EPA-460/3-76-010a
October 1976
                             SHORT TEST
            CORRELATION ANALYSES
                                   ON 300,
              1975 MODEL YEAR CARS
                                VOLUME I
       U.S. ENVIRONMENTAL PROTECTION AGENCY
           Office of Air and Waste Management
        Offiee of Mobile Source Air Pollution Control
           Emission Control Technology Division
              Ann Arbor, Michigan 48105

-------
                                      EPA-460/3-76-010a
               SHORT TEST
      CORRELATION  ANALYSES
ON  300,  1975 MODEL YEAR CARS
                  VOLUME I
                   Mobile Systems (iron))
                   Aerospace (Corporation
                     P.O. Box 929~>7
                 Los An Bolt's. California 90009
                   Contract No. 68-01-0417
              EPA Project Officer: F. Peter Hutchins
                      Prepared for

            ENVIRONMENTAL PROTECTION AGENCY
               Office of Air and Waste Management
            Office of Mobile Source Air Pollution Control
              Emission Control Technology Division
                 Ann Arbor, Michigan 48105

                      October 1976

-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers.   Copies are
available free of charge to Federal employees, current contractors ajid
grantees, and nonprofit organizations - as supplies permit - from the
Air Pollution Technical Information Center, Environmental Protection
Agency, Research Triangle Park, North Carolina 27711;  or,  for a fee,
from the National Technical Information Service, 5285 Port Royal Road,
Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by
Aerospace Corporation, Los Angeles, California 90009, in fulfillment
of Contract No. 68-01-0417.  The contents of this report are reproduced
herein as received from Aerospace Corporation.  The opinions, findings,
and conclusions expressed are those of the author  and not necessarily
those of the Environmental Protection Agency. Mention of company or
product names is not to be considered as an endorsement by the Environmen-
tal Protection Agency.
                  Publication No.  EPA-460/3-76-010a

-------
                              FOREWORD
      This report,  prepared by The Aerospace Corporation for the U.S.
Environmental Protection Agency, Emission Control Technology Division,
presents the results of a statistical analysis of the degree of correlation
between two short tests and the 1975 Federal Test Procedure. The correla-
tion analyses were based on experimental test data from 300 1975-model-year
vehicles,  composed of three groups of 100 vehicles each that were tested in
the cities of Chicago,  Houston, and Phoenix.
      The results of the study  are presented in seven sections.  A brief, con-
cise review of important findings and conclusions is presented in the Highlights
section.  Section 1  contains a summary and discussion of the study results.
The background, scope, objectives,  and method of approach are given in Sec-
tion 2.  The short tests  and test fleet composition are described and discussed
in Section 3. Section 4 describes the statistical screening procedures, the
primary statistical tools used  in the  correlation analyses,  and the results of
the conventional correlation analysis.  Similar results of a contingency table
analysis for the 300-vehicle fleet are presented in Section 5.  Section 6 pre-
sents an initial analysis of the  emission reductions  afforded by implementation
of a mandatory inspection and  maintenance program using  short test cut-points
based on (1) a selected error of commission rate for each of three engine
displacement groups and (2)  an average failure rate for the fleet such as is used
in existing I/M programs.
      Because of the numerical analysis nature of the study and the many
variables addressed, the results of the study are contained in many tables
and figures. Therefore, the principal discussion of results is contained in
the summary and discussion section; numerous selected tabular information
is presented in the summary to facilitate understanding of the stated results.
      The voluminous  tabular and graphical data are presented in an Appendix
of tables and figures .  Only Section 1 (the Summary and Discussion) has
tabular and graphical data integrated with the text.
                                   -111-

-------
                          ACKNOWLEDGMENT
      During the course of this study, Mr. F. P. Hutchins of the
Environmental Protection Agency's Emission Control'Technology Division,
who served as EPA Project Officer for the study, provided valuable guidance
and assistance. His efforts are gratefully acknowledged.
      Dr. John Thacker was principally responsible for the statistical analy-
sis effort reported herein.  The following technical personnel of The Aerospace
Corporation also made valuable contributions to the analyses performed under
this contract: B.  D.  Gregoire, A. M. Tirnmer, D. K. Sakaguchi,  and
C. A. Vejar.
                                    M. G. Hinton, Group Director
                                    Mobile Systems
Approved by
TTlura, General Manager
Environment and Energy
  Conservation Division
                                    -v-

-------
                               CONTENTS


FOREWORD  ............................... .......    iii
ACKNOWLEDGMENT ...................... ..........     v
HIGHLIGHTS  .....................................  H- 1
1.    SUMMARY AND DISCUSSION ..... ..................   1-1
      1. 1   Summary Overview
           1.1.1    Conventional Correlation Analyses
                    1.1.1.1  Short Test (ST) Comparison
                    1.1.1.2  Effect of Accumulated
                             Mileage  ....................   1-2
                    1.1.1.3  Effect of Engine Displacement .....   1-2
                    1.1.1.4  Effect of Inertia Test Weight
                             Group .....................   1-3
                    1.1.1.5  Effect of Emission Control
                             System Type .................   1-4
                    1.1.1.6  Effect of Carburetion vs
                             Fuel Injection  ................   1-4
                    1.1.1.7  Effect of Transmission Type  ......   I-5
                    1.1.1.8  Effect of Manufacturer ..........   1-5
           1. 1.2    Contingency Table Analysis ..............   1-6
                    1.1.2.1  Short Test Comparison ..........   1-1°
                    1.1.2.2  Effect of Engine
                             Displacement  ................   1-14
                    1.1.2.3  Effect of Inertia Test Weight
                             Group ................ .....   1-14
                    1.1.2.4  Effect of Emission Control
                             System .Type .................   1-15
                    1.1.2.5  Effect of Carburetion vs
                             Fuel Injection  ................   1-15
                    1.1.2.6  Effect of Transmission Type  ......   1-15
                    1. 1.2.7  Effect of Manufacturer ..........   I-16
                                  -Vll-

-------
                   CONTENTS (Continued)
              1.1.2.8   Effect of Cut-Point
                        Determination Method   	  1-16
      1.1.3    Simulation of 207(b) Implementation	  *-17
              1.1.3.1   Simulation Results	  1-21
1.2   Conventional Correlation Analysis Discussion	  1-29
      1.2. 1    Results for All Three Cities Combined	  1-29
              1.2. 1. 1   Short Test Comparison	  1-29
              1.2.1.2   Effect of Accumulated
                        Mileage	  1-29
              1.2.1.3   Effect of Engine Displacement	  1-32
              1.2. 1.4   Effect of Inertia Test Weight
                        Group	  1_32
              1.2.1.5   Effect of Emission Control
                        System Type	  1_35
              1.2. 1.6   Effect of Carburetion vs
                        Fuel Injection	  1-35
              1.2.1.7   Effect of Transmission Type	  1-38
              1.2.1.8   Effect of Manufacturer	  l_38
              1.2.1.9   Effect of Engine Displacement
                        and Manufacturer  	  1-41
      1.2.2    Results for Chicago	  .  1-45
              1.2.2. 1   Short Test Comparison	  1-45
              1.2.2.2   Effect of Engine Displacement  .....  1-48
              1.2.2.3   Effect of Inertia Test Weight
                        Group  .  .	  1-48
              1.2.2.4   Effect of Emission ControT
                        System Type	  1-51
              1.2.2.5   Effect of Carburetion vs
                        Fuel Injection	n.  1-51
              1.2.2.6   Effect of Transmission Type	  1-51
              1.2.2.7   Effect of Manufacturer	  1-55
                            -Vlll-

-------
                   CONTENTS (Continued)
      1.2.3    Results for Houston	   1-55
              1.2.3. 1  Short Test Comparison	  .   1-55
              1.2.3.2  Effect of Engine Displacement	   1-58
              1.2.3.3  Effect of Inertia Test Weight
                       Group	   1-58
              1.2.3.4  Effect of Emission Control
                       System Type	   1-58
              1.2.3.5  Effect of Carburetion vs
                       Fuel Injection	   1-62
              1.2.3.6  Effect of Transmission Type	   1-62
              1.2.3.7  Effect of Manufacturer	   1-62
      1.2.4    Results for Phoenix	   1-62
              1.2 4. 1  Short Test Comparison	   1-67
              1.2.4.2  Effect of Engine Displacement	   1-67
              1.2.4.3  Effect of Inertia Test Weight
                       Group	   1-69
              1.2.4.4  Effect of Emission Control
                       System Type	   1-69
              1.2.4.5  Effect of Carburetion vs
                       Fuel Injection	   1-72
              1.2.4.6  Effect of Transmission Type	   1-72
              1.2.4.7  Effect of Manufacturer	   1-72
1.3   Contingency Table Analysis Discussion	   1-77
      1.3.1    Hydrocarbon Emission	   1-77
      1.3.2    Carbon Monoxide Emission	   1-80
      1.3.3    Oxides of Nitrogen Emission	   1-80
      1.3.4    Multiple-Constituent Tests	   1-85
      1.3.5    Effect of Engine  Displacement	   1-88
      1.3.6    Effert of Inertia  Test Weight Group	   1-95
      1.3.7    Effect of Emission Control System
              Type	   1-103
                             -IX-

-------
                         CONTENTS (Continued)
            1.3.8    Effect of Carburetion vs Fuel Injection	  1-110
            1.3.9    Effect of Transmission Type  . .  .,	  1-116
            1.3. 10   Effect of Manufacturer	  1-122
            1.3. 11   Effect of Manufacturer and Engine
                    Displacement Group	  1-127
                    1.3. 11. 1   150 CID and Less Displacement
                               Group	  1-129
                    1.3.11.2   151 to 259 CID Group	  1-136
                    1.3. 11.3   260 CID and Greater Group	  1-143
            1.3. 12   Comparison of Short Test Cut-Points	  1-154
                    1.3. 12. 1   By City	  1-154
                    1.3. 12.2   By Engine Displacement  Group	  1-154
            1.3. 13   Effect of Cut-Point Selection Method
                    on STE Values	  1-159
2.    INTRODUCTION	2-1
      2. 1   Background and Objectives  	2-1
      2. 2   Study Scope	2-3
      2. 3   Method of Approach	2-3
3.    TEST CHARACTERISTICS AND FLEET
      COMPOSITION	3-1
      3. 1   Short Tests	."	3-1
           3. 1. 1    General	3-1
           3.1.2    Short Test Definition	3-2
                    3. 1. 2. 1     Federal Three Mode	3-2
                    3. 1.2.2     Federal Short Cycle	3-3
    •  3.2   Test Fleet Composition	'	3-3
           3. 2. 1    Types of  Cars	-  3-3
           3.2.2    Prior Use	3-5
                                   -x-

-------
                         CONTENTS (Continued)
4.     STATISTICAL SCREENING AND CORRELATION
      ANALYSES	4-1
      4. 1   Discussion of Methodology	4-1
           4. 1. 1    Conventional Correlation Analysis	4-1
           4. 1.2    Statistical Tests of Significance	4-2
      4.2   Results for All Three Cities Combined	4-3
      4.3   Results for Individual City Fleets	4-4
      4. 4   Discussion of Results	4-6
5.     CONTINGENCY TABLE ANALYSES	5-1
      5. 1   Discussion of Methodology	5-1
      5. 2   Variance Estimates	 5-4
           5.2. 1    Short Test Cut-Points	5-4
           5.2.2    Three-Constituent Test Percentages	5-5
      5.3   Specific Analyses and Their Results	5-6
           5.3.1    Analyses with Short Test Cut-Points
                    Based Upon 300 Cars	5-6
                    5.3. 1. 1    Hydrocarbon Emissions	5-6
                    5.3. 1.2    Carbon Monoxide Emission	5-7
                    5.3.1.3    Oxides of Nitrogen Emission .:.... 5-7
                    5.3. 1.4    Multiple Constituent Tests	5-7
                    5.3. 1. 5    Effects of Vehicle
                               Characteristics	5-8
                    5.3.1.6    Comparison of Short Test
                               Cut-Points	5-8
           5.3.2    Analyses with Short Test Cut-Points
                    Based Upon Engine Displacement	5-8
                    5.3.2. 1    150 CID and Less Displace-
                               ment Group	5-9
                    5.3.2. 2    151 to  259 CID Group	5-9
                    5.3.2.3    269 and Greater CID Group	5-10
                    5.3.2.4    Comparison of Short Test
                               Cut-Points . . .  '	5-10
                                   -XI-

-------
                          CONTENTS (Continued)


6.     207(b) IMPLEMENTATION SIMULATION ANALYSES	6-1
      6. 1   Objectives and Approach	6-1
      6. 2   Program Components	  6-2
           6. 2. 1    Initializations	  6-2
           6. 2. 2    Generation of Fleets	6-3
           6.2.3    Emission Inspection Medel	  6-3
                    6.2.3.1    General	  6-3
                    6.2.3.2    Measurement Component	6-3
                    6.2.3.3    Inspection Component	  6-4
           6.2.4    Vehicle Maintenance Model	  6-4
                    6. 2. 4. 1    General	6-4
                    6.2.4.2    Version  1: HC and CO Coupled
                               with a Lower  Bound	6-5
                    6.2.4.3    Version  2: HC,  CO and NOX
                               Independently Adjusted to the
                               FTP Standard	6-6
                    6.2.4.4    Versions: HC,  CO, and NOX
                               Independently Adjusted to the
                               Short Test Standard	6-6
                    6.2. 4. 5    Overview of Maintenance
                               Models	6-6
           6.2. 5    Mileage Accumulation and Deterioration
                    Model	6-7
                    6. 2. 5. 1    General	  6-7
                    6. 2. 5. 2    Deterioration Based Upon
                               Certification Data	6-7
                    6. 2. 5. 3    Deterioration Based Upon
                               Emission Factors Program
                               Data	   6-7
                    6.2.5.4    Deterioration Based Upon
                               Near-Term Effects of
                               Maintenance  ,	6-8
                    6.2.5. 5    Specific Models  Employed	6-8
                                   -Xll-

-------
                          CONTENTS (Continued)
            6.2.6   207(b) Effectiveness Calculations	   6-9
            6.2.7   Overview of Simulation Models	   6-9
      6.3   Simulation Scenarios and Results	   6-10
            6.3.1   General	   6-10
            6.3.2   Specific Scenarios and Their Results	   6-11
                    6.3.2. 1    Introduction	   6-11
                    6.3.2.2    . Simulation of 207(b)
                               Programs	   6-11
                    6.3.2.3    Simulation of I/M
                               Programs	   6-12
      6.4   Discussion and Conclusions	   6-12
GLOSSARY	   GL-1
REFERENCES	  R-l
APPENDIX - TABLES AND FIGURES
              (see detailed Contents list in Appendix)
                                   -xi 11-

-------
                                TABLES



1-1.   Comparison of ST Correlation Coefficients	   1-1

1-2.   Contingency Table	,	   1-7

1-3.   Comparison of ST Contingency Table Values for
       STE and STRR; All Three Cities  (Ec Set at
       5 Percent for  Pooled Fleet of 300 Vehicles)	   1-11

1-4.   Comparison of STE Values for Three Cities	   1-13

1-5.   Comparison of Cut-Point Determination Methods;
       5 Percent Ec Fleet Average vs 5 Percent Ec
       for Each of Three CID Groups	   1-18

1-6.   Average Deterioration Rates Used in 207(b)
       Effectiveness  Simulation	   1-20

1-7.   Annual Mileage Accumulations Used in 207(b)
       Effectiveness  Simulation	   1-23

1-8.   Variation of 207(b) Program Effectiveness with
       Various Maintenance Assumptions	   1-24

1-9.   207(b) Program Efficiencies with Maintenance
       Version 3	   1-25

1-10.  Variation of I/M Program Effectiveness with
       Various Maintenance Assumptions	   1-27

1-11.  Comparison of Effectiveness of 207(b) and I/M
       Program Approaches;  Maintenance Version  3;
       12-Month Maintenance Period Effectiveness	   1-28

1-12.  Correlation Coefficient Summary— All Three
       Cities	•.	   1-30

1-13.  Correlation Coefficient Summary by Accumulated
       Mileage — All  Three Cities	   1-31

1-14.  Correlation Coefficient Summary by Engine
       Displacement— All Three Cities	   1-33
                                   -xv-

-------
                           TABLES (Continued)
1-15.  Correlation Coefficient Summary by Inertia
       Test Weight Group— All Three Cities	•.	   1_34

1-16.  Correlation Coefficient Summary by Emission
       Control System Type— All Three Cities  	   1_36

1-17.  Correlation Coefficient Summary by Fuel
       System Type — All Three Cities	   1-37

1-18.  Correlation Coefficient Summary by
       Transmission Type — All Three Cities  	   1-39

1-19.  Correlation Coefficient Summary by Manu-
       facturer — All Three Cities	   1-40

1-20.  Correlation Coefficient Summary for 150 CID
       and Less by Manufacturer— All Three Cities	   1-42

1-21.  Correlation Coefficient Summary for 151 to
       259 CID by Manufacturer — All Three Cities	   1-44

1-22.  Correlation Coefficient Summary for 260 CID
       and Greater by Manufacturer— All Three Cities	   1-46

1-23.  Correlation Coefficient Summary— Chicago	   1-47

1-24.  Correlation Coefficient Summary by Engine
       Displacement— Chicago	   1-49

1-25.  Correlation Coefficient Summary by Inertia
       Test Weight Group— Chicago	   1_50

1-26.  Correlation Coefficient Summary by Emission
       Control System Type — Chicago	   1-52

1-27.  Correlation Coefficient Summary by Fuel
       System Type — Chicago	   1-53

1-28.  Correlation Coefficient Summary by Transmission
       Type — Chicago	   1-54

1-29.  Correlation Coefficient Summary by
       Manufacturer— Chicago	   1-56
                                  -xvi-

-------
                           TABLES (Continued)
1-30.  Correlation Coefficient Summary — Houston  ..,..•	   1-57

1-31.  Correlation Coefficient Summary by Engine
       Displacement— Houston	   1-59

1-32.  Correlation Coefficient Summary by Inertia
       Test Weight Group— Houston	   1-60

1-33.  Correlation Coefficient Summary by Emission
       Control System Type— Houston	   1-61

1-34.  Correlation Coefficient Summary by Fuel
       System Type — Houston . .	   1-63

1-35.  Correlation Coefficient Summary by
       Transmission Type — Houston	   1-64

1-36.  Correlation Coefficient Summary by
       Manufacturer — Houston .	   1-65
                          \
1-37.  Correlation Coefficient Summary— Phoenix	   1-66

1-38.  Correlation Coefficient Summary by Engine
       Displacement— Phoenix .	   1-68

1-39.  Correlation Coefficient Summary by Inertia
       Test Weight Group— Phoenix	   1-70

1-40.  Correlation Coefficient Summary by Emission
       Control System Type — Phoenix	   1-71

1-41.  Correlation Coefficient Summary by Fuel
       System Type — Phoenix	   1-73

1-42.  Correlation Coefficient Summary by
       Transmission Type — Phoenix	   1-74

1-43.  Correlation Coefficient Summary by
       Manufacturer — Phoenix .	   1-75

1-44.  Comparison of Short Test HC Results for Pooled
       Fleet and Individual Cities; Predicted Population;
       Bounded Errors of Commission Method (E   set
       at 5 percent)	c	   1-79
                                  -xvii-

-------
                           TABLES (Continued)
1-45.  Comparison of ST CO Results for Pooled Fleet and
       Individual Cities; Predicted Population; Bounded
       Errors of Commission  Method (E  set at
       5 percent)  	. . .  .C	   1-82

1-46.  Comparison of ST NOX  Results for Pooled Fleet
       and Individual Cities; Predicted Population; Bounded
       Errors of Commission  Method (E  set at 5 percent) .	   1-84

1-47.  Comparison of Short Test Multiple Constituent
       Results for Pooled Fleet  and Individual Cities;
       Actual Population; Bounded Errors of
       Commission Method (E  set at 5 percent)  	   1-86

1-48.  Comparison of Short Test HC Results by Engine
       Displacement;  All Three  Cities;  Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)	   1-89

1-49.  Comparison of Short Test CO Results by Engine
       Displacement;  All Three  Cities;  Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)	   1-91

1-50.  Comparison of Short Test NOX Results by Engine
       Displacement; All Three  Cities;  Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)	   1-93

1-51.  Comparison of Short Test Multiple Constituent
       Results by Engine Displacement; All Three
       Cities; Predicted Population;  Bounded Errors of
       Commission Method (E  set at 5 percent)  	   1-94

1-52.  Comparison of Short Test HC Results by Inertia
       Weight; All Three Cities; Bounded Errors of
       Commission Method; Predicted Population
       (E  set at 5 percent)  	   1-97

1-53.  Comparison of Short Test CO Results by Inertia
       Weight; All Three Cities; Bounded Errors of
       Commission Method: Predicted  Population
       (E  set at 5 percent)	 .   1-99
                                 -XVlll-

-------
                           TABLES (Continued)
1-54.  Comparison of Short Test NOX Results by Inertia
       Weight; All Three Cities; Bounded Errors of
       Commission Method; Predicted Population
       (E  set at 5 percent)  	  1-101

1-55.  Comparison of Multiple Constituent Results by
       Inertia Weight; All Three Cities; Bounded
       Errors of Commission Method; Actual Population
       (E  set at 5 percent)  	  1-102

1-56.  Comparison of Short Test HC Results  by  Emission
       Controls; All Three Cities; Predicted  Population;
       Bounded  Errors of Commission Method (E  set
       at 5 percent)	C	  1-104

1-57.  Comparison of Short Test CO Results  by  Emission
       Controls; All Three Cities; Predicted  Population;
       Bounded  Errors of Commission Method (E  set
       at 5 percent)	c	  1-106

1-58.  Comparison of Short Test NOX Results by Emission
       Controls; All Three Cities; Predicted  Population;
       Bounded  Errors of Commission Method (E  set at
       5 percent)  .	 . . ,c	  1-108

1-59.  Comparison of Short Test Multiple Constituent
       Results by Emission Controls; All Three Cities;
       Predicted Population; Bounded Errors of
       Commission Method (E   set at 5  percent)  	  1-109

1-60.  Comparison of Short Test HC Results  by  Fuel
       System Type;  All Three Cities; Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)	  1-111

1-61.  Comparison of Short Test CO Results  by  Fuel
       System Type;  All Three Cities; Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)  	  1-112

1-62.  Comparison of Short Test NOX Results by Fuel
       System Type;  All Three Cities; Predicted Population
       Bounded  Errors of Commission Method
       (E  set at 5 percent)	  1-114
                                  -xix-

-------
                           TABLES (Continued)
1-63.  Comparison of Short Test Multiple Constituent Results
       by Fuel System Type; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
      . (E  set at 5 percent)	  .  1-115

1-64.  Comparison of Short Test HC Results by
       Transmission Type:  All  Three Cities; Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)	  1-117

1-65.  Comparison of Short Test CO Results by  Trans-
       mission Type; All Three  Cities; Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)	  1-118
                 c

1-66.  Comparison of Short Test NOX Results by Trans-
       mission Type; All Three  Cities; Predicted Population;
       Bounded Errors of Commission Method (E set
       at 5 percent)	 . . . .C	  1-120

1-67.  Comparison of Short Test Multiple Constituent Results
       by Transmission Type; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)	  1-121

1-68.  Comparison of Short Test HC Results by  Manufacturer;
       All Three Cities; Predicted Population; Bounded
       Errors of Commission  Method (E  set at 5 percent)   	  1-123

1-69.  Comparison of Short Test CO Results by  Manufacturer;
       All Three Cities; Predicted Population; Bounded
       Errors of Commission  Method (E  set at 5 percent)	  1-124

1-70.  Comparison of Short  Test NOX Results by Manufacturer;
       All Three Cities; Predicted Population; Bounded
       Errors of Commission  Method (E  set at 5 percent)	  1-126

1-71.  Comparison of Multiple Constituent Results by
       Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission
       Method (E  set at 5 percent)"	  1-128
                                  -xx-

-------
                           TABLES (Continued)
1-72.  Comparison of HC Results for  150 CID and Less
       Group; by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)	  1-130

1-73.  Comparison of CO Results for  150 CID and Less
       Group; by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)  	  1-132

1-74.  Comparison of NOX Results for 150 CID  and Less
       Group; by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)	  1-134

1-75.  Comparison of Multiple Constituent Results for
       150 CID and Less Group; by Manufacturer; All
       Three Cities;  Predicted Population; Bounded Errors
       of Commission Method (E  set at 5 percent)	  1-137

1-76.  Comparison of HC Results for  151 to  259 CID Group;
       by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)  	  1-139

1-77.  Comparison of CO Results for  151 to  259 CID Group;
       by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)  	  1-141

1-78.  Comparison of NOX Results for 151 to 259 CID Group;
       by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)  .  .	  1-144

1-79.  Comparison of Multiple Constituent Results for
       151 to 259 CID Group; by Manufacturer;  All Three
       Cities; Predicted Population; Bounded Errors of
       Commission Method (E  set at 5 percent)  	  1-146

1-80.  Comparison of HC Results for  260 CID and Greater
       Group; by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)	  1-148
                                  -xxi-

-------
                          TABLES (Continued)
1-81.  Comparison of CO Results for 260 CID and Greater
       Group; by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)	  1-150

1-82.  Comparison of NOX Results for 260 CID and Greater
       Group; by Manufacturer; All Three Cities; Predicted
       Population; Bounded Errors of Commission Method
       (E  set at 5 percent)  	  1-151

1-83.  Comparison of Multiple Constituent Results for
       260 CID and Greater Group; By Manufacturer; All
       Three Cities; Predicted Population; Bounded Errors
       of Commission Method (E  set at 5 percent)   	  1-152

1-84.  Comparison of Federal Short Cycle Cut-Points by
       City (E  set at 5 percent)	  1-155

1-85.  Comparison of Federal Three-Mode Cut-Points by
       City (Ec set at 5 percent)	  1-156

1-86.  Comparison of Federal Short Cycle Cut-Points by
       Engine Displacement (Ec set at 5 percent)	  1-157

1-87.  Comparison of Federal Three-Mode Cut-Points by
       Engine Displacement (Ec set at 5 percent)	  1-158

1-88.  Federal Short Cycle STE Comparison; 5 Percent
       Ec Fleet Average vs 5  Percent Ec for Each of
       Three CID Groups	  1-160

1-89.  Federal Three-Mode STE Comparison; 5 Percent
       Ec Fleet Average vs 5  Percent Ec for Each of
       Three CID Groups	  1-161
                                  -xxii-

-------
                                FIGURES



1-1.   Contingency Table Representation  	   1-8

1-2.   Bounded Errors of Commission Method	   1-9

1-3.   Piecewise Linear Deterioration for FF Vehicles
       Shown Over Two-Year Period	   1-22

1-4.   Variation of Ec, Eo, and  FF with HC Cut-point;
       All Three Cities; Federal Short Cycle; Bounded
       Errors  of Commission Method	   1-78

1-5.   Variation of Ec, Eo, and  FF with HC Cut-point;
       All Three Cities; Federal Three-Mode, Bounded
       Errors  of Commission Method	   1-78

1-6.   Variation of Ec, Eo, and  FF with CO Cut-point;
       All Three Cities; Federal Short Cycle; Bounded
       Errors  of Commission Method	   1-81

1-7.   Variation of EC, EQ, and  FF with CO Cut-point;
       All Three Cities; Federal Three-Mode; Bounded
       Errors  of Commission Method .	   1-81

1-8.   Variation of Ec, E0, and  FF with NOX Cut-point;
       All Three Cities; Federal Short Cycle; Bounded
       Errors  of Commission Method	   1-83

1-9.   Variation of Ec, E0, and  FF with NOX Cut-point;
       All Three Cities; Federal Three-Mode; Bounded
       Errors  of Commission Method	   1-83
                                  -xxui-

-------
                              HIGHLIGHTS
      A series of statistical analyses was performed to determine the degree
of "correlation" that exists between two specific short tests (S,Ts) and the
Federal Emission Certification Test Procedure (FTP) for new vehicles.
This work was performed to determine if "reasonable  correlation with
certification test procedures" exists; this is a condition precedent to the
promulgation of regulations that impose the in-use warranty provisions of
Sec. 207(b) of the Clean Air Act of 1970 upon the motor vehicle
manufacturers.
      The basis for  the analyses was ST and FTP  test  data from three
100-vehicle fleets located in:
                             Chicago, Illinois
                             Houston, Texas
                             Phoenix, Arizona
Each of the vehicles in these  fleets was tested by the FTP and the following
STs:
                             Federal Short Cycle
                             Federal Three-Mode
The first of these STs is  a CVS (constant volume sampling) or bag-type test
wherein a test technician drives the car on the  dynamometer in accordance
with a prescribed driving pattern. The vehicle exhaust is diluted by the
CVS procedure, and a single  sample bag of diluted exhaust is collected for
the ST.  The latter ST is categorized as modal or  volumetric.  In this test,
the test technician operates the vehicle on a dynamometer at a fixed vehicle
speed and dynamometer load,  or at idle.  The vehicle  tailpipe.exhaust is
sampled directly, and the concentration of each pollutant  is measured and
recorded.  The Federal Three-Mode ST has high-speed,  low-speed, and
idle modes.  The idle mode can be conducted in two  ways, (1) idle with trans-
mission in drive (automatic transmission only), or (2) idle in neutral
(automatic and manual transmissions).
                                   H-l

-------
      Hydrocarbon (HC) and carbon monoxide (CO) measurements were
recorded with garage-type instruments for the Federal Three-Mode ST.
All oxides  of nitrogen (NO ) measurements were made with laboratory
analyzers.
      Two  different statistical analysis methods were used to assess
"correlation" -- a conventional correlation analysis,  and a contingency
table  analysis.
      The principal results of the study are briefly summarized below.
A summary overview and discussion are presented in Section 1, which
provides more general statements and findings as well as a detailed dis-
cussion and summarization of each study area, as a function of analysis
approach.
H-l.  SHORT TEST COMPARISONS; 300-VEHICLE
      FLEET AVERAGES 1
      a.     The Federal Short Cycle exhibited higher short-test-effectiveness
            (STE)   values  as compared with the Federal Three-Mode for all
            three pollutants (HC,  CO, and NO ).
                                            2C
      b.     For the Federal Three-Mode, the idle test modes in drive (auto-
            matic transmission) and neutral (both automatic and manual
            transmissions)  were superior to the high and low speed test
            modes  for tracking HC and CO.  The high-speed mode was best
            for NO  discrimination,  although the low-speed mode STE value
                  .5C
            (0. 301) was not significantly different from the high-speed value
            (0. 310) (some vehicles had higher NO STE values in the low-
                                                ,x
            speed mode than in the high-speed mode).
 Findings and values noted are based on a single set of ST cut-points
 selected to yield an error of commission (Ec) rate of 5 percent for each
 pollutant for the  300-vehicle fleet considered as a whole.
2       % of vehicles correctly failed by the short test
      ~        % FTP failures  in same fleet
                                   H-2

-------
      c.    In the case of CO, which was the dominant failure mode on the
            FTP for the 300-vehicle fleet, the idle-in-drive test mode STE
            value (0. 562) was  reasonably close to the Federal Short Cycle
            STE value (0. 690).
      d.    On a multiple-constituent test basis, *  the idle-in-drive test
            mode STE value (0. 704) of the Federal Three-Mode is reasonably
            close to the Federal Short Cycle STE value (0. 821).
      e.    The idle-in-neutral test mode, which would be required for cars
            with manual transmissions,  was lower in STE value than the
            idle-in-drive test  mode:  10 percent lower for HC; 7 percent
            lower for CO; and 15 percent lower on the multiple-constituent
            test basis.
      f.     Using a 5  percent  E   rate to select the short test cut points,  a
                              c            2
            short test rejection ratio (STRR)  value of one was not exceeded
            in any test mode; i.e. ,  the total number of vehicles  rejected by
            the ST as  failures  (E  +FF)  did not exceed the number  of vehicles
            actually failing the FTP (E  +FF).  By selecting  cut-points based on
            the emission characteristics of the pooled 300-vehicle fleet, then,
            the auto manufacturers  would not have been subjected to a warranty
            liability (in terms  of total number of vehicles rejected)  in excess
            of that properly justified on the basis of the FTP failure rate.
H-2  IMPACT OF CUT-POINT DETERMINATION METHOD
      a.     When the ST cut-points  were based on a 5 percent E  rate for
            the total 300-vehicle fleet, which was the  basis for the H-l
 A car fails the ST if any one of its HC,  CO, or NO  measurements
 exceed the cut-points for each pollutant.
2      _  number of  vehicles rejected by the ST
       - number of vehicles rejected by the FTP
3
 Ec  = error of commission
 Eo  = error of omission
 FF = vehicles failed by both the ST and the FTP
 PP = vehicles passed by both the  ST and the FTP
                                   H-3

-------
      findings above, higher ST failure rates were observed for the
      smaller vehicles than for the larger vehicles in terms of CO and
      NO  discrimination.  This was exemplified by generally higher
      E  values (and concomitant STRR values) for:
            (1)   the 150 CID and less  displacement group
            (2)   the 2500-lb and less weight group
            (3)   manual-transmission-equipped vehicles
            (4)   fuel-injected vehicles.
      For example,  in the case of CO the  150 CID and less group
      had a  12. 5 percent E  rate while the 260 CID and  above group
      had a  1.8 percent E  rate.
                                #
      The small CID, low  vehicle weight,  and manual transmission
      factors are all typical  attributes  of the small-vehicle  class.
      Fuel injection, at present, is also principally used in the small-
      vehicle class.   The observed higher short test failure  rate of the
      smaller vehicles  can be interpreted as inequitable to the small
      vehicle class.
b.    Similarly, there was a considerable variation between vehicle
      manufacturers with regard to E  rate  and  STE and STRR values.
      For example,  using  the idle-in-drive test  mode for  a multiple-
      constituent test:  E  rates varied from 1.6 to 9. 3 percent; STE
      values ranged  from 0. 500 to 0.800; and STRR values ranged
      from 0. 518 to  0.909.
c.    When  three sets of ST cut-points were used to  minimize engine
      displacement-related effects, based on a 5 percent E   rate for
      each of three CID  classes (150  CID and less; 151 to  259 CID;
      260  and greater CID), only very minor changes were noted to
      overall fleet STE values for the individual  STs or  test modes
      (as obtained  with  the single  set of ST cut-points based on a
      5 percent E  rate for the fleet).  However, there was  considerable
                             H-4

-------
variation in E  , STE, and STRR values between manufacturers
             c
within each CID class.  Also, for a given manufacturer, there
were similar variations between CID classes.
Even though the overall fleet STE values  were essentially
unchanged with the three sets of CID-related cut-points, 36 per-
cent more large vehicles  (260 CID and greater) were correctly
identified as CO failures than with the single set of cut points.
Since the mean CO level of this large vehicle class exceeds the
FTP CO standard by  14. 5  g/mi (compared to the  150 CID and
less group  with a mean CO level that exceeds the FTP standards
by only 1. 8 g/mi), the multiple cut-point approach based on CID
class represents a greater potential for overall CO reduction.
The above findings indicate that:
      (1)    A  single set of cut-points (based on a given  E  rate
            for the total fleet), while providing reasonable STE
            and STRR values on an overall fleet basis, results
            in inequities between vehicle  size classes and vehi-
            cle manufacturers.
      (2)    Multiple cut-points based on engine CID class largely
            resolve the inequities among  vehicle size  classes;
            however, large variations between manufacturers in
            E  , STE, and  STRR values still persist under  this
            scheme.
      (3)    It may be desirable to select ST cut-points on the
            basis of CID (or size) and vehicle manufacturer in
            order to minimize the E ,  STE, and STRR variations
                                   c
            noted above.  There may be the  further consideration
            of emission control and drive train technology, as
            discussed in H-3 below.
                        H-5

-------
                 (4)   Since the 5 percent E  rate resulted in STRR values
                       less than one, the number of correctly failed vehicles
                       in all classes could be increased merely by selecting
                       the ST Cut-points to result in an STRR value of one.
                       By this technique, the maximum number of correctly
                       failed vehicles would be obtained under the constraint
                       of not exceeding the number of vehicles  that should
                       have been failed based on the observed FTP failure
                       rate.
H-3.  EFFECT OF EMISSION CONTROL. FUEL SYSTEM.
      AND TRANSMISSION TYPE
      a.    Comparisons were made of the effects of three other sub-groups
           on the basis of a single  set of cut-points for the 300-vehicle fleet
           (5 percent E   rate):
                 1.    the use or non-use of catalysts and /or secondary
                       air injection
                 2.    carburetion vs fuel injection
                 3.    manual vs automatic transmissions
      b.    No significant STE patterns  or trends were observed for the
           catalyst /secondary air sub-group.  However, STRR values of
           one were exceeded by the "without catalyst /with secondary air"
           category for CO and by the "without either catalyst or secondary
           air" category for NO  .
                               X
      c.    Fuel-injected and carbureted vehicles had different and varying
           STE values; no significant patterns were observed.  The fuel-
           injected vehicles, however,  exceeded STRR values of one for
           both CO and NO  .
                          x
      d.    Vehicles with manual transmissions had lower HC STE values
           and higher CO STE values than vehicles with automatic trans-
           missions.   With the high-speed test mode,  the manual trans-
           mission NO  STE value was  150 percent higher than that of the
                                   H-6

-------
            automatic transmission.  NO  STRR values exceeded one for
                                       jC
            the manual transmission (high-speed test mode).
      e.    Thus, the STs examined appear to be applicable to the above
            control system and vehicle technological variations,  with the
            trends as noted.  However,  the  extent to which these trends
            would persist or change with other cut-point selection criteria
            (e.g., cut-points based on CID class,  or based on CID class
            and manufacturer) is not known  at this time because  of an
            insufficient test sample size to support the required  analyses.
H-4.  MULTIPLE-CONSTITUENT VERSUS SINGLE-
      CONSTITUENT TESTS
      a.    Single-constituent tests on the Federal Three-Mode require at
            least two test modes to discriminate all three regulated pollu-
            tants (e.g.,  idle mode  for HC and CO,  high or low speed mode
            for NO ).  A multiple-constituent test using only the  idle-in-
                  3t
            drive test mode produced a higher STE value (0.704) than the
            next highest single test mode (the idle-in-drive mode for CO
            discrimination, at 0.562).
H-5.  SIMULATION OF 207(b) IMPLEMENTATION
      a.    The implementation of a 207(b) program was simulated with
            computer models.  The 300-car data set was used to model the
            fleet  emission characteristics and to establish the ST inspection
            pass-fail levels.  Because of uncertainties in the  areas  of
            deterioration rates and maintenance effectiveness, sensitivity
            analyses were performed by using several different maintenance
            models and by varying  many other parameters.
      b.    With  the first inspection occurring when each vehicle is one
            year  old (with an  accumulated mileage of  17, 500 miles), and
            with succeeding inspection and maintenance cycles occurring
                                    H-7

-------
            annually,  it was determined that the following cumulative
            program efficiencies  would result at the end of 50,000 miles
            of travel:
                       HC   = 22. 1 to 34.6 percent
                       CO   = 24. 6 to 42. 5 percent
                       NO   = 0. 33 to 2.00 percent
            The above ranges of program effectiveness values are the result
            of specific assumptions made in the modeling process,  and
            principally reflect variations  in duration of the period of mainte-
            nance effectiveness ^  plus the degree of effectiveness of the
            maintenance performed at the time of inspection.
 Percent efficiency is the percent reduction in fleet emissions from that
 value which would have occurred without the  207(b) program, over the
 mileage interval 17, 500 to 50, 000 miles.
2
 The period of maintenance effectiveness is the  time interval (in months)
 required for emissions to return to pre-maintenance levels.
                                   H-8

-------
 1.1
 1.1.1
                     1.  SUMMARY AND DISCUSSION
               SUMMARY OVERVIEW
               Conventional Correlation Analyses
               A conventional correlation analysis was performed for the
three-city 300-vehicle 1975 model year fleet for both the Federal Short
Cycle and the Federal Three-Mode STs.  The correlation coefficient is
the quantitative measure of relatability between the  results of the short
test (ST) and the Federal Test Procedure (FTP).  The  closer  the coefficient
is to one, the better the relation.  No relationship is indicated by a coeffi-
cient value of zero. A negative  coefficient indicates an inverse  relation
between the observed  test results.
1. 1. 1.1
              Short Test (ST) Comparison
              A comparison of the Federal Short Cycle results and the
better modes of the Federal Three-Mode ST is shown below in Table  1-1.
          Table  1-1.  Comparison of ST Correlation Coefficients
                      (Pooled Fleet of 300 Vehicles)


Short Test

Federal Short Cycle
Federal Three-Mode

ST/FTP
Correlation Coefficient

HC
0.78
0.58(a)
0.45
CO
0.89
0.67(a)
0.69(b)
NOX
0.81
0.52
0. 48
         (a)
         (b)'
            Idle in drive test mode
            Idle in neutral test mode
(c)
(d)
High speed test mode
Low speed test mode
                                    1-1

-------
               The Federal Short Cycle exhibited superior FTP tracking
 characteristics (higher correlation coefficients)  as compared to the Federal
 Three-Mode for all three pollutants (HC,  CO and NO )*.  For the Federal
 Three-Mode ST, the idle test mode in neutral (both automatic and manual
 transmissions) and drive (automatic transmission only) was superior to the
 high and low-speed test modes for tracking HC and CO.  The high-speed
 mode was best for NO  discrimination, although the low-speed mode  coeffi-
                     5£
 cient was not significantly different from the high-speed value.
 1.1.1.2       Effect of Accumulated Mileage
               Seventy of the 300 vehicles had accumulated 4000 miles or
 less of travel prior  to being tested.  The only noticeable effect for the
 Federal Short  Cycle is that these 70 vehicles had a higher HC correlation
 and lower CO correlation than the  230 vehicles with greater than 4000 mileage
 accumulation.   In the case of the Federal Three- Mode, the results for the
 greater than 4000 mileage group were generally  similar to those  for the
 total 300-vehicle fleet.  Those vehicles with 4000 miles or less,  however,
 exhibited generally poorer  correlation coefficients than the fleet average,
 except for HC  in the idle in drive test mode which was  higher than the fleet
 average.  This is in agreement with the HC results noted above for these
 same low mileage vehicles when tested by the Federal  Short Cycle ST.
 1.1.1.3       Effect of Engine Displacement
               In the case of the  Federal Short Cycle,  NO correlations
                                                        X.
were largely unaffected by engine cubic inch displacement (CID) class and
were  similar to the total fleet values.  The CO correlations  of the 150 CID
or less group were significantly lower than the fleet average (and larger
 engine size  groups).  In  terms of HC discrimination,  the 260 CID or more
*
 HC  = hydrocarbon
 CO  = carbon monoxide
 NO  = oxides of nitrogen
                                   1-2

-------
engine group had a significantly higher correlation coefficient than the fleec
average,  the 150 CID or less group was slightly less than the fleet average,
and the 151 to 259 CID  group was significantly lower (0. 53) than the fleet
average (0. 78).  A possible explanation for the poorer HC and CO correlation
values for the small  CID group is that this group incorporated a multiplicity
of technologies for HC  and CO control (e. g. , with and without catalysts, both
carburetion and fuel  injection, with and without air injection), whereas the
large displacement vehicles were principally equipped with carburetors and
oxidation  catalysts.
              For the  Federal Three-Mode ST,  similar  trends  are observed
for NO  in the high-speed mode and for CO in the idle in  drive mode.  HC
      X.
correlation coefficients in the idle in  drive  mode for the two smaller CID
groups  are  higher than the 260 CID and above group, and also higher than
the fleet average.  Again,  the better test  modes of the Federal Three-Mode
(for a given pollutant) tend to show the same types of trends with engine
displacement as the Federal Short Cycle, except for HC.
1.1.1.4      Effect of Inertia Test Weight Group
              The Federal Short Cycle results indicate that NC  correlations
                                  '                           x
are relatively insensitive to inertia weight, whereas the HC and CO correla-
tion coefficients of the  2500 Ib and less group  (subcompacts) are significantly
lower than the heavier  weight vehicles.  As noted above,  the  smaller vehicles
incorporated varying technologies  for HC and CO control, whereas the larger
vehicles were principally limited to carburetors  and catalysts for HC and CO
control.
              For the  Federal Three-Mode, the NO__ correlation coefficients
                                                   5C
are similarly relatively insensitive to inertia weight. The discernable
trends for HC and CO (in idle in drive and neutral test modes) again indicate
lower coefficient values for the smallest test weight group, as was the  case
for the  Federal Short Cycle.
                                    1-3

-------
1.1.1.5        Effect of Emission Control System Type
               The correlation coefficients of the three-city vehicle fleet
were determined as a function of the usage or non-usage of oxidation cata-
lysts and/or secondary air injection for control of HC and CO.  In the case
of the Federal Short Cycle results, the data indicates generally higher HC
and CO correlations with the use of a catalyst (with or without secondary
air); however, the 25 vehicles without either a catalyst or secondary air
had a CO correlation coefficient (0.86) as high as the 89 vehicles with both
a catalyst and secondary air  injection  (0.86).   NO  correlations are rela-
                                                3£
tively unaffected by the use or non-use of catalysts and/or secondary air
injection.
               The Federal Three-Mode results are  less clear.  For exam-
ple,  in the idle in drive mode, the HC correlation coefficients for those
vehicles  without a catalyst (or secondary air) (0. 85 and 0. 86)  are higher
than for  those vehicles with a catalyst (and secondary air) (0.  46 and 0. 62).
The  twelve vehicles without a catalyst or secondary air had  a  higher CO
correlation coefficient (0. 88) than vehicles with a catalyst (and secondary
air)  (0. 77 and 0. 67).   NO   correlations (in the high speed and low speed
                         5C
test  modes) were  relatively little affected by catalyst and secondary air
usage, as in the case of the Federal Short Cycle ST.
1.1.1.6        Effect of Carburetion vs Fuel Injection
               Thirty vehicles of the 300-vehicle test fleet were equipped
with fuel injection.  The fuel injected vehicles had significantly poorer
Federal  Short Cycle correlation coefficients for all three pollutants than
the carbureted vehicles.   With the Federal Three-Mode ST, this was
reversed for NO  and HC; in the high speed test mode the fuel injected NO
                X                                                       X
coefficient (0. 63) was  higher than the carbureted value (0. 53); and in the
idle  in drive test mode the fuel injected HC  coefficient (0. 80)  was higher
than the  carbureted value  (0. 58).  When the idle in neutral test mode is
considered, the HC comparison is reversed,  with the fuel injected coeffi-
cient (0.  36) lower than the carbureted value (0.46).
                                    1-4

-------
1. 1. 1.7       Effect of Transmission Type
               Seventy-one of tho 3 00-vehicle test fleet were equipped with
manual transmissions; the remainder were equipped with automatic trans-
missions.   With the Federal Short Cycle ST,  the NO  correlation coefficient
                                                   Jt
was unaffected by transmission type, whereas both HC and CO correlation
coefficients were significantly reduced for manual transmission vehicles.
               In the case of the Federal Three-Mode ST, the situation is
similar for NO . For HC and CO,  the only comparative basis is the  idle
               5C
in neutral test mode,  since the manual transmission prevents the idle in
drive test mode.  Here the CO correlation coefficients are similar for both
transmissions (0.71 and 0.67),  and the HC coefficient for the manual  trans-
mission (0.33) is considerably lower than the automatic transmission value
(0.49).
1.1.1.8       Effect of Manufacturer
               On the  Federal Short Cycle,  General Motors and AMC vehi-
cles had uniformly high (0.9  and above) correlation  coefficients for all three
pollutants.  Ford had  the highest HC and CO coefficients (0.94 and  0.96,
respectively)  but a lower NO coefficient (0.77). Chrysler had CO and NO
                            X.                                            5C
coefficients similar to Ford values, but had a lower HC coefficient (0.81).
The "others"  category (primarily small imported vehicles) had the lowest
HC (0.51) and CO (0.65) correlations, with a NO value (0.85) slightly higher
                                               2£
than Ford and Chrysler  values.  Again, the multiplicity of  technologies used
for HC and CO control in the small import class could be responsible for the
observed lower correlations  when  grouped into the single category.
               On the  Federal Three-Mode ST,  the high speed test  mode
resulted in generally similar NO  correlation coefficient levels (0.64 to
                                Ji
0.72) for all manufacturers except General Motors,  which  had a value of
0.42.  General Motors,  Ford and "others" vehicles  had slightly higher NO
                                                                         .X
coefficients in the low speed  mode than in-the high speed mode. In the case
                                    1-5

-------
of CO in the idle in drive test mode,  Ford, General Motors, and AMC
relative rankings were similar to that of the Federal Short Cycle, however,
Chrysler CO coefficients were lower (0. 68) than the other domestic manu-
facturers (0. 73 to 0. 84).  The "others"  category had a very low (0. 32) CO
coefficient in the  idle in drive mode.   Conversely, the HC coefficient of the
"others"  category in the idle in drive mode was the highest  (0. 82) of all
manufacturer groupings.  AMC also maintained a high HC coefficient (0. 81)
in this same test  mode.  The  other three domestic-manufacturers had
similar, but lower, HC coefficients in the 0. 55 to 0. 63  range.
1.1.2           Contingency Table Analysis
                The contingency table analysis technique  was used to establish
the ST pass-fail levels for each pollutant.   The contingency table is defined
in Table 1-2, along with its associated parameters.  A  pictorial demonstra-
tion of its application to a given data  set is shown in Figure 1-1. It can be
seen that, for a given data set, part of the analysis is concerned -with the  cri-
teria used to select the ST cut-points.  In this regard, the bounded errors of
commission method was used extensively to establish trends for the  variations
in E ,  E  ,  FF, and PP (see definitions  in Table 1-2).   In this  method, the ST
cut-points are  selected to minimize E  while  holding the E  below a  specified
level.  It thus permits a direct answer to the  question,  "For a given per-
missible level  of  E ,  what level of E  is associated with the ST, and with
                   c                o
what impact on air quality (inferred from number of FF and E  vehicles)?"
This method is illustrated in Figure 1-2.  The policy  decision  is the maxi-
mum allowable E .
                 c
                With regard to procedural technique,  a bivariate normal or
log-normal distribution model was fitted to a  particular data set by  incor-
porating the correlation coefficient, mean values,  and standard deviations of
the data set.  The ST cut-points were then determined by using the model  for
the predicted population of the 300-vehicle fleet.
                                    1-6

-------
                      Table 1-2.  Contingency Table

Predicted =
Short Test
Pass
Fail
Total
True = FTP
Pass
a
c
a + c
Fail
b
d
b + d
Total
a + b
c + d
n = a + b
+ c -f d
                a - number of correctly passed vehicles (PP)
                b = number of errors of omission (E )
                c = number of errors of commission (E )
                                                      c
                d = number of correctly failed vehicles (FF)
          Sensitivity = a/(a + c)         \
          Specificity = b/(b + d)
          False positive error = b/(a + b)
          False negative  error =  c/(c + d)              Q
          „     ...    .  ,	ad - be	
          Correlation index =	;T>
                             [(a  + b)(a + c)(b + d)(c H
              Also determined in the contingency table analysis were the
short test "effectiveness" (STE) and the short test "rejection ratio" (STRR).
ST effectiveness is defined as:
        ST effectiveness = STE =
	% FF for the short test
% FTP failures in same population
    % FF
(1-D
                                 % FF + % E
Thus,  on this basis,  the ST is always less effective than the FTP, in propor-
tion to the percent of errors of omission (E ) associated with a given ST,
                                   1-7

-------
                    E  ERROR OF
                     v

                    COMMISSION
    ST CUT-POINT
                                           FF, CORRECTLY
                                           FAIlfD VEHICLES
a:
Z3
00
                                             EQ, ERROR OF

                                             OMISSION
          PP, CORRECTLY
          PASSED VEHICLES
                                    a.
                                    t
                   FTP MEASUREMENT
        Figure 1-1.  Contingency Table Representation
                           1-8

-------
    MINIMIZE E  SUBJECT TO E  < Y%
               o             c
  CO
         NOT TO

        EXCEED Y%
         ST

     CUT-POINT
•


•
                          MINIMIZE
        FTP MEASUREMENT
Figure 1-2. Bounded Errors of Commission Method
                   1-9

-------
and the STE value is a relative measure of the impact on air quality from ST
implementation.
              The short test rejection ratio is defined as:

                                  % EC + % FF
                          STRR = % E  + % FF                          (1 -2
                                     o

The denominator of this  ratio (% E  + % FF) represents the percent of the
vehicle population actually failing the  FTP.  The numerator (% E  + % FF)
represents the percent of the vehicle population failing the short test.  Thus,
the resultant ratio (STRR) is an indication to flag those situations where the
implementation of a short test or test mode could result in a greater number
of vehicles being failed than should be failed,  based upon FTP standards.
Whereas the STE  value is a measure of the  relative impact of the use of the
ST on air quality, the STRR value, which includes consideration of those E
vehicles incorrectly identified by the ST as  failing the FTP, can be con-
sidered a relative measure of the "fairness" of the ST to the vehicle; manu-
facturer or vehicle size  or  class.  For example, only when the STRR value
exceeds one would the manufacturer be required to implement warranty
procedures on a greater number of vehicles than he should have, based upon
the observed FTP failure rate.
1.1.Z.I       Short Test Comparison
              A comparison of the Federal  Short Cycle results and the better
modes of the Federal Three-Mode ST is shown in Table  1-3.
              As  was the case with conventional correlation analysis  (Sec-
tion 1.1.1.1), the Federal Short Cycle exhibited  superior STE values  as
compared to  the Federal Three-Mode for all three pollutants (HC, CO and
NO ).   However, in the case of CO, which was the dominant failure mode
   5C
on the FTP for the 300-vehicle fleet, the idle  test mode STE value (0. 562)
was reasonably  close to  the  Federal Short Cycle value (0. 690).  Further,
when  considering the idle mode in drive as a multiple  constituent test (i. e. ,
a car fails the ST  of any one of its HC, CO,  or NO   measurements exceed
                                   1-10

-------
  Table 1-3.  Comparison of ST Contingency Table Values for STE and STRR;
     All Three Cities (E   Set at 5 Percent for Pooled Fleet of 300  Vehicles)
Short
Test
Federal Short Cycle
Federal Three-Mode
High Speed Mode
Low Speed Mode
Idle in Drive
Idle in Neutral
Combination of
Best Modes (b>
Parameter
HC
STE
0. 589

-
-
0. 375
0. 338
-
STRR
0. 748

-
-
0. 533
0.495
-
CO
STE
0.690

-
-
0. 562
0. 524
"
STRR
0.783

-
-
0.652
0.614
-
NO
X
STE
0.604

0. 310
0. 301
-
-
-
STRR
0.840

0. 547
0. 538
-
-
-
Multiple . .
Constituents
STE
0.821

-
-
0.704
0. 597
0.663
STRR
0. 928

-
-
0.801
0.669
0. 729
(a)   Car fails the ST if any of its HC, CO,  or NO measurements exceed the cut-points for each
     pollutant.
(b)   Combination of best individual test modes (idle in drive for HC and CO,  high speed for NO ).

-------
the cut-points  for each pollutant), then the idle mode STE value (0.704) is
reasonably close to the  Federal Short Cycle value (0.821).  The idle in
neutral test mode, although not quite as effective as the idle in drive mode,
would be required for those cars with manual transmissions.
               Table 1-4 shows the variation of STE values achieved in the
three cities in which the vehicles were tested.  The most notable difference
is the Federal Three-Mode NO  STE value in Phoenix, which is approximately
                              3C
twice that of the other two cities.  The humidity in Phoenix is considerably
lower and more consistent in level than in the other two cities.  The NO
                                                                      ji
correction factor  for humidity was applied to the Federal  Short Cycle and
the FTP test results, but was not applied to  the Federal Three-Mode results.
This accounts  for the relative uniformity of correlation existing for the
Federal Short  Cycle in all three cities and the  non-uniformity of the Federal
Three-Mode results.  If the NO  humidity correction factor had been applied,
                               JX
the Chicago and Houston results would probably be similar to the results
obtained in Phoenix.
               With regard to short test rejection ratio, the STE values of
Table 1-3 were achieved without  exceeding an STRR value of one in any case,
i. e.,  the total number of vehicles rejected by the ST as failures (E + FF)
did not  exceed the number of vehicles actually  failing the FTP (E  + FF).
By selecting cut-points based on the emission characteristics of the pooled
300-vehicle fleet,  then,  the  auto manufacturers would not have been sub-
jected to a warranty liability (in terms  of total number of vehicles rejected)
in excess of that properly justified on the basis of FTP failure rate.
               The above findings are restricted solely to the case  of overall
fleet averages. Because the makeup of the  300-vehicle test fleet represented
a cross-section of the 1975 model year in-use  population,  the test fleet
included the many variables of manufacturer, vehicle  size (inertia  test weight),
engine size (displacement), emission control system type,  transmission.type,
and fuel system type.  As would be expected, individual findings for each of
these sub-groups of the total fleet varied somewhat  from that for the overall
                                    1-12

-------
Table 1-4.  Comparison of STE Values for Three Cities

Short Test
Federal Short Cycle
Federal Three-Mode
High Speed
Low Speed
Idle in Drive
Idle in Neutral
Short Test Effectiveness (STE)
Chicago
HC
0. 554

-
-
0.405
0. 352
CO
0.664

-
-
0. 539
0. 509
N0x
0. 618

0. Z04
0. 177
-
-
Houston
HC
0.691

-
-
0. 360
0. 348
CO
0. 731

-
-
0. 584
0. 536
NO
X
0. 562

0. 249
0.291
-
-
Phoenix
HC
0.530

-
-
0. 349
0.296
CO
0.676

-
-
0. 551
0. 517
NO
X
0.645

0.535
0.471
-
-

-------
fleet.  A summary of the more significant variations is presented in the
following sections.  They are discussed in depth in Sections 1.3.5 through
1.3. 11.
              The reader is reminded that the following remarks pertain to
the 300 vehicles being analyzed as a single group and using a 5 percent E
rate to select the cut-points.
1.1.2.2      Effect of Engine Displacement
              With the Federal Short Cycle,  the most  significant effect of
engine displacement was displayed by the 150 CID  and less group.  It has a
CO STE value (0.773) higher than the fleet average (0.690) and a NO  STE
                                                                 X.
value (0.530) below the fleet average (0.604).  The high CO STE value was
accompanied by a CO STRR value of 1.051.
              On the Federal Three-Mode ST (idle in drive mode) the same
higher CO STE trend of the 150 CID and less group was also evident; how-
ever, the NO  STE value (0. 578 high speed mode)  was  appreciably higher
than the fleet average value (0.310) and was  accompanied by an NO STRR
                                                                .X
value of 1.219, that is, approximately 22 percent more vehicles failed the
ST than failed the FTP.
1.1.2.3      Effect of Inertia Test Weight Group
              The most noticeable effects of inertia test weight were asso-
ciated with the 2500 Ib and less weight group.  On  the Federal Short Cycle,
this  group had a higher than fleet average CO STE value and had markedly
lower than fleet average HC and NO  STE values.  It had an STRR value of
1. 064 in the case of CO.  On a multiple constituent test basis this group also
exceeded an STRR value of one.
              On the Federal Three-Mode ST, these same trends for the
2500 Ib and less weight group were evident except  for NO .  Using the high
                                                       X.
speed test mode, the NO  STE value was the highest of the four weight groups,
                       .X
accompanied by a very high STRR value of 3.0, which was due to a very high
(13.2 percent) E   rate.
                                   1-14

-------
 1.1.2.4       Effect of Emission Control System Type
               With the  Federal Short Cycle ST,  there appear to be no readily
 observed STE trends attendant to the use or non-use of catalysts and/or
 secondary air injection.  However, STRR values in excess of one were
 associated with those vehicles (1) without a catalyst but with secondary air
 in the case of CO,  (2) with a catalyst but without secondary air in the  case
 of NO , and  (3) without  either a catalyst or secondary air in the case  of
      ji
 multiple constituents.
               On the  Federal Three-Mode, again no  significant STE patterns
 or trends were observed.   Those vehicles  with both a catalyst and secondary
 air injection did have lower STE values for all three pollutants and on a
 multiple constituent test basis.  In the case of rejection ratio,  an STRR value
 of one was exceeded by  the "without  catalyst/with secondary air" group for
 CO (idle in neutral test  mode) and by the "without either catalyst or secondary
 air" group for NO  (high speed test mode).
 1.1.2.5       Effect of Carburetion vs Fuel Injection
               With the  Federal Short Cycle ST,  the fuel-injected vehicles
 had STE values 20  percent lower than carbureted vehicles for HC, 30  percent
 higher for CO,  and 70 percent lower for NO .  In the case of both CO  and
                                          j£.
 multiple constituents,  the fuel injected vehicles had STRR values in excess
 of one.
              On the Federal Three-Mode ST, the above HC and NO   trends
                                                                 X.
were  reversed. The HC STE value of the fuel-injected vehicles was 8 per-
cent higher than that of the carbureted vehicles and the NO  STE value was
       &                                                x
42 percent higher (high speed test mode).   The fuel injected vehicles
exceeded an STRR value of one for both CO (1.082) and NO  (1 .480).
                                                        5C
1.1.2.6       Effect of  Transmission Type
              With the Federal Short Cycle ST, those  vehicles with manual
transmissions,  compared to those with automatic transmissions, had STE
values which were 30 percent lower for HC, 10 percent higher for CO, and
                                  1-15

-------
45 percent lower for NO .  They also exceeded an STRR value of one for
CO (1.032).
              On the Federal Three-Mode ST, the same HC and CO trends
were observed as noted above.  However, the NO  trend was reversed, with
the manual transmission STE value (high speed test mode) 150 percent higher
than that of the automatic transmission.  For NO  the manual transmission
                                               3t
had an STRR value  of 1. 538.
1.1.2.7       Effect of Manufacturer
              With the Federal Short Cycle ST, the most notable effect was
the very high NO  STE value (0.957) recorded by AMC.  This was accom-
                JC
panied by a 14.5 percent E rate and an STRR value of 2.463.  The "others"
                          c
category (principally small imports) had an STRR value of 1.121 for CO.
Both AMC and the "others" exceeded an STRR value of one on a multiple
constituents test basis.
              On the Federal Three-Mode, the above effects were not noted.
Chrysler had the highest CO STE value (0.898, idle in drive) accompanied by
an STRR value of 1.016.  In the case of NO  , the "others" category had the
                                         n
highest STE value in the high speed mode (0.467), whereas Chrysler had an
STRR value of 2.129 and the highest STE value (0. 755) in the low speed test
mode.
1.1.2.8       Effect of Cut-Point Determination Method
              The  foregoing analysis results indicated-that,  when the ST
cut-points were based on a 5 percent E  rate for the total 300-vehicle fleet,
there were possible inequities in terms  of CO and NO  discrimination for
the smaller vehicles.  This was exemplified by the higher E  and concomi-
                                                          C
tant STRR values for the 150 CID and less displacement group,  the 2500 Ib
and less  weight group, the fuel-injected vehicles,  and the manual-
transmission-equipped vehicles.  The small CID,  low vehicle weight, and
manual transmission factors are all typical  attributes of the  small vehicle
class.  Fuel injection, at its present degree of usage, is also principally
used in the small vehicle class.
                                   1-16

-------
               Therefore an additional contingency table analysis was made
wherein the ST cut-points were determined for each CID class, based on a
5 percent E  rate for that CID class alone,  not the pooled 300 vehicle fleet
as in all previously discussed results.  The results of this second analysis
is compared with the first contingency table analysis in Table 1-5.  As  can
be seen, selecting a 5 percent E  rate for each CID group resulted in very
minor changes to overall STE and STRR values for the individual STs or
modes.  In addition, all STE values were improved, except for (1) the multiple
constituent test basis on the Federal Three-Mode and  (2) NO   on the Federal1
                                                          Ji
Short Cycle and low speed test mode of the Federal Three-Mode.
               These results indicate that the more equitable CID class  cut-
point determination approach can produce the same benefit in air quality
(as indicated by STE value) while reducing the probability of excessive fail-
ures of small vehicles by ST implementation. Another approach, not
examined in the present study, would be to select cut-points on the basis
of a  5 percent E   rate for  each vehicle manufacturer within each CID class.
1.1.3          Simulation of 207(b) Implementation
               A computer program was developed to simulate the imple-
mentation of a 207(b) program on a  specific fleet of vehicles, the "controlled
fleet", and to compare the emission results to an uncontrolled fleet,
originally identical to the controlled fleet,  which deteriorates without the
influence of a mandatory 207(b) program.  The comparison was made over
a 50,000-mile period for each fleet, and the effectiveness of  the 207(b)
program was assessed annually.
               The two simulated fleets were  each divided into three mutually
exclusive groups based upon engine displacement:  150 CID and less;  151 to
259 CID; and 260 CID and greater.  Existing 300-car data was used to
establish a distribution model for the emission values of the three regulated
pollutants at the mean mileage for each displacement group.  This distribu-
tion  model was then used to  stochastically generate the two identical fleets
of vehicles with statistical attributes  similar  to the 300-car data set.
                                   1-17

-------
       Table 1-5.  Comparison of Cut-Point Determination .Methods;
             5 Percent Ec Fleet Average vs 5 Percent Ec for
                         Each of Three CID Groups

Short Test
Federal Short
Cycle


Federal Three -
Mode













Pollutant
HC
CO
NO
X
Multiple . .
Constituents (b)
HC

Idle in Drive
Idle in Neutral
CO
Idle in Drive
Idle in Neutral
NO
X
High Speed
Low Speed
Multiple . .
Constituents^
Idle in Drive
Idle in Neutral
Best Combina-
tion
Short Test Effectiveness (STE)
5% Ec Fleet
Average
0. 589
0.690
0.604
0. 821


0. 375
0. 338

0. 562
0. 524

0. 310
0. 301

0.704
0. 597

0.663
5% Ec for Each of, .
Three CID Groupsla'
0. 592
0.711
0. 573
0.847


0. 383
0. 339

0.605
0. 557

0. 341
0.291

0.653
0. 592

0.648
(a  150 CID and less
   151 to 259 CID
   260 CID and greater

   Car fails if any of HC, CO, or NO  measurements exceed cut-points for
   each pollutant                  x
                                     1-18

-------
               The existing 300-car data set was also used to establish the
ST inspection pass-fail levels.  Two methodologies in selecting the levels
were examined.  The first method, designated the  "Z07(b) approach",
established pass-fail levels for each engine displacement group at a 5  per-
cent error of commission rate for each pollutant.  The second method,
designated the "I/M approach", established, for each pollutant,  a  single pass-
fail level which applied to all groups.  The I/M (inspection maintenance)
pass-fail levels were determined using a 33 percent total failure  rate with a
three-constituent test.
               Because of uncertainties in the areas of deterioration rates
and maintenance effectiveness, a limited sensitivity analysis was performed
by using several different maintenance models and by varying many of the
program input parameters.  This provided visibility into the variability of
the results and identified the most sensitive components of the analysis
program.
               A maintenance policy was selected for each vehicle individu-
ally.  If the vehicle was classified by the ST  as an  E  or FF vehicle, the
FTP emission rates were adjusted to reflect the effect of corrective mainte-
nance.  If the vehicle was classified as  a PP or E  vehicle, the FTP
emission rates were  left undisturbed.  Three methods of adjusting the FTP
values were investigated, as discussed  below.
         a.     Version 1:  Proportionally Coupled with Lower Bounds
               For HC and CO, a pollutant  passing  the FTP is reduced
               by the same proportion as a pollutant failing the  FTP.
               The lower limit of any reduction,  beyond the FTP
               standard,  is one-half the FTP standard.  NO  is
                                                          x
              unaffected by HC and CO maintenance.
         b.     Version 2: Independent Maintenance to FTP Standard
               Each pollutant is adjusted independently of the values
               of the  other pollutants.  Failing pollutants are adjusted
               to the  respective FTP  standard.
         c.     Version 3: Independent Maintenance to Equivalent
               ST Standard.  Each pollutant is  adjusted independently
               of the  values of the other pollutants.  Failing pollutants
               are adjusted to an FTP value that corresponds to the
               ST cut-point.
                                   1-19

-------
               The general effects of age and wear were simulated by
deteriorating the emission rates from their value at the last inspection
and maintenance point to the next I/M point.  Deterioration effects on the
uncontrolled fleet were based upon linear models with mileage accumulation.
The vehicles were classified into two groups:
         a.    Vehicles passing the FTP
         b.    Vehicles failing  the FTP
Certification deterioration factors were applied to vehicles in group a, and
EFP factors were applied to vehicles in group b.  The values used are shown
in Table 1-6.

              Table 1-6.  Average Deterioration Rates  Used in
                      207(b) Effectiveness Simulation
Data Source
Certification Values
EFP Values
Rate of Deterioration
in Grams/Mile per 1000 Miles
HC
0.014
0.072
CO
0.084
0.800
NO
X
0.0
0.0
               Deterioration effects on the controlled fleet were based upon
piecewise linear models with mileage  accumulation.  The vehicles were
classified into four groups (E  ,  E , PP,  and FF vehicles) and the following
deterioration rates were applied.
               PP Vehicles: Certification rates as shown in Table 1-6
a.
b.
c.
               En Vehicles;  EFP rates  as shown in Table 1 -6
               EC Vehicles;  Certification rates shown in Table 1-6.
               Deterioration  occurs  after the maintenance adjustments.
                                     1-20

-------
          d.      FF Vehicles:  A maintenance effectiveness period is
                 specified as an input to the simulation program.  This
                 period is the number  of months (less than or equal to 12)
                 required for the emissions to return to their value just
                 prior to maintenance.  After this period, if less than
                 12 months, the emissions are deteriorated according to
                 the EFP rates as shown in Table 1-6.  This situation is
                 displayed in Figure 1-3.
                 Program effectiveness was based upon total mass of pollut-
 ants emitted since the inception of the program, reported annually.  For
. each pollutant, the total grams emitted for each vehicle were calculated for
 the average mileage accumulations shown in Table  1-7.  The total mass for
 each fleet is calculated by summing over each vehicle within the respective
 fleets.  The effectiveness is then expressible as a percent change  relative
 to  the uncontrolled fleet.
 1.1.3.1         Simulation  Results
                 Table 1-8 summarizes the simulation results for the  207(b)
 approach for the three-maintenance versions examined, which have a 12-
 month maintenance effectiveness period.  Maintenance version 3 produced
 the most conservative estimates of program efficiency, whereas version
 tended to be the most optimistic.   However, the differences between the
 maintenance versions  were quite small and rarely exceeded 10 percent.
                 Of greater  importance is the assumption as to the mainte-
 nance period of effectiveness.  This is illustrated in Table  1-9 for the 207(b)
 approach with maintenance  version 3.   If the  emissions of the maintained
 vehicles  returned to pre-maintenarice  levels in 6 months instead of 12
 months,  for example, then the effectiveness  of the  program at the  end
 of the 50,000 miles is substantially reduced (from 32.9 percent to 22. 1 per-
 cent for HC; from 38.3 percent to 24. 6 percent for  CO).  This result merely
 quantifies the intuitive knowledge that  "the quicker the erosion of maintenance
 effects, the less the overall impact of maintenance. " The differences  between
                                     1-21

-------
i
ro
ts)
               40-
            §30*
            cfl
            l-l
            o
c
a)
             O
            OH
               20-.
               10-.
                          O  -  Original emission rate

                          &  -  After application of I/M program
                                                             —•	 Rapid deterioration
                                                                   cancelling maintenance
                                                                   in 9 months


                                                             	— EFP deterioration
                                                                Adjustment due
                                                                to Maintenance
          Maintenance
         Effectiveness
            Period
                                                            Maintenance
                                                           Effectiveness
                                                              Period
                                              10    12     14

                                                 Months
                                                   16
18
20
22
24
                       Figure 1-3.   Piece-wise Linear Deterioration for FF Vehicles
                                        Shown Over Two-Year Period

-------
              Table 1-7.  Annual Mileage Accumulations Used in
                        207(b) Effectiveness Simulation
Age of
Vehicles
In Years
1
2
3
4(b)
Number of
Years In
Program
0
1
2
30>)
Accumulated Average
Mileage^3-)
Annual
17, 500
16, 100
13, 200
3, 200
Cumulative
17, 500
33, 600
46, 800
50,000
(a)
(b)
Nationwide Personal Transportation Study; Annual Miles of
Automobile Travel, Report No.  2,  April 1972,  U.S. Depart-
ment of Transportation

Not a full year - simulation stops at 50,000 miles.
                                    1-23

-------
                      Table 1-8.  Variation of 207(b) Program^ Effectiveness with
                                   Various Maintenance Assumptions
Maintenance
(b)
Assumptions
VERSION NO. 1: HC and CO
adjustments coupled, 'c' NOX
adjustment independent
VERSION NO. 2: HC, CO
and NOX adjusted independently
VERSION NO. 3: HC, CO
and NOX adjusted independently
Emission Values
Adjusted to:
FTP
Standards
X
X

Equivalent
ST Standards


X
Number of
Years
Since First
Inspection
1
2
3(d)
1
2
3(d)
1
2
3(d)
Estimated Program Efficiency
in Percent (Cumulative)*0'
HC
22.7
31.6
34.6
22.7
31.6
34.6
21. 1
30.0
32.9
CO
32.4
39.0
42.5
29.7
35.9
39.1
29.0
35.2
38.3
N0x
2.05
1.81
2.00
2.05
1.81
2.00
1.78
1.53
1.70
(a)
  Pass-fail ST levels set for 5 percent Ec rate for each pollutant for each of three engine CID groups
  (150 CID and less; 151 to 259 CID; 260 CID and greater).
(b)
(c)
(d)

(e)
All three versions shown have 12-month maintenance period effectiveness; i.e., _ emissions  return to
pre-maintenance levels in 12 months.

