vvEPA
United States
Environmental Protection
Agency
Office of Transportation EPA420-D-06-001
and Air Quality February 2006
Cold Temperature Effects on
Vehicle HC Emissions
Draft Report
-------
EPA420-D-06-001
February 2006
on HC
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
This Technical Report not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of issues using that are currently available.
The purpose in the of such reports is to facilitate an exchange of technical
information and to inform the public of technical developments which may form the
for a final EPA decision, position, or regulatory action.
-------
Background
The EPA highway mobile source emission factor model, MOBILE6 (MOBILE6.2.03),
applies a multiplicative adjustment to the base (75 degree Fahrenheit) HC emission rates for
engine starts from all 1981 and later model year light duty cars and trucks to account for the
effects of different temperatures. As a result, as the emissions from vehicles have been reduced
as a result of more stringent emission certification standards, MOBILE6 predicts a corresponding
proportional reduction in the increase in emissions resulting from engine starts at temperatures
below 75 degrees Fahrenheit.
In recent years, vehicle manufacturers have provided FTP test procedure results from
testing at 20 degrees Fahrenheit in compliance with requirements for certification for cold
temperature CO standards. The HC results included in some of these submissions demonstrate
clearly that the effect of more stringent standards at 75 degrees has not resulted in a proportional
reduction in engine start emissions at colder temperatures.
This analysis was done to evaluate an initial assessment of the impact of this new
information on expected HC emission rates at low temperatures for the most recent model year
vehicles. A more thorough analysis than we have made in this report will be needed in order to
fully account for this information in inventories based on the MOBILE6 model.
Data
Cold temperature measurements at 20 degrees Fahrenheit using the FTP testing
procedure are provided by the manufacturers as part of the cold temperature carbon monoxide
(CO) emissions certification procedure. In some cases, the data provided by the manufacturers
also includes measurements of total hydrocarbons (HC) as well as CO. The data was obtained
for this evaluation from the EPA Certification and Fuel Economy Information System1 (CFEIS)
for model years 2000 through 2006 in yearly spreadsheets. Each year contains the certification
emission class and the deteriorated 75 degree and 4k 20 degree Fahrenheit FTP HC results.
Certification for California is also done at 50 degrees Fahrenheit. Test results from
vehicles certified to the California standards were extracted from the CFEIS data for the 2005
model year. The data includes NMOG levels reported at 50 and 75 degrees Fahrenheit for
vehicles certified to the Low Emitting Vehicle (LEV), Ultra-Low Emitting Vehicle ULEV and
LEV2 standards. The values are from the undeteriorated reported results at both temperatures.
Expected emission trends at zero degrees Fahrenheit were also collected in two separate
EPA test programs. The most recent test program2 at Southwest Research Incorporated (SwRI)
comprised of four Tier 2 vehicles of different configurations (small car, minivan, pickup and
crossover). The testing done at the National Exposure Research Laboratory3 (NERL) included
two Tier 1 vehicles and two vehicles certified to LEV standards.
- 1 -
-------
Analysis
Not all engine families tested at cold temperatures included reported measurements of
HC emissions and not all engines are certified for sale in California, which would include
emissions testing at 50 degrees Fahrenheit. For this analysis, all available test results were
weighted equally, rather than devising a scheme to weight results based on projected or actual
engine sales, since not all engine families can be represented in any scheme. The variability of
the results does not suggest that just using the available vehicles will significantly affect the
averages we calculate.
The emission test results used in this analysis are all measurements using the emission
certification Federal Test Procedure (FTP). This procedure measures emissions in three "bags"
which represent different portions of typical driving. The first and third bags use the identical
driving schedules and include an engine start (key on), but the first bag is measured after a 12
hour "soak" (key off period) and the third bag is measured after a 10 minute soak.
Most of the CFEIS data reported to EPA is not separated by bag. The available testing
results with bag measurements show clearly that nearly all of the extra emissions over the
measurements at 75 degrees versus 20 degrees in the FTP test procedure come from the cold start
(Bag 1). There is comparatively little change in the emissions from either Bag 2 or Bag 3. This
suggests that once the engine becomes fully warmed, the ambient temperature is less significant.
