United States       Air and Radiation      EPA420-P-99-029
            Environmental Protection               November 1999
            Agency                     M6.EXH.006
vvEPA     Determination of Methane
            Offsets as a Function of
            Mileage for Light-Duty
            Cars and Trucks
                                 > Printed on Recycled Paper

-------
                                                                           EPA420-P-99-029
                                                                               November 1999
                         of                           as  a                of
                       for
                                     Phil Enns
                         Assessment and Modeling Division
                              Office of Mobile Sources
                        U.S. Environmental Protection Agency
                                     NOTICE

    This technical report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using data which are currently available.
         The purpose in the release of such reports is to facilitate the exchange of
      technical information and to inform the public of technical developments which
        may form the basis for a final EPA decision, position, or regulatory action.

-------
                                       - Draft -

           Determination of Methane Offsets as a Function of Mileage
                          for Light-Duty Cars  and Trucks

                           Report Number M6.EXH.006

                                    September 1999

                                       Phil Enns

                      U.S. EPA Assessment and Modeling Division
    1.0    Introduction

           Previous reports document how the MOBILE6 emission factor model will allocate
    vehicle exhaust emissions between engine start (start emissions) and travel (running
    emissions)1'2. This split allows the separate characterization of start and running emissions
    for correction factors such as fuel effects and ambient temperature. It also enables a more
    precise weighting of these two aspects of exhaust emissions for particular situations such as
    morning commute, parking lot and freeway driving.
           Because methane  does not contribute significantly to ozone formation, MOBILE
    attempts to separate it from total hydrocarbons in estimating vehicle emissions. This
    document describes methodologies for calculating this methane "offset" based on the
    separation of start  and running emissions proposed for MOBILE6. The procedure follows
    the scheme of modeling the in-use deterioration of emissions as a function of accumulated
    mileage. For model year 1981-1993 light-duty cars and trucks, stratified into carbureted
    versus fuel-injected, the method parallels that used for total hydrocarbons, carbon monoxide
    and oxides of nitrogen. This analysis utilizes actual methane data from emissions tests
    conducted on vehicles from those model years.
           For pre-1981 model years,  data of the type used in the newer vehicles is not
    available. Therefore, procedures are described for estimating methane as a function of
    JCarey, P., P. Enns, E. Glover, and M. Sklar, "Determination of Running Emissions as a
    Function of Mileage for 1981-1993 Model Year Light-Duty Cars and Trucks,"
    Report Number M6.EXH.001, October 1998.

    2Carey, P. and E. Glover, "Determination of Start Emissions as a Function of Mileage and
    Soak Time for 1981-1993 Model Year Light-Duty Vehicles," Report No. M6.STE.003,
    October 1998.


Methane                               -8-                              11/8/99

-------
    mileage using existing data in combination with methods that are applied in MOBILE
    version 5.

    2.0 Data

           The data underlying the analysis of 1981-93 light-duty vehicles are drawn from a
    subset of the Federal Test Procedure (FTP) tests described in the reports cited above. These
    tests were conducted by EPA, the American Automobile Manufacturers Association
    (AAMA), and the American Petroleum Institute (API). Most, but not all, of these tests
    produced measurements of methane. In particular, the sample sizes from the full data set and
    the reduced sets for which methane is recorded are compared in the table below, subdivided
    by vehicle type and the model year/technology groups used to determine basic emission rates
    of total HC, CO and NOx.
                      CARS



GROUP
81-82 GARB
81-82 FI
83-85 GARB
83-87 FI
86-93 GARB
88-93 PFI
88-93 TBI
DATA
REDUCED
N

580
88
203
688
93
1361
437
SET
FULL
N

1166
126
253
726
96
1605
444




GROUP
81-83 GARB
81-87 FI
84-93 GARB
88-93 PFI
88-93 TBI
TRUCKS
DATA
REDUCED
N

72
92
125
199
458

SET
FULL
N

180
94
134
330
467
           Another key data set is that based on a sample of FTP tests to which were appended a
    505-second cycle without an engine start. This cycle is identical to that of Bags 1 and 3, but
    contains no emissions associated with the cold start of Bag 1 or warm  start of Bag 3. It is
    referred to as the Hot Running 505 (HR505). Pure cold and warm start emissions are
    estimated by deducting HR505 emissions from the 505  bags that include a start.3
           The data from this test program were used to estimate the relation between the
    3Brzezinski, D. and P. Enns, "The Determination of Hot Running Emissions from FTP Bag
    Emissions", Report No. M6.STE.002, December, 1997.
Methane
11/8/99

-------
    HR505 and Bags 1 to 3 of an FTP. From this function, the portions of FTP emissions
    attributed to start and running are computed. This calculation then was applied to the large
    FTP data set described above for which Bags 1 to 3, but not the FIR505, are measured.