Percent efficiency is the percent reduction in fleet emissions from that value which would have occurred
without the 207(b) program,  over the time period shown.

Not a full year.   Program ended at 50, 000 miles.
When either HC or  CO is adjusted, the other is adjusted with same percentage reduction; neither HC
nor CO permitted to be less  than one-half their respective FTP standard.

-------
      Table 1-9.  207(b) Program Efficiencies with Maintenance Version 3
Maintenance
Period of
Effectiveness
12


9


6


Number of
Years
Since The
1st Inspection
1
2
3(a>
1
2
3(a)
1
2
3(a,
Estimated Program
Efficiencies in Percent
Cumulative
HC
21. 1
30.0
32.9
18.4
25.4
28. 5
13. 8
18.9
22. 1
CO NOX
29.0
35.2
38. 3
23. 6
28.9
32.3
17.0
20. 8
24.6
1.78
1. 53
1.70
1.05
0. 82
1.05
0. 32
0. 08
0.33
(a)
   Not a full year.  Program ended at 50, 000 miles.
                                   . 1-25

-------
the 12-month and 6-month periods of maintenance effectiveness were typically
about 50 percent for HC, 70 percent for CO, and 500 percent for NO .
                                                                  .X.
                Table 1-10 summarizes the simulation results for the I/M
approach for the same three maintenance  versions and the 12-month mainte-
nance period assumed for the 207(b) approach of Table 1-8.  The I/M
approach exhibits the same general behavior as the 207(b) approach, but
shows appreciably lower values of program effectiveness.  This is shown
more clearly in a direct comparison of the two approaches in Table 1-11.
These lower effectiveness values are principally a reflection of the method
of selecting the  ST  cut-points in these two examples and are not intrinsic to
207(b) or I/M, per  se.
                In the case of I/M, a 33. 3 percent total  failure rate was
selected, which resulted in a 2 percent E  rate for each pollutant,  based on
the total vehicle fleet.  In the case of 207(b), a 5 percent E  rate for three
engine size groups  was selected that resulted in a total failure rate of
42.4 percent.  This difference in total  failures and in E   rates is  the princi-
pal reason for the difference in effectiveness values between the two
approaches examined.  If they were compared at the  same E rate, it is
unclear as to which approach would be  more effective.
                In  either case,  for fixed maintenance effectiveness period
and a fixed maintenance version, the efficiency of the program increases as
the duration of the  program increases.  Hence,  the earlier the program can
be instituted, the greater the resulting benefit on air quality and the higher
will be the program efficiency at 50, 000 miles.
                It is emphasized that the  program effectiveness values
shown in Tables 1-8 through 1-11  are the result of the specific assumptions
made,  and are  highly dependent upon the duration of the  maintenance
effectiveness period as well as the estimated effectiveness of the mainte-
nance performed at the time of inspection.  One limitation inherent in the
three maintenance versions examined herein is that an E  vehicle is always
either an FF or E  vehicle the following year.  This  results because of the
deterministic approach employed; this  limitation can perhaps be eliminated
in future analyses through the use of stochastic maintenance models.
                                    1-26

-------
                              Table 1-10.   Variation of I/M Programs    Effectiveness with
                                             Various Maintenance Assumptions
Maintenance
As sumptions
VERSION NO. 1: HC and CO
adjustments coupled;
NOX adjustment independent
VERSION NO. 2: HC, CO
and NOX adjusted
independently
VERSION NO. 3: HC, CO
and NOX adjusted
independently
Emission Values
Adjusted to:
FTP
Standards
X
X

Equivalent
ST Standards


X
Number of
Years
Since First
Inspection
1
2
3(d)
1
2
3(d)
1
2
3(d)
Estimated Program Efficiency
in Percent (Cumulative)
HC
16.4
24.0
27.0
16.4
24.0
27.0
14.3
21.2
23.7
CO
24.3
30.3
33.8
22.2
27.7
30.8
18.4
23.6
26.1
N0x
1.59
1.58
1.78
1.59
1.58
1.78
1.04
0.99
1. 13
I
ro
-j
            'a'A single pass-fail ST level for each pollutant for all cars in fleet, regardless of engine CID.  ST levels
               set to provide a 33 percent total failure rate with a three-constituent test.
            'b'All three versions  shown have 12-month maintenance period effectiveness; i. e. , emissions return to
               pre-maintenance levels in 12 months.
            ^Percent efficiency  is the  percent reduction in  fleet emissions from that value which would have occurred
               without the I/M program, over the time period shown.
            '   Not a full year.  Program ended at 50, 000 miles.
            ^e'When either HC or CO is adjusted, the other is adjusted with same percentage reduction.  Neither HC
               nor CO permitted to be less  than one-half their respective FTP standard.

-------
                           Table 1-11.  Comparison of Effectiveness of 207(b) and
                              I/M Program Approaches; Maintenance Version 3;
                                  12-Month Maintenance Period Effectiveness
Program Type
207(b) Program^b)

I/M Program(c)

Number of
Years
Since First
Inspection
1
2
3(d)
1
2
3(d)
Estimated Program Efficiency in
Percent (Cumulative r
HC
21.1
30.0
32.9
14.3
21.2
23.7
CO
29.0
35.2
38.3
18.4
23.6
26. 1
NO
X
1.78
1.53
1.70
1.04
0.99
1. 13
I
ro
oo
       (a)
       (b)
       (c)
       (d)
Percent efficiency is the percent reduction in fleet emissions from that value which would have
occurred without the 207(b) program, over the time period shown.

Pass-fail ST levels  set for 5 percent E  rate for each pollutant for each of three engine CID groups
(150 CID and less; 151 to 259 CID;  260 CID and greater).  This resulted in a total failure rate of
42 percent.

A single pass-fail level for each pollutant for all cars in fleet, regardless of engine CID.  ST levels
set to provide a 33 percent total failure  rate with a three-constituent test.

Not a full year.  Program ended at 50, 000 miles.

-------
1.2            CONVENTIONAL CORRELATION ANALYSIS
               DISCUSSION
1.2.1          Results for All Three Cities Combined
               A summary of the ST/FTP correlation coefficients obtained
for the pooled fleets of all three cities (300 vehicles) is given in Table 1-12.
1.2.1.1        Short Test Comparison
               The Federal Short Cycle exhibited superior FTP tracking
characteristics (higher correlation coefficients) as compared to the Federal
Three-Mode for all three pollutants (HC,  CO and NO ).
               For the Federal  Three-Mode ST, the idle test mode in neutral
(both automatic and manual transmissions) and drive (automatic transmission
only) was superior to the high-and-low-speed test modes  for tracking HC
and CO.   The high-speed mode  was best for NO  discrimination, although
the low-speed mode coefficient  (0.48) was not significantly different than the
high-speed value (0. 52).,
1.2.1.2        Effect of Accumulated Mileage
               Seventy of the  300 vehicles had accumulated 4000 miles or
less of travel prior to being tested.  As shown in Table 1-13, the only
noticeable effect for the Federal Short Cycle is that these 70 vehicles had
a higher HC correlation and lower  CO correlation than the 230 vehicles  with
greater than 4000 mileage accumulation.  In the case of the Federal Three -
Mode, the results for the greater than 4000 mileage group were generally
similar to those for the total  300-vehicle fleet (Table 1-12).   Those
vehicles with 4000 miles or less,  however, exhibited generally poorer
correlation  coefficients than the fleet average,  except for HC in the idle
in drive test mode which was  higher than the fleet average.   This is in
agreement with the HC  results noted above for  these same low mileage
vehicles when tested by the Federal Short Cycle ST.
                                  1-29

-------
Table 1-12.  Correlation Coefficient Summary— All Three Cities
Short Test
Federal
Short Cycle
Federal
Three-
Mode


Test Mode

High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
. 300
300
300
300
229
Correlation ., .
Coefficients
HC
0.78
0. 31
0..47
0. 45
0. 58
CO
0.89
0.33
0. 46
0. 69
0.67
NO
X
0.81
0. 52
0.48
0. 11*
0. 12*'
(a)

(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the
95 percent confidence level except where indicated by an
asterisk
                              1-30

-------
             Table 1-13.  Correlation Coefficient Summary by
                  Accumulated Mileage— All Three Cities



Short
Test

Federal
Short
Cycle


Federal
Three-
Mode









Mileage

4000 or
less

Greater
than 4000
4000 or
less


Greater
than 4000





Test Mode






High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral



No. of
Vehicles

70


230 .

70
70
48
70
230
230
181
230
ST/FTP(b)
Correlation. .
Coefficients

HC

0.91


0.76

0. 13*
0.29
0. 80
0. 37
0. 33
0.52
0. 54
0.45
CO

0. 80


0.90

0. 16*
0. 31
0.61
0. 55
0. 37
0. 54
0.67
0.71
NO
X
0.79


0. 82

0.36
0.40
-0.08*
0.02*
0. 54
0.49
0.23
0. 14*
(a)
(b)
(c)
Number of accumulated miles on vehicles tested

ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent
confidence level except where indicated by an asterisk
                                    1-31

-------
1.2.1.3       Effect of Engine Displacement
              Table 1-14 displays the three-city correlation coefficients as
a function of engine  cubic inch displacement (CID) class.  In the case of the
Federal Short Cycle, NO  correlations were largely unaffected by CID class
                        X.
and were similar to the total fleet values of Table 1-12.   The CO  correla-
tions of the two larger  CID  groups were similar to the fleet average,
whereas the 150 CID or less group had a significantly lower  CO correlation
than the fleet average (and larger engine size groups).  In terms of HC
discrimination,  the  260 CID or more engine group had a significantly higher
correlation coefficient  than the fleet average; the 150 CID or less group was
slightly less than the fleet average; and the 151 to 259  CID group was signifi-
cantly lower (0. 53) than the fleet average (0.78).
              For the  Federal Three-Mode ST, similar trends are observed
for NO in the high-speed mode  and for CO in the  idle  in drive mode.  HC
correlation coefficients in the  idle in drive mode for the two smaller CID
groups are higher than the 260 CID and above group, and also higher than
the fleet average of Table 1-12.
              Again, the better  test modes of the Federal Three-Mode  (for
a given pollutant) tend to  show the same types of trends with engine displace-
ment as the Federal Short Cycle, except for HC.
1.2.1.4       Effect of Inertia Test Weight Group
              The correlation coefficients of the 300 vehicles (all three
cities) divided into four inertia test weight groups are shown in Table 1-15.
The Federal Short Cycle results indicate that NO  correlations are rela-
                    '                           x
tively insensitive to inertia weight,  whereas  the HC  and CO correlation
coefficients of the 2500 Ib  and less group (subcompacts) are significantly
lower than the heavier weight vehicles.
              For the Federal Three-Mode, the NO correlation coefficients
                                                   Ji.
are similarly relatively insensitive to inertia weight.  The discernible
trends for HC and CO (in idle in  drive and neutral test modes) again indicate
lower coefficient values for the smallest test weight group,  as was the case
for the Federal Short Cycle.
                                   1-32

-------
                 Table  1-14.  Correlation Coefficient Summary by
                      Engine Displacement— All Three Cities


Short Test


Federal
Short Cycle


Federal
Three-Mode













CID(a) Group

.
150 or Less

151 to 259
260 or More
1 50 or Loss




151 to 259



260 or More





Test Mode






High speed
Low speed
Idle in neutral
Idle in drive

High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in .drive


No. of
Vehicles


95

54
15-1
95
95
95
33

54
54
54
45
151
151
151
151
ST/FTP(b)
Correlation. .
Coefficients

HC

0. 71

0. 53
0.89
0.25
0.41
0.45
0.75
•A*
0.25"
0. 36
0. 38
0.78
0. 34
0.62
0.49
0.60

CO

0.63

0.90
0.92
0.40
0. 36
0.63
0.43

0. 34
0.72
0.65
0.66
0.43
0.47
0.82
0.80

NO
X
0. 82

0. 83
0. 79
0.61
0.63
-'f
0.06"
0. 30*

0. 59
0. 74
0.04*
0.13*
0. 58
0. 36
0.24
0. 10*
   CID = cubic inch displacement

'  'ST/FTP = short test/federal emission certification test procedure
(c)
  'Correlation coefficients are statistically significant at the 95 percent confidence
   level except where indicated by an asterisk

-------
                  Table 1-15.  Correlation Coefficient Summary by
                    Inertia Test Weight Group - All Three Cities
Short Test
Federal
Short Cycle



Federal
Three-Mode















Inertia Test
Weight Group
(Ib)
0 to 2500

2501 to 3500
3501 to 4500
450 1 or More

0 to 2500



2501 to 3500



3501 to 4500



450 1 or More



Test Mode






High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
50

119
98
33

50
50
50
10
119
119
119
88
98
98
98
98
33
33
33
33
ST/FTP(a)
Correlation ...
Coefficients
HC
0.49

0.76
0.92
0.73

0.20*
0. 33
0. 32
0. 37*
0.41
0. 55
0.49
0.66
0. 36
0. 51
0.42
0. 53
0.28*
0.48
0. 56
0.64
CO
0. 73

0.86
0.91
0.88

0. 53
0.42
0. 56
0.05*
0. 50
0.61
0. 59
0. 58
0.28
0.45
0.86
0.86
0.29*
0. 35
0. 75
0. 76
NO
X
0. 76

0.81
0.81
0. 77

0. 54
0. 50
0.08*
0. 33*
0.62
0.64
0.09*
0. 18*
0. 54
0. 33
0.20
0.05*
0. 53
0. 30*
0. 30*
0, 10*
(a)

(b)
ST/FTP = short  test/federal emission certification test procedure

Correlation coefficients are statistically  significant at the 95 percent confidence
level except where indicated by an asterisk

-------
1.2.1.5       Effect of Emission Control System Type
              The correlation coefficients of the three-city vehicle fleet are
shown in Table 1-16 as a function of usage of oxidation catalysts and/or
secondary air injection for control of HC and CO. In the case of the Federal
Short Cycle results, the data indicate generally higher HC and CO correla-
tions  with the use of a catalyst (with or without secondary air); however,
the 25 vehicles without either a catalyst or secondary  air had a CO correla-
tion coefficient (0. 86)  as high as the 89 vehicles with both a catalyst and
secondary air injection (0.86).  NO  correlations are  relatively unaffected
                                  3C
by the use or non-use  of catalysts and/or secondary air  injection.
              The Federal Three-Mode  results are less clear.  For  example,
in the idle in drive mode,  the HC correlation coefficients for  those vehicles
without a  catalyst (or  secondary air) (0. 85 and 0. 86) are higher than for
those vehicles with a catalyst (and secondary air) (0.46  and 0. 62).  The
twelve vehicles without a catalyst or secondary air had a higher CO corre-
lation coefficient (0.88) than vehicles  with a catalyst (and secondary air)
(0. 77 and 0. 67).   NO  correlations  (in the high speed and low speed test
modes) were relatively little affected by catalyst and secondary air usage,
as in  the case of the Federal Short Cycle ST.
1.2.1.6       Effect of Carburetion vs Fuel  Injection
              Thirty  vehicles of the 300-vehicle test fleet were equipped
with fuel injection.  As shown in Table 1-17, the fuel injected vehicles had
significantly poorer Federal Short Cycle Correlation Coefficients for all
three pollutants than did the carbureted vehicles.  With  the Federal Three-
Mode ST, this was reversed for NO   and HC; in the high speed test mode
                                   JC
the fuel injected NO  coefficient (0. 63) was higher than the  carbureted
value (0. 53); and in the idle in drive test mode the fuel injected HC coeffi-
cient (0. 80) was higher than the carbureted value (0. 58). When the idle in
neutral test mode is considered, it is seen that the HC comparison is
reversed, with the fuel injected coefficient (0.36) lower  than the carbureted
value (0.46).
                                   1-35

-------
                                Table  1-16.'  Correlation Coefficient Summary by

                                Emission Control System Type - All Three Cities
i
u>
Short Test
Federal
Short Cycle


Federal
Three-Mode














Catalyst/Air^
Yes/Yes
Yes/No
No/Yes
No /No
Yes/Yes



Yes /No



No/Yes



No /No



Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
89
127
59
25
89
89
89
76
127
127
127
117
59
59
59
24
25
25
25
12
ST/FTP*b)
Correlation, ,
Coefficients^0'
HC
0.94
0'. 84
0. 74
0. 32*
0. 73
0. 59
0.25
0.46
0. 13*
0. 50
0. 56
0.62
0. 34
0. 36
0.40
0.85
-0.01*
0. 16*
0.31*
0.86
CO
0.86
0.92
0. 54
0.86
0. 39
0. 54
0.80
0.77
0. 35
0. 51
0.65
0.67
0.40
0. 52
0. 50
0.40*
0. 39*
-0.02*
0.69
0.88
NO
X
0.83
0. 77
0.86
0.90
0.65
0.60
0.05*
-0.05*
0.48
0.47
0.12*
0.22
0.65
0. 74
0.29
0.55
0.28*
0.47
0.4Q
0.51*
              'Secondary air injection system


               ST/FTP = short test/federal emission certification test procedure

            (c)
              'Correlation coefficients are statistically significant at the 95 percent

               confidence level except where indicated  by an asterisk

-------
                                Table 1-17.   Correlation Coefficient Summary by

                                       Fuel System Type - All Three Cities
i
U>
-o


Short
Test
Federal
Short
Cycle
Federal
Three -Mode








Fuel
System Type
Fuel Injection
Carburetion

Fuel Injection



Carburetion





Test
Mode



High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive


No. of
Vehicles
30
270

30
30
30
11
270
270
270
218
ST/FTP(a)
Correlation _ .
Coefficients* }
HC
0.45
0.82

0. 11*
0.22*
0.36
0.80
0.32
0.50
0.46
0.58
CO
0.79
0.89

0.47
0.11*
0.61
0.54*
0.34
0.49
0.70
0.69
NOX
0.62
0.82

0.63
0.33*
-0.20*
-0.07*
0.53
0.49
0.13
0. 11*
               (a)


               (b)
ST/FTP = short test/federal emission certification test procedure


Correlation coefficients are statistically significant at the 95 percent

confidence level except where indicated  by an asterisk

-------
1.2.1.7       Effect of Transmission Type
               Seventy-one of the 300-vehicle test fleet were equipped with
manual transmissions; the remainder were equipped with automatic trans-
missions.  Table 1-18 shows the correlation coefficients as a function of
these two transmissions.  With the Federal Short Cycle ST, the NO   corre-
lation coefficient was  unaffected by transmission type,  whereas both HC and
CO correlation coefficients were significantly reduced for manual transmis-
sion vehicles.
               In the case of the Federal Three-Mode ST,  the situation is
similar for NO .  For HC and CO,  the only comparative basis  is the  idle in
               J\.
neutral test mode, because the manual transmission prevents the idle in
drive test mode.  Here the CO correlation coefficients  are similar for both
transmissions (0.71 and 0. 67), and the HC coefficient for  the manual trans-
mission (0. 33) is considerably lower than the automatic transmission
value (0.49).
1.2.1.8       Effect of Manufacturer
               Table 1-19 shows the variation of correlation coefficients as
a function of vehicle manufacturer.  On the Federal Short Cycle,  General
Motors and AMC vehicles had uniformly high (0. 9 and above) correlation
coefficients for all three pollutants.   Ford had the highest HC and CO coeffi-
cients (0.94 and 0.96,  respectively), but had a lower NO  coefficient (0.77).
Chrysler had CO and NO  coefficients similar to Ford values, but had a lower
HC coefficient  (0.81).  The "others" category (primarily small imported
vehicles) had the lowest HC (0. 51) and CO  (0. 65) correlations, with a NO
                                                                       n
value (0. 85) slightly higher than Ford and Chrysler values.
               On the Federal Three-Mode ST,  the high speed test mode
resulted in  generally similar NO   correlation coefficient levels (0.64 to 0.72)
for all manufacturers except  General Motors, which had a value of 0. 42.
General Motors, Ford  and "others"  vehicles had slightly higher NO  coeffi-
                                                                 Ji
cients in the low speed mode  than in the high speed mode. In the case of
CO in the idle in drive  test mode,  Ford,  General Motors, and AMC relative
                                   1-38

-------
                                Table 1-18.  Correlation Coefficient Summary by
                                     Transmission Types — All Three Cities
i
u>
sO

Short
Test
Federal
Short
Cycle
Federal
Three-Mode







Trans mis s ion
Type
Automatic
Manual

Automatic



Manual





Test Mode



High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

No of
Vehicles
229
71

229
229
229
229
71
71
71

Correlation .,.
Coefficients1 '
HC
0.87
0.42

0.31
0.48
0.49
0.58
0.28
0.42
0.33

CO
0.91
0.66

0.33
0.48
0.71
0.67
0.50
0.56
0.67

NOX
0.80
0.82

0.57
0.49
0. 11*
0. 12*
0.65
0.56
0.09*

                 (a)
                 (b)
                   ST/FTP = short test/federal emission certification test procedure
                    Correlation coefficients are statistically significant at the 95 percent
                    confidence level except where indicated by an-asterisk

-------
       Table 1-19.  Correlation Coefficient Summary by
                Manufacturer - All  Three Cities
Short
Test
Federal
Short
Cycle




Federal
Three-
Mode























Manu-
facturer
General
Motors
Ford
Chrysler
American
Motors .
Others
General
Motors


Ford



Chrysler




American
Motors





Others




Test Mode







High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed

Idle in neutral
Idle in drive
High speed
Low speed

Idle in neutral

Idle in drive

High speed
Low speed
Idle in neutral

Idle in drive
No. of
Vehicles
99

61
36
'18

86
99
99
99
92
61
61
61
55
36
36

36
36
18
18

18

16

86
86
86

30
STP/FTP(a)
Correlation ...
Coefficients1 '
HC
0.90

0.94
0.81
0.94

0. 51
0.35
0.61
0. 50
0. 55
0.33
0.40
0.42
0.63
0.48
0.60

0. 54
0. 56
0.23*
6. 58

0.56

0.81
*
0.21
0. 31
0.33

0.82
CO
0.90

0.96
0.94
0.91

0.65
0.29
0.65
0. 70
0.73
0.43
0.53
0.85
0. 84
0.57'
0.64

0.65
0. 68
0.68'
0.60

0. 84

0. 84

0. 47
0.49
0.53
&
0.32
NO
X
0.92

0. 78
0.77
0.93

0.85
0.42
0. 55
if
0. 17
0.28
0.65 .
0.67
-0. 10*
-0.25*
0.65
0. 58
*
0.09
-0.12*
0.72
0.39*
*
0. 24
*
0.29

0.64
0.69
0.11*

0.60
*a)ST/FTP = short test/federal emission certification test procedure
  'Correlation coefficients are statistically significant at the 95 percent
  confidence level except where indicated by an asterisk
                                 1-40

-------
rankings were similar to those of the Federal Short Cycle; however, Chrysler
CO coefficients were lower (0. 68) than the other domestic manufacturers
(0.73 to 0.84).  The "others" category had a very low (0.32) CO coefficient
in the idle in drive mode.  Conversely,  the HC coefficient of the "others"
category in the idle in drive mode was the highest (0. 82) of all manufacturer
groupings.  AMC also maintained a high HC coefficient (0.81) in this same
test mode.  The other three domestic manufacturers had similar, but lower,
HC coefficients in the 0. 55 to 0. 63 range.
1.2.1.9     •  Effect of Engine Displacement and Manufacturer
               The correlation coefficient summarization is extended to classi-
fication by vehicle  manufacturer and engine displacement in Tables  1-20
through 1-22.  The "others" category of Table 1-19 is expanded in Tables 1-20
and 1-21 to delineate the vehicles manufactured by Datsun,  Toyota,  VW, and
Honda.
1.2.1.9.1      150  CID and Less Engine Displacement
               Using the Federal Short Cycle for  comparison purposes (see
Table 1-20), the smaller displacement vehicles (150 CID and less) of General
Motors, Datsun and Toyota have representatively high (0. 82 to  0. 94) NO
                                                                      n
correlation coefficients; Ford (0.35), Volkswagen (0.49) and Honda (0. 74)
are considerably lower in coefficient values.  The CO coefficients of General
Motors, Ford, and Datsun are representatively high (0.83 to 0.93), whereas
the Toyota (0. 50),  Volkswagen (0. 71) and Honda (0.21) values are consider-
ably lower.  All HC coefficient values are  representatively high (0.84 to
0.95) except for Honda (0.20).
               The Federal Three-Mode correlation coefficient results are
more mixed, as might be  expected from previous effect trends.  In the case
of NO , General Motors vehicles have the highest correlation coefficient
     5C
(0. 84) in the high speed mode; Honda (0. 66) and Volkswagen (0. 77) also
exhibit their higher values in the high speed mode;  Ford (0. 6l), Datsun (0. 72)
and Toyota (0. 81) exhibit their best values in the low speed mode.
                                   1-41

-------
Table 1-20.   Correlation Coefficient  Summary for 150 CID*
          and Less by Manufacturer  - All Three Cities
Short
Test
Federal
Short
Cycle






Federal
Three -
Mode



























Manu-
facturer
General
Motors
Ford
Oatsun
Toyota
VW, Audi,
Porsche
Honda
Others
General
Motors


Ford




Datsun



Toyota



VW, Audi,
Porsche


Honda



Other



Test Mode









High speed
Low speed
Idle in drive
Idle in neutral
High speed

Low s peed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
12

6
8
9
21

6
33
12
12
6
12
6

6
3
6
8
8
2
8
9
9
2
. 9
21
21
5
21
6
6
0
6
33
33
15
33
ST/FTp(b)
Correlation . ,
Coefficients10'
HC
0.84

0.95
0.95
0.88
0.88

0.20*
0.74
0.25*
0.61
0.88
0.30*
-0. 17*

-0.22"
-0.81*
0.86
0. 53.'"
0.61*
...
0. 40 *
0.69
0.87
...
0.66°"
0.20*
0.53
0.77*
0.60
0.89
0.23*
	 	
0.94
0.15*
0.23*
0.84
0.37
CO
0.89

0.83
0.93
0. 50
0.71

-0.21*
0.48
0. 15*
0.70
0.96
0.66
0. 55*
V
0.76
0.93*
0.07"'
0.93.
0.93
--.
0. 42'"
-0.07*
-0. 17*
...
0. 42*
0. 59
0.43*
0.10*
0.53
0.29*
-0.26*
...
0.35*
0.47
0.31*
'0. 44
0.41
NO
•x
0.94

0.35*
0.89
0.82
0.49

0.74*
0.91
0.84
0.32*
0.04"'
0. 19*
0. 17*

0.61"
0.74"
0.31*
-0.49'"
0.72
...
-0. 07''
0. 56*
0.81
...
0.00*
0.77
0.66
0.77*
0.24*
0.66*
0.24*
	
0.62*
0.61
0.68
0.55
0.37
        (a)
        (b)
        (c)
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent
confidence level except where indicated by an asterisk
                                   1-42

-------
              In the case of CO, both General Motors and Ford vehicles had
high correlation coefficients (0.96 and 0.93) in the idle in drive test mode.
The other vehicles in this CID class were predominantly equipped with manual
transmissions and, therefore, only amenable to testing in the idle in neutral
mode.  On this basis,  the imports ranged from 0. 35 to 0. 53; General Motors
had 0. 66, and Ford had a very low coefficient (0. 07).
              HC correlation coefficient results on the Federal Three-Mode
ST are similarly mixed.  In the idle in neutral test mode, Honda  (0.94) and
Ford (0.86) had the highest coefficients,  Toyota (0.66) and Volkswagen
(0.60) ranged in the middle, and Datsun (0.40) and General Motors  (0. 30)
were at the lower end of the scale.
1.2.1.9.2     151  to 259 CID Engine Displacement
              The medium displacement vehicles  (151 to 259 CID)  are segre-
gated by manufacturer in Table 1-21.  On the Federal Short Cycle ST, all
domestic manufacturers have representatively high correlation coefficients
for all three pollutants,  except for Ford on NO (0.44).
                                             X.
              In the case of the  Federal Three-Mode ST, all manufacturers
except AMC had their highest NO correlation coefficients (0. 70  to 0. 94) in
                                3C
the low  speed test mode; AMC recorded 0. 69 in the high speed mode and
0.49 in  the low speed mode.  The differences between high speed and low
speed mode results for NO  were not great except in the case of General
Motors  vehicles which recorded  a NO  coefficient  of 0. 05 in  the high speed
                                    3C
mode versus 0. 81 in the low speed mode.
              General Motors, Chrysler, and the "others" category had
their highest CO correlation coefficients in the idle in drive mode,  while
Ford and AMC peaked in the idle in neutral mode.  However, the differences
in coefficient between these two idle test modes was small for this  size
engine class.
              In the case of HC, General Motors and Ford vehicles had their
highest  correlation coefficients in the idle in neutral mode, whereas Chrysler,
AMC, and  the "others" vehicles  peaked on the idle in drive mode.
                                  1-43

-------
Table 1-21.
        Correlation Coefficient Summary for  151 to 259 CID
               Manufacturer - All Three Cities
                                                                           (a)
by
Short
Test
Federal
Short
Cycle






Federal
Three -
Mode





















Manu-
facturer
General
Motors

Ford
Chrysler
American
Motors
Other

General
Motors



Ford




Chrysler'




American
Motors


Other'



Test Mode









High speed
Low speed
Idle ip drive

Idle in neutral
High speed
Low speed

Idle in drive
Idle in neutral
High speed

Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
13


13
7
12

9

13
13
12

13
13
13

10
13
7

7
7
7
12
12
10
12
9
9
6
9
ST/F'TP(b)
Correlation ..
Coefficients*
HC
0.93


0.92
0.80
0. 88

#
0. 18
*
0. 13
*
0. 50
0.79

0.82
0.79
0.79
ji£
0.21
0.73
0.70*
=>
0.71
0.89
0. 87
0.03*
0.35*
0.65
0.27*
0. 13*
0.07"
0.68*
0.02*
CO
0.98


0.93
0.94
0.89

0.90
#
-0. 40
0. 87
0.60

0.58
0.91
0.91
*
0.55
0.88
0.09*
#
0. 18
0.89
0.85
0.60
0.15*
0. 86
0. 89
0.08*
0. 86
0.93
0.92
NO
X
0.90

...
0 . 44""
0.95
0.92

0.90
j.
0.05""
0. 81
-0. 12*
*
-0.09
0.67
0.70
jV
-0.08
-0.35*
0.74*

0.94
0.09*
0. 18""
0.69
0. 49 "
0.32*
0.26"
0. 72
0. 78
0. 85
0.78
     (a)
     (b)
     (c)
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure
       Correlation coefficients are statistically significant at the 95 percent
       confidence level except where indicated by an asterisk
       "Other" category includes one Datsun and one Porsche
                                      1-44

-------
1.2.1.9.3     260 and Greater CID Engine Displacement
              Table 1-22 displays similar correlation coefficient summaries
for the larger displacement vehicles (260 CID and greater), which were all
domestically produced.  On the Federal Short Cycle ST, all manufacturers
have representatively high correlation- coefficients for all three pollutants,
except for AMC on CO (0. 54).
              Using the Federal Three-Mode ST, the high speed mode was
superior for NO  in the case of Chrysler and AMC vehicles.   Although the
               X.
low speed mode was superior for General Motors and Ford vehicles, the
difference between high and low speed mode NO  correlation coefficients was
                                              3C
not large except in the case of AMC, where the high speed coefficient (0.93)
was nearly three times higher than the low speed value (0. 33).
              The CO correlation coefficients were nearly the same in either
the idle in drive or idle in neutral  test modes  for all manufacturers; however,
the AMC values were much lower than the other manufacturers (0. 36 versus
0. 72 to 0.85).
              The idle in drive mode resulted in the highest HC correlation
coefficients for General Motors, Ford and Chrysler vehicles; however, the
differences  between drive and neutral coefficients was small except for Ford
(0. 37 in neutral versus 0. 64 in drive).
              The AMC vehicles had the lowest HC  and CO correlation coeffi-
cients and the highest NO  coefficients of the domestic manufacturers.  In
                        5C
the larger engine size vehicle group,  General Motors,  Ford and Chrysler
vehicles were generally similar in correlation coefficient level.
1.2.2         Results for Chicago
              The ST/FTP correlation coefficients  for the Chicago 100-vehicle
test fleet are summarized in Table 1-23.
1.2.2.1       Short Test Comparison
              In Chicago the Federal Short Cycle had higher correlation
coefficients than the Federal Three-Mode for all three pollutants, as was
                                  1-45

-------
Table 1-22.   Correlation Coefficient Summary for 260 CID(a' and
            Greater by Manufacturer - All Three Cities
Short
Test
Federal
Short
Cycle


Federal
Three-
Mode













Manu-
facturer
General
Motors
Ford
Chrysler
American
Motors
General
Motors


Ford
•


Chrysler



American
Motors


Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
74
42
29
6
74
74
74
74
42
42
42
42
29
29
29
29
6
/
6
6
6
ST/FTP(b>
Correlation. .
Coefficients'0'
HC
0.91
0.95
0.80
0.86
0.37
0.71
0. 59
0.51
*
0.20
0. 40
0.64
0.37
0.46
0.58
0. 59
0. 54
*
0. 11
*
-0.02
0.36*
0.42*
CO
0.88
0.97
0.94
0. 54"
0.40 '
0.63
0.79
0.84
0. 38
0.67
0.84
0.85
0. 55
0.64
0.74
0.72
0.30*
0.43*
0.3.6*
0.36*
NO
' x
0.92
0.88
0.78
0.95
0. 56
0.68
0. 37
0.42
0.65
0.73
-0. 19*
*
0.01
0.62
0.52
-0. 16*
*
0.06
0.93
0.33*
0.82
. . s'f
0.66
(a)
(b)
CID = cubic inch displacement
  ST/FTP = short test/federal emission certification test procedure

'C'Correlation coefficients are statistically significant at the 95 percent
  level except where indicated by an asterisk
                                                              confidence
                                  1-46

-------
     Table 1-23.  Correlation Coefficient Summary — Chicago
Short Test
Federal
Short Cycle
Federal
Three-
Mode


Test Mode

High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
100
100
100
100
78
ST/FTP(a)
Correlation ...
Coefficients^ '
HC
0. 88
0.12*
0.38
.0.42
0. 52
CO
0.82
0.31
0. 52
0. 52
0. 50
NO
X
0.81
0.43
0.40
A
0.03
0.01*
(a)
(b)
ST/FTP = short test/federal emission certification test
procedure

Correlation coefficients are statistically significant at the
95 percent confidence level except where indicated by an
asterisk
                              1-47

-------
the case for the entire 300-vehicle fleet (Table 1-12).  The HC coefficient
was higher than the pooled fleet (0. 88 versus 0. 78), the CO coefficient was
lower (0. 82 versus 0. 89), and the NO coefficient was identical (0. 81).
                                    J\.
              With the Federal Three-Mode ST,.  the same superior test
modes were identified for Chicago as for the 300-vehicle fleet: high speed
for NO  t and idle in drive or neutral for HC and CO.  Although the Federal
Short Cycle HC  coefficient was higher than the pooled fleet average, the idle
test mode (drive or neutral) indicated that both HC and CO were lower than
the fleet average. In the case of NO , the high speed mode indicated lower
than fleet-average coefficient values, whereas the Federal Short Cycle ST
predicted the same value as the fleet average.
1.2.2.2       Effect of Engine Displacement
              As shown in Table  1-24,  on the Federal Short Cycle ST the
150 and less CID group displayed  a higher NO   correlation coefficient and
a lower CO coefficient than the other two. engine size groups.  The 260 and
above CID group displayed the highest HC correlation.
              For the Federal Three-Mode ST, the NO  coefficient of the
150 CID or less  group  is lower than that of the larger engine  size groups,
as opposed to the opposite  findings of the Federal Short Cycle ST.,   On the
idle test mode (neutral or drive),  both the HC and CO coefficients of the
150 CID or less  group  are  lower than the larger size-engine groups.
1.2.2.3       Effect of Inertia Test Weight Group
              On the Federal Short Cycle ST, as shown in Table 1-25, the
smallest (2500 Ib and less) and largest (4500 and greater) inertia test weight
groups have the  highest NO  correlation coefficients  and the lower HC and
                          j£
CO coefficients.
              The Federal Three-Mode shows the same trends as mentioned
above for HC and CO in the idle test mode (neutral or drive), but indicates
lower rather than higher NO correlation coefficients for the 2500 Ib and less
                           3C
inertia test weight group.
                                  1-48

-------
               Table 1-24.   Correlation Coefficient Summary by
                       Engine Displacement — Chicago
1
Short
Test
Federal
Short
Cycle




Federal
Three-
Mode



















CID(a)
Group
150 or
less

151 to
259
260 or
more
150 or
less






151 to
259






260 or
more




Test Mode







High speed

Low speed

Idle in neutral

Idle in drive

High speed

'Low speed

Idle in neutral

Idle in drive

High speed

Low speed

Idle in neutral
Idle in drive
No. of
Vehicles
32


19

49

32

32

32

11

19

19

19

18

49

49

49
49
ST/FTP(b)
Correlation, .
Coefficients10'
HC
0.84


0.82

0.91

0.43

0. 59
*
0.31
*
0. 55
*
0.25
j.
0.36""

0.64

0.71
•A-
0.06"

0. 43

0. 55
0.61
CO
0.77


0.86

0.84

0. 51

0. 44

0. 58
*
0. 52
*
0.32

0.67

0.74

0. 71

0. 35

0. 52

0,76
0.80
NO
X
0.90


0.79

0.76

0. 41

0. 45
*
0. 12
*
-0.03
*
0.31

0.66
*
0.02
*
0.20

0. 57

0. 33
*
0.09
*
-0.01
(a)
(b)
CID = cubic inch displacement
   ST/FTP = short test/federal emission certification test procedure
   Correlation coefficients are statistically significant at the 95 percent
   confidence  level except where indicated by an asterisk
                                   1-49

-------
                                Table  1-25.  Correlation Coefficient Summary by

                                      Inertia Test Weight Group - Chicago
i
ui
o
Short Test
Federal
Short Cycle


Federal
Three-Mode















Inertia Test
Weight Group
db)
0 to 2500
2501 to 3500
350 1 to 4500
450 1 and mor

0 to 2500



2501 to 3500



3501 to 4500



450 1 and mor<



Test Mode



^

High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
3 High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
15
43
30
12

15
13
15
4
43
43
43
32
30
30
30
30
12
t>
12
12
ST/FTP*a)
Correlation ,, .
Coefficients^ '
HC
0.80
0.87
0.93
0.77

0.27*
0. 51
0. 14*
-0.77*
0.46
0. 52
0.61
0.79
0.01*
0. 38
0. 54
0. 59
0.62
0. 54*
0. 19*
0.43*
CO
0.64
0.83
0.88
0.67

0. 17*
0. 32*
0.45*
0.28*
0. 34
0. 55
0.63
0.66
0.28*
0.66
0.82
0.83
0.47*
0.42*
0. 50*
0.64
NO
X
0.91
0. 78
0. 77
0.90

0.26*
0.40*
-0.01*
-0.46*
0. 50
•0.40
0.05*
0. 33*
0.61
0.40
-0. 17*
-0.22
0. 58
0. 50*
0.60
0= 56*
              (a'sT/FTP = short test/federal emission certification test procedure


              '  'Correlation coefficients are statistically significant at the 95 percent confidence

                level except where indicated by an asterisk

-------
 1.2.2.4       Effect of Emission Control System Type
               As shown in Table 1-26 the use or non-use of oxidation
 catalysts and/or secondary air injection for control of HC and CO does not
 produce readily discernible trends with the Federal Short Cycle ST,  except
 that the lowest NO  correlation coefficients are achieved when no secondary
                  X
 air injection is present.
               When Federal  Three-Mode ST results are compared,  little
 difference in NO  discrimination is noted,  except for the catalyst without
 secondary air case, which has the lowest NO  correlation coefficient (when
                                           j£.
 comparing results from either the low or high speed test modes).  However,
 the presence of secondary air does seem to result in higher NO  coefficients
                                                             2C
 in the high and low speed test modes.  In the case of HC and CO, again little
 substantive trend information is present, except that the  catalyst with second-
 ary air configuration has the  lowest HC coefficient (in either idle in drive or
 idle in neutral test modes).
 1.2.2.5      Effect of Carburetion vs Fuel Injection
              In the Chicago fleet,  as shown in Table 1-27, there is little
 substantive difference between fuel injection and carburetion for the correla-
 tion coefficients of all three pollutants on the Federal Short Cycle ST.  On
 the Federal Three-Mode ST,  however,  fuel injected vehicles  achieved sig-
nificantly higher HC and CO coefficients in the idle test mode.
 1.2.2.6      Effect of Transmission Type
              The 22  vehicles equipped with manual transmissions and the
 78 with automatic transmissions in the Chicago  fleet had  essentially equiva-
lent HC, CO, and NO  correlation when tested on  the Federal Short Cycle ST
                    3c
(see Table 1-28),  although the manual transmission class had the highest
NO  coefficient and the lowest HC coefficient.  These same general trends
   5C
were observed on the  Federal Three-Mode ST,  using either the high or low
 speed mode for NO  and the idle in neutral mode for HC and CO comparative
purposes.
                                  1-51

-------
                                Table 1-26.  Correlation Coefficient Summary by
                                    Emission Control System Type - Chicago
i
Ul
ro
Short Test
Federal
Short Cycle


Federal
Three-Mode














Catalyst/Air(a)
Yes/Yes
Yes/No
No/Yes
No/No
Yes/Yes



Yes /No



No/Yes



No /No



Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
30
44
17
9
30
30
30
27
44
44
44
42
17
17
17
5
9
9
9
4
ST/FTP(b)
Correlation, .
Coefficients^0'
HC
0.91
0.86
0.91
0. 81
0. 30*
0. 14*
0.23*
0.47
0.03*
0. 52
0.46
0. 59
0. 52
0.68
0.75
0.88
-0. 10*
0. 36*
0. 75
0.88*
CO
0.76
0.85
0.92
*6.82
0.49*
0. 51
0.64
0. 58
0. 32
0. 56
0.48
0.60
0.48*
0. 72
0. 54
0. 38*
0. 36*
-0.07*
0.85
0.82*
NO
X
0.80
0.77
0.94
0.70
0.60
0.69
-0.05*
-0. 11*
0.44
0. 36
0. 10*
0.20*
0. 56
0.69
0. L8*
0.25*
0.04*
0.68
-0.04*
-0.74*
              (a)
              (b)
              (c)
Secondary air injection system
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent
confidence level except where indicated  by an asterisk

-------
                                Table 1-27.  Correlation Coefficient Summary by
                                           Fuel System Type — Chicago
UI
OJ


Short Test


Federal
Short Cycle

Federal
Three-Mode








Fuel
Injection



Yes
No
Yes



No





Test Mode





High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive


No. of
Vehicles



10
90
10
10
10
4
90
90
90
74
ST/FTP
a)
Correlation ,, .
Coefficients^ '

HC


0. 79
0.88
-0.07*
0.03*
0.9-1
0.95*
0. 11*
0. 39
0.40
0. 51

CO


0.82
0.82
0.01*
0. 31*
0.83
0.90*
0. 32
0. 53
0. 54
0. 53

NO
X

0. 75
0.81
0. 52*
0.15*
-0.64
-0.99
0.44
0.40
0.05*
0.02*
              (a)

              (b)
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent
confidence level except where indicated  by an asterisk

-------
                   Table 1-28.  Correlation Coefficient Summary by

                            Transmission Type — Chicago


Short Test


Federal
Short Cycle

Federal
Three-Mode








Transmission
Type


Automatic

Manual
Automatic



Manual





Test Mode





High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive


No. of
Vehicles


78

22
78
78
78
78
22
22
22

ST/FTP
a)
Correlation ,,.
Coefficients

HC

0.89

0.82
0.09*
0. 35
0. 50
0. 52 -
0. 57
0.73
0. 31



CO

0.82

0.84
0. 30
0. 53
0. 55
0. 50
0.61
0. 69
0.48



NO
X
0.80

0. 89
0.47
0. 38
0.01*
0.01*
0. 57
0. 54
0. 18*
	
(a)
  'ST/FTP = short test/federal emission certification test procedure


  'Correlation coefficients are statistically significant at the 95 percent confidence

  level except where indicated by an asterisk

-------
 1.2.2.7       Effect of Manufacturer
               Table 1-29 shows the variation of correlation coefficients as
a function of vehicle manufacturer for the Chicago fleet.  On the Federal
Short Cycle ST, the principal discernible trend is the lower HC and NO
                                                                     2t
coefficients achieved by Chrysler vehicles.   When tested on the Federal
Three-Mode ST, the Chrysler vehicles improved in NO  discrimination on
                                                     .X
a relative basis (when comparing readings of either the high or low speed
test modes).  The idle test mode (in neutral or drive) indicated that Chrysler
vehicles had the lowest HC and CO coefficients, with AMC vehicles having
very high HC and CO coefficients (the  same level of coefficient obtained by
AMC vehicles on the Federal Short Cycle ST).
1.2.3          Results for Houston
               The ST/FTP correlation coefficients for the Houston 100-
vehicle test fleet are summarized in Table 1-30.
1.2.3.1        Short Test Comparison
               In Houston the Federal Short Cycle ST had higher correlation
coefficients than the Federal Three-Mode ST for all three pollutants,  as
was the case for the entire 300-vehicle fleet (Table 1-12).  The HC coeffi-
cient was higher than the pooled fleet (0.91 versus 0.78); the CO coefficient
was also higher (0.97  versus 0.89); and the NO   coefficient was slightly
                                             .X
lower (0.78 versus 0.81).
               With the Federal Three-Mode ST, the same superior NO
test mode (high speed) was identified as for the  300-vehicle fleet and Chicago
vehicles.   The idle mode (neutral or drive) was  also the superior CO test
mode, as in the case of the pooled fleet and Chicago.  However,  in the case
of HC, both the low and high speed test modes were superior to the idle test
mode, with the low speed mode having the highest HC  coefficient (0.76);
this singular result was not obtained in any other city  but Houston.  For the
pooled 300-vehicle fleet the low speed mode had a slightly higher coefficient
(0.47) than the idle in neutral coefficient (0.45) but was lower than the  idle
in drive coefficient (0. 58).
                                  1-55

-------
        Table 1-29.  Correlation Coefficient Summary by
                      Manufacturer - Chicago
Short
Test
Federal
Short
Cycle





Federal
Three-
Mode
























Manu-
facture'r
General
Motors

Ford
Chrysler
American
Motors
Others
General
Motors


Ford







Chrysler





American
Motors


Others



Test Mode








High speed
Low s~peed
Idle in neutral
Idle in drive
High speed

Low speed

Idle in neutral

Idle in drive

High speed

Low speed
Idle in neutral

Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
33


20
12
6

29
33
33
33
31
20

20

20

19

12

12
12

12
6
6
6
6
29
29
29
10
ST/FTP(a)
Correlation ...
Coefficients1"'
HC •
0.91


0.96
0.74
0.94

0.80
0.04*
0.58
0.51
0.61
0.38*
*
0.02
1(1
0.38

0.61
*
-0.29
*'«
-0.37'"
0.33*
Og
0.30"
0. 50*
0. 54*
0.90
0.92
0.34*
0.57
0.40
0. 70
CO
0.88


0.75
0.88
0.93

0.86
0.34*
0.60
0.57
0.71
0.52

0.80

0. 70

0. 84
*
0.35
*
-0. 10
0.3.9*
jV
0. 47
0.64*
0.63*
0. 89
0.84
0. 19*
0.77
0.56
0.33*
NO
X
0.95


0.84
*
0.43
0.96

0.89
0.51
0.50
0. 13*
0.37
0.64

0.75
o.
-0. 38
5|<
-0. 15
#
0.23
.1.
0. 54""
-0. 3l'"
*
-0.43
0.96
0.06*
-0.38"
0.26*
0.42
0.61
*
0.12
0.08*
(a)
(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent
program confidence level except where indicated by an  asterisk
                                1-56

-------
    Table 1-30.  Correlation Coefficient Summary — Houston
Short Test
Federal
Short Cycle
Federal
Three-
Mode


Test Mode

High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
100
100
100
100
75
ST/FTP*a)
Correlation ...
Coefficients* '
HC
0.91
0.66
0.76
0.47
0.60
CO
0.97
0.45
0. 53
0.79
0.76
NO
X
0.78
0. 44
0.41
0. 34
0.36
(a)
(b)
ST/FTP = short test/federal emission certification test
procedure

Correlation coefficients are statistically significant at the
95 percent confidence level except where indicated by an
asterisk
                             1-57

-------
1.2.3.2       Effect of Engine Displacement
              On the basis of the Federal Short Cycle ST (see Table 1-31)
the only discernible trends for the Houston vehicles are the lower CO coeffi-
cient for the 150 CID or less group and the lower NO  coefficient for the
                                                  3C
151 to 159 CID group. When the high and low speed test modes of the Federal
Three-Mode ST are compared,  the NO  coefficient of the  151  to 159 CID
                                     X.
group is  essentially the same as the  260 or more CID group, but is lower
than the 150 or less CID group.  The three-mode results for HC arid  CO are
of such variability that trends with respect to engine size  are  not readily
apparent.
1.2.3.3       Effect of Inertia Test  Weight Group
              Examination of the Houston vehicles by inertia  test weight
grouping (see Table 1-32)  on the Federal Short Cycle  ST does  not reveal
significant trends,  other than the relatively poor NO  correlation coefficient
                                                  .X
of the 4500 Ib and above class.  This same characteristic is shown by the
Federal Three-Mode ST results in all test modes.  The three-mode HC  and
CO results also do not indicate significant trends with inertia  weight.
1.2.3.4       Effect of Emission Control  System Type
              As indicated in Table  1-33, the use or  non-use  of csitalysts
and/or secondary air injection for control of HC and CO in the Houston fleet
does not produce readily discernible  trends with the Federal Short Cycle ST.
              On the Federal Three-Mode ST,  however,  the presence of
secondary air does seem to result in higher NO correlation coefficients in
                                              Ji
the high and low speed test modes (which was also noted above for the
Chicago fleet).  Also,  the use of a catalyst plus secondary air seerns to
result in high HC correlation in both  high and low speed test modes and low
HC correlations in the idle test mode.  This latter trend was not seen in the
Chicago data, but is present in the pooled  fleet data of Table 1-16.
                                  1-58

-------
             Table 1-31.  Correlation Coefficient Summary by
                     Engine Displacement — Houston


Short
Test

Federal
Short
Cycle





Federal
Three-
Mode













CID(a)
Group

150 or
less

151 to
259
260 or
m ore

150 or
less



151 to
259


260 or
more




Test Mode









High speed
Low speed
Idle in neutral

Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive


No. of
Vehicles

33


16

51


33
33
33

11
16
16
16
13
51
51
51
51
ST/FTP(b)
Correlation. .
Coefficients^0'
HC

0.90


0.94

0.94

*
0.29
0. 53
0.61

0.92
-0. 14*
0.86
0. 82
0.90
0.75
0.83
0.40
0. 57
CO

0.82


0.96

0.98

*
0.33
0.38
0.79
*
0. 45
•j*
-0.31"
0.95
0.68
0.75
0.65
0.59
0. 84
0.76
NO
X
0.78


0. 56

0.79


0.61
0.66
0.42

0. 85
*
0. 32
0. 56
*
0. 34
0. 57
0. 54
0. 31
0. 35
0.31
   CID = cubic inch displacement
   ST/FTP = short/test  federal emission certification test procedure
'c'Correlation coefficients are statistically significant at the 95 percent
   confidence  level except where  indicated by an asterisk
                                  1-59

-------
            Table 1-32.   Correlation Coefficient Summary by
                   Inertia Test Weight Group - Houston
Short
Test
Federal
Short
Cycle




Federal
Three-
Mode












Inertia
Test
Weight
Group
(lb)
0 to 2500

2501 to
3500
3501 to
4500
4501 or
more
0 to 2500


2501 to
3500


3501 to
4500


4501 or
more


Test Mode







High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
19

38
33

10

19
19
19-
3
38
38
38
29
33
33
33
33
10
10
10
10
STP/FTP(a)
Correlation ...
Coefficients
HC
0.94

0.88
0.94

0.93

0.47
0. 81
0.82
1.00
0.66
0.75
0. 52
0.70
0.87
0.85
0.25*
0.43
0.49*
0.66
0.62*
0.75
CO
0.87

0.95
0. 98

0.98

0. 58
0. 57
0.67
-0.90*
0. 70
0.64
0.68
0.62
0. 42
0. 44
0.93
0.91
0.69
0.55*
0. 85
0.85
NO
X
0.89

0.73
0. 81

*
0.53

0.65
*
0. 44
0. 54
0.99*
V0.57
0.68
0.31*
0.33*
0. 45
0. 12*
0.41
0. 37
0.21*
0. 14*
0.33*
0.27*
(a)

(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent
confidence level except where indicated by an asterisk
                                 1-60

-------
            Table 1-33.   Correlation Coefficient Summary by
                Emission Control System Type — Houston
Short Test
Federal
Short Cycle


Federal
Three-Mode














Catalyst/
Yes/Yes
Yes/No
No/Yes
No/No
Yes/Yes


Yes/No



No/Yes



No/No



Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low spee
-------
1,2.3.5       Effect of Carburetion vs Fuel Injection
              In the Houston Fleet,  as shown in Table 1 -34,  the fuel
injected vehicles had substantially lower CO and NO  correlation coefficients
                                                  Ji
when tested on the Federal Short Cycle ST.  This was not evident in the
Chicago fleet, but was  reflected in the results for the pooled  fleet (Table 1-17).
On the Federal Three-Mode ST, the fuel injected vehicles achieved signifi-
cantly higher HC coefficients in the idle test mode, which was also  observed
in the Chicago fleet (Table 1-27).
1.2.3.6       Effect of Transmission Type
              The vehicles in the Houston fleet with manual and automatic
transmissions had essentially equivalent HC, CO and NO  correlations when
                                                       2C
tested in the Federal Short Cycle ST (see Table 1-35).  When tested on the
Federal Three-Mode ST,  they again had similar levels of NO coefficients
in the high and low speed test modes and similar levels of CO coefficients
in the idle in neutral mode.
1.2.3.7       Effect of Manufacturer
               Table 1-36 shows the variation of correlation coefficients as
a function of vehicle manufacturer for the Houston fleet.  On the Federal
Short Cycle ST, the principal discernible trends are the lower NO  correla-
tion coefficients achieved by the Chrysler and Ford  vehicles (Chrysler was
similarly low in the Chicago fleet).  On the Federal Three-Mode ST,  in
either high or low speed test modes,  the Chrysler and Ford vehicles  improved
in NO  discrimination relative to the other vehicles.  In the idle test  mode,
     J\.
there is little discernible difference an CO discrimination capability.  In
terms of HC coefficient in the idle mode,  however,  the AMC vehicles were
superior and General Motors vehicles had the lowest values.
1.2.4         Results for Phoenix
               The ST/FTP correlation coefficients for the Phoenix 100-
vehicle test fleet are summarized in Table 1-37.
                                   1-62

-------
            Table  1-34.  Correlation Coefficient Summary by
                      Fuel System Type — Houston
Short
Test
Federal
Short
Cycle
Federal
Three-
Mode





-




Fuel
System
Type
Fuel
injection
Carburetion
Fuel
injection




Carburetion





Test Mode


High speed
Low speed
Idle in
neutral
Idle in
drive
High speed
Low speed
Idle in
neutral
Idle in
drive
No. of
Vehicles
10
90
10
10
10

3

90
90
90

72

ST/FTP*a)
Correlation.
Coefficients^
HC
0.97
0.91
*
0.51
0.82
0.86

*
0.99

0.67
0.77
0. 45

0.60

CO
0.76
0.97
0.77
*!..
o. o'r
0.77

•f
0.47

0. 45
0.61
0.81

0.76

NO
X
«i*
0.53"
0.78
•V
0.48"
*,!*
-0.08"
.A.
0. 14"

/ *
0.57

0. 46
0.44
0.35

0.36

(a)
(b)
ST/FTP = short test/federal emission certification test procedure

Correlation Coefficients are statistically significant at the 95 percent
confidence level except where indicated by an asterisk
                                  1-63

-------
          Table  1-35.  Correlation Coefficient Summary by
                    Transmission Type - Houston
Short
Test
Federal
Short
Cycle
Federal
Three-
Mode






Trans-
mission
Type
Automatic
Manual
Automatic


Manual
\


Test Mode

High speed
Low speed
Idle in
neutral
Idle in
drive
High speed
Low speed
Idle in
neutral
Idle in"
drive
No. of
Vehicles
75
25
75
75
75
75
25
25
25

ST/FTP
Correlatic
Coefficien
HC
0.91
0.93
0.69
0.79
0.44
0.60
0. 40
0. 72
0.73
--
CO
0.97
0.91
0.47
0.60
. 0.84
0.76
0.24*
#
0. 31
0. 80

a)
;>
NO
X
0. 78
0.72
0. 55
0.45
0.34
0.36
0.49
0.46
0.48
--
(a)
(b)
ST/FTP = short test/federal emission certification procedure

Correlation coefficients are statistically significant at the 95 percent
confidence level except where indicated by an asterisk
                               1-64

-------
            Table 1-36.
Correlation Coefficient Summary by
Manufacturer — Houston
Short
Test
Federal
Short Cycle



Federal
Three-Mode


















Manufacturer
General Motors
Ford
Chrysler
American Motors
Others
General Motors



Ford



Chrysler



American Motors



Others



Test
Mx>de





High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. Oi
Vehicles
33
21
12
6
28
33
33
33
30
21
21
21
19
12
12
12
12
6
6
6
5
28
28
28
9
ST/FTP(a)
Correlation ,
Coefficients
HC
0.94
0.93
0.92
0.98
0.91
0.85
0.88
0.42
0.48
0.02*
0.72
0.39*
0.68
0.66
0.79
0.62
0.69
-0.68*
0.77*
0.82
0.86*
0.33*
0. 54
0. 59
0.92
CO
0.97
0.98
0.98
0.90
0.79
0.52
0.81
0.78
0.73
0.70
0.88
0.85
0.85
0.73
0.75
0.78
0.78
0.82
0.79*
0.77*
0.69*
0.58
0.32*
0.60
0.09*
NOX
0.90
0.66
0.57*
0.97
0.90
0.20*
0.51
0.50
0.51
0.66
0.69
0. 12*
-0.08*
0.34*
0. 53*
0.58
0. 59
0.55*
0.06*
0. 18*
0.41 *
0.71
0.73
0.48
0.92
ST/FTP = short test/federal emission certification test procedure

Correlation Coefficients are statistically significant at  the 95 percent
confidence level except where indicated by an asterisk
                                    1-65

-------
    Table 1-37.  Correlation Coefficient Summary — Phoenix
Short Test
Federal
Short Cycle
Federal
Three-
Mode


Test Mode
\
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
100
100
100
100
76
ST/FTP(a)
Correlation ,,.
Coefficients'
HC
0.45
0.25
0.28
0.44
0.71
CO
0.76
0.32
0.41
0.80
0.81
NO
X
0.85
0.73
0.66
0.03"
-0.05*
(a)
(b)
ST/FTP = short test/federal emission certification test
procedure

Correlation coefficients are statistically significant at the
95 percent confidence level except where indicated by an
asterisk
                              1-66

-------
1.2.4.1       Short Test Comparison
              The Federal Short Cycle ST resulted in a NO  correlation for
the Phoenix fleet that was higher than any other city and the pooled three-city
fleet (see Table 1-12),  0.85 versus the fleet average of 0.81.   Similarly, the
NO  coefficients for the high and low speed test modes of the Federal Three -
Mode ST were higher than any other city  or the pooled fleet average.  The
high speed mode value (0. 73) was not too far below the fleet average for NO
                                                                         3C
on the Federal Short Cycle  ST  (0.81).
              The HC coefficient on the  Federal Short Cycle was the lowest
observed in any city (0.45 versus 0.88 in Chicago  and 0.91 in Houston).  Con-
versely, the idle in drive HC coefficient  of the Federal  Three-Mode ST was
the highest observed in any city (0. 71  versus 0. 52  in Chicago and 0. 60 in
Houston).
              In terms of CO  correlation, the Phoenix  Federal Short Cycle
ST  value was the lowest observed (0. 76 versus 0. 82 in Chicago and 0. 97 in
Houston).  Again, the idle in drive CO coefficient of the Federal Three-
Mode ST (0.81)  was higher  than the Federal Short  Cycle value (0.76) and
higher than any idle in drive value in the  other cities (0. 50 in Chicago and
0. 76 in Houston).
              In Phoenix,  the  high  speed mode and idle in  drive mode were
clearly superior for NO and HC/CO discrimination, respectively, for the
Federal Three-Mode ST.
1.2.4.2       Effect of Engine Displacement
              For the Federal Short Cycle ST (see Table 1-38),  it is seen
that the two smaller size engine groups (150 or less CID and 151  to 259 CID)
were responsible for the low overall HC correlation coefficient observed in
Phoenix.  In the case of CO, the 150 or less CID had a very low (0.36)
coefficient.
              Using the Federal Three-Mode ST,  the NO  coefficients on
                                                       J£
the high and low speed test modes are largely insensitive to engine size,
                                   1-67

-------
                                Table 1-38.  Correlation Coefficient Summary by
                                        Engine Displacement -  Phoenix
oo
Short
Test
Federal
Short
Cycle

Federal
Three -Mode










CID(a) Group
150 or less
151 to 259
260 or more
150 or less



151 to 259



260 or more



Test
Mode


,
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
30
19
51
30
30
30
11
19
19
19
14
51
51
51
51
ST/FTP(b)
Correlation. .
Coefficients*0'
HC
0.42
0.33*
0.70
0.23*
0.29*
0.34*
0.65
0.36*
0.27*
0. 13*
0.80
0. 16*
0.28
0.71
0.76
CO
0.36
0.93
0.82
0.52
0.18*
0.74
0.46*
0.60
0.62
0.76
0.86
0.31
0.43
0.87
0.87
NOX
0.80
0.95
0.83
0.85
0.82
-0.18*
-0.21*
0.81
0.87
0.14*
-0. 14*
0.75
0.49
0.22*
' -0.02*
                (a)
                (b)
                (c)
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent-
confidence level except where indicated by an asterisk

-------
although there is some dropoff in coefficient (to 0.49) for the 260 and higher
CID group in the low speed mode.
               The idle in drive mode does not indicate the severe reduction
in HC coefficient observed with the Federal Short Cycle ST for the two
smaller engine size groups; however, the idle in neutral mode does indicate
this effect.
               The idle in drive mode, on the other hand,  does indicate a
reduction in CO coefficient for the 150 or less CID group,  whereas the idle
in neutral  mode does not. Thus, the idle in drive mode follows  the CO trends
of the Federal Short Cycle and the idle in neutral mode follows the HC trends
of the Federal Short Cycle.
1.2.4.3       Effect of Inertia Test Weight Group
               As shown in Table 1 -39,  the low HC  coefficient for the Phoenix
fleet in the Federal Short Cycle ST is seen to be associated with both small
(2500 Ib and less) and large  (4500 Ib and higher) vehicles (0. 21 and 0. 28
respectively).   This same effect is also  shown for  the CO  coefficient.  Also,
the NO  coefficient for the smallest vehicle size group is considerably below
the other inertia weight groups.
               Using the  Federal Three-Mode ST,  the small size vehicles
(2500 Ib and less) actually display  the highest NO   correlations of the four
                                               X.
vehicle weight groups in  the high and low speed test modes,  in opposition to
the findings of the Federal Short Cycle ST.  The HC and CO trends are less
clear, although the idle  in neutral  HC results tend  to follow the Federal Short
Cycle results (2500 and 4500 Ib groups have lower  HC coefficients) and the
idle in drive CO results also follow the-Federal Short Cycle results (the
2500 Ib group has a lower CO coefficient).
1.2.4.4       Effect of Emission Control System Type
              As indicated in Table 1 -40,  the use or non-use of catalysts
and/or secondary air injection for control  of HC and CO in the Phoenix fleet
provides no meaningful trend for NO  discrimination based on the  Federal
                                   1-69

-------
              Table  1-39.  Correlation Coefficient Summary by
                    Inertia Test Weight Group -  Phoenix
Short
Test
Federal
Short
Cycle

Federal
Three-Mode







Inertia Test
Weight Group
(Ib)
0 to 2500
2501 to 3500
3501 to 4500
4501 or more
0 to 2500

2501 to 3500

3501 to 4500

4501 or more

Test
Mode



High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
16
38
35
11
16
16
16
3
38
38
38
27
35
35
35
35
11
1 1
11
11
ST/FTP*a*
Correlation ,,.
Coefficients
HC
0.21*
0.52
0.86
0.28*
0. 15*
0. 13*
0. 15*
0.99*
0.56
0.64
0.49
0.71
0.29 *
0.29*
0.82
0.82
-0.03*
0.08*
0.38*
0. 52*
CO
0.57
0.71
0.83
0.68
0.68
0.17*
0.80
0.55*
0. 56
0.70
0.80
0.82
0.41
0.43
0.82
0.84
0.01*
0.09*
0.76
0.81
NOX
0.44*
0.91
0.86
0.80
0.85
0.89
0.02*
0.80*
0.80
0.79
0.03*
-0.25*
0.72
0.53
0.35
0.20*
0.76
0.36*
0.04*
-0.24*
  ST/FTP - short test/federal emission certification test procedure

*  'Correlation coefficients are statistically significant at the 95 percent
  confidence level except where indicated by an asterisk
(c)
   Only two data points used
                                    1-70

-------
               Table 1-40.  Correlation Coefficient Summary by
                   Emission Control System Type - Phoenix
Short
Test
Federal
Short
Cycle


Federal
Three-Mode














Catalyst/ Airla'
Yes/Yes
.Yes /No
No/Yes
No /No
Yes/Yes



Yes/No



No/Yes



No /No



Test
Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low spe.ed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
32
42
20
6
32
32
32
25
42
42
42
37
20
20
20
10
6
6
•6
4
ST/FTP*b)
Correlation, .
Coefficients10'
HC
0.92
0.62
0.51
-0.70*
0.54
0.61
0.73
0.87
0.20*
0.25*
0.63
0.66
0.36*
0.34*
0. 13*
0.55*
-0.29*
-0.56*
-0.92
0. 54*
CO
0.93
0.75
0.22*
0.83
0.47
0.54
0.96
0.96
0.31
0.44
0.73
0.73
0.38*
0.38*
0.45
0.57*
-0.21*
0 . 07 *
0.58*
0.38*
NOX
0.86
0.82
0.92
0.94
0.79
0.71
0. 17*
-0.21*
0.68
0.62
-0. 13*
rO. 17*
0.79
0.86
0.05*
0.41*
0.70*
0.86
0.48*
0.41*
(b)
Secondary air injection system
ST/FTP = short test/federal emission certification test procedure
^Correlation coefficients are statistically significant at the 95 percent confidence
  level except where indicated by an asterisk

-------
Short Cycle ST.  In the case of HC and CO,  it only appears that a catalyst
plus air injection improves both HC and CO correlation coefficients.
              With the Federal Three-Mode ST, observations similar to
those on the Federal Short Cycle are  apparent for NO  and HC/CO.
                                                    X.
1.2.4.5       Effect of Carburetion vs Fuel Injection
              In the Phoenix fleet, as shown in Table 1-41, the fuel injected
vehicles had decidedly poorer HC and NO  correlations when tested on the
                                        j£.
Federal Short Cycle.  Conversely, the CO correlation coefficient for fuel
injected vehicles was higher (0.88) than for the  carbureted vehicles (0.76).
              On the Federal Three-Mode ST,  the NO  trend was reversed,
with the fuel injected vehicles having  higher NO  coefficients. In the case of
                                              X.
HC, the idle in drive results  were essentially the same while the idle in
neutral results showed much  higher coefficients for the carbureted vehicles.
For CO, carbureted vehicles had higher coefficients in the idle test mode
(drive or neutral),  and thus displayed an effects trend opposite that observed
from the Federal Short Cycle  results.
1.2.4.6       Effect of Transmission Type
              Based on the Phoenix data shown in Table  1-42, the vehicles
with automatic transmissions had significantly higher HC and CO correlation
coefficients when tested on the Federal Short Cycle  ST.  The NO  coefficients
                                              7                x
were insensitive to transmission type.
              The Federal Three-Mode ST results  are similar in the case of
NO .  In the idle in neutral comparative case, the higher HC and CO  coeffi-
cients of the automatic transmission vehicles were in direct agreement with
the Federal Short Cycle trends for these pollutants.
1.2.4.7       Effect of Manufacturer
              Table 1-43 shows the variation of correlation coefficients as
a function of vehicle manufacturer for the Phoenix fleet.  On the Federal
Short Cycle ST,  the most discernible feature is the  extremely low HC
                                    1-72

-------
                                Table  1-41.  Correlation Coefficient Summary by
                                          Fuel System Type — Phoenix
i
-4
OO

Short Test
Federal
Short Cycle

Federal
Three -Mode




Fuel
System Type
Fuel
Injection
Carburetion
Fuel
Injection

Carburetion


Test Mode



High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

No. of
Vehicles
10

90
10
10
10
4
90
90
90
72
ST/FTP(a)
Correlation . .
Coefficients
HC
0.24*

0, 58
0.05*
0.01*
-0.05*
0.73
0. 32
0. 36
0.62
0.71
CO
0.88

0.76
0.60*
0. 13*
0.68
-0. 18*"
0. 34
0.42
0.81
0. 82
NOX
0. 50*

0, 86
0.83
0.75
-0. 11
0.09*
0.72
0.64
0. 10*
-0.07*
            (a)

