to obtain a
representative data set upon which the base emission rate equations
can be developed, various modeling strategies have been developed to
improve the predictive capability of these equations.  The following
discussion addresses some of the details that must be considered when
analyzing data for the development of the base emission rate
equations.

Data Sources - The raw data used to determine the base emission rates
(often referred to as "emission factors") are generally derived from
routine surveillance programs and occasional special studies
conducted by EPA.  The extent to which the data are adjusted prior to
their use can have a significant effect on the emission factors.  For
example, vehicles tested in surveillance programs are sometimes
subjected to pre-screening criteria that can "filter" the resulting
data set.  For example, CARB's practice of rejecting vehicles that
are considered "unsafe" for testing can eliminate a valid fraction of
the vehicle population that includes a relatively high fraction of
gross emitters. (However, EPA does not believe this to be a problem
with their data bases.) The selection of the results of a particular
test (or series of tests) for analysis can affect the final emission
factors.

Treatment of Gross Emitters - There is a clearly identifiable subset
of the vehicle population consisting of vehicles that, for a variety
of reasons, exhibit emission levels many times above the applicable
emissions standards.  Depending upon how this subset is defined
(e.g., vehicles with emissions more than ten times the standard), the
frequency of these vehicles in the overall population can be
relatively small.  If the frequency of these gross emitters is less
than 1 percent, then a total surveillance data set of even 5000
vehicles will include only 50 gross emitters spanning a wide range of
model years, emission control technologies, and age/odometer levels. 
The use of typical mobile source analytical techniques (such as
linear regressions of emissions vs. mileage, or population size vs.
mileage) can result in conclusions that depend heavily on the tests
of just a few vehicles.  That is because these vehicles exert a
disproportionately large influence on the fleet average emission
factors that are ultimately developed.

An alternative approach to estimating emissions from these vehicles
is to conduct a controlled series of tests on a small group of
vehicles, deliberately introducing component malfunctions and
determining their effect on vehicle emissions.  These data can then
be used with larger data sources (such as data from I/M programs)
that determine the frequency of these malfunctions to establish
emission factors for this small but important segment of the vehicle
population.  However, this


                                           -20-

approach can result in inaccuracies regarding synergistic and
antagonistic effects of multiple malfunctions.

For 1981 and subsequent model year vehicles, EPA utilizes a variant
of the latter approach by stratifying vehicles according to emission
level (i.e., 'emitter categories') and estimating the emission rate
and occurrence of each category as a function of vehicle mileage. 
This approach, embodied by the Tech IV6 and subsequent emission
factors models, is discussed in detail below.

Tampering - The portion of the vehicle fleet that has had some or all
of its emission control capabilities deliberately de-activated also
has a large effect on fleet emissions.  While the effect on emissions
is identical regardless of whether an EGR valve fails in the closed
position (no flow) or is disconnected, the different causal factors
can influence assumptions regarding the frequency of occurrence of
these events.  In addition, some components are more prone to
deliberate tampering (EGR valves, air injection pumps, catalysts),
while other components have problems that are more typically
associated with excessive wear or improper maintenance (evaporative
control systems, PCV valves, fuel system components).

The problems with the treatment of tampering are similar to those of
other gross emitters: a small number of vehicles can exert a
disproportionately large influence on fleet average emissions.  The
options for analyzing these vehicles include:

             implicitly including the vehicles in the overall
              surveillance population, with no separate explicit
              analysis;

             segregating the vehicle population based on whether
              tampering appears to have occurred, separately analyzing
              the two classes of vehicles, and combining the results of
              the analyses using assumed (or computed) frequencies; and

             eliminating vehicles that have been subjected to tampering
              from the surveillance vehicle population, and developing
              emission factors and frequency distributions for these
              vehicles based on special test programs and assumed (or
              computed) frequency distributions.


For the MOBILE models, EPA has chosen the third option and has
developed a "tampering offset" that is added to the base emission
rate equation prior to the application of correction factors.

4.3 TECH IV EMISSION FACTORS MODEL

As long as the data base used to develop the BER equations is
representative of the in-use fleet of vehicles, a simple regression
technique adequately models emissions as a function of vehicle
mileage.  However, EPA has found that the technology used to control
emissions from 1981 and later vehicles results in the random
occurrence of


                                           -21-

high-emitting vehicles that would not necessarily be properly
accounted for using the above methodology.6 In addition, a means to
account for the impact of an inspection and maintenance (I/M) program
was needed.  Thus, a separate model was developed that generates base
emission rates and I/M benefits for 1981 and later light-duty
vehicles.  This model is termed the 'Tech IV Credit Model" and is
described below. (This modeling procedure was used to determine only
HC and CO BER equations for MOBILE4 and MOBILE4.1, while NOx
emissions were determined by a simple weighting of regression
equations.  To estimate effects of I/M programs on NOx emissions,
MOBILE5 used the more complex process to calculate the NOx BER
equations as well.)

A flow chart outlining the Tech IV model, as it was applied in
developing the MOBILE4 base emission rate equations, is illustrated
in Figure 14. (The Tech IV model was modified in subsequent revisions
to the MOBILE model; however, the basic methodology as described here
has remained essentially unchanged.) In summary, the methodology
calls for stratification of FTP emissions data by model year,
technology group, and emission level (or 'regime").  The emission
rate (g/mi) for each emission regime (which is a function of
accumulated mileage and differs according to technology type) is then
multiplied by the fraction of the regimes making up the fleet (also
a function of vehicle mileage and technology type).  These values are
summed to arrive at the technology specific emission rates as a
function of mileage.  Model-year-specific emission rates (at each
mileage interval) are determined by weighting each technology group
by the anticipated sales fraction for each technology for the model
year of interest.  The above calculation is carried out at the
mileage intervals corresponding to 20 model years (25 model years for
MOBILE4.1 and MOBILE5), and the ZM, DR1, and DR2 components of the
base emission rate equation are determined from the resulting
emission values.

The 1981 and subsequent model year BER equations developed for
MOBILE4 were based on segregating the FTP emissions data into two
model-year groups (1981-1982 and 1983+), while three distinct
technology types were considered: carbureted closed-loop, fuel-
injected closed-loop, and open-loop (both carbureted and fuel-
injected).  These stratifications were chosen because, from an
engineering perspective, vehicles within each group should exhibit
similar emissions characteristics. (Note that in succeeding versions
of the model, the closed-loop fuel-injected category was further
stratified into multipoint (MPFI) and throttle-body (TBI) fuel
injection.)

The emission regimes utilized for the MOBILE4 BER equations were
defined as follows:

             "Normals" - emissions at or below the FTP certification
              level;

             "Marginals" - fail either HC or CO, but are not "Highs" or
              "Supers";

             "Highs" - either HC or CO emissions more than two standard
              deviations above the mean among vehicles with less than
              50,000 miles; and



                                           -22-

Click HERE for graphic.


                                           -23-

             "Supers" - emissions either above 10 g/mi HC or 150 g/mi CO
              for all model year groups and technology types.

The emission regime cutpoints are listed in Table 2 for the 1983+
model year group, and the HC emission rates for each regime (as a
function of mileage) are summarized in Table 3. As outlined in Table
3, the emission rate of the Super regime is the same for the


                                              Emission Level (g/mi)
Regime        Technology                      HC                    CO
Normal        All                             < 0.41              < 3.4
Moderate      All                       >0.41; < High             > 3.4; 0.815; 10.398; 0.965; 10.558; 0.837; 10.139;  10                       > 150
              Closed-Loop, FI

*No Supers are assumed for open-loop technology in TechIV.


Click HERE for graphic.
                                           -24-

closed-loop technologies and is independent of vehicle mileage
accumulation. (Because very few data points make up the Super regime,
the emissions were averaged and applied to both carbureted and fuel-
injected vehicles.) On the other hand, the remaining regime emission
rates differ according to technology and increase linearly with
mileage.  The zero-mile and deterioration rates for the Normal and
Marginal regimes were determined by performing least squares
regressions on data that fell within the regimes.  EPA assumed that
the Highs deteriorated at the same rate as the Marginal regime.

Once the emission rates of each regime are established as a function
of mileage, the incidence of emitter regimes (also as a function of
mileage) must be determined.  Data developed by EPA though its
emission test programs indicate that the fraction of non-Normal
regimes increases linearly with vehicle age (i.e., more vehicles are
in the high remitting regimes at higher mileages).  Thus, the
incidence of Moderates, Highs, and Supers is described by a zero-mile
level and a deterioration rate, while the Normal regime is determined
through subtraction (i.e., Normals = 1 - Moderates - Highs - Supers). 
Further, growth rate of the High emission category (expressed as the
fraction of vehicles expected in the fleet per 10,000 miles) is
assumed to increase by a factor of 3.103 beyond 50,000 miles. 

The incidence of emitter regimes (i.e., the fraction of the total
fleet) is shown graphically in Figures 15 and 16 for 1983+, closed-
loop, carbureted and fuel-injected vehicles, respectively.  As seen
in the figures, the assumed fraction of Highs in the fleet increases
very rapidly beyond 50,000 miles, and the percentage of Highs at a
given mileage interval is greater for carbureted vehicles than for
fuelinjected vehicles.  This last point is consistent with the
expectation that fuel-injected vehicles are more durable, from an
emissions perspective, than carbureted vehicles. (Note that the
Normal regime disappears beyond 70,000 miles.  At this point, the
Marginal regime growth equation is neglected, and the Marginal regime
is determined by subtracting the Highs and Supers, i.e., Marginals -
1 - Highs - Supers.  This approach avoids the sum of regime fractions
being greater than unity.)

Multiplying the emission rates - of each regime (Table 3) by the
regime size (Figures 15 and 16) results in a technology-specific
composite emission rate at each mileage interval.  This calculation
was performed for the 1983+, closed-loop, carbureted and fuel-
injected vehicles, and the results are presented in Figures 17 and
18.  As seen in the figures, the Normal regime is the largest
contributor to the overall emission rate at very low mileage, but at
high mileages, Highs and Supers are the predominant contributors. 
This is expected, given the greater emission rate and higher
incidence of these emission regimes as mileage increases.  It is
interesting to point out, however, that the contribution of the Super
regime is very significant, considering that it makes up such a small
fraction of the fleet (only 2.2 percent at 100,000 miles).  Thus,
errors in estimating the emission rate or regime size for the Super
emission category can have a dramatic impact on the overall emission
factor estimated by the Tech IV model.


                                           -25-


Click HERE for graphic.


Click HERE for graphic.



                                           -26-


Click HERE for graphic.


Click HERE for graphic.



                                           -27-

Finally, the model-year-specific emission rates (as a function of
mileage) are determined by weighting the technology-specific emission
rates by the appropriate technology split for the model year of
interest.  For MOBILE4, the technology splits shown in Table 4 were
used to establish the model-year-specific BER equations.  As an
example, the 1990 model-year HC BER equation was determined for
passenger cars using the methodology outlined above.  The emission
rate was determined at each mileage interval corresponding to vehicle
age.  This is depicted in Figure 19.  As seen, there is a change in
slope at 50,000 miles, which corresponds to the increase in the High
regime growth rate at 50,000 miles.  The data points shown in the
figure are used to determine the ZM, DR1, and DR2 values through a
regression technique.


Table 4. Technology Distribution by Model Year Used in MOBILE4
(Passenger Cars, LDGV)

                                              Technology Type*

Model Year          CLP-MPFI            CLP-TBI      CLP-CARB            OPLP

1981                0.057               0.027        0.635               0.281
1982                0.062               0.109        0.499               0.330
1983                0.086               0.195        0.478               0.241
1984                0.104               0.289        0.552               0.055
1985                0.280               0.265        0.393               0.062
1986                0.391               0.279        0.260               0.070
1987                0.483               0.264        0.239               0.014
1988                0.576               0.235        0.189               0.000
1989                0.594               0.243        0.163               0.000
1990                0.656               0.207        0.137               0.000
1991                0.742               0.174        0.084               0.000
1992+               0.785               0.172        0.043               0.000

*CLP-MPFI: Closed-Loop, Multiport Fuel-Injection
 CLP-TBI: Closed-Loop, Throttle-Body Fuel-Injection
 CLP-CARB: Closed-Loop, Carbureted
 OPLP: Open-Loop, All Fuel Systems


4.4 FINAL BASE EMISSION RATE EQUATIONS

Although the Tech IV methodology described above was specific to the
development of base emission rate equations for MOBILE4, the same
methodology was applied in developing base emission rates for
MOBILE4.1 and MOBILE5. (The subsequent revisions to the Tech IV model
have been named "Tech4.1" and "Tech5" to be consistent with the
MOBILE revision numbers.)


                                           -28-


Click HERE for graphic.


To illustrate the effect of vehicle mileage on emission rate, the
MOBILE5 HC and NOx BER equations for light-duty gasoline vehicles are
shown in Figures 20 and 21, respectively.  In addition to the fact
that emissions increase with vehicle age, it is obvious from these
figures that emissions are progressively lower for later model year
vehicles.  This is the result of more stringent emission standards as
well as the use of more durable technology (e.g., fuel injection). 
According to the data shown in the figures, however, the
implementation of more stringent emission standards on late-model
vehicles has a very small impact on the base emission rate.  For
example, comparing the 1990 and 1998+ modelyear HC emissions, only a
slight decrease in the base emission rate is observed even though the
BERs reflect a reduction in the HC emission standard from 0.41 to
0.25 g/mi.  It should be noted that the BER equations depicted in
Figures 20 and 21 do not include the effects of an inspection and
maintenance (I/M) program, nor do they include tampered vehicles.
(Both I/M benefits and the impact of tampered vehicles are accounted
for separately in the MOBILE models.)

A summary of the base emission rate equations for MOBILE5 is given in
Table 5. In comparing the uncontrolled 50,000-mile emission levels
(i.e., pre-1968 model year) to the 1998 and subsequent model year,
decreases of 93 percent, 90 percent, and 83 percent are observed for
HC, CO, and NOx, respectively.  Again, this indicates the significant
level of emission control that vehicles have been subjected to over
the years.  As an example, Table 6 lists the federal LDGV emission
standards with a summary of the predominant emission control systems
used to comply with those standards. (Similar tables for the LDGTI
and LDGT2 classes are contained in Appendix A.)


                                           -29-

Click HERE for graphic.


Click HERE for graphic.


                                           -30-

Click HERE for graphic.


                                           -31-


Click HERE for graphic.


Comparison of Base Emission Rate Equations - Figures 22 through 24
compare the HC, CO, and Nox base emission rates from MOBILE4,
MOBILE4.1, and MOBILE5 for 1990 model year LDGVS.  As seen, the base
emission rate equations for MOBILE4 and MOBILE4.1 are reasonably
similar, while the MOBILE5 emission rate equations are dramatically
higher, particularly for HC and CO.  This is primarily the result of
EPA's decision to use data collected at an operating I/M program to
develop base emission rates for MOBILE5.  These data were based on
transient, loaded mode I/M testing (i.e., IM240) from Indiana's I/M
program, which were then converted to "FTP-equivalent" values,
utilizing correlations developed from vehicles tested over both the
FTP and IM240 procedures.  According to EPA, the main reason for the
use of IM240 data is that the data allow for the compilation of a
much more robust data set that better represents the in-use fleet. 
In addition, biases in test vehicle selection are eliminated because
all vehicles in the fleet must undergo I/M testing.

Although Figures 22 through 24 illustrate the change in base emission
rates among the MOBILE revisions for a single model year, it is also
interesting to compare the differences for all model years.  This is
cumbersome to present as a series of base emission rate equations, so
specific mileage intervals were chosen to make the comparison. 
Figures 25 through 30 show HC, CO, and NOx emission rates,
respectively, for LDGVs at 50,000 and 100,000 miles for each of the
three model revisions.  These figures indicate only moderate
differences in the 50,000-mile emission rates among MOBILE versions. 
However, the 100,000-mile


                                           -32-

Click HERE for graphic.


Click HERE for graphic.



                                           -33-

Click HERE for graphic.

emission rates predicted by MOBILE5 are much higher than MOBILE4 and
MOBILE4.1 for 1980 and later model years.  This difference reflects
the much higher deterioration rates developed by EPA for MOBILE5 at
mileages greater than 50,000 (i.e., the DR2 values).  It should be
noted that the emission rates shown in Figures 25 through 30 do not
include the effects of an inspection and maintenance (I/M) program. 
Applying I/M credits to the MOBILE5 results dramatically reduces the
non-I/M deterioration. (This point is discussed in greater detail in
Section 7.) This same pattern is also observed for light-duty trucks,
and figures similar to those presented above are contained in
Appendix A.

Pre-1981 Base Emission Rates - For model years prior to 1981, the
Tech IV modeling methodology described above was not used.  Instead,
surveillance program vehicles were divided into groups according to
their certification emissions standards.  Linear regressions of
emissions vs. mileage were developed for each group to predict zero
mile emissions and deterioration rates.  The emission factors for
most of these pre-1981 vehicles have remained essentially unchanged
since MOBILE2 was released in early 1981.  These factors were
verified through a test program conducted by EPA in 1987.  Emission
data from that program, which consisted of 1972 to 1977 model years,
indicated very close agreement with the base emission rates developed
previously.