    3.0 Modeling Methane Deterioration in MOBILE6 - 1981-93 Vehicles

           To understand the current analysis, it is helpful to review how emissions
    deterioration is modeled in MOBILE6. The basic method involves separating start and
    running emissions.

        3.1 Running Emissions

           For the running component, simple linear functions of emissions versus mileage are
    fitted by the method of least squares regression. To improve the fit at low mileage, the mean
    of emissions is used. This produced piecewise continuous functions in which emission rates
    (in grams per mile) are constant at low mileage and increase when mileage accumulation
    exceeds approximately 20,000 miles. The higher mileage portion of the function has
    constant slope in most cases, but under certain conditions the slope may change, adding a
    second "corner" point to the graph of emissions as a function of mileage. There was a
    concern that the FTP data suffer from sample bias due to the self-selecting nature of the data
    collection. In order to account for such bias, adjustments were made to these running
    emission lines using data from a large sample of inspection and maintenance tests conducted
    in Dayton, Ohio. (See document (1) for details.) However, methane was not recorded in
    these tests, so it was not possible to compute this adjustment for the analysis described in
    this report. Instead, we propose using a proportional adjustment for sample bias
    corresponding to that used for total HC.
           Table 1 reports the deterioration coefficients derived for methane using the reduced
    data set. The first slope is always zero, reflecting the use of the low mileage mean (ZML
    Emissions). The first corner occurs at the mileage where the emission rate begins to slope
    positively. For total HC, only the 1983-87 fuel-injected car category has a second corner.
    With methane, several categories have two  corners, while the 1983-85 carbureted car group
    has zero deterioration at all mileages.
           The high emitter adjustment was determined as proportionately the same at a given
    mileage as for total HC. These adjustment factors are applied additively to the initial THC
    emissions. They raise or lower THC by an amount that is a linear function of mileage and is
    zero at mileage zero. In order to apply this factor to methane it is necessary to first compute
    unadjusted and adjusted THC at a given mileage. Their ratio is then multiplied by unadjusted
    methane to obtain adjusted methane.
           Figure 1 illustrates the unadjusted and adjusted methane deterioration lines for the
    case of 1988 to  1993 model year port fuel-injected cars. As described, the adjustment in
    methane is proportional to the corresponding HC adjustment, which also is illustrated.
    Figure 2 shows the final adjusted methane lines by model year-technology group for cars and
    trucks.
Methane                                 -8-  -3-                            11/8/99

-------
        3.2 Start Emissions

           Deterioration of start emissions (measured in grams) was modeled somewhat
    differently. First, the tested vehicles were stratified into categories of "normal" and "high"
    emitters depending on how their FTP values compare to applicable standards. For HC and
    CO, the normal emitter emissions were regressed against mileage, while high emitter
    emissions were fitted by the mean. At a given mileage, the normal and high emitter estimates
    were combined in a weighted average, where the weight reflects the proportion of high
    emitters at that mileage. Reference (2) gives a complete description of this procedure.
           Table 2 presents coefficients used to compute the start portion of emission rates at a
    given mileage. The majority of the groups actually show negative slope estimates for the
    normal emitter regression lines. However, in all but one of these, the slope is not statistically
    significant when tested against zero. Since negative deterioration is intuitively unexpected, it
    was decided to use the normal emitter mean value when the regression estimate is negative.
    It also is possible for estimated cold start emissions to be negative. This occurs when the
    fitted Hot Running 505 value exceeds the observed Bag 1 value. In the case of 1988-93 TBI
    cars, this produced an anomaly in which the mean methane of the high emitters is negative
    and the start emission deterioration line has negative slope.  For this case, the deterioration
    line was assigned a constant value equal to its zero mile level.
           Table 3 gives the fraction of high emitters estimated for THC which is used to
    compute the weighted average described above. These are the same values used for the full
    sample total HC calculations. Graphs of start emissions versus mileage for cars and trucks
    appear in Figure 3.
       3.3 FTP Emissions
           Final FTP deterioration functions are obtained by combining the start and running
    estimates according to their relative importance in that test. The weights used coincide with
    those used for HC,  CO and NOx. They form the equation