            (b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence
level except where indicated by an asterisk

-------
                                Table 1-42.   Correlation Coefficient Summary by
                                         Transmission Type — Phoenix

Short Test
Federal
Short Cycle





Transmission
Type
Automatic
Manual
Automatic

Manual


Test Mode


High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

No. of
Vehicles
76
24
76
76
76
76
24
24
24
--
ST/FTP(a)
Correlation ...
Coefficients
HC
0. 72
0. 19*
0.24
0.29
0.65
0.71
0.29*
0.29*
0. 18*

CO
0. 82
0. 53
0.26
0. 38
0.81
0.81
0.73
0.73
0.75
	
NO
X
0.83
0.89
0. 74
0.64
0.13*
0.05*
0.89
0.77
-0. 12*
- ---
I
-4
                      = short test/federal emission certification test procedure

           ^Correlation coefficients are statistically significant at the 95 percent confidence
             level except where indicated by an asterisk

-------
            Table  1-43.
                         Correlation Coefficient Summary by
                          Manufacturer -  Phoenix
Short Test
Federal
Short Cycle





Federal
Three-Mode



















Manufacturer
General
Motors
Ford
Chrysler
American
Motors Company
Others
General
Motors



Ford



Chrysler



AMC



Others



Test Mode








High speed
Low speed
Idle in neutral
Idle in d rive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in d rive
High speed
Low speed
Idle in neutral
Idle in d rive
High speed
Lo\v speed
Idle in neutral
Idle in drive
No. of
Vehicles
33

20
12
6

29

33
33
33
31
20
20
20
17
12
12
12
12
6
6
6
5
29
29
29
1 1
ST/FTP a*
Correlation
Coefficients'5'
HC
0. 80

0.94
-0.01*
0.95

0. 17*

0.25*
0. 32*
0.75
0. 83
0.60
0.64
0.78
0. 85
-0.05*
-0.09*
0. 19*
0. 13*
0. 69*
0. 89
0.96
0.99
0.13*
0. 13*
0.05*
0.68
CO
0. 82

0.97
0.72
0.89

0. 34*

0. 40
0. 59
0. 77
'0. 77
0.41*
0.49
0.96
0.96
0. 30*
0.22*
0.66
0.72
0.45*
0. 87
0.96
1.00
0. 63
0. 34*
0. 62
0. 32*
NO
X
0.93

0.95
0.94
0.91

0. 78

0. 60
0.67
-0.02*
-0.22*
0.87
0.73
-0.26*
-0.47*
0.91
0. 59
0. 36*
0. 34*
0. 82
0. 77*
0. 28*
-0. 38*
0.82
0. 86
0 . 04*
0. 59*
(a)
(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level
except where indicated by an asterisk
                                      1-75

-------
correlation coefficient of the Chrysler vehicles.   This same characteristic
is displayed by the idle test mode of the Federal Three-Mode ST for
Chrysler vehicles.
              The high and low speed mode NO  results of the Federal Three-
                                             .X
Mode agree with the Federal Short Cycle NO  trends, and the idle mode CO
results of the Three-Mode ST agree with the Federal "Short Cycle CO trends.
                                   1-76

-------
1 .3            CONTINGENCY TABLE ANALYSIS DISCUSSION
1.3.1          Hydrocarbon Emission
               The variation of E , E , and FF as a function of HC cut-point
was  determined for each ST.  The results for the pooled fleet with the
Federal Short  Cycle and the Federal Three-Mode STs are shown graphically
in Figures 1-4 and 1-5 to indicate the general nature of the tradeoffs avail-
able.  As can be seen, reducing  the E  increases E  and decreases FF.
               Table 1-44 summarizes the results for both STs  for the
pooled fleet (300 vehicles) and the individual city fleets (100  vehicles each)
at the specific E  analysis value of 5 percent.  This value was used in the
analysis to set HC ST cut-points based on the total pooled  fleet population.
These  same numerical cut-points were then used in subsequent  contingency
analyses on an individual city and test mode basis.  As a consequence, the
resultant actual EC values for the individual cities or test  modes vary from
the 5 percent Ec value for the pooled fleet due to differences in observed
emission characteristics between the pooled fleet and  the individual city
fleets.
               Also shown in Table 1-44 are the effectiveness values (STE)
and the rejection ratio values (STRR) for each test or  test mode.
               In the case of the  Federal Three-Mode  ST,  it is  clear that
the idle-in-drive test  mode is the superior mode for HC discrimination for
each city and the pooled fleet; the STE value ranges from 0.349 to 0.405 for
for the  city fleets and is 0.375 for the pooled fleet..  In comparison, the STE
value for the Federal  Short Cycle ST ranges from 0.530 to 0.691  for the
city  fleets and is 0.589 for the pooled fleet.
               Although the idle in neutral test mode is not as effective as
the idle in drive mode, its STE value is approximately 90  percent of the
idle  in drive STE value; thus, it  would appear appropriate to use that mode
for cars not equipped  with an automatic transmission.
               As can be  seen in Table 1-44, the STRR values are below
one for both STs for all cities and the pooled fleet.
                                   1-77

-------
 o
 <
  o
50

40

30

20
 o
                                   	FF
           1.0   1.2
               1.4   1.6   1.8   2.0
               HC CUT-POINT, gm/mi
                                       2.2   2.4  2.6
Figure  1-4.  Variation of Ec,  Eo, and FF with HC Cut-point;
      All Three Cities; Federal Short Cycle; Bounded
               Errors of Commission Method
  UJ
  O
  a:
  LLJ
  Q-
40

30

20

10

 0
                          c
                         Eo
                         FF
                                    H  -  HIGH SPEED
                                    L  -  LOW SPEED
                                    D  -  IDLE IN DRIVE
                                    N  -  IDLE IN NEUTRAL
          0   100   200   300   400   500    600   700

                     HC  CUT-POINT,  ppm

Figure 1-5.  Variation of Ec,  Eo, and FF with HC Cut-point;
      All Three Cities; Federal Three-Mode; Bounded
               Errors  of Commission Method
                           1-78

-------
            Table 1-44.  Comparison of Short Test HC Results for Pooled Fleet and
                  Individual Cities; Predicted Population; Bounded Errors of
                            Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle



Federal 3-Mode















City
all
Chicago
Houston
Phoenix
all



Chicago



Houston



Phoenix



Test Mode




high
low
idle-in-drive
idle-in-neutral
high
low
idle-in-drive
idle -in -neutral
high
low
idle-in-drive
idle -in -neutral
high
low
idle-in-drive
idle - in - neutral
Number
of
Vehicles
300
100
100
100
300



100



100



100



Parameter, %
E
c
5.0
4.2
4.0
6.3
5.0
5.0
5.0
5.0
8.4
6.9
8.1
7.5
1.3
1.3
2.2
3.2
5.8
7.5
4.3
3.6
E
o
13.0
12.7
11.4
13.6
26.2
24.1
19.8
21.0
20.6
19.5
16.9
18.4
33.7
31.0
23.6
24. 1
24.0
21.8
18.8
20.3
FF
18.7
15.8
25.5
15.3
5.4
7.5
11.9
10.7
7.5
8.5
1L. 5
10.0
3.3
6.0
13.3
12.8
4.9
7. 1
10. 1
8.5
STE(a)
.589
.554
.691
.530
. 172
.237
.375
.338
.266
.303
.405
.352
.088
. 162
.360
.348
.171
.246
.349
.296
STRR(b)
.748
.702
.799
.747
.329
.396
.533
.495
.566
.550
.690
.616
.124
.197
.420
.434
.370
.505
.498
.420
*a)Short Test Effectiveness =   % FF
                         % FF + % E
(b)
  Short Test Rejection Ratio =
7o
                                   FF

-------
1.3.2         Carbon Monoxide Emission
              The variation of E ,  E  and FF for the pooled fleet as a
                                co                r
function of CO cut-point are shown for the Federal Short Cycle and the
Federal Three-Mode in Figures 1-6 and 1-7,  respectively.  Table 1-45
summarizes the results for both STs for the pooled fleet (300 vehicles) and
the individual city fleets at the E analysis value of 5 percent.
              In the case of the Federal Three-Mode ST,  as was the case
for HC  emission,  the idle in drive test mode is superior for CO discrimina-
tion in each city and the pooled fleet; the STE value ranges from 0.539 to
0.584 for the city fleets and is 0.562 for the pooled fleet.  Again, the idle
in neutral STE value is above 90 percent that of the idle in drive mode and
could be substituted for those vehicles without an automatic transmission.
              The STE value for the Federal Short Cycle is approximately
23 percent higher than that of the idle in drive mode; it ranges from 0. 664
to 0.731 for the  city fleets and is 0.690 for the pooled fleet.
              With regard to the short test rejection ratio,  all STRR values
are below one, and relatively similar between cities and the pooled fleet.
1.3.3         Oxides of Nitrogen Emission
              The variation of EC,  EQ, and FF for the pooled fleet as a
function of NO cut points are  shown in Figures 1-8 and 1-9, respectively,
for the Federal Short Cycle and the Federal Three-Mode STs.  Table 1-46
summarizes the results for both STs for the pooled fleet and the individual
city fleets at the E  analysis value of 5 percent.
              With the Federal  Three-Mode ST, as opposed to the findings
for HC  and CO where the idle in drive mode was  superior, the high speed test
mode was superior in NO  discrimination for the pooled fleet and two of the
                         j£
three cities; in Houston the low  speed test mode was superior (STE = 0. 291
in the low speed mode and 0. 249 in the high speed mode).  In the other two
cities the low speed mode STE value was approximately 88 percent that of the
high speed mode,  and in the pooled fleet it was approximately 97 percent of
the high speed mode  value.
                                   1-80

-------
       o
   50


jj  40
?•

j° 30
 "o
r  20


   10
                                                0
                  1
                  4.0   8.0   12.0   16    20
                     CO CUT-POINT,  qmlml
                                       24    26
Figure 1-6.   Variation of Ec, EQ,  and FF with CO Cut-point;
       All Three Cities; Federal Short Cycle; Bounded
                Errors  of Commission Method
       UJ
       o
       LU
       Q-
    60

    50

    40

    30

    20

    10
                                       H -  HIGH SPEED
                                       L - LOW SPEED
                                       D -  IDLE IN DRIVE
                                       N -  IDLE IN NEUTRAL
              0       40,000     80,000     120,000

                       CO CUT-POINT, ppm

Figure 1-7.   Variation of Ec,  Eo,  and FF with CO Cut-point;
       All Three Cities;  Federal Three-Mode;  Bounded
                Errors of Commission  Method
                             1-81

-------
                     Table 1-45.  Comparison of Short Test CO Results for Pooled Fleet and
                           Individual Cities; Predicted Population; Bounded Errors of
                                    Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle



Federal 3-Mode















City
all
Chicago
Houston
Phoenix
all



Chicago



Houston



Phoenix



Test Mode




high
low
idle-in-drive
idle -in-neutral
high
low
idle-in-drive
idle -in -neutral
high
low
idle-in-drive
idle -in-neutral
high
low
idle-in-drive
idle -in-neutral
Number
of
Vehicles
300
100
100
100
300



100



100



100



Parameter, %
E
5.0
4.6
4. 1
5.9
•5.0
5. 0
5.0
5. 0
5.4
5. 8
6.9
6.8
3.7
4.4
4.5
4.5
4. 5
3.8
3.8
3.7
E
o
16.7
19.0
14.4
16.9
44.4
40.7
24.4
26.6
40.9
37.9
26.3
28.0
48.6
43.4
24.5
27.5
43.6
41.2
23.4
25. 1
FF
37. 1
37.6
39. 1
35. 1
10.6
14.4
31.4
29.3
15.7
18.7
30.8
29. 1
8.2
13.7
34.4
31.7
8.3
10.7
28.6
26.8
STE(a)
.690
.664
. 731
.676
. 193
.261
.562
.524
.277
.330
.539
.509
. 145
.239
.584
.536
. 160
.206
.551
. 517
STRR^b)
.783
.746
. 807
.788
.284
.352
.652
.614
.373
.433
.660
.629
.210
.317
.660
.611
.247
.279
.623
.588
I
oo
         (a)
           Short Test Effectiveness =
% FF
                                  % FF + % E
         (b)
           Short Test Rejection Ratio -
                                             FF
                                            FF

-------
  o
  <
o
LU
Q_
     30
     10
0
                                      	FF
                   1
       2.0    2.2   2.4   2.6   2.8   3.0    3.2    3.4   3.6   3.8
                       NOY CUT-POINT,  gm/mi
                          A
Figure  1-8.  Variation of EC, EQ, and FF with NOX Cut-point;
        All Three Cities; Federal Short Cycle; Bounded
                Errors of Commission Method
o
— ^
<
o
LU
"o
LlJ
1—
z
L^J
o
cm
UJ
O_
30

20

10
0
                             Ec
                             Eo
                             FF
                                           H - HIGH SPEED
                                           L - LOW SPEED
                                           D - IDLE IN DRIVE
                                           N - IDLE IN NEUTRAL
            N
        0    500   1000  1500  2000  2500   3000  3500  4000  4500  5000
                          NO  CUT-POINT,  ppm
                            A
Figure  1-9.  Variation of E ,  EQ, and FF with NOX Cut-point;
       All Three Cities;  Federal Three-Mode;  Bounded
                Errors of Commission Method
                             1-83

-------
                    Table 1-46.  Comparison of Short Test NOX Results for Pooled Fleet and
                          Individual Cities; Predicted Population; Bounded Errors of
                                   Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle



Federal 3-Mode















City
all
Chicago
Houston
Phoenix
all



Chicago



Houston



Phoenix



Test Mode




high
low
idle-in-drive
idle-in-neutral
high
low
idle-in-drive
idle -in-neutral
high
low
idle-in-drive
idle -in-neutral
high
low
idle-in-drive
idle -in-neutral
Number
of
Vehicles
300
100
100
100
300



100



100



100



Parameter, %
E
c
5.0
4.2
4.4
6.4
5.0
5.0
5.0
5.0
3.2
3. 1
5.4
5.7
4.2
4. 5
3.8
4.9
6.8
7. 1
5.2
2.9
E
o
8.4
7.5
10.3
7.2
14.6
14.8
17.6
18.4
15.6
16. 1
16.5
16.5
17.7
16.7
18.4
19.7
'9.4
10.7
18.0
19.2
FF
12.8
12. 1
13.3
13.0
6.6
6.4
3.5
2.8
4.0
3.5
3. 1
3.1
5.9
6.9
5.2
3. .9
10.8
9.5
2.2
1.0
STE
.604
.618
.562
.645
. 310
.301
. 167
. 130
.204
. 177
. 159
. 158
.249
.291
.221
. 166
.535
.471
. 108
. 049
STRR
.840
.832
.750
. 960
.547
.538
.. 403
.368
.367
.337
.434
.449
.428
.483
.381
.373
.871
.822
.366
.193
I
00

-------
               The STE value for NO  in the high speed mode was quite
                                   X.
variable with city fleet: 0. 310 for the pooled fleet, 0.204 in Chicago,
0.249 in Houston, and 0. 535 in Phoenix.  The variation in STE value for
the Federal Short Cycle was much less  variable: 0.604 for the pooled fleet,
0.618 in Chicago, 0.562 in Houston, and 0.645 in Phoenix.  Except for
Phoenix, the Federal Short Cycle NO STE values were approximately
double those of the Federal Three-Mode high speed test mode; in Phoenix
the high speed NO  STE value was approximately 83 percent of the Federal
Short Cycle value.
               In the case  of short test rejection ratio, all STRR values are
less than one.  Phoenix, however, displays considerably higher STRR values
than the other two cities (0.960 with Federal Short Cycle  and 0.871 with high-
speed mode of the Federal Three-Mode); this, of course, is a reflection of
the higher  E  values in Phoenix as compared with Chicago and Houston.
1.3.4          Multiple-Constituent Tests
               In addition to  analyzing each pollutant individually, an analy-
sis was made for three-constituent tests.  In a three-constituent test,  a car
fails the ST if any of its HC, CO, or NO  measurements  exceed the pre-
                                       j£
viously determined cut-points for each pollutant.
               The three-constituent test results for both STs for the  pooled
fleet and the individual city fleets are summarized in Table 1-47 for the
E  analysis value of 5 percent.  In the case of the Federal Short Cycle ST,
STE values over  0.8 were achieved in all cities and the pooled fleet.
               The "best" test mode category shown for the Federal Three-
Mode ST refers to the combination of best individual test modes:  the  idle-
in-drive for HC and CO discrimination and the high-speed mode  for NO ,  as
                                                                     -X.
was discussed in the previous  sections.  Although this combination of best
individual test modes had relatively high STE values in all three cities
(0.606 to 0.727) and in the pooled fleet (0.663),  the idle-in-drive mode
alone was slightly superior to this combination of modes:  0.652 in
Chicago, 0.773 in Houston, 0. 688 in Phoenix, and 0.704 for the pooled
                                  1-85

-------
                   Table 1-47.  Comparison of Short Test Multiple Constituent Results for
                        Pooled Fleet and Individual Cities; Actual Population; Bounded
                               Errors of Commission Method (E  set at 5 percent)


Short Test
Federal Short Cycle



Federal 3- Mode



















City
aU
Chicago
Houston
Phoenix
all




Chicago




Houston


'

Phoenix




Test Mode




high
low
idle-in-drive
idle -in-neutral
best(a>
high
low
idle-in-drive
idle -in-neutral
best
high
low
idle-in-drive
idle -in-neutral
bestx at high speed
00

-------
                   Table  1-47.   Comparison of Short Test Multiple Constituent Results for
                        Pooled Fleet and Individual Cities; Actual Population; Bounded
                        Errors of Commission Method (E   set at 5 percent) (Continued)
Short Test
Federal 3- Mode
(continued)

City
Phoenix
(continued)

Test Mode
idle- in- neutral

best(a)
Number
of
Vehicles



Parameter, %
E
4.0

4. 0
E
o
27.0

22. 0
FF
37.0

42.0
STE
0.578

0.656
STRR
.641

.719
oo
-vl
         (a)
           Best individual modes are HC in drive, CO in drive, and NOX at high speed

-------
fleet.  This is the result of the relatively high incidence of HC and CO FTP
failures in the vehicles and the relatively low incidence of NO  FTP
failures.  Thus,  the inclusion of the high-speed mode, which was  shown
to be the superior NO  discriminator, was relatively less effective on a
three-constituent test basis than the idle-in-drive mode alone.
              With regard to rejection ratio,  all STRR values in Table 1-47
are below one.  There is relatively little difference in STRR values between
the three cities.
1.3.5         Effect of Engine Displacement
              The variation of short test effectiveness and  rejection ratio
with engine displacement for the pooled fleet is summarized in Tables 1-48
through 1-51 for  the Federal Short Cycle and the  better individual test modes
of the Federal Three-Mode ST.
1.3.5.1       Hydrocarbon Emission
              In the  case  of HC (Table 1-48) on the Federal Short Cycle, the
STE value  of the  150  CID and less group (0.579) was essentially equivalent
to the fleet average (0.589), the 151 to 259 GID group value (0.471) was
20 percent lower than the fleet average, and the 260 CID and greater group
value (0.632) was 7 percent higher than the fleet average.
              Using  the idle-in-drive mode of the Federal Three-Mode ST,
the above CID-related trends were reversed.  The 150 CID  and less group
STE value (0.319) was 15 percent lower than the fleet average (0.375); the
151 to 259  CID group value (0.380) was essentially equivalent to the fleet
average; and the  260  CID and greater group value (0.356)  was 5 percent
lower than the fleet average.  Similar trends are  observed for the idle-in-
neutral test mode.  The actual STE value of the idle-in-neutral mode varies
from 3 percent higher than the idle-in-drive mode value with the 260 CID
and greater group to  18 percent lower than the idle-in-drive mode value with
the 151 to 259 CID  group.  For the pooled  fleet, the idle-in-neutral STE
value was 11 percent lower than the idle-in-drive STE value.
                                   1-88

-------
                  Table 1-48.  Comparison of Short TestHC Results by Engine Displacement;
                          All Three Cities; Predicted Population; Bounded Errors of
                                   Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle



Federal 3- Mode







CID Group
all
0 to 150
151 to 259
260 & greater
all

0 to 150

151 to 259

260 & greater

Test Mode




idle-in-drive
idle -in- neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle -in-neutral
Number
of
Vehicles
300
95
54
151
300

95

54

151

Parameter, %
E
c
5. 0
4.6
4. 6
5. 3
5.0
5.0
7. 1
5. 7
7. 1
6.8
3.2
4. 1
E
o
13.0
13.5
14.4
12. 1
19.8
21.0
19.6
22.7
16.9
18.5
21.2
20.9
FF
18.7
18.6
12.8
20. 8
11.9
10.7
12.6
9.4
10.3
8.8
11.7
12.0
STE
. 589
.579
.471
.632'
.375
.338
. 319
.293
. 380
.323
.356
.366
STRR
.748
. 723
.640
.793
.533
.495
.612
.470
. 640
.571
.453
.489
I
00

-------
              On a comparative basis, the Federal Three-Mode STE value
(idle in drive) was 36 percent lower than the Federal Short Cycle for the
pooled fleet.
              All rejection ratio (STRR) values in the table are less than
one; the 260 CID and greater group has relatively lower STRR values than
the other two GID groups on the Federal Three-Mode, and a relatively higher
STRR value on the Federal Short Cycle.
1.3.5.2       Carbon Monoxide Emission
              The results for CO are given in Table 1-49.  On the Federal
Short Cycle, the STE value of the  150 CID and less group (0.773) is 12 per-
cent higher than the fleet average  (0.690), the 151 to 259 CID group value
(0. 631) is 9 percent lower than the fleet average,  and the 260 CID and
greater group is 2 percent lower than the fleet average.
              Somewhat similar CO trends are observed for the idle-in-
drive mode of the Federal  Three-Mode ST.  The 150 CID and less group
STE value (0.639) is 14 percent higher than the fleet average (0.562); the
151 to 259 CID group value (0.585) is 4 percent lower than the  fleet: average;
and the 260 CID and greater group is  8 percent lower than the fleet average.
The idle-in-neutral mode exhibits similar trends, except for the 151 to 259
CID group where the STE value is above the  fleet average value instead of
lower, as was the case of the idle-in-drive mode.
              The idle-in-neutral STE values are lower than the idle-in-
drive values,  ranging from 4 percent lower for the 260 CID and greater
group to  13 percent lower for the  151 to 259  CID group.
              On a pooled fleet basis, the Federal Three-Mode ST (idle  in
drive) STE value was 19 percent lower than the Federal Short Cycle STE
value.
              With regard to short test rejection  ratio, the 150 CID  and
less group had STRR values considerably above the other two CID groups
on both STs; the STRR value  slightly exceeded one (1.051) on the Federal
Short Cycle.
                                  1-90

-------
                   Table 1-49.  Comparison of Short Test CO Results by Engine Displacement;
                           All Three Cities; Predicted Population; Bounded Errors of
                                    Commission Method (E   set at 5 percent)
Short Test
Federal Short Cycle



Federal 3 -Mode







CID Group
all
0 to 150
151 to 259
260 & greater
all

0 to 150

151 to 259

260 & greater

Test Mode




idle-in-drive
idle -in-neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle-in-neutral
Number
of
Vehicles
300
95
54
151
300

95

54

151

Parameter, %
E
5.0
12.5
2.7
2.7
5.0
5.0
12.5
12. 1
5.0
6.2
1.8
1.7
E
o
16.7
10.2
20.5
18.6
12.4
26.6
16.2
19.0
24.7
29.2
28.8
30.3
FF
37. 1
34.5
35.0
38.5
31.4
29.3
28.6
25.8
34.8
30.2
31. 1
29.8
STE
.690
.773
.631
.674
.562
.524
.639
.576
.585
.508
.519
.496
STRR
.783
1.051
.679
.722
.813
.614
.917
.846
.670
.613
. 549
.524
I
SO

-------
1.3.5.3       Oxides of Nitrogen Emission
              NO  results are shown in Table 1-50.  On the Federal Short
                 J?C
Cycle, the STE value of the 150 CID and less group (0.530) is 12 percent
lower than the fleet average (0.604),  the 151 to 259 CID group value (0.626)
is 4 percent higher than the fleet average,  and the 260 CID and greater group
value (0.600) is  essentially the same as the fleet average.
              The NO  versus CID trends for  the high-speed mode of the
Federal Three-Mode ST are markedly different from those noted above for
the Federal Short Cycle.  The 150 CID  and less group STE value (0.578) is
87 percent higher than the Federal Three-Mode fleet  average value (0.310)
and only 4 percent lower than the fleet average value  for the Federal Short
Cycle (0.604).  The 151 to 259  CID group STE  value (0.154) is  50 percent
lower than the fleet average, and the 260 CID and greater value is 22 per-
cent lower than the fleet average.
              The low-speed mode STE values exhibit trends similar to
the high-speed mode, except in the case of the 151 to 259 CID group wherein
the low-speed STE (0.372) is higher than the fleet average  (0.301) and higher
than the high-speed test mode value (0. 154).
              On a pooled-fleet basis,  the Federal Three-Mode ST (high-
speed mode) STE value was 50 percent lower than the Federal Short Cycle
STE value.
              In the case of rejection ratio, the  150 CID and less group had
the lowest STRR value on the Federal Short Cycle (0.729).   On the Federal
Three-Mode, this same group had considerably higher STRR values than the
other two CID groups; the value exceeded one (1 .219)  in the high-speed
mode.
1.3.5.4       Multiple Constituent Tests
              The three-constituent test results  for both STs on a CID
group basis are summarized in Table 1-51 .  In the case of the Federal
Short Cycle,  the 150 CID and less group STE value (0.782) was  5 percent
lower than the fleet average (0.821),  the 151 to 259 CID group value (0.950)
                                  1-92

-------
Table 1-50.  Comparison of Short Test NOX Results by Engine Displacement;
         All Three Cities;  Predicted Population; Bounded Errors of
                  Commission Method (E   set at 5 percent)
Short Test
Federal Short Cycle



Federal 3-Mode







CID Group
all
0 to 150
151 to 259
260 & greater
all

0 to 150

151 to 259

260 & greater

Test Mode




high
low
high
low
high
low
high
low
Number
of
Vehicles
300
95
54
151
300

95

54

151

Parameter, %
E
c
5. 0
3. 1
5.0
6.7
5:0
5. 0
9.9
4.7
2.8
3.5
2.8
4.6
E
o
8.4
7.3
7.5
9.9
14.6
14.8
6.5
9.2
17.0
12.6
18.8
19.7
FF
12.8
8.2
12.5
14.9
6.6
6.4
9.0
5.0
3. 1
7.4
6.0
5. 1
STE
.604
. 530
.626
. 600
. 310
.301
.578
.355
. 154
.372
.242
.206
STRR
.840
.729
.875
.871
. 547
.538
1.219
.683
. 294
.545
. 355
.391

-------
                     Table 1-51.  Comparison of Short Test Multiple Constituent Results by
                           Engine Displacement; All Three Cities; Predicted Population;
                           Bounded Errors of Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle



Federal 3-Mode











CID Group
all
0 to 150
151 to 259
260 & greater
all


0 to 150


151 to 259


260 & greater


Test Mode




idle-in-drive
idle -in-neutral
best (a)
idle-in-drive
idle -in-neutral
best 
idle-in-drive
idle -in-neutral
best
idle-in-drive
idle -in-neutral
best
Number
of
Vehicles
300
95
54
151
300


95


54


151


Parameter, %
E
7. 0
8.4
1.9
8.0
6.3
4.7
4. 3
8.4
9.5
10.5
9.3
5.6
1.8
4. 0
1. 3
1. 3
E
o
11.7
12.6
13.0
10.6
19.3
26.3
22.0
14.7
20.0
12.6
18.5
33.3
25.9
22. 5
27.8
26.5
FF
53.7
45.3
55.6
58.3
46.0
39.0
43.3
43.2
37.7
45.3
50.0
35.2
42.6
46.4
41.0
42.4
STE
.821
.782
. 950
.846
. 704
.597
.663
. 746
.655
.782
.729
.514
.622
.673
.596
.615
STRR
.928
.927
. 981
. 962
.801
.669
.714
. 891
.818
. 964
.866
.596
.648
.731
.615
.634
I
vO
         (a)
           Best Individual Modes are HC in Drive,
           CO in Drive and NO at High Speed
                            X

-------
was  16 percent higher than the fleet average, and the 260 CID and greater
group STE value (0.846) was 3 percent higher than the fleet average.
              In the case  of the Federal Three-Mode,  only the idle and best
(best combination of individual test modes: HC in drive, CO in drive, and NO
                                                                         5
in high speed) modes are shown,  because the high and low speed modes dis-
played markedly inferior discrimination capability on a three-constituent
basis (see Table 1-47, for example).   The idle-in-drive  mode is the
superior test mode in all cases except the 150 CID and less group, wherein
the best mode STE value (0.782) was 5 percent higher than the idle-in-drive
mode.  Although the idle-in-neutral test mode produced lower STE values
than the idle-in-drive mode (from 12 to 30 percent lower),  it would be
applicable to cars with standard transmissions.
              The CID-related trends of the idle-in-drive  test mode (and
best combination, as well) were not the  same as exhibited by the Federal
Short Cycle.  The 150 CID and less  group STE value (0.746) was 6 percent
higher than the fleet average (0.704); the 151 to 259 CID group value (0.729)
was  4 percent higher than the fleet average; and the 260 CID and greater
group value (0. 673) was  4 percent lower than the fleet average.
              With regard to rejection ratio, all  STRR values were below
one,  as shown in Table 1-51.  There is little discernible difference between
the CID groups on the Federal Short Cycle; on the Federal  Three-Mode ST,
the 150 CID and less group has the highest STRR values.
1.3.6         Effect of Inertia Test Weight Group
              The variation of short test effectiveness and rejection ratio
with  vehicle inertia test  weight group is summarized in Tables 1-52 through
1-55 for the Federal Short Cycle and the better individual test modes  of the
Federal Three-Mode ST.
                                   1-95

-------
1.3.6.1       Hydrocarbon Emission
              In the case of HC (Table 1-52) on the Federal Short Cycle,
the STE value of the 2500-lb and less weight group (0.431.) is 27 percent
lower than the fleet average (0.589), the 2501-lb to 3500-lb weight group
STE value (0.606) is 3 percent higher than the fleet average,  the 3501-lb
to 4500-lb weight group STE value is the highest (0.667,  13 percent above
the fleet average), and the 4501-Ib and greater weight group has an STE
value (0.575) which is 2 percent below the fleet average.
              With the idle-in-drive mode of the  Federal Three-Mode ST,
the above noted trends for three of the four weight groups are similar.
The 2500-lb and less  weight group STE value (0.293) is 22 percent lower
than the fleet average (0.375); the 2501-lb to 3500-lb weight group STE
value (0.439) is 17 percent higher than the fleet average; the  3501-lb to
4500-lb weight group  STE value (0. 340) is 9 percent below the fleet average
(it was 13 percent higher than fleet average  on the Federal Short Cycle);
and the 4501-lb and greater weight group has an STE value (0.385) which
is 3 percent higher than the fleet average (it was 2 percent below the average
on the Federal Short Cycle).
              Trends for the idle-in-neutral test mode are similar to those
noted above for the idle-in-drive test mode.  For the pooled fleet, the
idle-in-neutral STE value (0.338)  was 10 percent  lower than the idle-in-
drive STE value (0.375).
              On a comparative basis, the Federal Three-Mode ST (idle-
in-drive mode) was 36 percent lower in STE value than the Federal Short
Cycle, for  the pooled fleet.
              All inertia weight groups had STRR values less than one  on
both STs, with little discernible weight-related trend information.
                                   1-96

-------
Table 1-52.  Comparison of Short Test HC Results by Inertia Weight;
     All Three Cities;  Bounded Errors of Commission Method;
             Predicted Population (E  set at 5 percent)
Short Test
Federal Short Cycle




Federal 3-Mode









Inertia
Weight
all
0 to 2500
2501 to 3500
3501 to 4500
4501 & greater
all

'o to 2500

2501 to 3500

3501 to 4500

4501 & greater

Test Mode





idle-in-drive
idle -in-neutral
idle-in-drive
idle -in-neutral
idle-in-drive
idle -in-neutral
idle-in-drive
idle -in- neutral
idle-in-drive
idle-in-neutral
Number
of
Vehicles
300
50
119
98
33
300

50

119

98

33

Parameter, %
E
c
5. 0
3.5
4.5
4.5
8.3
5. 0
5. 0
4.4
4. 2
6.6
5.5
4.2
5. 0
2. 3
3.9
E
o
13.0
19.0
11.4
9.3
21. 1
19.8
21.0
24.0
25.0
16.3
19.0
18.5
18.7
30.6
29.4
FF
18.7
14.4
17.6
18.7
28.6
11.9
10.7
9.8
8.9
12.7
10.0
9.5
9.3
19. 1
20.3
STE
.589
.431
.606
.667
. 575
.375
.338
.293
.263
.439
. 344
. 340
. 333
. 385
.408
STRR
.748
.536
.567
.829
. 742
.533
.495
.420
. 386
.666
.534
.489
.511
.431
.487

-------
1.3.6.2       Carbon Monoxide Emission
              The results for CO are given in Table 1-53. On the Federal
Short Cycle, the STE value of the 2500-lb and less weight  group (0.715) is
4 percent higher than the fleet average (0. 690), the 2501-Ib to 3500-lb
weight group STE  value (0. 663) is 4 percent lower than the fleet average,
the 3501-Ib  to 4500-lb weight group value (0.718)  is 4 percent higher than
the fleet  average,  and the 4501-lb and greater  weight group has an STE
value (0.705) that  is 2 percent higher than the fleet average.  All weight
groups are similar in STE value (0.705 to 0.718)  except the 2501-lb to
3500-lb group, whose STE value (0.663) is  7 percent below the  average of
the other three weight groups.
              Somewhat different trends are observed  for the idle-in-drive
mode of the Federal Three-Mode ST.  Here, the STE values of the 2500-lb
and less  weight group (0.583) and the 2501-lb to 3500-lb group (0.617)  are
above the fleet average (0.562) by 4 and  10  percent, respectively; the STE
values of the two heavier weight groups (0.553 for 3501 Ib to 4500 Ib;
0.509 for 4501 Ib and greater) are 5 and 10  percent lower than the fleet
average, respectively.  The idle-in-neutral mode exhibits similar trends
except for the 4501-lb and  greater weight group which recorded the highest
STE value obtained (0.838) by either the Federal Short  Cycle  or the Federal
Three-Mode.
              The idle-in-neutral STE values  were lower than  the idle-in-
drive values (6 to  10 percent lower) except  for the 4501-lb and greatest
group, as noted above, which was 65 percent higher than the idle-in-drive
value.
              On  a pooled-fleet basis, the  Federal Three-Mode ST (idle
in drive), STE value was 19 percent lower than the Federal Short Cycle STE
value.
              With regard to rejection ratio, the higher E values of the
2500-lb and less group result in STRR values considerably higher than the
other weight groups for both STs. On the Federal Short Cycle, the STRR
value of this group exceeds one (1.064).
                                  1-98

-------
                     Table  1-53.  Comparison of Short Test CO Results by Inertia Weight;
                          All Three Cities;  Bounded Errors of Commission Method:
                                  Predicted Population (E  set at 5 percent)
Short Test
Federal Short Cycle




Federal 3-Mode









Inertia
Weight
all
0 to 2500
2501 to 3500
3501 to 4500
4501 & greater
all

0 to 2500

2501 to 3500

3501 to 4500

4501 & greater

Test Mode





idle-in-drive
idle -in-neutral
idle-in-drive
idle -in-neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle -in-neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
300
50
119
98
33
300

50

119

98

33

Parameter, %
E
c
5.0
13. 1
4.6
3. 1
2. 3
5. 0
5.0
14.6
14.5
5.6
5.8
2.0
1.6
1. 0
6. 1
E
0
16.7
10.7
16.5
16.0
22.3
24.4
26.6
15.8
17.8
19.3
22.4
27.9
30.0
40.2
12.3
FF
37. 1
26.9
32.6
40.8
53. 1
31.4
29.3
22.0
20. 0
31.0
28.2
31.8
30. 1
41.7
65.5
STE
.690
.715
.663
. 718
.705
.562
.524
.583
.528
.617
.557
.533
. 500
.509
. 838
STRR
.783
1.064
.758
.773
.735
.652
.614
.968
. 913
.728 .
.672
.566
.527
.521
.920
I
vQ
vO

-------
1.3.6.3       Oxides of Nitrogen Emission
              NO  results are shown in Table 1-54.  On the Federal Short
                 x
Cycle, all inertia weight groups except the 2500-lb and less group were
similar in STE value (from 0.585 to 0.618) and corresponded well with the
fleet average STE (0.604).  The smallest weight group had a very low  STE
value of 0.251,  68 percent lower than the fleet average.
              The high-speed mode of the Federal Three-Mode shows quite
different results.  Here, the smallest weight group (2500 Ib and less) has the
highest STE value (0.552); the 2501-lb to 3500-lb group value (0.315) is near
the fleet average (0.310); and the two largest weight groups have increasingly
poorer STE values (0.214 and 0.114, respectively).
              The low-speed mode results show the 2501-lb to 3500-lb group
with the highest STE value (0.417), with the  other three groups lower than
the fleet average value of 0.301.  Although essentially the same in STE value
for the pooled fleet as the high-speed mode,  the low-speed  mode more
closely agrees with the  Federal Short Cycle  with regard to the effect of
the smallest inertia test weight  group.
              On a pooled-fleet basis, the Federal Three-Mode (high-speed
mode) STE value was 49 percent lower than the Federal Short Cycle value.
This is largely due to-the relatively low STE values obtained for the two
largest inertia test weight  groups.
              With the  Federal Short Cycle, all STRR values are less than
one, and the 2500-lb and less weight  group has an STRR value less than
half that of the other weight groups.  This situation is reversed with the
Federal Three-Mode.  Here,  the STRR value is  3.0 for the 2500 Ib and
less group on  the high-speed mode and just under one (0.981) on the low-
speed mode.
1.3.6.4       Multiple  Constituent Tests
              The three-constituent test results for both STs on an inertia
test weight basis are summarized in Table 1-55.  In the case of the Federal
                                  1-100

-------
                     Table 1-54.   Comparison of Short Test NOX Results by Inertia Weight;

                          All Three Cities; Bounded Errors of Commission Method;
                                   Predicted Population (E   set at 5 percent)
Short Test
Federal Short Cycle




Federal 3-Mode









Inertia Weight
all
0 to 2500
2501 to 3500
3501 to 4500
4501 & greater
all

0 to 2500

2501 to 3500

3501 to 4500

4501 fc greater

Test Mode





high
LOW
high
low
high
low
high
low
high
low
Number
of
Vehicles
300
50
119
98
33
300

50

119

98

33

Parameter, %
EC
5.0
0.8
4.5
5.9
5.1
5.0
5.0
13.2
3.9
4.2
5.4
2.6
3.9
1.3
1.6
E
0
8.4
4.0
9.1
9.5
7.8
14.6
14.8
2.4
4.0
16.3
13.9
18.7
19.3
16.6
17.0
FF
12.8
1.3
14.7
14.3
11.0
6.6
6.4
3.0
1.3
7.5
10.0
5.1
4.6
2.1
1.8
STE
.604
.251
.618
.602
.585
.310
.301
.552
.250
.315
.417
.214
. 191
. 114
.095
STRR
.840
.396
.807
.849
.856
.547
.538
3.00
.981
.492
.644
.324
.356
. 182
. 182
I
>-*
o

-------
                  Table 1-55.  Comparison of Multiple Constituent Results by Inertia Weight;

                           All Three Cities; Bounded Errors of Commission Method;

                                     Actual Population (E  set at 5 percent)
Short Test
Federal Short Cycle




Federal 3-Mode
'








Inertia Weight
all
0 to 2500
2501 to 3500
3501 to 4500
4501 & greater
all

0 to 2500

2501 to 3500

3501 to 4500

4501 & greater

Test Mode





idle-in-drive
idle -in-neutral
best(a)
idle-in-drive
idle -in-neutral
best(a)
idle-in-drive
idle -in-neutral
best(a)
idle-in-drive
idle -in -neutral
best
idle-in-drive
idle - in-neutral
best
-------
Short Cycle, all weight groups had STE values of the  same approximate
level (0.756 to 0.885), compared with the fleet average value of 0.821.
The 2501-lb to 3500-lb weight group had the lowest value, which was only
8 percent below the fleet average.
               In the case of the Federal Three-Mode, only the idle and
best combination of individual test modes  are shown.   The idle-in-drive test
mode is the superior test mode in all cases except the 2500-lb and less
weight group, where the best mode  STE value (0.739) was 6 percent higher
than the idle-in-drive value.  The idle-in-neutral test mode values were
the same  as the idle-in-drive values for the 2500-lb and less and 4501-lb
and greater weight groups; for the other two weight groups, the idle-in-
neutral STE values were approximately 20 percent lower than the idle-in-
drive values.
               Whereas the 2501-lb to 3500-lb weight  group had the lowest
STE value (0.756) on the Federal Short  Cycle, it had the highest STE value
on the idle-in-drive test mode (0.771) and nearly the highest value (0.732)
with the best combination test mode.
              All STRR values  are relatively high on the Federal Short
Cycle; the 2500-lb and less weight group is the highest in STRR value
(1.087).  This weight group also has the highest STRR values  on the
Federal Three-Mode, reaching  1.0 with the combination of best individual
test modes.
1.3.7         Effect of Emission Control System Type
              The variation of short test effectiveness and rejection ratio
with emission control system type is summarized in Tables 1-56 through
1-59 for the Federal  Short Cycle and the better individual test modes of the
Federal Three-Mode ST.
1.3.7.1       Hydrocarbon Emission
              As  shown in Table  1-56,  on the Federal Short Cycle the use
or non-use of oxidation catalysts and/or secondary air injection for HC
                                  1-103

-------
                  Table 1-56.  Comparison of Short Test HC Results by Emission Controls;
                          All Three Cities; Predicted Population; Bounded Errors of
                                  Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle




Federal 3-Mode









Emission
Controls
Catalyst /2nd Air
all
yes/yes
yes/no
no/yes
no /no
aU

yes/yes

yes /no

no/yes.

no/no

Test Mode





idle-in-drive
idle -in-neutral
idle-in-drive
idle - in-neutral
idle-in-drive
idle - in -neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
300
89
127
59
25
300

89

127

59

25

Parameter, %
E
5.0
3.2
4.5
7.5
4.7
5.0
5.0
3.5
3.5
4.6
5.5
5.4
. 3.5
6.8
5.3
E
0
13.0
8.6
14.7
10.7
34.0
19.8
21.0
17.0
17.2
18.3
18.6
20.1
24.5
35.3
37.2
FF
18.7
11.4
22.3
18.9
20.4
11.9
10.7
3.0
2.7
18.6
18.4
9.5
5.1
19.1
17.2
STE
.589
.570
.602
.639
.376
.375
.338
.149
.136
.504
.498
.321
.172
.352
.316
STRR
.748
.730
.724
.892
.461
.533
.495
.325
.312
.629
.646
.503
.291
.476
.432
I
I-*
o

-------
control produces no discernible trends except that those vehicles which
did not have either a catalyst or  secondary air injection ("no/no") had a
definitely lower STE value (0. 376),  which was 36 percent lower than the
fleet average.
              Using the idle-in-drive mode of the Federal Three-Mode ST,
the above trend  is not seen.   Indeed, the no/no STE value (0.352) is second
only to the "yes/no"  category (0.504).  With both' a catalyst and secondary
air ("yes/yes"), the STE value is quite low (0. 149).
              The idle-in-neutral results parallel those of the idle-in-drive
test mode; on a  fleet average basis, the neutral mode STE  value is  10 per-
cent lower than  the drive mode STE value.
              All STRR values were below one,  with no readily discernible
trends by control system type.
1.3.7.2       Carbon Monoxide  Emission
              The results for CO are given in Table  1-57.  With the  Federal
Short Cycle ST, there appear to  be no readily observed trends  attendant
to the use or non-use of catalysts and/or secondary air.  However, the
                ">
yes/yes category does have the lowest STE value (0. 642, 7 percent below the
fleet average).
              With the idle-in-drive mode of the Federal Three-Mode ST,
the yes/yes category again has the lowest STE value (0.478,  15 percent
below the fleet average). The idle-in-neutral test mode results closely
parallel the idle-in-drive findings; on a fleet average basis, the neutral
mode is 7 percent lower in STE value than the drive mode.
              With regard to rejection-ratio, all groups were similar
except for the "no/yes" category. This  group had an STRR value in excess
of one  (1.234) on the  Federal Short Cycle and STRR values  of 0.964 and
1.029 for the idle-in-drive and neutral test modes, respectively.  These
STRR values reflect the high  E   values for this  control combination.
                             c
                                 1-105

-------
                   Table  1-57.  Comparison of Short Test CO Results by Emission Controls;
                          All Three Cities; Predicted Population; Bounded Errors of
                                   Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle




Federal 3 -Mode









Emission
Controls
Catalyst/2nd. Air
all
yes/yes
yes/no
no /yes
no /no
all

yes /yes

yes /no

no /yes

no/no

Test Mode





idle -in -drive
idle-in-neutral
idle-in-drive
idle - in-neutral
idle-in-drive
idle- in -neutral
idle- in-drive
idle -in -neutral
idle- in-drive
idle- in-neutral
Number
of
Vehicles
300
89
127
59
25
300

89

127

59

25

Parameter, %
E
5.0
4.8
2.1
17.0
8.1
5.0
5.0
2.5
2.5
2.8
3.2
15.0
13.3
10.7
9.2
E
o
16.7
14.3
20.0
7.3
15. 1
12.4
26.6
22.0
24.2
30. 1
32.0
16.5
22.1
13.6
13.3
FF
37.1
25.6
46.7
34.1
33.8
31.4
29.3
20.1
18.0
39.3
37.9
25.0
19.3
35.2
35.4
STE
.690
.642
.700
.823
.691
.562
.524
.478
.425
.566
.543
.602
.467
.721
.725
STRR
.783
.762
.732
1.234
.857
.831
.614
.537
.486
.607
.588
.964
1.029
.941
.916
I
h-fe
o

-------
Although less than one in STRR value, the no/no group was  also high on the
Federal Three-Mode with values of 0.941 and 0.916 on the idle-in-drive
and neutral test modes.
1.3.7.3       Oxides of Nitrogen Emission
              NO  results are shown in Table 1-58.  On the Federal Short
Cycle ST, the no/yes category has the lowest departure from the fleet
average STE value  of 0.604, being 14 percent lower at  STE = 0.522,  whereas
the no/no category  (0. 696) is 15 percent higher than the fleet average.  The
yes/yes category is within 4 percent of the fleet average, and the yes/no
category is essentially equivalent to the fleet average.
              Using the high-speed mode of the Federal Three-Mode, the
above trends are reversed. The no/no category (0.188) is 40 percent lower
than the fleet average STE value (0.310), and the no/yes category (0.495) is
60 percent higher than the fleet average.  Similar trends are also shown
with the low-speed  test mode.
              With regard to  rejection ratio, the  yes/no category exceeded
one in STRR value (1. 057) on the Federal Short Cycle,  and the no/no
category exceeded one (1 . 151)  on the high-speed mode of the Federal
Three-Mode.
1.3.7.4       Multiple Constituent Tests
              The  three-constituent test results for both STs on an emis-
sion control system basis are summarized in Table 1-59.  In the case of
the Federal Short Cycle, the most notable effects  are the yes/yes category
STE value of 0.764 (7 percent below the fleet average of 0.821) and the
no/no category STE value of 0.933 (14 percent above the fleet average).
              In the case of the Federal Three-Mode ST, only the idle and
best combination of individual test modes are shown.  The idle-in-drive
test mode indicates that the yes/yes category has  the lowest STE value
(0.471,  33 percent  below the fleet average of 0.704); the other categories
all have rather high STE values (0.703  to 0.817,  which approach the STE
                                 1-107

-------
                  Table 1-58.  Comparison of Short Test NO  Results by Emission Controls;
                          All Three Cities; Predicted Popula^on; Bounded Errors of
                                  Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle

,.
i

Federal 3-Mode









Erniss ion
Controls
Catalyst/2nd. Air
all
yes/yes
yes/no
no /yes
no /no
all

yes /yes

yes/no

no/yes

no/no

Test Mode





high
low
high
low
high
low
high
low
high
low
Number
of
Vehicles
300
89
127
59
25
300

89

127

59

25

Parameter, %
E
c
5.0
3.8
7.2
1.4
3.9
5.0
5.0
1.9
1.'3
5.4
8.1
6.6
3.6
12.9
6.2
E
o
8.4
14.8
6.3
8.8
4.2
14.6
14.8
.21.8
25.6
11.5
9.6
9.3
10.0
11.3
10.5
.FF
12.8
20.6
9.6
9.6
9.7
6.6
6.4
8.6
4.8
4.4
6.3
9.1
8.3
2.6
3.4
STE
.604
.582
.602
.522
.696
.310
o301
.283
. 158
.279
.395
.495
.451
.188
.244
STRR
.840
.689
1.057
.598
.978
.547
.538
.345
.201
.616
.906
.853
.650
1.151
.691
O
00

-------
                Table 1-59.  Comparison of Short Test Multiple Constituent Results by Emission
                      Controls; All Three Cities; Predicted Population; Bounded Errors of
                                    Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle




Federal 3-Mode








Emission
Controls
Catalyst/End. Air
all
yes /yes
yes /no
no /yes
no /no
all

yes/yes

yes /no

no/yes

no /no
Test Mode





idle-in-drive
idle -in -neutral
best(a>
idle-in-drive
idle-in-neutral
best
idle-in-drive
idle -in -neutral
best(a>
idle-in-drive
idle - in -neutral
best(a)
idle- in-drive
Number
of
Vehicles
300
89
127
59
25
300

89

127

59

25
Parameter, %
E
c
7.0
10.1
3.9
5.1
11.0
6.3
4.7
4.3
3.4
1.1
3.4
6.3
3.9
2.4
8.5
6.8
8.5
12.0
E
0
11.7
13.5
13.4
8.5
4.0
19.3
26.3
22.0
30.3
36.0
30.3
13.4
19.7
18.9
18.6
30.5
20.3
12.0
FF
53.7
43.8
59.8
54,2
56.0
46.0
39.0
43.3
.27.0
21.3
27.0
59.8
53.5
54.3
44. 1
32.2
42.4
48.0
STE
.821
.764
.816
.864
.933
.704
.597
.663
.471
.372
.471
.817
.731
.742
.703
.514
.676
.800
STRR
.928
.941
.870
.946
1. 120
.801
.669
.729
.531
.391
.531
.903
.784
.775
.839
.622
.812
1.00
I
H*-
o
          (a)
            Best individual modes are HC in drive,
            CO in drive, and NOX at high speed.

-------
value range of the Federal Short Cycle).  The idle-in-neutral and best
combination test modes follow the  idle-in-drive trends  at a slightly lower
STE value.
              All STRR values were less than one, except for  the no/no
category.  On the Federal Short Cycle, the STRR value was l.ZO and on the
idle-in-drive mode of the Federal  Three-Mode it was 1.0.
1.3.8         Effect of Carburetion vs Fuel Injection
              The variation of short test effectiveness  and rejection ratio
with fuel system type is  summarized in Tables 1-60 through  1-63 for the
Federal Short Cycle and the better individual test modes of the Federal
Three-Mode ST.
1.3.8.1       Hydrocarbon Emission
              In the case of HC (Table 1-60) on the Federal Short Cycle,
the carbureted vehicles achieved an STE  value (0.600) 20 percent higher
than the STE value of the fuel injected vehicles (0. 501).
              Using the idle in drive test mode  of the Federal  Three-Mode,
however,  the roles are reversed and the  fuel-injected vehicles had an STE
value (0.401) 8 percent higher than that of the carbureted vehicles (0.370).
The idle in neutral results paralleled the idle in drive results.
              All STRR values were less than one, with no evident trends
   q
by fuel system type.
1.3.8.2       Carbon Monoxide Emission
              The results for CO are given in Table 1-61.  On the Federal
Short Cycle, the fuel injected vehicles achieved  a 27 percent higher STE
value (0.868) than the  carbureted vehicle STE value (0.682).
              In the idle in drive test mode of the Federal Three-Mode
ST, the fuel injected vehicles again had a higher STE value (0.781),  with
a 70 percent increase  over  the carbureted value (0.458).  The idle in neutral
results  were similar.
                                  1-110

-------
Table 1-60.  Comparison of Short Test HC Results by Fuel System Type;
       All Three Cities; Predicted Population; Bounded Errors of
                Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle


Federal 3-Mode





Fuel System
all
injection
carburetion
all

injection

carburetion

Test Mode



idle-in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
idle- in-drive
idle -in -neutral
Number
of
Vehicles
300
30
270
300

30

270

Parameter, %
EC
5.0
3.3
5.2
5.0
5.0
7.0
6.2
4.8
4.9
E
o
13.0
20.7
12.3
19.8
21.0
24.8
26.5
1.9.3
20.4
FF
18.7
20.7
18.4
11.9
10.7
16.6
14.9
11.3
10.2
STE
.589
.501
.600
.375
.338
.401
^.360
.370
.334
STRR
.748
.580
.769
.533 '
.495
.570
.510
.526
.493

-------
                  Table 1-61.  Comparison of Short Test CO Results by Fuel System Type;
                         'All Three Cities; Predicted Population; Bounded Errors of
                                  Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle



Federal 3-Mode





Fuel System
aU
injection
o
carburetion
all

injection

carburetion

Test Mode




idle -in -drive
idle -in -neutral
idle -in -drive
idle - in - neut r al
idle-in-drive
idle -in-neutral
Number
of
Vehicles
300
30

270
300

30

270

Parameter, %
E
5.0
8.9

4.4
5.0
5.0
13.2
10.7
4.1
4.3
E
0
16.7
5.8

17.4
24.4
26.6
9.6
12.0
25.8
28.0
FF
37.1
38.1

37.3
31.4
29.3
34.3
32.0
31.2
29.2
STE
.690
.868

.682
.562
.524
.781
.728
.547
.511
STRR
.783
1.071

.762
.652
.614
1.082
.971
.602
.586
ro

-------
               With regard to rejection rate, the fuel-injected vehicles had
much higher STRR values than the carbureted vehicles, because of higher
E  values.  On the Federal Short Cycle the fuel-injected STRR value was
1.071, and on the Federal Short Cycle (idle-in-drive mode) it was 1.082.
1.3.8.3        Oxides of Nitrogen Emission
               NO  results are  shown in Table 1-62. On the Federal Short
                 x
Cycle, the fuel-injected vehicles had a very low STE value (0.201); the
carbureted vehicle STE value (0.629)  was over three times  higher.
               This pattern is reversed using the high-speed mode of the
Federal Three-Mode ST, where the fuel injected STE value (0.426) is
42 percent higher than the carbureted value (0.301).  When  the low-speed
test mode is used, the situation reverses  again, and the carbureted
vehicles then have 1.3 percent higher STE  value.
               As in the case of CO noted  above,  the fuel-injected vehicles
had a considerably higher STRR value (1.480) on the Federal Three-Mode.
On the Federal Short-Cycle, however,  the fuel-injected vehicles were lower
in STRR value (0.494) than the carbureted vehicles (0.854).
1.3.8.4        Multiple Constituent Tests
               The three-constituent test results on a fuel system basis are
summarized in Table  1-63.  On the Federal Short Cycle, the fuel injected
vehicles had an STE value (0.876)  7 percent higher than the carbureted
vehicles (0.816).  This parallels the finding  for CO alone, as noted  above,
because CO was the principal pollutant causing the majority of FTP  failures
observed in the 300-car test fleet.
               Using the idle in drive,  idle in neutral,  or best combination
of test modes,  in all cases the fuel injected vehicles had higher  STE values
than the carbureted vehicles. For example,  with the idle in drive mode,
the  fuel injected STE value (0.812) was  17 percent higher than the carbureted
STE value (0.694).
                                  1-113

-------
Table 1-62.  Comparison of Short Test NOX Results by Fuel System Type;
        All Three Cities; Predicted Population; Bounded Errors of
                Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle


Federal 3-Mode





Fuel System
all
injection
carburetion
all
\
injection

carburetion

Test Mode



high
low
high
low
high
low
Number
of
Vehicles
300
30
270
300

30

270

Parameter, %
E
5.0
2.2
5. 1
5.0
5.0
5.3
6.9
4.6
4.8
E
o
8.4
6.1
8.4
14.6
14.8
2.9
5.6
15.7
15.7
FF
12.8
1.6
14.2
6.6
6.4
2.1
2.1
6.8
6.9
STE
.604
.201
.629
.310
.301
.426
.271
.301
.305
STRR
.840
.494
.854
.547
.538
1.480
1.169
.507
.518

-------
                     Table 1-63.  Comparison of Short Test Multiple Constituent Results by
                           Fuel System Type;  All Three Cities; Predicted Population;
                           Bounded Errors of Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle


Federal 3-Mode





Fuel System
all
injection
carburetion
all

injection

carburetion

Test Mode



idle-in-drive
idle -in-neutral
best
Number
of
Vehicles
300
30
270
300

30

270

Parameter, %
E
7.0
10.0
6.7
6.3
4.7
4.3
13.3
16.7
13.3
5.6
3.3
3.3
E
o
11.7
6.7
12.2
19.3
26.3
22.0
10.0
13.3
6.7
20.4
27.8
23.7
FF
53.7
46.7
54.4
46.0
39.0
43.3
43.3
40.0
46.7
46.3
38.9
43.0
STE
.821
.876
.816
.704
.597
.663
.812
.751
.875
.694
.583
.645
STRR
.928
1.062
.917
.801
.669
.729
1.062
1.064
1.124
.778
.633
.694
I
H*
h-*-
Ul
          (a)
            Best individual modes are HC in drive,
            CO in drive, and NOX at high speed.

-------
               With regard to rejection ratio, the fuel-injected vehicles had
higher STRR values than the carbureted vehicles on both STs, and exceeded
one in both cases.
1.3.9          Effect of Transmission Type
              The variation of short test effectiveness and rejection ratio
with transmission type is summarized in Tables 1-64 through 1-67 for the
Federal Short Cycle and the better individual test modes of the Federal
Three-Mode ST.
1.3.9.1      Hydrocarbon Emission
              As shown in Table  1-64, on the Federal Short Cycle,  the
vehicles with automatic transmissions achieved an STE value (0.674) which
was  40 percent higher than that achieved by vehicles with manual
transmissions (0.481).
              With the idle in drive mode of the Federal Three-Mode ST,
the above pattern remained unchanged; the automatic case received an STE
value (0.401) 35 percent higher than the manual case (0.296). These same
trends were  shown by the idle in neutral test mode, however the margin
of difference was reduced from 35 to 16 percent.
              All STRR values were less than one and of the same relative
level for both transmission types.
1.3.9.2       Carbon Monoxide Emission
              The results for CO are shown in Table 1-65.   On the Federal
Short Cycle,  the  vehicles with manual transmissions resulted in a STE
value of 0.752, 11 percent higher  than the STE value (0.678) for automatic
transmissions.
              This same relationship held on the idle in neutral mode of the
Federal Three-Mode ST (0.578 versus 0.512), although the  difference
increased from 11 percent to  1 3 percent.  Comparisons on the idle in drive
mode cannot  be made because the  vehicles with manual transmissions could
not be tested in this mode.
                                  1-116

-------
Table 1-64.  Comparison of Short Test HC Results by Transmission Type;
        All Three Cities; Predicted Population; Bounded Errors of
                Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle


Federal 3-Mode





Transmission
Type
all
automatic
manual
aU

automatic

manual

Test Mode



idle-in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle - in - neut ral
Number
of
Vehicles
300
229
71
300

229

71

Parameter, %
EC
5.0
5.1
4.8
5.0
5.0
4.9
4.9
5.4
5.6
E
o
13.0
11.4
18.9
19.8
21.0
18.2
19.8
25.6
25.3
FF
18.7
19.0
17.4
11.9
10.7
12.2
10.6
10.7
10.9
STE
.589
.624
.481
. 375
.338
.401
.348
.296"
.301
STRR
.748
.793
.639
.533
.495
.563
.510
.444
.456

-------
                  Table 1-65.  Comparison of Short Test CO Results by Transmission Type;
                          All Three Cities; Predicted Population;  Bounded Errors of
                                  Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle


Federal 3rMode





Transmission
Type
all
automatic
manual
all

automatic

manual

Test Mode



.idle- in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
300
229
71
300

229

71

Parameter, %
E
5.0
3<4
12.2
5.0
5.0
3.6
3,6
10.8
10.7
E
o
16.7
18.1
10,8
24.4
26.6
26.1
28.7
18.3
18.4
FF
37.1
38.0
32.8
31.4
29.3
32.7
30.2
25.4
25.3
STE
.690
.678
.752
.562
.524
.556
.512
.582
.578
STRR
.783
.738
1.032
.652
.614
.617
.574
.828
.824
oo

-------
         In the case of CO, the manual transmission equipped vehicles
had the higher STRR values and exceeded one (1.032) on the Federal Short
Cycle.
1.3.9.3      Oxides of Nitrogen Emission
              NO  results are  shown in Table 1-66.  On the Federal
Short Cycle, the STE value of the automatic transmission equipped vehicles
(0. 610) was 83 percent higher than that of the manual transmission vehicles
(0.334).
              When tested on the Federal Three-Mode ST,  however, the
vehicles with manual transmissions recorded higher STE values in both
the high and low  speed test modes.  In the high speed mode, the manual
case STE value (0.647) was over twice as large as the automatic case STE
value (0.251).  In the low speed mode, the higher STE value (0.496) of the
manual was 77 percent higher than the automatic transmission STE value
(0.281).
              On the Federal Short Cycle, the vehicles with manual
transmissions had lower STRR values than those equipped with automatic
transmissions.  On the Federal Three-Mode, this was reversed; the manual
transmission case was  higher by a factor or two or more and exceeded one
(1. 538) on the high-speed test mode .
1.3.9-4      Multiple Constituent Tests
              The three-constituent test results on a transmission type
basis are summarized in  Table  1-67.  In the Federal Short Cycle, the
vehicles with manual transmissions had an STE value (0.870) 7 percent
higher than that of the vehicles with automatic transmissions (0.813).
              With the idle in neutral and best combination of test modes
on the Federal Three-Mode ST, the manual transmission case had higher
STE values, as on the Federal Short Cycle.  The idle in neutral STE
advantage of the manual (0. 694) over the automatic (0. 575) was 21 percent.
                                  1-119

-------
                  Table 1-66.  Comparison of Short Test NOX Results by Transmission Type;
                         All Three Cities; Predicted Population; Bounded Errors of
                                  Commission Method (E  set at 5 percent)
Short Test

Federal Short Cycle


Federal 3-Mode



'7

Transmission
Type

all
automatic
manual
all

automatic

manual

Test Mode




high
low
high
low
high
low
Number
of
Vehicles

300
229
71
300

229

71

Parameter, %
E

5.0
5.6
1.0
5.0
5.0
3.4
4.4
10.6
6.0
E
0

8.4
9.4
6.3
14.6
14.8
18.0
17.3
4.2
6.0
FF

12.8
14.7
3.2
6.6
6.4
6.0
6.8
7.7
6.0
STE
C
.604
.610
.334
.310
.301
.251
.281
.647
.496
STRR

.840
.842
.442
.547
.538
.392
.465
1.538
1.00
CO
o

-------
          Table 1-67.  Comparison of Short Test Multiple Constituent Results by
                Transmission Type; All Three Cities; Predicted Population;
                Bounded Errors of Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle


Federal 3-Mode








Transmission
Type
all
automatic
manual
all


automatic


manual


Test Mode



idle-in-drive
idle -in-neutral
best
idle-in-drive
idle -in -neutral
best'a'
idle-in-drive
idle -in-neutral
best
Number
of
Vehicles
300
279
71 •
300


229


71


Parameter, %
E
7.0
7.0
7.0
6.3
4.7
4.3
5.7
2.6
2.2
8.4
11.3
11.3
E
o
11.7
13.1
7.0
19.3
26.3
22.0
20.5
29.7
24.4
15.5
15.5
14.1
FF
53.7
56.8
46.7
46.0
39.0
43.3
49.3
40.2
45.4
35.2
35.2
36.6
STE
.821
.813
.870
.704
.597
.663
.706
.575
.650
.694
.694
.722
STRR
.928
.913
1.00
.801
.669
.714
.788
.613
.682
.860
.917
.945
(a)
  Best individual modes are HC in drive,
  CO in drive, and NOX at high speed.