For this analysis, we assumed that all of the change in FTP emissions from 75 degrees to 20
degrees Fahrenheit is a result of an increase in Bag 1 emissions. All of the increase in Bag 1 can
be attributed to the engine start, since the emissions that occur once the engine becomes warm
are not significantly affected by the change in temperature. The calculation of grams per engine
start increase from 75 to 20 degrees Fahrenheit from the FTP grams per mile then becomes:
Additional grams per cold start at 20 degrees = (FTP20-FTP75)*7.5/0.43
Where :
FTP20 is the FTP grams per mile at 20 degrees Fahrenheit.
FTP75 is the FTP grams per mile at 75 degrees Fahrenheit.
The nominal length of the FTP driving schedule is 7.5 miles.
The weighting of cold start trips versus hot start trips in the FTP is 0.43.
The resulting grams represent the extra emissions emitted by a vehicle after a 12 hour
soak at 20 degrees over emissions emitted at 75 degrees. All of these emissions are assumed to
occur during the engine start. For emission factor modeling, the added HC emissions would be
added to the estimate of engine start emissions at 75 degrees to model the impact of temperature
on engine start emissions.
This is significantly different from the methods used in MOBILES4 and MOBILE65,
where exhaust emissions for HC are increased using a multiplicative adjustment as a function of
-2-
-------
temperature. However, the temperature adjustments for CO emissions already use an additive
factor for temperatures below 75 degrees Fahrenheit. Modeling temperature adjustments as an
additive adjustment makes the increase in emissions attributed to changes in temperature
independent of other changes (i.e., gasoline sulfur content) made to the base emission rate at 75
degrees.
Results at 20 Degrees Fahrenheit
Matched sets of FTP emission results, both at 20 degrees and 75 degrees Fahrenheit for
the same vehicle engine family (test group), were used in the analysis. The average HC start
emissions were determined separately for vehicles (passenger cars and light trucks combined)
certified to the Tier 1 standards, the transitional low emission vehicle (TLEV) standards, the low
emission vehicle (LEV) standards, the ultra-low emission vehicle (ULEV) standards and Tier 2
standards.
The HC emissions data reported by manufacturers at 20 degrees Fahrenheit is always in
units of total hydrocarbons (THC), reflecting use of the flame ionization detection (FID) method
of hydrocarbon measurement. The HC emission measurements at 75 degrees Fahrenheit were
usually adjusted to reflect the speciation of the hydrocarbons required for emission certification.
Most results were reported in units of non-methane organic gases (NMOG), which includes
ethane and aldehydes, but excludes methane. However, some results at 75 degrees were reported
in units of THC and in units of non-methane hydrocarbons (NMHC), which only excludes
methane. Since the grams of hydrocarbons from engine starts is large, compared to running
emission rates, and mostly NMHC, rather than attempt to adjust all of the data to a common
definition of hydrocarbons, all data was treated as THC for this analysis. Adjustments would
have made only small differences in the average results.
Since Tier 2 standards phase in over a number of years, the average HC start emissions
for Tier 2 vehicles were determined separately for each model year, beginning with the 2004
model year. This more appropriately accounts for the fact that "early" conversions to Tier 2
standards reflect the ability of specific engine families to more easily achieve Tier 2 emission
performance.
FTP HC Data From Federal Certified Vehicles
(grams per mile)
75 Degrees 20 Degrees
Emission Standard Sample Size Mean Std. Dev. Mean Std. Dev.
Tier 1
TLEV
LEV
ULEV
LEV2
2004 Tier 2
2005 Tier 2
2006 Tier 2
410
64
695
132
119
172
190
90
0.1190
0.0804
0.0501
0.0335
0.0296
0.0406
0.0415
0.0408
0.0553
0.0286
0.0209
0.0214
0.0123
0.0169
0.0203
0.0239
0.8630
0.6996
0.6402
0.4675
0.5035
0.5641
0.5651
0.5502
0.7269
0.2778
0.3723
0.2727
0.2549
0.3269
0.3247
0.3107
-------
Results at 50 Degrees Fahrenheit
The following table shows the ratios of the average 50 degree measurements divided by
the average 75 degree measurements. The measurements were extracted from the 2005 CFEIS
data for vehicles certified to California standards. It is based on the HC emission levels in grams
per mile reported at 50 and 75F for LEV, ULEV and LEV2 California certified vehicles. The
ratios are based on the undeteriorated reported results at both temperatures.