                                           -34-

Click HERE for graphic.



Click HERE for graphic.


                                           -35-


Click HERE for graphic.



Click HERE for graphic.


                                           -36-


Click HERE for graphic.



Click HERE for graphic.


                                           -37-


4.5 CORRECTION FACTORS

As discussed above, a number of correction factors are needed to
tailor the base emission rates to the specific conditions of speed,
temperature, etc. being modeled.  These are detailed below.

Operating Mode Corrections - Also referred to as bag correction
factors, the operating mode correction factors adjust the FTP
composite base emission rates to account for the mode under which the
vehicle is operating (i.e., cold start, stabilized, hot start). 
Because the mass emission rate during the cold start mode is
typically much higher compared to during stabilized and hot start
modes, the BCF for Bag 1 is greater than unity.  On the other hand,
the BCFs for Bags 2 and 3 are generally less than one (with the
exception of Bag 3 NOx).  MOBILE predicts that the bag correction
factors for catalyst-equipped vehicles are a function of vehicle
mileage, with a decrease in weight for Bags I and 3 as the vehicle
ages.  This is seen graphically for HC, CO, and NOX in Figures 31
through 33 for 1990 model year LDGVs based on the MOBILE5 bag
correction factor coefficients.

Note that at any mileage point, the equation

       0.206 *BCF1 + 0. 521 *BCF2 + 0.273 *BCF3 = 1            [5-4]


is valid.  This occurs because the base emission rate for a vehicle
operating according to the FTP bag weightings (which were derived in
equations 5-1 and 5-2) does not need to be corrected for non-FTP
operating mode fractions, and the sum of the FTP-weighted BCFs is
unity.

Click HERE for graphic.

                                           -38-

Click HERE for graphic.



Click HERE for graphic.



                                           -39-



Temperature and RVP Corrections - Under the Federal Test Procedure,
emissions tests are performed within a temperature window of 68 to
86F, in order to ensure consistency and repeatability of the test
results.  However, many air quality problems are associated with
temperatures that lie far outside this range, and the performance of
different emission control components can vary substantially with
temperature as well.  For example, at cold temperatures common in CO
nonattainment areas, catalyst warm-up times are far longer than at
the standard temperature conditions.  On the other hand, when
temperatures are above 86F, evaporative emission control system
purges can take longer (because more HC vapors have been stored in
the canister), resulting in higher exhaust HC and CO emissions. 
Consequently, the MOBILE models contain factors to adjust the base
emission rates to reflect the temperature for which the model is
being run.

The impact of fuel volatility, measured as Reid Vapor Pressure (RVP),
on vehicle emissions is closely related to temperature.  Thus, there
also is an RVP adjustment contained in the MOBILE models.  The RVP
adjustment at temperatures below 75F is a simple multiplicative
factor; however, at temperatures above 75F, the temperature and RVP
adjustment is combined into a single correction factor. (It should be
noted that RVP impacts exhaust emissions primarily by its influence
on vapor storage in the evaporative emission control system.)

The MOBILE models calculate a correction factor to adjust the base
emission rate for fuel volatility and temperature according to two
basic vehicle groups.  Group 1 consists primarily of non-catalyst and
open-loop technology vehicles.  Group 2 consists primarily of
vehicles using closed-loop three-way catalyst technology.  The model-
year breakdown of these groups is:


              Group 1                   Group 2
       1970 - 1979 LDGV          1980 + LDGV
       1970 - 1980 LDGT1         1981 + LDGT1
       1978 - 1980 LDGT2         1981 + LDGT2
              All HDGV


Low-Temperature Correction Factors - Because temperature affects
emissions differently for each operating mode, temperature correction
factors (TCFS) are derived separately for Bags 1, 2, and 3. In
addition, emissions at lower temperatures are dependent on the type
of fuel delivery system with which the vehicle is equipped (i.e.,
carburetion, throttle-body fuel-injection, or multipart fuel-
injection).  For example, the Bag 1 HC low-temperature correction
factors for the three fuel delivery systems are shown in Figure 34. 
The figure shows that as the fuel delivery system is improved
(resulting in better atomization and fuel distribution among
cylinders), cold-temperature emissions are improved.

The temperature correction factors (TCFS) by fuel delivery system
remained unchanged among the MOBILE4, MOBILE4.1, and MOBILE5
revisions; however, slight changes to model-year-specific TCFs have
occurred because of changes to the fuel delivery system technology
mix.  Using


                                           -40-



Click HERE for graphic.


Using the MOBILE4 assumptions on technology mix, TCFs (by bag) were
calculated for HC, CO, and NOx, and the results are presented in
Figures 35 through 37 for 1992 and subsequent model year LDGVS.  As
seen in the figures, the impact of temperature is greatest on cold
start emissions, increasing the HC emission rate by a factor of
almost 4 at 20F.  For stabilized and hot start operation, the HC
emission rate is increased by roughly 50 percent at 20F.  A very
high emission rate during cold start at low temperatures is simply
the result of a magnification of the factors that lead to high
emissions during cold start at standard temperature, i.e., incomplete
fuel vaporization and an increased warm-up period for the emission
control system.

Modeling of low-temperature CO emissions for Bag 1 is handled
somewhat differently in the MOBILE models.  Instead of being a
multiplicative correction factor, the Bag 1 CO temperature correction
is an offset which is added to the base emission rate before
corrections for speed, fuel, etc. are applied.  This offset is a
linear function of temperature and varies according to fuel delivery
system.  As an example, the cold start CO offset for a 1982 model
year LDGV at 30F is roughly 60 grams per mile. (Recall, however,
that this only contributes to the cold start operating mode, and the
impact on the composite emission factor would be correspondingly
less.)

Low-Temperature RVP Correction - The RVP correction factors for low-
temperature operation are calculated independently from the
temperature correction factors and are developed for each operating
mode separately.


                                           -41-

Click HERE for graphic.



Click HERE for graphic.



                                           -42-



Click HERE for graphic.


(For 1971 to 1979 model years, the RVP correction is not bag
specific.  Further, it is not applied to vehicles older than the 1971
model year.) Figures 38 and 39 illustrate the low-temperature HC and
CO RVP correction factors for a series of fuel RVPS.  Note that the
impact falls as the temperature declines, and no correction is
applied at temperatures below 45F.  While not shown, there is also
an RVP correction applied to NOx; however, its magnitude is very
small.

High-Temperature Combined RVP and Temperature Correction - At high
temperatures, the temperature and RVP correction factor is combined. 
Figures 40 and 41 show the MOBILE5 RVP/temperature correction factors
for HC and CO, respectively.  Although the model calculates these
factors independently for each operating mode, they have been
composited in the figures to reflect the FTP operating mode
fractions.  As seen, the correction factors increase with increasing
temperature and with increasing RVP. (Although not shown, NOx is also
corrected; however, the correction factor is much smaller for NOx.)

Speed Corrections - Emission factors are very sensitive to the
average speed that is assumed.  In general, emissions tend to
increase as average speeds decrease from the 19.6 mph value employed
in the FTP.  The MOBILE models do not assume an average speed: it is
a requirement that an estimate of the speed experienced by vehicles
operating in the area of analysis be specified.  MOBILE adjusts the
emission factors for speeds other than 19.6 mph through the use of
speed correction factors.  These multiplicative adjustments to the
base emission factors follow a non-linear relationship that increases
the emission levels as speeds decline.



                                           -43-

Click HERE for graphic.



Click HERE for graphic.



                                           -44-



Click HERE for graphic.



Click HERE for graphic.



                                           -45-


There are several forms of speed correction factors contained in the
MOBILE models:

              SCF - exp(A + B*S + C*S2 + D*S3 + E*S4 + F*S5) (Form 1)

              SCF - (A + B*S + C*S2 + D*S3 + E*S4 + F*S5) (Form 2)

              SCF - (A/S + B) / (A/FTPS + B) (Form 3)

              SCF - (exp(A + B*S + C*S2)) / (exp(A + B*FTPS + C*FTPS2))
              (Form 4)
where:


SCF    = speed correction factor,
FTPS   = operating mode adjusted FTP speed,
S      = vehicle speed in mph,
A,B,..        = constants specific to each pollutant, model year group,
and vehicle category.

The exponential equation (Form 1) is the form that was used in EPA's
MOBILE2 emission factor model for HC and CO emissions.  The second
form of the equation is used for NOx emissions.  These forms are
still generally applicable to older model year vehicles.  The third
and fourth forms were introduced with MOBILE4, and are generally used
for 1979 and later model vehicles.  Form 3 is used for HC and CO
emissions, while Form 4 is used for NOx emissions.

The data used to develop the speed correction factors are based on a
number of different test cycles in which the speed-time profile is
changed to reflect different average speeds.  A total of eight
different test cycles are used by EPA in modeling the effect of speed
on emissions:

             EPA   Low Speed #3              2.5 mph
             EPA   Low Speed #2              3.6 mph
             EPA   Low Speed #1              4.0 mph
             New   York City Cycle           7.1 mph
             EPA   Speed Cycle 12            12.1 mph
             FTP                             19.6 mph
             EPA   Speed Cycle 36            35.9 mph
             Highway Fuel Economy Test    47.9 mph


In addition to these cycles, EPA has used data from CARB's high-speed
testing to develop the high-speed (i.e., over 48 mph) correction
factors.  The speed-time traces for all of the above speed correction
cycles are contained in Appendix B.

The impact of speed on emissions is observed in Figures 42 through
44.  These figures present the speed correction factors (SCFS)
calculated by MOBILE4, MOBILE4.1, and MOBILE5 for HC, CO, and NOx,
using 19.6 mph as the reference point.  There are three speed regimes
modeled in MOBILE: low speed (under 19.6 mph), mid-speed (19.6 to 48
mph) and high speed (48 to 65 mph).



                                           -46-

Click HERE for graphic.


Click HERE for graphic.



                                           -47-
Click HERE for graphic.


The low-speed regime results in the highest SCFs (and thus emissions)
for HC and CO, with MOBILE4 predicting an SCF of over 8 for HC at 2.5
mph, while MOBILE4.1 and MOBILE5 estimate an SCF of over 4 for this
speed.  (MOBILE4.1 and MOBILE5 speed correction factors are
equivalent for the low-speed regime.) On the other hand, low-speed
NOx emissions are predicted to increase by only about 50 percent from
the reference speed.  Higher emissions at these very low speeds occur
because the vehicle is operating over a longer time period to cover
the same distance, and the test cycles used to develop the low-speed
corrections have a higher fraction of time spent under acceleration
modes.

The mid-range HC and CO SCFs are very similar among the MOBILE
revisions, steadily decreasing from 19.6 to 48 mph.  The results for
NOx, however, are quite different for the three MOBILE revisions,
with MOBILE5 showing a steady increase throughout the mid-speed
regime.  Finally, MOBILE4.1 and MOBILE5 incorporated SCF revisions
that resulted in a significant increase at high speed.  High-speed
data developed by the CARB were used in developing the high-speed
correction factors.



                                           -48-



Additional Correction Factors - Before the basic emission factors for
the contributing model years can be summed, they must be further
corrected for area-specific air conditioner use, trailer towing, and
extra load.  Although the MOBILE models contain the option to include
this correction factor, it is generally based on few data and
typically it is not utilized.

Tampering - Tampering effects are also accounted for in MOBILE.  When
basic emission rates are developed from surveillance vehicles, the
effects of tampering are removed by deleting tampered vehicles from
the test sample.  An additive emissions impact (in g/mi) and a rate
of occurrence have been developed by EPA for each tampered component
(including air pump, fuel inlet restrictor and EGR system
disablement, catalyst removal and misfueling), based on tests of
tampered vehicles and EPA surveys.  Although MOBILE contains default
tampering rates, these rates can be changed by the user to reflect
area-specific tampering rates and antitampering programs.  The
estimated increase in emissions due to tampering is adjusted to
reflect the reduced tampering rates expected under an antitampering
program in the development of the tampering offset.

4.6 IDLE EMISSION RATES

Although not generally used for inventory preparation purposes, idle
emission rates are sometimes needed for more specialized analyses
(e.g., intersection modeling).  Idle emissions are not calculated by
MOBILE5, but idle rates are estimated by MOBILE4 and MOBILE4.1.
Because the interactions of some of the new features included in
MOBILE5 (e.g., Tier I emission standards, reformulated gasoline, low-
emission vehicles) with the idle emission calculations resulted in
uncertain estimates, the idle algorithm was disabled in MOBILE5.  EPA
initially recommended continued use of MOBILE4.1 for idle emission
estimates; however, EPA has recently issued guidance proposing a
method to convert MOBILE5a exhaust emission rates to idle rates.* The
methods used to calculate idle emission rates in MOBILE4 and
MOBILE4.1, as well as the MOBILE5a procedure, are described below.

MOBILE4 - Idle emission rates are calculated by the MOBILE models in
units of grams per hour (g/hr).  As with running exhaust emission
estimates, a fleet-average idle emission rate is calculated by
weighting the model-year-specific idle emission rates by each model
year's travel fraction.  The sum of these VMT-weighted idle rates
gives the fleetaverage idle emission rate for the calendar year of
interest.  The idle rates calculated by MOBILE4 assume that the
vehicle is warmed-up and that the ambient temperature is 75F.

The methodology used to generate model-year idle emission rates in
MOBILE4 varies according to vehicle class and model year.  For the
older model years (i.e., pre-1977 for LDGVS, pre-1979 for LDGT1s and
LDGT2s),


__________________________

* EPA will allow the continued use of MOBILE4.1 for one year. 
However, EPA intends to use the method developed for MOBILE5a in the
future.


                                           -49-

the model employs a relatively simple algorithm.  Idle emission
factors, which were based on "bagged" idle data, are stored in the
model in terms of a zero-mile component (g/hr) and deterioration rate
(g/hr/10,000 mi).  An idle emission rate is computed by multiplying
the deterioration rate and the assumed mileage for the model year
being analyzed, and adding the results to the zero-mile component. 
As an example, consider a 1975 model-year LDGV in 1990.  Such a
vehicle would have accumulated an estimated 135,705 miles; thus, its
CO idle emission rate would be:

  Idle_CO1975MY = 360.00 g/hr + (51.00 g/hr/10,000 mi * 135,705 mi)
  Idle_CO1975MY = 1,052.10 g/hr


This rate is then corrected for I/M (if specified) and a tampering
offset is added.  For this model year and vehicle age, the tampering
offset (under a no-I/M scenario) is calculated to be 90.02 g/hr. (The
tampering offset accounts for catalyst removal, air pump tampering,
and misfueling.) Thus, the total CO idle rate estimated by MOBILE4
for this model year is 1,052.10 + 90.02 = 1,142.12 g/hr.

For 1977 and subsequent model years, a significantly different method
is used to generate g/hr idle emission rates.  That method is based
on the g/mi running exhaust emission rate which is converted to g/hr. 
The calculation is somewhat complicated and is best illustrated with
an example, which is provided below.

Consider a 1986 model-year LDGV in calendar year 1990.  It would have
accumulated 45,572 miles; its running exhaust CO base emission rate
is described by the following ZM and DR values:

       ZM:    2.764 g/mi
       DR:    0.771 g/mi/10,000 mi;

and the CO emission rate is calculated to be:

CO1986 = 2.764 g/mi + (0.771 g/mi/10,000 mi * 45,572 mi) - 6.28 g/mi.

(Because the vehicle has under 50,000 miles, a second DR is not
necessary to calculate the running exhaust emission rate.)

The 6.28 g/mi value above is based on FTP operating mode fractions
(i.e., 20.6% cold start, 52.1% stabilized, and 27.3% hot start). 
However, EPA chose to calculate idle emissions for MOBILE4 only under
stabilized conditions.  Thus, operating mode correction factors (or
"bag correction factors") are applied to the FTP-based base emission
rate so that it represents stabilized operation.  For the purposes of
calculating idle emission rates, the bag 2 (stabilized) correction
factor is weighted 52.1% and the bag 3 (hot start) correction factor
is weighted 47.9%. For this example (i.e., a 1986 model year LDGV in
1990), the bag correction factor applied to the exhaust CO emission
rate is:

       COBCF = 0.521*BCF2 + 0.479*BCF3
       COBCF = 0.521*0.666 + 0.479*0.864 - 0.761,



                                           -50-


and the corrected running exhaust emission factor is 6.28 * 0.761 =
4.78 g/mi.

The next step in the process is to transform the g/mi emission rate
calculated above to a g/hr basis.  This is accomplished by
multiplying the 4.78 g/mi value by 19.6 mi/hr, the average speed of
the FTP.  For this example, the calculation results in an idle CO
rate of:

       Idle_CO1986MY = 4.78 g/mi * 19.6 mi/hr - 93.69 g/hr.

An offset is then added to the above result so that the calculated
values match observed "bagged" idle data.  In this example, the
offset is -42.13 g/hr, and the idle CO rate is:

       Idle_CO1986MY = 93.69 g/hr - 42.13 g/hr - 51.56 g/hr.