-------
              With regard to rejection ratio, the vehicles with manual
transmissions had higher STRR values on both STs,  and reached 1.0 on
the Federal Short Cycle.
1.3.10        Effect of Manufacturer
              The variation of short test effectiveness and rejection ratio
as a function of vehicle manufacturer is summarized in Tables 1-68 through
1-71 for the Federal Short Cycle and the better individual test modes of
the Federal Three-Mode ST.
1.3.10.1     Hydrocarbon Emission
              In the case of HC (Table 1-68) on the Federal Short Cycle, the
STE values of GM, Ford, and Chrysler were all similar (in 0.638 to 0.676
range).  AMC was considerably lower (0.491),  and the "others" category
(principally small imports) were intermediate in value (0.544).
              The above characteristics were not present with the idle in
drive test mode of the Federal Three-Mode ST.  Here,  the GM vehicles were
considerably higher (0.511) and Ford considerably lower (0.195) than the
fleet average STE value (0.375).  Chrysler, AMC, and "others" were very
similar (0.318 to 0.364) and near the fleet average.  The  idle in neutral
trends were similar to the idle in drive trends.
              With  regard to rejection ratio, the manufacturers were all
roughly comparable on the Federal Short Cycle except for Ford, with an
STRR value of 0.992.  On the Federal Short Cycle (idle  in drive),  Ford had
the lowest STRR value (0.286) and AMC had the highest  (1.00).  AMC had
had the lowest STRR value on the  Federal Short Cycle (0. 509).
1.3.10.2      Carbon Monoxide Emission
              The  results for CO are given in Table 1-69.  On the Federal
Short Cycle, GM,  Ford, and Chrysler vehicles (as in the  case of HC), all
had similar STE values (in 0.652 to 0.732 range).  AMC vehicles ha.d the
lowest STE (0.510) and the "others" category had the highest STE  (0.790).
                                  1-122

-------
                     Table  1-68.  Comparison of Short Test HC Results by Manufacturer;

                          All Three Cities; Predicted Population; Bounded Errors of
                                  Commission Method (E  set. at 5 percent)
Short Test
Federal Short Cycle





Federal 3-Mode











Manufacture r
all
GM
Ford
Chrysler
AMC
other
all

GM

Ford

Chrysler

AMC

other

Test Mode






idle- in -drive
idle -in-neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle- in -neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle - in -neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
300
99
61
36
18
86
300

99
•
61

36

18

86

Parameter, %
EC
5.0
3.9
8.8
3.6
0.4
5. 1
5.0
5.0
4.5
5.2
2.3
3.6
3.7
7.8
5.1
3.9
6.1
4.6
E
0
13.0
10.1
9.0
20.7
11.3
15.0
19.8
21.0
15.2
15.9
20.3
19.8
36.4
29.1
14.6
16.3
22.5
25.8
FF
18.7
21.0
16.2
36.5
10.9
17.9
11.9
10.7
15.9
15.2
4.9
5.5
20.8
17.6
7.5
5.8
10.5
7. 1
STE
.589
.676
.642
.638
.491
.544
.375
.338
.511
.489
.195
.217
.364
.377
.340
.263
.318
.216
STRR
.748
.801
.992
.701
.509
.699
.533
.495
.656
.656
.286
.360
.428
.448
0. 570
.439
.503
.356
I
t-^

to

-------
Table 1-69.  Comparison of Short Test CO Results by Manufacturer;
     All Three Cities; Predicted Population; Bounded Errors of
             Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle





Federal 3-Mode











Manufacturer
all
GM
Ford
Chrysler
AMC
other
all

GM

Ford

Chrysler

AMC

other

Test Mode






idle- in-drive
idle -in -neutral
idle -in -drive
idle - in -neutral
idle- in-drive
idle - in-neutral
idle -in-drive
idle- in-neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle -in-neutral
Number
of
Vehicles
300
99
61
36
18
86
300

99

61

36

18

86

Parameter, %
E
c
5.0
2.4
4.8
0.2
1.2
14.0
5.0
5.0
2.0
2.6
4.2
3.1
9.5
10. 1
2.5
1.0
12.7
12. 1
	 	
E
o
16.7
18.2
14.4
28. 1
14.7
8.9
24.4
26.6
26.7
28.9
28.3
31.7
8.2
8.4
14.2
14.3
16.6
20.3
FF
37.1
35.8
39.4
52.5
15.3
33.3
31.4
29.3
31. 1
28.9
29.3
26.0
71.8
71.7
15.7
15.6
25.6
21.9
STE
.690
.663
.732
.652
.510
.790
.562
.524
.538
.500
.509
.451
.898
.896
.525
.522
.607
.519
STRR
.783
.707
.822
.654
.550
1.121
.652
.614
. 573
.545
.582
.504
1.016
1.021
.609
.555
.908 '
.806

-------
              Using the idle in drive test mode of the Federal Three-Mode
ST, all vehicles except Chrysler haid STE values (0.509 to 0.607) reasonably
close to the fleet average (0.562).  The Chrys.ler vehicles had a much higher
STE value (0.898). The idle-in-neutral STE results paralleled the idle-in-
drive results .
              In the case of CO rejection ratio,  on the Federal Short Cycle
the major domestic manufacturers were similar in STRR value range; the
"others" category (principally small imported cars) exceeded an STRR value
of one (1.121).  On the Federal Three-Mode,  the "others" category had an
STRR value of slightly less than one, while Chrysler vehicles exceeded  one
(1.016 in idle-in-drive test mode).
1.3.10.3      Oxides  of Nitrogen Emission
              NO  results  are shown in Table 1-70.  Or. the  Federal Short
                 x
Cycle, the AMC (0.957)  and Chrysler (0.817) vehicles were considerably
above the  fleet average (0. 604).  Ford had the lowest  STE value (0.463),
followed by "others" .(0.492) and  GM (0.564).
              The NO versus manufacturer  trends for the high speed mode
                      j£
of the Federal Three-Mode ST are considerably different.  AMC, the highest
on the Federal Short Cycle, has the lowest STE value (0.087).  Only
Chrysler (0.349) and  "others" (0.467) are above the fleet average (0.310).
Ford (0.210) and GM (0.229) are both below the fleet average.
              The low speed  test mode results are similar only for  GM,
AMG,  and the "others" category.   Chrysler had a very high STE value
(0.755) in the low speed  mode, while Ford had a very low STE value (0.065)
in the low speed mode.
              In terms of NO  rejection  ratio, on the Federal Short Cycle
                             *•» •
all manufacturers were similar in STRR  value except for AMC which had
a value of 2.463.  On the Federal Three-Mode, AMC  was quite low in STRR
(0. 322 in low-speed test mode), while Chrysler recorded the highest STRR
value (2.129 on the low-speed mode).
                                  1-125

-------
Table 1-70.  Comparison of Short Test NOX Results by Manufacturer;
      All Three Cities; Predicted Population; Bounded Errors of
   —x          Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle





Federal 3-Mode











Manufacturer
all
GM
Ford
Chrysler
AMC
other
all

GM

Ford

Chrysler

AMQ

other

Test Mode






high
low
high
low
high
low
high
low
high
low
high
low
Number
of
Vehicles
300
99
61
36
18
86
300

99

61

36

18

86

Parameter, %
E
c
5.0
1.6
3.5
23.6
14.5
2.5
5.0
5.0
4.2
2.8
1.4
0.4
7. 1
22.4
0.2
4.2
7.5
5.4
E
o
8.4
8.8
16.1
2.9
I-4
7.9
14.6
14.8
15.5
15.3
23.7
28.1
10.6
4.0
28.7
24.0
7.6
7.9
FF
12.8
11.4
13.9
13.3
30.1
7.6
6.6
6.4
4.6
4.9
6.3
2.0
5.7
12.3
2.7
5.2
6.6
7.6
STE
.604
.564
.463
.817
.957
.492
.310
.302
.229
.243
.210
.065
.349
.755
.087
. 177
.467
.489
STRR
.840
.644
.580
2.278
1.416
. 652
.547
.538
.438
.381
.257
.078
.790
2. 129
.092
.322
.993
.839

-------
1.3.10.4      Multiple Constituent Tests
               The three-constituent test results for both STs on a manu-
facturer basis are summarized in Table  1-71.  In the case of the Federal
Short Cycle, AMC (0.916), GM (0.877),  and "others" (0.875) were above
the fleet average (0.821), whereas Ford  (0.700) and Chrysler (0.742) were
below  the fleet average STE value.
               These trends are changed using the idle in drive test mode
of the Federal Three-Mode ST (and the best combination of test modes,  as
well).  On this basis,  GM(0.800), Chrysler (0. 775), and "others" (0.729)
have the highest STE values, and Ford (0.500) and AMC (0.583) are the
lowest.  The idle  in neutral trends are similar to the idle in drive trends.
               With regard to rejection ratio,  AMC (1.165) and the "others"
category (1.041) both exceeded STRR values of one on the Federal Short
Cycle.  All manufacturers had STRR values less than one on the Federal
Three-Mode ST, with AMC and Ford having the  lower STRR  values.
1.3.11         Effect of Manufacturer and Engine Displacement Group
               The foregoing analyses indicated that, when the ST cut-points
were based on a 5 percent Ec rate for the total 300-vehicle pooled fleet,
there were possible inequities in terms of CO and NO  discrimination for
                                                   5C
the smaller vehicles with low displacement engines.  This was exemplified
by the higher  E and concomitant STRR values for the 150-CID and less
displacement  group, the 2500-lb and less weight group, the fuel injected
vehicles, and the manual transmission-equipped vehicles. The  small CID,
low vehicle weight, and manual transmission factors are all  typical attributes
of the small vehicle class.  Fuel injection, at its present degree of usage,
is also principally restricted to the small vehicle class.
               Therefore, in order to gain further insight into this phenome-
non, an additional contingency table analysis was made wherein the ST  cut-
points were determined for each CID class and were based on a  5 percent
Ec rate for that CID class alone, not the  pooled  300 vehicle fleet as in all
                                  1-127

-------
                   Table 1-71.  Comparison of Multiple Constituent Results by Manufacturer;
                          All Three Cities; Predicted Population; Bounded Errors of
                                    Commission Method (E   set at 5 percent)
Short Test
Federal Short Cycle





Federal 3 -Mode











Manufacturer
all
GM
Ford
Chrysler
AMC
other
all

GM

Ford

Chrysler


AMC

other
Test Mode






idle -in -drive
idle -in-neutral
best
idle-in-drive
idle-in-neutral
best
idle-in-drive
idle -in -neutral
best^
idle-in-drive
idle-in-neutral
best
idle-in-drive
idle -in-neutral
best(a)
idle-in-drive
idle - in - ne ut ral
best(a)
Number
of
Vehicles
300
99
61
36
18
86
300

99

61

36


18

86
Parameter, %
E
c
7.0
2.0
8.2
8.3
16.7
9.3
\
6.3
4.7-
4.3
7.1
1.0
1.0
1.6
1.6
1.6
5.6
5.6
5.6
5.6
5.6
0.0
9.3
10.5
10.5
E
0
11.7
8.1
19.7
22.2
5.6
7.0
19.3
26.3
22.0
13.1
20.2
20.2
32.8
37.7
31.2
19.4
19.4
19.4
27.8
50.0
44.4
15.1
23.3
14.0
FF
53.7
57.6
45.9
63.9
61.1
48.8
46.0
39.0
43.3
52.5
45.5
45.5
32.8
27.9
34.4
66.7
66.7
66.7
38.9
16.7
27.2
40.7
32.6
41.9
STE
.821
.877
.700
.742
.916
.875
.704
.597
.663
.800
.693
.693
.500
.425
.524
.775
.775
.775
.583
.250
.408
.729
.583
.749
STRR
.928
.907
.825
.839
1.165
1.041
.801
.669
.714
.909
.708
.708
.518
.450
.534
.840
.840
.840
.667
.334
.408
.896
.771
.937
00
         (a)
           Best individual modes are HC in drive, CO in drive, and NOX at high speed

-------
previously discussed results.  The results of this investigation are
reported below.
1.3.11.1       150 CID and Less Displacement Group
1.3.11.1.1    Hydrocarbon Emission
              Table 1-72 presents the HC results for the 95-vehicle 150 CID
and less group.  On the Federal Short Cycle,  all manufacturers except
Toyota (0.273) and VW (0.098) had STE values reasonably close to the
fleet average value of 0.595.  None had STRR values of one or above.
              On  the Federal Three-Mode ST (idle in drive test mode),
the STE values of  GM (0.575) and Honda (0.683) were well above the
fleet average (0.318).  Those of Toyota (0.127) were considerably below
the fleet average.   All STRR values were below one except Honda, which
had a value of 1 .037.
1.3.11.1.2    Carbon Monoxide Emission
              As  shown in Table  1-73, on the Federal Short'Cycle all
manufacturers had STE values reasonably close to the fleet average of
0.567 except Ford (0.070) and Toyota (0.251); Datsun was intermediate
in value (0.374).   None had STRR values of one or above.
              On  the Federal Three-Mode ST (idle in drive test mode),
Ford again had a very low STE value  (0.002) compared to the fleet average
of 0.411.  The other manufacturers had STE values near the fleet average
except for GM (0.806) and Datsun (0.722), who were  well above the average.
Again, no manufacturer  class had STRR  values exceeding one.
1.3.11.1.3    Oxides of Nitrogen Emission
              In terms of STE value  on the Federal Short Cycle (Table 1-74),
both Honda (0) and VW (0. 154) registered notable exceptions to the fleet
average value of 0.628.  With regard to rejection .ratio,  both Ford (2.039)
and VW (1.077) had STRR values  in excess of one.
              On  the Federal Three-Mode ST (high speed mode),  the STE
values of Ford (0.031),  Datsun (0.105), and Toyota (0.104) were all well
                                  1-129

-------
             Table  1-72.  Comparison of HC Results for  150 CID and Less Group; by Manufacturer;
                          All Three Cities; Predicted Population; Bounded Errors of
                                   Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode









Manufacturer
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

GM

Ford

Datsun

Toyota

Test Mode










idle- in-drive
idle - in -neutral
idle- in-drive
idle- in -neutral
idle -in -drive
idle -in -neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
95
12
6


8
9
21
6
33
95
95
12

6

8

9

Parameter, %
E
5.0
2.9
2.1


3.9
0.3
2.7
4.9
7.2
5.0
5.0
11.3
10.2
• 3.3
5.2
5.4
.9
2.8
1.5
E
0
13.0
13.4
5.4


6.7
31.5
18.9
7.4
10.3
21.9
23.4
14.4
13.6
13.3
12.5
30.3
44.9
37.8
40.2
FF
19.1
20.5
13.9


41.7
11.8
17.7
0.8
19.4
10.2 '
8.7
19.5
20.3
6.0
6.7
18. 1
3.6
5.5
3. 1
STE
.595
.605
.720


.862
.273
.484
.098
.653
.318
.271
.575
.599'
.311
.349
.374
.074
. 127
.072
STRR
.439
.690
.829


.942
.279
.557
.695
.896
.474
.427
.909
.900
.482.
.620
.486
.092
. 192
. 106
OO
o

-------
   Table 1-72.  Comparison of HC Results for 150 CID and Less Group; by Manufacturer;
                All Three Cities; Predicted Population; Bounded Errors of
                   Commission Method (E  set at 5 percent) (Continued)
Short Test
Federal 3-Mode
(continued)





Manufacturer
vw(a)

Honda

others


Test Mode
idle -in -drive
idle -in-neutral
idle-in-dr ive
idle-in-neutral
idle-in-drive
idle -in -neutral

Number
of
Vehicles
21

6

33


Parameter, %
E
c
4.2
4.6
2.9
5.6
4.2
3.7

E
o
23.4
22.3
2.6
1.5
21.9
25.1

FF
13.1
14.2
5.6
6.6
7.8
4.7

STE
.359
.389
.683
.815
.263
.158

STRR
.474
.515
1.037
1.488
.404
.282

(a)
  VW category includes VW,
  Audi, and Porsche

-------
             Table 1-73.  Comparison of CO Results for  150 CID and Less Group; by Manufacturer;
                         All Three Cities; Predicted Population; Bounded Errors of
                                   Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3 -Mode









Manufacturer
all
CM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

CM

Ford

Datsun

Toyota

Test Mode










idle-in-drive
idle -in -neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle - in -neutral
idle-in-drive
idle - in-neutral
Number
of
Vehicles
95
12
6


8
9
21
6
33
95

12

6

8

9

Parameter, %
E
5.0
0.1
0.0


0.4
3.7
4.9
0
6.8
5.0
5.0
6.7
6.6
.1
.06
10.9
4.0
4.7
3.9
E
0
19.4
36.9
23.8


33.1
25.6
24.5
0
10.7
26.4
28.8
15.4
15.7
25.6
25.6
14.7
39.5
21.6
29.0
FF
25.4
40.5
1.8


19.8
8.6
23.4
0
30.8
18.4
16.0
64. 1
63.8
0.05
0.03
38.2
13.3
12.6
5.2
STE
.567
.523
.070


.374
.251
.489
0
.742
.411
.357
.806
.803
.002
.001
.722
.252
.368
. 152
STRR
.679
.525
.070


.382
.360
.591
0
.906
.522
.469
.903
.886
.006
.004
.928
.328.
.506
.266
I
I-*-
OO

-------
             Table 1-73.  Comparison of CO Results for 150 CID and Less Group; by Manufacturer;
                          All Three Cities; Predicted Population; Bounded Errors of
                            Commission Method (E  set at 5 percent) (Continued)
Short Test
Federal 3-Mode
(continued)




Manufacturer
VW

Honda

others

Test Mode
idle-in-drive
idle- in -neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
21

6

33

Parameter, %
E
c
8.4
8.9
0.0
0.0
2.8
2.9
E
o
30.0
31.6
0.0
0.0
26.5
30.0
FF
18.3
16.7
0.0
0.0
15.0
11.5
STE
.379
.346
0
0
.361
.277
STRR
.552
.530
0
0
.429
.347
CO
CO

-------
              Table 1-74.  Comparison of NOX Results for 150 CID and Less Group; by Manufacturer;
                          All Three Cities; Predicted Population; Bounded Errors of
                                  Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode









Manufacture r
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

GM

Ford

Datsun

Toyota

Test Mode










high
low
high
low
high
low
high
low
high
low
Number
of
Vehicles
95
12
6


8
9
21
6
33
95

12

6

8

9

Parameter, %
E
c
5.0
3.9
19.2


2.9
0.7
1.2
0
1.7
5.0
5.0 '
9.9
11.0
.8
.01
27.2
.9
.34
3.8
'E
o
5.8
2.2
5.8


10.8
35.9
1. 1
0. 1
3.8
9.0
9.0
2.7
5.9
.12.5
12.9
23.8
21.0
57.0
21.7
FF
9.8
8.8
7.1


14.7
28.6
0.2
0
8.8
6.5
5.2
8.3
5.2
.4
.04
2.8
4.4
6.6
37.7
STE
.628
.800
.550


.577
.443
.154
0
.698
.419
.366
.755
.468
.031
.003
.105
.173
. 104
.635
STRR
.949
1.155
2.039


.690
.454
1.077
0
.833
.742
.789
1.655
1.459
. 093
.004
1. 128
.209
. 109
.698
I
h^
(JO

-------
             Table 1-74.  Comparison of NOX Results for 150 CID and Less Group; by Manufacturer;
                          All Three Cities; Predicted Population; Bounded Errors of
                             Commission Method (E set at 5 percent) (Continued)
Short Test
Federal 3 -Mode
(continued)




Manufacturer
VW

Honda

others

Test Mode
high
low
high
low
high
low
Number
of
Vehicles
21

6

33

Parameter, %
EC
.6
10.3
.7
.8
2.4
2.4
E
o
.9
.3
. 1
.O'l
9.8
8.6
FF
.4
.9
.03
0.0
2.8
4.0
STE
.308
.750
.231
0
.222
.318
STRR
.769
9.333
5.615
8.00
.413
.508
oo
Ul

-------
below the fleet average of 0.419, whereas the GM value (0.755) was quite
high. In:terms of ST rejection ratio, GM (1.655),  Datsun (1 . 128), and
Honda (5. 615) all exceeded an STRR value of one.
1.3.11.1.4    Multiple Constituent Tests
              As  shown in Table 1-75, on the Federal Short Cycle, all
manufacturers were reasonably close to the fleet average STE value of
0.763 except Honda, which registered a value of zero. In terms of STRR
value,  all were below one except for Datsun (1 .200).
              On  the Federal Three-Mode  ST (idle in drive test mode),
Ford (0.250) and Toyota (0.285) had the lowest STE values.  GM (1.00),
Honda (1.00), and VW (0.700) had the highest values.  With  regard to
rejection ratio,  GM (1.111) exceeded an STRR value of 1.00.
1.3.11.2     . 151 to  259 CID Group
1.3.11.2.1    Hydrocarbon Emission
              Table 1-76 presents the  HC results  for the 54-vehicle 151 to
259 CID group.  On the  Federal Short Cycle,  Chrysler (0.241) was well
below the fleet average  STE value of 0.487, and  Ford  (0.671) was well
above the average. Ford (1.043) also exceeded the STRR value of one.
              On  the Federal Three-Mode  ST (idle in drive test mode),
the STE values of  GM (0.703) and Chrysler (0.707) were  quite  high.
Ford (0.062), AMC (0.170),  and the others were all very low.  Only
Chrysler (1.063) exceeded an STRR value of one.
1.3.11.2.2    Carbon Monoxide Emission
              As  shown in Table 1-77, on  the Federal Short Cycle  all
manufacturers had STE values reasonably close  to the fleet average of
0.737.  Only Ford (1.051) slightly exceeded an STRR value  of one.
              On the Federal Three-Mode ST (idle in drive test mode),
General Motors (0.781) and Chrysler (0.803) had the highest STE values;
the rest fell in the 0.360 to 0.551 range.  Only Chrysler  (1.025) exceeded
an STRR value of one.
                                  1-136

-------
             Table 1-75.  Comparison of Multiple Constituent Results for  150 CID and Less Group;
                  by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                                   Commission Method (E  set at  5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode








Manufacturer
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

GM

Ford

Datsun


Test Mode










idle-in-drive
idle -in -neutral
best(a)
idle- in-drive
idle -in -neutral
best
-------
             Table  1-75.  Comparison of Multiple  Constituent Results for 150 CID and Less Group;
                  by Manufacturer; All Three  Cities; Predicted Population; Bounded Errors of
                             Commission Method (E  set at 5 percent) (Continued)
Short Test
Federal 3 -Mode
(continued)











Manufacturer
Toyota


VW


Honda



others


Test Mode
idle-in-drive
idle -in-neutral
best*
idle-in-drive
idle-in-neutral
best
idle-in-drive
idle -in-neutral
/a\
best* '
idle-in-drive
idle -in -neutral
best
-------
                       Table  1-76.   Comparison of HC Results for 151 to 259 CID Group;
                  by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                                   Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3 -Mode







Manufacturer
all
CM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
others(a)
all

GM

Ford

Chrysler

Test Mode










idle-in-drive
idle -in -neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle-in-neutral
Number
of
Vehicles
54
13
13
7
12




7
9
54

13

13

7

Parameter, %
Ec
5.0
1.8
6.0
1.5
1.9




4.9
3.6
5.0
5.0
6.5
8.4
1.0
.44
6.8
3.2
E
o
14.0
13.1
5.3
14.5
16.9




0.7
21.1
18.8
20.0
9.4
14.4
15.1
15.5
5.6
7.7
FF
13.3
18.6
10.8
4.6
21.0




9.1
12.0
8.4
7.2
22.3
17.3
1.0
1.0
13.5
11.6
STE
.487
.587
.671
.241
.554




.929
.363
.309
.265
.703
.546
.062
.061
.707
.601
STRR
.670
, 44
1.043
.319
.604




1.556
.471
.493
.449
.909
.811
.124
.087
1.063
.767
I
l-ub
u>
          (a)
'This category includes 1 Datsun and 1 VW

-------
             Table 1-76.  Comparison of HC Results for  151 to 259 CID Group;
        by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                   Commission Method (E  set at 5 percent) (Continued)
Short Test
Federal 3-Mode
(continued)




Manufacturer
AMC

others(a)

others

Test Mode
idle-in-drive
idle -in-neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle-in-neutral
Number
of
Vehicles
12

9

7

Parameter, %
E
c
5.3
4.1
1.7
1.7
3.1
3.1
E
o
27.3
28.2
31.1
31.5
8.3
9.2
FF
5.6
4.7
2.0
1.6
1.5
.7
STE
.170
.143
.060
.048
.153
.071
STRR
.331
.267
.111
.100
.469
.384
(a)
  This .category includes 1 Datsun and 1 VW

-------
              Table 1-77.  Comparison of CO Results for  151 to 259 CID Group;
         by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                         Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3 -Mode







Manufacturer
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
others(a)
all

GM

Ford

Chrysler

Test Mode










idle-in-drive
idle -in-neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
54
13
13
7
12




7
9
54

13

13

7

Parameter, %
EC
5.0
0.5
8.7
2.6
4.1




4.6
5.5
5.0
5.0
7.4
9.0 .
1.5
1.7
15.3
12.3
E
o
14.6
20.4
6.4
30.8
16.4




5.6
7.8
24.6
32.3
14. 1
30.8
36.2
41. 1
13.6
3. 1
FF
41.0
44. 1
38.6
38.3
34.5




48.4
47.9
34.8
27.1
50.4
33.6
20.4
15.5
55.5
68.5
STE
.737
.684
.858
.554
.678




.896
.860
.586
.456
.781
.522
.360
.274
.803
.957
STRR
.827
.691
1..051
.592
.758




.981
.959
.670
.540
.896
.661
.387
.304
1.025
1. 128
(a)
  This category includes  1 Datsun and 1 VW

-------
                      Table  1-77.  Comparison of CO Results for 151 to 259 CID Group;
                 by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                             Commission Method  (E   set at 5 percent) (Continued)
Short Test
Federal 3 -Mode
(continued)




Manufacturer
AMC

other s(a)

others

Test Mode
idle-in-drive
idle -in-neutral
idle-in-drive
idle -in -neutral
idle-in-drive
idle -in -neutral
Number
of
Vehicles
12

9

7

Parameter, %
EC
5.3
1.1
1.0
. 1
.8
.03
E
o
26.1
30.3
25.0
35.0
24.5
35.5
FF
24.8
20.6
30.7
20.7
29.5
18.4
STE
.487
.405
.551
.372
.546
.341
STRR
.591
.426
.569
.373
.561
.342
ro
         (a)
           This category includes
           1 Datsun and 1 VW.

-------
 1.3.11.2.3   Oxides of Nitrogen Emission
              In terms of STE value on the Federal Short Cycle (Table 1-78),
 Ford (0. 359) had the lowest value and AMC (0.939) had the highest.  AMC
 (1.429) also exceeded an STRR value of one.
              On the Federal Three-Mode ST,  the low speed test mode was
 superior for the fleet averaged 151 to 259  CID vehicle group (STE = 0.445
 versus  0.230 for the high speed test mode).  In this test mode,  GM (0.648)
 and Chrysler (0.819) had the highest  STE values,  whereas Ford (0.244) and
 the "others" category (0. 038 to  0.122) were quite  low in STE value.  Both
 GM (1.032) and Chrysler (1.023) slightly exceeded an STRR'value of one.
 1.3.11.2.4   Multiple Constituent Tests
              As  shown in Table 1-79, on the Federal Short Cycle, all
 manufacturers were reasonably close to the fleet average STE value  of
 0.812 except Chrysler, which registered a value of 0.334.  In terms of STRR
 value, both Ford and AMC had values of 1.0.
              On the Federal Three-Mode ST (idle in drive test mode), GM
(1.0)  and Chrysler (0.833) had the highest  STE values.  Ford (0. 333), AMC
 (0. 500), and the "others" category (0.400  to 0. 500) were below the fleet
 average value of 0.648.  Chrysler had an STRR value of 1. 0 and GM  recorded
 a value  of 1 .223.
 1.3.11.3      260 CID and Greater Group
 1.3.11.3.1    Hydrocarbon Emission
              Table 1-80 presents the HC results for the 151-vehicle
260 CID and greater group. On the Federal Short Cycle, all manufacturers
(domestic only) were fairly close together in terms of STE value.  None
exceeded an STRR value of one.
              On the Federal Three-Mode (idle in drive test mode),  the STE
value of Ford (0.299) was well below that of GM (0.531)  and Chrysler (0.520).
No manufacturer exceeded an  STRR value of one.
                                  1-143

-------
          Table 1-78.  Comparison of NOX Results for 151 to 259 CID Group;
     by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                     Commission Method (E   set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode







Manufacturer
all
CM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
others(a)
all

GM

Ford

Chrysler

Test Mode










high
lo'w
high
low
high
low
high
low
Number
of
Vehicles
54
13
13
7
12




7
9
54

13

13

7

Parameter, %
E
5.0
0.5
6.2
5.0
13.6




0.6
0.6
5.0
5.0
5.8
8.4
2.9
1.5
7.4
7.0
E
o
7.5
10.8
8.4
6.9
1.7




7.5
7.6
15.4
11.1
20.1
7.7
9.0
9.9
20.7
6.2
FF
12.5
11.0
4.7
27.4
26.0




3.0
3.9
4.6
8.9
1.8
14.2
4.1
3.2
13.5
28.0
STE
.625
.505
.359
.799
.939




.286
.339
.230
.445
.082
.648
.313
. 244
.395
.819
STRR
.875
.528
.832
.945
U429




.343
.391
.480
.695.
.347
1.032
.534
.359
.611
1.023
This category includes 1 Datsun and 1 VW

-------
            Table 1-78.  Comparison of NOX Results for 151 to 259 CID Group;
      by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                  Commission Method (E   set at 5 percent) (Continued)
Short Test
Federal 3-Mode
(continued)




Manufacturer
AMC

others fa)

others

Test Mode
high
low
high
low
high
low
Number
of
Vehicles
12

9

7

Parameter, %
EC
2.0
5.2
4.8
1.3
2.0
.4
E
0
20.0
17.6
7.8
10. 1
9.0
10. 1
FF
7.7
8.2
3.7
1.4
1.5
.4
STE
.278'
.318 '
.322
.122
.143
.038
STRR
.350
.536
.739
.234
.333
.076
(a)
  This category includes
  1 Datsun and 1 VW.

-------
    Table 1-79.  Comparison of Multiple Constituent Results for 151 to 259 CID Group;
        by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                        Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode



x

Manufacturer
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
others(a)
all

GM

Ford

Test Mode










idle-in-drive
idle -in -neutral
best
idle-in-drive
idle -in -neutral
best
-------
    Table 1-79.  Comparison of Multiple Constituent Results for 151 to 259 CID Group;
        by Manufacturer; All Three Cities; Predicted Population;  Bounded Errors of
                   Commission Method (E  set at 5 percent) (Continued)
Short Test
Federal 3 -Mode
(continued)












Manufacturer
Chrysler



AMC


others


others
.


Test Mode
idle-in-drive
idle - in - neut ral
/M
best(DI
idle-in-drive
idle -in -neut ral
bestfb)
idle-in-drive
idle - in - ne ut ral
best(b)
idle-in-drive
idle -in -neut ral
(b)
best1 '
Number
of
Vehicles
7



12


9


7



Parameter, %
Ec
14.3
14.3

14.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

0.0
E
0
14.3
14.3

14.3
41.7
66.7
41.7
33.3
44.4
33.3
42.9
57. 1

42.9
FF
71.4
71.4

71.4
41.7
16.7
41.7
33.3
22.2
33.3
28.6
14.3

28.6
STE
.833
.833

.833
.500
.200
.500
.500
.333
.500
.400
.200

.400
STRR
1.000
1.000

1.000
.500
.200
.500
.500
.333
.500
.400
.200

.400
(a)
(b)
This category includes
1 Datsun and 1 VW

Best individual modes are HC
  and CO in drive and NOx at
  high speed.

-------
                   Table 1-80.  Comparison of HC Results for 260 CID and Greater Group;
                 by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                                 Commission Melhod (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode









Manufacturer
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

GM

Ford

Chrysler

AMC

Test Mode










idle-in-drlve
Ldle-in-neutral
idle-in-drive
idle- in-neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle -in-neutral
idle-in-drive
idle -in-neutral
Number
of
Vehicles
151
74
42
29
6





151

74

42

29

6

Parameter, %
E
5.0
4.0
2.8
4.0
0





5.0
5.0
4.0
4,2
3.9
5.1
5.9
6.2
0.0
0,0
E
o
12.5
8.9
11.4
20.3 '
0





18.3
19.6
14.4
15.9
20.6
21.2
29.2
31.1
0.0
0.0
FF
20.4
21.8
18.0
40.5
0





14.6
13.3
16.3
14.8
8.8
8.3
31.6
29.7
0.0
r\ n
w . vy
STE
.620
.710
.612
.666
0





.444
.404
.531
.482
.299
.281
.520
.488
0
A
\J
STRR
.772
.840
.707
.731
0





.596
.556
.661
.619
.432
.454
.617
.590
0
0
00

-------
1.3.11.3.2    Carbon Monoxide Emission
              As shown in Table  1-81, on the Federal Short Cycle, the STE
values of all manufacturers were  reasonably well grouped, with Chrysler
the highest at STE =  0.920.  Ford (1.028) slightly exceeded an STRR value
of one.
              On the Federal Three-Mode (idle in drive test mode), Chrysler
again had the highest STE value (0.937) and GM had the lowest (0. 631).  Both
Ford (1.058) and Chrysler (1.041) slightly exceeded an STRR value of one.
1.3.11.3.3    Oxides  of Nitrogen Emission
              In terms of STE value on the Federal Short Cycle (Table  1-82),
AMC had the highest value (0.942) and Ford the lowest (0.385).  Both AMC
(1.1 59)  and  Chrysler (2. 554) exceed an STRR value of one.
              On the Federal Three-Mode ST, the high speed test mode was
superior for the fleet average.  In this test mode,  Chrysler had the highest
STE value (0.479), AMC the lowest (0.86); GM (0.288) and Ford (0.335) were
near the fleet average value of 0. 343. Only Chrysler  (1.488) had an STRR
value in excess of one.
1.3.11.3.4    Multiple Constituent Tests
              As shown in Table 1-83, on the Federal Short Cycle all manu-
facturers were reasonably close to the fleet average STE value of 0.904.
All manufacturers except Chrysler exceeded an STRR value of one and AMC
had an STRR value of 2.0.
              On the Federal Three-Mode ST (idle in drive test mode), all
manufacturers except AMC were closely grouped in STE value near the  fleet
average (0.721); AMC recorded an STE value of one.  No manufacturer
exceeded an STRR value of one.
                                   1-149

-------
                   Table 1-81.  Comparison of CO Results for 260 CID and Greater Group;
                 by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                                 Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode









Manuf actur e r
all
CM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

GM

Ford

Chrysler

AMC

Test Mode










idle-in-drive
tdle-in-neutral
idle-in-drive
idle -in-neutral
idle-in-drive
idle - in -neutral
idle-in-drive
idle-in-neutral
idle-in-drive
idle-in-neutral
Number
of
Vehicles
151
74
42
29
6





151

74

42

29

6

Parameter, %
E
c
5.0
5.2
10.1
2.0
0.4





5.0
5.0
3.0
3.5
13.0
10.2
8.6
8.9
.2
.4
E
o
13.0
12.2
8.4
6.6
0





17.8
18.8
20. 1
21.2
9.4
11.6
5.2
5.5
0.0
0.0
FF
44.2
38.1
52.0
76.4
0





42. 1
41.3
34.4
33.3
52.1
49.8
77.8
77.4
0.0
0.0
STE
.773
.758
.861
.920
0





.703
.687
.631
.611
.847
.811
.937
.934
0
0
STRR
.860
.861
1.028
.945
0





.790
.770
.686
,675
1.058
.977
1.041
1.041
0
0
(Jl
o

-------
                   Table 1-82.  Comparison of NO% Results for 260 CID and Greater Group;
                 by Manufacturer; All Three Cities; Predicted Population; Bounded Errors of
                                 Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode









Manufacturer
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

GM

Ford

Chrysler

AMC

Test Mode










high
low
high
low
high
low
high
low
high
low
Number
of
Vehicles
151
74
42
29
6





151

74

42

29

6

Parameter, %
E
c
5.0
1.1
0.6
20.9
9.0





5.0
5.0
4.0
1.0
3.0
.2
12.2
28.2
0.0
2.8
E
o
11.5
10.9
23.5
2.1
2.4





16.3
19.4
15.3
17.6
25.4
35.4
6.3
2. 1
34.9
33.4
FF
13.3
10.7
14.7
10.0
38.9





8.5
5.4
6.2
3.9
12.8
2.8
5.8
10.0
3.3
4.8
STE
.536
.495
.385
.827
.942





.343
.218
.288
. 181
.335
.073
.479
.826
.086
.126
STRR
.738
.546
.401
2.554
1.159





.544
.419
.474
.228
.414
.078
1.488
3.157
.086
.199
I
h-^
Ul

-------
            Table 1-83.  Comparison of Multiple Constituent Results for 260 CID and Greater Group;
                   by Manufacturer; All Three Cities; Predicted Population; Bonded Errors of
                                    Commission Method (E  set at 5 percent)
Short Test
Federal Short Cycle









Federal 3-Mode





Manufacturer
all
GM
Ford
Chrysler
AMC
Datsun
Toyota
VW
Honda
others
all

GM

Ford

Test Mode










idle-in-drive
idle -in-neutral
best
idle-in-drive
idle -in-neutral
best(a)
idle-in-drive
idle -in-neutral
best
Number
of
Vehicles
151
74
42
29
6





151

74

42

Parameter, %
EC
8.0
4.1
14.3
3.5
33.3





6.0
5.3
4.0
4.0
1.3
0.0
11.9
14.3
11.9
E
o
6.6
4. 1
7.1
13.8
0.0





19.2
17.9
19.9
17.6
17.6
23. '0
19.0
16.7
16.7
FF
62.3
59.5
64.3
72.4
33.3





49.7
51.0
49.0
46.0
46.0
40.5
52.4
54.8
54.8
STE
.904
.937
.901
.840
1.00





.721
.740
.711
.723
.723
.638
.734
.766
.766
STRR
1.020
1.000
1.101
.881
2.000





.808
.817
.769
.786
.744
.638
.900
.966
.933
ro
          (a)
            Best individual modes are HC in drive, CO in drive, and NO  at high speed

-------
            Table 1-83.  Comparison of Multiple Constituent Results for 260  CID and Greater Group;
                  by Manufacturer; All Three Cities; Predicted Population; Bonded Errors of
                             Commission Method (E  set at 5 percent) (Continued)
                                                    c
Short Test
Federal 3-Mode
(continued)




Manufacturer
Chrysler


AMC


Test Mode
idle -in -drive
idle -in -neutral
best
idle-in-drive
idle-in-neutral
best
Number
of
Vehicles
29


6


Parameter, %
E
c
3.4
3.4
3.4
0.0
0.0
0.0
E
o
20.7
17.2
13.8
33.3
33.3
33.3
FF
65.5
69.0
72.4
0.0
0.0
0.0
STE
.760
.800
.840
0
0
0
STRR
.799
.840
.879
0
0
0
Ul
         (a)
           Best individual modes are HC and
           CO in drive, and NOX at high speed.

-------
1.3.12         Comparison of Short Test Cut-Points
1.3.12.1       By City
               Table 1-84 summarizes the variation of FSC cut-point value
for each emission constituent required to result in a 5 percent E  rate in
each individual city and for the pooled 300-vehicle fleet.  The pooled fleet
values were those used in the contingency table analyses of Sections 1.2.1
through 1.2.10.  The cut-spoint values for all three pollutants are very
similar for Chicago and Houston; Phoenix cut-points are higher in each case,
with a corresponding impact on the pooled fleet average.
               Table 1-85 summarizes similar cut-point values for the Federal
Three-Mode ST (idle in drive for HC and CO,  and high speed mode for NO ).
In contrast to  the Federal Short Cycle, here the HC and CO cut-points for
Phoenix are lower than the pooled fleet average and the Chicago cut-points are
much higher.
1.3.12.2       By Engine Displacement Group
              Table 1-86 summarizes the variation of FSC cut-point value
for each emission constituent required to result in a 5 percent E  rate for
each individual CID  group and for the pooled 300-vehicle fleet.  The CID
group values shown  were those used in the contingency table analyses  of
Section 1.2.11.  As  can be noted, the CO cut-point for the 150 CID and less
group is considerably higher than the pooled fleet average.  Thus,  in the
analyses made using pooled fleet cut-points (Sections 1.2.1 through 1.2.10),
the small vehicle class should have been and was consistently identified as
having a high E  value and correspondingly high STRR values.
              Table 1-87 summarizes similar cut-point values for the
Federal Three-Mode ST (idle in drive for HC and CO,  and high speed  mode
for NO  ).  Here both CO and NO  cut-points for the 150 CID and less group
      Jt.                       J£
are higher than fleet average values.  Again, in the analyses of Sections
1.2.1 through 1.2.10, this CID group was consistently identified with  high
E  and STRR values because of these cut-point distribution characteristics.
                                 1-154

-------
             Table 1-84.  Comparison of Federal Short Cycle
                 Cut-Points by City (E  set at 5  percent)
City
All
Chicago
Houston
Phoenix
No. of
Vehicles
300
100
100
100
ST Cut Points and Standard Errors
(gm/mi)
HC
CP
1.32
1.24
1.24
1.44
SE
0. 145
0.283
0. 181
0.267
CO
CP
8.40
7.99
7.49
9.48
SE
0.866
A . 240
1.327
1.916
NO
X
CP
2.62
2.54
2.56
2.77
SE
0.219
0.355
0.360
0.412
(a)
(b)
CP = cut point.
SE = standard error.
                                 1-155

-------
             Table 1-85.  Comparison of Federal Three-Mode
                Cut-Points by City (Ec set at 5 percent)
City
All
Chicago
Houston
Phoenix
No. of
Vehicles
300
100
100
100
ST Cut-Points and Standard Errors
(ppm) (a)
HC
CP
26
78
27
31
CO
CP
7,081
10, 145
6, 175
5,494
SE
1, 251
2, 627
2,432
1, 510
NOX
CP
2,265
2,026
2',172
2,439
SE
243
410
393(e)
379
(a)
(b)
Idle in drive test mode
(c)
(d)

(e)
High speed test mode
Cut-point

Standard error

The best mode for NOX for Houston was found to be the low speed
mode.   The cut-point and standard error for NOX low  speed are
1500 ppm and  372 ppm, respectively.
                                 1-156

-------
        Table 1-86.  Comparison of Federal Short Cycle Cut-Points
               by Engine Displacement (Ec set at 5 percent)

Engine
Displacement
Group
All
150 CID and
Less
151 to 259 CID
260 CID and
Greater

No. of
Vehicles
300
95

54
151

ST Cut-Points and Standard Errors
(gm/mi)
HC
CP
1.32
1.28

1.27
1.35

SE^
0.145
0.234

0.464
0. 192

CO
CP
8.40
13.10

6.24
5.61

SE
0.866
1.920

1. 124
0.821

NOX
CP
2.62
2.37

2.62
2.78

SE
0.219
0.439

0.440
0.285

(a)
(b)
CP = cut point.

SE  = standard error.
                                   1-157

-------
        Table 1-87.  Comparison of Federal Three-Mode Cut-Points
               by Engine Displacement (E  set at 5 percent)


Engine
Displacement
Group
All Three
Cities
150 CID and
Less

151 to 259 CID
260 CID and
Greater



No. of
Vehicles
300

95


54
151

ST Cut -Points and Standard Errors
(ppm)
HC

. »
cp(c>
184

219


213
151

/ j \
SEV
26

57


79
26

co(a)


CP
7,081

18, 382


7,063
2,655


SE
1, 251

5, 863


2,488
607

NO (b)
X


CP
2,265

2,805

(e)
1,985
1,986


SE
243

665

/ \
446
252

(a)
(b).
(c)
(d)
(e)
Idle in drive test mode
High speed test mode
Cut-point
Standard error
The best mode for NOX on this engine displacement was found to
be the  low speed mode.   The cut-point and standard error for the
low speed NOX are  1396 ppm and 423 ppm, respectively.
                                  1-158

-------
1.3.13        Effect of Cut-Point Selection Method on STE.
               Values
               As discussed in Sections 1.3.1 through 1.3.11 and elsewhere,
the use of cut-points based on pooled fleet characteristics can result in
higher than average rejection ratios (STRRs) for the smaller cars with
small engines, particularly for CO amd NO  pollutants.  Since it might prove
                                         jC
beneficial, for instance, to select cut-points as a function of engine size
(i.e.,  CID group) to minimize such potential inequities, an analysis was made
whereby each of the three  CID-groups  was tested with cut-points selected to
result in a 5 percent E  rate for each CID-group, rather than 5 percent for
the pooled fleet irrespective of CID group.   These results are compared with
the results of Sections 1. 3.1 through 1.3.4 in Table 1-88 for the Federal Short
Cycle and in Table 1-89 for the  Federal Three-Mode.
               In the case  of the Federal Short Cycle,  use of the CID-grouping
for cut-point selection resulted in very minor improvements in the STE value
for HC, CO, and multiple  constituents and a very slight decrease in NO
                                                                     jf,
STE value.  In the case of the Federal Three-Mode, HC, CO, and NO  (high
                                                                   3C
speed only)  STE values were slightly improved and the multiple constituents
and NO  (low speed only) STE values were slightly decreased.  The use of
CID-grouped cut-points would have the advantage of avoiding possible
inequities associated with  excessive ST failures of small vehicles.
                                  1-159

-------
         Table 1-88.  Federal Short Cycle STE Comparison;
             5 Percent EC Fleet Average vs 5 Percent EC
                    for Each of Three CID Groups
Pollutant
HC
CO
NO
X
Multiple , .
Constituents
Short Test Effectiveness (STE)*a*
5% Ec Fleet
Average
0.589
0.690
0.604
0.821
5% for Each of Three
CID Groups(b)
0.592
0.711
0.573
0.847
(a)
  STE =
           %FF
(b).
         %Eo + %FF

  '150 CID and less
   151 to 259 CID
   260 and greater CID

(c)Car fails if any of the HC, CO,  or
  exceed cut-points for each pollutant.
                                       measurements
                                 1-160

-------
           Table 1-89.  Federal Three-Mode STE Comparison;
              5 Percent Ec  Fleet Average vs  5 Percent E
                      for Each of Three CID Groups
Pollutant and
Test Mode
HC Idle in Drive
Idle in Neutral
C° Idle in Drive
Idle in Neutral
NO
xHigh Speed
Low Speed
Multiple Constituents
Idle in Drive
Idle in Neutral
Best Combination
Short Test Effectiveness (STE)^
5% Ec Fleet
Average
0.375
0.338
0. 562
0. 524
0.310
0. 301
0.704
0. 597
0.663
5% for Each of Three
CID Groups(b)
0. 383
0.339
0.605
0.557
0.341
0.291
0.653
0. 592
0.648
(a)

(b)
'STE =
           %FF
(c)
        %Eo + %FF
150 CID and less
151 to 259 CID
 260 and greater CID

Car fails if any of HC, CO,  or NO  measurements exceed
cut-points for each pollutant.     x
                                 1-161

-------
                           2. INTRODUCTION
2. 1           BACKGROUND AND OBJECTIVES

              With regard to compliance by vehicles and engines in actual

use with the certification emission standards established for a vehicle at

the time of its  manufacture,  the  Clean Air Act  of  1970 stipulates in
Section 207(b):

              "If the Administrator determines that

    (i)        there are available  testing methods and procedures to
              ascertain whether,  when in actual use  throughout its
              useful life	  each vehicle and engine to which
              regulations ....  apply complies with the emission
              standards of such regulations,

    (ii)       such methods and procedures are in accordance with
              good engineering practices,  and

    (iii)      such methods and procedures are reasonably capable of
              being correlated with tests conducted under section 206(a)(l),
              then - -
                    (1)   he shall  establish such methods and procedures
                         by regulation,  and

                    (2)   at such time as he determines that inspection
                         facilities or equipment are  available for purposes
                         of carrying out testing methods and procedures
                         established under paragraph (1), he shall pre-
                         scribe regulations which shall require manufac-
                         turers to warrant the  emission control device or
                         system of each new motor vehicle or new motor
                         vehicle engine . . . for its useful life."

Thus, there are the essential requirements of "availability," "conformance

with good engineering practices," and "reasonable correlation with certifica-

tion test procedures" which must be met prior to the  promulgation of regu-
lations which impose the in-use warranty provisions of Section 207(b) upon

the motor vehicle manufacturers.
                                    2-1

-------
               The states of New York and New Jersey have developed short
emission tests for potential use  in inspection/maintenance (I/M) programs in
their areas.   The Clayton Manufacturing Company also developed a short
test procedure for use in I/M programs.  More recently, the EPA has devel-
oped short tests  similar to  those of New-York, New Jersey, and Clayton.
Thus, there are  a number of tests "available" to determine the exhaust emis-
sions of in-use vehicles; these test methods and procedures "conform with
good engineering practices" in that they utilize well-recognized emission-
measurement equipment and techniques.
               These tests  are "short" in duration (approximately 3 to 5 min-
utes) in order to (a) minimize the inconvenience of the motoring public (and
thereby maximize cooperation),  and  (b) minimize capital costs  of inspection
stations  by maximizing the  number of vehicles a given facility could test.
They have been structured  for "simplicity" in order to (a) reduce the  poten-
                                                                      Ca
tial  for procedural errors,  and (b) to reduce test costs.   As a result,  all
such tests require that the  vehicle be tested in a "hot" condition; i. e. , at its
normal operating temperature.
               There remains the requirement to demonstrate "reasonable
correlation with  certification test procedures;" i.e., with the Federal Test
Procedure (FTP) used in the certification of new motor vehicles.  An initial
evaluation of the degree of  "correlation" between five  short tests (STs)  and
the FTP was previously conducted (Ref.  1) utilizing test data from a 40-car
catalyst-equipped experimental vehicle fleet and a 147-car  in-use 1974
model year vehicle fleet.   Based on the results of that evaluation, two of
the five STs were selected  by EPA for further test  and evaluation in a
three-city, 300-car,  1975 model year vehicle fleet. The present study was
performed,  therefore,  to analyze the resultant emission data from the short
tests and FTP tests of this  300-car fleet to determine  the degree of       *
"correlation"  between the ST and FTP test results.  The same computational
and  statistical analysis techniques developed in the initial evaluation (Ref. 1)
were used in the present study.
                                   2-2

-------
2.2           STUDY SCOPE
              The basis for the analyses was ST and FTP data from three
100-vehicle fleets located in the following cities:
                        Chicago,  Illinois
                        Houston,  Texas
                        Pho.enix,  Arizona
              Each of the 1975 model year vehicles was tested by the FTP
and the following STs:
                        Federal Short Cycle
                        Federal Three Mode
For the undiluted  exhaust type test (Federal Three Mode), garage-type instru-
ments were used to record HC and CO measurements.  Garage-type instru-
ments were used to simulate the working environment of a typical automotive
garage or a large-scale vehicle testing station.  All the NO  readings were
                                                         X.
made with laboratory analyzers due to the unavailability of an appropriate
garage-type NO  instrument.
               X,
2.3           METHOD OF APPROACH
              The primary thrust  of the work performed under this contract
was statistical in nature.  Two  complementing methods were employed to
assess Section 207(b) correlation --a conventional  correlation analysis and
a contingency table analysis.  The  conventional correlation analysis addresses
the question of direct relatability between the ST and the FTP by examining
the relationships present in the data.  The results are of great usefulness in
indicating the extent to which each  ST tends to track the FTP.  The  con-
tingency table analysis addresses the relatability of ST and FTP on  a pass-
or-fail level.  Each data point is examined,  and a determination is made as
to whether the auto passed or failed the  FTP and passed or failed the ST.
Thus,  errors of commission (E  ),  errors of omission (E ), correct passes
by each test (PP), and correct fails by each test (FF) are identified. Hence,
the technique allows for  the study of the tradeoffs between errors and correct
identifications.
                                   2-3

-------
              The conventional correlation analysis,  being purely an
analysis of the data,  does not permit policy decision as a variable or param-
eter.  Contingency table analysis, on the other hand,  permits the integration
of policy decision in that it provides for  the determination of the ST pass/fail
cut-points.  One important method  reflecting impact to policy is that of the
method of bounded errors of commission.  In this scheme, limits are  set on
the maximum permissible percentage of errors of commission,  and the  ST
cut-points are selected to yield minimum errors of omission within this
constraint.  This analysis permits  a direct answer to  the question,  "For a
given permissible level of errors of commission, what level  of errors of
omission is associated with a given test, and with what impact on air quality
(inferred from the percentage of FF and E  vehicles)?"
              These two methods of analysis,  each representing different
interpretations of Section 207(b) correlation, were applied to both the full
300-vehicle fleet and to each 100-vehicle city fleet.
                                   2-4

-------
        3.  TEST CHARACTERISTICS AND FLEET COMPOSITION
              In this program, two short tests (STs) and the 1975 Federal
Test Procedure (FTP) were performed on three 100-vehicle test fleets.
This section defines the two STs and describes the composition of the test
fleets.
3. 1           SHORT TESTS
3.1.1         General
              Two classes of STs were involved, and these may be cate-
gorized as (1) modal or volumetric and (2) as driving trace or  CVS.  Both
sets of nomenclature are used in this report, depending upon the aspect of
the test structure that is  pertinent to the discussion.  In the modal tests, the
test technician operates the vehicle on a dynamometer at a fixed  vehicle
speed and dynamometer load, or at idle.  The vehicle tailpipe  exhaust is
sampled directly, and the concentration of each pollutant is measured and
recorded in percent,  or in parts per million, of the undiluted exhaust.  One
modal  ST was used, the Federal Three Mode.   The Federal Three Mode ST
has high-speed, low-speed,  and idle modes.  The idle mode can  be conducted
in two  ways, (1) idle with transmission in drive (automatic transmission
only),  or (2) idle in neutral (automatic and manual transmissions).
              For the second class of ST, the test technician drives the; car
on the  dynamometer in accordance with  a prescribed  driving pattern on a
driving trace.  The vehicle exhaust is diluted by the  constant volume sam-
pling (CVS) procedure,  and a single sample bag of diluted exhaust is  collected
for  the whole ST.  The dilute sample is analyzed and  the results  usually
expressed  in grams per mile.  This procedure requires the same equipment,
sampling procedure,  and analytical equipment as the  Federal Test Procedure
(FTP) used in the certification of new vehicles.  The  difference is that the
driving trace for the ST is much shorter and simpler.  The CVS-type ST
used was the Federal Short Cycle.
                                   3-1

-------
              Both classes of ST involved approximately two or three minutes
of driving time on the dynamometer,  and both STs were performed with the
engine at its normal operating temperature; i.e., hot tests.
              The HC and CO content of the exhaust gas in the volumetric
tests was  measured by garage instruments, a lower-cost, lower-accuracy
and precision instrument of the type currently in use by many automotive
service stations for routine diagnostic work.  The structure of each  test is
given below.
3.1.2         Short Test Definition
3.1.2.1
Federal Three Mode
              The Federal Three Mode uses dynamometer loadings simulating
the average power that occurs  at the appropriate speeds in the FTP where
the vehicle is accelerating (decelerations are not included).  The speed and
dynamometer loadings for the high-speed, low-speed,  and idle test modes
are shown below.
Vehicle
Weight
Class, Ib
Up to
2500
2501 to
3500
3501 to
4500
Above
4500
Transmission
In lower gear
for 30-mph
test (3rd gear)
Drive or
high gear
Drive or
high gear
Drive or
high gear
High Speed
Mode
Speed,
mph
50
50

50

50

Load,
hp
21
26

31

36

Low Speed
Mode
Speed,
mph
30
30

30

30

Load,
hp
9
12

15

18

Idle Mode
Automatic
transmission
in neutral






                                   3-2

-------
3.1.2.2       Federal Short Cycle
               The Federal Short Cycle was derived from the FTP.
Accelerations and decelerations are representative of those encountered in
the FTP, and average speed is  nearly the same as the three-bag FTP driving
cycle (21.70 mph and 21.27 mph, respectively).
               This is a nine-mode, 125-sec CVS test that follows the driving
                                                            >[s
schedule shown below and plotted in Figure 57 in the Appendix.
                        Mode            Time in Mode, sec
                0 - 16 mph acceleration          6
               16-29 mph acceleration         23
                   29 mph cruise               10
               29 - 37 mph acceleration         18
               37 - 42 mph acceleration          4. 5
              42 - 37 mph deceleration          2.5
               37 - 20 mph deceleration         32
              20-0  mph deceleration          7. 5
              Idle                              21.5
                                              125.0
The test does not include engine startup or shutdown. The dynamometer
loadings follow the  procedure as required for the FTP.
3.2           TEST FLEET COMPOSITION
3.2.1         Types of Cars
              All 300 vehicles were 1975 model year vehicles representing
a typical cross-section of that model year with regard to manufacturer
(domestic and foreign), vehicle  size (weight),  engine  size (displacement: and
number of cylinders),  fuel system type, transmission type, and emission
control system type.  Tables 1 through 13 in the Appendix list cross-
tabulations of specific features of the vehicle fleet in the following context.
 Figure and table numbers  refer to figures and tables contained in the
 Appendix.
                                   3-3

-------
                                                          Table
                    Cross-Tabulation Feature            Number*
         No. of cylinders vs engine displacement              1
         Inertia test weight ys engine displacement            2
         Emission control system vs engine displacement      3
         Inertia test weight vs emission control system        4
         Fuel system vs emission control system              5
         Fuel system vs inertia test weight                   6
         Fuel system vs engine displacement                  7
         Transmission vs  engine  displacement                8
         Transmission vs  inertia test weight                  9
         Manufacturer vs engine displacement (all 3 cities)   10
         Manufacturer vs engine displacement (Chicago)      11
         Manufacturer vs engine displacement (Houston)      12
         Manufacturer vs engine displacement (Phoenix)      13

Engine displacement was subdivided into three  sizes:
         a.     150 cubic inches displacement (CID) or less
         b.     151 to 259 CID
         c.     260 or more CID
Inertia test weight breakdown was limited to four categories:
         a.     2500 Ib or less
         b.     2501 Ib to 3500 Ib
         c.     3501 Ib to 4500 Ib
         d.     4501 Ib or greater
               Emission control  system description was limited to the use
or non-use of oxidation catalysts  and/or  secondary air injection.
               Fuel system type description refers to  the use of either
carburetion or fuel injection.
 Table numbers refer to tables contained in the Appendix.
                                   3-4

-------
              Transmission type refers to use of manual versus automatic
transmissions.
3.2.2         Prior Use
              At the time of receipt of these vehicles by the testing
laboratory (Automotive Testing Laboratories, Inc.),  the odometer readings
ranged from less than 100 to 44, 000, with an average of 8, 390.  Specific
vehicle use patterns, in terms of city driving vs highway driving,  were not
quantified.  The type and extent of vehicle and emission  system maintenance
were similarly undefined. It was assumed that the use and state-of-
maintenance characteristics of this large test sample (300 vehicles from
three different locations) would be adequately representative of  the 1975
model year vehicle population.
                                   3-5

-------
     4.  STATISTICAL SCREENING AND CORRELATION ANALYSES
4. 1           DISCUSSION OF METHODOLOGY
              In studying the degree of correlation that exists between a
particular short test (ST)  and the Federal Test Procedure (FTP), a major
concern is the variability  in the correlation with vehicle population charac-
teristics.  That is,  it is important to know which classes of vehicles, if any,
show inadequate ST/FTP correlations.  Thus,  correlations on individual
groupings of the data were assessed.  In addition to  assessing differences
in correlation, differences in group means and standard deviation were also
reviewed.  Thus,  in presenting the results of the analyses, means,  standard
deviations, and correlation coefficients were computed for each group of
interest.  Statistical tests of significance were performed as  discussed
below.
4.1.1         Conventional Correlation Analysis
              A conventional correlation analysis includes the calculation
of the sample correlation  coefficient r,  and an  a-percent confidence interval
for the population correlation coefficient p,  on  paired observations.  Letting
(x.,  y.)i = l,	,  N denote the observations, r is  defined by

                            N
                               (x. - Mx)(y. - My)
                        * = -	s-s	                     (4-
                                     x y

where M , S  and M ,  S  are the mean and standard deviation of the
        xx       Y   y
observations x.  and y., respectively. An a-percent confidence interval is
given by (r-, r+), where the probability that the interval covers  p is cc/100.
For the 95 percent interval used in this study
                                  4-1

-------
                                    (**—}
                                    \    N/NVT /
r±  =  tanhfzi  A'7U  }                     (4-2)
where z  = 1/2 ln(l + r/l-r),  (Fisher's Z statistic, Ref. 2).
               The sample correlation coefficient is used as the prime
quantitative measure of relatability.  The closer r  is to 1, the better the
relation.  A lack of relationship is indicated by r = 0.  Negative  r  indicates
an inverse relation between the observations,  i. e. , if one observation  is
high, the other is low, and vice versa. The confidence interval is viewed
as reflecting the sensitivity of the calculations to the data.  The wider  the
interval,  the less predictable is the correlation coefficient and, hence,
the relatability.
4.1.2          Statistical Tests of Significance
               The ST/FTP correlation was classified as statistically
deficient if the 95 percent confidence interval,  given by equation (4-2), con-
tains the value zero.  Statistically,  the computed value of r  cannot be  dis-
tinguished from zero due to the existing variability in the data.
               To test for statistical differences in group means,  a classical
F-test for equality of group means was performed  (Ref. 3).  Both the com-
puted F-value and the level of significance were reported.  The level of
significance can be interpreted as the probability of making an error when
deciding  to reject the hypothesis of  equality.
               To test for statistical differences in group standard devia-
tions,  Bartlett's  test for homogeneity of group variances was performed,
(Ref. 3).  Both the computed  statistic and the  corresponding level of signifi-
cance were  reported.
               Tables* 20a, 20b, and 21 provide excellent examples to
illustrate the procedure. Table 20a shows the FTP means and  standard
deviations for each pollutant on each city group.  In examining these means
and standard deviations,  differences in the values from group to group are
*
 Table numbers refer to tables contained in the Appendix.
                                   4-2

-------
quite noticeable.  The results of F-tests and Bartlett's tests are shown in
Table 20b (note that the significance tests were also performed on the
logarithms of the FTP data).  Thus, the differences in mean CO would be
attributed to statistical variations,  unless one is willing to accept a 20 per-
cent chance of error in attributing the differences to differences in the cities.
However, significant differences in variance do occur for HC and CO, and
these differences can be attributed to systematic differences in the data  from
Chicago, Houston, and Phoenix.  These findings are repeated with varying
levels of significance for an analysis of the logarithms of the data.
               Table 21  shows the ST/FTP  constituent correlation coefficients
for both STs using the pooled fleet of 300 vehicles.  The coefficients were
computed for each constituent and the logarithm of each constituent.  The
NO  correlation for the Federal Three-Mode idle in neutral (0. 11)  and in
   x
drive (0. 12)  are statistically deficient and are noted as .deficient with an
asterisk.  All the other coefficients  in Table 21 are statistically significant,
which means that they are statistically distinguishable from zero with 95 per-
cent confidence.  Whether the significant coefficients are high enough to be
of practical value is a judgmental matter.
4.2            RESULTS FOR ALL THREE CITIES COMBINED
               Preliminary analyses were made of the ST data.  Means and
variances were tested for homogeneity between cities and between engine
displacement groups.   Tables 14 and 15 give the results for the Federal
Short Cycle by city and engine displacement,  respectively.  The Federal
Three-Mode results are shown in Tables 16 and 17 for city and engine dis-
           >,
placement grouping,  respectively.
               The effect of vehicle mileage was also investigated.   The
vehicles were grouped into those with 4000  miles  or less and those with
more than 4000 miles.   Tables 18 and 19  show significant differences
between the groups in their  FTP means and standard deviation,  and their
ST/FTP correlation coefficients, respectively.
                                    4-3

-------
               The results of an analysis of city effects are given in
Tables 20, 21, 22, 23, 24.   The differences in mean and standard deviations
are addressed in Table 20,  whereas the correlation coefficients are reported
in Tables 21  through 24 for the pooled fleet (300 vehicles) and the individual
city fleets (100 vehicles each).
               An analysis to determine the effects of engine displacement,
inertia test weight,  emission control system,  fuel system,  transmission
type,  and manufacturer was performed on the pooled fleet of 300 vehicles.
The table numbers for both STs correspond to the following statistics
spectrum:
For the Effects of
Engine Displacement
Inertia Test Weight
Emission Control System
Fuel System
Transmission Type
Manufacturer
Table Number
FTP Mean and
Standard Deviations
25a
27a
29a
31a
33a
35a
Tests of
Significance
25b
27b
29b
31b
33b
35b
ST/FTP
Correlation
Coefficients
26
28
30
32
34
36
              Additionally the ST/FTP correlation coefficients on each
manufacturer's vehicles within an engine displacement group were computed.
These results for both STs are shown in Tables 37,  38,  and 39 for the 150
CID and less, 151 to 259 CID,  and 260 CID and greater engine groups,
respectively.
4.3           RESULTS FOR INDIVIDUAL, CITY FLEETS
              An analysis to determine the effects of engine displacement,
inertia test weight,  emission control system, fuel system,  transmission
type,  and manufacturer was performed on the individual city fleets of
                                    4-4

-------
100 vehicles each.  The Chicago fleet numbers,  for both STs, correspond
to the following statistics spectrum:
For the Effects of
Engine Displacement
Inertia Test Weight
Emission Control System
Fuel System
Transmission Type
Manufacturer
Table Number
FTP Mean and
Standard Deviations
40a
42 a
44 a
46a
48a
50a
Tests of
Significance
40b
42b
44b
46b
48b
50b
ST/FTP
Correlation
Coefficients
41
43
45
47
49
51
For the Houston fleet the Table correspondence is:
For the Effects of
Engine Displacement
Inertia Test Weight
Emission Control System
Fuel System
Transmission Type
Manufacturer
Table Number
FTP Mean and
Standard Deviations
52a
54 a
56a
58a
60a
62a
Tests of
Significance
52b
54b
56b
58b
60b
62b
ST/FTP
Correlation
Coefficients
53
55
57
59
61
63
                                   4-5

-------
For Phoenix the table number correspondence is given by:
For the Effects of
Engine Displacement
Inertia Test Weight
Emission Control System
Fuel System
Transmission Type
Manufacturer
Table Number
FTP Mean and
Standard Deviations
64 a
66a
68a
70a
72a
74a
Tests of
Significance
64b
66b
68b
7 Ob
72b
74b
ST/FTP
Correlation
Coefficients
65
67
69
71
73
75
4.4
DISCUSSION OF RESULTS
              The HC Federal Short Cycle/FTP correlation coefficient
varies considerably between cities.  It is 0.88 in Chicago, 0.91 in Houston,
and 0.45 in Phoenix.  This non-uniformity in the values of the correlation
coefficient has been attributed in part to city differences in the FTP measure-
ments.  Vehicles 5520 and 5524 in Phoenix had extraordinarily high (relative
to remaining data) HC accumulation in bag 1 of the FTP, which is the cold
start portion of the FTP.  This resulted in a high FTP composite.  These
vehicles had comparatively low Federal Short Cycle readings for the high
FTP values.  Although these vehicles were identified as having anomalous
readings,  they were not discarded for any of the  subsequent analyses.
                                    4-6

-------
                   5.  CONTINGENCY TABLE ANALYSES
5. 1            DISCUSSION OF METHODOLOGY
               The contingency table analysis technique was used to establish
the  short test (ST) pass-fail levels for each pollutant.  The contingency table
is defined in Table 133*, along with its associated parameters.  A  pictorial
demonstration of its application to a given data set is shown in Figure 58.
It can be  seen that, for a given data set, part of the analysis is concerned
with the  criteria used to select the ST cut-points.  In this regard, the
bounded  errors of commission method was used extensively to establish
trends for the variations in E  ,  E ,  FF,  and PP.  In this method,  the ST
cut-points are  selected to minimize E  while holding the E  below a speci-
fied level.  It thus permits a direct answer to the question,  "For a given
permissible level of E , what level of E  is  associated with the ST, and
with what impact on air quality (inferred from number of FF and E  vehi-
cles)? "  This method  is illustrated in Figure 59.   The policy decision is
the  maximum allowable E .
                         c
               The Federal Test Procedure (FTP) levels used throughout
the  analyses were 1. 5 grams/mile for HC, 15. 0 grams/mile for CO, and
3. 10 grams/mile for NO .
               With regard to  procedural technique,  a  bivariate normal or
log-normal  distribution model was fitted to a particular data set by incor-
porating the correlation coefficient,  mean values,  and standard deviations
of the data set.  The bivariate normal distribution is shown in Figure 60.
The mathematical format of the log-normal distribution is identical to that
shown in Figure 60 except the statistics and variables  are in terms of the
natural logarithm of the measurements.  Figure 61 shows the pertinent
probability  equations used for predicting the table  entries.
 Table numbers refer to tables contained in the Appendix.
                                     5-1

-------
               The ST cut-points were determined by using the model as
determined from the data and by solving the equation in Figure 62.  Once
the cut-points have been  established, the predicted table entries for each
pollutant can be computed using the formulas  shown in  Figure_ 61.
               Also determined in,the contingency table analysis were  the
short test effectiveness and the short test rejection ratio.  ST effectiveness
is defined as:
         em  t,  ,.         C^TP        % FF for the short test
         ST effectiveness = STE =	   	
                                  % FTP failures in same population   (5-1)

                                     % FF
                                  % FF + % E
                                             o

Thus,  on this basis,  the ST is always less effective than the FTP, in pro-
portion to the percent of errors of omission (E  ) associated with a given
ST, and the short test effectiveness  (STE) value is a relative measure of
the impact on air quality from ST implementation.
               The short test rejection  ratio (STRR), is defined as:

                                    % E  +  % FF
                           STRR =  % E' +  % FF                       <5-2>
The denominator of this ratio (% E  + % FF) represents the percent of the
vehicle population actually failing the FTP.  The numerator (% E  + % FF)
represents the percent of the- vehicle population failing the short test.  Thus,
the resultant ratio  (STRR) is an indication to flag those situations where the
implementation of a short test or test mode could result in a  greater number
of vehicles being failed than should be failed based upon FTP standards.
Whereas the STE value is a  measure of the relative impact of the use of
the ST on air quality,  the short test rejection ratio (STRR) value (which
                                    5-2

-------
includes consideration of those E  vehicles incorrectly identified by the ST
as failing the FTP) can be considered a relative measure of the "fairness"
of the ST to the vehicle manufacturer or vehicle size or class.  For exam-
ple,  only when the STRR value exceeds  one would the manufacturer be
required to implement warranty procedures on a greater number of vehi-
cles than he should have based upon the observed FTP failure rate.
               In addition to analyzing each pollutant individually, an analy-
sis was made for three-constituent tests, using the  contingency table
approach.   In a three-constituent test, a car fails the ST if any of its HC,
CO,  and NO measurements exceed the previously determined cut-points.
            5C
This test is applicable to the Federal Short Cycle and the individual modes
of the Federal Three-Mode.  A car fails the ST if any one of the modes
fails on its  three-constituent tests. A model for predicting population
results was not available; hence the actual data was  used along with pre-
established cut-points to count the number of vehicles in each cell of the
contingency table.
               The computational procedure followed in determining the
three constituent tests is shown in Figure 63.  Note  that the cut-point
selection policy is applied at the pollutant level and  not at the multiple-
constituent  test level.  For example,  the percent E  is bounded for indi-
                                                 c
vidual pollutants in the method of bounded errors of commission, and this
bound can possibly be exceeded on a three-constituent test.  In forming the
three-constituent test contingency table, the following definitions apply:
         Correctly passed (PP):      Car passes the ST and the FTP
         Correctly failed (FF):       Car fails  the  ST  and the FTP
         Error of  commission (E ):   Car fails the  ST  and passes the FTP
                                C
         Error of  omission (E ):      Car passes the ST and fails the FTP
where FTP or  ST  failure occurs if any one of the test constituents exceeds
its respective cut-points.  A car  is counted once in  forming the table and
falls into one,  and only one, of the above categories.
                                   5-3

-------
               The "best" three-constituent test was formed by selecting
those individual test modes for HC,  CO, and NOX for which the correlation
coefficient was maximum.  A typical example might by HC and CO at idle and
NO  at the high speed mode.
5. 2            VARIANCE ESTIMATES
5. 2. 1          Short Test Cut-Points
               The problem of estimating the ST cut-point, for a fixed FTP
level,  is analogous to estimating the quantiles  of a distribution function
(Ref. 4).  Thus,  the  large sample standard deviation is given by
                        dp(y)
                         dy
v:
                         JL_            /v(PF -
N
                              y = LS                                 (5-3)
where
          PF = probability of passing the FTP
         p(y) = probability of an error of commission for ST cut-point
               set at value y
           Y = upper bound on probability of errors of commission
           N = sample size or number of cars in the data set
         LS  = true cut-point for the population
As PF, p(y), and LS . are unknown,  they can only be approximated from the
data.  LS  is,  of course,  approximated by the cut-point estimated from the
data.  PF is estimated on the  percent passed by FTP divided by 100, p(y)
is taken to be the locus of E  versus cut-point and dp(y)/dy|LS  is taken
as the derivative of the E   versus cut-point curve evaluated at the cut-point
of interest (LS ).  Equation (5-3) will be used to discuss variability of the
predicted population.
                                   5-4

-------
              For a fixed FTP level,  the standard deviation of the estimated
cut-point can be independently controlled by increasing the sample  size.
Once the sample size is fixed, this standard deviation varies inversely with
the magnitude of the  derivative of the E  versus cut-point curve.   Thus, in
                                      c
regions where the curve is steep,  the variability of the predictions will be
less than in regions where the curve is flat. Figure  64 illustrates  the eflect
of the cut-point uncertainty on the  other  computed quantities of E  and FF.
It shows that the uncertainty in the predicted results  increases with decreasing
errors of commission bounds.
              Standard deviations were  calculated for each ST cut-point com-
puted and are displayed herein as necessary.
5. 2. 2         Three-Constituent Test Percentages
              For fixed ST and FTP cut-points,  the  cell counts  in a 2 x 2
contingency table are binomially distributed when the observations  are
independent (Ref.  5). Since the  ST cut-points are computed from data prior
to forming the contingency table, there is statistical  dependence between the
ST cut-points and the resulting table.  Hence, the binomial distribution will
be an approximation  to the true distribution.  Thus,  the approximate  standard
deviation is
                               X     -                              (5-4,
where
         X =  cell count in percent
         N =  total table count
              This procedure was used to calculate the standard error of the
estimates for the  three-constituent test results reported herewithin.  Table
104 in the Appendix lists  the standard errors as a function of sample  size
and cell percentage.
                                    5-5

-------
5.3             SPECIFIC ANALYSES AND THEIR RESULTS
5.3.1           Analyses with Short Test Cut-Points Based
                Upon 300 Cars
                Usingthe 300 car data set, ST cut-points for HC, CO, and
NO  were determined for various rates of commission errors.  The differ-
   x
ences between various stratum within the data set were assessed by predic-
ting the contingency entries for the strata using the above pre-established
cut-points .  The effects of varying engine displacement, inertia test weight,
emission controls, carburetion, transmission, and manufacturer were
investigated in this manner.
                The results are presented in  graphical and tabular  form in
the Appendix and are referenced in the discussions to follow. The results
are presented first by individual constituent,  followed by the results of multi-
ple'constituent tests, and, finally, the results indicating the effects of engine
displacement, inertia weight,  emission systems,  carburetion, transmission,
and manufacturer are given.
5.3.1.1         Hydrocarbon Emissions
                The variation of E  , E ,  and  FF as a function of HC cut-
                                  c'   o
point was determined for each ST.  The results for the pooled fleet, using
the Federal Short Cycle and the Federal Three-Mode, are shown graphically
in Figures  1 and 17 to indicate the general nature  of the trade-offs available.
                Shown in Figures 4, 7, and 10 are the Federal Short Cycle
results for each individual city fleet (100 vehicles each).  The corresponding
Federal Three-Mode results are displayed in  Figures 20,  23, and 26.  Note,
however, that the cut-points used in these figures were determined by using
the pooled fleet of 300 vehicles.
                Table 76 summarizes the results  for both STs for the pooled
fleet and individual city fleets  at the specific E  analysis value of 5 percent.
This value was used in the analysis to  set HC  ST cut-points based on the total
pooled fleet population.  These same numerical cut-points were then used
                                    5-6

-------
in subsequent contingency analyses on an individual city and test mode
basis.  As a consequence, the resultant actual E  values  for the individual
cities or test modes vary from the 5 percent E  value for the pooled fleet
                                             c
due to differences in observed emission characteristics between the pooled
fleet and the individual city fleets.  Also shown in Table 76 are the effective-
ness values (STE).
5.3.1.2-      '  Carbon Monoxide Emission                \i
                The variation of E  , E  ,  and FF for the  pooled fleet as a
function of CO cut-point are shown for the Federal Short  Cycle and the
Federal Three-Mode in Figures  2 and 18, respectively.  The individual city
fleet results are shown in Figures 5,  8, and 11 for the Federal Short Cycle
and Figures  21, 24, and 27 for the Federal  Three-Mode.   Table 77  sum-
marizes the  results for both STs for the pooled fleet and  individual city fleets
at the E analysis value of 5 percent.
5.3.1.3        Oxides of Nitrogen Emission
                The pooled fleet variations  of E , E   and FF as a function
                    r                         co
of NO   cut-point are shown in Figures 3 and 19 for the Federal Short Cycle
     X.
and the  Federal Three-Mode,  respectively.  The results  for the individual
city fleets are  shown in Figures  6, 9, and 12 for the Federal Short Cycle and
Figures 22,  25 and 28 for the Federal Three-Mode.  Table 78 summarizes
the results for both ST for the pooled fleet and individual  city fleets at the
E  analysis value of 5 percent.
5.3.1.4        Multiple  Constituent Tests
                The variation of actual E  ,  E  , and FF as a function of
                                        c   o
predicted E  are shown for the three-constituent Federal Short Cycle in
Figures 13,  14, 15, and 16 for the pooled fleet and individual city fleets,
respectively.  The corresponding Federal Three-Mode idle in drive mode
and best mode three-constituent  results are shown in Figures 29, 30, 31,
and 32 for the pooled fleet and individual city fleets respectively. The best
                                    5-7

-------
mode refers to the combination of the best individual test modes: the
idle-in-drive for HC and CO discrimination and the high speed for NO  .
  /..'...                               •                     jt
Table 79 shows the three-constituent results,for both ST for the pooled and
individual city fleets at the E  analysis value of 5  percent.
          • . .;'•,.•         ' •  • - C •             '    •
5. 3. 1. 5       Effects of Vehicle  Characteristics
              The variation of E , E  , and FF with  engine displacement,
inertia test  weight,  emission control system, fuel system,  transmission
and manufacturer were individually determined for the E  analysis value of
5 percent.   The results  are presented  for both ST in  Tables 80 through 103
and correspond to the following constituent spectrum:
For the Effects of
Engine Displacement
Inertia Test Weight
Emission Control System
Fuel System
Transmission Type
Manufacturer
Constituent Table Number
HC
84
80
88
92
96
100
CO
85
81
89
93
97
101
NOX
86
82
90
94
98
102
Three
Constituents
87
83
91
95
99
103
5.3.1.6       Comparison of Short Test Cut-Points
              Table 105 summarizes the variation of the Federal Short
Cycle cut-point  value for each emission constituent required to result in a
5 percent E  rate in each individual city for the pooled 300-vehicle fleet.
Table 106 summarizes similar cut-point variations for the Federal Three-
Mode ST (idle in drive for HC  and CO, and high speed in NO ).
5.3.2
Analyses with Short Test Cut-Points Based Upon
              Engine Displacement
              An additional contingency table analysis was made wherein
the ST cut-points were determined for each CID class and were based on a
                                    5-8

-------
5 percent E  rate for the CID class alone, not the pooled 300-vehicle fleet
as in the results presented in Section 5. 3. 1.  The results of this investiga-
tion are reported below.
5.3. 2. 1
150 CID and Less Displacement Group
Tables 107 through 110 present the results of this analysis
for both ST while showing the variation in E ,  E  and FF by manufacturer.
Tables  121 through 124 show the variation in E ,  E  and FF by city.   The
analysis was made for an E  value of 5 percent and the tables correspond to
                          c
the following constituent spectrum;
For the Effects of
Manufacturer
City
Constituent Table Number
HC
107
121
CO
108
122
NOX
109
123
Three
Constituents
110
124
5.3.2.2
151 to 259 CID Group
              Tables 111  through  114 show the variation in E ,  E  , and FF
by manufacturer for both ST while  Tables 125 through 128  show the results
by city.  The following correspondence shows the table numbers for the
constituent spectrum:
For the Effects of
Manufacturer
City
Constituent Table Number
HC
111
125
CO
112
126
NOX
113
127
Three
Constituents
114
128
                                    5-9

-------
5.3.2.3
              260 and Greater CID Group
              Tables 115 through 118 and Tables 129 through 132 show the
                    and FF for both ST as a function of manufacturer  and
city according to the following correspondence:
variation of E ,  E
             c   o
For the Effects of
Manufacturer
City
Constituent Table Number
HC
115
129
CO
116
130
NOX
117
131
Three
Constituents
118
132
5.3.2.4       Comparison of Short Test Cut-Points
              Tables 119 and 120 summarize the variation of cut-point
value for each emission constituent required to result in a 5 percent E  rate
in individual displacement group and for the pooled 300-vehicle fleet for the
Federal Short Cycle and the Federal Three-Mode, respectively.
                                  5-10

-------
          6.   207(b) IMPLEMENTATION SIMULATION ANALYSES
6. 1           OBJECTIVES AND APPROACH
              The objective of this effort was to assess the effectiveness
of a 207(b) program in reducing total emissions from light-duty vehicles
as characterized by the data collected on the 300 1975 model year cars.
Engineering judgment was to be used where firm data was lacking.
              A computer program was developed which follows the logical
                             ->'<                              *
flow chart shown in Figure 65.'   The purpose of the  computer program was
to simulate the implementation of a 207(b) program on a specific fleet of
vehicles, the  "controlled fleet", and compare the emission results to  an
uncontrolled fleet,  originally identical, which deteriorates without the
influence of a mandatory 207(b) program.  The comparison was made  over  a
50,000 mile period for each fleet,  and the effectiveness of the 207(b) program
was assessed annually.
              The two simulated fleets were  each divided into three mutually
exclusive groups based upon engine displacement:  150 CTD and less,  151 to
259 CID,  and  260 CID and greater.   Existing 300-car data was used to estab-
lish a distribution model for the emission values of the three regulated
pollutants at the mean mileage  for each displacement group.   This distribu-
tion model was then used to stochastically generate the two identical fleets
of vehicles with statistical attributes similar to the 300-car data set.
              The controlled fleet was aged under the influence of a manda-
tory inspection/maintenance program which included an annual inspection of
each vehicle,  mandatory maintenance of those vehicles failing the inspection,
and simulated deterioration based on mileage accumulation models.  The
uncontrolled fleet was aged using deterioration models based on mileage
accumulation.  The deterioration models all assumed linear  deterioration
with accumulated mileage.
 Figure  and table numbers refer to tables and figures contained in the
 Appendix.
                                    6-1

-------
               The existing 300-car data set was used to establish the
 short test (ST) inspection pass-fail levels. 'Two methodologies in selecting
 the levels were examined.  The first method, designated the 207(b) approach,
 established pass-fail levels for each engine displacement group at a 5 per-
 cent  error of commission rate for each pollutant.  The second method
 designated the I/M approach, established,  for each pollutant,  a single pass-
 fail level which applied to all groups.   The I/M  pass-fail levels were deter-
 mined using a 33 percent total failure rate with  a three constituent test.
               Because of uncertainties in the areas of deterioration rates'
 and maintenance  effectiveness,  a limited sensitivity analysis was performed
 by using several  different maintenance models and by varying many of the
 program input parameters.  This provided visibility into the variability of
 the results and identified the  most sensitive components of the analysis
 program.                                 ,
 6.2           PROGRAM COMPONENTS
               In the  sections below, the following  components of the  simu-
 lation (see Figure 65) are dis.cussed:
         6.2.1 Initializations (of vehicle mix, mileage,  etc.)
         6. 2. 2 Generation of Fleets (vehicles to be tracked)
         6. 2. 3 Emission Inspection Model—
         6. 2. 4 Vehicle Maintenance Models                          •
         6. 2. 5 Mileage Accumulation and Deterioration Model
         6.2.6 207(b) Effectiveness Calculations
 6. 2. 1         Initializations
.               Given  the (distributional parameters of the 300-car data, the
 use-in deterioration factors  from the  Emission Factors Program (EFP)
 were applied to these parameters to establish a distribution of fleet emis-r
 sions at the  average mileage  equivalent to the first inspection point (e. g. ,
 17, 500 miles). Linear deterioration was used to accomplish this
 initialization.
                                    6-2

-------
 6. 2. 2         Generation of Fleets
               The adjusted distribution model from the initialization com-
 ponent was used to stochastically generate a fleet of vehicles (e. g. ,  3000
 cars).  A tri-variate log-normal probability distribution model was used to
 produce,  for each vehicle, values of the HC,  CO, and NOX emission rates in
 grams/miles.   The values represent the emission rates that would be mea-
 sured by a Federal  Test Procedure (FTP) on the particular vehicle.  The
 statistical properties of the simulated fleet and the  300-car data set were
 quite similar.   This set of simulated vehicles was tracked simultaneously
 with and without the influence of a mandatory inspection/maintenance
 program.
 6. 2. 3         Emission Inspection Model
 6.2.3.1        General
               The model is composed of two components  which are executed
 sequentially for each vehicle of the controlled fleet.  The  first component,
 measurement,  consists of generating the ST measurements,  given the emis-
 sion rates which characterize the vehicle.   Once the measurements for HC,
 CO,  and NO have been noted, the second component is entered which is
designated as the inspection component. The inspection component consists
 of determining whether the vehicle passes  or fails the ST  and classifies the
 vehicle as an FF, E , E , or PP.
 6. 2. 3. 2        Measurement Component
               The ST measurements for the three regulated pollutants of
 each vehicle were generated stochastically using a conditional tri-variate
log-normal probability model, the parameters of which were determined
from the original 300-car data.  Thus,  given the FTP emission rates which
characterize the vehicle, three ST emission values  (one per pollutant) were
generated to simulate the effects of the measurement procedure  and  testing
equipment.  In this manner, the statistical properties inherent in the 300-car
                                   6-3

-------
data were approximated by the model and data generated which was
characteristic of the observed 300-car ST/FTP data.  The pertinent mathe-
matical equations are contained in "An introduction to Multivariate Statistical
Analysis, " by T. W. Anderson,  (Ref.  6).
6. 2. 3. 3        Inspection Component
               The ST measurement generated in the above component and
the FTP emission rates are used to classify the vehicle.  The vehicle fails
the ST inspection and the FTP inspection if any one of the three pollutant
measurements exceed their respective ST cut-points.  The vehicle is thus
classified according to the following:
           Type                        Inspection Results
           PP                      Passes ST and FTP
           FF                      Fails ST and FTP
           E                        Fails ST, passes FTP
           E                        Passes ST,  fails  FTP
This classification,  the FTP emission rates,  and  the ST measurements are
considered as  pertinent data for the various alternative  maintenance models,
and, hence, were carried through to the vehicle maintenance model and the
deterioration model.
6. 2. 4          Vehicle  Maintenance Models
6. 2. 4. 1        General
               The results of the  inspection model were used to select a
maintenance  policy for each vehicle individually.  The general philosophy
of the maintenance model was to perform corrective  maintenance on all
those vehicles failing the ST by adjusting the FTP failure  rates on each
emission in accordance with some pre-established rules or methods.  If a
vehicle passes the ST,  no such maintenance action was taken.  Specifically,
if the vehicle is classified as an E  or FF vehicle, the FTP emission rates
                                    6-4

-------
are adjusted to reflect the effect of corrective maintenance.  If the vehicle
is classified as a PP or E  vehicle, the FTP emission rates are left undis-
turbed.  Several  methods of adjusting the FTP values were investigated
and are discussed in the following sections.

