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
------- |