If an I/M program is specified, the I/M credits are applied at this
point in the calculation. (I/M was not specified for this example.)
Finally, a tampering offset is added, which in this case is 4.46
g/hr.  Thus, the final CO idle emission rate is 51.48 + 4.46 - 55.94
g/hr.

MOBILE4.1 - MOBILE4.1 uses a methodology very similar to that
employed in MOBILE4 to calculate idle emission rates.  Several
modifications were made, however, so that the idle rates from
MOBILE4.1 reflect user-input operating mode fractions, temperature,
RVP, and fuel oxygen content.  For the most part, these corrections
are based on data collected over the FTP.  In some cases,
modifications are made within MOBILE4.1 to get the corrections on an
idle basis (e.g., the cold-temperature CO offset, which is calculated
in units of g/mi, is converted to g/hr), while others (e.g., the RVP
correction factor, which is a multiplicative correction factor) are
used directly to adjust the idle emission rate.  It is unclear that
idle emission rates and running exhaust emission rates will react the
same to changes in some of these model input parameters.  However,
there appears to be an acute lack of data to develop idle-specific
emission rates and correction factors.*

As an example of the above corrections, the impact of temperature and
operating mode on idle CO emission rates for the 1996 calendar year
is shown graphically in Figure 45 for the LDGV vehicle class.  As
with running exhaust emission rates, the idle CO rates increase with
decreasing temperature, particularly during cold start operation.

It is also interesting to compare idle emission rates computed by
MOBILE4.1 to the running exhaust emission rates estimated by that
model.  This is illustrated in Figure 46, which shows the LDGV
50,000-mile running exhaust and idle CO emission rates for pre-1968
to 1995 and subsequent model years. (Similar charts for LDGTls and
LDGT2s are in Appendix C, which also contains HC and NOx results.) As
shown, the idle
______________________

*Although it would appear that idle emission rates could be easily
obtained from the numerous I/M programs operating throughout the
U.S., those data are collected in concentration units (e.g., percent
CO, ppm HC), which are not directly transferrable to a mass per unit
time basis.



                                           -51-

Click HERE for graphic.




Click HERE for graphic.

                                           -52-

rates generally track the decrease in running exhaust rates as new
standards and technologies have been implemented.

MOBILE5a - As outlined above, the idle emissions algorithm was
disabled in MOBILE5 (and MOBILE5a) because the addition of new
features (e.g., Tier I emission standards, reformulated gasoline,
low-emission vehicles) resulted in uncertain idle emission estimates
from the algorithm utilized in MOBILE4.1. When MOBILE5 was released,
EPA recommended the continued use of MOBILE4.1 for calculating idle
emission rates.  Recent EPA guidance, however, provides a method to
convert MOBILE5a running exhaust emission rates to idle emission
rates, and EPA recommends that approach for all future emission
estimates requiring idle emission rates.

EPA's recommended method to convert MOBILE5a g/mi running exhaust
emission rates to g/hr idle emission rates is somewhat similar to the
approach already used within MOBILE4 and MOBILE4.1 to develop idle
emission rates for 1977 and subsequent model ears.  The MOBILE5a
method requires the user to run MOBILE5 at 2.5 mph and then multiply
the resulting g/mi running exhaust rates by 2.5 mph to obtain idle
rates in terms of g/hr. (The 2.5 mph speed is utilized because it
represents the lowest speed modeled by MOBILE5a, and the test cycles
used to develop speed correction factors for low speeds contain the
highest fraction of idle operation.) Developing idle rates in this
manner inherently accounts for all of the corrections (e.g., I/M,
temperature, fuel, etc.) that are applied to the running exhaust
emission rates.  The idle CO emission rates calculated according to
the methodology outlined above are compared to running exhaust
emission rates in Figure 47 for LDGVS.  As with the MOBILE4.1
results, the idle rates generally track the running exhaust emission
rates as new standards are implemented. (This is not unexpected since
the idle rates are based on the running exhaust emission rates.)
However, there is a slight discontinuity in the idle results for the
1975 and 1976 model years.  That is because the 2.5 mph speed
correction factor for those model years is lower than the SCF for the
1977-1979 model years, while the FTP-based emission rates are roughly
the same. (Charts similar to Figure 47 for LDGTls and LDGT2s are in
Appendix C, which also contains the HC and NOx results.)

Comparison of MOBILE4.1 and MOBILE5a Idle Emission Rates - A
comparison of model-year-specific idle CO emission rates calculated
with MOBILE4.1 and MOBILE5a is illustrated in Figure 48 for LDGVs at
50,000 miles.  This figure shows reasonable agreement between the
models, with the MOBILE5a results being slightly higher for later
model years.  Some of this difference is attributable to the in-use
fuel correction applied in MOBILE5 (explained in Section 8.5) and the
differences in methodologies used to develop the idle emission rates. 
The comparison made in Figure 48, however, is based on vehicles with
50,000 miles.  The difference for late-model vehicles at higher
mileages will be greater because of the higher emission deterioration
rates assumed in MOBILE5 for those vehicles.  Thus, comparisons of
fleet-average idle emission rates will show greater differences, as
demonstrated in Figure 49 for LDGVS.  This is particularly true of
future-year analyses (e.g.,


                                           -53-

Click HERE for graphic.


Click HERE for graphic.



                                           -54-

Click HERE for graphic.

calendar year 2010), where the MOBILE5a idle CO predictions are twice
those calculated by MOBILE4.1.

4.7 HEAVY-DUTY VEHICLES

Heavy-duty vehicles consist of vehicles that exceed 8500 lbs. gross
vehicle weight (GVW).  This general class is further segregated into
specific classes according to GVW.  The additional segregation is
necessary because of the differing characteristics of engines making
up these classes.

Standard Test Procedure - Because of the large number of applications
for which heavy-duty engines are utilized, emissions testing is
normally engine-specific and is performed on an engine dynamometer.
Additionally, the heavier GVW rating of heavy-duty vehicles precludes
testing on most chassis dynamometers.  Therefore, transient engine
dynamometer test cycles have been developed that simulate average
urban driving for gasoline and Diesel heavy-duty engines.  These test
cycles specify RPM and torque by second and are roughly 20 minutes
long.  The test procedure, outlined in 40CFR86, calls for a "cold
start" portion which is initiated after the engine temperature is
stabilized at 68 to 86F. (This can be accomplished through an
extended "cold soak" period or with a "forced cool-down" procedure.)
A "hot start" portion, commencing 20 minutes after the end of the
cold start test, is also required and follows the same dynamometer
test cycle.  The test results are reported in units of grams per
brake-horsepower-hour (g/Bhp-hr), and


                                           -55-



the composite emission rate is determined by weighting the cold
portion of the test by 1/7 and the hot portion of the test by 6/7. 
The test procedure is diagrammed in Figure 50.

Click HERE for graphic.


                       Figure 50.  Heavy-Duty Engine Test Procedure



The transient test cycles developed for heavy-duty gasoline and
Diesel engines are the result of a significant effort by EPA and
industry to simulate heavy-duty vehicle operation in urban areas. 
The data used to develop the cycles were collected from instrumented
heavy-duty trucks that operated in New York City and the Los Angeles
Basin in the mid1970s.8 The complete 20-minute cycles were formulated
from four separate 5-minute cycles that represented freeway driving
in New York City, non-freeway driving in New York City, freeway
driving in Los Angeles, and non-freeway driving in Los Angeles.9,10 An
optional chassis dynamometer test cycle was also developed in this
program which covers 5.73 miles at an average speed of 19.45 mph.9

Basic Emission Rates - The basic emission rates developed for the
MOBILE models are based on engine dynamometer test results.  For
MOBILE4, data from a cooperative test program with engine
manufacturers formed the basis of the g/Bhp-hr emission estimates
(for 1979 and later model years).  For heavy-duty Diesel vehicles, a
total of 30 engines were tested that were representative of the 1979
to 1984 model years.  For heavy-duty gasoline emission rates, 18
engines were tested by EPA.  These were from the 1979 to 1982 model
years.  Unfortunately, heavy-duty engine testing is very expensive,
so data are generally sparse.  No new data since the MOBILE4 effort
have been developed with which to update heavy-duty emission factors.



                                           -56-

Conversion Factors - Because the exhaust emission test procedure
results in emissions reported in units of g/Bhp-hr, it is necessary
to convert the results into g/mi units to be consistent with
available travel information.  Therefore, conversion factors (in Bhp-
hr/mi) are developed to represent the emission results obtained from
engine dynamometer testing in units appropriate for inventory
purposes.  The derivation of heavy-duty conversion factors is
described in a 1984 EPA technical report11 which was updated in 1988.12
Only a summary of the methodology is presented here.

Because it is difficult to measure Bhp-hr/mi directly, a methodology
was developed to calculate this parameter with available data.  In
equational form, the conversion factor is represented by:


              CF =               P
                           --------------
                              BSFC x FE


where         CF    = conversion factor (Bhp-hr/mi),
              p     = fuel density (lb/gal),
              BSFC  = brake-specific fuel consumption (lb/Bhp-hr), and
              FE    = fuel economy (mi/gal).


Thus, by obtaining estimates of fuel density, brake-specific fuel
consumption, and fuel economy, it is possible to estimate CF.  Once
the values of CF are obtained for each GVW class, a fleet composite
CF is established by weighting the class-specific values by an
estimated VMT mix.  These calculations are carried out separately for
gasoline and Diesel vehicles.

The data sources used in developing the conversion factors for 1982
and previous model years are summarized as follows:

             Fuel Density - Gasoline fuel density was based on data
              published by the National Institute for Petroleum and
              Energy Research (NIPER), and Diesel fuel density was based
              on surveys by the Motor Vehicle Manufacturers Association
              (MVMA). (The same value for fuel density was used for all
              model years.)

             Fuel Economy - Estimates for heavy-duty trucks were
              obtained from the Department of Commerce's "1982 Truck
              Inventory and Use Survey" (TIUS), while bus data came from
              documents published by the Federal Highway Administration
              (FHWA).

             Brake-Specific Fuel Consumption (BSFC) - The pre-1978
              heavy-duty truck data came from a 1983 report prepared by
              Energy and Environmental Analysis (EEA) for the MVMA.  For
              model years 1978 to 1982, the BSFC values were interpolated
              from the above 1977 estimates and more recent data
              submitted to EPA by manufacturers for 1987 model year
              engines.  Bus data were based on EPA testing of high sales-
              volume bus engines.


                                           -57-

The 1983 and subsequent conversion factors were developed by
incorporating fuel economy improvements into the 1982 class-specific
conversion factors. (Because engine-related fuel economy improvements
also result in decreased BSFC (which cancels the overall impact on
the CF), only non-engine-related fuel economy improvements were
considered.) Specific improvements accounted for in the analysis
included weight reduction, radial tires, aerodynamic add-on devices,
drive-train lubricants, improved fan drives, overdrive, electronic
transmission controls, and speed controls.

The sales and VMT estimates used to composite the class-specific
conversion factors for each model year were based on information
contained in various publications authored by MVMA, the Department of
Energy, the American Public Transit Association, and FHWA, as well as
data in the TIUS.

Speed and Temperature Adjustment - As with light-duty vehicles, the
heavy-duty emission rates are corrected for average speed and
temperature outside of those encountered in the test procedure.
(Gasoline vehicles are corrected for both speed and temperature,
while Diesel vehicles are corrected only for speed.  The effect of
nonstandard temperature on Diesel emissions is assumed to be
negligible.) For speed corrections, the basic emission rate is
multiplied by the appropriate speed correction factor that is
calculated from the speed correction coefficients according to the
following formula:

              SCF = exp (a + bs + cs2)

where         a,b,c        = coefficients,
              s            = vehicle speed, and
              exp          = exponential function.


Figures 51 through 53 show a graphical representation of the heavy-
duty Diesel and gasoline speed correction factors for HC, CO, and
NOx, respectively. (The same speed correction factors apply to all
model year vehicles.)

Temperature correction is only applied to heavy-duty gasoline
vehicles.  Further, this correction is not operating-mode specific,
which is a departure from the light-duty gasoline vehicle
methodology.  In equation form, the temperature correction factor is
represented by:

              TCF = exp(a * (T - 75))


where         a     = coefficient,
              T     = temperature (F), and
              exp   = exponential function.


The value of the coefficient for a particular model year group
differs, depending upon whether the temperature range being modeled
is above or below 75F.  The heavy-duty gasoline vehicle temperature
correction factors for 1985 and later model years are depicted in
Figure 54 for


                                           -58-

Click HERE for graphic.



Click HERE for graphic.



                                           -59-



Click HERE for graphic.



Click HERE for graphic.



                                           -59-

HC, CO, and NOx for temperatures below 75F.  As seen, the influence
of temperature is greatest for HC at low temperatures.

Changes from MOBILE4 to MOBILE4.1 to MOBILE5 - Since MOBILE4, very
little work has gone into updating the parameters that make up the
heavy-duty emission factors.  The only substantive revision was the
incorporation of the Clean Air Act requirements for a 4.0 g/Bhp-hr
Nox emission standard beginning with the 1998 model year.  This
change was introduced into the model by simply scaling the 1998 and
subsequent zm component of the basic emission rate equation by the
ratio of the proposed-to-existing standard (i.e., 4.0/5.0).




                                           -61-

                           5. EVAPORATIVE EMISSION METHODOLOGIES

Evaporative emissions are treated quite differently than exhaust
emissions in the MOBILE models, with six separate categories of
evaporative emissions being considered: hot soak, diurnal, running
losses, resting losses, refueling, and crankcase emissions.  Only hot
soak, diurnal, and crankcase emissions are currently regulated by
EPA.  Refueling losses, if controlled, are regulated by local air
pollution control agencies through Stage II vapor recovery
requirements.* Running and resting loss emissions will require
control in response to the recently revised evaporative test
procedures.

5.1 EMISSION STANDARDS AND TEST PROCEDURES

Evaporative emissions from gasoline-fueled, light-duty vehicles have
been controlled since the 1971 model year. (Heavy-duty vehicles were
first controlled beginning in 1985.) Evaporative emission standards,
listed in Table 7, are based on the sum of diurnal and hot soak
emissions through 1995, while vehicles built after 1995 will also be
required to meet a running loss standard.  The standards summarized
in Table 7 are based on three different test procedures: the 'carbon
trap" method, the "SHED' method, and an enhanced evaporative test
procedure that is a modification of the original SHED method.  Each
of these procedures has required an increased level of emission
control.

The discussion that follows summarizes the basic function of the
evaporative emission control system used to meet the standards in
Table 7. In addition, each of the test procedures used to certify
vehicles to those standards is briefly described.

Evaporative Emission Control Systems - The evaporative emission
control system generally consists of a series of hoses that direct
hydrocarbon vapors to a canister filled with activated carbon.  The
carbon adsorbs these vapors during periods of vapor generation (i.e.,
diurnal emissions resulting from heating of the fuel tank during the
day; hot soak

______________________________

*The 1990 Clean Air Act Amendments (CAAA) require the implementation
of Stage II vapor recovery in all areas of the country that are
classified as moderate, serious, severe, and extreme nonattainment
with respect to the National Ambient Air Quality Standards for ozone. 
These requirements should be fully implemented by November 15, 1995. 
Further, the CAAA direct EPA to develop regulations requiring the use
of on-board refueling vapor recovery systems on new light-duty
vehicles.  Although EPA has not yet promulgated these requirements,
it appears that it will do so in 1994, and ORVR systems are likely to
first appear with the 1998 model year.

                                           -63-
                          Table 7. Light-Duty Vehicle Evaporative
                                    Emission Standards


                    Diurnal +
       Model        Hot Soak     Running Loss               Test
       Year         (g/test)     (g/mi)                     Procedure

       1971         6.0                 -                   Carbon Trap
       1972-77      2.0                 -                   Carbon Trap
       1978-80      6.0                 -                   SHED
       1981-95      2.0                 -                   SHED
       1996a        2.0                 0.05                Enhanced Evap
       

       aThese standards are phased-in beginning with the 1996 model
       year.



emissions resulting from evaporation of fuel exposed to a hot engine
after the engine is turned off).  When the engine is running, the
canister is "purged" of these stored vapors by drawing fresh air
through the canister and routing these vapors back into the engine's
fuel intake system where the vapors are burned in the combustion
process.  Although evaporative emission standards and test procedures
have changed through the years, the basic method of control has
remained essentially the same.  To meet increasingly stringent
evaporative standards, manufacturers have relied on better materials
(e.g., less permeable vapor hoses) and larger carbon canister storage
capacity (e.g., more adsorptive activated carbon or increased
canister volume).

"Carbon Trap" Method - The first evaporative emission standards were
based on a test procedure known as the "carbon trap" method.  That
method of measuring evaporative emissions utilized activated carbon
traps connected to the fuel system at selected locations where vapors
would be expected to escape.  The traps adsorbed vapor emitted at
those locations, and the increased weight of the carbon traps
represented evaporative emissions.  Emissions were measured during a
one-hour diurnal test, over a running loss test designed to represent
evaporative losses during urban driving, and during a one-hour hot
soak test following engine shut-off.  The sum of these measurements
represented the total evaporative emission rate.