6.2.4.2       Version 1;  HC and CO Coupled with a
              Lower Bound

              For vehicles failing the ST, the following checks  and adjust-
ments were made in the order presented:

         a.    NOX Check and Adjustment.  If NOX exceeded the federal
              standard of 3.  1 grams per mile, then the vehicle's FTP
              rate of NOX emission was adjusted to  3. 1 grams per mile.
              If  NO   was below 3. 1 grams per mile, then the vehicle's
              FTP rate of NOY emission was undisturbed.
                              J^,
         b.    CO Check and  Adjustments.  If CO exceeded the federal
              standard of 15.0 grams /per mile,  then the FTP CO rate
              of emission was adjusted to 15. 0 grams per mile, and the
              FTP HC rate of emission was adjusted as follows:


                                     /                     C°  A-\
                        HC  ,. = MAX 1/2 HC _ ; HC . , x  ^ * J      (6-1)
                           adj       I       std     old   CO  . , /     v
                                     \                       old/


              where

                    HC  ,. = adjusted value of HC

                    HC  , = HC federal  standard,  1. 5 grams per mile

                    HC , , = HC before adjustment

                    CO  ,. = adjusted value of CO which is  15. 0 grams  mile
                       adj

                    CO , , = CO before adjustment


              After these adjustments, the vehicle was checked to see  if
              HC exceeded its federal standard.
              If  CO was below its standard, then the vehicle was  checked
              for HC failing  its standard.
                                    6-5

-------
         c.    HC Check and Adjustments.  If HC exceeded the federal
              standard of 1. 5 grams per mile,  then HC was adjusted to.
              1.5 grams per mile, and CO was adjusted in accordance
              with Equation (6-1) where HC and CO have been interchanged.
              If HC was below its standard, then HC and CO were left
              undisturbed.
              The adjustment of NOX is  independent of the values of HC and
CO.  However,  when either  CO or HC is adjusted the other, is  adjusted with
the same percentage reduction.   This naturally  couples the maintenance of
CO and HC.  In  any event the adjusted values are not allowed to be less than
one-half the value of their respective standard.
6.2.4.3      Version 2: HC,  CO, and NOX Independently
              Adjusted to the FTP Standard
              For vehicles  failing the ST, each pollutant was  checked
individually.  Those pollutants exceeding their respective  FTP standard
were adjusted to their FTP  standard.  If a pollutant was below its standard,
it was left undisturbed.   Thus pollutants  not exceeding their standard were
not affected by those pollutants  which did exceed their standards.
6.2.4.4      Version 3: HC,  CO, and NOx Independently
              Adjusted to the Short Test Standard
              For vehicles  failing the ST, each pollutant was  checked
individually.  Those pollutants exceeding their respective  ST standard and
their FTP standard  (FF vehicles), were  adjusted to an FTP value in grams
per mile equivalent  to their  ST  standard.   Those pollutants exceeding their
respective ST standard,  but falling below their FTP standard (Ec vehicles),
were adjusted to their FTP  standard.  If a pollutant was below its ST
standard, it was left undisturbed and these pollutants were not affected by
those pollutants failing their ST  standards.
6. 2. 4. 5      Overview of Maintenance Models
              Maintenance options and adjustment procedures used in each
version of the maintenance model, are shown in Table 134 of the Appendix.
                                    6-6

-------
6. 2. 5         Mileage Accumulation and Deterioration Model
6, 2. 5. 1       General
              In this  model the  general  effects of age and wear were
simulated by deteriorating the emission  rates from their value at the last
inspection and maintenance point to the next I/M point.  Several  alternative
deterioration  concepts are available to be incorporated into such a model.
Three of these important deterioration types are discussed below.
6. 2. 5. 2       Deterioration Based Upon Certification Data
              The deterioration factors  estimated from the certification
program represent somewhat optimal conditions with respect to  maintenance,
driving habits, and time span of mileage accumulation.  The  maintenance is,
of course,  performed promptly in accordance with the manufacturers' speci-
fications and is most likely of the highest quality.  The driving of the dura-
bility vehicles is performed by professional drivers who probably inflict
less abuse on the engine than the average driver.  The mileage  is accumu-
lated in a somewhat optimal manner.  Typically, the vehicle would be
operated approximately four hours  at a time,  twice a day.  The total
elapsed time is typically on the order of  three months, thus eliminating
age and seasonal effects.
              Hence, deterioration based on  certification data may not be
representative of typical deterioration within  the population of in-use
vehicles  and is probably not representative of any large segment of the
in-use vehicle population.   However, it can be interpreted as an idealized
extreme.
6. 2. 5. 3       Deterioration Based Upon Emission Factors
              Program Data
              The deterioration factors  estimated from the Emission
Factors Program (EFP) represent  typical deterioration of vehicles owned
and operated by the public.  Maintenance habits, driving characteristics,
                                    6-7

-------
and time span of mileage accumulation are all quite .variable from vehicle
to vehicle and are difficult effects to isolate.   Thus,  the deterioration rates
are interpreted as being based on the average  of maintenance habits, driving
characteristics, and time span of mileage accumulation.  No attempt has
been made to iso'late the effects upon deterioration of these and other impor-
tant variables.  Thus,  these deterioration rates were taken as representative
of voluntary maintenance on in-use vehicles.
6. 2. 5.4      Deterioration Based Upon Near-Term Effects
              of Maintenance
              The disadvantage common to the two above types of deteriora-
tion is that they are long-term, in effect and apply over a large range in
mileage, i.e.,  50,000 miles.   However,  maintenance requirements are
considerably more frequent, which suggests the presence of a short-term
and possibly more rapid deterioration effect.  In fact, studies  (Refs. 7 and 8)
have shown that the  effect of maintenance is short-lived, six months to a
year, with respect to emission reductions.  This  is an important effect
which would support an optimistic position on inspection and maintenance.
6. 2. 5. 5      Specific Models Employed
6.2.5.5.1     Uncontrolled Fleet
              Deterioration effects on this fleet are based upon linear
models with mileage accumulation.  The-vehicles are classified into two
groups:
         a.    Vehicles passing the FTP
         b.    Vehicles failing the FTP
Certification deterioration factors were applied to vehicles in group a,  and
EFP factors were applied to vehicles in group b.  The actual values of  the
deterioration factors are shown in Table 135.
                                    6-8

-------
6.2.5.5.2     Controlled Fleet
               Deterioration effects on this fleet are based upon piecewise
linear models with mileage  accumulation.   The vehicles are classified into
four groups (E , E  ,  PP, and FF vehicles), with the following deterioration
rates:
         a.     PP Vehicles.  Certification rates as shown in Table 135.
         b.     EQ Vehicles.  EFP rates  as  shown in Table  135.
         c.     EC Vehicles.  Certification rates shown in Table 135.
               Deterioration occurs after the maintenance adjustments.
         d.     FF Vehicles.  A maintenance effectiveness period is
               specified as an input to the simulation program.  This
               period is the  number of months (less than or equal to 12)
               required for  the emissions to return to their value just
               prior to maintenance.  After this period, if  less than 12
               months, the emissions are deteriorated according to the
               EFP rates as shown in Table 135.   This situation  is dis-
               played in Figure 66.           ;
6. 2. 6          207(b) Effectiveness Calculations
               Program effectiveness was based upon total mass of pollutants
emitted since the inception of the program, reported annually.  For each
pollutant the total grams emitted for each vehicle were calculated for  the
average mileage accumulations shown in Table  136.  The total mass for
each fleet is calculated by summing over each vehicle within the  respective
fleets.  The effectiveness is then expressible as a percent, change relative to
the uncontrolled  fleet.
6. 2. 7          Overview  of Simulation Models
               The initial values of the emission rates are  established
stochastically for each vehicle in the simulated fleets.  Beyond this,  the
changes in emission rates of individual vehicles in the uncontrolled fleet are
purely deterministic.  However,  the changes in emission rates and the
sequence of events causing these changes are inherently stochastic for
individual vehicles in the controlled fleet.  This is due entirely to the
stochastic behavior of the inspection model.
                                    6-9

-------
              If a vehicle has been classified as an E  or FF,  then at least
one pollutant exceeded its  standard.  Since the combined effects of  mainte-
nance and deterioration over one year leave the year's end emission value
greater than or  equal to the year's beginning emission value, a vehicle
classified as an E  or FF  the previous year will be an E  or FF the following
year.  For example,  an FF one year will be an FF or E  the following year
and the probability of each is  determined by the inspection model.
              If a vehicle has been classified as an E  or PP vehicle, then
all of its emission rates were below the standards. An E  vehicle, because
                                                       c
of the maintenance method, will become  an E  or FF in the following year
regardless of the rate of deterioration.   A PP vehicle  may be a PP, E ,  FF,
or E  vehicle the following year, depending on the rate of deterioration.  If
deterioration is rapid enough  so that the  vehicle fails the FTP in the follow-
ing year,  then the vehicle  will be an FF or E .  However,  if deterioration is
not enough to fail the vehicle  on the FTP, then the vehicle will be an E or
PP vehicle the following year.
              These patterns,  summarized in Figure  67, may not  be  real-
istic and should be further investigated with data from other sources  such
as Reference 8.
6. 3           SIMULATION SCENARIOS AND RESULTS
6. 3. 1         General
              With each case or scenario treated, many input parameters
remained unchanged.  For the current set of analyses these inputs  were:
         a.    Mileage driving schedule  as given in Table 136.
         b.    The 1975 FTP standards which were 1.5 g/mi HC,  15. 0 g/mi
              CO, and 3. 1 g/mi NOx
         c.    Number of  simulated vehicles in each fleet which was
              3000 each.
         d.    The number of engine displacement groups which was three.
         e.    The vehicle mix which was 31. 7 percent 150 CID and less,
               18 percent  151 to 259 CID, and 50. 3 percent 260 CID and
              greater.
                                   6-10

-------
         f.    Starting seed for the simulation.
         g.    Deterioration rates are given in Table 135.
        • h.    Means and covariance  matrices for the logarithm of FTP
              emissions for each engine group.
         i.    Means and covariance  matrices for the logarithm of ST
              emissions for each engine group.  The ST was formed
              using the best three constituent test of the Federal Three-
              Mode, i.e. ,  HC and CO idle  in drive,  NOX at the high speed,
              (the 151 to 259 CID group used  NOX at the low speed).
         j.    Cross-covariance  matrices between the FTP and ST  for
              each engine group.
              These inputs generated an initial fleet with the characteristics
shown in Table 137.  The means and standard deviations  shown in Table 137
have been corrected to reflect a sample with average mileage of 17, 500
miles.  Below are presented the individual cases which were  simulated along
with the additional inputs required and the results obtained.
6. 3.2         Specific Scenarios  and Their  Results
6. 3. 2. 1      Introduction
              As previously indicated analyses  were conducted to assess  the
effectiveness  of a 207(b) program which establishes ST cut-points on the
basis of engine family and an I/M program which establishes for the entire
fleet a single,  ST  cut-point for each pollutant.  Each of these analyses
included an analysis using each of the three  maintenance models  for three
different maintenance  effectiveness periods.  The results, presented below,
are grouped according to the type of program  being simulated.
6. 3. 2. 2      Simulation of 207(b) Programs
              Shown in Table  138 are the ST cut-points used in the  207(b)
analyses.  The cut-points were computed from the 300-car  data set by using
a 5  percent E  rate for each individual pollutant on each engine classification.
For each version of the maintenance model  analyzed, three maintenance
effectiveness  periods were used:  12 months, 9 months, and 6 months.
                                   6-11

-------
Tables  139 through 141 show the results  of the 207(b) analyses.  The FTP
values equivalent to the ST cut-points are shown in Table 142.  These rates
were used in.maintenance model version 3 as the levels to which emissions
were reduced for maintained vehicles.  The program efficiences were calcu-
lated by accumulating and comparing the total emissions from each fleet
beginning at the first inspection point.  Hence these efficiencies do not
represent the full 50, 000 miles of operation since the emission during the
first year of the  fleet population's life was not included in this computation.
6. 3. 2. 3      Simulation of I/M Programs
              Shown in Table  143 are the ST  cut-points used in the I/M
analyses.  The cut-points were computed so  that 33. 33 percent of the original
300 vehicles  would fail the ST.  For each version of the three maintenance
models  analyzed, three maintenance effectiveness periods were used;
12 months, 9 months,  and 6 months.  Tables  144 through 146 show the results
of the I/M analyses.  The FTP values equivalent to the ST cut-points are
shown in Table 147.  These  rates were used  in maintenance  model version 3
as the levels to which emissions were reduced for the maintained vehicles.
The program efficiencies were calculated as explained above for the 207(b)
analyses.
6.4           DISCUSSION AND CONCLUSIONS
              The superiority of the 207(b) approach or the  I/M approach
cannot be determined from this set  of analyses.  Due to the method of
selecting  the I/M ST cut-points,  (i.e. , a 33. 33 percent total ST failure  rate),
the  error of commission  rate for the I/M program is predicted at 2 percent
for  each pollutant on the total fleet.  The 207(b) program,  with a 5 percent
E  rate, produces higher ST total failures,  42.4 percent.  This difference in
  c
total failures and in E  rate, alone, will shown 207(b) to be more effective
than I/M.  This is substantiated by  comparing the results shown in Tables
139 through 141  to the results  shown in Tables 144 through 146.  At a given
E  rate it is unclear which approach would be more effective.
                                   6-12

-------
               For a 207(b) program the effects of the maintenance  model
and the period of effective maintenance are readily compared.  These effects
are discussed below for each pollutant.
         HC.   For a fixed period of effective maintenance, verision 3 of the
         maintenance model had the lowest efficiencies with the efficiency
         of version  1 and 2 being equal.  The phenomenon is thought to be
         due to the fact that the vehicles are maintained to lower FTP levels
         in versions 1 and 2 than in version 3.
         CO.  For a fixed period of effective maintenance, version  1 had the
         highest efficiencies, and version  3 had the lowest efficiencies.  The
         differences in the efficiency using version 2 and version 3  were less
         than 1 percent.  Version 1 uses a coupled scheme for maintaining
         HC and CO, thus CO can benefit from improvements in HC which
         will produce higher CO efficiencies than  an independent method of
         adjustment.  The difference in efficiencies between versions 2 and
         3 are  attributed to the difference in the FTP values to which the
         vehicles were maintained.
         NOX.  For  a fixed period of maintenance  effectiveness,  version 3
         showed lower efficiencies than version 1  and 2 which are identical.
         Again this  is attributed to  the lower FTP maintenance levels in
         version 1 and 2.
               For a fixed version of maintenance, program efficiencies
decrease with decreasing periods of maintenance  effectiveness for each
pollutant.  Intuitively, the quicker the  erosion  of maintenance, the less the
impact of maintenance.
               For a fixed maintenance effectiveness period and a fixed
maintenance version,  the efficiency of the  program increases as the dura-
tion of the program increases.  Although this behavior is  expected, it has
several important consequences. The earlier  the program can be insti-
tuted, the greater the resulting benefit on air quality and the  higher will be
the program efficiency at 50, 000 miles.  Since emissions on controlled
vehicles cannot be reduced to zero, the maximum program efficiency is less
than 100  percent and can be approximated by allowing program duration to
increase beyond the 50, 000 mile limit.   This maximum will still be dependent
upon the time at which the program is  established.
                                   6-13

-------
   .            The behavior observed for 207(b) is likewise found in the I/M
results,  Tables 144 to 146.  The only exception occurs for CO.- For a fixed
value of  the maintenance effectiveness period, version 1 is highest and
version  3 is lowest, as shown in the 207(b)  results.  However, the  maximum
difference between the CO efficiencies of version  2 and  3 is 4;9 percent,
and the differences are consistent above about 2 percent.  This is most
likely due to the inverse  relationship in the CO ST standard of the 207(b)
simulation and the I/M simulation^  Notice  that for the 207(b)  simulation the
CO ST standards, in FTP units  as shown in Table  142, decrease with
increasing engine size.   For the I/M simulation the CO ST  standards,
Table 147, increase with increasing engine size.   This difference,  taken
with the  fact that the 260 CID and greater engine class accounts for  50 per-
cent of the total fleet, offers an explanation.
               The following general comments are offered to aid the inter-
preter of these results and,  hopefully, increase their utility.  Program
impact on air quality can be meaningfully measured on an annual basis or on
a cumulative basis.   The cumulative impact should probably be measured
over the  total life of the vehicle up to 50, 000 miles.  The efficiencies pre-
sented herein have been computed over the  17, 500 to 50, 000 mile segment of
the fleet population's life.  Thus the percent improvement in air  quality on  a
cumulative basis will be  lower than the above calculated efficiencies (with
estimates running approximately 50  percent lower).
              In any simulation, the interpreter of the results must realize
that the .properties and general behavior observed in the results are directly
attributable to the simulator.   Since simulation is based primarily on belief
and secondarily on fact,  the degree to which the results can be extrapolated
is dependent upon the interpreter's belief in the underlying models used to
construct the  simulator.  Sensitivity analyses can be used to increase the
diversity of beliefs or models and hence reflect the sensitivity of the results
to these  underlying models.
                                   6-14

-------
              A limited attempt at a  sensitivity analysis was made during
this study through the use  of three models of maintenance and three periods
of the effectiveness  of maintenance.  Although this provided a much needed
diversity in models,  there are many other different approaches to mainte-
nance and deterioration which could be represented.   One deficiency
inherent in all three models of maintenance examined in the present analy-
sis is that an  E   vehicle always is an FF or E  vehicle the following year.
This deficiency  is traceable to the deterministic approach employed which
can perhaps be eliminated through the use of stochastic maintenance models.
                                   6-15

-------
                              REFERENCES
1.    "Federal Test Procedure and Short Test Correlation Analyses,"
      Report No. EPA-460/3-76-011, The Aerospace Corporation,
      El Segundo, California, April 1976.

2.    T. W.  Anderson, Multivariate Statistical Analysis, John Wiley and
      Sons,  Inc. , New York (1958).

3.    K. A. Brownlee,  Statistical Theory and Methodology - In Science
      and Engineering,  John Wiley and Sons, Inc.,  New York (1965).

4.    H. Cramer, Mathematical Methods of Statistics, Princeton University
      Press, New Jersey (1971).

5.    C. R. Rao, Linear Statistical Interference and Its  Applications,
      John Wiley and Sons,  Inc., New York  (1965).

6.    High Altitude Vehicular Emission Control Program - Volume III,
      TRW,  Redondo Beach, California and  Automotive Testing Laboratories,
      Denver, Colorado, June 1975.

7.    Degradation Effects on Motor Vehicle  Exhaust Emission, Olson
      Laboratories, Inc.,  Anaheim"Calif., March 1976.
                                  R-l

-------
      APPENDIX




TABLES AND FIGURES

-------
                                 TABLES
1.    Cross Tabulation of Vehicles by Number of Cylinders   •  *   .
      and Engine Displacement - All  Three Cities	     1

2.    Cross Tabulation of Vehicles by Inertia Test Weight
      and Engine Displacement - All  Three Cities	     1

3.    Cross Tabulation of Vehicles by Engine Displacement
      and Emission Control System Type —  All Three Cities	     2

4.  .  Cross Tabulation of Vehicles by Inertia Test Weight and
      Emission Control System Type  -  All Three  Cities	     2

5.    Cross Tabulation of Vehicles by Fuel  System and
      Emission Control System Type.  —  All Three  Cities	     3

6.    Cross Tabulation of Vehicles by Inertia Test Weight
      and Fuel System Type - All Three Cities	     3

7.    Cross Tabulation of Vehicles by Engine Displacement
      and Fuel System Type - All Three Cities	     4

8.    Cross Tabulation of Vehicles by Engine Displacement
      and Transmission Type - All Three Cities	     4

9.    Cross Tabulation of Vehicles by Inertia Test Weight
      and Transmission Type - All Three Cities	     5

10.   Cross Tabulation of Vehicles by Manufacturer and
      Engine Displacement — All  Three Cities	     5

11.   Cross Tabulation of Vehicles by Manufacturer and
      Engine Displacement — Chicago	     6

12.   Cross Tabulation of Vehicles by Manufacturer and
      Engine Displacement — Houston	     7

13.   Cross Tabulation of Vehicles by Manufacturer and
      Engine Displacement — Phoenix	     8

-------
                          TABLES (Continued)
14a.    Federal Short Cycle — Means and Standard

14b.
15a.

15b.
I6a.

I6b.
17a.

17b.
18a.

18b.
19.

20a.
20b.
21.
22.
23.
24.
25a.

25b.
Deviations by City . . 	 	 	
Associated Tests of Significance 	 	
Federal Short Cycle — Means and Standard Deviations
by Engine Displacement - All Three Cities 	
Associated Tests of Significance 	
Federal Three Mode - Means and Standard
Deviations by City 	 ' .
Associated Tests of Significance 	
Federal Three-Mode - Means and Standard Deviations
by Engine - All Three Cities 	
Associated Tests of Significance 	
FTP Means and Standard Deviations by Mileage
Group - All Three Cities 	 	
Associated Tests of Significance 	
Correlation Coefficient Summary by Accumulated
Mileage - All Three Cities 	 	
FTP Means and Standard Deviations - by City 	
Associated Tests of Significance 	
Correlation Coefficient Summary - All Three Cities 	
Correlation Coefficient Summary — Chicago 	
Correlation Coefficient Summary — Houston 	
Correlation Coefficient Summary - Phoenix 	
FTP Means and Standard Deviations by Engine
Displacement — All Three Cities 	
Associated Tests of Significance 	
9
9

10
10

11
12

13
14

15
15

16
17
17
18
18
19
19

20
20
                                   11

-------
                            TABLES (Continued)
26.    Correlation Coefficient Summary by Engine
       Displacement — All Three Cities 	    21

27a.   FTP Means and Standard Deviations by Inertia
       Test Weight - All Three Cities  	    22

27b.   Associated Tests of Significance	    22

28.    Correlation Coefficient Summary by Inertia Test
       Weight Group - All Three Cities 	    23

29a.   FTP Means and Standard Deviations by Emission
       Control System Type - All Three Cities	    24

29b.   Associated Tests of Significance	    24

30.    Correlation Coefficient Summary by Emission
       Control System Type - All Three Cities	    25

3 la.   FTP Means and Standard Deviations by Fuel System
       Type  - All Three Cities	    26

31b.   Associated Tests of Significance	    26

32.    Correlation Coefficient Summary by Fuel System
       Type  - All Three Cities	    27

33a.   FTP Means and Standard Deviations by Transmission
       Type  - All Three Cities	    28

33b.   Associated Tests of Significance	    28

34.    Correlation Coefficient Summary by Transmission
       Type  - All Three Cities	    29

35a.   FTP Means and Standard Deviations by Manufacturer —
       All Three Cities 	    30

35b.   Associated Tests of Significance	    30

36.    Correlation Coefficient Summary by Manufacturer -
       All Three Cities 	    31
                                    111

-------
                            TABLES (Contiriue'd)
37.    Correlation Coefficient Summary for  150 CID        '.  •-   <: '•
  • •'    and Less by Manufacturer — All Three Citifes .	.-.' . .-J    32

38.    Correlation Coefficient Summary for  IS*, to 259 CID   .',,
       by Manufacturer — All Three Cities  :•*.'.. r.'.. '.'	>'.-••". ;. •/(    33

39-.-'    Correlation Coefficient Summary for  2,60. CID and     ...!.-•• .:•-..:'
       Greater by Manufacturer — All Three Cities	    34

40a.   Means and Standard. Deviations by Engine
       Displacement — Chicago	    35

40b.   Associated Tests of Significance	•    35

41.    Correlation Coefficient Summary by Engine
       Displacement - Chicago	    36

42a.   FTP Means and Standard Deviations by Inertia
       Test Weight - Chicago	    37

42b.   Associated Tests of Significance	    37

43.    Correlation Coefficient Summary by Inertia Test
       Weight Group - Chicago	'	    38

44a.   FTP Means and Standard Deviations by'Emission
       Control System Type  - Chicago	    39

44b.   Associated Tests of Significance	    39

45.'    Correlation Coefficient Summary by Emission
       Control System Type  - Chicago	    40
                   .'..-..•    •"•(">'    -         .      ,
46a.   FTP Means and Standard Deviations by Fuel
       System Type — Chicago	    41

46b.   Associated Tests of Significance	    41

47.    Correlation Coefficient Summary by Fuel
       System Type - Chicago	    42

-------
                           TABLES (Continued)
48a.   FTP Means and Standard Deviations by
       Transmission Type - Chicago	    43

48b.   Associated Tests of Significance	    43

49.    Correlation Coefficient Summary by Transmission
       Type — Chicago	    44

50a.   FTP Means and Standard Deviations by
       Manufacturer - Chicago	    45

50b.   Associated Tests of Significance	    45

51.    Correlation Coefficient Summary by
       Manufacturer — Chicago	    46

52a.   FTP Means and Standard Deviations by Engine
       Displacement - Houston	;    47

5Zb.   Associated Tests of Significance	    47

53.    Correlation Coefficient Summary by Engine
       Displacement - Houston	    48

54a.   FTP Means and Standard Deviations by
       Inertia Test Weight - Houston	    49

54b.   Associated Tests of Significance	    49

55.    Correlation Coefficient Summary by Inertia
       Test Weight Group - Houston	    50

56a.   FTP Means and Standard Deviations by Emission
       Control System Type - Houston	    51

56b.   Associated Tests of Significance	    51

57.    Correlation Coefficient Summary by Emission
       Control System Type - Houston  . .  .	    52

-------
                            TAB-LJE'S' (Continued)
58a.   FTP Means and Standard Deviations by - " ' '•' "•   '    •    ' "    '|
       FuelSystem Type -  Houston ......:." ."•.'-": ...... .  .".'." ;.    53

58b.   Associated Tests of Significance .... :T .  . .'	.'  . .  .J;.'    53

59.    Correlation Coefficient Summary by Fuel';             •' •'"'"';*'
  '' -    System Type - Houston. :  .	::..;...... :  . '•.-; ;' '.    54

60a.   FTP Means and Standard Deviations by
       Transmission Type — Houston	    55

60b.   Associated Tests of Significance	    55

61.    Correlation Coefficient Summary by                         .
       Transmission Type — Houston	 .    56

62a.   FTP Means and Standard Deviations by
       Manufacturer — Houston	..'.....    57

62b.   Associated Tests of Significance	    57

63.    Correlation Coefficient Summary by
       Manufacturer - Houston	 .  . .    58

64a.   FTP Means and Standard Deviations by Engine
       Displacement — Phoenix	    59

64b.   Associated Tests of Significance ;	 .    59

65.    Correlation Coefficient Summary by Engine
       Displacement - Phoenix		    60

66a.   FTP Means and Standard Deviations by
       Inertia Test Weight - Phoenix	;    61

66b.   Associated Tests of Significance	.'  .'-.'  . :    61

67.    Correlation Coefficient Summary by Inertia
       Test Weight Group - Phoenix .  . .  ;	    62

68a.   FTP Means and Standard Deviations by Emission
       Control System — Phoenix	    63

68b.   Associated Tests of Significance	    63
                                    VI

-------
                            TABLES (Continued)
69.    Correlation Coefficient Summary by Emission     .,   . ,  ,
       Control System Type - Phoenix	•	    64

70a.   FTP Means and Standard Deviations by Fuel
       System Type — Phoenix	    65

70b.   Associated Tests of Significance*.  .... ;	  .r	    65

71.    Correlation Coefficient Summary by Fuel
       System Type - Phoenix		    66

72a.   FTP Means and Standard Deviations by
       Transmission Type - Phoenix  . .  .•	    67

72b.   Associated Tests of Significance	    67

73.    Correlation Coefficient Summary by
       Transmission Type — Phoenix	  .    68

74a.   FTP Means and Standard Deviations by
       Manufacturer - Phoenix	  .    69

74b.   Associated Tests of Significance	    69

75.    Correlation Coefficient Summary by
       Manufacturer — Phoenix	. .  . .	    70

76.    Comparison of ST HC Results by City - Predicted
       Population, Bounded Errors of Commission
       Method (Ec Set at 5 Percent)	  . .	    71

77.    Comparison of ST CO Results by City - Predicted
       Population, Bounded Errors of Commission
       Method (Ec Set at 5 Percent)	    72

78.    Comparison of ST NOX Results by City -  Predicted
       Population, Bounded Errors of Commission
       Method (Ec Set at 5 Percent)	    73

79.    Comparison of ST Multiple Constituent Results by
       City - Actual Population,  Bounded Errors of
       Commission Method (Ec Set at 5 Percent)  	    74
                                    VII

-------
                           TABLES (Continued)
80.    Comparison of ST HC Results by Inertia Weight -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)	    75

81.    Comparison of ST CO Results by Inertia Weight -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  ......    76

82.    Comparison of ST NOX Results by Inertia Weight -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)	    77

83.    Comparison of ST Multiple Constituent Results of
       Inertia Weight - Actual Population, Three  Cities,
       Bounded Errors of Commission Method
       (Ec Set at 5 Percent)  	    78

84.    Comparison of ST HC Results by Displacement -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (E., Set at 5 Percent)	    79
                                       v*

85.    Comparison of ST CO Results by Displacement -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  	    80

86.    Comparison of ST NOX Results by Displacement -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (E  Set at 5 Percent)	    81

87.    Comparison of ST Multiple Constituent Results by
       Displacement — Predicted Population,  Three Cities,
       Bounded Errors of Commission Method
       (Ec Set at 5 Percent)	 .    82

88.    Comparison of ST HC Results by Emission Controls
       Type — Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  	    83

89.    Comparison of ST CO Results by Emission Controls
       Type - Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)	    84
                                    Vlll

-------
                           , TABLES (C ontinued)
90.    Comparison of ST NOX Results by Emission Controls
       Type — Predicted Population,  Three Cities, Bounded
       Errors of Commission Method (E  Set at 5 Percent) .......     85

91.    Comparison of ST Multiple Constituent Results by
       Emission Controls Type — Predicted Population,
       Three Cities, Bounded Errors of Commission      ;,•
       Method (Ec Set at 5 Percent)	     86

92.    Comparison of ST HC Results by Fuel System Type -
       Predicted Population, Three Cities, Bounded Errors
       of Commission Method (Ec Set at 5 Percent)	     87

93.    Comparison of ST CO Results by Fuel System Type -
       Predicted Population, Three Cities, Bounded Errors
       of Commission Method (Ec Set at 5 Percent)	     88

.94.    Comparison of ST NOX Results by Fuel  System Type -
       Predicted Population, Three. Cities, Bounded Errors
       of Commission Method (Ec Set at 5 Percent)	     89

95.    Comparison of ST Multiple Constituent Results by
       Fuel System Type -  Predicted Population, Three
       Cities, Bounded  Errors of Commission Method
       (Ec Set at 5 Percent)  	     90

96. .    Comparison of ST HC Results by Transmission Type —
       Predicted Population, Three Cities, Bounded Errors
       of Commission Method (Ec Set at 5 Percent)	     91

97.    Comparison of ST CO Results by Transmission Type —
       Predicted Population, Three Cities, Bounded Errors
       of Commission Method (Ec Set at 5 Percent)	     92

98.    Comparison of ST NOX Results by Transmission  Type —
       Predicted Population, Three Cities, Bounded Errors
       of Commission Method (Ec Set at 5 Percent)	     93

99.    Comparison of ST Multiple Constituent Results by
       Transmission Type — Predicted Population, Three
       Cities, Bounded  Errors of Commission Method
       (Ec Set at 5 Percent)  	;	     94
                                     IX

-------
                           TABLES (Continued)
100.    Comparison of ST HC Results by Manufacturer -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)	    95

101.    Comparison of ST CO Results by Manufacturer -
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  ......    96

102.    Comparison of ST NOX Results by Manufacturer —
       Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)	    97

103.    Comparison of ST Multiple Constituent Results by
       Manufacturer - Predicted Population, Three Cities,
       Bounded Errors of Commission Method
       (E_ Set at 5 Percent)  	    98
         \_.

104.    Estimated Standard Errors for Contingency Table
       Percentages  in Multiple Constituent Test Results	    99

105.    Comparison of Federal Short Cycle Cut-Points by
       City (Ec Set at 5 Percent)	    99

106.    Comparison of Federal Three-Mode Cut-Points by
       City (Ec Set at 5 Percent)	    100

107.    Comparison of HC Results for CID = 150  and Less,
       by Manufacturer — Predicted Population, Three
       Cities, Bounded Errors of Commission Method
       (Ec Set at 5 Percent)	    101

108.    Comparison of CO Results for CID = 150  and Less,
       by Manufacturer — Predicted Population, Three
       Cities, Bounded Errors of Commission Method
       (Ec Set at 5 Percent)  	    102

109.    Comparison of NOX Results for CID =  150 and Less,
       by Manufacturer - Predicted Population, Three
       Cities, Bounded Errors of Commission Method
       (Ec Se.t at 5 Percent)  	    103

110.    Comparison of Multiple Constituent Results for
       CID =150 and Less, by Manufacturer -  Predicted
       Population,  Three Cities, Bounded Errors of
       Commission  Method (Ec Set at 5  Percent) 	    104

-------
                           TABLES (Continued)
111.    Comparison of HC Results for CID =  151 to 259,
       by Manufacturer — Predicted Population,  Three
       Cities, Bounded Errors of Commission Method
       (Ec Set at 5 Percent)  	     105

112.    Comparison of CO Results for CID =  151 to 259,
       by Manufacturer — Predicted Population,  Three
       Cities, Bounded Errors of Commission Method
       (Ec Set at 5 Percent)	     106

113.    Comparison of NOX Results for CID = 151 to 259,
       by Manufacturer — Predicted Population,  Three
       Cities, Bounded Errors of Commission Method
       (Ec Set at 5 Percent)  	     107

114.    Comparison of Multiple Constituents  Results for
       CID = 151 to 259, by Manufacturer - Predicted
       Population,  Three Cities,  Bounded Errors of
       Commission Method (Ec Set at 5 Percent)  	     108

115.    Comparison of HC Results for CID =  260 and
       Greater, by Manufacturer — Predicted Population,
       Three Cities,  Bounded Errors of Commission
       Method (Ec Set at 5 Percent)	     109

116.    Comparison of CO Results for CID =  260 and
       Greater, by Manufacturer — Predicted Population,
       Three Cities,  Bounded Errors of Commission
       Method (Ec Set at 5 Percent)	     110

117.    Comparison of NOX Results for CID = 260 and
       Greater, by Manufacturer — Predicted Population,
       Three Cities,  Bounded Errors of Commission
       Method (Ec Set at 5 Percent)	     Ill

118.    Comparison of Multiple Constituent Results for
       CID = 260 or Greater,  by Manufacturer - Predicted
       Population,  Three Cities,  Bounded Errors of
       Commission Method (Ec Set at 5 Percent)  	     112

119.    Comparison of Federal Short Cycle Cut-Points by
       Engine Displacement (Ec  Set at 5 Percent)	     113

-------
                           TABLES, (Continued)
120.   Comparison of Federal Three-Mode Cut-Points
       by Engine Displacement (Ec Set at 5 Percent) ..... .......    114

121.   Comparison of HC Results for CID = 150 and Less,
       by City - Predicted Population, Three Cities,
       Bounded Errors of Commission Method
       (Ec Set at 5 Percent)  ......... ....... ......... ,.-....    H5
122.    Comparison of CO Results for CID = 150 and Less,
       by City —  Predicted Population, Three Cities,
       Bounded Errors of Commission Method
       (Ec Set at 5 Percent) . . .  . .......................    116

123.    Comparison of NOX Results for CID = 150 and Less,
       by City -  Predicted Population, Three Cities,
       Bounded Errors of Commission Method
       (Ec Set at 5 Percent) . . .  . ..... ...................    117

124.    Comparison of Multiple Constituent Results for
       CID =150 and Less, by City - Predicted Population,
       Three  Cities,  Bounded Errors of Commission
       Method (Ec Set at 5 Percent)  . ............. ..... ....    118

125.    Comparison of HC Results for CID = 151 to 259,  by
       City -  Predicted Population,  Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  ......    119

126.    Comparison of CO Results for CID = 151 to 259,  by
       City —  Predicted Population,  Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  ......    120

127.    Comparison of NOX Results for CID = 151 to 259, by
       City -  Predicted Population,  Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  ......    121

128.    Comparison of Multiple Constituent Results for
       CID = 151 to 259, by City - Predicted Population,
       Three  Cities,  Bounded Errors of Commission
       Method (Ec Set at 5 Percent)  ........... . . .........    122

129.    Comparison of HC Results for GID . = 260 and Greater,
       by City -  Predicted Population, Three Cities,  Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  ......    123
                                    XII

-------
                           TABLES (Continued)
130.    Comparison of CO Results for CID = 260 and Greater
       by City -  Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  	    124

131.    Comparison of NOX Results for CID = 260 and Greater,
       by City -  Predicted Population, Three Cities, Bounded
       Errors of Commission Method (Ec Set at 5 Percent)  	    125

132.    Comparison of Multiple Constituent Results for
       CID = 260 and Greater, by City - Predicted
       Population,  Three Cities,  Bounded Errors of

133.
134.

135.

136.

137.
138.
139.

140.

141.

142.
143.
144.
Commission Method (Ec Set at 5 Percent) . . . •. 	
Contingency Table 	 	
Maintenance Options and Adjustment Procedures for
Each Version of the Maintenance Model 	
Average Deterioration Rates Used in 207(b)
Effectiveness Simulation 	
Annual Mileage Accumulations Used in 207(b)
Effectiveness Simulation 	
Initial Simulated Fleet Statistics 	
207(b) ST Cut- Points for Each Engine Family 	
207(b) Program Efficiencies with Maintenance
Version 1 	
207(b) Program Efficiencies with Maintenance
Version 2 	
207(b) Program Efficiencies with Maintenance
Version 3 	
FTP Values Equivalent to 207(b) ST Cut- Points 	
I/M ST Cut- Points for Each Engine Family 	
I/M Program Efficiencies with Maintenance
126
127

128

128

129
129
130

, . 131

132

133
134
134

       Version 1	    135
                                   Xlll

-------
                           TABLES (Continued)
145.    I/M Program Efficiencies with Maintenance
       Version 2 ...'...........••••••	    136

146.    I/M Program Efficiencies with Maintenance .  .
       Version 3 . V ... . ^  ...... \	    137

147.    FTP Values  Equivalent to I/M ST Cut-Points  ...........    138
                                   xiv

-------
                                 FIGURES.
1.    Variation of EC, EQ,  and FF with HC Cut-point;
      All Three Cities; Federal Short Cycle; Bounded
      Errors  of Commission Method	  .    139

2.    Variation of EC, EQ,  with FF with CO Cut-point;
      All Three Cities; Federal Short Cycle; Bounded
      Errors  of Commission Method	    139

3.    Variation of Ec, Eo,  and FF with NOX Cut-point;
      All Three Cities; Federal Short Cycle; Bounded
      Errors  of Commission Method	    140

4.    Variation of EC, EQ,  and FF with HC Cut-point;
      Chicago; Federal Short Cycle; Bounded Errors
      of Commission Method	    140

5.    Variation of EC, EQ,  and FF with CO Cut-point;
      Chicago; Federal Short Cycle; Bounded Errors
      of Commission Method	  .    141

6.    Variation of EC, EQ,  and FF with NOX Cut-point;
      Chicago; Federal Short Cycle; Bounded Errors of
      Commission Method	    141

7.    Variation of Ec, EQ,  and FF with HC Cut-point;
      Houston; Federal Short Cycle; Bounded Errors
      of Commission Method	    142

8.    Variation of EC, EQ,  and FF with CO Cut-point;
      Houston; Federal Short Cycle; Bounded Errors
      of Commission Method	    142

9.    Variation of EC, E    and FF with NOX Cut-point;
      Houston; Federal Short Cycle; Bounded Errors
      of Commission Method	    143

10.   Variation of Ec, Eo,  and FF with HC Cut-point;
      Phoenix; Federal Short Cycle; Bounded Errors  .
      of Commission Method	    143

11.   Variation of Ec, EQ,  and FF with CO Cut-point;
      Phoenix; Federal Short Cycle; Bounded Errors
      of Commission Method	    144
                                    xv

-------
                           FIGURES (Continued)
12.   Variation of Ec, Eo, and FF'with NOX Cut-point;
      Phoenix; Federal Short Cycle; Bounded Errors
      of Commission Method	. .	     144

13.   Variation of Actual EC, EQ, and FF with Predicted
      Ec; Federal Short Cycle; All Three  Cities	     145

14.   Variation of Actual Ec, Eo, and FF with Predicted
      Ec; Federal Short Cycle; Chicago	     145

15.   Variation of Actual Ec, E0, and FF with Predicted
      Ec; Federal Short Cycle,  Houston	     146

16.   Variation of Actual Ec, Eo, and FF with Predicted
      Ec; Federal Short Cycle; Phoenix	     146

17.   Variation of EC, EQ, and FF with HC Cut-point;
      All Three Cities; Federal Three-Mode; Bounded
      Errors of Commission Method	     147

18.   Variation of EC, E , and FF with CO Cut-point;
      All Three Cities; Federal Three-Mode; Bounded
      Errors of Commission Method	     148

19.   Variation of EC, EQ, and FF with NOX Cut-point;
      All Three Cities; Federal Three-Mode; Bounded
      Errors of Commission Method	     149

20.   Variation of EC, EO, and FF with HC Cut-point;
      Chicago; Federal Three-Mode; Bounded Errors
      of Commission Method	     149

21.   Variation of EC, EQ, and FF with CO Cut-point;
      Chicago; Federal Three-Mode; Bounded Errors
      of Commission Method	     150

22.   Variation of EC, EQ, and FF with NO  Cut-point;
      Chicago; Federal Three-Mode; Bounded Errors
      of .Commission Method	     151

23.   Variation of EC, EQ, arid FF with HC Cut-point;
      Houston; Federal Three-Mode; Bounded Errors
      of Commission Method	•	     151
                                   xvi

-------
                           FIGURES (Continued)
24.   Variation of Ec, E^,  and FF with CO Cut-point;
      Houston; Federal Three-Mode; Bounded Errors
      of Commission Method	     152

25.   Variation of Ec, Eo,  and FF with NOX Cut-point;
      Houston; Federal Three-Mode; Bounded Errors
      of Commission Method	     153

26.   Variation of EC, EQ,  and FF with HC Cut-point;
      Phoenix; Federal Three-Mode; Bounded Errors
      of Commission Method	     153

27.   Variation of EC, EQ,  and FF with CO Cut-point;
      Phoenix; Federal Three-Mode; Bounded Errors
      of Commission Method	     154

28.   Variation of EC, EQ,  and FF with NOX Cut-point;
      Phoenix; Federal Three-Mode; Bounded Errors
      of Commission Method	     155

29.   Variation of Actual EC, EQ, and FF with Predicted
      Ec; Federal Three-Mode; All Three Cities	     156

30.   Variation of Actual Ec, Eo, and FF with Predicted
      Ec; Federal Three-Mode; Chicago	     157

31.   Variation of Actual EC, EQ, and FF with Predicted
      Ec; Federal Three-Mode; Houston	     158

32.   Variation of Actual EC, EQ, and FF with Predicted
      Ec; Federal Three-Mode; Phoenix	     159

33.   Variation of EC, E^  and FF with HC Cut-point;
      150 CID and Less; Federal Short Cycle; Bounded
      Errors of Commission Method	     160

34.   Variation of EC, EQ,  and FF with CO Cut-point;
      150 CID and Less; Federal Short Cycle; Bounded
      Errors of Commission Method	     160

35.   Variation of EC, E  .  and FF with NOX Cut-point;
      150 CID and Less; Federal Short Cycle; Bounded
      Errors of Commission Method	;	     161
                                   xvii

-------
                           FIGURES. (Continued)
36.   Variation of EC, E ,  and FF with HC Cut-point;
      151 to 259 CID; Federal Short Cycle; Bounded
      Errors of Commission Method	     161

37.   Variation of Ec, Eo,  and FF with CO Cut-point;
      151 to 259 CID; Federal Short Cycle; Bounded
      Errors of Commission Method .	     162

38.   Variation of Ec, Eo,  and FF with NOX Cut-point;
      151 to 259 CID; Federal Short Cycle; Bounded
      Errors of Commission Method	     162

39. .  Variation of Ec, Eo,  and FF with HC Cut-point;
      260 CID and Greater; Federal Short Cycle;
      Bounded  Errors of Commission Method	     163

40.   Variation of EC, EQ,  and FF with CO Cut-point;
      260 CID and Greater; Federal Short Cycle;
      Bounded  Errors of Commission Method	     163

41.   Variation of EC, EQ,  and FF with NOX Cut-point;
      260 CID and Greater; Federal Short Cycle;
      Bounded  Errors of Commission Method	     164

42.   Variation of Ec, Eo,  and FF with Predicted
      Ec; Federal Short Cycle; All Three Cities;
      Engine Displacement 150 CID and Less	     165

43.   Variation of Actual Ec, Eo, and  FF with Predicted
      Ec; Federal Short Cycle; All Three Cities; Engine
      Displacement  1 5 1 to 259 CID	     165

44.   Variation of Actual Ec, Eo, and  FF with Predicted
      Ec; Federal Short Cycle; All Three Cities; Engine
      Displacement  260 CID and Greater	' . .  . .     166

45.   Variation of EC, E^  and FF with HC Cut-point;
      150 CID and Less; Federal Three-Mode; Bounded
      Errors of Commission Method	     16?

46.   Variation of EC, EQ,  and FF with CO Cut-point;
      150 CID and Less; Federal Three-Mode; Bounded
      Errors of Commission Method	 .     16?
                                   XVlll

-------
                           FIGURES (Continued)
47.   Variation of Ec, Eo, and FF with NOX Cut-point;
      150 CID and Less; Federal Three-Mode; Bounded
      Errors of Commission Method ......................    168

48.   Variation of EC, E  , and FF with HC Cut-point;
      151 to 259 CID; Federal Three-Mode; Bounded
      Errors of Commission Method .......................    168

49.   Variation of EC, EQ, and FF with CO Cut-point;
      151 to 259 CID; Federal Three-Mode; Bounded
      Errors of Commission Method ................ ......    169

50.   Variation of EC, Eo, and FF with NOX Cut-point;
      151 to 259 CID; Federal Three-Mode; Bounded
      Errors of Commission Method ......................    170

51.   Variation of EC, EQ, and FF with HC Cut-point;
      260 CID and Greater; Federal Three-Mode;
      Bounded Errors of Commission Method ...............  .    170

52.   Variation of Ec, Eo, and FF with CO Cut- point;
      260 CID and Greater; Federal Three- Mode;
      Bounded Errors of Commission Method ................    171

53.   Variation of EC, EQ, and FF with NOX Cut-point;
      260 CID and Greater; Federal Three-Mode;
      Bounded Errors of Commission Method ................    171
54.   Variation of Actual EC,  EQ, and FF with Predicted
      Ec; Federal Three-Mode; All Three Cities; Engine
      Displacement 150 CID and  Less .....................    172

55.   Variation of Actual EC,  EQ, and FF with Predicted
      E  ; Federal Three-Mode; All Three Cities; Engine
      Displacement 151 to 259 CID .......................    173

56.   Variation of Actual EC,  EQ, and FF with Predicted
      Ec; Federal Three- Mode; All Three Cities; Engine
      Displacement 260 CID and  Greater ...................    174

57.   Federal Short Cycle Test Driving Schedules   ............    175

58.   Contingency Table Representation  ...................    176

59.   Bounded Errors of Commission Method   ...............    177
                                   xix

-------
                            FIGURES, (Continued)



60.   Parametric Model  . . .: .-.-..•."•'. .. .... .,-.... .	_•...-.- . .     178
                                r . • •                              - .  ,
61.   Probability Equations   .	 .  . ,./. ... .	  .:	-..     178

62.   Equations for ST Cut-Poiht..Determination	, >... .  . .     179

63:.   Computation Flow Chart	,.  . . , .•.-.	...-. ....     180

64.   Typical Variability of Predicted Population Results   	     181

65.   Logical Flow Chart for 207(b) Effectiveness Estimation	     182

66.   Piecewise Linear Deterioration for FF Vehicles
      Shown Over Two-Year Period   .....'..	  . .  :	     183

67.   Vehicle Classification Properties of the 207(b)
      Effectiveness Simulator	     184
                                     xx

-------
               Table  1.  Cross Tabulation of Vehicles by
  Number of Cylinders and Engine Displacement— All Three Cities
                                                               (a)
No. of
Cylinders
4
6
8
Engine Displacement
150 CID(b)
or Less
95
0
0
151 to
259 CID
0
54
0
260 CID or
Greater
0
0
151
(a)
(b)
300 vehicles total
CID = cubic inch displacement
               Table 2.  Cross Tabulation of Vehicles by
     Inertia Test Weight and Engine Displacement— All Three Cities
Inertia Test
Weight Group
db)
2500 or
Less
2501 to 3500
3501 to 4500
4501 or
Greater

Engine Displacement
150 CID(a)
or Less
49
46
0
0
95
151 to
259 CID
1
46
7
0
54
260 CID or
Greater
0
27
91
33
151
Total Number
of Vehicles
50
119
98
33
300
(a)
  CID = cubic inch displacement

-------
                 Table 3»  Cross Tabulation of Vehicles by
Engine Displacement and Emission Control System Type — All Three Cities
: ( V
; 150 or Less
151 to 259
260 or More
. Catalyst(b)
Yes
(c)
Secondary Airv '
Yes
15
14
60
89
No 	
15
'"• 23
89
127
•• No •••
Secondary Air
Yes
50
7
2
59
...No
15
10
0
25
Total Number
of -Vehicles
95
54
151
300
(a)
(b)
(c)
CID = cubic inch displacement
Oxidation catalyst
Secondary air injection system
                 Table 4.  Cros.s Tabulation of Vehicles by
 Inertia Test Weight and Emission Control System Type — All Three Cities
Inertia Test
Weight Group
db)
2500 or Less
2501 to 3500 •
3501 to 4500
. Greater than 4501

Catalyst(a)
Yes
Secondary Air* '
Yes
10
35
33
11
89
No
3
44
58
22
127
No
Secondary Air
Yes
23
29
7
0
•.. 59
No
14
11
0
0
25
Total Number
of Vehicles
50
119
98.
33
300
(a)
(b)
Oxidation catalyst
Secondary air injection system

-------
                 Table 5..  Cross Tabulation of Vehicles by
     Fuel System and Emission Control System Type — All Three Cities



Fuel System Type
Fuel Injection
Carburetion


Ye
Seconda
Yes .
1
88
89
Catalys
s
ry Air '
No
3
124
127
5t(a) ,
Nc
Seconda
Yes
11
48
59

>
ry Air
No
15
10
25


Tnf a 1 "NT-ii T-rtVic* t*
of Vehicles
30
270
300
(a)
(b)
Oxidation catalyst

Secondary air injection system
                 Table 6.  Cross Tabulation of Vehicles by
        Inertia Test Weight and Fuel System Type — All Three Cities
Inertia Test
Weight Group
(lb)
0 to 2500
2501 to 3500
3501 to 4500
4501 or Greater

Fuel System Type
Fuel
Injection
19
7
4
0
30
Carburetion
31
112
94
33
270
Total Number
of Vehicles
50
119
98
33
300

-------
                 Table 7.  Cross Tabulation of Vehicles by
       Engine Displacement and Fuel System Type — All Three Cities
CID(a) Group
150 or Less
151 to 259
260 or More

Fuel System Type.
Injection
24
6
0
30
Carburetion
71
48
151
270
Total Number
of Vehicles
95
54"
151
300
(a)
  CID = cubic inch displacement
                 Table 8.  Cross Tabulation of Vehicles by
      Engine Displacement and Transmission Type — All Three Cities
CID(a) Group
150 or Less
151 to 259
260 or More
Transmission Type
Automatic
33
45
151
229
Manual
62
9
0
71
Total Number
of Vehicles
95
54
151
300
(a)
   CID = cubic inch displacement

-------
              Table 9.  Cross Tabulation of Vehicles by
    Inertia Test Weight and Transmission Type — All Three Cities
Inertia Test
Weight Group
(lb)
0 to 2500
2501 to 3500
3501 to 4500
4501 or Greater
Transmission Type
Automatic
10
88
' 98
33
229
Manual
40
' .' 31
1 •'*'• 0,
0
71
Total Number
of Vehicles
50
119
98
33
300
              Table 10.  Cross Tabulation of Vehicles by
      Manufacturer and Engine Displacement— All Three Cities
Manufacturer
General Motors
Ford
Chrysler
American Motors
Da t sun
Toyota
VW, Audi,
Porsche
Honda
Others
Engine Displacement
150 CID(a)
or Less
12
6
0
0
8
9
21
6
33
95
151 to
259 CID
13
13
7
12
1
0
1
0
7
54
260 CID
or Greater
74
42
29
6
0
0
0
0
0
151
Total Number
of Vehicles
99
61
36.
18
• 9
9
22
6
40
300
(a)
  CID = cubic inch displacement

-------
              Table 1U  Cross Tabulation of Vehicles by
          Manufacturer arid Engine Displacement— Chicago
Manufacturer
General Motors
Ford
Chrysler
American Motors
Da t sun
Toyota
VW, Audi,
Porsche
Honda
Others

Engine Displacement
150 CID(a)
or Less
4
2
0
0
2
3
7
2
12
32
1 151 to
259 CID
5
4
3
4
1
0
0
0
2
19
260 CID
or Greater
24
14
9
2
0
0 .
0
0
0
49
Total Number
of Vehicles
33
20
12
6
3
3
7
2
14
100
(a)
  CID = cubic inch displacement

-------
                Table 12.  Cross Tabulation of Vehicles by
             Manufacturer and Engine Displacement — Houston
Manufacturer
General Motors
Ford
Chrysler
American Motors
Datsun
Toyota
VW, Audi,
Porsche
Honda
Others
Engine Displacement
150 CID(a)
or Less
4
2
0
0
3
3
8
2
11
33
151 to
259 CID
4
5
2
4
0
0
0
0
1
16
260 CID
or Greater
25
14
10
2
0
0
0
0
0
51
Total Number
of Vehicles
33
21
12
6
3
3
8
2
12
100
(a)
  CID = cubic inch displacement

-------
             Table 13.  Cross Tabulation of Vehicles by
          Manufacturer and Engine Displacement — Phoenix
Manufacturer
General Motors
Ford
Chrysler
American Motors
Datsun
Toyota
VW, Audi,
Porsche
Honda
Others
-
Engine Displacement
150 CID(a)
or Less
4
2
0
0
3
3
6
2
10
30
151 to
259 CID
4
4
2
4
0
0
1
0
4
19
260 CID
or Greater
25
14
10
2
0
0
0
0
0
51
Total Number
of Vehicles
33
20
12
6
3
3
7
2
14
100
(a)
  CID = cubic inch displacement

-------
               Table 14a.  Federal Short Cycle — Means and
                       Standard Deviations by City
City
Chicago
Hou s to n
Phoenix
No. of
Vehicles
100
100
100
Emission (gm/mi)
HC
Mean
0.854
1. 194
0.934
SD(a)
0. 739
1.204
0.919
CO
Mean
13. 16
17.84
14.62
SD
15. 52
25.69
18.21
NOX
Mean
1..761
1.796
1.818
SD
0.831
0.882
0.940
(a)
   SD = standard deviation
               Table 14b.  Associated Tests of Significance

Emission

HC
CO
NO
X
Log1()(HC)
Log1()(CO)
Log1()(NOx)
Equality of Means


F-Value(a)
3. 33
1.39
0. 11

2.51
0. 10
0.02

Level of
Significance
0.04
0.25
0.90

0.08
0.90
0.98
Homogeneity of Variance


Bartlett
Test
23.6
27. 1
1. 52

0.48
0.98
1.59

Level of
Significance
0
0
0.47

0.79
0.62
0.45
(a)
  Degrees of freedom are 2,  297

-------
  Table 15a.  Federal Short Cycle.—Means "and Standard Deviations by
                Engine Displacement—All Three Cities
CID(a) :
Less than 150
150 to 260 , :
Greater than 260
No, of
Vehicles
• • 95 ,
54
151 >
	 Emission "(grri /mi)"" r"
HC
Mean
0.978
0.781
1.080
SD(b>,
1.020
0.665
1.041
CO
Mean
11.97
12. 33
18.27
SD '
11.80
18.76
24. 30
1 NOX
Mean
l;454
1.845
1.985
'" SD
0.754
0.899
0.895
(a)
(b)
CID = cubic inch displacement
SD = standard deviation
              Table  15b.  Associated Tests of Significance
Emission
HC
CO
NO
Log10(HC)
Log1Q(CO)
Log1()(NOx)
Equality of Means
F-Value(a)
1.87
3. 53 '
11.4
1.29
1.61
13.1
Level of
Significance
0. 16
0.03
0
0.28
0. 20
0
; Homogeneity of Variance
Bartlett
Test
13.9
51.3.
3.67
2.37
49. 1
, 6. 37
;Level of
Significance
0
0
0. 16
0.31
0
0.04
(a)
  Degrees of freedom are 2, 297
                                   10

-------
                      Table I6a.  Federal Three-Mode — Means and
                              Standard Deviations by City

Test
Mode

High Speed


Low Speed


Idle in Drive


Idle in Neutral



City

Chicago
Houston
Phoenix
Chicago
Houston
Phoenix
Chicago
Houston
Phoenix
Chicago
Houston
Phoenix

No. of
Vehicles

100


100


78
75
76
100


Emission (ppm)
HC

Mean
61
40
53
57
45
59
114
104
115
112
114
99

SD(a)
72
46
34
47
43
47
116
125
91
121
151
90
CO

Mean
4439
2345
2864
4036
2784
2442
16015
16687
15324
14049
14018
12546

SD
7907
6202
5571
7116
6427
,5099
22949
21182
21155
20671
18592
19169
NO
X

Mean
1131
1288
1540
649
793
963
160
149
159
74
68
73

SD
549
714
781
485
587
697
226
163
114
79
43
33
(a)
  SD = standard deviation

-------
              Table I6b.  Associated Tests of Significance
Test
Mode
High Speed





Low Speed





Idle in Drive





Idle in Neutral





Emission
HC
CO
NO
Lof (HC)
Log^(CO)
Log1Q(NOx)
HC
CO
NO
Lof (HC)
LogJ^(CO)
Log1Q(NOx)
HC
CO
NO
Lof (HC)
Log "(CO)
Log^NOJ
HC
CO
NO
Log* (HC)
Log|g(CO>
Log10(NOx)
Equality of Means
F-Value(a)
3.87
2.70
9.00
12.4
9.81
7.56
2.71
1.79
6.99
6.61
4.21
6.19
0.23
0.07
0.09
2.63
0. 14
1.72
0.44
0. 19
0. 32
1.07
0. 59
1.85
Level of
Significance
0.02
0.07
0
0
0
0
0.07
0. 17
0
0
0.02
0
0.80
0.93
0.92
0.07
0.87
0. 18
0.65
0.82
0.73
0. 34
0. 55
0. 16
Homogeneity of Variance
Bartlett
Test
57.7
13. 1
12.5
3.0
8. 11
0.06
0.88
10.9
12.7
2.64
10. 3
0. 14
7.61
0.68
33.7
9.77
7.33
9.22
25. 3
1.29
82.2
14. 5
8.62
22.3
Level of
Significance
0
0
0
0.22
0.02
0.97
0.64
0
0
0.27
0.01
0.93
0.02
0.71
0
0.01
0.03
0.01
0
0. 55
0
0
0.01
0
(a)
  Degrees of freedom are 2, 297
                                     12

-------
                    Table  17a.  Federal Three-Mode — Means and Standard Deviations by
                                  Engine Displacement — All Three Cities


Test
Mode
High Speed,


Low Speed


Idle in Drive


Idle in Neutral



(a)
CIDia; Group

Less than 150
150 to 260
Greater than 260
Less than 150
150 to 260
Greater than 260
Less than 150
150 to 260
Greater than 260
Less than 150
150 to 260
Greater than 260


No. of
Vehicles
95
54
151
95
54
151
33
45
151
95
54
151
Emission (ppm)
HC


Mean
58
47
49
60
61
47
148
120
100
109
105
108

SD(b>
41
40
64
44
60
40
155
104
100
134
106
122
CO


Mean
4602
1633
2911
3521
2722
2945
19619
18828
14625
12620
15180
15530

SD
7875
4000
6480
4488
7479
6798
24106
25125
20117
18820
22000
18950
NOX


Mean
1511
1160
1257
834
813
778
125
198
151
68
95
66

SD
856
598
610
562
693
606
94
151
191
33
100
41
u>
         (a)
         (b)
CID = cubic inch displacement
SD = standard deviation

-------
               Table 17b.   Associated Tests of Significance
Test
Mode
High Speed