                            FTP=(7.5*Run + .43*CS + .57*HS)/7.5

     where Run is emissions in grams per mile from the running LA4 portion; CS is emissions in
    grams from cold start; and HS is hot start, computed as simply 0.16*CS.  The factor 7.5 is the
    driving mileage corresponding to Bags 1 and 2 (the LA4 cycle) in the FTP.
    4.0 Other Model Years

       4.1 Pre-1981 Model Years

           In MOBILES, start and running emissions are not separated. MOBILES contains
    deterioration functions for methane for each of the three FTP bags. This enables the


Methane                                -8- -4-                           11/8/99

-------
    calculation of bag values at selected vehicle mileages. These values can in turn be used to
    compute running and start emissions when combined with the regression coefficients
    derived from the 77-car test data. The validity of this approach rests on the assumption that
    start and running emissions in older vehicles occur in a manner similar to that of late model
    vehicles. For carbureted cars, graphs of running and start emissions are shown in Figure 4.
    The graphs include the proposed MOBILE6 functions for 1981-82 and for 1983-85. It is
    apparent that the deterioration functions for the two versions of the model  are reasonably
    compatible. Similar graphs for fuel injected cars and for trucks yield the same impression.
    Therefore, in the absence of new data, EPA proposes to separate start and running methane
    emissions in pre-1981 light-duty vehicles and trucks using the methodology of MOBILE6
    applied to the deterioration functions employed in MOBILES.
           For diesel vehicles and motorcycles, FTP bag values are not provided in MOBILES.
    Thus, it is not possible to compute running and start emissions by the method outlined
    above. In these cases, EPA proposes to use fractions described in a separate report4 that are
    applied to total hydrocarbon deterioration functions to determine methane  function
    coefficients.

       4.2 Model Years 1994 and Later

           For model year 1994 and later vehicles subject to Tier I standards, MOBILE6
    predicts nonmethane hydrocarbon (NMHC) emissions directly. In the case of LEVs, the
    model predicts nonmethane organic gas (NMOG). No adequate methane data exist for
    estimating methane deterioration functions for these model years. Therefore, EPA proposes
    developing multiplicative methane offsets from these projections in a manner consistent with
    the treatment of NMHC and NMOG. In general, emission rates for those pollutants are
    obtained using ratios of standards that apply to the newer vehicles to those of 1993 model
    year vehicles. For a given class of vehicles, there exists a ratio between NMHC  (or NMOG)
    in the two model year ranges. The proposed method would apply the same ratio to compute
    methane for newer  vehicles.
    5.0 Summary

           EPA proposes combining available methane data with methods employed for
    methane in MOBILES and total hydrocarbons in MOBILE6 to develop basic emission rates
    for methane in MOBILE6. This results in an eclectic approach that attempts to fully utilize
    the data while making reasonable assumptions where insufficient data exists. In general, the
    assumptions are consistent with methods applied to total and nonmethane HC.
       4Brzezinski, D. and J. Gilmore, "Basic Exhaust Emission Rates of Open Loop Vehicles
forMOBILE6," Report No. M6.EXH.005, May 1999.


Methane                                -8-                                11/8/99

-------
 Table 1: Methane Running Emission Deterioration Model Coefficients for Light-Duty Vehicles
                                         Cars

Model Year/
Technology
88-93 PFI
88-93 TBI
83-87 FI
86-93 CARB
83-85 CARB
81-82 FI
81-82 CARB
ZML Emission
(gr/m)

0.0167
0.0240
0.0365
0.0405
0.0721
0.0271
0.0845
First
Slope
(gr/m/lOOOm)

0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
First
Corner
(1000 miles)

15.47
32 .18
14.12
15.19
N/A
13 .92
22.11
Second
Slope
(gr/m/lOOOm)

0.0003
0.0002
0.0006
0.0002
N/A
0.0005
0.0005
Second
Corner
(1000 miles)

67.89
N/A
81.29
71.91
N/A
265.40
N/A
Third
Slope
(gr/m/lOOOm)

0.0003
N/A
0.0004
0.0001
N/A
0.0005
N/A
                                        Trucks

Model Year/
Technology
88-93 PFI
88-93 TBI
84-93 CARB
81-87 FI
81-83 CARB
ZML Emission
(gr/m)

0.0291
0.0253
0.1118
0.0594
0.1033
First
Slope
(gr/m/lOOOm)

0.0000
0.0000
0.0000
0.0000
0.0000
First
Corner
(1000 miles)

19.18
16 .25
36.51
29.76
12.35
Second
Slope
(gr/m/lOOOm)

0.0005
0.0004
0.0008
0.0006
0.0002
Second
Corner
(1000 miles)

N/A
54 .46
N/A
N/A
80.35
Third
Slope
(gr/m/lOOOm)