After implementing this procedure for several years, it became
apparent that it was deficient in several ways.  Most notably, the
carbon trap method did not accurately measure evaporative emissions
from all possible sources.  Using the Sealed Housing Evaporative
Determination (SHED) method (described below), both EPA and the
California Air Resources Board found that vehicles certified
according to the carbon trap method greatly exceeded the emission
standards to which they were certified. 13 Thus, both agencies
adopted evaporative emission


                                           -64-

regulations based on the SHED procedure beginning with the 1978 model
year.

SHED Procedure - The evaporative test procedure by which
manufacturers have certified their vehicles since the 1978 model year
is based on emissions measurements conducted in an enclosed test cell
(SHED) and consists of two parts: a diurnal test and a hot soak test.
(The original SHED-based standards did not include running loss
measurements as these were, at the time, considered to be
negligible.) These evaporative tests are conducted in conjunction
with the FTP exhaust emissions test.  The diurnal test is performed
prior to the exhaust test and consists of first draining and filling
the fuel tank (to a 40 percent fill level) with standard test fuel
that is at a temperature of 60F.  The vehicle is then placed in the
SHED, and the temperature of the fuel is raised from 60 to 84F in
a one-hour time period. (A heat blanket is placed on the outside of
the fuel tank to effect the temperature increase.) Emissions are
collected during that time period, and test results are reported in
terms of grams/test.

The hot soak test is performed after the FTP exhaust emissions test. 
Immediately following the exhaust test (i.e., within two minutes of
engine shutdown and within seven minutes of the completion of the
exhaust emissions test), the vehicle is placed in the SHED for a
period of one hour.  Emissions are collected during this time, and
results are reported in grams/test (i.e., grams/trip).

Enhanced Evaporative Test Procedures - Through a variety of test
programs, EPA found that the above SHED test procedure does not
adequately describe evaporative emissions under all conditions.  This
was also recognized by Congress in the 1990 Clean Air Act Amendments,
which directed EPA to promulgate a new evaporative test procedure to
better represent evaporative emissions under in-use, summertime
conditions.  These requirements were finalized in March 199314 and
include an extended, multi-day diurnal period with a more severe
temperature rise (72 to 96F) and a requirement for evaporative
emissions measurement while the vehicle is operating on a chassis
dynamometer (i.e., a running loss test).  MOBILE5a models the effects
of the new evaporative emissions test procedure on vehicles certified
under the new requirements, which are to be phased-in beginning with
the 1996 model year. (For 'a discussion of how MOBILE5a models the
effects of these regulations, refer to Section 8.3.)

5.2 EVAPORATE EMISSIONS DATA

MOBILE4 - The evaporative emission factors used in MOBILE4 were
derived from the results of EPA's in-use emission factor test
program.  For the development of the emission factors, the test
vehicles were divided into two categories: tampered and non-tampered.

The vehicles that had evaporative emission control system
malfunctions attributed to deliberate tampering were placed in the
category of tampered vehicles.  The remaining vehicles were placed in
the non-tampered vehicle category.  Within each of the above
categories, the


                                           -65-
vehicles were also separated into carbureted and fuel-injected
vehicles.  In the MOBILE4 evaporative emissions calculations, the
fuel-injected vehicles were further divided into throttle-body-
injected (TBI) vehicles and multipart fuel-injected (MPFI) vehicles. 
The non-tampered emissions estimates are based on the emissions of
the vehicles in the emission factor database.  Tampered emissions
estimates were developed based on a separate testing program in which
vehicles were tested in a tampered state.

For calculating hot-soak and diurnal emissions, the light-duty
vehicles were divided into the following model year categories based
on the different evaporative emission standards for these model
years:

             1970 and earlier,
             1971,
             1972-1977,
             1978-1980, and
             1981+.


The estimates for 1981+ vehicles are the most refined as they are
based on the largest amount of data.  The emissions estimates for
older years become increasingly less accurate because the earlier
data were based on less stringent test procedures and performed on a
smaller sample of vehicles.  This issue is of lesser importance for
the future years as the population of the pre-1981 vehicles becomes
smaller.

MOBILE4.1 and MOBILE5 - With the release of MOBILE4.1, the
methodology used to calculate evaporative emission estimates changed. 
Although the data were still stratified according to fuel delivery
system and model year, the distinction between non-tampered and
tampered vehicles was not made.  Rather, the data were stratified
according to whether the vehicle passed an evaporative system
functional check.  This test, which was developed by EPA for use in
inspection and maintenance programs, consists of two parts: a
pressure test and a purge test.  The pressure test is designed to
assess the integrity of the fuel tank (including the gas cap) and
vapor line leading from the fuel tank to the evaporative canister. 
Under this procedure, the fuel tank vapor line is disconnected at the
canister, a pressure gauge is connected, and the tank is pressurized
with nitrogen (through the vapor line) to 14 inches of water.  If the
pressure drops below 8 inches of water in a two-minute period, the
system is considered to be defective.  As the name implies, purge
testing is intended to identify defects in the evaporative purge
system.  EPA's procedure consists of placing a flow meter in the
vapor purge line between the canister and the engine, and monitoring
cumulative flow (in liters) over a dynamometer-based test cycle
(i.e., EPA's IM240 test).  If the cumulative flow during the test is
less than 1.0 liter, the system fails.

It is very difficult to detect evaporative control system defects
based on a visual inspection (which was the basic approach in
defining tampered and non-tampered vehicles for the MOBILE4
methodology), and the functional checks outlined above provide a much
better indication of malfunctioning systems.  For that reason, EPA
opted to stratify data according to three failure regimes in
MOBILE4.1 and MOBILE5: pass


                                           -66-

pressure/purge, fail pressure, and fail purge. (Vehicles that fail
both pressure and purge tests are treated as pressure failures in the
model because emissions from pressure failures are generally higher
than emissions from purge failures.) The emission data collected for
use in MOBILE4.1 and MOBILE5 were, therefore, tested according to the
pressure/purge procedure prior to emissions measurements.

5.3 PRESSURE/PURGE FAILURE RATES

Model-year-specific evaporative emission rates are calculated within
MOBILE4.1 and MOBILE5 by weighting the emission rates of the three
failure regimes according to the expected occurrence of failures in
the fleet.  Thus, it was necessary for EPA to develop estimates of
the fraction of vehicles failing these tests.  The failure rates
estimated by MOBILE4.1 and MOBILE5 are a function of vehicle age, and
were based on data collected from the Hammond, Indiana, I/M program. 
The fraction of pressure and/or purge failures predicted in MOBILE4.1
and MOBILE5 are compared in Figure 55.  As demonstrated in the
figure, a substantial fraction of vehicles are expected to have
malfunctioning evaporative control systems beyond about 10 years of
age. (As explained in Section 7, segregating emissions according to
pressure/purge failure status facilitates estimates of pressure/purge
check effectiveness in I/M program modeling.)

Click HERE for graphic.

                                           -67-
                       5.4 HOT SOAK AND DIURNAL EMISSION MODELING

MOBILE4 - Both non-tampered and tampered emission estimates for hot-
soak and diurnal emissions are based on EPA's emission factor data
and EPA data on special studies of RVP and temperature effects. 
These data are from tests conducted by EPA at Ann Arbor and at
Automotive Testing Labs in Ohio.  The hot-soak emissions are based on
the one-hour soak test following the driving portion of the Federal
Test Procedure (FTP).  The diurnal test is performed before the
driving portion of the FTP when the temperature of the fuel tank is
increased from the minimum temperature to the maximum temperature
(i.e., 60 to 84F) over a one-hour time period.

EPA has analyzed these data to establish relationships in MOBILE that
calculate hot-soak and diurnal emissions on a gram per test basis as
a function of RVP, temperature (minimum and maximum for diurnal), and
fuel tank level.  The diurnal estimate is a function of the
Uncontrolled Diurnal Index (UDI).

The UDI is based on the Wade equation.  This equation estimates the
uncontrolled diurnal emissions from a tank depending on the
atmospheric pressure, the RVP of the fuel, the size of the vapor
space in the tank, and tank temperature.' The Wade estimate is first
calculated for standard conditions of 9.0 psi RVP fuel, 60-84F
diurnal temperature, and 40 percent tank fill level.  The UDI is then
calculated as a ratio of the Wade estimate for the selected
conditions of RVP, temperature, and tank level, and the Wade estimate
for the standard conditions.  Hotsoak emissions are calculated as a
function of RVP and then corrected for temperature.

Table 8 gives the various estimates for non-tampered hot-soak and
diurnal emissions used in MOBILE4 for 1981+ LDGVS.  The RVP value
used in evaporative emissions estimates is assumed to have weathered
(i.e., lowered) from the RVP at the fill level to the in-use tank
level.  Fuel weathering is based on the fill-up RVP level and ambient
temperature.  The tampered estimates for hot-soak and diurnal
emissions are also based on functions of RVP and temperature.  The
tampering rates were calculated for fuel cap removal and canister
disconnection in MOBILE4.


Table 8. 1981 + LDGV Non-Tampered Emission Equations
Used in MOBILE4


Fuel System                Hot Soak                  Diurnal
Carb                       -4.51 + 0.72 x RVP        2.35 x UDI
TBI                        -2.04 + 0.32 x RVP        2.10 x UDI
MPFI                       -2.26 + 0.39 x RVP        1.75 x UDI


                                           -68-

once the gram/trip hot-soak and full-day diurnal emissions are
calculated as described above, daily values are estimated.  These are
required because the daily estimates correct the emissions for
driving effects.  These effects are based on driving pattern data
from NPD diary data.

In MOBILE4, daily hot-soak emissions are calculated as:

              HS = HS * TPD * 0.76

where TPD = trips per day (function of vehicle age).

The factor of 0.76 is used because EPA estimates that no trips are
taken on 24 percent of all days.  Therefore, if there are no trips,
then there can be no hot-soak emissions, so only 76 percent of the
vehicle days have hot-soak emissions.

The daily diurnal emissions are calculated in three components:

             full diurnals,

             partial diurnals, and

             multiple diurnais.

Full diurnals occur as vehicles experience the entire diurnal
increase in temperature.  They are calculated as:

              FDI = DI * 0.34

where 0.34 is the frequency of occurrence of these days.

Partial diurnals are caused as vehicles experience diurnal
temperature increases between trips.  EPA assumes that these occur
between specific times of the day and are calculated as:

              PDI = (D1 x 0.32) + (D2 x 0.071) + (D3 x 0.037)

where         D1 = Diurnal from 8 a.m. to 11 a.m.,
              D2 = Diurnal from 10 a.m. to 3 p.m., and 
              D3 = Diurnal from 8 a.m. to 2 p.m.


Multiple diurnals occur as vehicles experience successive days
without a trip, resulting in canister overloading and higher
emissions.  In MOBILE4, these are calculated as:

       MDI = TPDI * 0.8 * 0.16

where TPDI = tampered diurnal emissions.

The factor of 0.8 is used because EPA assumed that the second day
diurnals are 80 percent of the tampered emission level.  The 0.16
factor


                                           -69-

is the EPA estimate for the frequency of occurrence of these type of
diurnals.

The total diurnal emission rate is the sum of FDI, MDI, and PDI.  It
should be noted that the fractions of FDI, MDI, and PDI do not sum to
1. That is because a certain fraction of vehicles are assumed not to
be parked long enough to undergo a diurnal emission cycle (e.g.,
delivery vehicles).

Although MOBILE calculates diurnal and hot soak emissions
independently, it reports these emission categories as a combined
gram per mile "evaporative" emission rate. (Also included in the gram
per mile "evaporative" rate are crankcase emissions.) To generate
these gram per mile estimates, the following formula is used:

{(HS(g/trip) * trip fraction * trips/day)+DI(&/day)) + CC(g/mi) [6-1]
                           mi/day       

where CC represents crankcase emissions in g/mi.

MOBILE4.1 and MOBILE5 - To model hot soak emissions, MOBILE4.1 and
MOBILE5 also segregate the vehicle fleet according to fuel delivery
system (i.e., carbureted, throttle-body fuel-injected, and multipart
fuel-injected).  However, rather than defining non-tampered and
tampered vehicles, the distinction is made as to whether the vehicle
has passed the pressure/purge functional evaporative system check. 
The baseline emission estimates are calculated by determining the
emission rate of the passing and failing vehicles separately, and
then weighting the results by the expected occurrence in the fleet.

Hot soak emissions are adjusted to account for the input RVP and
temperature. (The RVP used in determining emission rates is also
adjusted downward to account for fuel tank weathering, and the
temperature used to determine hot soak emissions is a function of the
input minimum and maximum daily temperature.) As an example, the
MOBILE5 hot soak emission rate for "passing' multipart fuel-injected
vehicles as a function of RVP is shown in Figure 56, while the
multiplicative temperature correction factor is illustrated in Figure
57.

As an example of the weighting procedure, MOBILE5a was run for the
1995 calendar year under typical summer conditions (i.e., 70 - 95F;
7.8 RVP fuel).  For a 1992 model year LDGV, the model predicts 0.3%
would be carbureted, 20.2% would have TBI, and 79.5% would be
equipped with MPFI.  Emission rates from vehicles passing the
pressure/purge test are predicted to be as follows:

              Carb  - 1.996 g/trip
              TBI   - 0.824 g/trip
              MPFI  - 0.283 g/trip


                                           -70-

Click HERE for graphic.



Click HERE for graphic.


Thus, the composite hot soak emission rate for passing vehicles is
calculated as:

       HSpass       = 1.966 g/trip * 0.003                  (Carb)
                    + 0.824 g/trip * 0.202 * 0.88           (TBI)
                    + 0.283 g/trip * 0.795 * 0.88           (MPFI)

       HSpass       = 0.350 g/trip

Note that the TBI and MPFI vehicles include a correction factor of
0.88. This accounts for the difference in average fuel tank fill
level observed in-use (55%) versus the test procedure fill level
(40%).  Lower emissions are predicted from the higher fuel tank level
because there is less vapor space in the tank, and therefore less
vapor generation.

For the hot soak estimates, failed pressure and failed purge emission
rates are not stratified according to fuel delivery system.  Under
the temperature and RVP conditions specified above for this MOBILE
run, the hot soak rates are:

       HSfail purge = 4.305 g/trip, and
       HSfail pres  = 4.357 g/trip.


Finally, the above emission rates are composited according to the
expected fraction of pass, fail purge, and fail pressure vehicles in
the fleet.  In this example (i.e., a 1992 model year vehicle analyzed
in 1995), the pass/fail regime sizes are:

       Pass                = 93.15%
       Fail Purge          = 2.11%
       Fail Pressure       = 4.74%


Thus, the 1992 model-year hot soak emission rate is:

       HS1992MY     = (0.9315 0.350 g/trip) + (0.0211 * 4.305 g/trip)
                      (0.0474 4.357 g/trip) = 0.623 g/trip.


In modeling diurnal losses, MOBILE4.1 and MOBILE5 also segregate the
fleet according to fuel delivery system and whether the vehicle has
passed the pressure/purge functional system check.  Emissions are
determined as a function of temperature (minimum and maximum, input
by the user) and fuel RVP in a manner similar to that described above
for MOBILE4.  The model calculates diurnal rates for "partial"
diurnals (i.e., diurnals that are not completed before the vehicle is
driven again), full diurnals, and "multiple" diurnals (i.e., cases in
which a vehicle sits idle for more than one day).  These values are
weighted according to their expected occurrence in the fleet to
arrive at a single diurnal emission rate in grams/day.

For MOBILE4.1 and MOBILE5, the fraction of vehicles undergoing
partial, full, and multiple diurnals was modified to reflect changes
in this


                                           -72-

distribution as a function of vehicle age.  This is illustrated in
Figure 58, which shows the fractional occurrence of diurnal episodes
as a function of vehicle age.  As seen in the figure, the fraction of
fullday and multi-day diurnals increases with vehicle age as these
vehicles are used less.

To illustrate, the diurnal emission rates (already weighted for fuel
delivery system and pressure/purge failure status) and fraction of
each diurnal episode are summarized below for the 1992 model-year
LDGV considered in the above example:

       Diurnal Episode           Fraction     Emission Rate (g/event)
       Full-Day                  0.3325       2.180
       8 a.m. - 11 a.m.          0.3408       0.398
       10 a.m. - 3 p.m.          0.0747       0.547
       8 a.m. - 2 p.m.           0.0390       0.718
       Multi-Day                 0.1423       7.554
       Composite:                0.9293       2.004


It is interesting to note that for this example, the multi-day
diurnal episode is the largest contributor to the composite daily
emission rate, representing 54% of the composite rate. (Recall that
the sum of the diurnal episodes does not equal 1.)


Click HERE for graphic.


                                           -73-

As with the MOBILE4 methodology, MOBILE4.1 and MOBILE5 combine hot
soak and diurnal emissions with crankcase emissions into a single
"evaporative" gram per mile emission rate.  This is accomplished by
multiplying the hot soak emission rate by the number of trips per day
and the daily trip fraction; adding this daily hot soak rate to the
daily diurnal rate; dividing this sum by the miles per day; and
adding the crankcase emissions.  This calculation, represented by
equation 6-1, is shown below for the example presented above (i.e.,
a 1992 model year LDGV in 1995), and the average trips per day and
miles per day assumed in MOBILE5a are illustrated graphically in
Figure 59.