Low Speed





Idle in Drive





Idle in Neutral





Emission
HC
CO
NOX
Log (HC)
Logj0(CO)
Log1Q(NOx)
HC
CO
NO
Lof (HC)
LogJgfCO)
Log1QNOx)
HC
CO
NO
Lof (HC)
Logjg(CO)
Log10(NOx)
HC
CO
NO
Lof (HC)
Loglg(CO)
Log10(NOx)
Equality of Means
F-Value(a)
0.94
3.80
5.64
2.58
11.9
2. 30
3.20
0. 36
0.26
3.28
9.66
0.20
2.76
0.95
1.95
2.95
3.08
4.58
0.02
0. 30
5.82
0. 30
2.61
4.68
Level of
Significance
0. 39
0.02
-0-
0.08
-0-
0. 10
0.04
0.70
0.77
0.04
-0-
0.82
0.07
0. 39
0. 15
0.05
0.05
0.01
0.98
0. 74
-0-
0. 74
0.08
0.01
Homogeneity of Variance
Bartlett
Test
28.7
26.3
16.3
1.87
2.74
8.99
15.8
22.8
3. 11
1. 55
5. 53
2. 14
12.6
4. 18
21.0
2. 19
4.76
0.48
3.68
2. 15
1 14.0
2. 35.
11.8
6. 30
Level of
Significance
-0-
-0-
-0-
0. 40
0.26
0.01
-0-
-0-
0.21
0.46
0.06
0. 34
-0-
0. 13
-0-
0. 34
0.09
0. 79
0. 16
0. 34
-0-
0. 31
-0-
0.86
(a)
  Degrees of freedom are 2, 297
                                     14

-------
            Table 18a.  FTP Means and Standard Deviations by
                     Mileage Group — All Three Cities
Mileage
Group (a)
4000 Miles or Less
Greater than
4000 Miles
No. of
Vehicles
70
230
FTp(t>) Emission
(gm/mi)
HC
Mean
1.09
1.45
SD
0.657
1. 141
CO
Mean
18.66
25.71
SD
14.80
26.85
NOX
Mean
2. 11
2.48
SD
0.801
1.005
 (a)
 (b)
 (c)
Number of accumulated miles on vehicles tested
Federal emission certification test procedure
SD = standard deviation
                Table 18b.  Associated Tests of Significance



HC
CO
NO
X
Log1Q (HC)
Log1Q (CO)
Log,» (NO )
1U X
Equality of Means

F-Value*a)
6.29
4.42
8.08

7.59
1.89
7.01

Level of
Significance
0.01
0.04
-0-

0.01
0. 17
0.01

Homogeneity of Variance

Bartlett
Test
26.0
29.8
4.97

0. 385
5.60
0.398

Level of
Significance
1 -0-
-0-
0.03

0.54
0.02
0. 53

(a)
  Degrees of freedom are 1, 298
                                     15

-------
                       Table 19.   Correlation Coefficient Summary by
                           Accumulated Mileage — All Three Cities
Short
Test
Federal
Short
Cycle

Federal
Three-
Mode




Mileage
Group*3-)
4000 or
less
Greater
than 4000
4000 or
less

Greater
than 4000


Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
70

230

70
7.0
48
70
230
230
181
230
ST/FTP*b) Correlation Coefficients *c)
HC
0.91

0.76

0.13*
0.29
0.80
0.37
0.33
0.52
0. 54
0.45
CO
0.80

0.90

0. 16*
0.31
0.61
0.55
0.37
0.54
0.67
0.71
NO
X
0.79

0.82

0.36
0.40
-0.08*
0.02*
0. 54
0.49
0.23
0. 14*
Log
(HC)
0.85

0.78

0.21
0.33
0.51
0.37
0. 34
0.46
0.69
0.63
Log
(CO)
0.82

0.82

0. 17
0.21
0.68
0. 54
0.24
0.38
0.75
0.70
Log
(N0x)
0.84

0.82

0.42
0.47
-0. 30
0. 12
0. 59
0. 57
0.31
0,27
(a)
(b)
(c)
Number of accumulated miles on vehicles tested
ST/FTP  =  short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence level
except where indicated by an asterisk

-------
        Table 20a.  FTP Means and Standard Deviations —by City
City
Chicago.
Houston
Phoenix
No. of
Vehicles
100
100
100
FTp(a) Emission (gm/mi)
HC
Mean
1.283
1.534
1.288
SD
0.834
1.335
0.930
CO
Mean
21.665
27.598
22.934
SD
17. 145
32.812
21.378
NO
X
Mean
2.375
2.446
2.357
SD
0.896
1.050
0.974
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
              Table 20b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1() (HC)
Log1Q (CO)
Log1Q (N0x)
Equality of Means
F- Value (a)
1.841
1.602
0.234
0.941
0. 185
0. 135
Level of
Significance
0. 160
0.203
0.792
0.392
0.831
0.874
Homogeneity of Variance
Bartlett
Test
25. 162
44.224
2.506
5.423
12.967
2.333
Approximate Level
of Significance
0.0
0.0
0.287
0.067
0.002
0.313
 (a)
   Degrees of freedom are 2,  297
                                   17

-------
     Table 21.   Correlation Coefficient Summary —All Three Cities
Short Test
Federal
Short Cycle
Federal
Three-
Mode


Test Mode


High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
300

300
300
300
229
ST/FTP^a) Correlation Coefficients ^b)
HC
0.78

0.31
0.47
0.45
0.58
CO
0.89

0.33
0.46
0.69
0.67
NO
X
0.81

0.52
0.48
-i,
0. 11""
0. 12*
Log
(HC)
0.80

0.31
0.43
0.58
0.66
Log
(CO)
0.82

0.22
0.35
0.67
0.74
Log
(N0x)
0.82

0. 56
0.55
0.23
0.27
(a)
(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence
level except where indicated by an asterisk
          Table 22.   Correlation Coefficient Summary — Chicago
Short Test
Federal
Short Cycle
Federal
Three-
Mode


Test Mode


High speed

Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
100

100

100
100
78
ST/FTP(a) Correlation Coefficients^'
HC
0.88

0.12*

0.38
0.42
0.52
CO
0.82

0.31

0. 52
0.52
0.50
NO
X
0.81

0.43

0.40
0.03V
0.01*
Log
(HC)
0.81

0.33

0.45
0. 50
0. 54
Log
(CO)
0.80

0. 34

0. 39
0. 57
0.63
Log
(N0x)
0. 85

0. 51

0.47
0.27
0.22*
 (a)
 (b)
 ST/FTP = short test/federal emission certification test procedure

'correlation coefficients are statistically significant at the 95 percent confidence
 level except where indicated by an asterisk
                                         18

-------
         Table 23.   Correlation Coefficient Summary — Houston
Short Test
Federal
Short Cycle
Federal .
Three-
Mode


Test Mode


High speed
Low speed
Idle in neutral
Idle in drive
No. -of
Vehicles
100
<
100
100
100
75
ST/FTP(a) Correlation Coefficient (b)
HC
0.91

0.66
0. 76
0. 47
0.60
CO
0.97

0.45
0.53
0.79
0.76
NO
X
0.78

0.44
0.41
0. 34
0. 36
Log
(HC)
0.88

0.42
0.58
0.67
0. 76
Log
(CO).
0.87

0. 19*
0.33
0.68
0. 79 •
Log
(N0x)
0.81

0.52
0.56
0. 3Z
0. 46
(a)
(b)
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence
level except where indicated by an asterisk
         Table 24.  Correlation  Coefficient Summary — Phoenix ,
Short Test
Federal
Short Cycle
Federal
Three -
Mode


•Test Mode


High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
100

100
100
100
76
ST/F.TP(a) Correlation Coefficient (b>
HC
0.45

0.25
0.28
0.44
0.71
CO
0.76

0.32
0.41
0.80
0.81
NO
X
0.85

0.73
0.66
0.03*
*
-0.05
Log
(HC)
0.73

0.27
0.34
0.61
0.75'
Log
(CO)
0.79

0.20
0.36
0.74
0.78
Log
(N0x)
0.82

0.73
0.66
0.08*
0.03*
(a)
(b)
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence
level except where indicated by an asterisk
                                         19

-------
          Table 25a.  FTP Means and Standard Deviations by
               Engine Displacement — All Three Cities

^a)
Group

150 and
Less
151 to
259
260 and
Greater

No. of
Vehicles

95

54

151

. FTP* ' Emission (gm/mi)
HC
Mean
1.327

1.271

1.429

SD(C)
0.712

1.021

1.241

CO
Mean
16.789

21.718

29.484

SD
11.065

19.348

30.744

NO
x
Mean
2.095

2.428

2.568

SD
0.939

0.838

0.999

(a)
(b)
(c)
CID = cubic inch displacement
FTP = federal emission certification test procedure
SD = standard deviation
              Table 25b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1Q (HC)
Log1() (CO)
Log,n (NO )
1 U X
Equality of Means
F-Value(a)
0.547
8.373
7.235
0.319
2.832
10.133
Level of
Significance
0.579
0.0
0.001
0.727
0.061
0.0
Homogeneity of Variance
Bartlett
Test
31.651
97.662
2.336
3.277
25.554
10.476
Approximate Level
of Significance
0.0
0.0
0.313
0. 196
0.0
0.006
(a)
  Degrees of freedom are 2, 297
                                   20

-------
                         Table 26.  Correlation Coefficient Summary by
                             Engine Displacement — All Three Cities
Short Test
Federal
Short Cycle


Federal
Three-Mode




CID(a) Group
150 or Less

151 to 259
260 or More
150 or Less

151 to 259
260 or More

Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
95

54
151
95
95
95
33
54
54
54
45
151
151
151
151
ST/FTP(6)
Correlation Coefficients
HC
0.71

0.53
0.89
0.25
0.41
0.45
0.75
0.25*
0.36
0.38
0.78
0. 34
0.62
0.49
0.60
CO
0.63

0.90
0.92
0.40
0. 36
0.63
0.43
0.34
0.72
0.65
0.66
0.43
0.47
0.82
0.80
NO
X
0.82

0.83
0.79
0.61
0.63
0.06*
0. 30*
0. 59
0.74
0.04*
0.13*
0. 58
0. 36
0.24
0. 10*
Log(HC)
0.81

0.74
0.82
0. 35
0.45
0.48
0.57
0.23*
0.41
0.49
0.7.3
0. 30
0.44
0.65
0.69
Log(CO)
0.-76

0.84
0.86
0. 19*
0. 35
0. 55
0. 54
-0.03*
0.27
0.58
0.75
0. 37
0.46
0.79
0.80
Log(NOx)
0.83

0.84
0.78
0.68
0.71
0.20
0.41
0.45
0.71
0.22*
0.28*
0.60
0.42
0. 30
0.23
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an  asterisk

-------
            Table 27a.  FTP Means and Standard Deviations by
                  Inertia Test Weight— All Three Cities
Inertia Test
Weight Group
(lb)
2500 and Less
2501 to 3500
3501 to 4500
Greater than
4500
No. of
Vehicles
50
119
98
33
, FTP(a
(8m
HC
Mean.
1. 382
1 . 29 1
1. 309
1.799
-SD
1.020
0.921
1. 191
1. 107
Emission
/mi)
CO
Mean
15.071
20.909
25.859
43.753
SD
10.563
22. 370
24.066
37.019
NO
X
Mean
1.868
2.486
2. 557
2. 365
SD
0.763
1.043
0.944
0.829
(a)
(b)
FTP = federal emission certification test procedure

SD  = standard deviation
                Table 27b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1()(HC)
Log1()(CO)
Log1()(NOx)
Equality of Means
F-Value(a*
2. 162
10.996
6.498
3. 101
8.225
8. 305
Level of
Significance
0.093
' 0.0
0.0
0.027
0.0
0.0
Homogeneity of Variance
Bartlett
Test
7. 324
58.248
7. 384
0.413
8.787
7. 120
Approximate
Level of
Significance
0.064
0.0
0.062
0.938
0.033
0.070
(a)
   Degrees of freedom are 3, 296
                                     22

-------
                        Table  28.  Correlation Coefficient Summary by
                          Inertia  Test Weight Group — All  Three Cities
Short Test
Federal
Short Cycle



Federal
Three-Mode















Inertia Test
Weight Group
db)
0 to 2500

2501 to 3500
3501 to 4500
4501 or More

0 to 2500



2501 to 3500



3501 to 4500



4501 or More



Test Mode






High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
50

119
98
33

50
50
50
10
119
119
119
88
98
98
98
98
33
33
33
33
ST/FTP(a) ,
Correlation Coefficients^ '
HC
0.49

0.76
0.92
0.73

0.20*
0.33
0.32
0. 37*
0.41
0.55
0.49
0.66
0.36
0.5-1
0.42
0.53
0.28*
0.48
0.56
0.64
CO
0.73

0.86
0.91
0.88

0.53
0.42
0.56
0.05*
0. 50
0.61
0.59
0. 58
0.28
0.45
0.86
0.86
0.29*
0. 35
0.75
0.76
NO
X
0.76

0.81
0.81
0.77

0.54
0.50
0.08'1'
0. 33*
0.62
0.64
0.09*
0. 18*
0.54
0.33
0.20
0.05*
0. 53
0.30*
0.30*
0. 10*
Log(HC)
0.74

0.83
0.87
0.63

0.39
0.49
0. 51
0.58*
0. 30
0.46
0. 57
0.72
0.33
0.41
0.58
0.63
0.34*
0.47
0.64
0.71
Log (CO)
0.71

0.82
0.88
0.81

0.46
0. 36
0.47
-0. 13*
0.20
0.45
0.70
0.81
0.32
0.40
0.80
0.80
0.26*
0.39
0.62
0.72
Log(NOx)
0.78

0.84
0.80
0.81

0.64
0.68
0. 18*
0.40*
0.61
0.66
0.31
0. 31
0.57
0.44
0.24
0. 14*
0.60
0.43
0.36
0.31*
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
           Table 29a.  FTP Means  and Standard Deviations by
           Emission Control System  Type — All Three Cities
Catalyst
Yes
Yes
No
No
Secondary
Air Injection
Yes
No
Yes
No
No. of
Vehicles
89
127
59
25
FTp(a) Emission
(gm/mi)
HC
Mean
1. 118
1.498
1.284
1.799
SD
1. 144
.1.078
0.641
1.241
CO
Mean
17.493
34.025
15.213
17. 767
SD.
18.083
31. 538
7,607
11.844
NO
X
Mean
2.670
2.283
2.246
2. 312
SD
1. 150
0.832
0.998
0.700
(a)
(b)
FTP = federal emission certification test procedure

SD = standard deviation
               Table 29b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1()(HC)
Log1Q(CO)
Log10(NOx)
Equality of Means
-F-Value(a)
3. 905
13. 533
3. 539
7. 873
11. 755
2.252
Level of
Significance
0.009
0.0
0.015
0.0
0.0
0.082
Homogeneity of Variance
Bartlett
Test
23.895
130.884
15. 31.1
7.838
27. 375
13.093
Approximate
Level of
Significance
0.0
0.0
0.002
0.051
0.0
0.005
 (a)
    Degrees of freedom are 3,  296
                                     24

-------
                                      Table 30.   Correlation Coefficient Summary by
                                    Emission Control System Type — All Three Cities
Short Test
Federal
Short Cycle


Federal
Three-Mode














Catalyst/Air(a)
Yes/Yes
Yes /No
No/Yes
No /No
Yes/Yes



Yes /No



No /Yes



No /No



Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
89
127
59
25
89
89
89
76
127
127
127
117
59
59
59
24
25
25
25
12
ST/FTP(b)
Correlation Coefficients
HC
0.94
0.84
0.74
0. 32*
0.73
0.59
0.25
0.46
0. 13*
0.50
0.56,
0.62
0. 34
0.36
0.40
0.85
-0.01*
0. 16*
0.31*
0.86
CO
0.86
0.92
0. 54
0.86
0. 39
0. 54
0.80
0.77
0. 35
0.51
0.65
0.67
0.40
0.52
0.50
0.40*
0.39*
-0.02*
0.69
0.88
NO
X
0.83
0.77
0.86
0.90
0.65
0.60
0.05*
-0.05*
0.48
0.47
0. 12*
0.22
0.65
0.74
0.29
0.55
0.28*
0.47
0.40
0.51*
Log(HC)
0.85
0.82
0.78
0.54
0. 38
0.42
0. 35
0.40
0.23
0.46
0.71
0.75
0.41
0.42
0.41
0.58
0.15*
0.36*
0.42
0.77
Log(CO)
0.83
0.87
0.75
0. 76
0.62
0.66
0.76
0.82
0.22
0.41
0.68
0.72
0.41
0. 39
0.40
0. 58
0. 32*
0. 12*
0.76
0.70
Log(NOx)
0.83
0.76
0.90
0.89
0.70
0. 59
0. 13*
0. 15*
0. 51
0. 55
0.22
0.26
0.70
0.77
0. 50
0.62
0.08*
0.41
0.46
0.51*
ts)
(Jl
           (a)
           (b)
Secondary air injection system
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an  asterisk

-------
            Table 3la.  F-TP Means and Standard Deviations by:
                  Fuel System Type - All Three Cities
Fuel System
Type
Fuel Injection
Carburetion
No. of
Vehicles
30
270
FTp(a) Emission
(gm/mi)
HC
Mean
1.598
1.344
SD
1.224
1.038
CO
Mean
17. 158
24.833
SD
12.181
25.650
NOX
Mean
2.036
2.433
SD
0.647
0.995
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
                Table 31b. Associated Tests of Significance
Emission
HC
CO
N0x
Log1Q (HC)
Log1Q (CO)
L°glO (NO*>
Equality of Means
F- Value (a)
1 . 452
2.614
4 . 534
2.336
1.706
2 . 892
Level of
Significance
0.229
0. 107
0.034
0.128
0. 193
0.090
Homogeneity of Variance
Bartlett
Test
1.550
19.842
7.740
0.043
1.469
3.600
Approximate
Level of
Significance .
0.216
0.0
0.006
0.837
0.228
0.059
 (a)
    Degrees of freedom are 1, 298
                                    26

-------
                                     Table  32.   Correlation Coefficient Summary by
                                          Fuel System Type — All Three Cities

Short
Test
Federal
Short
Cycle
Federal
Three-Mode




Fuel
•System Type
Fuel Injection
Carburetion

Fuel Injection

Carburetion


Test
Mode


High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

- No. of
Vehicles
30
270

30
30
30
11
270
270
270
218
ST/FTP(a)
Correlation Coefficients
HC
0.45
0.82

0.11*
0.22*
0.36
0.80
0.32
0.50
0.46
0.58
CO
0.79
0.89

0.47
0.11 *
0.61
0. 54*
0.34
0.49
0.70
0.69
NOX
0.62
0.82

0.63
0.33*
-0.20*
-0.07*
0.53
0.49
0.13
0.11*
Log (HC)
0.78
0.81

0.35*
0.38
0.51
0.64
0.29
0.43
0.58
0.66
Log (CO)
0.90
0.82

0.52
0.45
0.75
0.58*
0.23
0.37
0.68
0.76
Log (NOX)
0.57
0.84

0.59
0.42
-0. 19*
-0.01*
0.57
0.57
0.27
0.27
ts)
         ^a'ST/FTP = short test/federal emission certification test procedure
         (b)
            Correlation coefficients are statistically significant at the  95 percent confidence level except where indicated
            by an asterisk

-------
              Table 33a.  FTP Means and Standard Deviations by
                    Transmission Type - All Three Cities
Transmission
Type
Automatic
Manual
No. of
Vehicles
229
71
FTp(a) Emission
(gm/mi)
HC
Mean
1.345
1.442
SD
1.086
0.967
CO
Mean
26.301
16.854
SD
26.979
13.142
NOX
Mean
2.513
2.005
SD
0.983
0.834
(a)

(b)
FTP = federal emission certification test procedure

SD = standard deviation
                Table 33b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1() (HC)
Log1() (CO)
Log1() (N0x)
Equality of Means
F- Value (a*
0.045
8.096
15.558
2.052
4.379
17.961
Level of
Significance
0.502
0.005
-0-
0. 153
0.037
-0-
Homogeneity of Variance
Bartlett
Test
1.399
41.861
2.729
0.758
11.922
0.992
Approximate
Level of
Significance
0.238
-0-
0.099
0.385
0.001
0.321
  (a)
    Degrees of freedom are 1,  298
                                    28

-------
                                    Table 34.  Correlation Coefficient Summary by

                                         Transmission Type —All Three Cities

Short
Test
Federal
Short
Cycle
Federal
Three -Mode







Transmis sion
Type
Automatic
Manual

Automatic



Manual




Test
Mode



High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

No. of
Vehicles
229
71

229
229
229
229
71
71
71

ST/FTP(a) .
Correlation Coefficients
HC
0.87
0.42

0.31
0.48
0.49
0.58
0.28
0.42
0.33
--
CO
0.91
0.66

0.33
0.48
0.71
0.67
0.50
0.56
0.67
--
NOX
0.80
0.82

0.57
0.49
0.11*
0. 12*
0.65
0.56
0.09*
--
Log (HC)
0.83
0.72

0.29
0:41
0.60
0.66
0.33
0.47
0.48
--
Log (CO)
0.85
0.75

0.27
0.39
0.70
0.74
0.14*
0.40
0.60
--
Log (NOX)
0.81
0.82

0.57
0.55
0.23
0.27
0.72
0.71
0.33
--
to
sO
           (a)

           (b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where indicated
by an asterisk

-------
             Table 35a.  FTP Means and Standard Deviations by
                      Manufacturer  — All Three Cities
Manufacturer
General Motors
Ford
Chrysler
American Motors
Others
No. of
Vehicles
99
61
36
18
86
FTp(a) Emission
(gm/mi)
: HC
. Mean
1.378
r.245
1.726
1.052
1.361
SD
1. 180
1.003
1.248
0.583
0.911
CO
Mean
26. 165
22.550
45.648
13.076
15,990
SD
24.772
26. 106
36.359
9.466
10.254
NOX
Mean
2.400
2.742
2.352
2.764
2.076
SD
0.956
1.045
0.811
0.696
0.957
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
                Table 35b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Logjg (HC)
Log1() (CO)
Log1Q (N0x)
Equality of Means
F- Value (a)
1.649
11.747
5.204
2.238
8.926
7.087
Level of
Significance
0.162
0.0
0.001
0.065
0.0
0.0
Homogeneity of Variance
Bartlett
Test
16.848
108.921
5.456
3.529
21.243
19.122
App r oximate
Level of
Significance
0.002
0.0
0.249
0.479
0.0
0.001
  (a)
    Degrees of freedom are 4,  295
                                    30

-------
                Table 36.  Correlation Coefficient Summary by
                         Manufacturer —All Three  Cities
Short
Test
Federal
Short
Cycle




Federal
Three-
Mode



















Manu-
facturer
General
Motors
Ford
Chrysler
Ame rican
Motors
Others
General
Motors


Ford



Chrysler




American
Motors


Others



Test Mode







High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
99

61
36
18

86
99
99
99
92
61
61
61
55
36
36
36
36

18
18
18
16
86
86
86
30
ST/FTP(a) Correlation Coefficients^'
HC
0.90

0.94
0.81
0.94

0. 51
0.35
0.61
0. 50
0.55
0.33
0.40
0.42
0.63
0.48
0.60
0.54
0. 56
;*{
0.23
0.58
0. 56
0.81
0.21*
0. 31
0.33
0.82
CO
0.90

0.96
0.94
0.91

0.65
0.29
0.65
0.70
0.73
0.43
0.53
0.85
0. 84
0.57
0.64
0.65
0.68

0. 68
0.60
0.84
0.84
0.47
0.49
0. 53
0. 32"
NO
X
0.92

0.78
0.77
0.93

0.85
0.42
0.55
0. 17*
0.28
0.65
0.67
-0. 10*
-0.25*
0.65
0. 58
0.09*
-0.12*

0.72
0.39*
0.24*
0.29*
0.64
0.69
0. 11*
0.60
Log
(HC)
0.88

0.87
0.72
0.97

0.77
0.34
0.53
0.73
0.77
0.26
0.35
0. 50
0. 55
0. 30*
0.39
0.44
0.42
*
0.23
0. 59
0. 76
0.90
0. 34
0.42
0.41
0. 58
Log
(CO)
0.88

0.85
0.90
0.90

0.76
0.30
0. 55
0. 75
0.81
0.38
0. 51
0.69
0.73
0.21*
0. 27*
0.57
0.63

0.80
0.78
0.91
0.85
0.47
0. 46
0. 50
0.48
Log
(N0x)
0.91

0.77
0.69
0.95

0.84
0.50
0.60
0.33
0.32
0.66
0.58
-0.01*
-0.09*
0.53
0.64
0.26*
0.12*

0. 51
0.38*
0.26*
0.37*
0.69
0.73
0.25
0.58
(a)
  ST/FTP = short test/federal emission certification test procedure

  Correlation coefficients are statistically significant at the 95 percent confidence
  level except where indicated by an asterisk
                                            31

-------
Table 37.   Correlation Coefficient Summary for  150 CID
                        Manufacturer — All Three Cities
                                                                    (a)
                                                                      and Less by
Short
Test
Federal
Short
Cycle






Federal
Three -
Mode


























Manu-
facturer
General
Motors
Ford
Datsun
Toyota
VW, Audi,
Porsche
Honda
Others
General
Motors


Ford



Datsun



Toyota



VW, Audi,
Porsche


Honda



Other



Test Mode









High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
12

6
8
9
21

6
33
12
12
6
12
6
6
3
6
8
8
2
8
9
9
2
9
21
21
5
21
6
6
0
6
33
33
15
33
ST/FTp'b' Correlation Coefficients'0 '
HC
0.84

0.95
0.95
0.88
0.88

0.20*
0.74
0.25"
0.61
0.88
0. 30"
-0.17*
-0.22*
-0.81"
0.86
0.53"
0.61*
...
0.40"
0.69
0.87
....
0.66*
0.20*
0.53
0.77"
0.60
0.89
0.23*
	
0.94
0. 15"
0.25
0. 84
0. 37
CO
0.89

0.83
0.93
0. 50 ""
0.71

-0.21*
0.48
0. 15*
0.70
0.96
0.66
0.55*
0.76*
0.93"
0.07"
0.93
0.93
...
0.42"
-0.07"
-0. 17"
	
0.42"
0. 59
0.43"
0. 10"
0.53
0.29*
-0.26*
	
0. 35*
0. 47
0. 51
0. 44
0. 41
NO
X
0.94

0.35"
0.89
0.82
0.49

0.74*
0.91
0.84
0.32*
0.04*
0. 19"
0.17*
0.61*
0.74*
0.31*
-0.49*
0.72
...
-0.07*
0. 56"
0.81
	
0.00*
0.77
0.66
0.77*
0.24"
0.66*
0.24*
	
0.62"
0.61
0.6M
0. 55
0. 57
Log
(HC)
0.87

0.94
0.98
0.84
0.81

0. 15*
0.80
0.46"
0.68
0.92
0.67
-0. 10*
-0. 16*
0.89"
0.60*
0.76
0.77
...
0.70*
0. 53"
0.84
	
0.28*
0. 41 "
0.60
0.71*
0.64
0. 82
0. 18"
	
0. 90
0.20*
0.27 '
0. 51
0. 56
Log
(CO)
0.89

0.95
0.92
0. 52*
0.68

-0. 18*
0.85
0.05"
0.67
0.93
0.80
0.79*
0.91
1.0*
0. 56*
0.91
0.95
...
0.74
-0. 13*
-0.25*
	
0. 33*
0. 53
0. 42"
-0.23*
0. 33*
0. 30"
-0.22*
	
0. 43"
0. 53
0. 49
0. 71
0. 59
Log
(N0x)
0.93

0.46
0.87
0.80
0.67

0.66*
0.94
0. 81
0. 54"
0.02*
0. 14"
0.26*
0. 55*
0. 55*
0.04*
-0. 50*
0.68
...
0.05*
0. 50"
0.79
	
-0. 02
0.78
0. 78
0.67'
0. 31*
0. 56*
0. 18
	
-0. 17
0.60
0. 71
0. 44
0. 55
(b)
(c)
CID = cubic inch displacement
ST/FTP = short test/federal emission certification tost procedure--
Correlation coefficients are statistically significant at the 95 percent confidence
level except where indicated by an asterisk
                                           32

-------
   Table  38.   Correlation Coefficient Summary for 151 to  259 CID
                         Manufacturer —All  Three Cities
                                                                                   (a)
                                                                                     by
Short
Test
Federal
Short
Cycle





Federal
Three-
Mode


















Manu-
facturer
General
Motors

Ford
Chrysler
American
Motors
Other
General
Motors


Ford



Chrysler



American
Motors


Other*



Test Mode








High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
13


13
7
12

9
13
13
12
13
13
13
10
13
7
7
7
7
12
12
10
12
9
9
6
9
ST/FTP(b) Correlation Coefficients (c)
HC
0.93


0.92
0.80
0.88

0. 18*
0.13*
0.50*
0.79
0.82
0.79
0.79
0.21*
0.73
0.70*
0.71*
0.89
0.87
0.03*
0.35*
0.65
0.27*
0. 13"'
0.07*
0.68"
0.02*
CO
0.98


0.93
0.94
0.89

0.90
-0.40*
0.87
0.60
0.58
0.91
0.91
0.55*
0.88
0.09*
0. 18*
0.89
0.85
0.60
0.15*
0.86
0.89
0.08*
0.86
0.93
0.92
NO
X
0.90

k
0.44*
0.95
0.92

0.90
0.05*
0. 81
-0. 12*
-0.09*
0.67
0.70
-0.08*
-0.35*
0.74*
0.94
0.09*
0. 18*
0.69
0. 49'"
0.32*
0.26*
0.72
0.78
0.85
0.78
Log
(HC.)
0.90


0.84
0.71
0.90

0.67
-O.Ol"'
0.57
0.84
0.68
0.37*
0. 46*
0.31*
0.42*
0. 54*
0. 53*
0.85
0.78
0.13*
0.40"°
0.61*
0.29*
0.26'"
0.24*
0.46*
0. 17*
Log
(CO)
0.94


0.89
0.73
0.84

0.91
-0.48*
0.38*
0.80
0.41""
0.36*
0.41*
0.73
0.61
0. 18*
0.25*
0. 49""
0.44*
0.38*
0. 10*
0.67
0.81*
0.50*
0.85
0.93
0.89
Log
(N0x)
0.94


0. 57
0.92
0.95

0.87
0.03*
0.77
0. 22*
0.23*
0.67
0.71*
-0.09*
-0.32*
0. 54*
0.90*
0.25*
0.38*
0.43*
0.56*
0.46*
o. 29"
0. 58*
0. 54*
0.72*
0.68
(b)
(c),
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure
  Correlation coefficients are statistically significant at the 95 percent confidence
  level except where indicated by an asterisk
  "Other" category includes one Datsun and one Porsche
                                             33

-------
    Table 39.   Correlation Coefficient Summary for  260 CID*a' and
                Greater by Manufacturer —All Three Cities
Short
Test
Federal
Short
Cycle



Federal
Three-
Mode













Manu-
facturer
General
Motors
Ford
Chrysler
American
Motors
Gene ral
Motors



Ford



Chrysler



Ame rican
Motors


Test Mode






High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
74

42
29
6

74
74
74
74
42
42
42
42
29
29
29
29
6
6
6
6
ST/FTP(b) Correlation Coefficients (c)
HC
0.91

0.95
0.80
0.86

0.37
0.71
0. 59
0.51
0.20*
0.40
0.64
0.37
0.46
0.58
0.59
0. 54
0. if"
-0.02*
0. 36""
0.42"
CO
0.88

0.97
0.94
0.54*

0.40
0.63
0.79
0.84
0. 38
0.67
0.84
0.85
0.55
0.64
0.74
0.72
0. 30"
0. 43'"
0.36*
0.36*
NO
X
0.92

0.88
0.78
0.95

0. 56
0.68
0. 37
0.42
0.65
0.73
-0. 19*
o.of'
0.62
0. 52
-0. 16'"
0.06"
0.93
:'(
0. 33
0.82
0.66*
Log
(HC)
0.90

0.88
0.71
0.89

0.37
0. 52
0.80
0.75
0.17*
0. 35
0. 59
0. 50
0.26*
0. 34*
0. 40
0.41
0.07"
-0. 10*
0.33*
0 . 4 1 "
Log
(CO)
0.87

0. 82
0.91
0. 38*

0.42
0. 57
0,83
0. 80
0. 38
0.60
0.74
0.76
*
0. 11
0. 20*
0.72
0.67
0.60"
0.69*
0.68*
0.68*
Log
(N0x)
0.90

0. 87
0.73
0.96

0. 59
0. 70
0.40
0.43
0.66
0.64
0.03*
*
0. 11
0. 53
0.64
0.05*
0.20*
0.96
0.33*
0.73*
0.62*
(a)

(b)

(O
CID - cubic inch displacement

ST/FTP - short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level
except where indicated by an asterisk
                                          34

-------
           Table 40a.  FTP Means and Standard Deviations by
                    Engine Displacement — Chicago

CID(a)


150 and
Less :
151 to
259
260 and
Greater

No. of
Vehicles


32

19

49

FTp(b) Emissions (gm/mi)
HC

Mean
1.346

1.071

1.324

(r )
SDIC'
0.469

0.421

1.098

CO

Mean
17.089
-:
19. 333

25.557


SD
7.928

12.558

21.794

NO
X

Mean
2.241

2.294

2.494


SD
0.883

0.698

0.969

(a)
(b)
(c)
CID = cubic inch displacement
FTP = federal emission certification test procedure
SD = standard deviation
              Table 40b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1Q (HC)
Log10 (CO)
Log1Q (N0x)
Equality of Means
F- Value (a)
0. 764
2.664
0.868
1.745
0.692
0. 845
Level of
Significance
0.469
0.075
0.423
0. 180
0.503
0.433
-Homogeneity of Variance
Bartlett
Test
34.475
32.790
2.537
13.481
10.203
1.867
Approximate Level
of Significance
0.0
0.0
0.287
0.001
0.007
0.399
(a)
  Degrees of freedom are 2, 97
                                   35

-------
OO
                                      Table 41.   Correlation Coefficient Summary by
                                               Engine  Displacement— Chicago
Short
Test
Federal
Short
Cycle



Federal
Three-
Mode










CID(a)
Group
150 or
less
151 to
259
260 or
more
150 or
less


151 to
259


260 or
more



Test Mode






High speed • •
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed

Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
32

19

49

32
32
32
11
19
19
19
18
49

49
49
49
ST/FTP^b) Correlation Coefficients (c)
HC
0.84

0.82

0.91

0.43
0.59
0.31*
0.55*
0.25*
0.36*
0.64
0.71
0.06*

0.43
0.55
0.61
CO
0.77

0.86

0.84

0.51
0.44
0.58
0.52*
0.32*
0.67
0.74
0.71
0.35

0.52
0.76
0.80
NO
X
0.90

0.79

0.76

0.41
0.45
0.12*
-0.03*
0.31*
0.66
0.02*
0.20*
0.57

0.33
0.09*
-0.01*
Log
(HC)
0.75

0.81

0.84

0.40
0.59 .
0.45
0.44*
0.22*
0. 19*
0.55 '
0.64
0.32

0.46
0.50
0.55
Log
(CO)
0. 73

0.90

.0.83

0.24*
0.17*
0.51
:0.27*
0.31*
0.43*
0,62
.0.70
0.44

0.52
0.69
0.73
Log
(N0x)
.0.92

0.81

0.81

0*49
•0.56
J0.29*
-0.04*
0.38*
0.65
:o.n*
;0.28*
0.62

0.39
0.35
0.29
                 (a)
                 (b)
                 (c)
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence level
except where indicated by an asterisk

-------
            Table 42a.  FTP Means and Standard Deviations by
                       Inertia Test Weight— Chicago


Inertia Test
Weight Group
db)

2500 and Less
2501 to 3500
3501 to 4500
Greater than
4500


No. of
Vehicles


15
43
30
12

(a)
FTPV 'Emission
(gm/mi)
HC


Mean
1.35
1. 11
1.49
1. 31

Ch\
SD(b)
0.463
0. 527
1.270
0.653

CO


Mean
15. 1
17.4
27.9
29.4


SD
7. 11
12. 3
27.9
20.2

NO
X

Mean
2.11
2.44
2.46
2.26


SD
0.786
0.914
0.943
0.865

(a)
(b)
FTP = federal emission certification test procedure

SD = standard deviation
                Table 42b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1()(HC)
Log1()(CO)
Log10(NOx)
Equality of Means
F-Value^
1.33
4. 13
0.67
1. 11
4.55
0.69
Level of
Significance
0.269
0.008
0. 572
0. 349
0.005
0. 558
Homogeneity of Variance
Bartlett
Test
35. 1
25.2
0.657
5.56
3. 30
0. 346
App roximate
Level of
Significance
0.0
0.0
0.887
0. 142
0. 358
0.953
(a)
  Degrees of freedom are 3, 96
                                     37

-------
                                       Table 43.   Correlation Coefficient Summary by
                                             Inertia Test Weight Group  - Chicago
Short Test
Federal
Short Cycle



Federal
Three-Mode















Inertia Test
Weight Group
(lb)

0 to 2500
2501 to 3500
3501 to 4500
Test Mode




4501 and more

0 to 2500



2501 to 3500



3501 to 4500



450 1 and mor




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
5 High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles

15
43
30
12

15
15
15
4
43
43
43
32
30
30
30
30
12
12
12
12
ST/FTP(a) . .
Correlation Coefficient
HC

0.80
0.87
0.93
0.77

0.27*
0.51
0. 14*
-0.77*
0.46
0. 52
0.61
0.79
0.01*
0.38
0. 54
0.59
0.62
0 . 54*
0. 19*
0.43*
co

0.64
0.83
0.88
0.67

0. 17*
0. 32*
0.45*
0.28*
0. 34
0.55
0.63
0.66
0. 28*
0.66
0.82
0.83
0.47*
0.42*
0. 50*
0.64
NO
X

0.91
0.78
0.77
0.90

0.26*
0.40*
-0.01*
-0.46*
0. 50
0.40
0.05*
0.33*
0.61
0.40
-0. 17*
-0.22
0.58
0. 50*
0.60
0. 56*
Log(HC)

0. 70
0.83
0.88
0.67

0. 13*
0.48*
0.22*
-0. 69*
0.41
0.44
0.66
0.75
0.26*
0.46*
0.46
0.49
0.67'
0. 56*
0. 33*
0. 54*
Log(CO)

0.71
0.84
0.81
0.78

0..08*
0 . 0 3*
0. 39*
-0.09*
0;4'1
0.48
0.77
0.83
0. 20*
0. 38
0.66
0.65
0.55*
0.58
0.53*
0.65
Log(NOx)

0.94
0.84
0.77
0.95

0. 37*
0. 53
0.22*
-0. 38*
0.56
0.49
0. 26*
0. 33*
0.61
0.46
0.01*
-0. 07*
0.65
0.49*
0.79
0.76
OJ
00
          (a)
          (b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
            Table  44a.  FTP Means and Standard Deviations by
                 Emission Control System Type — Chicago


Catalyst


Yes
Yes
No
No

Secondary
Air
Injection


Yes
No
Yes
No

No of
Vehicles


30
44
17
9
/a\
FTPV ' Emission
(gm/mi)

HC

Mean
1. 139
1.293
1.364
1.563
SD
1.005
0.875
0.475
0.455
CO

Mean
19.995
25.528
14.749
21.408
SD
16.407
20.473
7.024
10.773
NO
X
Mean
2.612
2.229
2.450
2. 158
SD
1.074
0.778
0.964
0.489
  FTP = federal emission certification test procedure

  SD = standard deviation
                Table 44b.  Associated Tests of Significance
Emission
HC
CO
NO
Log1Q (HC)
Log1Q (CO)
Log.- (NO )
lu x
Equality of Means
— (a>
0.686
1.805
1.320
2.404
1.639
0.775
Level of
Significance
0. 563
0. 151
0.272
0.072
0. 185
0. 511
Homogeneity of Variance
Bartlett
Test
13.783
20.971
7.994
8.931
4. 109
3.943
App r oximate
Level of
Significance
0.004
0.0
0.051
0.034
0.261
0.279
(a)
   Degrees of freedom are  3, 96
                                   39

-------
                            Table 45.   Correlation Coefficient Summary by
                               Emission Control System Type — Chicago
Short Test
Federal
Short Cycle


Federal
Three-Mode














Catalyst/Air(a)
Yes/Yes
Yes /No
No /Yes
No/No
Yes/Yes



Yes /No



No /Yes



No /No



Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
30
44
17
9
30
30
30
27
44
44
44
42
17
17
17
5
9
9
9
4
ST/FTP(b) ,.
Correlation Coefficients
HC
0.91
0.86
0.91
0.81
0.30*
0. 14*
0.23*
0.47
0.03*
0.52
0.46
0. 59
0.52
0.68
0.75
0.88
-0. 10*
0. 36*
0.75
0.88*
co
0. 76
0.85
0.92
0.82
0.49*
0. 51
0.64
0. 58
0. 32
0. 56
0.48
0.60
0.48*
0.72
0. 54
0. 38*
0. 36*
-0.07*
0.85
0.82*
NO
X
0.80
0.77
0.94
0.70
0.60
0.69
-0.05*
-0. 11*
0.44
0. 36
0. 10*
0.20*
0. 56
0.69
0. 18*
0.25*
0.04*
0.68
-0.04*
-0. 74*
Log(HC)
0.85
0.82
0.77
0.86
0. 36
0. 34*
0.24*
0.21*
0.20*
0.41
0. 55
0.66
0.43
0.65
0.65
0.84*
-0.03*
0.47*
0.79
0.75*
Log(CO)
0.81
0.86
0.94
0.84
0.70
0.61
0.80
0.87
0. 31
0.47
0. 51
0. 56
0. 32*
0 . 44*
0.53
-0. 13*
0.23*
-0.05*
0.90
0.96
Log(NOx)
0.83
0.86
0.93
0. 74
0.64
0.61
0. 16*
-0.04*
0. 50
0.48
0. 38
0.42
0.70
0.73
0.43*
0. 50*
0.01*
0.60*
0. 14*
-0.68*
(a)
(b)
(c)
Secondary air injection system
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
            Table 56a.  FTP Means and Standard Deviations by
                        Fuel System Type — Chicago
Fuel Injection
Yes
No
No. of
Vehicles
10
90
FTp(a) Emission
(gm/mi)
HC
Mean
1.588
1.249
SD
0.448
0. 862
CO
Mean
15. 506
22. 349
SD-
7.991
17.770
NOX
Mean
2. 378
2. 375
SD
0.600
0.925
(a)
(b)
FTP = federal emission certification test procedure

SD =  standard deviation
               Table 46b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log (HC)
Log1Q (CO)
Log1() (N0x)
Equality of Means
F-Value(a)
1.492
1.440
0.0
4. 340
0.900
0. 163
Level of
Significance
0.225
0.233
. 0.991
0 . 040
0. 345
0.687
Homogeneity of Variance
Bartlett
Test
4.967
6.930
2.433
6.518
1.627
3.785
Approximate
Level of
Significance
0.029
0.010
0. 126
0.012
0.211
0.056
(a)
  Degrees of freedom are 1, 98
                                      41

-------
                                       Table 47.   Correlation Coefficient Summary by
                                                 Fuel System Type — Chicago

Short Test

Federal
Short Cycle

Federal
Three-Mode







Fuel
Injection


Yes
No
Yes



No




Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Los speed
Idle in neutral
Idle in drive

No. of
Vehicles


10
90
10
10
10
4
90
90
90
74
ST/FTP(a) .
Correlation Coefficients

HC

0.79
0.88
-0.07*
0.03*
0.91
0.95*
0. 11*
0.39
0.40
0.51

CO

0.82
0.82
0.01*
0. 31*
0.83
0.90*
0. 32
0. 53
0.54
0. 53

NO
X

0.75
0.81
0.52*
0. 15*
-0.64
-0.99
0.44
0.40
0.05*
0.02*

Log(HC)

0.81
0.80
0.08*
0. 17*
0.86
0.81*
0.30
0.43
0.48
0. 52

Log(CO)

0.81
0.81
0.05*
0.20*
0.76
0.64*
0. 38
0.43
0.59
0.67

Log(NOx)

0.67
0.86
0. 36*
0. 10*
-0.73
-0.96
0.52
0.49
0.29
0.23
ro
          (a)
          (b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
              Table 48a.  FTP Means and Standard Deviations by
                         Transmission Type — Chicago
Transmission
Type
Automatic
Manual
No. of
Vehicles
78
22
FTp(a) Emission
(gm/mi)
HC
Mean
1.280
1.294
SD
0.919
0.430
CO
Mean
23.547
14.992
SD
18.689
6.732
NOX
Mean
2.408
2.258
SD
0.920
0.812
  (a)
  (b)
FTP = federal emission certification test procedure
SD = standard deviation
                 Table 48b.  Associated Tests of Significance
Emission
HC
CO
NO
Log1Q (HC)
Log1Q (CO)
Log1() (N0x)
Equality of Means
F-Value(a)
0.005
4.420
0.481
1. 197
2. 561
0.474
Level of
Significance
0.946
0.038
0.490
0.277
0. 113
0.493
Homogeneity of Variance
Bartlett
Test
13.933
22.660
0.486
8.333
7.601
0.158
Approximate
Level of
Significance
0.0
0.0
0.489
0.004
0.006
0.693
(a)
  Degrees of freedom are 1, 98
                                     43

-------
                            Table 49-  Correlation Coefficient Summary by
                                     Transmission Type — Chicago
Short Test
Federal
Short Cycle
Federal
Three-Mode


Transmission
Type
Automatic
Manual
Automatic

Manual
Test Mode


High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
78
22
78
78
78
78
22
22
22
ST/FTP(a) „.
Correlation Coefficients
HC
0.89
0.82
0.09*
0.35
0.50
0.52
0. 57
0.73
0.31

CO
0.82
0.84
0. 30
0.53
0. 55
0.50
0.61
0.69
0.48

NO
X
0.80
0.89
0.47
0. 38
0.01*
0.01*
0.57
0. 54
0. 18*

Log(HC)
0.83
0.70
0. 31
0.41
0.48
0. 54
0.43
0. 74
0.56

Log(CO)
0.82
0.85
0. 37
0.46
0.61
0.63
0. 39*
0. 34*
.0.54
Log(NOx)
0.84
0.90
0.49
0.42
0. 18*
0.22*
0.75
0.74
0.59
(a)

(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients, are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
           Table 50a.  FTP Means and Standard Deviations by
                        Manufacturer — Chicago
Manu-
facturer
General
Motors
Ford
Chrysler
American
Motors
Others
No. of
Vehicles
33
20
12
6
29
FTp(a) Emission (gm/mi)
HC
Mean
1.319
1.230
1.091
1.053
1.406
SD
0.980
1. 148
0.621
0. 524
0.472
CO
Mean
26.071
19.618
25.013
19.017
17.224
SD
22.743
12.710
19.604
13.519
10. 115
NO
X
Mean
2.249
2.707
2.243
2.497
2.319
SD
0.967
0.958
0.631
0.592
0.902
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
              Table 50b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log.Q (HC)
.Log1() (CO)
Log10(NOx)
Equality of Means
F - Value ^
0.455
1.266
0.973
1.377
0.488
1. 125
Level of
Significance
0.768
0.289
0.426
0.248
0.744
0.349
Homogeneity of Variance
Bartlett
Test
22.706
21.250
4.059
10. 189
9.744
5.762
Approximate Level
of Significance
0.0
0.0
0.416
0.043
0.051
0.233
(a)
  Degrees of freedom are 4, 95
                                  45

-------
                 Table 51.   Correlation Coefficient Summary by
                               Manufacturer — Chicago
Short
Test
Federal
Sho rt •
Cycle-







Federal
Three- .
Mode
























Manu-
facturer
General
Motors

Ford

Chrysler
American
Motors
Others

General
Motors


Ford





Chrysler






American
Motors



Others



Test Mode










High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed

Idle in neutral
Idle in drive

High speed

Low speed

Idle in neutral

Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
33


20

12
6

29

33
33
33
31
20
20

20
19

12

12

12

12
6
6
6
6

29
29
29
10
ST/FTP(a) Correlation Coefficients^'
HC
0.91


0.96

0.74
0.94

0.80
*
0.04
0.58
0.51
0.61
0.38*
0.02*

0.38'"
0.61
;•.
-0.29""

-0.37'"
•f
0.33
,..
0. 30""
0. 50*
0. 54*
0.90
0.92
•''
0. 34
0. 57
0. 40
0.70
CO
0.88


0.75

0.88
0.93

0.86
' *
0. 34
0.60
0.57
0.71
0.52
0.80

0.70
0.84
&
0.35

-0. 10'

0. 39"
jj^
0.47
0.64*
o:63*
0.89
0.84
sis
0. 19
0.77
0. 56
0. 33"
NO
X
0.95


0.84
*
0.43
0.96

0. 89

0.51
0.50
A
0. 13
0.37
0.64
0.75
A
-0.38
-0.15*
*
0.23
*
0. 54

-0.31""
*
-0.43
0.96
0.06*
-0.38*
0.26*

0.42
0.61
0. 12*
0.08*
Log
(HC)
0.89


0.87

0.66
0.97

0.78
*
0.30
0.52
0.65
0.73
0.35*
-0. 10*
*
0.43
0. 55
...
0. 16""

0. 10

-0.04'"
,t»
-0.04""
0. 54*
0.51*
0.90
0.92
.,
0.32'"
0.58
0.47
0. 43!"
Log
(CO)
0.88


0.82

0.83
0.97

0. 85 .

0.48
0.62
0.69
0.77
0.45
0.60

0.62
0.77
-i
0. 36"
*
0.04

0. 34'"
0,
0.33""
0.68*
0.71*
0. 88
0.74*
..-
0. 19'"
0.43
0.59
0.21*
Log
(NO )
0.95


0.86
...
0.41'"
0,96

0.91

0. 53
0. 57
0.43
0. 53
0.69
0.65

-0.42""
*
-0. 17
A
0.41

0.61

-0.26'"
0-
-0. 39'"
0.95
-0. 16*
-0. 37*
0.23*

0. 51
0.65
0. 31*
0.09*
(a)

(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent program confidence
level except where indicated by an asterisk
                                             46

-------
          Table 52a.  FTP Means and Standard Deviations by
                    Engine Displacement — Houston

CID(a)


1 50 and,
Less
151 to
259
260 and
Greater

No. of
Vehicles


33

16

51

FTp(b) Emission (gm/mi)
HC

Mean
1.518

1.227

1.640

(c)
SD(C)
0.903

0. 842

1.659

CO

Mean
18.930

22.014

34.959


SD
13.750

26.480

41.052

NO
X

Mean
2.044

2.469

2.700


SD
0.941

0.754

1. 129

(a)
(b)
(c)
CID = cubic inch displacement
FTP = federal emission certification test procedure
SD = standard deviation
              Table 52b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log (HC)
Log1Q (CO)
Log,- (NO )
1U X
Equality of Means
F- Value (a)
0.580
2.761
4. 154
0.549
0.264
5.380
Level of
Significance
0.562
0.068
0.019
0.579
0.768
0.006
Homogeneity of Variance
Bartlett
Test
17.919
36.282
3.762
1.577
13.801
5.741
Approximate Level
of Significance
0.0
0.0
0. 158
0.461
0.001
0.060
(a)
  Degrees of freedom are 2,  97
                                  47

-------
                                    Table 53.  Correlation Coefficient Summary by
                                            Engine Displacement — Houston
Short
Test
Federal
Short
Cycle



Federal
Three -
Mode









CID(a)
Group
150 or
less
151 to
259
260 or
more
150 or
less


151 to
259


260 or
more


Test Mode






High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
33

16

51

33
33
33
11
16
16
16
13
51
51
51
51
ST/FTP*b) Correlation Coefficients (c)
HC
0.90

0.94

0.94

•Jf
0.29"
0, 53
0.61
0.92
-0. 14*
0.86
0.82
0.90
0.75
0.83
0.40
0.57
CO
0.82

0.96

0.98

0.33*
0.38
0.79
0.45*
-0.31*
0.95
0.68
0.75
0.65
0. 59
0.84
0.76
NO
X
0.78

0. 56

0.79

0.61
0.66
0.42
0.85
0.32*
0.56
*
0.34
0.57
0. 54
0.31
0.35
0.31
Log
(HC)
0.86

0.83

0.90

0.46
0.56
0.63
0. 76
;|c
-0. 11
0.66
0. 58
0.79
0.46
0. 58
0.70
0.77
Log
(CO)
0.77

0.78

0.92

0.03*
0.37
0. 40
0.48*
-0.61
0.11*
0.33
0.66
0.48
0.48
0.86
0. 85
Log
(N0x)
0.88

0.65

0.73

0.76
0.79
0.45
0.85
0. 19*
0.65
0.47*
0.64
0.59
0.39
0.34
0.34
oo
          (a),
          (b)
          (c)
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence level
except where indicated by an asterisk

-------
           Table 54a.  FTP Means and Standard Deviations by
                     Inertia Test Weight —Houston
Inertia
Test
Weight
Group


2500 and
less
2501 to
3500
3501 to
4500
Greater
than 4500

No. of
Vehicles


19

38

33

10

FTp(a) Emission (grn/mi)

HC

Mean
1.437

1. 531

1.385

2.219

(b)
SD1D'
0.777

1. 330

1. 523

1. 500

CO

Mean
17. 143

25.795

26.318

58. 541


SD
12.383

32.411

29.035

53.905

NO
X

Mean
1.818

2.477

2.822

2.286


SD
0.785

1. 149

1.010

0.680

(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
              Table 54b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1() (HC)
Log1() (CO)
Log1() (N0x)
Equality of Means
F-Value(a)
1.051
3.995
4. 120
1.582
1.849
5.490
Level of
Significance
0.374
0.010
0.009
0.199
0. 143
0.002
Homogeneity of Variance
Bartlett
Test
9.076
26.544
5.481
2.303
8.080
7. 163
Approximate Level
of Significance
0.031
0.0
0. 148
0.522
0.048
0.072
(a)
  Degrees of freedom are 3, 96
                                   49

-------
                 Table 55.   Correlation Coefficient Summary by
                       Inertia  Test Weight Group —Houston
Short
Test
Federal
Short
Cycle





Federal
rT>L.
Mode















Inertia
Test
Weight
Group
0 to 2500

2501 to
3500
3501 to
4500
4501 or
more
0 to 2500




2501 to
3500


3501 to
4500



4501 or
more


Test Mode








High speed

Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed

Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
19

38

33

10

19

19
19
3
38
38
38
29
33

33
33
33
10
10
10
10
ST/FTP(a) Correlation Coefficients (b)
HC
0.94

0. 88

0.94

0.93

0.47

0.81
0.82
1.00
0.66
0.75
0. 52
0.70
0.87

0.85
0.25*
0.43
0.49*
0.66
0.62*
0.75
CO
0.87

0.95

0.98

0.98

0.58

0.57
0.67
-0.90*
0.70
0.64
0.68
0.62
0.42

0.44
0.93
0.91
0.69
0.55*
0.85
0.85
NO
X
0.89

0.73

0. 81

0. 53*

0.65
j.
0.44'"
0.54
0.99*
0.57
0.68
0. 31*
0. 33*
0.45
.1.
0. 12'"
0.41
0. 37
0.21*
0. 14*
0.33*
0.27*
Log
(HC)
0.82

0.89

0. 88

0.90

0.57

0.72
0.76
1.00
0.32*
0.56
0.60
0.77
0.52

0. 57
0.65
0. 76
0.42""
0.61*
0.79
0.82
Log
(CO)
0.70

0.85

0.90

0.96

0.45*
.'.
0.33""
0.26*
-0.91*
0.04*
0.45
0.63
0.82
0.35

0. 35
0.82
0. 81
0.49*
0.41*
0.90
0.91
Log
(N0x)
0.92

0.72

0.80

0.57*

0. 84

0.79
0. 52
1.00
0.47
0.69
0. 50
0. 56
0.54
-I*
0.31""
0.39
0. 35
0.45*
0.32*
0. 19*
0. 12*
(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence le\
except where indicated by an asterisk
                                                                           el
                                            50

-------
           Table 56a.   FTP Means and Standard Deviations by
                Emission Control System Type — Houston
Catalyst
Yes
Yes
No
No
Secondary
Air
Injection
Yes
No
Yes
No
No. of
Vehicles
^ 27
41
22
10
FTp(a) Emission (gm/mi)
HC
Mean
1. 185
1.800
1.383
1.713
SD
-------
                   Table 57.   Correlation  Coefficient Summary by
                      Emission Control System Type —Houston
Short Test
Federal
Short Cycle


Federal
Three-Mode














Catalyst/
Air(a)
Yes/Yes
Yes/No
-No/Yes
No/No
Yes/Yes


Yes/No



No/Yes



No /No



Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
27
41
22
10
27
27
27
24
41
41
41
38
22
22
22
9
10
10
10
4
ST/FTP(b) Correlation Coefficients^
HC
0.97
0. 91
0. 94
0.98
0. 94
0. 94
0. 14*
-0.40*
0. 56
0.76
0. 59
0.68
0.33*
0.32*
0. 40*
0.92
0. 55*
0.86
0.80
0. 70*
CO
0. 90
0. 98
0. 87
0.91
0. 41
0. 57
0.79
0.82
0. 64
0.63
0.75
0.69
0. 56
0. 64
0.64
0.76
0.93
-0. 01*
0.90
0. 56*
NO
X
0. 84
0.70
0.75
0.94
0. 78
0. 64
0. 09*
0. 17*
0. 27*
0.36
0. 52
0. 54
0. 61
0.74
0. 62
0. 84
-0.09*
0. 11*
0. 62*
0. 43*
Log
(HC)
0.80
0.93
0.83
0.98
0.48
0. 55
0.35*
0.48
0.46
0. 64
0.79
0.83
0.52
0.52
0. 50
0.78
0.49*
0.74
0. 69
0. 63*
Log
(CO)
0.87 .
0.90
0.90
0.72
0. 52
0.56
0.62
0.76
0.49
0. 55
0.76
0..81
0. 50
0.43
0. 51
0.78
0. 68
0. 25*
0.75
0.41*
Log
(NOX)
0. 89
0..66
0. 84
0.91
0. 81
0. 67
0. 10*:
0.48
0. 35
0. 55
0.37
0/44
0.66
0.80
0.75
0.76
-0. 19*
-0.02*
0. 53*
0. 27*.
(a)
(b)
(c)
Secondary air injection system
ST/FTP = short test/federal emission certification test procedure
Correlation coefficients are statistically significant at the 95 percent confidence level
except where indicated by an asterisk

-------
           Table 58a.  FTP Means and Standard Deviations by
                      Fuel System Type — Houston
Fuel
System Type
Fuel injection
Carburetion
No. of
Vehicles
10...
90
FTp(a) Emissions (gm/mi)
HC
Mean SD(b)
1.652 0.937
1.520 1.375
CO
Mean
20. 900
28.343
SD
14. 022
34. 236
NOX
Mean SD
2.032 0.489
2.493 1.09
(a)
(b)
FTP = federal emissions certification test procedure

SD = standard deviation
              Table 58b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1Q (HC)
Log.Q (CO)
L°g10 (N°x}
Equality of Means
F-Value(a)
0. 087
0.461
1.744
0.840
0. 026
0. 727
Level of
Significance
0.769
0.499
0. 190
0. 362
0. 872
0.369
Homogeneity
of Variance
Bartlett
Test
1.961
8. 275
6.916
0. 800
2. 568
4.641
Approximate
Level of
Significance
0. 169
0. 005
0. 010
0.380
0. 116
0. 034
 (a)
   Degrees of freedom are 1,  98
                                  53

-------
                        Table 59.  Correlation Coefficient Summary by
                                  Fuel System Type — Houston
Short
Test
Federal
Short
Cycle
Federal
Three -
Mode










Fuel
System
Type
Fuel
injection
Carburetion
Fuel
injection




Carburetion





Test Mode


High speed
Low speed
Idle in
neutral
Idle in
drive
High speed
Low speed
Idle in
neutral
Idle in
drive
No. of
Vehicles
10
90
10
10
10

3

90
90
90

72

ST/FTP(a) Correlation Coefficients (t>)
HC
0.97
0.91
0.51*
0.82
0.86

0.99*

0.67
0.77
0.45

0.60

CO
0.76
0.97
0.77
0.0*
0.77

0.47*

0.45
0.61
0.81

0.76

NO
X
0.53*
0.78
0.48*
-0.08*
0. 14*

0.57*

0.46
0.44
0.35

0. 36

Log
(HC)
0.95
0.88
0.64
0.72
0.83

*
1.00

0.39
0.57
0.65

0.76

Log
(CO)
0.92
0. 87
0.77
0.32*
0. 82

5JC
0.58

0. 19*
0.36
0.68

0. 80

Log
(N0x)
0.40*
0.82
0. 35*
0.02*
0.08*

0.48*

0. 54
0.61
0. 34

0. 46

(a)
(b)
ST/FTP = short test/federal emission certification test procedure

Correlation Coefficients  are statistically significant at the 95 percent confidence level
except where indicated by an asterisk

-------
           Table 60a.  FTP Means and Standard Deviations by
                     Transmission Type —Houston
Trans-
mission
Type
Automatic
Manual
No. of
Vehicles
75
25
FTp(a) Emission (gm/mi)
HC
Mean SD(b)
1.514 1.477
1.591 0.786
CO
Mean
30.718
18.239
SD
36.511
14.461
NO
X
Mean SD
2.607 1.094
1.965 0.735
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
               Table 60b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1Q (HC)
Log1Q (CO)
Log,n (NO )
1 v X
Equality of Means
F- Value (a)
0.061
2.760
7.457
1.872
0.547
6.764
Level of
Significance
0.805
0., 100
0.008
0. 174
0.461
0.011
Homogeneity of Variance
Bartlett
Test
11.357
21.790
4.934
3.070
7.249
0. 143
Approximate Level
of Significance
0.001
0.0
0.028
0.082
0.008
0.707
(a)
  Degrees of freedom are 1,  98
                                   55

-------
                                 Table 61.  Correlation Coefficient Summary by
                                          Transmission Type — Houston
Short
Test
Federal
Short
Cycle
Federal
Three-
Mode










Trans-
mission
Type
Automatic

Manual
Automatic






Manual





Test Mode



High speed

Low speed
Idle in
neutral
Idle in
drive
High speed
Low speed
Idle in
neutral
Idle in
drive
No. of
Vehicles
75

25
75

75
75

75

25
25
25

__

ST/FTP*a) Correlation Coefficients (b)
HC
0.91

0.93
0.69

0.79
0.44

0.60

0.40
0.72
0.73

__

CO
0.97

0.91
0.47

0.60
0. 84

0.76

*
0.24
0.31*
0.80

._

NO
X
0. 78

0.72
0.55

0.45
0.34

0.36

0.49
0.46
0.48

__

Log
(HC)
0.89

0.86
0.39

0.55
0.67

0.76

0. 51
0.68
0.65

__

Log
(CO)
0.89

0.72
0.31

0.40
0.75

0.79

*(*
-0. 17
0. 34*
0. 55

__

Log
(N0x)
0.80

0. 82
0.60

0.62
0. 31

0.46

0. 52
0.63
0. 58

__

(Jl
          (a)

          (b)
ST/FTP = short test/federal emission certification procedure

Correlation coefficients are statistically significant at the 95 percent confidence level
except where  indicated by an asterisk

-------
           Table 62a.  FTP Means and Standard Deviations by
                        Manufacturer — Houston
Manu-
facturer
General
Motors
Ford
Chrysler
American
Motors
Others
No. of
Vehicles
33
21
12
6
28
FTp(a) Emission (gm/mi)
HC
Mean
1.581
1.227
2.325
1. 142
1.453
SD
1.616
1.037
1.790
0.737
0.916
CO
Mean
26.897
23. 531
70. 437
10.377
16.806
SD
29.390
34.954
46.412
5.661
10.936
NO
X
Mean
2.530
3.004
2. 136
2.852
1.976
SD
0.980
1.278
0.583
0.613
0.964
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
              Table 62b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log,Q (HC)
Log1Q (CO)
Log,-. (NO )
1 U x
Equality of Means
F- Value (a)
1.529
8.228
3.809
1.767
5.984
4.330
Level of
Significance
0.200
0.0
0.007
0. 142
0.0
0.003
Homogeneity of Variance
Bartlett
Te.st
15.656
49.453
9.348
3. 128
13.365
10.286
Approximate Level
of Significance
0.004
0.0
0.060
0.553
0.012
0.041

  Degrees of freedom are 4, 95
                                   57

-------
                                      Table 63.  Correlation Coefficient Summary by
                                                  Manufacturer — Houston
Short
Test
Federal
Short Cycle



Federal
Three -Mode









Manufacturer
General Motors
Ford
Chrysler
American Motors
Others
General Motors

Ford

Chrysler

American Motors

Others

Test
Mode





High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
33
21
12
6
28
33
33
33
30
21
21
21
19
12
12
12
12
6
6
6
5
28
28
28
9
ST/FTP(a) (b)
Correlation Coefficients
HC
0.94
0.93
0.92
0.98
0.91
0.85
0.88
0.42
0.48
0.02*
0.72
0.39*
0.68
0.66
0.79
0.62
0.69
-0.68*
0.77*
0.82
0.86*
0.33*
0.54
0.59
0.92
CO
0.97
'0.98
0.98
0.90
0.79
0.52
0.81
0.78
0.73
0.70
0.88
0.85
0.85
0.73
0.75
0.78
0.78
0.82
0.79*
0.77*
0.69*
0.58
0.32*
0.60
0.09*
NOX
0.90
0.66
0.57*
0.97
0.90
0.20*
0.51
0.50
0.51
0.66
0.69
0. 12*
-0.08*
0.34*
0.53*
0.58
0.59
0.55*
0.06*
0. 18*
0.41*
0.71
0.73
0.48
0.92
Log (HC)
0.92
0.88
0.95
0.97
0.84
0.52
0.69
0.75
0.81
0.02*
0.43*
0.50
0. 56
0.64
0.70
0.79
0.75
-0.70*
0.61*
0.87
0.89
0.50
0.58
0.58
0.77
Log (CO)
0.90
0.92
0.95
0.96
0.73
0.30*
0.55
0.79
0.84
0.20
0.34
0.62
0.67
0.39*
0.61
0.73
0.86
0.94
0.94
0.93
0.91
0.48
0.37*
0.21*
0. 16*
Log (NOX)
0.86
0.62
0.59
0.96
0.91
0.31*
0.58
0.52
0.56
0.68
0.57
0. 19*
0.13*
0.34*
0.75
0.58
0.55*
0.27*
0.27*
0.23*
0.48*
0.81
0.80
0.52
0.90
(J\
oo
          (a)

          (b)
ST/FTP = short test/federal emission certification test procedure

Correlation Coefficients are  statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
             Table 64a.  FTP Means and Standard Deviations by
                      Engine Displacement — Phoenix
CID(a) Group
1 50 and Less
151 to 259
260 and
Greater
No. of
Vehicles
30
19
51
FTp( ) Emission
(gm/mi)
HC
Mean
1.095
1.509
1.319
SD
0.641
1.482
0.802
CO
Mean
14.112
23.853
27.782
SD
10.407
18.668
25.410
NOX
Mean
1.995
2.527
2.508
SD
1.005
1.037
0.889
(a)
(b)
(c)
CID = cubic inch displacement
FTP = federal emission test certification procedure
SD = standard deviation
                Table 64b.  Associated Tests of Significance
Emission
HG
CO
NO
X
Log1() (HC)
Log1Q (CO)
Log1() (N0x)
Equality of Means
F- Value (a)
1.213
4. 128
3.106
1.098
3.552
5.030
Level of
Significance
0.302
0.019
0.049
0.338
0.033
0.008
Homogeneity of Variance
Bartlett
Test
19.320
23.415
0.897
0.522
3.645
5.804
Approximate
Level of
Significance
0.0
0.0
0.643
0.773
0.167
0.058
  (a)
    Degrees of freedom are 2, 97.
                                    59

-------
                            Table 65.   Correlation Coefficient Summary by
                                   Engine  Displacement — Phoenix
Short
Test
Federal
Short
Cycle
Federal
Three -Mode





CID(a) Group
150 or less
151 to 259
260 or more
150 or less

151 to 259

260 or more

Test
Mode


High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
30
19
51
30
30
30
11
19
19
19
14
51
51
51
51
ST/FTp(b)
Correlation Coefficients
HC
0.42
0.33*
0.70
0.23*
0.29*
0.34*
0.65
0.36*
0.27*
0.13*
0.80
0.16*
0.28
0.71
0.76
CO
0.36
0.93
0.82
0.52
0.18*
0.74
0.46*
0.60
0.62
0.76
0.86
0.31
0.43
0.87
0.87
NOX 1
0.80
0.95
0.83
0.85
0.82
-0.18*
-0.21*
0.81
0.87
0.14*
-0.14*
0.75
0.49
0.22*
-0.02*
Log (HC)
0.83
0.67
0.70
0.29*
0.33*
0.38
0.56*
0.44*
0.43*
0.41*
0.76
0. 15*
0.28
0.80
0.83
Log (CO)
0.75
0.88
0.84
0.34*
0,40
0.67 .
0.83
0.21*
0,30*
0.82
0.91
0.27*
0.46
0.77 •
0.80
Log (NOX)
0.72
0.95
0.80
0.87
0.87
-0.06*
0.11*
0.77
0.80
0.22*
-0.04*
0.68
0.51
0.21*
-0.01*
(a)
(b)
(c)
CID = cubic inch displacement
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
             Table 66a.  FTP Means and Standard Deviations by
                      Inertia Test Weight — Phoenix

Inertia Test
Weight Group
(Ib)
2500 or less
2501 to 3500
3501 to 4500
Greater than
4500

No. of
Vehicles
16
38
35
11

FTp(a) Emission
(gm/mi)
HC
Mean
1.344
1.262
1.081
1.957

SD
1.571
0.723
0.631
0.964

CO
Mean
12.578
19.983
23.642
45.946

SD
11.013
18.245
20.792
29.192

NOX
Mean
1.701
2.544
2.392
2.557

SD
0.704
1.094
0.847
0.947

  'FTP = federal emission certification test procedure
^  'SD = standard deviation
                Table 66b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log'(HC)
Log1Q (CO)
Log10(NOx)
Equality of M^eans
F- Value (a)
2.629
6.759
3.274
3.039
5.976
3.779
.Level of
Significance
0.055
0.0
0.024
0.033
0.001
0.013
Homogeneity of Variance
Bartlett
Test
24.023
11.845
4.677
5.758
0.679
4.388
Approximate
Level of
Significance
0.0
0.009
0.206
0.131
0.882
0.232
  (a)
    Degrees of freedom are 3, 96
                                    61

-------
                           Table 67.  Correlation Coefficient Summary by
                                 Inertia Test Weight Group — Phoenix
Short
Test
Federal
Short
Cycle


Federal
Three -Mode














Inertia Test
Weight Group
(Ib)
0 to 2500
2501 to 3500

3501 to 4500
4501 or more
0 to 2500



2501 to 3500



3501 to 4500



4501 or more



Test
Mode





High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
16
38

35
11
16
16
16
3
38
38
38
27
35
35
35
35
11
11
11
11
ST/FTP(a) (b) ;
Correlation Coefficients '
HC
0.21*
0.52

0.86
0.28*
0. 15*
0.13*
0. 15*
0.99*
0.56
0.64
0.49
0.71
0.29*
0.29*
0.82
0.82
-0.03*
0.08*
0.38*
0.52*
CO
0.57
0.71

0.83
0.68
0.68
0.17*
0.80
0.55*
0.56
0.70
0.80
0.82
0.41
0.43
0.82
0.84
0.01*
0.09*
0.76
0.81
NOX
0.44*
0.91

0.86
0.80
0.85
0.89
0.02*
0.80*
0.80
0.79
0.03*
-0.25*
0.72
0.53
0.35
0.20*
0.76
0.36*
0.04*
-0.24*
Log (HC)
0.72
~0.82