N/A
0.0003
N/A
N/A
0.0001
Methane
11/8/99

-------
   Table 2: Methane Start Emission Deterioration Model Coefficients for Light-Duty Vehicles
                                         Cars

GROUP
88-93 PFI
88-93 TBI
86-93 CARB
83-87 FI
83-85 CARB
81-82 FI
81-82 CARB
ZML
(gr/m)

0.102
0.084
0.108
0.115
0.174
0.077
0.177
SLOPE
(gr/m/lOOOm)

-0.0002
-0.0003
-0.0001
-0.0004
-0.0001
0.0008
0.0007
MEAN
METHANE
HIGHS
(gr/m)

0.178
-0.073
0.242
0.151
0.601
0.335
0.551
MEAN
METHANE
NORMALS
(gr/m)

0.095
0.071
0.105
0.097
0.172
0.116
0.211
                                        Trucks

GROUP
88-93 PFI
88-93 TBI
84-93 CARB
81-87 FI
81-83 CARB
ZML
(gr/m)

0.151
0.148
0.302
0.098
0.423
SLOPE
(gr/m/lOOOm)

-0.0022
-0.0002
-0.0001
-0.0000
-0.0001
MEAN
METHANE
HIGHS
(gr/m)



0.556
0.183
0.686
MEAN
METHANE
NORMALS
(gr/m)

0.050
0.140
0.297
0.098
0.417
Methane
11/8/99

-------
                    Table 3: Fraction of High Emitters for Light-Duty Cars


	 •
MILES
(xlOOO)
2 .142
12.823
29.335
50
60.006
74.239
87.786
100.01
112 .948
124.625
135.738
146 .315
156 .38
165.96
175.077
183 .753
192 .01
199.869
207.349
214 .466
221.241
227.688
233.823
239.663
245.22
250.509

81-82 CARB


0.0282
0.0543
0.1580
0.2906
0.3560
0.4503
0.5416
0.6253
0.7152
0.7976
0.8772
0.9539
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000

81-82 FI
• 	

0.0203
0.0654
0.1613
0.2861
0.3485
0.4393
0.5275
0.6094
0.6986
0.7812
0.8620
0.9407
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000

83-85 CARB
• 	

0.0232
0.0158
0.0047
0.0917
0.1348
0.1972
0.2578
0.3135
0.3737
0.4290
0.4826
0.5345
0.5847
0.6332
0.6801
0.7253
0.7690
0.8111
0.8516
0.8907
0.9284
0.9646
1.0000
1.0000
1.0000
1.0000
GROUP
83-87 FI
• 	

0.0223
0.0157
0.0406
0.1003
0.1298
0.1723
0.2078
0.2346
0.2634
0.2898
0.3153
0.3400
0.3638
0.3868
0.4089
0.4303
0.4508
0.4706
0.4896
0.5079
0.5255
0.5425
0.5587
0.5743
0.5893
0.6036

86-93 CARB
• 	

0.0052
0.0197
0.0526
0.1042
0.1296
0.1661
0.2012
0.2334
0.2678
0.2992
0.3295
0.3586
0.3866
0.4135
0.4393
0.4641
0.4879
0.5108
0.5327
0.5537
0.5738
0.5931
0.6116
0.6293
0.6462
0.6624

88-93 PFI
• 	

0.0184
0.0227
0.0422
0.0800
0.0987
0.1260
0.1525
0.1770
0.2036
0.2280
0.2518
0.2748
0.2972
0.3189
0.3398
0.3601
0.3798
0.3988
0.4171
0.4348
0.4519
0.4683
0.4842
0.4994
0.5141
0.5283

88-93 TBI
• 	

0.0239
0.0251
0.0270
0.0386
0.0458
0.0561
0.0661
0.0753
0.0851
0.0940
0.1026
0.1110
0.1190
0.1267
0.1341
0.1412
0.1480
0.1546
0.1609
0.1669
0.1727
0.1782
0.1836
0.1887
0.1936
0.1982
                                                                                  (Cont.)
Methane
-10-
11/8/99

-------
                   Table 3: Fraction of High Emitters for Light-Duty Trucks

MILES
(xlOOO)
2 .142
12.823
29.335
45.05
60.006
74.239
87.786
100.678
112 .948
124.625
135.738
146 .315
156 .38
165.96
175.077
183 .753
192 .01
199.869
207.349
214 .466
221.241
227.688
233.823
239.663
245.22
250.509
GROUP
81-83 CARB| 81-87 FI | 84-93 CARB | 88-93 PFI