{(O.623 g/tr*0.802*4.48 tr/day) + 2.00 9/day) + 0.006 g/mi=0.128 g/mi
                    34.8 mi/day


The above calculation is carried out over the 25 model years
comprising the fleet.  The fleet-average "evaporative' emission rate
is then calculated by weighting the model-year-specific values by the
assumed travel fraction for each model year.

Click HERE for graphic.



                                           -74-

5.5 RUNNING LOSSES

Running losses, in g/mi, are calculated by MOBILE with a
comparatively simple algorithm.  Although a distinction is not made
by fuel delivery system, vehicles are stratified according to whether
they pass the functional pressure/purge test. (In MOBILE4, the
distinction was made between non-tampered and tampered.) Emissions
are determined by interpolating among data tables that list running
loss emission rates (in g/mi) at four temperatures (80, 87, 95,
and 105F) and four different fuel RVP levels (7.0, 9.0, 10.4, and
11.7). Once emission rates are derived for the two emissions
categories, these are weighted according to their expected occurrence
in the fleet.

MOBILE4 - Figure 60 shows the MOBILE4 non-tampered running loss
estimate for 1981+ light-duty vehicles at various RVP and
temperatures.  These data are used to estimate running losses at
other RVPs and temperatures by linear interpolation.  No running
losses are calculated for temperatures below 40F.  RVP values of 7.0
psi and 11.7 psi are used as extreme ranges for allowable RVPs.  RVPs
below or above this range are considered to be equal to end points of
the range.  Similarly, temperatures above 105F are considered to be
equal to 105F.

Click HERE for graphic.

                                           -75-

A comparison of MOBILE4 non-tampered running loss emissions for LDGV
by model year is given in Figure 61.  This figure demonstrates a
significant reduction in running loss emissions (for non-tampered
evaporative emission control systems) as a result of the SHED-based
2.0 gram/test emission standard that became effective with the 1981
model year.

Click HERE for graphic.

In MOBILE4, the tampered running losses are calculated for two
tampering categories: gas cap removal and canister disconnection. 
Table 9 shows the tampered emissions used for all vehicle types at
various RVP and temperature levels.  The tampering rates for gas cap
removal and canister disconnect are calculated as a function of
vehicle miles based on EPA's tampering data.  Figure 62, which shows
a comparison of tampered to non-tampered running loss emissions,
graphically illustrates the large impact of gas cap and canister
removal.



                                           -76-

Click HERE for graphic.


Click HERE for graphic.


                                           -77-

MOBILE4.1 and MOBILE5 - The basic methodology used to estimate
running loss emission rates remained the same between MOBILE4 and the
MOBILE4.1 and MOBILE5 models.  However, rather than segregating
emissions according to tampering status, MOBILE4.1 and MOBILE5
segregate emission rates by pressure/purge failure status.  Running
loss emission rates from the MOBILE5 model are shown in Figures 63
and 64 for vehicles passing and failing the pressure/purge test,
respectively. (For running losses, pressure failures and purge
failures are assumed to have the same emission rate.) These figures
indicate that higher temperatures and RVPs result in higher running
loss emissions.

A comparison of MOBILE5 running loss predictions for 'passing" and
"failing' vehicles is presented in Figure 65 for 9.0 RVP fuel.  As
seen, the emission rates from vehicles failing the pressure/purge
test are much higher (by nearly an order-of magnitude) than emissions
from those passing the test.

Finally, in comparing the MOBILE5 results to those obtained with
MOBILE4, the following is observed:

             The "passing' emission rates in MOBILE5 are slightly lower
              than the "untampered" rates estimated by MOBILE4.

             The MOBILE5 "failing" emission rates are similar to the
              MOBILE4 gas-cap-tampered estimates at temperatures below
              95F, and lower at temperatures above 95F.


Click HERE for graphic.

                                           -78-

Click HERE for graphic.




Click HERE for graphic.


                                           -79-
             The MOBILE5 "failing" emission rates are higher than the
              MOBILE4 canister-tampered emission rates.

The above emission differences are not unexpected, given the
different methods used to identify the MOBILE4 "tampered" and MOBILE5
"failing' vehicles.

5.6 RESTING LOSSES

Resting losses were incorporated into the MOBILE model with the
MOBILE4.1 release.  The model calculates resting losses in
grams/hour, and emissions are a function of temperature and the type
of canister in the evaporative control system (i.e., open bottom
versus closed bottom).  As an example', MOBILE5 estimates resting
loss emissions for a vehicle with an open bottom canister to be 0.29
grams/hour at 100F, while the value for a closed bottom canister
under the same conditions is 0.13 grams/hour.  Resting losses are
also converted to an equivalent g/mi value by summing emissions
throughout the day and dividing by miles/day.

5.7 REFUELING LOSSES

Two components of refueling emissions are considered in the MOBILE
models: vapor space displacement and spillage.  Vapor space
displacement is a function of fuel tank temperature, dispensed fuel
temperature, and dispensed fuel RVP, and emissions are reported as
grams/gallon (g/gal).  Spillage is considered a constant at 0.31
g/gal.  The user has the option of specifying uncontrolled rates or
including the impact of a vapor recovery system in the output.  EPA
has chosen to model the impact of an on-board vapor recovery system
by assuming a 98 percent reduction in uncontrolled displacement
emissions and a 50 percent reduction in spillage.  Stage II vapor
recovery is modeled by applying a user-specified percent reduction in
uncontrolled displacement emissions (no benefit is assumed for
spillage).  The g/gal values are then converted to g/mi by dividing
by the average fuel economy for the vehicle class and model year
being considered.

5.8 CRANKCASE EMISSIONS

Crankcase emissions are largely controlled by positive crankcase
ventilation (PCV) systems that have been installed on new vehicles
since the early 1960s.  Crankcase emissions, therefore, are the
result of tampered or defective PCV systems.  MOBILE models crankcase
emissions by assuming a certain tampering rate (which increases with
vehicle age) and multiplying that value by an uncontrolled crankcase
emission rate.  Overall, crankcase emissions are very small,
generally contributing only a few hundredths of a gram per mile to
the HC emission rate.


                                           -80-

                                6. MODELING OF I/M PROGRAMS


On-road motor vehicles have been required to meet increasingly
stringent emission standards when new for over 25 years, but they
still are significant contributors of ozone precursors and carbon
monoxide in most urban areas.  Much of this contribution is
attributed to vehicles that exceed certification standards in
customer service.  Although EPA has implemented programs to improve
in-use emission control system durability (e.g., in-use recall,
100,000-mile certification standards), a properly designed inspection
and maintenance (I/M) program remains one of the most effective means
of ensuring that high-emitting vehicles are identified and repaired. 
Congress recognized the importance of I/M programs in reducing
vehicular emissions in urban areas when drafting the Clean Air Act
Amendments of 1990, and directed EPA to develop performance
requirements and other standards for basic and enhanced I/M programs. 
These guidelines were published in November 1992,15 and specify a
number of features associated with both program types.

The methodology by which MOBILE accounts for I/M benefits differs
between exhaust and evaporative emissions.  The exhaust procedure
follows from the discussion of the Tech IV model contained in Section
5, and I/M credits (by model year) are estimated by recalculating the
emitter regime emission levels as a result of an I/M repair.  The
emission reductions calculated by MOBILE are a function of I/M
program type (i.e., centralized versus decentralized), I/M test type
(i.e., idle versus loaded mode), inspection frequency (annual versus
biennial), and I/M emission cutpoints.  The overall effectiveness of
a particular program is strongly influenced by assumptions made by
EPA regarding how well a program identifies and repairs a failing
vehicle. (Although the calculation methodology is similar among
MOBILE revisions, the discussion that follows has utilized
information from MOBILE5/Tech5.)

An I/M benefit for evaporative emissions can be calculated by the
MOBILE4.1 and MOBILE5 versions of the model, and it is applied
according to whether or not the functional pressure and purge tests
are in place. (For MOBILE4, a small evaporative benefit is calculated
if the evaporative system is included in an ATP program.) The
evaporative emissions benefit is calculated in MOBILE4.1 and MOBILE5
by changing the fraction of vehicles passing and failing the
pressure/purge test and recomputing the evaporative emission rate. 
That process follows from the discussion of pressure/purge failure
and emission rates contained in Section 6.

The discussion below summarizes the methodologies used in MOBILE5 to
estimate emission benefits (both exhaust and evaporative) from I/M
programs.  In addition, EPA's performance standards for basic and
enhanced I/M programs are reviewed.

                                           -81-

                             6.1 EXHAUST EMISSION I/M MODELING

When an I/M program option is invoked in a MOBILE run, the base
emission rate ("BER" in equation 4-1) is adjusted to reflect the
presence of that program.  The adjustment takes into account the
features of the I/M program being modeled, such as:

             program type (centralized or decentralized),
              inspection frequency (annual or biennial),
             test type (idle, idle/2500, loaded idle, or IM240),
             emission cutpoints (for IM240),
             waiver rate, and
             compliance rate (i.e., the fraction of vehicles subject to
              the program that complete the process to the point of
              receiving a certificate of compliance or a waiver).

To perform the I/M adjustment on the BER, the following methodology
is employed:

              BERI/M = BERNon-I/M * {1 - [CREDI/M * (1  WVR)] * ADJcomp1}

where CREDI/M is the I/M credit for the test type, cutpoints,
frequency, model year, and vehicle age being considered; WVR is the
user-input waiver rate; and ADJcomp1 an adjustment that accounts for
the user input compliance rate.  The above calculation is valid for
centralized program types; if a decentralized program is specified,
the calculation includes an adjustment that reduces the overall
effectiveness of the program by 50%.

As an example, consider a calendar year 2000 MOBILE5a run in which an
annual, centralized IM240 program is specified with 0.8 g/mi HC and
20.0 g/mi CO cutpoints.  Further, assume that the I/M compliance rate
is 96% and the waiver rate is 3%. (These are the MOBILE parameters
that EPA has chosen for developing the enhanced I/M performance
standards.) The non-I/M HC base emission rate for a 1992 model-year
LDGV (analyzed in the year 2000) is 2.153 g/mi, and the base emission
rate including the effects of the above I/M program is:

       BERI/M = 2.153 * {1 - [0.476 * (1 - 0.03)] * 0.92) 
       BERI/M = 2.153 * {1 - 0.425) = 1.238 g/mi.


Several items are worth noting in the calculation above.  First, the
I/M credit is reduced from 47.6% to 42.5% when accounting for the
effects of the waiver rate and the compliance rate.  Second, the
adjustment performed to account for the compliance rate is nonlinear
(i.e., although the compliance rate was 96% in the above example, the
I/M benefits are reduced by 8% rather than 4%).  This adjustment
inherently assumes that the I/M failure rate of non-complying
vehicles will be -higher than that of the rest of the fleet.16

                                           -82-

The I/M credits (i.e., CREDI/M in the above equation) are stored in
data matrices that are read by MOBILE.  For 1981 and later model
years, these credits are developed by EPA's "Tech" models, and are a
function of test type, inspection frequency, model year, and vehicle
age.  A summary of how the Tech models develop the I/M credits is
provided below.

Tech5 Model - The Tech5 model was used to generate the I/M credits
(i.e., "CREDI/M in the above equation) for use in MOBILE5.  To develop
those credits, Tech5 compares emission rates under no I/M program to
the emission rates that would occur in the presence of an I/M program
(neglecting compliance and waiver rates, which are accounted for
within MOBILE).  The emission rates for a particular I/M case are
calculated by Tech5 (by technology class and emitter regime)
according to the following approach:

Emission RateIM - {(1-IDRATE)*EMIS} + (IDRATE*EMIS*(1-REF))

where IDRATE is the identification rate specific to the I/M test type,
emitter group, and technology group; EMIS is the baseline emission
rate of the emitter group being analyzed; and REF is the assumed
repair effectiveness.

As with the Tech IV model, emissions data are stratified by
certification standard, technology (i.e., closed-loop multipart fuel-
injection, closed-loop throttle-body injection, closed-loop
carbureted, and open-loop), and emitter group (or "regime").  The
emitter groups utilized in the Tech5 model are:

             Normal:       2 x HC Std. and  3 x CO Std.

             High:        > 2 x HC Std. or > 3 x CO Std.

             Very High:   > 4 x HC Std. or > 4 x CO Std.

             Super:       > 10 g/mi HC or > 150 g/mi CO

In addition, NOx emissions are stratified into low and high NOx
regimes, with high NOx emitters being greater than 2.0 g/mi.

Baseline Emission Rates - Once the emission data are stratified as
outlined above, the baseline emission rates (as a function of vehicle
mileage or age) are determined by multiplying the emission rate of
each regime by the fraction of each regime making up the fleet at the
mileage intervals corresponding to vehicle age.  As an example, the
incidence of emitter groups in the fleet is illustrated in Figure 66
for 1983 to 1993 model year multipart fuel-injected (MPFI) light-duty
gasoline vehicles, and the exhaust HC emission rate of each regime
(as a function of vehicle age) is shown in Figure 67.  Combining the
data illustrated in Figures 66 and 67 results,in the baseline HC
emission rate shown in Figure 68.

After-I/M Emission Rates - The Tech5 model then determines an "after-
I/M" emission rate (for each vehicle age) that is based on the
identification rate and repair effectiveness of the specific I/M
program


                                           -83-

Click HERE for graphic.



Click HERE for graphic.



                                           -84-

Click HERE for graphic.


type (e.g., IM240) and cutpoints being modeled.  The identification
rate and repair effectiveness developed by EPA are a function of
emitter category, with the identification of failing vehicles being
close to 100 percent for the Super and Very High categories for most
cutpoints.  As an example, the identification rates assumed for 1983
to 1993 model year vehicles by emitter group and technology are shown
in Figure 69 based on 0.8/15.0 HC/CO cutpoints (i.e., vehicles with
an emission rate of over 0.8 g/mi HC or 15.0 g/mi CO on the IM240
test are considered failing vehicles under these cutpoints).

The after-repair emission rates are variable, according to I/M test
type and model year, but for 1983 to 1993 model years, the after-
repair HC rates (based on IM240 testing) are generally about 0.5 g/mi
for Normals, 0.8 g/mi for Highs, 1.0 g/mi for Very Highs, and 0.9
g/mi for Supers. The after-repair emission rates are depicted
graphically in Figure 70 for this model year group.

The after-repair composite emission rates from the Tech5 model for
light-duty MPFI vehicles are shown in Figure 71 as a function of
emitter category (based on IM240 testing and 0.8/15.0 HC/CO
cutpoints).  The figure indicates a significant reduction in emission
level as a result of an I/M program, and most of the reduction is
from the Super and Very High emitter categories.



                                           -85-

Click HERE for graphic.


Click HERE for graphic.


                                           -86-

Click HERE for graphic.

To develop model-year specific emission rates and I/M program 
benefits, the above calculations are performed for each standard
group and technology group.  These results are then weighted
according to the expected fraction of each group in the fleet.  As an
example, the baseline and after-I/M emission rates as a function of
vehicle age are given in Figure 72 (again based on IM240 testing and
0.8/15.0 HC/CO cutpoints).  The I/M credit matrices developed for use
in the MOBILE models are determined from these two lines through an
algorithm that applies the non-I/M deterioration rate (i.e., the
slope of the top line) to the I/M emission rate at each vehicle age
(for an annual program) to arrive at the final I/M emission rate. 
This results in the traditional "stair-case" emission rate normally
associated with modeling of I/M programs.

6.2 EVAPORATIVE EMISSIONS I/M MODELING

Pressure/Puree Testing - EPA's recommended evaporative system
functional test procedure consists of two distinct test types: a
pressure test and a purge test.  The pressure test is designed to
assess the integrity of the fuel tank (including the gas cap) and
vapor line leading from the fuel tank to the evaporative canister. 
Under this procedure, the fuel tank vapor line is disconnected at the
canister, a pressure gauge is connected, and the tank is pressurized
with nitrogen (through the vapor


                                           -87-

Click HERE for graphic.



line) to 14 inches of water.  If the pressure drops below 8 inches of
water in a two-minute period, the system fails.  As the name implies,
purge testing is intended to identify defects in the evaporative
purge system.  EPA's test procedure consists of placing a flow meter
in the vapor purge line between the canister and the engine, and
monitoring the cumulative purge flow (in liters) over the 4-minute
IM240 transient test.  If the cumulative flow during the test is less
than 1.0 liter, the system fails.

Benefits of Evaporative System Functional Tests - Section 6 described
the process used in MOBILE4.1 and MOBILE5 to generate baseline (i.e.,
no pressure/purge test in effect) evaporative emission rates.  That
method relies on estimates of the fraction of vehicles (by vehicle
age) passing the pressure/purge test, failing the pressure test, and
failing the purge test.  Those fractions are then applied to the
emission rates applicable to each of the three failure categories to
arrive at a composite model-year-specific emission rate.  Under an
I/M program, the benefits of pressure/purge testing are estimated by
modifying the distribution of vehicles expected to pass, fail purge,
or fail pressure, and then re-computing the model-year emission rate.