0.84
0.22*
0.43*
0.41*
0.50
i.oo(c)
0.37
0.47
0.50
0.66 .
0.20*
0.25*
0.79
0.83
-0.03*
-0.16*
0 . 51 *
0.68
Log (CO)
0.74
0.81

0.87
0.70
0.66
0.49*
6.72:
0.67*
0.33.
0.49
0.75
0.84;
0.32*
0.43
0.86
0.88
-0.0*
0.25*
0.42*
0.62
Log (NOX)
0.48*
0.94

0.83
0.74
0.84
0.90
0*
0.83*
0.82
0.83
0.14*
-0.08*
0.69
0.54
0.33*
0.17*
0.66
.0.47*
0.18*
-0. 10*
(a)
(b)
(c)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except.where
indicated by an asterisk
Only two data points used

-------
            Table  68a.  FTP Means and Standard Deviations by
                    Emission Control System — Phoenix
Catalyst
Yes
Yes
No
No
Secondary
Air Injection
Yes
No
Yes
No
No. of
Vehicles
32
42
20
6
FTp(a) Emission
(gm/mi)
HC
Mean
1.043
1.417
1.108
2.300
SD
0.802
0.816
0.425
2.263
CO
Mean
17.739
31.434
16.200
13.612
SD
22.571
23.288
8.267
10.871
NO
Mean
2.545
2.282
2.189
2.448
SD
0.996
0.919
1.049
1.054
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
                Table 68b.  Associated Tests of Significance
Emission
HC
CO
N0x
Log1Q (HC)
Log1Q (CO)
Log1() (N0x)
Equality of Means
F- Value (a)
3.930
4.271
0.693
4.040
4.647
0.757
Level of
Significance
0.011
0.007
0.559
0.009
0.005
0.521
Homogeneity of Variance
Bartlett
Test
32.607
23.411
0.577
6.122
5.881
1.530
Approximate
Level of
Significance
0.0
0.0
0.906
0. 115
0.128
0.687
  (a)
     Degrees of freedom are 3, 96
                                    63

-------
                            Table 69.  Correlation  Coefficient Summary by
                               Emission  Control System Type — Phoenix
Short
Test
Federal
Short
Cycle

Federal
Three -Mode







Catalyst/ Air(a)
Yes/Yes
Yes/No
No/Yes
No /No
Yes/Yes

Yes /No

No/Yes

No/No

Test
Mode



High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
32
42
20
6
32
32
32
25
42
42
42
37
20
20
20
10
6
6
6
4
ST/FTP(b) . .
( C 1
Correlation Coefficients
HC
0.92
0.62
0.51
-0.70*
0.54
0.61
0.73
0.87
0.20*
0.25*
0.63
0.66
0.36*
0.34*
0. 13*
0.55*
-0.29*
-0. 56*
-0.92
0.54*
CO
0.93
0.75
0.22*
0.83
0.47
0.54
0.96
0.96
0.31
0.44
0.73
0.73
0.38*
0.38*
0.45
0.57*
-0.21*
0.07*
0.58*
0.38*
NOX
0.86
0.82
0.92
0.94
0.79
0.71
0. 17*
-0.21*
0.68
0.62
-0. 13*
-0.17*
0.79
0.86
0.05*
0.41*
0.70*
0.86
0.48*
0.41*
Log (HC)
0.90
0.66
0.80
-0.71*
0.41
0.51
0.62
0.79
0.22*
0.37
0.80
0.78
0.29*
0.16*
0.03*
0. 35*
-0.16*
-0.58*
-0.93
0 . 57 *
Log (CO)
0.81
0.84
0.62
0.75*
0.71
0.81
0.87
0.87
0. 12*
0.34
0.71
0.74
0.48
0.39*
0.35*
0.86
-0. 17*
0.20*
0 . 62 *
0.36*
Log (NOX)
0.81
0.75
0.95
0.97
0.84
0.65
0.18*
-0.22*
0.66
0.62
-0. 13*
-0.20*
0.77
0.80
0.23*
0.57*
0.70*
0.86
0.63*
0.57*
(a)
(b)
(c)
Secondary air injection system
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
            Table 70a.  FTP Means and Standard Deviations by
                        Fuel System Type — Phoenix
Fuel
Injection
Yes
No
No. of
Vehicles
10
90
FTp(a) Emission
(gin /mi)
HC
Mean
1.527
1.262
SD(b)
1.934
0.760
CO
Mean
15.068
23.809
SD
13.934
21.932
NO
X
Mean
1.700
2.431
SD
0.703
0.975
   FTP = federal emission certification test procedure
*  ' SD = standard deviation
                Table 70b.  Associated Tests of Significance
Emission
HC
CO
NO
Log1()(HC)
Log1Q(CO)
Log10(NOx)
Equality of Means
F-Value(a)
0.731
1. 512
5. 304
0.027
2.858
5.903
Level of
Significance
0.395
0.222
0.023
0.869
0.094
0.017
Homogeneity of Variance
Bartlett
Test
23. 128
2.646
1.470
2.098
0. 165
0. 150
Approximate
Level of
Significance
0.0
0. 110
0.234
0. 155
0.690
0.704
(a)
  Degrees of freedom are 1,  98
                                    65

-------
                             Table 71.  Correlation  Coefficient Summary by
                                       Fuel System Type  - Phoenix

Short Test

Federal
Short Cycle

Federal
Three-Mode







Fuel
System Type

Fuel
Injection
Carburetion
Fuel
Injection


Carburetion




Test Mode




High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

No. of
Vehicles

10

90
10
10
10
4
90
90
90
72
ST/FTP(a*
Correlation Coefficients

HC
0.24*

0. 58
0.05*
0.01*
-0.05*
0. 73
0. 32
0. 36
0.62
0.71

CO
0.88

0.76
0.60*
0. 13*
0.68
-0. 18*
0. 34
0.42
0.81
0.82

NOX
0. 50*

0.86
0.83
0.75
-0. 11
0. 09*
0.72
0.64
0. 10*
-0.07*

Log (HC)
0.67

0. 75
0.28*
0. 19*
0.22*
0. 36*
0.27
0. 36
0.67
0.77

Log (CO)
0.90

0.80
0. 55*
0. 53*
0.70
0.64*
0.21
0. 38
0.75
0. 79

Log (N0x)
0. 50*

0.83
0.79
0.71
-0. 15*
0.26*
0.71
0.64
0. 14*
0.01*
(a)

(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where
indicated by an asterisk

-------
            Table 72a.  FTP Means and Standard Deviations by
                       Transmission Type — Phoenix



Transmission
Type

Automatic
Manual



No. of
Vehicles

76
24
(a)
FTPV ' Emission
(grri/mi)
HC
Mean
1.246
1.423
SD(b)
0.725
1.409
CO
Mean
24.771
17. 118
SD
22. 564
16. 143
NOX
Mean
2. 529
1.813
SD
0.931
0.921
(a)
(b)
FTP = federal emission certification test procedure
SD = standard deviation
                Table 72b.  Associated Tests of Significance
Emission
HC
CO
NO
X
Log1()(HC)
Log1()(CO)
Log1()(NOx)
Equality of Means
F- Value (a)
0. 661
2. 370
10.848
0.001
2. Ill
15. 789
Level of
Significance
0.418
0. 127
0.001
0.978
0. 149
0.0
Homogeneity of Variance
Bartlett
Test
18. 597
3.480
0.004
2.91
0.680
2. 108
Approximate
Level of
Significance
0.0
0.064
0.948
0.091
0.413
0. 150
(a)
  Degrees of freedom are 1,  98
                                    67

-------
                                      Table 73.   Correlation Coefficient Summary by
                                                Transmission Type  -  Phoenix

Short Test

Federal
Short Cycle










Transmission
Type

Automatic

Manual
Automatic



Manual




Test Mode




High speed
Low speed
Idle Vi neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive

No. of
Vehicles

76

24
76
76
76
76
24
24
24
--
ST/FTP(a)
Correlation Coefficients

HC
0.72

0. 19*
0.24
0.29
0.65
0.71
0.29*
0.29*
0. 18*
	

CO
0.82

0.53
0.26
0.38
0.81
0.81
0.73
0.73
0.75
----

NO
X
0.83

0.89
0. 74
0.64
0. 13*
0.05*
0.89
0.77
-0. 12*
-- —

Log (HC)
0.76

0.67
0.25
0. 35
0.71
0. 75
0. 32*
0. 32*
0. 38*
	

Log (CO)
0.82

0.74
0.20*
0. 36
0.75
0.78
0.41
0. 53
0.68
	

Log (N0x)
0.80

0.79
0.69
0.62
0. 16*
0.03*
0.88
0.85
-0.05*
	
00
             ST/FTP = short test/federal emission certification test procedure

             Correlation coefficients are statistically significant at the 95 percent confidence level except where
             indicated by an asterisk

-------
               Table 74a.  FTP Means and Standard Deviations by
                            Manufacturer — Phoenix
Manufacturer
General Motors
Ford
Chrysler
American
Motors
Others
No. of
Vehicles
33
20
12
6
29
FTp(a) Emission
(gm/mi)
HC
Mean
1.233
1.280
1.761
0.962
1.228
.SD
0. 790
0.853
0.732
0. 563
1.208
CO
Mean
25.527
24.451
41.495
9.833
13.969
SD
22.299
26.279
22.679
5. 113
9.749
NO
X
Mean
2.423
2. 502
2.678
2.945
1.929
SD
0.929
0.820
1.084
0.886
0.990
(a)
(b)
FTP = federal emission certification test procedure
SD  = standard deviation
                  Table 74b.  Associated Tests of Significance
Emission
HC
CO
NO
Log1()(HC)
Log1()(CO)
Log1()(NOx)
Equality of Means
F- Value (a)
1.019
4.918
2. 579
2.092
4.778
3.898
Level of
Significance
0.402
0.001
0.042
0.088
0.002
0.006
Homogeneity of Variance
Bartlett
Test
9.349
32.471
1. 340
6. 138
3.495
7.844
Approximate
Level of
Significance
0.060
0.0
0.862
0.204
C.496
0.108
  (a)
     Degrees of freedom are 4, 95
                                       69

-------
                                Table  75.  Correlation Coefficient Summary by
                                             Manufacturer - Phoenix
Short Test
Federal
Short Cycle





Federal
Three-Mode



















Manufacturer
General
Motors
Ford
Chrysler
American
Motors Company
Others
General
Motors



Ford



Chrysler



AMC



Others



Test Mode








High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
High speed
Low speed
Idle in neutral
Idle in drive
No. of
Vehicles
33

20
12
6

29

33
33
33
31
20
20
20
17
12
12
12
12
6
6
6
5
29
29
29
11
ST/FTP(a)
Correlation Coefficients* '
HC
0.80

0.94
-0.01*
0.95

0. 17*

0.25*
0. 32*
0.75
0.83
0.60
0.64
0.78
0.85
-0.05*
-0.09*
0. 19*
0. 13*
0.69*
0.89
0.96
0.99
0. 13*
0. 13*
0.05*
0.68
CO
0.82

0.97
0.72
0.89

0. 34*

0.40
0. 59
0.77
0.77
0.41*
0.49
0.96
0.96
0. 30*
0.22*
0.66
0.72
0.45*
0.87
0.96
1.00
0.63
0. 34*
0.62
0. 32*
NO
X
0.93

0.95
0.94
0.91

0.78

0.60
0.67
-0.02*
-0.22*
0.87
0.73
-0.26*
-0.47*
0.91
0.59
0. 36*
0. 34*
0.82
0.77*
0.28*
-0. 38*
0.82
0.86
0.04*
0. 59*
Log (HC)
0.85

0.88
0.01*
0.98

0. 73

0.23*
0.42
0.84
0.84
0.44*
0. 52
0.66
0.74
0.03*
0.02*
0.38*
0. 30*
0.59*
0.95
0.96
1.00
0.26*
0.26*
0.23*
0 . 54*
Log (CO)
0.87

0.85
0.86
0.88

0.73

0.20*
0. 56
0.80
0.83
0. 58
0.67
0.79
0.80
0. 31*
0.27*
0.62
0.66
0.73*
0.90
0.98
0.97
0.61
0.55
0.61
0.79
Log (N0x)
0.90

0.91
0.90
0.95

0. 72

0.61
0.66
-0.04*
-0.28*
0.82
0.62
-0.27*
-0. 52
0.88
0.60
0.44*
0. 37*
0.88
0.80*
0.46*
-0. 33*
0.83
0.84
0.09*
0.68
(a)

(b)
ST/FTP = short test/federal emission certification test procedure

Correlation coefficients are statistically significant at the 95 percent confidence level except where indicated
by an asterisk

-------
       Table 76.  Comparison of ST HC Results by City —
 Predicted Population,  Bounded Errors of Commission Method-
                       (E  Set at 5 Percent)
Short Test
Federal
Short Cycle



Federal
Three-Mode















City

All
Chicago
Houston
Phoenix

All



Chicago



Houston



Phoenix



Test Mode






^igh speed
Low speed
die in drive
die in neutral
•Jigh speed
LiOw speed
Idle in drive
Idle in neutral
High speed
Low speed
[die in drive
Idle in neutral
riigh speed
Low speed
Idle in drive
[die in neutral
No. of
Vehicles

300
100
100
100

300



100



100



100



Parameter
(percent)
Ec

5.0
4.2
4.0
6. 3

5.0
5.0
5.0
5.0
8.4
6.9
8. 1
7. 5
1. 3
1. 3
2.2
3.2
5.8
7. 5
4. 3
3.6
Eo

13.0
12.7
11.4
13.6

26.2
24. 1
19. 8
21.0
20. 6
19.5
16.9
18.4
33.7
31.0
23.6
24. 1
24.0
21.8
18.8
20. 3
FF

18.7
15.8
25.5
15.3

5.4
7. 5
11.9
10.7
7. 5
8. 5
11. 5
10.0
3. 3
6.0
13. 3
12.8
4.9
7. 1
10. 1
8.5
STE

0. 589
0. 554
0.691
0. 530

0. 172
0.237
0. 375
0. 338
0.266
0. 303
0.405
0. 352
0.088
0. 161
0. 360
0. 348
0. 171
0.246
0. 349
0.296
(a)
  Short Test Effectiveness =
                            % FF for short test
                      % FTP failures in same population
                                 71

-------
       Table 77.  Comparison of ST CO Results by City—
 Predicted Population,  Bounded Errors of Commission Method
                       (E  Set at 5  Percent)
Short Test
"ederal
Short Cycle



Federal
Three-Mode















City

All
Chicago
Houston
Phoenix

All



Chicago



Houston



Phoenix



Test Mode






High speed
Lx>w speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Lo\v speed
Idle in drive
Idle in neutral
No. of
Vehicles

300
100
100
100

300



100



100



100



Parameter
(percent)
Ec

5.0
4.6
4. 1
5.9

5.0
5.0
5.0
5.0
5.4
5.8
6.9
6.8
3.7
4.4
4.5
4. 5
4. 5
3.8
3.8
3.7
Eo

16.7
19.0
14.4
16.9

44.4
40.7
24.4
26.6
40.9
37.9
26. 3
28.0
48.6
43.4
24. 5
27. 5
43.6
41.2
23.4
25. 1
FF

37. 1
37.6
39. 1
35. 1

10.6
14.4
31.4
29.3
15.7
18.7
30.8
29. 1
8.2
13.7
34.4
31. 7
8. 3
10.7
28.6
26. 8
STE(a)

0.690
0.664
0.731
0.676

0. 193
0.261
0. 562
0. 524
0.277
0. 330
0. 539
0. 509
0. 145
0.239
0. 584
0. 536
0. 160
0.206
0. 551
0. 517
(a)
                           % FF for short test
 („,  . ^   _,,  ..             70 t t tor snore test
  Short Test Effectiveness = fl FTP failures in same population
                                   72

-------
         Table 78.   Comparison of ST NOX Results by City —

  Predicted Population,  Bounded Errors of Commission Method

                        (E  Set at 5 Percent)
                           c
Short Test
Federal '
Short Cycle



Federal
Three-Mode















City

All
Chicago
Houston
Phoenix

All



Chicago



Houston



Phoenix



Test Mode






High speed
Low speed
[die in drive
[die in neutral
rfigh speed
Low speed
[die in drive
Idle in neutral
High speed
Low speed
[die in drive
[die in neutral
High speed
Low speed
Idle in drive
[die in neutral
No. of
Vehicles

300
100
100
100

300



100



100



100



Parameter
(percent)
Ec

5.0
4.2
4.4
6.4

5.0
5.0
5.0
5.0
3.2
3. 1
5.4
5.7
4.2
4. 5
3.8
4.9
6.8
7. 1
5.2
2.9
Eo

8.4
7. 5
10. 3
7.2

14.6
14.8
17.6
18.4
15.6
16. 1
16. 5
16.4
17.7
16.7
18.4
19.7
9.4
10.7
18.0
19.2
FF

12.8
12. 1
13. 3
13.0

6.6
6.4
3. 5
2.8
4.0
3. 5
3. 1
3. 1
5.9
6.9
5.2
3.9
10.8
9. 5
2.2
1.0
STE(a)

0.604
0.618
0. 562
0.645

0. 310
0. 301
0. 167
0. 130
0.204
0. 177
0. 159
0. 158
0.249
0.291
0. 22 1
0. 166
0. 535
0.471
0. 108
0.049
(a!
  Short Test Effectiveness =
                              FF for short test
                       FTP failures in same population
                                   73

-------
Table 79.  Comparison of ST Multiple Constituent Results by City—
     Actual Population,  Bounded Errors of Commission Method
                          (Ec Set at 5 Percent)
Short Test
Federal
Short Cycle



Federal
Three-Mode




















City

All
Chicago
Houston
Phoenix

All




Chicago




Houston





Phoenix




Test Mode






High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
(hi
Best' '
High speed
Low speed
Idle in drive
Idle in neutral
Best
No. of
Vehicles

300
100
100
100

300




100




100





100




Parameter
(percent)
Ec

7.0
10.0
4.0
7.0

3. 3
6.3
6. 3
4.7
4. 3
5.0
10.0
8.0
7.0
6.0
1.0
1.0
6.0
3.0

3.0
4.0
8.0
5.0
4.0
4.0
Eo

11.7
12.0
1 1.0
12.0

42. 7
37.0
19. 3
26. 3
22.0
44.0
36.0
23.0
28.0
26.0
46.0
40.0
15.0
24.0

18.0
38.0
35.0
20.0
27.0
22.0
FF

53.7
54.0
55.0
52.0

22.7
28. 3
46.0
39.0
43. 3
22.0
30.0
43.0
38.0
40.0
20.0
26.0
51.0
42.0

48.0
26.0
29.0
44.0
37.0
42.0
STE(a)

0.821
0.818
0. 833
0. 813

0. 348
0.429
0. 704
0. 597
0. 663
0. 333
0.455
0. 652
0. 576
0.606
0. 303
0. 394
0. 773
0. 636

0. 727
0.406
0.453
0.688
0. 578
0. 656
    Short Test Effectiveness =
                              % FF for short test
  (b)
                       % FTP failures in same population
    Best mode is: HC and CO at idle in drive
               NO at high speed
                                     74

-------
         Table 80.  Comparison of ST HC Results by Inertia Weight —
Predicted Population, Three Cities, Bounded Errors of Commission" Method
                              (Ec Set at 5 Percent)
Short Test
Federal
Short Cycle




Federal
Three-Mode



















Inertia Test
Weight Group
Ub)

All
0 to 2500
2501 to 3500
3501 to 4500
4501 and greater

All



0 to 2500



2501 to 3500



3501 to 4500



4501 and greater



Test Mode







High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles

300
50
119
98
33

300



50



119



98



33



Parameter
(percent)
. Ec

5.0
3. 5
4. 5
4. 5
8. 3

5.0
5.0
5.0
5.0
9.5
7. 5
4.4
4.2
4. 5
5. 5
6.6
5. 5
5.6
4. 5
4.2
5.0
0.0
1. 1
2. 3
3.9
Eo

13.0
19.0
il.4
9. 3
21.1

26. 1
24. 1
19. 8
21.0
22.0
21.5
24.0
25.0
24. 6
21.3
16. 3
19.0
22. 3
22. 1
18. 5
18.7
60.6
43.8
30. 6
29.4
FF

18.7
14.4
17.6
18.7
28. 6

5.4
7. 5
11.9
10.7
11.0
12.0
9.8
8.9
4.4
7.6
12.7
10.0
5. 3
5. 5
9.5
9. 3
0.01
5.9
19. 1
20. 3
STE(a)

0. 589
0.431
0.606
0. 667
0. 575

0. 172
0.237
0. 375
0.338
0. 340
0. 357
0.293
0.265
0. 151
0.264
0.439
0. 344
0. 192
0. 200
0. 340
0. 333
0.000
0. 119
0. 385
0.408
      (a)
        Short Test Effectiveness =
                                 % FF for short test
                           % FTP failures in same population
                                        75

-------
       Table 81.  Comparison of ST CO Results by Inertia Weight-
Predicted Population, Three Cities, Bounded Errors of Commission Method
                            (Ec Set at 5 Percent)
Short Test
Federal
Short Cycle




Federal
Three-Mode



















Inertia Test
Weight Group
(lb)

All
0 to 2500
2501 to 3500
3501 to 4500
4501 and greater

All



0 to 2500



2501 to 3500



3501 to 4500



4501 and greater



Test Mode







High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles

300
50
119
98
33

300



50



119



98



33



Parameter
(percent)
EC

5.0
13. 1
4.6
3. 1
2. 3

5.0
5.0
5.0
5.0
12.6
14. Z
14.6
14. 5
4.2
3.9
5.6
5.8
2.8
3.2
2.0
1.6
1.9
1.7
1.0
6. 1
Eo '

16.7
10.7
16.5
16.0
22. 3

44.4
40.7
12.4
26.6
19.5
21.0
15.8
17.8
42.7
37. 1
19.3
22.4
48.7
45.8
27.9
30.0
63.4
59.2
40.2
12. 3
FF
... .
37. 1
26.9
32.6
40.8
53. 1

10.6
14.4
31.4
29.3
18.0
'16.7
22.0
20.0
7. 3
12.7
31.0
28.2
10.4
13.8
31.8
30. 1
14.0
18.2
41.7
65.5
STE(a>

0.690
0.715
0.663
0.718
0.705

0. 193
0.261
0. 562
0. 524
0.483
0.443
0. 583
0. 528
0. 146
0.255
0.617
0. 557
0. 175
0.231
0. 53*3
0. 500
0. 181
0. 235
0.509 .
0.838
    (a)
      Short Test Effectiveness =
                               % FF for short test
                         Vo t' if tailures in same population
                                       76

-------
        Table 82.  Comparison of ST NOX Results by Inertia "Weight —
Predicted Population,  Three Cities,  Bounded Errors of Commission Method
                            (Er Set at 5 Percent)
Short Test
Federal
Short Cycle




Federal
Th ree-Mode



















Inertia Test'.
Weight Group
(lb)

All
0 to 2500
Z501 to 3500
3501 to 4500
4501 and greater

All



0 to 2500



2501 to 3500



3501 to 4500



4501 and greater



Test Mode







High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idl'e in d rive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in d rive
Idle in neutral
High speed
Low speed
Idle drive
Idle in neutral
No. of
Vehicles

300
50
119
98
33

300



50



119



98



33



Parameter
(percent)
Ec

5.0
0.8
4. 5
5.9
5. 1

5.0
5.0
5.0
5.0
13.2
3.9
0.0
4. 3
4.2
5.4
5.9
3.4
2.6
3.9
6.2
1.4
1. 3
1.6
4.7
0.01
Eo..

8.4
4.0
9. 1
9. 5
7. 8

14.6
14.8
17.2
18.4
2.4
4.0
5. 3
5.0
16. 3
13.9
19.2
23. 5
18.7
19. 3
20.9
22.8
16.6
17.0
16.0
18.7
FF

12.8
1. 3
14. 7
14. 3
11.0

6.6
6.4
3/5
2.8
3.0
1. 3
0.0
0.4
7. 5
10.0
4. 7
1.00
5. 1
4.6
2.9
1.0
2. 1
1. 8
2.9
0:02
STE(a)

0.604
0.251
0. 618
0.602
0. 585

0. 310
0. 301
0. 167
0. 130
0. 552
0.250
0.000
0.066
0.315
0.417
0. 195
0. 157
0.214
0. 191
0. 124
0.044
0. 114
0.095
0. 152
0.000
     (a)
      Short Test Effectiveness =
                                 FF for short test
                         % FTP failures in same population
                                      77

-------
Table 83.  Comparison of ST Multiple Constituent Results by Inertia Weight
 Actual Population,  Three  Cities, Bounded Errors of Commission Method
                              (E_ Set at 5 Percent)
Short Test
Federal
Short Cycle




Federal
Three-Mode
























Inertia Test
Weight Group
(lb)

All
0 to 2500
2501 to 3500
3501 'to 4500
4501 and greater

All




0 to 2500




2501 to 3500




3501 to 4500




4501 and greater




Test Mode







High speed
Low speed
Idle in drive
Idle in neutral
Best(b>
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in d rive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Be3t '
No. of
Vehicles

300
50
119
98
33

300









119




98




33




Parameter
(percent)
Ec

7.0
10.0
6,7
6. 1
6. 1

3. 3
6. 3
6. 3
4.7
4. 3
16.0
26.0
10.0
12.0
12.0
1. 7
3.4
7.6 ,
5.0
4.2
0.0
2.0
5. 1
2.0
2.0
0.0
0.0
0.0
0.0
0.0
Eo

11.7
6.0
16.0
10.2
9. 1

42.7
37.0
19. 3
26. 3
22.0
20.0
18.0
14.0
14.0
12.0
42.0
35. 3
15.0
26.0
17.6
51.0
43.9
25. 5
33.7
30.6
54.5
51.5
24.2
24.2
27.2
FF

53.7
40.0
49.6
60.2
69.7

3. 3
28. 3
46.0
39.0
43. 3
26.0
28.0
32.0
32.0
34.0
23. 5
30.2
50.4
40.0
48.0
19.4
26. 5
44.9
36. 7
39.8
24.2
27. 3
54.6
54. 5
51.5
STE(a)

0.821
0.870
0.756
0.855
0.885

0.072
0.433
0.704
0. 597
0.663
0. 565
0.609
0.696
0.696
0.739
0. 359
0.461
0. 771
0. 606
0.732
0.276
0. 376
0.637
0. 521
0. 565
0. 307
0. 346
0. 693
0.692
0.654
      (a)

      (b)
Short Test Effectiveness =;
                          7o FF for short test
                    % FTP failures in same population

Best mode is: HC and CO at idle in drive
               at high speed
                                          78

-------
        Table  84.   Comparison of ST HC Results by Displacement —
Predicted Population,  Three Cities, Bounded Errors of Commission Method
                              (Ec Set at 5 Percent)
Short Test
Federal
Short Cycle



Federal
Three-Mode















CID(a) Group

All
. 0 to 1 50
151 to 259
260 and greater

All



0 to 150



151 to 259



260 and greateT



Test Mode





~"
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed '
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles

300
95
54
151

300



95



54



151



Parameter
(percent)
Ec

5.0
4.6
4.6
5. 3

5.0
5.0
5.0
5.0
7. 3
7.0
7. 1
5.7
4.9
7.9
7. 1
6.8
3.6
2-9.
3.2
4. 1
Eo

13.0
13. 5
14.4
12. 1

26. 1
24. 1
19.8
21.0
25. 3
22.0
19.6
22.7
23.6
19. 1
16.9
18. 5
28. 3
27.2
21.2
20.9
FF

18.7
18.6
12.8
20.8

5. 4
7. 5
11.9
10.7
20.0
10. 1
12.6
9.4
3.6
8.2
10. 3
8.8
4. 3
5. 5
11.7
12.0
STE(b)

0. 589
0. 579
0.471
0.632

0. 172
0.237
0. 375
0. 338
0.258
0. 315
0. 319
0.293
0. 133
0. 300
0. 380
0. 323
0. 133
0. 168
0. 356
0. 366
      (a)
      (b)
CID = cubic inch displacement
Short Test Effectiveness = $?
% FF for short test
                           
-------
        Table 85.   Comparison of ST  CO Results by Displacement —
Predicted Population, Three Cities,  Bounded Errors of Commission Method
                             (Ec  Set at 5 Percent)
Short Test
Federal
Short Cycle



Federal
Three-Mode















CID(a) Group

All
0 to 1 50
1 5 1 to 2 59
260 and greater

All



0 to 150



151 to 2 59



260 and greater



Test Mode






iigh speed
Low speed
Idle in drive
.d\e in neutral
-ligh speed
Low speed
[die in d rive
[die in neutral
High speed
Low speed
[die in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles

300
95
54
151

300



95



54



151



Parameter
(percent)
Ec

5.0
12. 5
2. 7
2.7

5.0
5.0
5.0
5.0
11.0
10.4
12. 5
12. 1
3.9
3. 7
5.0
6.2
2.4
2. 5
1. 8
1.7
Eo

16.7
10.2
20. 5
18.6

44.4
40.7
12.4
26.6
31.6
27. 8
16.2
19.0
53.4
47.9
24.7
29.2
48. 5
45.4
28.8
30. 3
FF

37. 1
34. 5
35.0
38. 5

10.6
i4.4
31.4
29. 3
13. 1
17.0
28.6
25.8
5. 1
11. 1
34.8
30.2
10.2
13.4
31.1
29. 8
STE(b)

0.690
0. 773
0.631
0.674

0. 193
0.261
0. 562
0. 524
0.293
0. 379
0.639
0. 576
0.087
0. 188
0. 585
0. 508
0. 174
0.228
0. 519
0.496
      (a)

      (b)
CID = cubic inch displacement

Short Test Effectiveness =
% FF for short test
                          % FTP failures in same population
                                        80

-------
        Table 86.  Comparison of ST NOX Results  by Displacement —
Predicted Population,  Three Cities, Bounded Errors of Commission Method
                              (Ec Set at 5 Percent)
Short Test
Federal
Short Cycle



Federal
Three-Mode















CID(a) Group

All
0 to 150
151 to 259
260 and greater

All



0 to 150



1 5 I "to 2 59



260 and greater



Test Mode






High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of-
Vehicles

300
95
54
151

300



95



54



151



Parameter
(percent)
Ec

5.0
3. 1
5.0
6.7

5.0
5.0
5.0
5.0
9.9
4.7
1.2
1.0
2.8
3. 5
10. 1
10. 3
2.8
4.6
6. 5
3. 3
Eo

8.4
7. 3
7. 5
9.9

14.6
14.8
17.2
18.4
6.5
9.2
14.6
14.0
17.0
12.6
14.4
15.4
18.8
19.7
20.7
21.9
FF

12.8
8.2
12..5
14.9

6.6
6.4
3. 5
2.8
9.0
5.0
1.0
0.2
3. 1
7.4
5.6
4. 5
6.0
5. 1
4.0
2.8
STE(b)

0.604
0. 530
0.626
0.600

0. 310
0.301
0. 167
0. 130
0. 578
0. 355
0.062
0.016
0. 154
0. 372
0.280
0.228
0.242
0.206
0. 163
0. 115
     (a)
     (b)
CID = cubic inch displacement

Short Test Effectiveness =
% FF for short test
                          % FTP failures in same population
                                         81

-------
Tabl   87.  Comparison of ST Multiple Constituent Results  by Displacement —
Predicted Population,  Three Cities,  Bounded Errors of Commission Method
                              (Ec Set at 5 Percent)
Short Test
Federal
Short Cycle



Federal
Three-Mode



















CID(a) Group

All
0 to 150
1 5 1 to 2 59
260 and greater

AH




0 to 150




151 to 259




260 and greater




Test Mode






iigh speed
L.OW speed
Idle in drive
die in neutral
Best
Bigh speed
Los speed
[die in drive
[die in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
-------
   Table 88.  Comparison of ST HC  Results by Emission Controls Type —
Predicted Population,  Three Cities, Bounded Errors of  Commission Method
                              (E   Set at  5  Percent)
Short Test
Federal
Short Cycle



Federal
Three-Mode


















Emission
Controls
Catalyst/Air(a)
All
Yes/Yes
Yes/No
No/Yes
No /No
All



Yes/Yes



Yes/No



No /Yes



No /No



Test Mode





High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
89
127
59
25
300



89



127



59



25



Parameter
(percent)
Ec
5.0
3.2
4.5'
7.5
4.7
: 5.0
5.0
5.0
5.0
4.7
.4.5
• 3'. 5
3.5
3.4
2.8
4.6
5.5
7. 1
8.0
5.4
' 3,5
8.8
9.3
6.8
5.3
Eo
13.0
8.6
14.7
10.7
34.0
26. 1
'24. 1
19.8
21.0
15.9
15.6
17.0
17.2
33.0
30.4
'18.3
18.6
21.2
20.3
20. 1
24.5
40. 1
32. 1
35.3
37.2
FF
18.7
11.4
. 22.3
18.9
20.4
'5.4
7.5
11.9
10.7
3.6
3. 8
3.0
2.7
3.9
6.6
18.6
18.4
8.4
9.3
9.5
5. 1
14.3
22.3
19. 1
17.2
STE(b)
0.589
0.570
0.602
0.639
0.376
0/172
0.237
0.. 375
0.338
0.183
0. 197
0. 149
0. 136
0.106
0. 179
0.504
0.498
0.285
0.315
0.321
0. 172
0.263
0.410
0.352
0.316
     (a)
     (b)
Secondary air injection system

Short Test Effectiveness  =
FF for short test
                          %-FTP failures in same population
                                        83

-------
   Table 89.  Comparison of ST;CO Results by Emission Controls  Type —
Predicted Population,  Three Cities, Bounded Errors of Commission' Method
                             (E  ;Set at 5 Percent)
Short Test
Federal
Short Cycle



Federal'
Three-Mode


















Emission
Controls;
Catalyst/ Air'a)
All
Yes/Yes
Yes/No
No/Yes
No /No
All



Yes/Yes



Yes/No



No /Yes



No /No



Test Mode





High speed
Low speed
Idle in drive
idle in neutral
Hfcgh speed
Low speed
Idle in drive
Idle In neutral
High speed
Low, speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
89
127
59
25
300



89



127



59



25



', Parameter
(percent)
•' Ec :
5.0;
'4.8i
2.1
17.0
8.1
5.0
5.0
5.0
5.0
2. 1
3.0
2.5
2.5
2.1
1.6
2.8
3.2
7.6
9.6
15.0
13.3
7.4
15.3
10.7
9.2
Eo '
16.7
14.3
20.0
7.3
15.1
44.4
40.7
12.4
26.6
31. 1
27.2
22.0
24.2
59.2
56.0
30.1
32.0
27.6
25.9
16.5
22. 1
34.3
30. 8
13.6
13.3
FF
37. 1
25.6
46.7
34. 1
33.8
10.6
14.4
31.4
29.3
10.4
14.2
20. 1
18.0
8.8
12.7
39.3
37.9
13.9
15.5
25.0
19.3
14.5
18.0
35.2
35.4
. sra^'
.0.690
. 0.642
0.700
0.823
0.691
0.193
0.261
0.562
0.524
0.250
0.344
0.478
0.425
0. 129
0. 185
0.566
0.543
6.334
0.375
0.602
0.467
0.298
0.369
0.721
0.725
     (b)
      Short Test Effectiveness = -a
% FF for short test
                          % FTP failures in same population
                                       84

-------
   Table 90.  Comparison of ST NOX Results by Emission Controls Type —
Predicted Population,  Three'Cities, Bounded Errors of Commission Method
                               (E  Set at 5 Percent)
Short Test.
• Federal
Short Cycle



Federal
Three-Mode


















Emission
Controls
Catalyst/Air(a)
All
Yes/Yes
Yes/No
.No/Yes
No /No
All



Yes/Yes



Yes/No



No/Yes



No /No



Test Mode





High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
89
127
59
25
300



89



127



59



25



Parameter
(percent)
Ec
5.0
3.8
7.2
1.4
3.9
5.0
5.0
5.0
5.0
1.9
1.3
4.7
4.2
5.4,
8. 1
8.7
6.7
6.6
3.6
0.5
0.03
12.9
•6.2
2.8
4.9
Eo
8.4
14.8
6.3
8.8
4.2
14.6
14.8
17.2
18.4
21.8
25.6
25.6
27.6
11.5
9.6
12.8
13.5
9.3
10.0
17. 1
19.6
11.3
10.5
10.4
10.7
FF
12.8
20^6
9J6
9.6
9.7
6:6
6.4
3.5
2.8
8.6
4.8
4.8
2:7
4.4
6.3
3. 1
2.4
9. 1
8.3
113
0.04
2.6
3.4
3.6
3.2
STE
0.604
0.582
0.602
0.522
0.696
0.310
0.301
0. 167
0. 130
0.283
0. 158
0. 157
0.090
0.279
0.395
0. 195
0. 153
0.495
0.451
0.068
0.002
0. 188
0.244
0.256
0.230
      (a)
      (b)
Secondary air injection system

Short Test Effectiveness  =
% FF for short test
                           % FTP failures in same population
                                         85

-------
   Table 91.  Comparison of ST Multiple Constituent Results by
  Emission Controls Type— Predicted Population, Three Cities,
   Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Teet
Federal
Short Cy •



Federal
Three-Mode























Emission
Controls
Catalyst/Air(a)
All
Yes/Yes
Yes/No
No/Yes
No /No
All




Yes/Yes




Yes/No




No/Yes




No /No




Test Mode





iigh speed
Low speed
Idle in drive
Idle in neutral
Best(C)
riigh speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best(C>
No. of
Vehicles
300
89
121
59
25
300




89




127




59




25




Parameter
(percent)
Ec
7.0
10.1
3.9
5. 1
16.0
3.3
6.3
6.3
4.7
4.3
1. 12
5.6
3.4
t.l
3.4
1.0
2.4
6.3
3.9
2.4
6.8
10.2
8.5
6.8
8/5
16.0
20.0
12.0
16. 0
8. 0 '
Eo
11.7
13.5
13.4
8.5
4.0
42.7
37.0
19.3
26.3
22.0
38.2
38.2
30.3
36.0
30.3
52.0
43.3
13.4
19.7
18.9
32.2
30.5
18.6
30.5
20.3
36.0
16.0
12.0
16. 0
12. 0
FF
53.7
43.8
59.8
54.2
56.0
3.3
28.3
46.0
39.0
43.3
19. 1
19. 1
27.0
21.3
27.0
21.3
30.0
59.8
53.5
54.3
30.5
32.2
44. 1
32.2
42.4
24.0
44.0
48.0
44. 0
48. 0
STE(b)
0. 821
0.764
0. 816
0. 864
0.933
0.072
0. 433
0.704
0. 597
0. 663
0. 333
0. 333
0. 471
0. 372
0. 471
0. 291
0.40P
0. 817
0.731
0. 742
0. 486
0.514
0.703
0. 514
0. 676
0.400
0. 733
0. 800
0. 733
0. 800
(b).

(c)
Secondary air injection system

'Short Test Effectiveness =
                            % FF for short test
                    70 FTP failures in-same population

Best mode is: HC and CO at idle in drive; NOX at high speed

-------
      Table 92.  Comparison of ST HC Results by Fuel System Type —
Predicted Population,  Three Cities,  Bounded Errors of Commission Method
                             (E  Set at 5 Percent)
Short Test
Federal
Short Cycle

Federal
Th ree-Mode










Fuel
System Type
All •
Fuel Injection
Carburetion
All



Fuel Injection



Carburetion



Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
rligh speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
30
270
300



30



Z70



Parameter
(percent)
Ec
5.0
3.3
5.2
5.0
5.0
5.0
5.0
10.7
11.9
7.0
6.2
4.3
4.3
4.8
4.9
Eo
13.0
20.7
12.3
26.1
24. 1
19.8
21.0
26.0
24.0
24.8
26.5
26. 1
24.2
19.3
20.4
FF
18.7
20.7
18.4
5.4
7.5
11.9
10.7
15.3
17.4
16.6
14.9
4.4
6.4
11.3
10.2
STE(a)
0.589
0.501
0.600
0. 172
0.237
0.375
0.338 •
0.370
0.420
0.401
0.360
0.145
0.209
0.370
0.334
     (a)
       Short test effectiveness =
                              % FF for short test
                         % FTP failures in same population

-------
      Table 93.  Comparison of ST CO, Results by.Fuel,System Type —...
Predicted Population, Three Cities,  Bounded Errors, of Commission Jvlethod
                              (E  Set,at .5 Percent)
Short Test
Federal
Short Cycle

Federal
Three-Mode










Fuel
System Type
All (
Fuel Injection
Carburetion
All



Fuel Injection



Carburetion .'



Test Mode ,



High speed
Low speed
!dle in drive
!dle' in neutral
High speed
Low speed
[die in drive
Idle in neutral
-ligh speed
Low speed
[die in drive
[die in neutral
.;No. of
Vehicles
300
30
. 270
300



30



270



Parameter
: (percent) ;
Ec
5.0
8.9
. 4.4
5.0
5.0
5.0 ,
5.0
9.3
12.2
13.2'
10.7
4.'1
4.1 ;
4. 1
4.3
Eo
16.7
5. .8 ...
17.4
44.4
40.7
12.4
26.6
23.3
22. 1
9.6
12.0
46.3
42.5
25.8
28.0
.. FF
37. 1
38. 1
37.3
10.6
14.4
31.4
29.3
20.6
21.8
34.3
32.0
9. '8
13.7
31.2
29.2
STE(a>
0.690
: 0.868
0.682
0. 193
0.261 • .
0.562
0.524
0.470
0.495
0.781
0.728
0. 175
0.244
0.458
0.511
      'Short Test Effectiveness =•
                               % FF for short test
                          % FTP failures in same population
                                         88

-------
      Table  94.   Comparison of ST NOX Results by Fuel System Type —
Predicted Population,  Three Cities, Bounded Errors of Commission Method
                              (E   Set  at 5 Percent)
Short Test
Federal
Short Cycle

Federal
Three-Mode










Fuel
System Type
All
Fuel Injection
Carburetion
All



Fuel Injection



Carburetion


•-
Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
30
270
300



30



270



Parameter
(percent)
Ec
5.0
2.2
5. 1
5.0
5.0
5.0
5.0
5.3
6.9
0.16
5.3
4.6
4.& .
5.5
4.6
Eo
8.4
6.1
8.4
14.6
14.8
17.2
18.4
2.9
5.6
7.7
7. 5
15.7
15.7
18.3
19.4
FF
12.8
1.6
14.2
6.6
6.4
3.5
2.8
2. 1
2. 1
0.01
0. 18
6.8
6.9
4.2
3. 1
STE(a)
0.604
0.201
0.629
0.310
0.301
0. 167
0. 130
0.426
0.271
0.001
0.024
0.301
0.305
0. 187
0. 137
     (a)
       Short Test Effectiveness =
                               % FF for short test
                          % FTP failures in same population
                                        89

-------
    Table 95.   Comparison of ST Multiple Constituent Re suits;b>y
      Fuel System Type — Predicted Population, Three Cities,
    Bounded Errors of Commission'Method  (E  Set at 5 Percent)
Short Test
Federal
Short Cycle

Federal
Three-Mode













• Fuel '
System Type-
All
Fuel Injection
Carburetion .
All




Fuel Injection




Carburetion




Test Mode

'.
,
High speed
Low speed
Idle in drive
Idle in neutral
Best(b)
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
No. ,of
.Vehicles
300
• 30
270
300




30




270




t
' Parameter
(percent)
Ec
7.0
10.0 :
6.7
3.3
6.3
6.3
4.7
4.3
16.7
23.3
13.3
16.7
13.3.
1.8
4.4
5.6
3.3
3.3
Eo
11.7
6.7
12. 2
42.7
37.0
19.3
26.3
2Z.O
30.0
23.3
10.0
13.3
6.7
44.0
38.5
20.4
27.8
23.7
. FF .. .
53.7
' 4.6.7
5'4.4
3.3
28.3
46.0
39.0
43.3
23.3
30.0 .
43.3
40.0
46.7
22.6
28. 1
46.3
38.9
43.0
'.STE(a)
0. 821
0. 876
0.816
0.072 .
0. 433
0.704
0. 597
0. 663
0. 437
0. 563
0. 812
0.751
0.875 '
0. 339
0. 422
0. 694
0. 583
0. 645
(a)
  Short Test Effectiveness = •
                             FF for short test
                      % FTP failures in same population

'  'Best mode is: HC and CO at idle in drive
             NO  at high speed
                                    90

-------
      Table 96.  Comparison of ST HC Results by Transmission Type-
Predicted Population, Three Cities-, Bounded Errors of Commission Method
                              (E  Set at  5 Percent)
Short Test
Federal
Short Cycle

Federal
Three-Mode










Transmission
Type
All
Automatic
Manual
All



Automatic



Manual



Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
High speed
'Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles'
300
229
71
300



229



71



Parameter
(percent)
- Ec
5.0
5. 1
4.8
5.0
5.0
5.0
5.0
4.5
4.8
4.9
4.9
6.5
5.6
5.4
5.6
Eo
13.0
11.4
18.9
26.1
24.1
19.8
21.0
24.0
23.7
18.2
19.8
27.7
25.7
25.6
25.3
FF
18.7
19.0
17.4
5.4
7.5
11.9
10.7
4.5
6.6
12.2
10.6
8.6
10.6
10.7
10.9
STE(a)
0.589
0.624
0.481
0.172
0.237
0.375
0..338
0. 150
0.217
0.401
0.348
0.236
0.292
0.296
0.301
     (a).,,   _   „,,  .             % FF for short test
     *  Short Test Effectiveness = ... ~TT-, t -i	'•	TT	
                          % FTP failures in same population
                                       91

-------
      Table 97.  Comparison of ST CO Results by Transmission Type
Predicted Population,  Three Cities, Bounded Errors of Commission Method
                             (E  Set at  5  Percent)
Short Test
Federal
Short Cycle

Federal
Three-Mode










.Transmission
Type
All
Automatic
Manual
All



Automatic



Manual



Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
229
71
300



229



71



Parameter
(percent)
Ec
5.0
3.4
12.2
5.0
5.0
5.0
5.0
3.5
3.4
3.6
3.6
10.8
10.1
10.8
10.7
Eo
16.7
18. 1
10.8
44.4
40.7
12.4
26.6
47.6
44.5
26.1
28.7
32.4
26.2
18.3
18.4
FF
37. 1
38..0
32.8
10.6
14.4
31.4
29.3
10. 1
13.4
32.7
30.2
11.2
17.4
25.4
25.3
STE(a)
0.690
0.678
0.752
0.193
0.261
0.562
0.524
0. 175
0.231
0.556
0.512
0.257
0.399
0..582
0.578
     (a),
       Short Test Effectiveness =
                               % FF for short test
                          % FTP failures in same population
                                      92

-------
     Table 98.   Comparison of ST NOX Results by Transmission Type—
Predicted Population,  Three Cities,  Bounded Errors of Commission Method
                             (E  Set at 5 Percent)
Short Test
Federal
Short Cycle

Federal
Three-Mode










Transmission
Type
All
Automatic
Manual
All



Automatic



Manual



Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
3'00
229
71
300



229



71



Parameter
(percent)
Ec
5.0
5.6
1.0
5.0
5.0
5.0
5.0
3.4
4.4
7.3
5. 1
10.6
6.0
1.0
4. 1
Eo
8.4
9.4
6.3
14.6
14.8
17.2
18.4
18.0
17.3
19.2
20.9
4.2
6.0
11.9
10.2
FF
12.8
14.7
3.2
6.6
6.4
3.5
2.8
6.0
6.8
4.8
3.2
7.7
6.0
1.0
1.7
STE(a)
0.604
0.610
0.334
0.310
0.301
0. 167
0. 130
0.251
0.28'1
0.200
0. 132
0.647
0.496
0.006
0. 145
    (a)0,  . _  . „,.  ..        	% FF for short test	
      Short Test Effectiveness = ^ FTp failurgs to same pc^tion
                                       93

-------
   Table 99.  Comparison of ST Multiple Constituent Results by
     Transmission Type — Predicted Population, Three Cities,
   Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Test
Federal.
Short Cycle

Federal
Three-Mode













Transmission
Type
All
Automatic
Manual
All




Automatic




Manual




Test- Mode



High speed
Low speed
Idle in drive
Idle in neutral
Best(b>
High speed
Low speed
Idle in drive
Idle in neutral
Best(b)
High speed
Low speed
Idle in drive
Idle in neutral
Best
0. 821
0.813
0. 870
0.072
0.433
0.704
0. 597
0. 663
0.319
0.419
0.706
0. 575
0. 650
0.472
0. 500
0. 694
0. 694
0.722
 'Short Test Effectiveness = •
                           % FF for-short test
(b)
                    % FTP failures-in same population

Best mode is: HC and CO at idle in drive
           NO  at high speed
                                   94

-------
        Table 100.  Comparison of ST HC Results by Manufacturer —
Predicted Population, Three Cities, Bounded Errors of Commission Method
                              (E  Set  at  5 Percent)
Short Test
Federal
Short Cycle




Federal
Three-Mode





















Manufacturer
All
General Motors
Ford
Chrysler
American
Motors
Other
All


General Motors



Ford



Chrysler



American
Motors


Other



Test Mode





High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed'
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
99
61
36
18
86
300


99



61



36



18


86



Parameter
(percent)
Ec
5.0
3.9
3.8
3.6
0.4
5. 1
5.0
5.0
5.0
5.0
3. 1
2.9
4. 5
5.2
4. 1
4. 1
2. 3
3.6
2.4
2.0
3.7
7.8
6.0
4. 5
5. 1
3.9
8.6
8.9
6. 1
4.6
Eo
13.0
10. 1
9.0
20.7
11.3
15.0
26.2
24. 1
19.8
21.0
27. 1
24.4
15.2 •
15.9
22. 1
21. 3
20. 3
19.8
40. 8
40.2
36.4
29. 1
16.7
13. 5
14.6
16. 3
23.6
21.6
22, 5
25.8
FF
18.7
21.0
16.2
36.5
10.9
17.9
5.4
7. 5
11.9
10.7
4.0
6.7
15.9
15.2
3. 1
4.0
4.9
5. 5
5. 4
5.9
20.8
17.6
3.0
6. 1
7. 5
5.8
9.4
11.3
10.5
7. 1
STE(a)
0. 589
0.676
0.642
0.638
0. 491
0. 544
0. 172
0.237
0. 375
0. 338
0. 129
0.215
0. 511
0. 489
0. 123
0. 157
0. 195
0.217
0. 1 16
0. 129
0. 364
0. 377
0. 150
0. 312
0.340
0.263
0.285
0. 343
0. 318
0.216
     (a).
       Short Test Effectiveness =
                                 FF for short test
                          % FTP failures in same population
                                        95

-------
        Table 101.  Comparison of ST  CO Results by Manufacturer —
Predicted Population, Three Cities, Bounded Errors  of Commission Method
                             (E  Set at 5 Percent)
Short Teat
Federal
Short Cycle




Federal
Three -Mode





















Manufacturer
All
General Motors
Ford
Chrysler
American
Motors
Other
All


General Motors



Ford



Chrysler



American
Motors


Othe r '



Test Mode





High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
99
61
36
18
86
300


99



61



36



18


86



Parameter
(percent)
Ec
5.0
2.4
4.8
0.2
1.2
14.0
5.0
5.0
5.0
5.0
1.5
1.4
2.0
2.6
3.0
3.9
4.2
3. 1
1. 5
0.4
9. 5
10. 1
0. 4
0.0
2. 5
1.0
8.7
10.9
12.7
12. 1
Eo
16.7
18.2 .
14.4
28. 1
14.7
8.9
44. 4
40. 7
24. 4
26.6
51.3
45.0
26.7
28.9
45. 0
'37.7
28.3
31.7
69.0
76. 3
8.2
8.4
23.7
41.9
14.2
14. 3
24.9
22.9
16.6
20. 3
FF
37. 1
35.8
39.4
52. 5
15.3
33. 3
10.6
14.4
31.4
29.3
5. 1
11.8
31. 1
28.9
11.2
18.6
29.3
26.0
11.9
5.6
71.8
71.7
6.3
0.0
15. 7
15.6
17. 2
19.3
25.6
21.9
STE(a)
0.690
0.663
0.732
0.652
0. 510
0.790
0. 193
0.261
0. 562
0. 524
0.091
0.208
0.538
0. 500
0.200
0. 330
0. 509
0. 451
0. 148
0.069
0. 898
0.'896
0. 210
0
0. 525
0. 522
0. 409
0.457
0.607
0.519
    •(a)
      Short Test Effectiveness =
                               % FF for short test
                         % FTP failures in same population
                                       96

-------
       Table 102.  Comparison of ST NOX Results by Manufacturer —
Predicted Population, Three Cities, Bounded Errors of Commission Method
                             (E   Set  at 5 Percent)
Short Test
Federal
Short Cycle




Federal
Three-Mode





















Manufacturer
All
General Moto.rs
Ford
Chrysler
American
Motors
Other
All


General Motors



Ford



Chr.ys ler



Ame rican
Motors


Other



Test Mode





High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in d rive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
300
99
61
36
18
86
300


• 99



61



36



18


86



Parameter
(percent)
Ec
5.0
1.6
3. 5
23.6
14.5
2.5
5.0
5.0
5.0
5. 0
4.2
2. 8
9.6
5.2
1. 4
0. 4
4.9
0.0
7. 1
22.4
4.8
2. 5
0.2
4.2
14.3
6.2
7. 5
5.4
0. 7
1.0
Eo
8.4
8.8
16. 1
2.9
1.4
7.9
14.6
14.8
17.6
18.4
15. 5
15. 3
14. 9
16.6
23.7
28. 1
28. 3
36.6
10.6
4.0
14.9
15. 1
28.7
24.0
17. 8
24.0
7.6
7.9
14.6
14. 0
FF
12.8
11.4
13.9
13.3
30. 1
7.6
6.6
6.4
3. 5
2.8
4.6
4.9
5. 2
3. 5
6. 3
2.0
1.8
0.0
5.7
12. 3
.I-4
1.2
2.7
5.2
13.7
5. 3
6.6
7.6
1.0
0.3
STE(a)
0.604
0. 564
0.463
0. 817
0.957
0.492
0. 310
0. 302
0. 168
0. 130
0.229
0.243
0.259
0. 175
0.210
0.065
0.059
0
. 0. 349
0.755
0.085
0.076
0.087
0. 177
0. 434
0. 181
0.467
0. 489
0.062
0.020
      )_,   „   _,,  ..            To FF for short test	
      'Short Test Effectiveness = % FTP faiiures in same population
                                       9?

-------
Table 103.  Comparison of ST Multiple Constituent Results  by Manufacturer —

Predicted Population, Three Cities, Bounded Errors of Commission Method

                               (E  Set at 5 Percent)
                                 c
Short Test
Federal
Short Cycle




Federal
Th ree -Mode





















Manufacturer
All
General Motors
Ford
Chrysler
American Motors
Other
All


General Motors



Ford



C h r y s 1 c r



American
Moto rs


Other



Test Mode





High speed
Low speed
Idle in drive
Idle in neutral
Best(b)
High speed
Low speed
Idle in drive
Idle in neutral
Bestfb)
High speed
Low speed
Idle in drive
Idle in neutral
Best(b)
High speed
Low speed
Idle in drive
Idle in neutral
Best161
High speed
Low speed
Idle in drive
Idle in neutral
Best'6'"
I ligh speed
Low speed
Idle in drive
Idle in neutral
Best(bl -.
No. of
Vehicles
300
99
61
36
18
86
300


99



61



36



18


86



Parameter
(percent)
Ec
7.0
2.0
8.2
8. 3
16.7
9-3
3. 3
6. 3
6. 3
4.7
4. 3
1.0
1.0
7. 1
1.0
1.0
0.0
4.9
1.6
1.6
1.6
0.0
0.0
5.6
5.6
5.6
0. 0
0.0
5.6
5.6
0.0
10.5
17.4
9. 5
10. 5
10. 5
Eo
11.7
8. 1
19.7
22.2
5.6
7.0 '
42.7
37.0
19.3
26. 3
22.0
48. 5
41. 4
13. 1
20.2
20.2
45.9
44. 3
32. 8
37.7
31.2
58. 3
41. 7
19.4
19.4
19.4
50.0
44. 4
27. 8
50. 0
44. 4
25.6
23. 3
15. 1
23. 3
14. 0
FF
53.7
57.6
45.9
63.9
61. 1
48.8
22. 7
28. 3
46.0
39.0
43. 3
17.2
24. 2
52. 5
45. 5
45. 5
19.7
21.3
32. 8
27.9
34. 4
27. 8
44. 4
66.7
66.7
66.7
16.7
22. 2
38. 9
16.7
22.2
30. 2 .
32.6
40.7
32.6
41. 9
STE(a)
0. 821
0. 877
0. 700
0. 742
0.916
0. 875
0. 347
0. 433
0. 704
0. 597
0.665
0. 262
0. 369
0. 800
0. 693
0.693
0. 300
0. i25
0. 500
0. 425
0. 524
0. 323
0.516
0. 775
0. 775
0. 775
0. 250
0.408
0. 581
0.250
0. 408
0. 541
0. 58i
0.729
0. 583
0. 749
      (a)_.   _    cr  ,.       	% FF for short test
        Short lest Lucrtiveness = TT.—,^~.. -.—r:	:	-.
                            "(i FTP failures in same popul

      (h)
                                          ation


Bc-st mode is: HC and CO al idle in drive; NO  at high speed
                                         98

-------
 Table 104.  Estimated Standard Errors for Contingency Table Percentages
                    in Multiple Constituent Test Results
No. of
Vehicles
300
151
100
95
54
Estimated Percentage
5
1.26
1.77
2.18
2.24
2.97
10
1.73
2.44
3.00
3.08
4.08
20
2.31
3.26
4.00
4. 10
5.44
30 or 70
2.65
3.73
4.58
4.70
6.24
40 or 60
2.82
3.99
4.90
5.03
6.67
50
2.89
4.07
5.00
5.13
6.80
               Table 105.  Comparison of Federal Short Cycle
                  Cut-Points by City (E  Set at 5 Percent)
City
All
Chicago
Houston
Phoenix
No. of
Vehicles
300
100
100
100
ST Cut Points and Standard Errors
(gm/mi)
HC
CP
1.32
1.24
1.24
1.44
SE
0. 145
0.283
0. 181
0. 267
CO
CP
8.40
7.99
7.49
9.48
SE
0.866
1. 240
1.327
1.916
NO
X
CP
2.62
2.54
2.56
2.77
SE
0.219
0.355
0.360
Oo412
(a)

(b)
CP = cut point

SE = standard error
                                    99

-------
    Table 106.  Comparison of Federal Three-Mode Cut-Points by City
                           (E  Set at 5 Percent)
City
All
Chicago
Houston
Phoenix
No. of
Vehicles
300
100
100
100
ST Cut-Points and Standard Errors
(ppm) (*)
HC Idle in Drive
CP
184
241
132
174
SE •
26
78
27
31
CO Idle in Drive
CP
7,081
10, 145
6, 175
5,494
SE
1, 251
2, 627
2,432
1, 510
NO High Speed
3t
CP
2,265
2,026
2,172(d)
2,439
SE
243
410
393(d)
379
(a)
(b)
(c)
(d)
ppm = parts per million
CP = cut point
SE  = standard error
The best mode for NOX for Houston was found to be the low speed mode.
The cut point and standard error for NOX low speed are 1500 ppm and
372 ppm,  respectively
                                   100

-------
 Table 107.  Comparison of HC Results for CID    = 150 and Less,
       by Manufacturer — Predicted Population, Three Cities,
    Bounded Errors of Commission Method (E • Set at 5 Percent)
Short Test
Federal
Short Cycle









Federal
Three-Mode






























Manufacturer
All
General Motors
Ford
Chrysler
American Motors
Datsun
Toyota
VW, Porsche,
Audi
Honda
Others
All



General Motors



Ford



Datsun



Toyota



VW, Porsche,
Audi


Honda



Others



Test Mode











High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
95
1Z
6


8
9
21

6
33
95



12



6



8



9



21



6



33



Parameter
(percent)
Ec
5.0
2.9
2. 1


3.9
0.3
2.7

4.9
7.2
5.0
5.0
5.0
5.0
0.65
0. 30
11.3
10. 2
0.8
1.2
3. 3
5. 2
10.7
10.6
5.4
0.9
2.6
4.0
2.8
1. 5
8. 1
5.8
4.2
4.6
0.0
0.07
2.9
5.6
5. 4
4. 5
4.2
3.7
Eo
13.0
13.4
5.4


6'. 7
31. 5
18.9

7.4
10. 3
25.9
24. 0
21.9
23.4
31.8
30.6
14. 4
13.6
19. Z
19.2
13. 3
12. 5
22. 3
19.0
30. 3
44.9
33. 5
15.9
37. 8
40.2
24.6
21.9
23.4
22. 3
8.2
8.2
2.6
1. 5
25.6
25.6
21.9
25. 1
FF
19. 1
20. 5
13.9


41.7
11.8
17.7

0.8
19.4
6.2
8. 1
10.2
8.7
2. 1
3. 3
19.5
20. 3
0. 1
0. 1
6.0
6.7
26. 1
29.4
18. 1
3.6
9.9
27. 5
5. 5
3. 1
12.0
14.6
13. 1
14.2
0.0
0.02
5.6
6.6
4. 1
4. 1
7.8
4.7
STE(b)
0. 595
0.605
0.720


0.862
0.273
0.484

0.098
0.653
0. 193
0. 252
0.318
0. 271
0.062
0.097
0. 575
0. 599
0.005
0.005
0.311
0. 349
0. 539
0.607
0. 374
0.074
0.228 '
0.633
0. 127
0.072
0. 328
0.400
0. 359
0. 389
0
0.002
0.683
0.815
0. 138
0. 138
0.263
0. 158
(a),
  CID = cubic inch displacement
(b),
  Short Test Effectiveness =
                            FF for short test
                     %FTP failures in same population
                                   101

-------
 Table 108.   Comparison of CO Results for CID^ .= 150 and.Less,
      by Manufacturer .— Predicted Population,  Three Cities,
   Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Test
Federal
Short Cycle









Federal
Three -Mode






























.Manufacturer
All -
General Motors
Ford
Chrysler
American Motors
Datsun
Toyota
VW, Audi,
Porsche
Honda
Others
All



General Motors



Ford



Datsun



Toyota



VW, Audi,
Porsche


Honda



Others



Test Mode











High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
95
12
6


8
9
21

6
33
95



12



6



8



9



21



6



33



• Parameter"
(percent)
Ec
5.0
0. 1
0.0


0.4
3. 7
4.9

0:0
6.8
5.0
5.0
5.0
5.0
0. 8
0.05
6.7
6.6
0.0
0.0
0. 1
0.06
0.0
0.0
10.9
4.0
0.03
7. 1
4.7
3.9
5. 1
5.6
8.4
8.9
q.o
1.3
0.0
0.0
2. 5
4.6
2. 8
2. .9
E0
19.4
36.9
23.8


33. 1
25.6
24. 5

Q.O
10.7
38. 0
34.7
26.4
28.8
75. 1
68.4
15.4
15. 7
25.6
25.6
25.6
25.6
40.9
32. 3
14. 7
39.5
34.2
32. 5
21.6
29.0
30.7
33. 5
30. 0
31.6
0.0
0.0
0.0
0.0
<2.6
29.7
26.5
30.0
FF
25.4
40. 5
1.8


19.8
8.6
23.4

0.0
30.8
6.7
10.0
18.4
16.0
2. 3
9.0
64. 1
63.8
o.'o
0.0
0.05
0.03
6.0
14.6
38.2
13. 3
0.01
1.7
12.6
5.2
17.2
14.4
18. 3
16.7
0.0
0.0
0.0
0.0
8.9
11.8
15.0
11.5
STE
-------
Table 109.  Comparison of NOX Results for CID^a' =  150 and Less,
      by Manufacturer — Predicted Population,  Three Cities,
   Bounded Errors of Commission Method (Ec Set at 5 Percent)
Short Test
Federal
Short Cycle









Federal
Three -Mode






























Manufacturer
All
General Motors
Ford
Chrysler
American Motors
Datsun
Toyota
VW, Audi,
Porsche
Honda
Others
All



General Motors



Ford



Datsun



Toyota



VW, Audi,
Porsche


Honda



Others



Test Mode











High speed
I_»ow speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
95.
12
6


8
9
21

6
33
95



12



6



8



9



21



6



33



Parameter
(percent)
Ec
5.0
3.9
19.2


2.9
0.7
1.2

0.0
1.7
5.0
5. 0
5.0
5.0
9.9
11.0
18.0
4. 7
0.8
0.01
1. 3
0.07
27.2
0.9
0.0
0.01
0.34
3.8
1.0
0.8
0.6
10.3
5.3
25.7
0.7
0.8
0.06
10.5
2.4
2.4
2.0
0.2
Eo
5.8
2.2
5.8


10.8
35.9
1. 1

0. 1
3.8
9.0
9.0
12.6
13.2
2.7
5.9
10.3
10.7
12. 5
12.9
11.2
12. 5
23.8
21.0
26.7
26.7
57.0
21.7
52.6
58.2
0.9
0. 3
0.8
0.6
0. 1
0.01
0. 1
0. 1
9.8
8.6
10.4
11.9
FF
9.8
8.8
7. 1


14.7
28.6
0.2

0.0
8.8
6.5
5.2
2.9
1.0
8.3
5.2
0.7
0.31
0.4
0.04
1.7
0.01
2.8
4.4
0.0
0.0
6.6
37.7
6.7
1. 1
0.4
0.9
0.4
0.6
0.03 '
0.0
0.0
0.0
2.8
4.0
2.2
0.2
STE
0.628
0.800
0.550


0. 577
0. 443
0. 154

0
0.698
0. 419
0. 366
0. 187
0.070
0.755
0. 468
0.064
0.028
0.031
0.003
0. 132
0
0. 105
0. 173
0
0
0. 104
0.635
0. 113
0.019
0. 308
0.750
0. 333
0. 500
0.231
0
0
0
0.222
0.318
0. 175
0.017
(b)
CID = cubic inch displacement'
Short  Test Effectiveness =
                          % FF for short test
                     % FTP failures in same population
                                   103

-------
Table 110.  Comparison of Multiple Constituent Results for CJXr'  =  150
   and Less,  by Manufacturer —Predicted Population,  Three Cities,
       Bounded Errors of Commission. Method (E  Set at 5 Percent)
Short Test
Federal
Short Cycle









Federal
Three -Mode






































Manufacturer
All
General Motors
Ford
Chrysler
American Motors
Datsun
Toyota
VW, Audi,
Porsche
Honda
Othe rs '
All




General Motors




Ford




Datsun




Toyota




VW, Audi,
Porsche



Honda




Others




Test Mode











High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
No. of
Vehicles
95
12
6


8
. 9
21

6
33
95




12




6




8




9




21




6




33




Parameter [
(percent)
Ec
1. 1
0.0
0.0


12.5
0.0
0.0

0.0
0.0
6.3
8.4
6.3
5. 3
5.3
8.3
8.3
8.3
0.0
8. 3
0.0
0.0
0.0
0.0
0.0
25.0
12:5
12. 5
12.5
12. 5
0.0
0.0
0.0
0.0
0.0
9.5
19.0
14. 3
14. 3
9.5
0.0
0.0
0.0
0.0
0.0
3.0
6. 1
3.0
3.0
3.0
Eo
13.7
25.0
33.3


0.0
22.2'
9.5

16.7
9. 1
38.9
36.8
27.4
35.8
28.4
50.0
58.3
0.0
16.7
16.7
66.7
66.7
50.0
50.0
50.0
37. 5
37. 5
37. 5
50.0
25.0
55.6
33.3
55.6
55.6
55.6
33. 3 '
19.0
14. 3
14. 3
14. 3
16.7
16.7
0.0
0.0
0.0
33.3
39.4
36.4
51. 5
36.4
FF
44^2
50; 0
33.3


62i 5
55.6
38. 1

0.0
48.5
18.9
21.0
30. 5
22. 1
29.5
25.0
16.7
75.0
58. 3
58. 3
0.0
0.0
16.7
16.. 7
16.7
25.0
25.0
25.0
12.5
37. 5
0.0
44. 4
22. 2
22. 2
22.2
14. 3
28.6
33.3
33. 3
33.3
0.0
0.0
16.7
16.7
16.7
24.2
6. 1
21,2
6. 1
21.2
STE(b)
0.763
0.667
0. 500


1.00
0.715
0.800

0
0.842
0. 327
0. 363
0. 527
0. 382
0.509
0. 333
0.223
1.00
0.777
0. 777
0
0
0.250
0.250
0.250
0. 400
0.400
0.400
0.200
0.600
0
0. 571
Q. 285
0.285
0.285
0.300
0.601
0. 700
0.700
0.700
0
0
1.00
1.00
1.00
0. 421
0. 134
0. 368
0. 106
0. 368
   '
CID = cubic inch displacement
Short Test Effectiveness =
                             % FF for short .test
   (c!
                    % FTP failures in same population
Best mode is: HC.and CO at idle in drive; NO  at high speed
                                      104

-------
  Table 111.  Comparison of HC Results for CID^* = 151  to  259,
      by Manufacturer —Predicted Population, Three Cities,
   Bounded Errors  of Commission Method (E   Set at 5  Percent)
                                                 c
Short Test
Federal
Short Cycle





Federal
Three -Mode


























Manufacturer
All
General Motors
Ford
Chrysler
American Motors
Others, ,
Others (c)
All



General Motors



Ford



Chrysler



American
Motors


Others



Others (c)



Test Mode







High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
54
13
13
7
12
7
. 9
54



13



13



7



12



7



9



Parameter
(percent)
Ec
5.0
1.8
6.0
1. 5
1.9
4.9
3.6
5.0
5.0
5.0
5.0
3.0
3.9
6. 5
8.4
6.9
6.4
1.0
0. 44
0.03
0.0
6.8
3.2
15.0
3.0
5.3
4. 1
6.2
16.0
3. 1
3. 1
6.6
10. 3
1.7
1.7
Eo
14.0
13. 1
5. 3
14.5
16.9
0,7
21. 1
23.6
21. 3
18.8
20.0
30.3
22.7
9:4
14.4
12. 3
11.6
15. 1
15.5
18.3
19.2
5.6
7.7
28.0
32. 5
27. 3
28.2
9.2
6.2
8. 3
9.2
26.6
24.2
31. 1
31. 5
FF
13. 3
18.6
10.8
4.6
21.0
9. 1
12.0
3.7
6.0
8.4
7.2
1.4
9.0
22. 3
17.3
3.8
4.5
1.0
1.0
1.0
0.02
13. 5
11.6
9.8
5.4
5.6
4.7
0.6
3.7
1.5
0.7
6.5
8.8
2.0
1.6
STE
0.487
0. 587
0.671
0.241
0. 554
0. 576
0.363
0. 136
0.220
. 0. 309
0.265
0.045
0.284
0.703
0.546
0.236
0.280
0.062
0.061
0.052
0.001
0.707
0.601
0.259
0, 143
0. 170
0. 143
0. 061
0. 374
0. 153
0.071
0. 196
0.267
0.060
0.048
(a),
(b).