0.0500
0.0405
0.0733
0.1181
0.1637
0.2101
0.2573
0.3052
0.3538
0.4030
0.4528
0.5032
0.5540
0.6052
0.6568
0.7087
0.7608
0.8132
0.8656
0.9181
0.9706
1.0000
1.0000
1.0000
1 . 0000


0.0659
0.0515
0.0283
0.0370
0.0883
0.1393
0.1900
0.2405
0.2905
0.3401
0.3891
0.4376
0.4855
0.5328
0.5793
0.6251
0.6702
0.7144
0.7578
0.8004
0.8420
0.8827
0.9225
0.9614
0.9993


0.0000
0.0000
0.0000
0.0000
0.0164
0.0697
0.1230
0.1761
0.2291
0.2819
0.3343
0.3864
0.4381
0.4892
0.5399
0.5899
0.6393
0.6881
0.7361
0.7833
0.8297
0.8753
0.9200
0.9638
1.0000


0.0301
0.0294
0.0356
0.0546
0.0734
0.0919
0.1102
0.1281
0.1458
0.1631
0.1801
0.1968
0.2131
0.2291
0.2446
0.2599
0.2747
0.2891
0.3032
0.3169
0.3302
0.3431
0.3556
0.3677
0.3795
88-93 TBI


0.0094
0.0066
0.0202
0.0381
0.0555
0.0721
0.0884
0.1043
0.1198
0.1349
0.1497
0.1641
0.1781
0.1918
0.2050
0.2178
0.2303
0.2424
0.2541
0.2655
0.2764
0.2870
0.2973
0.3072
0.3167
1.0000 1.0000 1.0000 0.3909 0.3259
Methane
-10-
11/8/99

-------
 (g/mfl

  0.75
  0.50-
  025
  0.00
                figure 1: METHANE and THC vs. MILEAGE, RUNNING  LA4, PFI CARS
                           50
                     100

                 MILES (xlOOO)
150
Methane
UNADJUSTED THC
UNADJUSTED METHANE

        -10-
                                                      ADJUSTED THC
                                                      ADJUSTED METHANE
200
 11/8/99

-------
          0.0
                             Figure 2(a): METHANE vs. MILEAGE, RUNNING LA4 CARS
                                                   MILES (xlOOO)
Methane
                        81-82 GARB
                        86-93 GARB
-10-
        81-82 Fl
        88-93 PR
83-85 GARB
        1/8/99
83-87 Fl

-------
                    Figure 2(b): METHANE vs. MILEAGE, RUNNING LA* TRUCKS
 (g/mi)

  025


  020


  0.15


  0.10


  0.05


  0.00
      \  '   '  '  '  '  ^
      0
       81 -83 GARB
  i
 50
81-87 Fl
    100
MILES (X1000)
 84-93 GARB
Methane
          -10-
   150


-93 PR
    11/8/99
      200
1-93 TBI

-------
 grams

  0.6
                   Rgure 3(a): METHANE vs. MILEAGE, START EMISSIONS, CARS
  0.4
  0.0
                         50
            100

          MILEAGE
              150
                   200
                81 -82 GARB
                86-93 GARB
Methane
81-82 R
88-93 PR
-10-
83-85 GARB
88-93 TH
 • 83-87 R


11/8/99

-------
 grams

  0.75
  0.50
  0.25
                 Rgure 3(b): METHANE vs.  MILEAGE, START EMISSIONS, TRUCKS
  0.00
       81-83 GARB
Methane
                           i
                          50
81-87 Fl
   100
 MILEAGE
84-93 GARB
         -10-
     150


88-93 PR
      11/8/99
                                                              200
88-93 TBI

-------
 0.4
                        Figure 4(a): MOBILES and MOBILES METHANE DETERIORATION
                              CARBURETED CAR RUNNING EMISSIONS (g/mQ
 0.3
 0.2
       MY

Methane
                                             MILES 0(1000)
M51975-79
M5PRE-1975
 M51980
 M61981-82
10-
M5:198l-82
M61983-85
M51983+
                                                       11/8/99

-------
 0.2
 0.0
       MY

Methane
                        Figure 4(b): MOBILES and MOBILES METHANE DETERIORATION
                               CARBURETED CAR START EMISSIONS (grams)
                           50
M51975-79
M5PRE-1975
                               100
                           MILES (XIOOO)
                                  150
 M5:1980
 M6:1981-82
10-
M5:19B1-82
M6:1983-85
                                    200
M5:1983+
                                                         11/8/99

-------
Methane                                 -10-                                     11/8/99

-------