The effect of vehicle repair on pressure/purge failure rates as a
function of vehicle age is depicted graphically in Figure 73.  It
shows the MOBILE5a baseline failure rate (from Figure 55) and the
after-repair failure rate as a result of an annual pressure/purge
check.  This figure demonstrates the classical saw-tooth pattern
associated with inspection



                                           -88-


Click HERE for graphic.


and maintenance cycles (i.e., failing vehicles are repaired, but in
the next cycle there are additional failures).  It is assumed that
the failure rate between inspections will increase at the same rate
as that observed for the same-age vehicles under the baseline case,
and the after-repair failure rate is 5 percent of the before-repair
failure rate (i.e., 95 percent identification and repair
effectiveness is assumed).  This can be thought of as translating the
slope of the top line of Figure 73 (e.g., between years 10 and 11)
onto the after-repair failure rate depicted by the lower points of
the bottom line in Figure 73 (e.g., at year 10).

Note that Figure 73 indicates that there is a cyclical nature to the
pressure/purge failure rate under an I/M scenario.  That is because
EPA assumes that a certain fraction of vehicles repaired early in
their lives fail again at a later date.  However, this approach was
incorrectly applied in the MOBILE5 version, which resulted in near-
zero failure rates at years 14 and 23.  Conversations with EPA staff
revealed that this error has only a small impact on the MOBILE5
results, and it will likely be corrected in the next revision to the
model.


                                           -89-

6.3 EPA'S I/M PERFORMANCE STANDARDS

EPA's performance standard for enhanced I/M programs is commonly
reported to be 28% for HC, 31% for co, and 9% for NOx, the emission
reductions cited in the I/M final rule.15 However, these percentages
were based on the MOBILE4.1 version of EPA's emission factor model
and reflect the reduction in emissions for the entire on-road motor
vehicle fleet when compared to a non-I/M case.  The actual
performance standards for both basic and enhanced I/M programs are to
be calculated for each nonattainment area using the latest version of
the MOBILE model run under a very specific set of I/M input
parameters.  The performance standard, calculated as g/mi (or
tons/day), is then compared to the MOBILE results that reflect the
proposed local I/M program being modeled.  If emission rates from the
proposed program equal or fall below the performance standard, the
program can be approved by EPA.

This approach allows nonattainment communities the flexibility to
develop an I/M program different from the "model" programs set forth
by EPA in the I/M final rule.  For example, although the "model"
programs specify annual testing, the benefits from biennial testing
are nearly as great at a substantially reduced cost.  Thus, if a
biennial program is chosen, the reduced benefits can be made up by
expanding.the I/M program to include additional vehicle classes
(e.g., HDGVS) or extending the model year coverage for some test
types (e.g., require pressure testing on 1971 and later vehicles
rather than 1983 and later vehicles as specified in the "model"
program).

MOBILE Input Parameters - Although the I/M final rule lists the I/M
program inputs to be used in establishing I/M performance standards,
some of these inputs (e.g., 1.4 g/mi IM240 NOx standard for Tier I
vehicles under enhanced I/M) cannot be used with the MOBILE model.
Subsequent guidance 17 , however, has established the proper inputs
for communities to use in developing I/M performance standards for
basic and enhanced programs.  These inputs are summarized in Table
10, which shows that only 1968 and later model-year LDGVs subject to
an idle-based emissions test are considered in developing the
performance standards for the basic program.  On the other hand,
LDGVS, LDGTls, and LDGT2s are included in the enhanced I/M
requirements.  These requirements also specify IM240 testing for 1986
and later vehicles, evaporative system pressure testing for 1983 and
later, and evaporative purge testing for 1986 and later.

In addition to the parameters listed in Table 10, the MOBILE input
files prepared to establish the I/M performance standard should
include area specific inputs for the following:

             registration distributions by vehicle class;
             VMT mix (i.e., percent travel by vehicle class);
             minimum and maximum temperature;
             ASTM class (used to model the effects of reformulated
              gasoline);
             fuel RVP;


                                           -90-


Table 10. MOBILE5a Parameters Used to Model the Performance
Standard
Targets for Basic and Enhanced I/M Programs


Input Parameter            Basic I/M          Enhanced I/M

Network:                   Centralized        Centralized
Frequency:                 Annual             Annual
Model Years:               1968+              1968+
Vehicle Types:             LDGV               LDGV,LDGT1,LDGT2
Emission Test(s):          Idle on 1968+      Idle on 1968-80
                                              Idle/2500 on 81-85
                                              IM240 on 1986+a
Start Date:                1983/1994b         1983/1995b
Pressure Test:                   None                1983+
Purge Test:                      None                1986+
Visual Check:                    None         Catalyst & Fuel Inlet
Pre-81 Stringency:               20%                 20%
Pre-81 Waiver Rate:              0%                  3%
Post-80 Waiver Rate:             0%                  3%
Compliance Rate:                 100%                96%

       a IM240 cutpoints: 0.8 g/mi HC, 20.0 g/mi CO, 2.0 g/mi NOx.
       b1983 is used for areas with existing I/M programs, 1994 or 1995
       is used for areas newly subject to I/M under the 1990 CAAA.


             altitude; and
             all other mobile source programs modeled by MOBILE that are
              applicable for the case being analyzed (e.g., oxygenated
              fuels, Stage II vapor recovery).


I/M Performance Standards - The MOBILE input parameters specified in
Table 10 were used to generate I/M performance standards for basic
and enhanced I/H programs based on national default MOBILE5a runs -
for the year 2000.  The MOBILE5a output for HC and CO is summarized
in Table 11, which indicates an 8% reduction in on-road motor vehicle
HC emissions under the basic program and a 34% reduction in HC
emissions under an enhanced program.  The CO benefits are 11% and 36%
for basic and enhanced programs, respectively.


                                           -91-

Table 11. Basic and Enhanced I/M Performance Standards Based on
MOBILE5a (Year 2000, 750F, 9.0 RVP, 19.6 mph)



                        Hydrocarbonsa         Carbon Monoxide
I/M Scenario        g/mi   Reduction          g/mi          Reduction
No I/M              2.47         -            21.62               -
Basic I/M           2.28         8%           19.28               11%
Enhanced I/M        1.64         34%          13.83               36%

a      Includes exhaust and all evaporative components except refueling
       losses.


                                           -92-

                               7. CLEAN AIR ACT REQUIREMENTS


The Clean Air Act Amendments (CAAA) of 1990 included a number of new
requirements for on-road motor vehicles.  In addition, Section 130 of
the CAAA directed EPA to review and revise emission factors used to
estimate emissions of VOC, CO, and NOx.  Thus, many of the revisions
included in MOBILE4.1 (released November 1991) and MOBILE5 (released
December 1992) were in direct response to the directives of the CAAA. 
Among the requirements necessitating revisions to the model are
heavy-duty truck emission standards, light-duty vehicle cold CO
emission standards, revised evaporative emission test procedures,
more stringent ("Tier I") emission standards for light-duty vehicles,
and reformulated and oxygenated gasolines.  The following describes
how the model was revised to account for the effects of these CAAA
requirements. (Although not a CAAA requirement, a description of how
the California Low-Emission Vehicle program is modeled in MOBILE5 is
also included in this section.)

7.1 HEAVY-DUTY TRUCK EMISSION STANDARDS

Promulgation of new NOx emission standards for heavy-duty gasoline
and Diesel vehicles is required by the CAAA.  Beginning with the 1998
model year, the NOx standard will be changed from 5.0 grams per brake
horsepower-hour (g/bhp-hr) to 4.0 g/bhp-hr.  This new standard was
modeled in MOBILE5 by simply applying the ratio of the standards
(i.e., 4.0/5.0) to the zero-mile component of the 1998 model year
base emission rate equation.

7.2 COLD CO EMISSION STANDARDS

Section 204 of the CAAA directs EPA to promulgate low-temperature CO
standards for light-duty vehicles and light-duty trucks beginning
with the 1994 model year.  The regulations are to be phased-in, with
40 percent compliance in 1994, 80 percent in 1995, and full
compliance with the 1996 model year.  Under these requirements,
light-duty vehicles must comply with a 10.0 g/mi CO emission level at
20F while operating over the three-bag FTP.  Light-duty trucks will
comply with a 12.0 g/mi standard.

To model the impact of the cold temperature CO emission standards,
EPA assumed that all of the benefit of the regulation would occur
during the cold start portion of the exhaust emission test. 
Therefore, the Bag 1 cold temperature CO offset (discussed in Section
5) was modified, while the stabilized and hot start modeling
procedure remained unchanged.  Recall that the CO emission factor
(EFCO) is represented by:


                                           -93-

       EFCO = BER*BCF + COoff] * CFn                             [8-1]

where BER is the basic emission rate, BCF is the bag (or operating
mods) correction factor, COoff is the cold temperature CO offset, and
CFn represents the various multiplicative correction factors included
in the model (e.g., speed, in-use fuel, etc.). The cold CO standard
is modeled by decreasing the CO offset by an amount that varies
linearly with ambient temperature.  The offset correction, OFFcor' is
specified as:

       OFFcor = (75-T *   fcncs ) * (13. 2 -  2.7* STD)                  [8-2]
                ( 55       0.206)   (         3.4     )

where T is the temperature at which the exhaust emission factor is
being calculated, fcncs' is the VMT fraction occurring in the cold
start mode, and STD is the numerical cold CO standard applicable to
the particular vehicle type being modeled (i.e., 10.0 g/mi for light-
duty vehicles and 12.0 g/mi for light-duty trucks).

As a numerical example, a 1997 model year LDGV was considered in the
year 2000.  Further, the vehicle was assumed to be operating 100
percent in the cold start mode at a temperature of 30F.  The
emission rate (neglecting CF. in equation (8-1]) is therefore
calculated as:

       EFCO =       [(BER * BCF) + (COoff - OFFcor)]
       EFCO =       [(7.60 *1.59) + (34.83 -20.89)]
       EFCO =       26.02 g/mi

Thus, the effect of the cold CO standard is to reduce the CO emission
level of this vehicle at 30F from 46.91 g/mi to 26.02 g/mi (a 45
percent reduction) during cold start operation.  The overall impact
of the standard is reduced when including hot start and stabilized
operating modes in the composite emission factor.

7.3 REVISED EVAPORATE TEST PROCEDURE

The CAAA directed EPA to promulgate new evaporative standards and
test procedures to control running loss emissions and multi-day
diurnal emissions under summertime, ozone conditions (i.e., elevated
temperatures).  Although the CAAA were not explicit in terms of
actual requirements or implementation schedule, the new procedures
and standards will be effective with the 1996 model year and will be
phased in over a four-year period (i.e., 20 percent compliance in
1996, 40 percent compliance in 1997, 90 percent compliance in 1998,
full compliance in 1999).

A review of the MOBILE5a code indicates that the following
methodology is used to model these requirements for vehicles passing
the pressure/purge functional test:

             An 80 percent reduction is applied to the baseline running
              loss emission rate;

                                           -94-

             Resting losses are reduced by 75 percent;

             Hot Soak emissions are reduced by 50 percent; and

             Partial and full diurnals are reduced by 50 percent, two-
              to three-day diurnals are reduced by 75 percent, and
              diurnals beyond three days are reduced by 40 percent.


For vehicles failing the functional pressure/purge test, the benefits
from the new test procedures are substantially reduced. (A review of
the MOBILE5 code indicates that no diurnal benefit is ascribed to
failing vehicles, while failing vehicle running loss and hot soak
emissions are reduced by about 20 to 30 percent.  However, the actual
reductions for running loss and-hot soak emissions are dependent upon
RVP and temperature.)

7.4 TIER I EMISSION STANDARDS

The CAAA call for more stringent (i.e., "Tier I") exhaust emission
standards for light-duty vehicles (LDVS) and light-duty trucks (LDTS)
beginning with the 1994 model year.  The standards are to be phased
in, with 40 percent compliance in 1994, 80 percent compliance in
1995, and full compliance in 1996. (Intermediate in-use compliance
standards are also included in the requirements.) LDVs and LDTs under
3750 lbs. loaded vehicle weight (LVW) would comply with 0.25 g/mi
NMHC, 3.4 g/mi CO, and 0.4 g/mi NOx, while LDTs from 3751 to 5750
lbs.  LVW would meet 0.31 g/mi NMHC, 4.4 g/mi CO, and 0.7 g/mi NOx.
(All standards above are 50,000 mile requirements; the CAAA also
specifies corresponding 100,000 mile standards.)

To model the effects of the Tier I standards, EPA adjusted the ZM
component of the base emission rate equation.  This adjustment was
based on the ratio of the Tier I standard to the existing standard
for the vehicle class being modeled.  For example, the LDGV zero-mile
level for NOx was multiplied by 0.4/1.0 to model the effects of the
0.4 g/mi NOx standard (the current NOx standard is 1.0 g/mi).  No
changes were made to the deterioration rates.

7.5 REFORMULATED AND OXYGENATED GASOLINE

Modeling the effects of oxygenated fuels was first included in
MOBILE4.1. That version of the model, however, only estimated the
impacts on exhaust CO and evaporative emissions (evaporative
emissions are impacted by virtue of the change in RVP from
commingling effects).  MOBILE5 extended the modeling of oxygenated
fuels to also include HC emissions.  MOBILE5 can model the effects of
oxygenated fuels on NOx emissions, but at the time of the MOBILE5
release, the impact of oxygenated fuels on NOx was estimated to be
negligible.

The MOBILE5 subroutine that estimates the exhaust emission benefits
from oxygenated fuels first calculates a factor that increases the
base


                                           -95-

emission rate to account for the differences in certification (i.e.,
Indolene) and in-use fuels.  This correction for HC and CO is applied
to all catalyst-equipped vehicles, whereas a correction for NOX is
only applied to vehicles equipped with three-way catalysts. (No in-
use correction is applied to non-catalyst vehicles.) These correction
factors are:

             HC = 1.157 (applied to all catalyst vehicles),

             CO = 1.087 (applied to all catalyst vehicles), and

             NOx = 1.160 (applied to three-way catalyst vehicles).

These factors were calculated from data collected in EPA's Phase I
Reformulated Gasoline Study.

The exhaust emission benefits of oxygenated fuels modeled in
MOBILE4.1 and MOBILE5 are a function of fuel oxygen content, emission
rate, and model year. (Imbedded in the model-year-specific
corrections are assumptions regarding vehicle technology.  These data
matrices were developed from a subroutine contained in the TECH5
model.) For HC emissions, the model has been structured to account
for a difference in benefits between ether and alcohol blends;
however, MOBILE5 assumes the same exhaust HC effects for ether and
alcohol blends.

To illustrate the magnitude of the projected impact of oxygenated
fuels determined by MOBILE5, Figures 74 and 75 show the calculated
benefits (as a percent reduction from the base emission rate) for CO
and HC for 1990 model year light-duty vehicles.  The benefits
increase with increasing fuel oxygen content and with increasing base
emission rate, resulting in a maximum benefit of approximately 18
percent for HC and 29 percent for CO at an oxygen content of 2.7
weight percent. (Note that MOBILE5 does not calculate HC benefits for
fuels with an oxygen content above 2.7 percent.  For cases in which
a higher oxygen content is input by the user, a value of 2.7 percent
is utilized in the calculation.)

The changes in CO benefits estimated by the MOBILE4.1 and MOBILE5
versions of the emission factors model are relatively minor as
illustrated in Figure 76.  The data contained in the figure indicate
a slightly higher benefit at lower emission rates for MOBILE4.1. What
is more interesting about this figure, however, is the fact that the
maximum CO emission rate predicted by MOBILE5 is nearly 3 times that
predicted by MOBILE4.1. Again, this demonstrates the impact of
changes to the methodology and data used in TECH5 to develop the
exhaust base emission rate equations for MOBILE5.


                                           -96-


Click HERE for graphic.


Click HERE for graphic.

                                           -97-

Click HERE for graphic.


7.6 CALIFORNIA LOW-EMISSION VEHICLE PROGRAM

Because of interest by a variety of states in adopting the California
Low-Emission Vehicle (LEV) standards (notably the northeast states),
EPA included an algorithm in MOBILE5 to model the effects of those
emission standards (including the capability to specify an
alternative start date).  Basically, EPA has taken two approaches in
modeling emission rates from LEVS:

             Standard (or No) I/M: In this case, the only adjustment
              made to the emission factor equation that accounts for the
              standard change was to the zero-mile level.  EPA has argued
              for many years that more stringent emission standards do
              not impact emission control system deterioration rates. 
              Thus, under this scenario, the zero-mile level for the
              various LEV classes (i.e., TLEVS, LEVS, ULEVS) was
              developed from the existing Tier I zero mile levels by
              applying the ratio of the LEV to Tier I standards. (Also
              included in this methodology were a variety of adjustments
              to account for NHOG versus NMHC differences, reactivity
              differences, and certification versus in-use fuel
              differences).