(c)
CID = cubic inch displacement
'Short Test Effectiveness =
FF for short test
                    % FTP failures in same population
  Includes one Datsurr, one VW
                                    105

-------
  Table 112.   Comparison of CO Results for CID(a) =,151 to 259,
      by Manufacturer —Predicted Population,  Three Cities,
    Bounded Errors of Commission Method  (E  Set at 5 Percent)
Short Test
Federal
Short Cycle






Federal
Three -Mode













1












Manufacturer
All
General Motors
Ford
Chrysler
American
Motors
Others.
Others (c)
All



General Motors



Ford



Chrysler



American
Motors


Others



Others (C)



Test Mode








High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
54
13
13
7
12

7
9
54



13



13



7



12



7



9



Parameter
(percent)
Ec
5.0
0.5
8.7
2.6
4. 1

4.6
5. 5
5.0
5.0
5.0
5.0
0.04
1.0
7.4
9.0
6.2
6.1
1. 5
1.7
0.07
•0.27
15.3
12. 3
1.8
0.07
5.3
1. 1
1. 4
0.7
0.8
0.03
2.0
2.6
1.0
0. 1
Eo
14.6
20.4
6.4.
30.8
16.4

5.6
7.8
51.9
45.0
24.6
32.3
64.5
59. 4
14. 1
30.8
36.9
35. 5
36.2
41. 1
67.7
66.3
13.6
3. 1
44.6
50.8
26. 1
30. 3
44.2
19.3
24. 5
35.5
43.9
20.2
25.0
35.0
FF
41.0
44. 1
38.6
38.3
34.5

48.4
47.9
6.5
14.0
34. 8
27. 1
0.0
7. 4
50.4
33.6
16.4
17.7
.20.4
15. 5
0.38
K7
55. 5
68. 5
6.2
0. 12
24.8
20.6
9.8
34.7
29.5
18.4
11.8
35. 5
30.7
20.7
STE
-------
  Table  113.  Comparison of NOX Results for CID(a) = 151 to 259,
      by Manufacturer — Predicted Population,  Three Cities,
   Bounded Errors  of Commission Method (E   Set at 5 Percent)^
Short Test
Federal
Short Cycle






Federal
Three -Mode


























Manufacturer
All
General Motors
Ford
Chrysler
American
Motors
Others
Others
All



General Motors



Ford



Chrysler



American
Motors


Others



Others(c)



Test Mode








High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
54
13
13
7
12

7
9
54



13



13



7



12



7



9



Parameter
(percent)
Ec
5.0
0.5
6.2
5.0
13.6

0.6
0.6
5.0
5.0
5.0
5. 0
5.8
8.4
9.8
13. 1
2.9
1.5
0.03
0.0
7. 4
7.0
0. 12
0. 3
2.0
5.2
6.2
0.8
2.0
0.4
0. 1
0.0
4. 8
1.3
0.04
0.0
Eo
7.5
10.8
8.4
6.9
1.7

7. 5
7.6
15. 4
11.1
16.7
17. 5
20. 1
7.7
16. 7
15. 7
9.0
9.9
10.9
10.9
20.7
6.2
34.0
33.4
20.0
17.6
20. 1
26.7
9.0
10. 1
9.9
10. 5
7.8
10. 1
11. 3
11.5
FF
12.5
11.0
4.7
27.4
26.0

3.0
.3.9
4.6
8.9
3. 3
2.5
1.8
14.2
5. 1
6. 7
4. 1
3. 2
0.0
0.0
13. 5
28.0
0.2
1.0
7.7
8.2
7. 5
0.9
1. 5
0.4
0.6
0.0
3.7
1.4
0.2
0.0
STE
0.625
0. 505
0. 359
0.799
0.939

0.286
0. 339
0.230
0.445
0. 165
0. 125
0.082
0. 648
0.234 '
0. 299
0. 313
0. 244
0
0
0. 395
0.819
0. 006
0.029
0. 278
0. 318
0.272
0.033
0. 143
0.038
0.057
0
0. 322
0. 122
- 0.017
0
(a)
(b)
CID = cubic inch displacement

Short Test Effectiveness =
% FF for short test
                      % FTP failures in same population
(O
  Includes one Datsun, one VW
                                   107

-------
                                                                         (a) _
Table 114.   Comparison, of Multiple Constituents Results for CID
       259, by Manufacturer —Predicted Population,  Three  Cities,
       Bounded Errors of Commission Method (E   Set at 5  Percent)
                                                                       = 151
Short Test
Federal
Short Cycle






Federal
Three -Mode

































Manufacturer
All
General Motors
Ford
Chrysler
American
Motors
Others
Others'0'
All




General Motors




Ford




Chrysler




American
Motors



Othe rs




Others (c)




Test Mode








High speed
Low speed
Idle in drive
Idle in neutral
Best(d)
High speed
Low speed
Idle in drive
Idle in neutral
Best(d)
High speed
Low speed
Idle in drive
Idle in neutral
Best(d)
High speed
Low speed
Idle in drive
Idle in neutral
Best(d)
High speed
Low speed
Idle in drive
Idle in neutral
Best(d)
High speed
Low speed
Idle in drive
Idle in neutral
Best(d) •
High speed
Low speed
Idle in drive
Idle in neutral
Best(d)
No. of
Vehicles
54
13
13
7
12

7
9
54




13




13




7




12




7




9




Parameter
(percent)
Ec
5.6
0.0
0.0
0.0
8.3

28.6
22.2
1.8
3. 7
5.6
3.7
1.8
0.0
0.0
15.4
13. 1
7.7
0.0
7.7
0.0
0.0
0.0
0.0
0.0
14. 3
14. 3
14. 3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
11. 1
11. 1
0.0
0.0
0.0
Eo
13.0
15.4
0.0
57. 1
8.3

0.0
0.0
55.6
35.2
20. 1
35.2
22.2
61.5
23. 1
0.0
15.7
15.4
30.8
15.4
30.8
30.8
23. 1
71.4
57. 1
14. 3
14. 3
14.3
58.3
58. 3
41.7
66.7
41. 7
71.4
42.9
42.9
57. 1
42.9
66.7
33.3
33.3
44.4
33.3
FF
55.6
53.9
46.2
28.6
75.0

71.43
66.7
13.0
33.3
44.4
33. 3
46. 3
7.7
46. 1
69.2
6.2
53.8
15.4
30. 8
15.4
15. 4
23. 1
14.3
28.6
71.4
71.4
71.4
25.0
25.0
41.7
16.7
41.7
0.0
28.6
28.6
14.3
28.6
0.0
33.3
33. 3
22.2
33.3
STE
0.812
0.779
1.00
0.334
0.900

1.00
1.00
0. 190
0.486
0.648
0.486
0.676
0. Ill
0.666
1.00
0. 283
0. 777
0. 333
0.667
0. 333
0. 333
0. 500
0. 167
0.334
0. 833
0..833
0.833
0.300
0. 3*00
0. 500
0.200
0. 500
0
0.400
0. 400
0.200
0. 400
0
0. 500
0. 500
0. 333
0. 500
    (a)
    (b)
CID = cubic inch displacement
Short Test Effectiveness =
% FF for short test
                          % FTP failures in same population
(c)
(d)
Includes one Datsun, one VW
Bost mode is: HC and CO at
idle in drive; NOx at high speed
                                        108

-------
Table 115.  Comparison of HC Results for CID    = 260 and Greater,
        by Manufacturer — Predicted Population, Three Cities,
    Bounded Errors  of Commission Method (JH  Set at 5  Percent)
Short Test
Federal
Short Cycle




Federal
Three-Mode


















Manufacturer
All
General Motors
Ford
Chrysler
American
Motors.'
All



General Motors



Ford



Chrysler



American
Motors


Test Mode






High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speedv
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
151
74
42
29
6

151



74



42



29



• 6



Parameter
(percent)
Ec
5.0
4.0
2.8
4.0
0.0

5.0
5.0
5.0
5.0
4.8
4.7
4.0
4.2
5. 1
5. 2
3.9
5. 1
4.2
4. 3
5.9
6.2
0.0 '
0.0
0.0
0.0
Eo
12.5
8.9
1.1 . 4
20. 3
0.0

27.0
24.7
18. 3
19.6
24.7
22. 1
14. 4
15.9
26.0
23.8
20.6
21.2
41.8
39.7
29.2
31.1
0.0 '
0.0
0.0
0.0
FF
20.4
21.8
18.0
40. 5
0.0

5.6
8.0
14.6
13. 3
6.0
8.6
16.3
14.8
3.4
5.7
8.8
8.3
8.7
10. 8
31.6
29.7
0.0
0.0
0.0
0.0
STE
0.620
0.710
0.612
0.666'
--

0. 172
0.245
0. 444
0. 404
0. 195
0.280
0. 531
0.482
0. 116
0. 193
0.299
0. 281
0. 172
0.214
0. 520
0. 488

--
--

 (b)
  CID = cubic inch displacement
                            % FF for short test
  l_.   . „   „,, ..       	70 r J lor snuri icai	
  Short Test Effectiveness - ~ r-**-.-, c—~,	:	;—r:—
                      % FTP failures in same population
                                    109

-------
Table 116.  Comparison of CO Results for CILVa' = 260 and Greater,
        by Manufacturer — Predicted Population, Three Cities,
    Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Test
Federal
Short Cycle




Federal
Three -Mode


















Manufacturer
All
General Motors
Ford
Chrysler
American
Motors
All



General Motors



Ford



Chrysler



American
Motors


Test Mode






High speed
Low speed •
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
151
74
42
29
6

151



74



42



29



6



Parameter
(percent)
Ec
5.0
5.2
10. 1
2.0
0.4

5.0
5.0
5.0
5.0
3.4
3.6
3.0
3. 5
7.7
8.6
13.0
10.2
3.8
1.4
8.6
8.9
0.0
0.2
0.2
0:4
Eo
13.0
12.2
8.4
6.6
0.0

41.4
38.0
17.8
18.8
40. 7
34.7
20. 1
21.2
35.2
22.9
9.4
11.6
58.6
70.2
5. 2
5. 5
0.06
0.0
0.0
0.0
FF
44.2
38. 1
52.0
76.4
0.0

17.4
21.9
42. 1
41.3
12. 3
18. 5
34. 4
33. 3
25. 1
37.5
52. 1
49.8
24.6
14.4
77.8
77.4
0.0
0.0
0.0
0.0
STE(b>
0.773
0.758
0.861
0.920
--

0.296
0. 366
0.703
0.687
0.232
0. 348
0.631
0.611
0.416
0.621
0.847
0. 81 1
0.296
0. 170
0.937
0. 934
	
--
--

 (b)
CID = cubic inch displacement

Short Test Effectiveness =
                           % FF for short test
                      % FTP failures in same population
                                    110

-------
Table 117.  Comparison of NOX Results for CID^a'  = 260 and Greater,
        by Manufacturer — Predicted Population, Three Cities,
     Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Teat
Federal
Short Cycle




Federal
Three -Mode


















Manufacturer
All
General Motors
Ford
Chrysler
Ame rican
Motors
All



General Motors



Ford



Chrysler



Ame rican
Motors


Test Mode






High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
151
74
42
29
6

151



74



42



29



6



Parameter
(percent)
Ec
5.0
1. 1
0.6
20.9
9.0

5.0
5.0
5.0
5.0
4.0
1.0
6. 5
4.3
3.0
0.2
0.. 01
0.2
12.2
28.2
4.0
4.7
0.0
2.8
0.6
0. I
Eo
11.5
10.9
23.5
2. 1
2.4

16.3
19.4
21. 5
20.8
15. 3
17.6
16.3
17.2 .
25.4
35.4
45. 3
45.0
6. 3
2. 1
11.4
10.8
34.9
33. 4
30.6
38.6
FF
13.3
10.7
14.7
10.0
38.9

8. 5
5. 4
3.2
4.0
6.2
3.9
5. 3
4. 3
12.8
2.8
0.0
0.2
5.8
10. 0
0. 7
1.3
3.3
4.8
7.6
2.7
STE(b)
0. .536
0.495
0.385
0.827
0.942

0. 343
0.218
0. 130
0. 161
0.288
0. 181
0. 245
0.200
0. 335
0.073
0
0.004
0.479
0.826
0.058
0. 107
0.086
0. 126
0. 199
0.065
  (a)
  (b)
CID = cubic inch displacement
Short Test Effectiveness =
                              FF for short test
                            failures in same population
                                    ill

-------
Table 118.  Comparison of Multiple Constituent Results for CID^a' = 260
  or Greater, by Manufacturer —Predicted Population, Three Cities,
      Bounded Errors of Commission Method (E   Set at 5  Percent)
Short Test
Federal
Short Cycle




Federal
Three-Mode























Manufacturer
All
General Motors
Ford
Chrysler
American
Motors
All




General Motors




Ford




Chrysler




American
Motors



Test Mode






High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)'
High speed
Low speed
Idle in drive
Idle in neutral
Best(c) .
No. of
Vehicles
151
74
42
29
6

151




74




42




29




6




Parameter
(percent)
Ec
8.0
4. 1
14.3
3. 5
33.3

2.6
2.6
6.0
5. 3
4.0
0.0
1.3
4.0
1.3
0.0
7. 1
7. 1
11.9
14. 3
11.9
0.0
0.0
3. 4
3.4
3.4
0.0
0.0
0.0
0.0
0.0
Eo
6.6
4. 1
7. 1
13.8
0.0

24. 5
33.8
19.2
17.9
19.9
43.2
39.2
17.6
17.6
23.0
35.7
23.8
19.0
16.7
16.7
31.0
34. 5
20.7
17.2
13. 8
33. 3
33. 3
33.3
33. 3
33. 3
FF
62.3
59.5
64. 3
72.4
33. 3

10.0
2.6
49.7
51.0
49.0
20. 3
24. 3
46. 0
46.0
40. 5
35.7
47.6
52. 4
54. 8
54. 8
55.2
51.7
65. 5
69.0
72. -4
0.0
0.0
0.0
0.0
0.0
STE
0.904
0.937
0.901
0.840
1.00

0.290
0.071
0. 721
0. 740
0. 71 1
0. 320
0. 383
0. 723
0.723
0.638
0. 500
0.667
0.734
0.766
0.766
0.640
01600
0.760
0.800
0.840
0
0
0
0
0
   (b)
CID = cubic inch displacement

Short Test Effectiveness =
                                                     (c)
                              % FF for short test
                         %FTP failures in same population
Best mode- is: HC.and CO at
idle in drive; NOX at hi^h spe
                                      112

-------
         Table 119.  Comparison of Federal Short Cycle Cut-Points
               by Engine Displacement (E  Set at 5 Percent)

Engine
Displacement
Group
All
150 CID and
Less
151 to 259 CID
260 CID and
Greater

No. of
Vehicles
300
95

54
151

ST Cut-Points and Standard Errors
(gm/mi)
HC
CP
1.32
1.28

1.27
1.35

SE
-------
        Table 120.  Comparison of Federal Three-Mode Cut-Points


              by Engine Displacement^' (E  Set at 5 Percent)
Engine
Displacement
Group
All Three
Cities
150 CID and
Less
151 to 259 CID
260 CID and
Greater
No. of
Vehicles
300
95
54
151
ST Cut-Points and Standard Errors
(ppm)
HC(b)
Cp(d)
184
219
213
151
SE
26
57
79
26
co(b>
CP
7,081
18, 382
7, 063
2,655
SE
1, 251
5, 863
2,488
607
NO (c)
X
CP
2,265
2, 805
l,985(f)
1,986
SE
243
665
446(f)
252
(al
  'Engine cubic inch displacement = CID

(b)


(c)
Tested at idle in drive
   Tested at high speed


(d)CP = cut point

(e)
v  'SE = standard error


'f'Best NOX mode is at low speed.  CP and SE for the low speed NOX

   are  1396 ppm and 423 ppm,  respectively
                                    114

-------
Table 121.  Comparison of HC Results for CID    = 150  and Less,
          by City — Predicted Population,  Three Cities,
   Bounded Errors of Commission Method (E  Set  at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode














City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix



Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
95
32
33
30
95



32



33



30



Parameter
(percent)
Ec
5.0
4. 8
4.4
5.3
5.0
5.0
5.0
5.0
9.9
6.8
9.5
10.0
1.2
1.7
2.0
2.4
4.7
6.3
3. 1
2.4
Eo
13.0
15.7
12.9
8.6
25.9
24. I
22.0
23. 5
20.7
18.9
18. 1
19.6
35.9
32. 5
26.7
29.0
19.3
18.0
19.2
20. 1
FF
19. 1
16.3
27.0
14. 1
6.2
8. 1
10.2
8.7
11.4
13. 1
13.9
12. 5
4.'0
7.4
13.2
10.8
3. 3
4-. 6
3.4
2. 5
STE(b)
0. 595
0. 509
0.677
0.621
0. 193
0.252
0.317
0.270
0.355
0.409
0.434
0.389
0. 100
0. 185
0.331
0. 271
0. 146
0.204
0. 150
0.111
(a)
(b)
  CID = cubic inch displacement
  Short Test Effectiveness =
                           % FF for short test
                      % FTP failures in same population
                                  115

-------
Table 122.  Comparison of CO Results for CID^ =  150 and Less,
           by City — Predicted Population,  Three Cities,
   Bounded Errors of Commission Method (E  Set at 5  Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode














City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix



Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
L*ow. speed
Idle in drive
Idle in neutral
No. of
Vehicles
95
32
.33
30
95



32



33



30



Parameter '
(percent)
Ec
5.0
3.0
4.7
7.2
5.0
5.0
5.0
5.0
5.0
7.3
7.7
6. 1
5.4
4.7
5.4
6.6
3. 1
2. 5
2.7
2.4
Eo
19.4
27.7
20.6
13.3
38.0
34.7
26.4
28.8
41.7
40.0
29.4
31. 1
44.7
38.3
31.0
33.8
29. 1
29. 1
21.0
23.7
FF
25.4
23.8
30. 1
20.3
6.7
10.0
18.4
16.0
9.9
11.5
22. 2
20. 5
6.0
12.4
20.0
17.2
4. 5
4. 5
12.6
9.9
STE
0.567
0.462
0. 594
0.604
0. 150
0.224
0.411
0.357
0. 192
0.223
0.430
0.397
0. 118
0.245
0.392
0. 337
0. 134
0. 134
0.375
0.295
(b)
1 CID = cubic inch displacement

 Short Test Effectiveness =
                           % FF for short test
                      % FTP failures in same population
                                  116

-------
Table 123.  Comparison of NOX Results  for CID    =  150 and Less,
           by City — Predicted Population,  Three Cities,
   Bounded Errors of Commission Method  (E  Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode














City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix



Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in. drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
95
32
33
30
95



32



33



30



Parameter
(percent)
Ec
5.0
2.9
3.6
7.4
5. 0
5.0
5.0
5.0
0.7
0.9
0. 1
2. 1
3.8
5. 0
3.0
1. 3
4.6
7. 7
8.4
10.4
Eo
5.8
5.0
5. 5
5.9
9.0
9.0
12.6
13.2
15.8
15.4
16.5
15.9
8.8
7.9
9.7
12.0
4. 8
2. 8
11.8
12. 1
FF
9.8
12.3
10. 5
7.5
6. 5
5.2
2.9
1.0
0. 8
1. 1
0.03
0. 7
7. 1
5.2
6.2
1 . 0
8.8
10.6
1.6
1. 3
STE
0.628
0.711
0.656
0.559
0.419
0. 366
0. 187
0.070
0.048
0.067
0.002
0.042
0.447
0. 397
0. 390
0.077
0.647
0.791
0. 119-
0.097
'  CID = cubic inch'displacement
(b)        „..   .            % FF for short test	
  'Short Test Effectiveness = % FTp failures Jn same pOpuiation
                                   117

-------
Table-124.  Comparison of Multiple Constituent Results for CIDV  ' = 150
         and Less,  by City — Predicted Population,  Three Cities,
       Bounded Errors of Commission  Method (E   Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three -Mode


















City
All
Chicago
Houston
Phoenix
All




Chicago




Houston




Phoenix




Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
-------
 Table 125.  Comparison of HC Results for CID   =  151 to 259,
          by City — Predicted Population, Three Cities,
  Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode















City
All
Chicago
Houston
Phoenix
All


Chicago



Houston



Phoenix




Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral

No. of
Vehicles
54
19
16
19
54


19



16



19




Parameter
(percent)
Ec
5.0
8.3
3.8
5.0
5.0
5.0
5.0
5.0
8.8
6.3
6.6
5.8
0.0
1.0
2.8
2.8
6.3
10. 1
5.7
5.9

Eo
14. 0
4.4
11.6
20. 1
23.6
21.3
18.8
20.0
14. 1
15.2
10.0
11.8
37.3
25.9
19.4
21.5
25.0
21. 5
25.5
26.0

FF
13.3
11.0
15.3
15. 1
3.7
6.0
8.4
7.2
3.5
2.4
7.6
5.8
0.0
1.1
7.6
5.5
10.2
13.7
9.7
9.2

STE
0.487
0.714
0. 569
0.429
0. 136
0.220
0.309
0.265
0. 199
0. 136
0.432
0.330
0
0.041
0.281
0.204
0.290
0. 389
0.276
0.261

(a)

(b)
CID = cubic inch displacement

Short Test Effectiveness =
% FF for short test
                      % FTP failures in same population
                                  119

-------
 Table 126.  Comparison of CO Results for CID    = 151 to 259,
          by City— Predicted Population,  Three Cities,
  Bounded Errors of Commission Method (E^  Set at 5  Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode















City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix




Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral

No. of
Vehicles
54
19
16
19
54



19



16



19




Parameter
(percent)
Ec
5.0
3. 1
6.6
4.0
5.0
5.0
5.0
5.0
2.4
3.7
7.9
7.4
8.2
6.2
6.0
6.0
3.8
5.0
1.7
1.3

Eo
14.6
14.0
16.0
12.7
51.9
45.0
24.7
32.4
47.3
40.6
18.4
24.0
57.9
48.8
30.3
41.0
50.4
44.8
24.2
30.0

FF
41.0
40. 5
34.9
48. 1
6.5 .
14. 1
34.8
27. 1
7.2
13.9
38. S
32.9
1. 5
12.8
29.7
19.0
10.4
16.0
36.6
30.8

STE(b)
0.737
0.743
0.686
0.791
0.111
0.239
0. 585
0.456
0. 132
0.255
0.677
0. 578
0.025
0.208
0.495
0. 317
0. 171
0.263
0.602
0. 507

(a)

(b)
CID = cubic inch displacement
                         % FF for short test
    „.   „,,   .            7o FF for short test
Short Test Effectiveness = % FTP failures in same population
                                  120

-------
 Table  127.  Comparison of NOX Results for CID*^ = 151  to 259,
          by  City — Predicted Population,  Three Cities,
  Bounded Errors of Commission Method  (E  Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode














City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix



Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
54
19
16
19
54



19



16



19



Parameter
(percent)
Ec
5.0
3.4
10.3
2.4
5.0
5.0
5.0
5.0
0.8
1.4
8. 2
12. 1
1.6
6.9
0.6
0.4
5.2
6.9
2.7
0.3
Eo
7.5
6.8
10.8
5. 5
15.4
11.1
16.7
17.5
13.3
11.4
11.5
11.3
18.8
11.1
19.2
19.7
11.7
9.2
22.9
25.0
FF
12.5
7. 1
9.4
19.8
4.6
8.9
3.3
2. 5
0.7
2.6
2. 5
2.7
1.3
9.3
0.9
0.4
13.6
16. 1
2.4
0.3
ST£(b)
0.625
0. 511
0.465
0.783
0.230
0.445
0. 165
0. 125
0.050
0. 186
0. 179
0. 193
0.065
0.456
0.045
0.020
0.538
0.636
0.095
0.012
(a)
(b)
CID = cubic inch displacement
Short Test Effectiveness =
FF for short test
                      % FTP failures in same population
                                  121

-------
Table 128.   Comparison of-Multiple Constituent Results for  CILr'  -  151 to 259,
                   by City — Predicted Population,  Three Cities,
           Bounded Errors of Commission Method (E  Set  at  5  Percent)
Short Test
Federal
Short Cycle
•

•Federal
Three -Mode


















City
All
Chicago
Houston
Phoenix
AH




Chicago




Houston




Phoenix




Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best
0.810
0.834
0.727
0.858
0. 190
0.486
0.648
0.486
0.676
0. 166
0.417
0.834
0. 583
0.750
0
0.363
0.637
0. 363
0.727
0. 357
0.643
0. 500
0. 500
0. 571
        (a)
        (b)
        (c)
CID = cubic inch displacement

Short Test Effectiveness =
% FF for,short test.
                     , FTP failures in same population
          Best mode is:  HC and CO at idle in drive; NO at high speed
                                           122

-------
Table 129.  Comparison of HC Results for CLEr' = 260  and Greater,
            by City — Predicted Population,  Three Cities,
    Bounded Errors of Commission Method  (E  Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode














City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix



Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
151
49
51
51
151



49



51



51



Parameter
(percent)
Ec
5.0
3.7 '
3.8
7.4 •
5.0
5.0
5.0
5.0
6.8
6.6
7. 1
6.4 .
1. 7
1.8
2.4
4. 3
6.8
6.8
4.0
2.4
Eo
12.5
12.2
10.2
14.0
27.0
24.7
18.3
19.6
22.3
19.9
18. 1
19.8
33. 1
30.7
21.9
21.2
26. 1
24.4
13. 1
16.9
FF
20.4
17.0
27.9
16.6
5.7
8.0
14.6
13.3
6.2
8.6
11.0
9.3
S.O
7.4
16.2
16. .9
4. 5
6.2
17^6
13.7
STE(b)
0.620
0.582
0.732
0. 542
0. 174
0.245
0.444
0.404
0.218
0.302
0.378
0.320
0. 131
0. 194
0.425
0.444
0. 147
0.203
0. 573
0.448
   CID = cubic inch displacement
  (b)c
   Short Test Effectiveness =
% FF for short test
                       '% FTP failures in same population
                                    123

-------
Table 130.  Comparison of CO Results  for CLLr' = 260 and Greater,
            by City — Predicted Population, Three Cities,
    Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode














City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix



Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
No. of
Vehicles
151
49
51
51
151



49



51



51



Parameter
(percent)
Ec
5.0
5.2
4. 1
5.0
5.0
5.0
5.0
5.0
7.4
6.3
5.8
6. 1
1. 7
3. 1
3.8
3.0
5.6
4.6
5.9
6.4
Eo
13.0
15.0
9.3
14.8
41.4
38.0
17.8
18.8
33.2
32.0
21.4
22.9
47.8
42.9
16.8
17.8
43.0
38.8
16.5
17.0
FF
44.2
44.5
45.9
43.0
17.4
20.9
42. 1
41.3
26.3
27.4
38. 1
36.5
11.5
16.6
45.6
45.3
14.8
19.0
41.4
40.9
STE
0.773
0.748
0.832
0.744
0.296
0.355
0.703
0.687
0.442
0.461
0.640
0.615
0. 194
0.279
0.731
0.718
0. 256
0. 329
0.715
0.706
    CID = cubic inch displacement
  (b).,,   _   .-,,.  ..       	% FF for short test
   'Short Test Effectless = % FTP /ailures in same pop.llation
                                    124

-------
Table 131.  Comparison of NOX Results  for CID^ = 260 and Greater,
             by City — Predicted Population,  Thr.ee Cities,
     Bounded Errors of Commission Method (E  Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode










r




City
All
Chicago
Houston
Phoenix
All



Chicago



Houston



Phoenix




Test Mode




High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in. drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral
High speed
Low speed
Idle in drive
Idle in neutral

No. of
Vehicles .
151
49
• 51
51
151



49



51



51




Parameter
(percent)
Ec
5.0
5.2
4.. 2
5.7
5. 0
5. 0
5.0
5. 0
.6
3.2
4.4
3.2
2.8
3.7
4. 3
5.9
8.8
8.2
5. 2
3. 5
t
Eo
11.5
9.5
16.3
8.8
16.3
19.4
21.5
20.8
24.4
19.9
20.0
20.2
22.3
24.5
24.7
23. 1
10.0
13.6
20.3
19.7

FF
13.3
13.8
13.0
12.9
8. 5
5.4
3.2
4.0
2. 2
3.4
3.2
3.0
7. 1
4.9
4.7
6.2
11.7
8. 1
1.4
2.0

STE(b)
0. 536
0. 592
0.44>4
0. 594
0. 342
. 0.218
0. 130
0. 161
0.083
0. 146
0. 138
0. 129
0. 241
0: 167
0. 160
0. 212
' 0.539
0.373
0.065
0.092
•
    CID = cubic incn displacement
  (b)
    Short Test Effectiveness =
                             % FF for short test
                        % FTP failures in samo population
                                    125

-------
Table 132.  Comparison of Multiple Constituent Results for CID^a' = 260 and Greater,
                     by City — Predicted Population,  Three Cities,
             Bounded Errors of Commission Method (E   Set at 5 Percent)
Short Test
Federal
Short Cycle


Federal
Three-Mode














City
All
Chicago
Houston
Phoenix
All


Chicago



Houston



Phoenix



Test Mode



High speed
Low speed
Idle in drive
Idle in neutral
Best
High speed
Low speed
Idle in drive
Idle in neutral
Best'c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c)
High speed
Low speed
Idle in drive
Idle in neutral
Best(c>
No. of
Vehicles
151
49
51
51
151


49



51



51



Parameter
(percent)
Ec
8.0
10. 2
7.8
5.9
2.0
2.7
6.0
5. 3
4.0
4. 1
4. 1
4. 1
6. 1
4. 1
0.0
2. 0
7.8.
3.9
3.9
2.0
2.0
5.9
5.9
3.9
Eo
6.6
10.2
3.9
5.9
38.4
33.8
19.2
17.9
19.9
38.8
34.7
28.6
28.6
34.7
39.2
37.3
9.8
7.8
11.8
37.3
29.4
19.6
17.7
13.7
FF
62.3
63.3
60.8
62.8
30. 5
35. 1
49.7
51.0
49.0
34.7
38.8
44.9
44.9
38.8
25. 5
27. 5
54.9
56.9
52.9
31.4
39.2
49.0
51.0
54.9
STE                    % FTP failures in same population
            Best mode is:  HC and CO at idle in drive;  NO at high speed
                                             126

-------
             Table  133.  Contingency Table
                                True = FTP
                      Pass
                      Fail
                         Total
   D  O
   h X
    co
          Pass
                                     a + b
          Fail
                                     c + d
Total
a + c
b + d
n = a + b
 + c + d
      a = number of correctly passed vehicles (PP)

      b = number of errors of omission (E  )
                                       v  o'
      c = number of errors of commission (E  )
                                            c
      d = number of correctly failed vehicles (FF)

Sensitivity = a/(a + c)
Specificity = b/(b + d)
False positive error = b/(a + b)
False negative error = c/(c +  d)
                               ad - be
Correlation index =
                    [(a + b)(a + c)(b + d)(c + d)]
                                              1/2
                           127

-------
  Table 134.  Maintenance Options and Adjustment Procedures
           for Each Version of the Maintenance Model

Options and Procedures
HC, CO, NO Adjusted Independently
X
NO Adjustment Independent
KC and CO Adjustment Coupled
Emission Values Adjusted to
FTP Standard
Emission Values Adjusted to
Equivalent ST Standard
Version of
Maintenance Model
1

X

X

2
X


X

3
X



X
        Table 135.  Average Deterioration Rates Used in
                 207(b) Effectiveness Simulation
Data Source
Certification^'
EFP
-------
              Table  136.  Annual Mileage Accumulations Used in
                        207(b) Effectiveness Simulation
Age of
Vehicles
In Years
1
2
3
4(b)
Number of
Years In
Program
0
1
2
3
-------
       Table 138.   207(b) ST(a) Cut-Points for Each Engine Family
CID(b)
150 and Less
151 to 259
260 and G.reater
Cut-Points ' in PPM
HC
219
213
151
CO
18, 332
7, 062
2, 655
NOX
2, 805
1, 395
1,986
(a)
(b)
(c)
The ST was defined from the Federal Three-Mode with HC and CO
idle in drive,  and NOx in high speed (low speed NOX was used for
151 to  259 CID group)

CID - cubic inch displacement

Cut-points  were computed  from 300 car data using a 5 percent E
rate on each pollutant for each engine class
                                  130

-------
               Table 139.  207(b) Program Efficiencies with
                           Maintenance  Version 1
Maintenance
Period of
Effectiveness
12


9


6


Number of
Years
Since The
1 st Inspection
1
2
3 U)
1
2
3 (a)
1
2
3 (a)
Estimated Program
Efficiencies in Percent
Cumulative
HC
22. 7
31.6
34. 6
19.5
26. 7
30. 0
14. 6
19. 8
23. 3
CO
32. 4
39. 0
42. 5
26.2
31.9
36. 1
18. 7
23.0
27. 6
NOX
2.05
1. 81
2. 00
1.25
1. 04
1.29
0.46
0.22
0. 50
(a)
   Not a full year.  Program ended at 50, 000 miles
                                   131

-------
                Table 140.  207(b) Program Efficiencies with
                           Maintenance Version 2
Maintenance
Period of
Effectiveness
12


9


6


Number of
Years
Since The
1st Inspection
1
2
3(a)
1
2
3(a)
1
2
3(a)
Estimated Program
Efficiencies in Percent
Cumulative
HC
22. 7
31. 6
34. 6
19.5
26. 7
30. 0
14.6
19. 8
23. 3
CO
29. 7
35.9
39. 1
24.2
29.4
33.0
17. 3
21.2
25. 1
NOX
2.05
1.81
2. 00
1.25
1. 04
1.29
0.46
0.22
0. 50
(a)
   Not a full year.  Program ended at 50, 000 miles
                                    132

-------
                Table 141.  207(b) Program Efficiencies -with
                           Maintenance Version 3
Maintenance
Period of
Effectivenes s
12


9


6


Number of
Years
Since The
1st Inspection
1
2
3(a)
1
2
3 (a)
1
2 '
3(a)
Estimated Program
Efficiencies in Percent
Cumulative
HC
21. 1
30.0
32.9
18. 4
25. 4
28. 5
13. 8
18. 9 :
22. 1
CO
29. 0
35. 2
38. 3
23. 6
28. 9
32. 3
17. 0
20. 8
24. 6
NOX
1.78
1. 53
1.70
1.05
0.82
1.05
0. 32
0.08
0. 33
(a)
  Not a full year.  Program ended at 50, 000 miles
                                    133

-------
        Table 142.  FTP Values Equivalent to 207(b) ST Cut-Points
CID
150 and Less
151 to 259
260 and Greater
Equivalent ST Cut-Points - Grams /Mile
HC
1. 574
1.633
1.809"
CO
18.084
16.66 •
15. 0
NOX
3. 100
3. 100
( 3.262
      Table  143.  I/M
                               Cut-Points for Each Engine Family
CID(b)
150 and Less
151 to 259
260 and Greater
Cut-Points^ in PPM
HC
269
269
269
CO
17105
17105
17105
NOX
2823
2823
2823
(a)
(b)
(c)
The ST was defined from the Federal Three-Mode with HC and  CO
idle in drive,  and NOX at high  speed

CID = cubic inch displacement

Cut-points were computed from the 300 car data  set using a
33. 33 percent total failure rate on 300 cars
                                   134

-------
                 Table 144.  I/M Program Efficiencies with
                           Maintenance Version 1
Maintenance
Period of
Effectiveness
12


9


6


Number of
Years
Since The
1 st Inspection
1
2
3(a)
1
2
3 1 a)
1
2
3(a)
Estimated Program
Efficiencies in Percent
Cumulative
HC
16.4
24. 0
27.0
14. 1
20. 1
23.2
10. 5
14. 8
18. 1
CO
24. 3
30. 3
33. 8
19.5
24. 5
28. 5
13.8
17.6
21. 8
NOX
1.59
1.58
1.78
1.04
1.03
1.28
0.50
0.45
0.73
(a).
  Not a full year.  Program ended at 50, 000 miles
                                    135

-------
                Table 145.   I/M Program Efficiencies with
                          Maintenance Version 2
Maintenance
Period of
Effectiveness
12


9


6


Number
Years
Since The
1 st Inspection
1
2
3(a)
1
2
3(a)
1
2
3 (a)
Estimated Program
Efficiencies in Percent
Cumulative
HC
16.4
24. 0
27.0
14. 1
20. 1
23.2
10. 5
14. 8
18. 1
CO
22.2
27.7
30.8
18.0
22.5
25.9
12. 8
16.2
19.7
NOX
1.59
1.58
1.78
1.04
1.03
1.28
0. 50
0.45
0.73
(a)
  Not a full year.  Program ended at 50, 000 miles
                                   136

-------
               Table  146.  I/M Program Efficiencies \yith
                         Maintenance Version 3
Maintenance
Period of
Effectiveness
12


9


6


Number of
Years
Since The
1st Inspection
1
2
3(a)
1
2
3(a)
1
2
3(a)
Estimated Program
Efficiencies in Percent
Cumulative
HC
14.3
21.2
23. 7
12. 5
17.9
20. 3
9.4
13. 3
15.7
CO
18.4
23.6
26. 1
15. 1
19.3
22.0
10.9
14.0
16.8
NOX
1.04
0.99
1. 13
0.63
0.58
0.76
0.23
0. 15
0.34
(a)
   Not a full year.  Program ended at 50, 000 miles
                                  137

-------
Table 147.  FTP Values Equivalent to
         I/M ST Cut-Points
CID
150 and Less
151 to 259
260 and Greater
Equivalent ST Cut-Points - Grams/Mile
HC
1. 71
1. 84
2. 69
CO
17.66
21. 82
32. 51
NOX
3. 10
3.79
4. 06
                 138

-------
  o
O
OS
50

40

30

20

10

 0
                                          	Eo
                                          	FF
 0.8   1.0   1.2   1.4   1.6   1.8   2.0   2.2
                  HC CUT-POINT, gm/mi
                                                 2.4   2.6
Figure 1.  Variation of Ec,  Eo, and FF with HC Cut-point;
     All Three Cities; Federal Short Cycle; Bounded
              Errors of Commission Method
       o
     o
     50

     40

     30

     20

     10

                4.0   8.0   12.0  16    20
                   CO CUT-POINT, gm/mi
                                        24    26
Figure 2.  Variation of Ec,  Eo, and FF with CO Cut-point;
      All Three Cities; Federal Short Cycle; Bounded
              Errors of .Commission Method
                           139

-------
I 30
c
_°20
 "o
- 10
o
Q_
                                          	Eo
                                          _______  cp
    0
     2.0    2.2   2.4   2.6   2.8   3.0   3.2    3.4   3.6   3.8
                     IMOY CUT-POINT, gm/mi
                       A
 Figure 3.  Variation of Ec, Eo,  and FF with NOX Cut-point;
      All Three Cities; Federal  Short Cycle; Bounded
              Errors of Commission Method
                                                  Ec
                                                  Eo
                                                  FF
                     1.4    1.6.   1.8   2.0
                     HC CUT-POINT, gm/mi
                                           2.2    2.4   2.6
 Figure 4.  Variation of Ec,  EQJl and FF with HC Cut-point;
       Chicago, Federal Short Cycle, Bounded Errors
                  of Commission Method
                            140

-------
        o
50

40

30

20

10
                                           C

                                      -Eo
                                      — FF
                                                I
             0   4.0   8.0   12.Q  16.0  20.Q 24.0  28.0
                      CO CUT-POINT,  gm/mi
Figure 5.  Variation of Ec, Eo,  and FF with CO Cut-point;
      Chicago; Federal Short Cycle; Bounded Errors
                  of Commission Method
                                                   Ec
                                                    0
                                                   FF
                                        3.2   3.4    3.6   3.8
                      NO  CUT-POINT,  gm/mi
                        A
Figure 6.  Variation of Ec, Eo, and FF with NOX Cut-point;
      Chicago; Federal Short Cycle; Bounded Errors
                  of .Commission Method
                            141

-------
o_
40

30

20

10

 0

     0.8   1.0   1.2   1.4   1.6   1.8   2.0  2.2   2.4   2.6
                     HC CUT-POINT, gm/mi

Figure  7.  Variation of Ec, Eo,  and FF with HC Cut-point;
      Houston; Federal Short Cycle; Bounded Errors
                  of Commission Method
                                   --- FF
                4.0   8.0  12.0  16.0  20.0   24.0  28.0
                     CO CUT-POINT, gm/mi
Figure 8.  Variation of Ec, EQ, and FF with CO Cut-point;
      Houston; Federal Short Cycle; Bounded Errors
                  of Commission Method
                            142

-------
                                                   0
                                                  FF
                                                  3.6    3.8
                     NO  CUT-POINT, gm/mi
                        A
 Figure 9.   Variation of Ec,  Eo,  and FF with NOX Cut-point;
       Houston; Federal Short Cycle; Bounded Errors
                  of Commission Method
  o
    30 r-
    20
    10
o
LU
Q_
                                          	FF
0
0.8   1.0
1.2   1.4   1.6   1.8   2.0   2.2
     HC CUT-POINT, gm/mi
                                                   2.4   2.6
 Figure 10.  Variation of Ec,  Eo, and FF with HC Cut-point;
       Phoenix; Federal Short Cycle; Bounded Errors
                   of Commission Method
                            143

-------
       Q
       - and FF with NOX Cut-point;
      Phoenix; Federal Short Cycle; Bounded Errors
                  of Commission Method
                            144

-------
           15
           10
        o  '
              oo
                                           H
                                               10
                                                                    201-
                                                             15
                                                                  o
                                                             10
                                      0   o
                                                                                                        o
                                                                                               0   O
                     o
                                                                                                                 10
in
           60r
         °40
o

t-Lj
Q_
           20
2468
      PREDICTED  PERCENT E
                                                       10
        Figure 13.  Variation of Actual Ec,  E , and  FF
           with Predicted Ec;  Federal Short Cycle;
                       All Three Cities
                                                                 O
                                                                 C£
                                                                 LU
                                                                 Q-
                                                            60
                                                           '40
                                                            20
                                                                                                    FF
— /
                                                                      2        4        68
                                                                          PREDICTED  PERCENT E.
                                          10
                                                         Figure 14.  Variation of Actual Ec,  EQ, and FF
                                                            with Predicted Ec;  Federal Short Cycle;
                                                                            Chicago

-------
   15
- 10
LlJ
CJ
OL
I  5
o
0
                                               0
                                               10
O
o;
   60
  '40
   20
                           F£-^-—
   4        6       8
PREDICTED PERCENT E.
                                              10
Figure 15.   Variation of Actual Ec,  EQ, and FF
    with Predicted Ec; Federal Short Cycle;
                    Houston
                          10
                       o
                       <
                                                             Ofroo-
                                                                                                o
                                                                                   o
                                                                      10
                                                        
-------
  O

  "b
O

LU
Q_
                                           HIGH SPEED
                                           LOW  SPEED
                                           IDLE IN DRIVE
                                           IDLE IN NEUTRAL
         0     100   200   300   400   500    600  70,0

                     HC CUT-POINT, ppm
   Figure 17.  Variation of Ec,  Eo, and FF with HC Cut-point;
        All Three Cities; Federal Three-Mode; Bounded
                Errors of Commission Method
                            147

-------
  o

 "u
o
UJ
Q_
60

50

40

30

20

10
                       	FF
H  -  HIGH  SPEED
L  -  LOW SPEED
D  -  IDLE  IN DRIVE
N  -  IDLE  IN NEUTRAL
                   40,000       80,000      120,000

                      CO CUT-POINT, ppm
     Figure 18.  Variation of Ec, EQ, and FF with CO Cut-point;
          All Three Cities; Federal Three-Mode; Bounded
                  Errors of Commission Method
                              148

-------
o
UJ
Q_
     30
     20
                                          H  - HIGH SPEED
                                          L  - LOW SPEED
                                          D  - IDLE IN DRIVE
                                          N  - IDLE IN NEUTRAL
                           	FF
           N
            500   1000  1500  2000  2500  3000  3500  4000  4500  5000
                          NO CUT-POINT, ppm
                            A
 Figure 19.   Variation of Ec, Eo,  and FF with NOX Cut-point;
       All Three Cities; Federal Three-Mode; Bounded
                Errors  of Commission Method
     o
     UJ
     Q_
                                           HIGH SPEED
                                           LOW SPEED
                                           IDLE JN DRIVE
                                            OLE IN NEUTRAL
                 100   200  300   400   500   600  700

                       HC CUT-POINT, ppm
 Figure 20.  Variation of Ec, Eo,  and FF with HC Cut-point;
           Chicago; Federal Three-Mode; Bounded
                Errors of Commission Method
                              149

-------
  o
o
UJ
a.
                      	FF

H  -  HIGH  SPEED
L -  LOW SPEED
D -  IDLE  IN DRIVE
N -  IDLE  IN NEUTRAL
                  40,000       80,000       120,000

                     CO  CUT-POINT, ppm
 Figure 21.  Variation of Ec, Eo,  and FF with CO Cut-point;
         Chicago; Federal Three-Mode; Bounded
             Errors of Commission Method
                         150

-------
o
a:
     30
     20
     10
      0
                            	FF
H - HIGH  SPEED
L -  LOW SPEED
D -  IDLE  IN  DRIVE
N -  IDLE  IN  NEUTRAL
        0    500  1000  1500  2000   2500  3000  3500  4000  4500  5000

                          NO  CUT-POINT,  ppm
  Figure 22.  Variation of Ec, Eo,  and FF with NOx Cut-point;
            Chicago; Federal Three-Mode; Bounded
                 Errors of Commission Method
                         	pp
     o
     o;
                                        H - HIGH SPEED
                                        L - LOW  SPEED
                                        D - IDLE IN DRIVE
                                        N - IDLE IN NEUTRAL
                       200   300  400   500

                        HC CUT-POINT, ppm
   Figure 23.  Variation of Ec,  Eo, and FF with HC Cut-point;
             Houston; Federal Three-Mode; Bounded
                 Errors of Commission Method
                               151

-------
UJ
UJ
UJ
O

UJ
ex.
60 i-

50

40

30
    //
   IT
,i
      0
       0
                            Ec
                            Eo
                            FF
                                H  -  HIGH SPEED
                                L -  LOW SPEED
                                D -  IDLE  IN  DRIVE
                                N -  IDLE  IN  NEUTRAL
                        ~3
            40,000       80,000      120,000
              CO CUT-POINT, ppm
   Figure 24.  Variation of Ec, Eo, and FF with CO Cut-point;
           Houston; Federal Three-Mode; Bounded
               Errors of Commission Method
                           152

-------
o
LU
0-
                                  "  Eo
                             	FF
                                       .H - HIGH  SPEED
                                       L - LOW SPEED
                                       D - IDLE IN DRIVE
                                       N - IDLE IN NEUTRAL
                           L
                                  H
 30


 20


 10


  0
    0    500   1000  1500  2000  2500  3000  3500  4000  4500  5000

                        NO  CUT-POINT, ppm
                          A

Figure 25.  Variation of Ec, Eo, and FF with NOX Cut-point;
          Houston; Federal Three-Mode; Bounded
              Errors of Commission Method
     o

     LLJ
     0-
    40


I   30
j°
;o  20
*
    10


     0
r
Ec
0
	 FF
N-w
H -
L -
D -
N -
HIGH SPEED
LOW SPEED
IDLE IN DRIVE
IDLE IN NEUTRAL
             0   100   200   300   400    500   600   700

                        HC CUT-POINT,  ppm

     Figure 26.  Variation of Ec, Eo, and FF with HC Cut-Point;
              Phoenix; Federal  Three-Mode; Bounded
                   Errors of Commission Method
                                153

-------
  o
o
LLJ
Q_
60

50

40

30

20

10

 0
                            	H
                                      H  - HIGH SPEED
                                      L  -  LOW  SPEED
                                      D  -  IDLE IN DRIVE
                                      N  -  IDLE IN NEUTRAL
        0
            40,000       80,000       120,000

               CO CUT-POINT, ppm
   Figure 27.  Variation of Ec,  EQ, and FF with CO Cut-point;
            Phoenix; Federal Three-Mode; Bounded
                Errors of Commission Method
                            154

-------
Ul
Ul
             o
             OS!
30

20

10

 0
                                                             H
                                                             L
                                                             D
                                                             N
                                               HIGH SPEED
                                               LOW  SPEED
                                               IDLE IN DRIVE
                                               IDLE IN NEUTRAL
                     0    500   1000  1500  2000  2500   3000   3500 4000   4500  5000


                                          NO  CUT-POINT,  ppm
                                            A

                       Figure 28.  Variation of Ec, Eo, and FF with NOX Cut-point;
                                Phoenix; Federal Three-Mode; Bounded
                                    Errors of Commission Method

-------
O
o;
UJ
Q_
    15
    10
       o  DRIVE MODE
       A  BEST MODE
      RANGE OF VARIATION
       -  DRIVE MODE
       ^  BEST MODE
A
vl/
                                                  10
O
or
        o  DRIVE MODE
60•-    v  BEST MODE
   40
    20
                               	FF
                               • DRIVE MODE
                               A BEST MODE
                        I
                             I
                       4        6
                  PREDICTED  PERCENT E.
                                              10
    Figure 29.  Variation of Actual Ec, Eo,  and FF
       with Predicted Ec;  Federal Three-Mode;
                  All Three Cities
                        156

-------
Q_
—I
<
    20
    15
    10
         o   DRIVE MODE
         A   BEST MODE
         RANGE OF VARIATION
         -   DRIVE MODE
         rt   BEST MODE
                            o
                                     I
t
                                                  10
 p40 -
o
C£
UJ
Q_
O
<
              o DRIVE MODE
              v BEST MODE
                                 	FF
                                 •  DRIVE  MODE
                                 A  BEST MODE
     0
                      4         6,
                   PREDICTED PERCENT E
    Figure 30.  Variation of Actual Ec,  EQ,  and FF
       with Predicted Ec; Federal Three-Mode;
                      Chicago
                         157

-------
    15 -
    10 -
UJ
Q_
o DRIVE MODE
A BEST MODE
RANGE OF VARIATION
- - DRIVE MODE
- BEST MODE
1'
<
(

;
i ^ i
<

' t
\

i
i

(
)
t
\
i

i
\


> (

» ^

k L

f \


(
>
/
\
i
i

f
* >

> (
1 t
'
\
i
b

V


O

UJ
Q_





O
   80
   60
   40
   20
                                                  10
                      :0	FF
               o DRIVE MODE  •  DRIVE  MODE
               v BEST MODE    A  BEST MODE
                                1
                       4        6
                   PREDICTED  PERCENT E
   Figure 31.  Variation of Actual Ec,  Eo, and FF
       with Predicted Ec; Federal Three-Mode;
                      Houston
                                                  10
                         158

-------
 o
 DC
 O
 
-------
   o
     40
     30
     20
     10
      0
	FF
       0.8   1.0   1.2   1.4   1.6    1.8   2.0
                    HC CUT-POINT,  gm/mi
        2.2   2.4
Figure 33.   Variation of Ec,  E0,  and FF with HC Cut-point;
         150 CID and Less; Federal Short Cycle;
         Bounded Errors of Commission Method
                                            	FF
                   14    16          20
                   CO CUT-POINT,  gm/mi
Figure 34.  Variation of Ec, Eo, and FF with CO Cut-point;'
          150 CID and Less; Federal Short Cycle;
          Bounded Errors of Commission Method
                           160

-------
 30

 20

,10

'  0
                                                      c

                                                      0
  1.8    2.0   2.2
                    NO
                      x
                           2.6          3.0
                         CUT-POINT, gm/mi
             3.6
  Figure 35.  Variation of Ec, Eo,  and FF with NOX Cut-point;
           150 CID and Less; Federal Short Cycle;
           Bounded Errors of Commission Method
o
LU
Q_
30
20
10
 0
                                                     0
                                          	FF
 0.8    1.0    1.2
                    1.4   1.6    1.8   2.0
                    HC CUT-POINT,  gm/mi
2.2   2.4   2.6
  Figure 36.  Variation of Ec, EQ, and FF with HC Cut-point;
         151 to 259 CID; Federal Short Cycle; Bounded
               Errors of Commission Method
                           161

-------
o
50

40

30

20

10

 0
                                                  0
                                       	FF
      0     2.0    4.0    6.0   8.0    10     12
                    CO CUT-POINT, gm/mi
                                             14     16
 Figure 37.   Variation of Ec, Eo, and FF with CO Cut-point;
        151 to 259 CID; Federal Short Cycle; Bounded
              Errors of Commission Method
                                                    0
                                          	FF
                    NO  CUT-POINT,  gm/mi
                      A

 Figure 38.  Variation of Ec, Eo, and FF with NOX Cut-point;
        151  to 259 CID; Federal Short Cycle; Bounded
               Errors of Commission Method
                           162

-------
Q

<
   30
                                            	FF
   10
    0
                                                          I
     0.8   1.0   1.2  1.4   1.6   1.8   2.0  2.2   2.4   2.6   2.8
                       HC CUT-POINT,  gm/mi
Figure 39.   Variation of Ec,  EQ, and FF with HC Cut-point;
         260 CID and Greater,  Federal Short Cycle;
          Bounded Errors of Commission Method
o
LLJ
Q_
60

50

40

30

20

10
                                             	FF
     0
      0.0   2.0   4.0   6.0    8.0   10.0  12.0  14.0   16.0  18.0  20.0
                       CO CUT-POINT,  gm/mi
Figure 40.   Variation of Ec,  Eo, and FF with CO Cut-point;
         260 CID and Greater; Federal Short Cycle;
          Bounded Errors of Commission Method
                             163

-------
    30
    20

o
ex.
    10
     0
      2.0
                                                                0
                                                       	FF
2.4          2.8          3.2
        NO   CUT-POINT,  gm/mi
3.6
4.0
      Figure 41.  Variation of Ec,  Eo,  and FF with NOX Cut-point;
              260 CID and Greater; Federal Short Cycle;
               Bounded Errors of Commission Method

-------
           15
           10
        CtL
                                 i
                                                      10
             10
                                                                a:
                                                                 o
                                                                                                               6
                                                                                              6
                                                        10
Ul
           60I-
        Q
        ^
         ,o40
        o
        en.
           20
                     246
                         PREDICTED PERCENT E.
10
          o
          a:
                                                                <
                                                                I—
                                                                <
             60
            '40
             20
                                                                                       FF	-- —
2468
     PREDICTED PERCENT E
10
        Figure 42.  Variation of Actual Ec, Eo, and FF
           with Predicted EC,  Federal Short Cycle;
           All Three Cities; Engine Displacement
                       150 CID and Less
          Figure  43.  Variation of Actual Ec, Eo,  and FF
              with Predicted Ec, Federal Short Cycle;
              All Three  Cities; Engine Displacement
                          151 to 259 CID

-------
o
01
LiJ
Q_

	I
<
   15
   10
OM>
 0
                                                    10
   60
   40
   20
                          FF	.-	—
           2         4        6.8
                PREDICTED  PERCENT E.
                                                    10
   Figure 44.  Variation of Actual Ec, EQ, and FF
       with Predicted Ec; Federal Short Cycle;
        All Three Cities; Engine Displacement
                260 CID and Greater
                         166

-------
u_
o
<
LU°
0
LU
1—
§
Qi
LkJ
0-
40

30

20

10

n
—

H ->--'
,;XXL


- \L x
^^.
1^
                                  Ec
                                  Eo
                                  FF
H - HIGH SPEED
L -  LOW SPEED
D -  IDLE IN DRIVE
N -  IDLE IN NEUTRAL

     N
         50   100   150  200  250   300   350  400  450   500  550

                          HC CUT-POINT, ppm
Figure 45.   Variation of Ec, Eo, and FF with HC Cut-point;
           150 CID and Less; Federal Three-Mode;
          Bounded Errors of Commission Method
                                          H  - HIGH SPEED
                                          L - LOW SPEED
                                          D - IDLE IN DRIVE
                                          N - IDLE IN NEUTRAL
    o
    a:
                  20,000     40,000     60,000

                          CO CUT-POINT, ppm
      80,000    100,000
Figure 46.  Variation of Ec,  Eo, and FF with CO Cut-point;
           150 CID and Less;  Federal Three-Mode;
           Bounded Errors of Commission Method
                               167

-------
 o
 Qi

r
Ec
	 E
0
	 FF
H
L
D
N

                                             HIGH SPEED
                                             LOW SPEED
                                             IDLE IN DRIVE
                                             IDLE IN NEUTRAL
                                            		H
                 1000
                     2000
3000
4000
5000
                          NO CUT-POINT,  ppm
                            A
Figure  47.  Variation of Ec, Eo, and FF with NOX Cut-point;
          150 CID and Less Federal Three-Mode;
          Bounded Errors of Commission Method
  o
  a:
30

20

10

 0
                                  E

                                  Eo
                                  FF
                                   H - HIGH SPEED
                                   L - LOW SPEED
                                   D - IDLE IN DRIVE
                                   N - IDLE IN NEUTRAL
             50   100   150   200   250   300   350   400   450   500

                          HC CUT-POINT, ppm
Figure 48.  Variation of Ec, Eo, and FF with HC Cut-point;
           151 to 259 CID;'Federal Three-Mode;
           Bounded Errors of Commission Method
                             168

-------
LU
UJ
UJ
O
CX.
     60 r~
                                    E
                                    FF
                                               H -
                                               N -
                     HIGH SPEED
                     LOW SPEED
                     IDLE  IN  DRIVE
                     IDLE  IN  NEUTRAL
                                        H
                                                       _-	N
                                                   ~-~ - " D
                                                            D
                                                                N
                                                               1
                  10,000
20,000
30,000
40,000
                          CO  CUT-POINT,  ppm
        Figure 49.  Variation of Ec, Eo, and FF with CO Cut-point;
                 151 to 259 CID; Federal Three-Mode;
                Bounded Errors of Commission Method
                                 169

-------
  o
  o;
                                       H -  HIGH SPEED
                                       L -  LOW SPEED
                                       D -  IDLE IN DRIVE
                                       N -  IDLE IN NEUTRAL
        0
         0   400  800  1200  1600  2000  2400  2800 3200

                      NO  CUT-POINT,  ppm
                  ,      A'
Figure 50.  Variation of Ec, EQ,  and FF with NOX Cut-point;
           151 to 259 CID; Federal Three-Mode;
          Bounded Errors of Commission Method
 UJ

 a;
 UJ
 D-
40

30

20

10

 0
	 EC
-Eo
 — FF
                 /L
                                         H
                                         L
                                         D
                                         N
HIGH SPEED
LOW SPEED
IDLE  IN DRIVE
IDLE  IN NEUTRAL
             40.  80   ,120   160   200   240   280  320  360
                       HC  CUT-POINT, ppm
Figure 51.   Variation of Ec,  EQ, and FF with HC Cut-point;
        260 CID and Greater, Federal Three-Mode;
          Bounded Errors of Commission Method
                             170

-------
                                        H - HIGH SPEED
                                        L - LOW SPEED
                                        D - IDLE IN DRIVE
                                        N - IDLE IN NEUTRAL
        0   2000       6000      10,000     14,000

                          CO CUT-POINT, ppm
18,000
Figure 52.  Variation of Ec,  EQ, and FF with CO Cut-point;
        260  CID and Greater;  Federal Three-Mode;
          Bounded Errors  of Commission Method
    o
    LU
    D-
                         	FF
                                        H -  HIGH SPEED
                                        L -  LOW SPEED
                                        D -  IDLE IN DRIVE
                                        N -  IDLE IN NEUTRAL
              400   800  1200  1600  2000  2400  2800  3200
                       N0v CUT-POINT, ppm
                         A
Figure 53.   Variation of Ec,  Eo, and FF with NOX Cut-point;
        260 CID and Greater; Federal Three-Mode;
          Bounded Errors of Commission Method
                             171

-------
   15
   10
oc.
Q.
C_3
 o  DRIVE MODE
 A  BEST MODE
RANGE OF VARIATION
 -  DRIVE MODE
 ~  BEST MODE
      11
            T
        51
                                        o   o
                                            10
o
LU
Q-
__l
<
   80
   60
   40
   20
            o
     o DRIVE MODE
     v BEST MODE
	FF
 DRIVE MODE
> BEST MODE
            468
         PREDICTED PERCENT E_
                                           10
 Figure 54,  Variation of Actual Ec, Eo, and FF
    with Predicted Ec; Federal Three-Mode;
     All Three Cities; Engine Displacement
               150 CID and Less
                      172

-------
    15
   10
o
o

-------
     15
     10
  UJ
  Q_
  o  DRIVE MODE
  A  BEST MODE
RANGE OF VARIATION
  -  DRIVE MODE
  ~  BEST MODE
           I
                                                 10
O
o
<
     80
     60
     40
     20
                  o
           o DRIVE MODE
           A BEST  MODE
                     FF
                 DRIVE  MODE

                 BEST MODE
                                 I
                        4        6
                    PREDICTED PERCENT E,
                                       10
  Figure 56.  Variation of Actual Ec, EQ, and FF
     with Predicted Ec; Federal Three-Mode;
      All Three Cities; Engine Displacement
               260 CID and Greater
                        174

-------
                    FEDERAL SHORT CYCLE
0       20
40       60       80
       TIME, sec
100
         Figure 57.  Federal Short Cycle Test
                   Driving Schedules
                          175

-------
                                           FF, CORRECTLY
                                           FAIlfD VEHIClfS
                    E , ERROR OF

                    COMMISSION
    ST CUT-POINT
to
          PP, CORRECTLY
          PASSED VEHICLES
                                    Q_
                                    t
                                             E ,  ERROR OF

                                             OMISSION
                   FTP MEASUREMENT
         Figure 58.  Contingency Table Representation
                           :17.6

-------
      MINIMIZE EQ SUBJECT TO EC < /%
    CO
           NOT TO
          EXCEED /%
           ST     •
       CUT-POINT  *
• _ •
                            MINIMIZE
          FTP MEASUREMENT
Figure 59.  Bounded Errors of Commission Method
                   177

-------
BIVARIATE  NORMAL DISTRIBUTION
                     'X, -AA2
               1
D(Xr X2) -
WHERE
           /z2
                                  277- CTj 0*2

                   POPULATION MEANS
                   POPULATION STANDARD DEVIATIONS
                   CORRELATION COEFFICIENT
                   FTP MEASUREMENT
                   ST MEASUREMENT

                 Figure 60.   Parametric Model
             PROBABILITY OF ERROR OF COMMISSION
                                  Ci
                                     0
             Pr { Xj < Cr X2 > C2 } =       D (Xr X
             PROBABILITY OF ERROR OF OMISSION
                                +00 G£
             Pr{ X^Cj. X2C1,X2>C2( - f
                                Cl C2
             WHERE   C, =  FTP STANDARD
                     C2 =  ST CUT-POINT
               Figure  61.  Probability Equations
                               178

-------
Bounded Errors of Commission:  Solve for


                    Cl ,r\
y
                   I  I  D(XI(
where
    Cj = FTP Standard
    C2 = ST Cut-Point
      •y = Maximum allowable probability of an error of
          commission
 D(. ,  . ) = Statistical model as determined by the data
        Figure 62.  Equations for ST Cut-Point
                    Determination
                          179

-------
START
1
Determine ST
cut-points for
HC. CO. and NO
n.

Set
Policy
1

  Write Results of
  Contingency  Table
    Analysis for
  HC.  CO. and NO
      Determine
    3-Constituent
     Test Results
    for each Mode,
    Write Results
           YES
     Determine
        Best
   3-Constituent
        Test
     Determine
    Test Results
   Write Results
                        NO
                                   FINISH
      FINISH
Figure 63.   Computation Flow Chart
                  180

-------
80
70
60
50
40
30
20
 10
                            I
ERROR BAR INDICATES PLUS AND MINUS ONE

STANDARD DEVIATION OF THE ESTIMATE
                    o

              	FF
                                   »
            lx      A
                                         s
                                       S  -L
         I      234567

                    HC FEDERAL SHORT CYCLE, gm/mi
Figure 64. Typical Variability of Predicted Population Results
                            181

-------
                               r
     INPUT
      DATA
                                INITIALIZE
                               VEHICLE MIX,
                              AVE.  MILEAGE
                                   ETC.
                                GENERATE
                                 VEHICLES
                                  TO BE
                                 TRACKED
                     UPDATE
                  CONTROLLED
                     FLEET
                      (I/M)
                   UPDATE
               UNCONTROLLED
                   FLEET
                  (NON-I/M)	
 I/M MILEAGE
ACCUMULATION
     AND
DETERIORATION
    MODEL
  CALCULATE
     207(b)
EFFECTIVENESS
'OUTPUT
 RESULTS
/
   VEHICLE
MAINTENANCE
    MODEL
   EMISSION
 INSPECTION
    MODEL
                            NON-I/M
                            MILEAGE
                         ACCUMULATION
                              AND
                         DETERIORATION
                             MODEL
   CORRECT
   RESULTS
      TO
  50,000 MILES.
                               C
     STOP
  Figure 65.  Logical Flow Chart for 207(b) Effectiveness Estimation
                                  182

-------
                 40-
oo
UJ
o
4->
c

I  204.
'c
OH
                 10-.
                           O  -  Original emission rate

                           &  -  After application of I/M program
                                                          	 Rapid deterioration
                                                                cancelling maintenance
                                                                in 9 months


                                                          -—- — EFP deterioration
                                                                  Adjustment due
                                                                  to Maintenance
                        Maintenance
                       Effectiveness
                          Period
                                                Maintenance
                                               Effectiveness
                                                  Period
—I	1	1—
 10    12    14

    Months
                                                                 16
                                                         18
20
22
24
                          Figure 66.  Piecewise Linear Deterioration for FF Vehicles
                                          Shown Over Two-Year Period

-------
     Current Year's

      Classification
Possible  Classifications

       Next Year(a'
           FF



           E
            o


           E
            c


           PP
      E  ,  FF
       o


      E  ,  FF
       o


      E  ,  FF
       o


   PP, E ,  E ,  FF
         c    o
(a)
   The probability of each classification is determined by

   the inspection model based upon the 300-car data
      Figure 67.  Vehicle Classification Properties of the

                207(b) Effectiveness Simulator
                             184

-------
                              GLOSSARY







CID           cubic inch displacement



CO           carbon monoxide



CVS          constant volume sampling



E             error of commission
 c


EFP          Emission Factors Program



E             error of ommision
 o


FF           vehicles failed by both the  ST and the FTP



FTP          Federal  Test Procedure



HC           hydrocarbon



I/M           inspection and maintenance



NO           oxides of nitrogen
   x


PP           vehicles passed by both the ST and the FTP



ST            short test



STE          short test effectiveness



STRR         short test rejection ratio



207(b)         reference to section 207(b) of the  1970 Clean Air Act
                                  GL-1

-------
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
1. REPORT NO.
  EPA-460/3-76-010a
                            2.
                                                         3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
    Short Test Correlation Analyses on 300
    1975 Model Year Cars
        Volume I
             5. REPORT DATE
                   October  1976
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                         8. PERFORMING ORGANIZATION REPORT NO.
           M. G.  Hinton and John C. Thacker
              ATR-77(7356)-l
9. PERFORMING ORGANIZATION NAME AND ADDRESS
    The Mobile Systems Group
    Environment & Energy Conservation Division
    The Aerospace Corporation
    El Segundo,  California 90245
                                                         10. PROGRAM ELEMENT NO.
             11. CONTRACT/GRANT NO.
                68-01-0417
12. SPONSORING AGENCY NAME AND ADDRESS
    EPA Office of Air and Waste Management
    Office of Mobile Source Air Pollution Control
    Emission Control Technology Division
    Ann Arbor, Michigan  48105
             13. TYPE OF REPORT AND PERIOD COVERED
                Final
             14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
    A series of statistical analyses was performed to determine the degree of
    "correlation" that exists between two specific short tests  (STs) and the Federal
    Emission Certification Test Procedure (FTP) for new vehicles.  This work was
    performed to determine if "reasonable correlation with certification test
    procedures" exists;  this is a condition precedent to the promulgation of
    regulations that impose the in-use warranty provisions of Sec. 207(b) of the
    Clean Air Act of 1970 upon the motor vehicle manufacturers.

    The basis for the analyses was ST and FTP test data from three  100-vehicle
    fleets located in (a) Chicago, Illinois,  (b) Houston, Texas, and (c)  Phoenix,
    Arizona.  Each of the vehicles in these fleets was tested by the FTP,  the
    Federal Short Cycle, and the Federal  3-Mode.  Two different statistical
    analysis methods •were used to assess "correlation"  -- a conventional
    correlation analysis, and a contingency table analysis.
17.
                               KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                            b.lDENTIFIERS/OPEN ENDED TERMS
                          c. COSATl Field/Group
       Air Pollution
       Emission Testing
       Short Test Procedures
       Test Correlations
  Air Pollution Control
  Conventional Corre-
    lation Analysis
  Contingency Table
    Analysis
     13 B
    14 B
18. DISTRIBUTION STATEMENT

       Unlimited
19. SECURITY CLASS (This Report)
     Unclassified
21. NO. OF PAGES
    439
20. SECURITY CLASS (This page)
     Unclassified
                          22. PRICE
EPA Form 2220-1 (9-73)

-------
                                                         INSTRUCTIONS

    1.   REPORT NUMBER
        Insert the EPA report number as it appears on the cover of the publication.

    2.   LEAVE BLANK

    3.   RECIPIENTS ACCESSION NUMBER
        Reserved for use by each report recipient.

    4.   TITLE AND SUBTITLE
        Title should indicate  clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smaller
        type or otherwise subordinate it to main title. When a report  is prepared in more than one volume, repeat the primary title, add volume
        number and include subtitle for the specific title.

    5.   REPORT DATE
        Each report shall carry a date indicating at least month and year.  Indicate the basis on which it was selected (e.g., date of issue, date of
        approvcl, date of preparation, etc.).

    6.   PERFORMING ORGANIZATION CODE
        Leave blank.

    7.   AUTHOR(S)
        Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.).  List author's affiliation if it differs from the performing organi-
        zation.

    8.   PERFORMING ORGANIZATION REPORT  NUMBER
        Insert if performing organization wishes to assign this number.

    9.   PERFORMING ORGANIZATION NAME AND ADDRESS
        Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.

    10.  PROGRAM ELEMENT NUMBER
        Use.the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.

    11.  CONTRACT/GRANT NUMBER
        Insert contract or grant number under which  report was prepared.

    12.  SPONSORING AGENCY NAME AND ADDRESS
        Include ZIP code.

    13.  TYPE OF REPORT AND PERIOD COVERED
        Indicate interim final, etc., and if applicable, dates covered.

    14.  SPONSORING AGENCY CODE
        Leave blank.

    15.  SUPPLEMENTARY  NOTES
        Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of. Presented at conference of,
        To be published in, Supersedes, Supplements, etc.

    16.  ABSTRACT
        Include a brief (200 words or less) factual summary of the most significant information contained in the report. If the report contains a
        significant bibliography or literature survey, mention it here.

    17.  KEY WORDS AND DOCUMENT ANALYSIS
        (a) DESCRIPTORS - Select from the Thesaurus Of Engineering and Scientific Terms the proper authorized terms that identify the major
        concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.

        (b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc.  Use open-
        ended terms written in descriptor form for those subjects for which no descriptor exists.

        (c) COS ATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COS ATI Subject Category List. Since the ma-
        jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
        endeavor, or type of physical object.  The application(s) will be cross-referenced with secondary  Field/Group assignments that will follow
        the primary posting(s).

    18.  DISTRIBUTION STATEMENT
        Denote releasability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
        the public, with address and price.

    19. & 20. SECURITY CLASSIFICATION
        DO NOT submit classified reports to the National Technical Information service.

    21.  NUMBER OF PAGES
        Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.

    22.  PRICE
        Insert the price set by the National Technical Information Service or thc-Government Printing Office, if known.
EPA Form 2220-1 (9-73) (Revcru)

-------