FTP HC Emissions Data from California Certified Vehicles
(grams per mile)
Emission
Standard
LEV
ULEV
LEV2
Sample
Size
53
14
21
75 Degrees
Mean
0.0397
0.0162
0.0346
Std.
Dev.
0.0259
0.0043
0.0097
50 Degrees
Mean
0.0988
0.0403
0.0843
Std.
Dev.
0.0631
0.0176
0.0310
Ratio of
Averages
2.49
2.48
2.44
The emission impact of temperature on engine start emissions was estimated for Tier 1
and LEVs by applying this ratio to the average emissions (in grams at 75 degrees) from the
larger sample of federally certified vehicles used in the 20 degree analysis to determine the
increase (in grams) due to temperature at 50 degrees Fahrenheit.
Additional grams per cold start at 50 degrees = ((Ratio-1.0)*FTP75)*7.5/0.43
Tier 1 and TLEV certified vehicles our analysis will assume the LEV ratio, but Tier 1
vehicles may actually perform worse, since there is no 50 degree Fahrenheit test requirement for
Tier 1 vehicles in California.
For Tier 2 vehicles, 25.5% of vehicles were assumed to already be certified to the
California 50 degree standard. This number is based on the 2003 model year new registrations
and accounts for registrations in California and states which require California vehicles. We
assume these vehicles will achieve the 0.0843 gram per mile emission rate achieved by the
California LEV2 vehicles in the sample. The remaining vehicles (74.5%) were assumed to be
unaffected by California standards. The non-California vehicles are assumed to have emissions
at 50 degrees Fahrenheit which are a linear interpolation between their 75 degree emissions (0.71
grams or 0.041 grams per mile) and their 20 degree emissions (0.56 grams per mile), which was
determined to be 0.277 grams per mile.
Additional grams per cold start for Tier 2 vehicles at 50 degrees =
((0.0843*0.255+0.277*0.745)*7.5/0.43)-0.71 = 3.27 grams
-4-
-------
The adjustment to account for vehicles not certified to California standards was not made
for vehicles standards other than Tier 2. Making the adjustment for all standards would tend to
increase the overall emission effect of temperature at 50 degrees. The following table shows the
effects assumed for each emission standard.
Increase in HC Emissions From California Certified Vehicles at 75
and 50 Degrees Fahrenheit
Standard
Tier 1**
TLEV**
LEV
ULEV
LEV2
Tier 2**
Sample
Size
N/A
N/A
53
14
21
N/A
Ratio of
Averages
2.49*
2.49*
2.49
2.48
2.44
N/A
FTP at
75 Degrees
(grams)
2.08
1.40
0.87
0.58
N/A
0.71
Increase
(grams)
3.09
2.09
1.30
0.87
N/A
3.27
* Ratios assumed from other emission standard groupings.
** No vehicles of this type tested for California standards at 50 degrees.
Results Below 20 Degrees Fahrenheit
The effects of temperatures below 20 degrees Fahrenheit is taken from a small sample of
vehicles tested at temperatures down to zero degrees Fahrenheit. Two Tier 1 vehicles and two
NLEV vehicles were tested in the NERL Program and four Tier 2 vehicles were tested in the
SWRI Program.
Increase
in Vehicle HC
Emissions From 20 to Zero Degrees
(FTP grams per mile)
20 Degrees
Vehicle
Tier 1 /
Tier 2
All
Sample
LEV
Sample Size
4
4
8
Mean at 20
degrees
0.7375
0.3593
0.5484
std dev
0.2455
0.1783
0.2834
Fahrenheit
Zero Degrees
Mean at zero
degrees
1.2500
0.7958
1.0229
std dev
0
0
0
3909
4730
4694
Ratio of
Means
1
2
1
69
22
87
-5-
-------
The increase in the average FTP HC emissions from these vehicles at zero degrees ranged
from 1.7 to 2.2 times the emissions at 20 degrees. The sample sizes are very small but the
increase is reasonably consistent across all of the vehicles in the samples. For this analysis, the
increase in engine start emissions at zero degrees will be assumed to be twice the emission
increase measured at 20 degrees Fahrenheit for all of the vehicle technologies (Tier 1, LEV and
Tier 2).