             "Appropriate I/M": As an alternative for those users that
              want more emission benefit associated with the LEV program,
              EPA has


                                           -98-

              developed a second methodology to estimate its benefits. 
              Under this "Appropriate I/M" approach, it is assumed that
              these vehicles meet their certification standards in
              customer service due to the presence of a very strong I/M
              program (i.e., the maximum possible I/M benefits).  Thus,
              the deterioration rates of the LEV emission factor
              equations were also adjusted downward so that emission
              standards are met in-use.  Although this results in
              significant LEV benefits, historical in-use data have never
              indicated that standards are met in customer service for an
              entire fleet of vehicles.


The difference in the emission factor equation for LEVs developed
through these two methodologies is quite dramatic, as illustrated in
Figure 77.  Clearly, the presence of just a fraction of a percent of
high-emitting vehicles could cause a significant change in the
emission rate depicted by the bottom line in Figure 77.


Click HERE for graphic.

                                           -77-


                              8. EMISSIONS VERSUS VEHICLE AGE


There has been considerable interest recently in the fraction of
emissions attributable to older vehicles versus newer vehicles.  This
fraction has ramifications in terms of where efforts should be placed
to control emissions from motor vehicles (i.e., should the focus be
on new vehicle standards or on in-use control programs).  For
example, many communities are considering vehicle scrappage programs
in which owners of older vehicles (e.g., pre-1975 model year) are
offered a bounty for their vehicle.  These vehicles, which have high
emissions relative to the fleet average, are then removed from
service (i.e., they are destroyed).  Presumably, such vehicles are
replaced by lower-emitting vehicles resulting in a net reduction in
emissions.

The emission reduction potential of such measures depends on the
fractional contribution to the fleet-average emission rate of the
model year groups included in the program.  That contribution, in
turn, depends on the vehicle class (e.g., passenger cars versus light
trucks) and the type of I/M program in place (I/M programs may have
a different impact on different model years).  This section presents
results from MOBILE4.1 and MOBILE5a depicting the contribution of
specific model year groups to the fleet-average emission rate.  The
calculations were performed for the light-duty gasoline vehicle
classes (i.e., LDGV, LDGT1, LDGT2) for calendar years 1985 through
2010 under no I/M, "basic" I/M (assumed to be idle/2500 rpm testing
for this analysis), and enhanced" I/M (i.e., IM240 and pressure/purge
testing) cases.

8.1 ANALYTICAL APPROACH

MOBILE Model-Year Output - The MOBILE4.1 and MOBILE5 versions of the
model offer the user an option to specify output on a model-year-
specific basis, and an example of this output format is contained in
Figure 78.  This output also contains the estimated travel fraction
for each of the 25 model years making up the fleet. (Section 4 of
this report describes how the travel fraction is determined.) To
generate the fleet-average emission rate, the model-year-specific
emission rates are multiplied by the corresponding travel fraction
and summed over all model years.

For this analysis, it was necessary to process the model-year output
in a series of spreadsheets that combined the emission rates and
travel fractions (i.e., VMT fractions) associated with certain model-
year groups.  In this way, the fractional contribution of model-year
groups to the fleet-average emission rate was determined.


                                           -101-

Click HERE for graphic.

                                           -102-


Model-Year Groupings - The model-year groupings chosen for this
analysis were based on the emission standards to which the vehicles
were certified, as well as the expected emission control technology
used to meet those standards.  This was discussed previously in
Section 5.4, and Table 6 lists the emission standards and predominant
emission control technologies for LDGV as a function of model year.
(Similar tables for LDGT1s and LDGT2s are contained in Appendix A.)
Based on this table, the model year groupings chosen for analyzing
LDGV emission rates are as follows:

             pre-1968,
             1968-1971,
             1972-1974,
             1975-1980,
             1981-1993, and
             1994+.

The same approach was also used to develop the model year groupings
for LDGT1s and LDGT2s.

MOBILE Inputs - Because the comparisons presented in this section are
intended to reflect differences in emission rates, FTP conditions
were used when developing the input files for the MOBILE runs (i.e.,
75F, 19.6 mph, FTP operating mode fractions, 9.0 RVP fuel).  For
cases in which an I/M program was specified, a waiver rate of 3% and
a compliance rate of 96% was assumed.  For the Purposes of this
analysis, "basic" I/M was considered to be an idle/2500 test and
"enhanced" I/M was assumed to be an IM240 test. (An enhanced I/M
program was only modeled with MOBILE5a in this analysis.) An
evaporative system functional check was included under the "enhanced"
I/M scenarios (pressure testing was assumed for 1971 and later model
years, purge testing was assumed for 1981 and later model years). 
Finally, all I/M programs were assumed to be conducted on an annual,
centralized basis.

8.2 MOBILE4.1 RESULTS

The analysis described above was performed on MOBILE4.1 output for a
no I/M case and a basic I/M case.  The results of that analysis are
presented below.

No I/M - Figures 79 and 80 illustrate the fleet-average LDGV emission
rate by model year group for exhaust HC and CO, respectively, for
calendar years 1985 to 2010 (in five-year increments).  As seen, the
overall fleet-average emission rate drops dramatically between 1985
and 2000 and then levels off beyond that point.  This is the result
of fleet turnover (i.e., higher-emitting, older model years
disappearing from the 25-year fleet considered in MOBILE4.1) and the
fact that emission rates beyond the 1981 model year are substantially
similar. (Recall that MOBILE4.1 does not include the effects of Tier
I vehicles.)



                                           -103-

Click HERE for graphic.



Click HERE for graphic.


                                           -104-

It is also interesting to compare the fractional contribution of each
model year group to the fleet-average emission rate versus that
group's VMT fraction.  This is depicted in Figures 81 and 82 for
exhaust HC and CO, respectively, for the 1995 calendar year.  Those
figures indicate that although pre-1981 LDGVs account for only 5.1%
of the VMT in 1995, they contribute 29.8% and 26.3% of the exhaust HC
and CO emissions, respectively. (The spreadsheets used to develop
these figures are contained in Appendix D, which also includes
results for LDGTls and LDGT2s.)

Basic I/M - Under the "basic" program modeled in this section, the
MOBILE4.1 results are very similar to those presented above for the
no I/M case, but the relative contribution of the pre-1981 model
years is slightly less.  This is the result of I/M being more
effective, on a percentage basis, for older vehicles than newer
vehicles.  The results for this I/M case are summarized for LDGVs,
LDGT1s, and LDGT2s in Appendix D.


Click HERE for graphic.


                                           -105-


Click HERE for graphic.


8.3 MOBILE5a RESULTS

The analysis presented above was also carried out using MOBILE5a as
the basis for the emission calculations, and the results from that
evaluation are summarized below.  Only results for the no I/M case
are presented in this section, as the I/M scenarios revealed similar
trends. (Results from the I/M cases are contained in Appendix D.)

The model-year contributions to the fleet-average LDGV exhaust HC and
CO emission rates calculated by MOBILE5a under a no I/M case are
illustrated in Figures 83 and 84, respectively.  As with the
MOBILE4.1 results, the overall fleet-average emission rates decline
substantially between 1985 and 2010.  However, with the MOBILE5a
estimates, the exhaust HC emission rates continue to decrease beyond
2000 because of the Tier I emission standards that are phased-in
beginning in 1994.  Figures 85 and 86 show the travel fraction and
the emission contribution of the model-year groups for exhaust HC and
CO, respectively, for a 1995 analysis year.  Again, the pre-1981
emission contribution is much greater than indicated by this group's
travel fraction.



                                           -106-

Click HERE for graphic.


Click HERE for graphic.

                                           -107-

Click HERE for graphic.


Click HERE for graphic.

                                           -108-


8.4 COMPARISON OF MOBILE4.1 AND MOBILE5a

Table 12 compares the emission contribution of pre-1981 and 1981+
modelyear groups for LDGVs for a 1995 analysis year computed with
MOBILE4.1 and MOBILE5a.  As seen, MOBILE5a is predicting a higher
contribution from the newer (1981+) vehicles.  This is primarily the
result of the changes to the base emission rate equations in MOBILE5a
that predict much higher deterioration rates beyond 50,000 miles.


Table 12.  Comparison of Model Year Contribution to the Fleet-
Average LDGV Emission Rate for Calendar Year 1995 
(MOBILE4.1 Versus MOBILE5a)



                           MOBILE4.1                        MOBILE5a
Pollutant/
I/M Scenarioa              Pre-1981     1981+        Pre-1981     1981+
HC
   No I/M                  0.297        0.703        0.214        0.786
   Basic I/M               0.262        0.738        0.192        0.808
   Enhanced I/M                                      0.224        0.776

CO
   No I/M                  0.285        0.715        0.183        0.817
   Basic I/M               0.244        0.756        0.151        0.849
   Enhanced I/M                                      0.176        0.824

NOx
  No I/M                   0.222        0.778        0.135        0.865
  Basic I/M                0.226        0.774        0.137        0.863
  Enhanced I/M             0.168        0.832


For this analysis, "basic" I/M refers to an idle/2500 RPM test;
enhanced" I/M refers to an IM240 test.



                                           -109-

9. EFFECT OF MODEL CHANGES ON
EMISSION FACTOR ESTIMATES


This section presents a comparison of the fleet-average emission
rates calculated by MOBILE4, MOBILE4.1, and MOBILE5a.  The baseline
emissions estimates are presented first, followed by an investigation
of changes made to individual parameters while holding the remaining
parameters constant.  Included in this comparability analysis are
changes in operating mode fractions; the effect of temperature on
emissions; the effect of speed on emissions; the impact of RVP
control, oxygenated fuel, and reformulated gasoline; and the impact
of inspection and maintenance programs.

As detailed in the preceding sections, many parameters go into the
fleet-average emissions estimates provided by the MOBILE model, and
many of these have changed among revisions to the model.  This is
further complicated by the fact that not all vehicle types included
in the emissions estimates are treated in the same manner.  For
example, gasoline vehicles are corrected for temperatures outside the
standard FTP temperature of 75F, whereas Diesel vehicles are not;
and multiple day diurnals are calculated for LDGV, LDGT1, LDGT2, and
HDGV, but not for motorcycles or Diesel vehicles. (A summary of the
various model parameters and the applicability to each vehicle class
is given in Table 13.)

When comparing the following fleet-average g/mi output from MOBILE4,
MOBILE4.1, and MOBILE5a, it must be recognized that each vehicle
class contributes to the overall emission rate, with LDGVs generally
having the most significant impact because of their large numbers in
the fleet.  Thus, changes that have occurred to LDGVs are most
noticeably reflected in the fleet-average emission rate.  On the
other hand, changes to the heavy-duty vehicle classes have a
relatively smaller impact on fleetaverage emissions.  In addition,
measures that impact all model years generally have a larger impact
on fleet-average emissions than a measure that is applied to new
vehicles and must wait for fleet turnover before it is fully
implemented.  Thus, measures such as the more stringent NOX standard
for HDGV and HDDV coming from the 1990 CAAA have a relatively minor
impact on the fleet-average emission rate, whereas measures such as
oxygenated and reformulated fuels have a significant effect on fleet-
average emissions.

9.1 BASELINE EMISSIONS

The fleet-average baseline emissions estimates for MOBILE4,
MOBILE4.1, and MOBILE5a are illustrated in Figures 88 through 90 for
HC, CO, and NOx, respectively.  No I/M or ATP programs were specified
for these

                                           -111-

Click HERE for graphic.


                                           -112-

Click HERE for graphic.


Click HERE for graphic.


                                           -113-


Click HERE for graphic.


estimates, and they were conducted at an average speed of 19.6 mph
and a temperature of 75F (diurnal emissions assumed a 60 to 84F
temperature rise), and the FTP operating mode fractions (i.e., "bag
splits," 20.6 percent cold start, 27.3 percent hot start, 52.1
percent stabilized) were utilized.  The large differences in the
baseline emissions estimates observed in these figures are the result
of modifications to the base emission rate equations made among the
model revisions, the use of an in-use fuel correction in MOBILE5, and
higher mileage accumulation rates assumed in MOBILE5.  This is
especially noticeable for HC and CO, where the MOBILE4 and MOBILE4.1
estimates are fairly close, but the MOBILE5a results are much higher. 
For NOx, the emission rates have increased by about the same amount
between the release of MOBILE4 and MOBILE4.1, and the release of
MOBILE4.1 and MOBILE5.

The HC emissions estimates depicted in Figure 87 include exhaust and
evaporative emissions.  Figure 90 shows a breakdown of HC emissions
for the 1995 and 2005 calendar years.  As seen, the increase in the
HC emission rate observed with MOBILE5a is primarily the result of an
increase in exhaust emissions (at least for the 75F case modeled
here), although an increase in running loss emissions also occurred
with the release of MOBILE5a.



                                           -114-

Click HERE for graphic.



9.2 OPERATING MODE FUNCTIONS

Differences in operating mode have significant impacts on the
emission rates calculated by the MOBILE models.  This difference is
limited, however, to light-duty vehicles, as no operating mode
correction is applied to the HDGV and HDDV vehicle classes. 
Differences in cold start, hot start, and stabilized operation are
shown in Figures 91 through 93 for HC, CO, and NOx emissions,
respectively.  This comparison is made for the year 2000, and only
exhaust HC emissions are presented since operating mode does not
impact non-exhaust HC components.

As seen in the figures, the relative difference in emission rates
between cold start, hot start, and stabilized operation remained
fairly constant among the MOBILE revisions.  As expected, cold start
operation results in the highest g/mi emissions for all pollutants,
followed by hot start and stabilized operation.  The higher emission
rates observed for MOBILE5a are again the result of higher estimates
of base emission rates.



                                           -115-


Click HERE for graphic.



Click HERE for graphic.

                                           -116-

Click HERE for graphic.



9.3 TEMPERATURE

The ambient temperature used in a MOBILE run also plays a critical
role in the fleet-average emission rate.  As temperature decreases
(below 75F), emissions increase.  The increase in emissions is
explained by increased cold start emissions as the engine and
emission control system take longer to warm up and additional fuel
enrichment is necessary for smooth combustion.  Additionally, as the
temperature rises above the standard 75F point, an increase in HC
and CO emissions is observed, but the increase is not as severe. 
This increase is primarily the result of increased vapors being
purged from the evaporative emission control system, leading to rich
operation.

The impact of temperature on the year 2000 fleet-average emission
rate is shown in Figures 94 through 96 for exhaust HC, CO, and NOx,
respectively.  The general shape of the curves is very similar, which
was expected since no substantive differences in the temperature
correction factors exist among the MOBILE4, MOBILE4.1, and MOBILE5a
versions of the model.  The relative differences among the emission
estimates are again the result of higher base emission rates
predicted for MOBILE5a.

The comparison of emission estimates for non-exhaust HC was performed
with a slightly different basis for each of the evaporative
components modeled.  Figure 97 displays the year 2000 diurnal
emission rate (in


                                           -117-

Click HERE for graphic.


Click HERE for graphic.

                                           -118-


Click HERE for graphic.


Click HERE for graphic.

                                           -119-


grams/day, not accounting for multiple-day diurnals) for a diurnal
temperature swing of 25F with a variety of different maximum
temperatures. (Note that only the LDGV class is shown in the figure.)
The general shape of the curve is similar for each of the MOBILE
revisions, and the values of the results are also very close.  As
expected, an increase in emission rate is observed for higher maximum
temperatures.  Hot-soak emissions for the year 2000 (in grams/trip)
are illustrated in Figure 98.  Widely varying results among the model
revisions are seen, particularly at the higher temperatures. 
Finally, a comparison of running loss emissions is given in Figure
99.  The results for MOBILE5a and MOBILE4.1 are very close, while the
MOBILE4 estimates are much lower at higher temperatures.  The
differences in the evaporative component emission estimates are
primarily the result of new data incorporated by EPA, however, the
MOBILE5a estimates also include the effect of the new evaporative
standards and test procedures.


Click HERE for graphic.


                                           -120-

Click HERE for graphic.


9.4 SPEED

Figures 100 through 103 illustrate the impact of speed on the year
2000 MOBILE4, MOBILE4.1, and MOBILE5a emission estimates for exhaust
HC, CO, and NOx, respectively.  Because the speed correction factors
have not changed dramatically among the MOBILE revisions, the shape
of the curves is very similar.  However, the MOBILE5a results are
higher in terms of absolute g/mi values.  This is the result of the
higher base emission rates assumed for the MOBILE5a model.  It is
interesting to note, however, that the NOx emissions depicted in
Figure 103 may have significant implications for conformity analyses. 
Because NOx emissions are projected to increase beyond 20 mph, any
transportation measure that increases speed by reducing congestion
increases the NOx inventory.  This result is counter to historical
thinking that assumed as speed increased, emissions decreased.

9.5 FUEL EFFECTS

There are three fuels-related inputs to the MOBILE models that impact
emission estimates: Reid vapor pressure (RVP), oxygenate content, and
the presence or absence of reformulated gasoline.  RVP, which is a
measure of the fuells volatility, most significantly impacts
evaporative emissions, but it also affects exhaust emission
estimates.  Oxygenated


                                           -121-

Click HERE for graphic.



Click HERE for graphic.



                                           -122-


Click HERE for graphic.


fuels are now required in many communities as a cold-weather CO
control strategy, and the capability to model the effects of
oxygenated fuels was first included with MOBILE4.1. Finally, an
option included in MOBILE5 is modeling of reformulated gasoline.  As
part of the Clean Air Act Amendments of 1990, reformulated gasoline
will be required in ozone nonattainment areas classified as "severe"
and "extreme" beginning in 1995.

RVP - As discussed above, the impact of RVP is most pronounced on
evaporative emission estimates.  This is evidenced in Figure 103,
which illustrates MOBILE5a nonexhaust hydrocarbon emission rates for
a series of RVPs ranging from 15.0 psi down to 7.0 psi. (Included in
these estimates are hot soak, diurnal, running loss, and resting loss
emissions.) Substantial emission reductions can be achieved by
controlling fuel RVP, and EPA has promulgated fuel volatility
regulations that limit summertime RVP to 9.0 and 7.8 psi, depending
on state and month, beginning in 1992.18 (This represents Phase II of
EPA's fuel volatility regulations; Phase I, which had less stringent
RVP requirements, was implemented in 1989.) As demonstrated in the
figure, nonexhaust HC emissions from 9.0 psi fuel are roughly 50%
lower than 11.0 psi fuel under typical summertime conditions. (Figure
103 is based on a 70 to 95F diurnal temperature rise.)

Although not as significant, RVP level also impacts exhaust emission
estimates calculated by MOBILE.  At higher RVP, HC and CO emissions


                                           -123-

Click HERE for graphic.

increase.  That is because higher volatility fuel leads to more
hydrocarbon vapors being stored in the evaporative control system
canister.  This, in turn, causes rich operation during canister purge
which elevates HC and CO emission rates.  Figures 104 and 105
illustrate this effect for HC and CO, respectively.

Figure 104 shows MOBILE5a exhaust HC results for RVPs ranging from
12.5 to 7.0 psi. (The maximum RVP chosen for this figure was 12.5 psi
because MOBILE does not correct for RVP above that point.*) This
figure indicates a 10% to 20% reduction in exhaust HC emission rates
as a result of changing RVP from 11.0 to 9.0 psi.  Also of note in
Figure 104 is the very small change in emission rate when RVP is
lowered to 7.0 from 9.0. This is because MOBILE does not allow the
exhaust RVP correction factor to fall below 1.0, which would occur
under these ambient conditions at just below 9.0 psi. (This is shown
in Figures 40 and 41.)


_____________________________
* In actuality, MOBILE considers the maximum RVP for purposes of
exhaust corrections to be 11.7 psi.  However, MOBILE has an algorithm
that decreases the user-input RVP to account for RVP loss as fuel
sits in vehicle fuel tanks (this phenomenon is known as 'fuel tank
weathering').  Under the ambient conditions specified in Figure 104
(i.e., 70 to 95F), the 12.5 psi fuel has been "weathered" to 11.7
psi; thus, 12.5 represents the maximum RVP for a 70 to 95F case.


                                           -124-


Click HERE for graphic.


Click HERE for graphic.
 

                                           -125-

The effect of RVP on CO emission estimates from MOBILE5a is
illustrated in Figure 105.  Because CO is typically a wintertime
problem, these runs were performed for a diurnal temperature rise of
45 to 65F. (Because MOBILE does not calculate an RVP correction at
temperatures below 45F, this was chosen as the lower temperature for
developing Figure 105.  Also note that at temperatures below 75F, no
RVP correction is calculated for RVPs below 9.0 psi.) The results
shown in Figure 105 indicate fairly significant CO reductions (e.g.,
30% for the year 2000, although the reduction would be slightly
diminished if an I/M program was specified) from lowering RVP from
11.0 to 9.0 psi.  Although it is clearly not an option for many CO
nonattainment areas, some of the warm weather CO nonattainment areas
(e.g., Phoenix, Arizona) are investigating wintertime RVP limits as
a CO control strategy.

Oxygenated Fuels - As discussed in Section 8, the capability to model
oxygenated fuels for cold-temperature CO control was included in
MOBILE4.1 and MOBILE5a.  The impact of oxygenated fuels on cold-
temperature CO emissions is shown in Figure 106 for MOBILE5a run at
an ambient temperature of 30F. (Because modeling of oxygenated fuels
is very similar between MOBILE4.1 and MOBILE5a, only the MOBILE5a
results are shown in the figure.) A considerable decrease in CO
emissions is observed when comparing the oxygenated fuels case to the
baseline case.  As expected, the effects are roughly linear as a
function of oxygen content (i.e., higher oxygen content results in a
higher benefit).

Click HERE for graphic.

                                           -126-
Reformulated Gasoline - The capability to model the effects of
reformulated gasoline was included with the MOBILE5 version, and the
effect of reformulated gasoline on HC emissions is shown in Figure
108.  In this analysis, the MOBILE5a model was run at a relatively
high temperature (i.e., a diurnal temperature swing of 70 to 95F,
with 90F serving as the exhaust analysis temperature), which roughly
corresponds to the ambient conditions required for the baseline
modeling analysis contained in the notice of proposed rulemaking on
reformulated gasoline.19 The relative difference between the 1995
estimate and the 2000 and subsequent calendar years reflects the
difference between Phase 1 and Phase 2 of the regulations.

Click HERE for graphic.

9.6 INSPECTION AND MAINTENANCE

Inspection and maintenance programs remain an attractive motor
vehicle pollution control option for many local communities.  Because
all model years can fall under inspection requirements, the impact on
fleet emissions is much faster compared to new vehicle regulations
that take many years to become fully effective.


                                           -127-

Exhaust Emissions - Figures 108 through 110 illustrate the baseline
and after-I/M LDGV exhaust HC emission factors for MOBILE4,
MOBILE4.1, and MOBILE5a, respectively.  Because MOBILE4 did not model
a transient I/M program (i.e., EPA's IM240) and the MOBILE4.1 IM240
credits were based on little IM240 data, only a single I/M program is
shown in Figures 108 and 109.  The program modeled for these figures
is an annual, centralized, idle program.  This program was also
modeled by MOBILE5a and is shown in Figure 110; however, Figure 110
also contains an estimate for an annual, centralized, transient
chassis dynamometer based program (IM240 based on 0.8/15.0/2.0
HC/CO/NOx cutpoints).  Significant reductions are predicted for I/M
programs, particularly the IM240 program shown in Figure 110. 
Applying the IM240 program to the MOBILE5a emission factors results
in estimates similar to the MOBILE4 and MOBILE4.1 non-I/M cases.

The I/M results for CO are shown in Figures 111 to 113 for MOBILE4,
MOBILE4.1, and MOBILE5a, respectively.  As with HC, significant
reductions are also predicted as a result of I/M.  In fact, MOBILE5a
estimates that the LDGV CO emission rate beyond 1995 will be roughly
cut in half as a result of the IM240 program modeled in this
analysis.

Finally, the NOx benefits ascribed to IM240 are shown in Figure 114. 
Although the NOx benefits are not as large as HC and CO, this still
represents a significant reduction in in-use emissions. (Because
idlebased programs result in only minor changes to NOx emissions, the
impact of I/M on NOx emissions for MOBILE4 and MOBILE4.1 was not
shown.)

Click HERE for graphic.

                                           -128-

Click HERE for graphic.


Click HERE for graphic.


                                           -129-

Click HERE for graphic.




Click HERE for graphic.


                                           -130-

Click HERE for graphic.




Click HERE for graphic.


                                           -131-

Evaporative Emissions - In conjunction with the transient, loaded-
mode exhaust emissions test, EPA is strongly encouraging states to
adopt functional testing of evaporative emission control systems. 
This check consists of a pressure test in which the fuel tank is
pressurized with nitrogen to a preset limit, and a significant fall
in pressure over time indicates a defective system.  A purge test has
also been developed which consists of placing a flow meter in the
purge line.  If the flow registered is not sufficient to properly
purge the charcoal canister, the vehicle fails the test.

MOBILE5a has been structured to estimate the benefits of a
pressure/purge test, and the LDGV baseline and after-test results are
given in Figures 115 and 116 for evaporative emissions (i.e., hot
soak, diurnal, and crankcase emissions) and running loss emissions. 
The test modeled was an annual, centralized program including 1981
and later model year vehicles.  Significant emission reductions are
estimated as a result of this program, with reductions on the order
of 50 percent being ascribed for the later calendar years.

Click HERE for graphic.



                                           -132-

Click HERE for graphic.



                                           -133-


                                        APPENDIX A


Emission Control Technologies and Comparison of Base Emission Rates
at 50,000 and 100,000 Miles for the LDGTI and LDGT2 Vehicle Classes


This appendix presents a summary of emission standards and emission
control technologies used to most those standards for the LDGT1 and
LDGT2 vehicle classes.  It also contains a series of illustrations
that show a comparison of LDGT1 and LDGT2 base emission rates at
50,000 and 100,000 miles computed by the MOBILE4, MOBILE4.1, and
MOBILE5a models.




                                           -135-



                                           LDGT1


                                           -137-

Click HERE for graphic.


                                           -138-


Click HERE for graphic.


Click HERE for graphic.


                                           -139-

Click HERE for graphic.


Click HERE for graphic.


                                           -140-


Click HERE for graphic.


Click HERE for graphic.


                                           -141-

                                           LDGT2


                                           -143-

Click HERE for graphic.


                                           -144-


Click HERE for graphic.


Click HERE for graphic.


                                           -145-


Click HERE for graphic.


Click HERE for graphic.


                                           -146-


Click HERE for graphic.


Click HERE for graphic.


                                           -147-



APPENDIX B


Speed-Time Profiles of Cycles Used
in Speed Correction Factor Development




                                           -149-


Click HERE for graphic.


Click HERE for graphic.


                                           -150-

Click HERE for graphic.


Click HERE for graphic.


                                           -151-

Click HERE for graphic.


Click HERE for graphic.


                                           -152-

Click HERE for graphic.


Click HERE for graphic.


                                           -153-

APPENDIX C

Comparison of Running Exhaust and Idle Emission Rates
Computed with MOBILE4.1 and MOBILE5a



This appendix presents a series of illustrations that show a
comparison of running exhaust emission rates and idle emission rates
computed by the MOBILE4.1 and MOBILE5a models. (Note that the
MOBILE5a model does not calculate idle emission rates directly. 
EPA's recommended procedure to convert g/mi running exhaust emission
rates to g/hr idle rates, which is described in the text, was used to
develop the MOBILE5a idle rates for these figures.)



                                           -155-




                                      MOBILE4.1; LDGV


                                           -157-



Click HERE for graphic.


Click HERE for graphic.


                                           -158-

Click HERE for graphic.


                                           -159-



                                     MOBILE4.1; LDGT1



                                           -161-


Click HERE for graphic.


Click HERE for graphic.


                                           -162-

Click HERE for graphic.


                                           -163-


                                     MOBILE4.1; LDGT2



                                           -165-

Click HERE for graphic.


Click HERE for graphic.


                                           -166-

Click HERE for graphic.


                                           -167-

                                      MOBILE5a; LDGV



                                           -169-



Click HERE for graphic.


Click HERE for graphic.


                                           -170-

Click HERE for graphic.


                                           -171-
                                     MOBILE5a; LDGT1



                                           -173-


Click HERE for graphic.


Click HERE for graphic.


                                           -174-

Click HERE for graphic.


                                           -175-

                                      MOBILE5a; LDGT2



                                           -177-


Click HERE for graphic.


Click HERE for graphic.


                                           -178-

Click HERE for graphic.


                                           -179-


                                        APPENDIX D


                  Contribution of Model Year Groups to the fleet-Average
                    Emission Rates Computed with MOBILE4.1 and MOBILE5a


This appendix presents a series of spreadsheets that summarize the
contribution (in grams per mile) of model-year groups to the HC, CO,
and NOx fleet-average emission rates.  The calculations were carried
out with the MOBILE4.1 and MOBILE5a models under a series of I/M
cases and include the LDGV, LDGT1, and LDGT21 vehicle classes.  For
this analysis, "basic" I/M refers to an idle/2500 rpm test, while
"enhanced" I/M refers to the IM240 procedure and includes
pressure/purge testing. (Recall, however that MOBILE assumes an idle
test for pre-1981 vehicles, regardless of the I/M test applied.to the
1981 and later model years.)

The spreadsheets are organized in the following manner:

1.     MOBILE4.1; No I/M
2.     MOBILE4.1; "Basic" I/M
3.     MOBILE5a; No I/M
4.     MOIBLE5a; "Basic" I/M
5.     MOBILE5a; "Enhanced" I/M



                                           -181-

                                   1. MOBILE4.1; No I/M


                                          - 183 -

Click HERE for graphic.



                                           -184-

Click HERE for graphic.



                                           -185-

                                 2. MOBILE4.1; "Basic" I/M



                                           -187-

Click HERE for graphic.



                                           -188-

Click HERE for graphic.



                                           -189-


                                     MOBILE5a; No I/M


                                           -191-
Click HERE for graphic.



                                           -192-

Click HERE for graphic.



                                           -193-


                                 2. MOBILE4.1; "Basic" I/M





                                          - 195 -

Click HERE for graphic.



                                           -196-

Click HERE for graphic.



                                           -197-

                                5. MOBILE5a; "Enhanced" I/M


                                           -199-


Click HERE for graphic.



                                           -200-


Click HERE for graphic.



                                           -201-


                                        REFERENCES


1.     U.S. Environmental Protection Agency.  User's Guide to MOBILE5. 
       Chapter 2. December 1992.

2.     Sierra Research, Inc.  Estimating the Effect of Driving Pattern
       on Exhaust Emissions Using a Vehicle Simulation Model (draft). 
       Prepared for the U.S. Environmental Protection Agency, Mobile
       Source Certification Division.  October 22, 1990.

3.     U.S. Environmental Protection Agency.  User's Guide to MOBILE4.1
       (Mobile Source Emission Factor Model).  Office of Mobile
       Sources.  July 1991.

4.     Platte, Lois.  Personal Communication.  U.S. Environmental
       Protection Agency.  March 1993.

5.     Huls, T.A. Evolution of Federal Light-Duty Mass Emission
       Regulations.  U.S. Environmental Protection Agency.  SAE Paper
       No. 730554.  May 1973.

6.     Brzezinski, D.J. Tech IV Credit Model: Estimates for Emission
       Factors and Inspection and Maintenance Credits for 1981 and
       Later Vehicles for MOBILE3.  U.S. Environmental Protection
       Agency.  October 1985.

7.     U.S. Environmental Protection Agency.  Estimating Idle Emission
       Factors Using MOBILE5.  MOBILE5 Information Sheet #2.  July 30,
       1993.

8.     Systems Control, Inc. [Formerly Olson Laboratories].  Heavy-Duty
       Vehicle Cycle Development.  For the U.S. Environmental
       Protection Agency.  July 1978.

9.     Wysor, T. and C. France.  Selection of Transient Cycles for
       Heavy-Duty Vehicles.  U.S. Environmental Protection Agency. 
       June 1978.

10.    France, C. Transient Cycle Arrangement for Heayy-Duty Engine and
       Chassis Emission Testing.  U.S. Environmental Protection Agency. 
       August 1978.

11.    Smith. M. Heavy-Duty Vehicle Emission Conversion Factors 1962 -
       1977.  U.S. Environmental Protection Agency.  August 1984.

12.    Machiele, P. Heavy-Duty Vehicle Emission Conversion Factors II
       1962 - 2000.  U.S. Environmental Protection Agency.  October
       1988.

13.    California Air Resources Board.  Public Hearing to Consider Fuel
       Evaporative Emission Regulations for Light-Duty Vehicles.  Staff
       Report No. 75-7-6.  April 16, 1975.




                                           -203-

14.    U.S. Environmental Protection Agency.  Control of Air Pollution
       from New Motor Vehicles and New Motor Vehicle Engines:
       Evaporative Emission Regulations: Final Rule.  Federal Register,
       Vol. 58, No. 55, March 24, 1993.

15.    U.S. Environmental Protection Agency.  Inspection/Maintenance
       Program Requirements: Final Rule.  Federal Register.  Vol. 57,
       no. 215, Thursday, November 5, 1992.

16.    U.S. Environmental Protection Agency.  MOBILE5 User's Guide. 
       Chapter 2, Draft 4a, December 3, 1992.

17.    U.S. Environmental Protection Agency.  Checklist for Completing
       the Inspection/Maintenance SIP.  March 1993.

18.    U.S. Environmental Protection Agency.  Volatility Regulations
       for Gasoline and Alcohol Blends Sold in Calendar Years 1992 and
       Beyond: Final Rule.  Federal Register.  Vol. 55, No. 112.  June
       11, 1990.

19.    U.S. Environmental Protection Agency.  Regulations of Fuels and
       Fuel Additives: Standards for Reformulated Gasoline: Proposed
       Rule.  Federal Register.  February 26, 1993.



                    *U.S. GOVERNMENT PRINTING OFFICE:1994-600-572/00050


                                           -204-

(mob.html)
Jump To Top