Results for High Emitting Vehicles
A significant amount of HC inventory emissions comes from vehicles referred to as
"high" emitters. These vehicles are not operating properly and may have emission levels many
times that of normally emitting vehicles. None of the low temperature data available included
the effects of engine starts on vehicles which grossly exceed their emission certification
standards (high emitters). The assumptions about the temperature effects on HC emissions for
Tier 1 and LEV high emitting vehicles used in MOBILE6 were not changed for this analysis.
However, since any proposed control strategies will affect Tier 2 vehicles, the effect of cold
temperatures on the HC emissions from Tier 2 high emitters was estimated as well.
The default MOBILE6 additive impact on high emitter engine start emission rates for
Tier 2 vehicles was made for 20 degrees Fahrenheit by running the model. It was assumed that,
in light of the higher emission impact at cold temperatures for normal emitting vehicles, the
emission impact for high emitters would be increased as well. Twice the impact of low
temperatures on normal emitting vehicles was added to the base (MOBILE6 default) temperature
impact to represent the overall impact of temperature on the HC emissions of high emitting
vehicles at low temperatures. The emission impact at zero degrees was assumed to be twice the
impact at 20 degrees and the impact at 50 degrees was assumed to be half the impact at 20
degrees. The following table shows the engine start emission impacts used for high emitting Tier
2 vehicles.
Tier 2 High Emitter HC Adjustment
Based on 2005 Model Year MOBILE6 Results in Calendar Year 2005
Temperature Degrees Fahrenheit
Default MOBILE6 grams
Normal Emitter Adjustment grams
Total High Emitter Adjustment grams
0
62.68
18.26
99.20
20
31.34
9.13
49.60
50
15.67
3.27
22.21
-6-
-------
Final Temperature Effects
The following tables show a summary of the results of the analysis. Each values
represents the average increase in engine start HC emissions in grams per engine start from a
base emission rate of a 12 hour soak before key on at 75 degrees Fahrenheit for Tier 1, LEV,
ULEV and Tier 2 standard vehicles. This value is added to the base engine start emission rate to
account for the impact of the colder temperature on the total HC emissions from an engine start.
This adjustment would completely replace the default temperature adjustment used by
MOBILE6 for these vehicles.
Increase in Engine Start Hydrocarbon Emissions
Over the 75 Degree Fahrenheit Baseline at Low Temperatures
(grams per engine start after 12 hours of soak)
Description
Tier 1
Transitional Low Emission Vehicle
Low Emission Vehicle (LEV)
Ultra Low Emission Vehicle (ULEV)
Tier 2 Model Year 2004
Tier 2 Model Year 2005
Tier 2 Model Year 2006 and Later
Tier 2 High Emitting Vehicles
Degrees Fahrenheit
0
25.96
21.60
20.59
15.14
18.26
18.27
17.77
99.20
20
12.98
10.80
10.29
7.57
9.13
9.13
8.88
49.60
50
3.09
2.09
1.30
0.87
3.27
3.27
3.27
17.71
Comparison with the default MOBILE6 case is difficult, since the multiplicative
approach used by MOBILE6 makes the effect of temperature (in grams) vary significantly
depending on the base emission rate on which it is applied. The default multiplicative
temperature correction factor equation for bag 1 (cold start) HC emissions for 1992 and later
model years is:
TCF(l) = EXP [ TC(1)*(T - 75.0) ] = 3.7
Where:
TCF(l) is the multiplicative temperature correction factor for bag 1 HC emissions.
TC(1) is the coefficient for segment 1 (bag 1) for HC emissions (-0.023768)
T is the ambient temperature of interest in degrees Fahrenheit. The base emission
rate is at 75 degrees.
-7-
-------
When compared to the base engine start emission rates at 75 degrees Fahrenheit
estimated by the MOBILE6 model6, the new estimates are very large increases in emissions.
Tier 2 (Bin 5) engine start emissions at zero miles is 0.8 grams (10.752*0.075, NMOG) per
engine start at 75 degrees. Adding 9.13 grams to adjust the engine start emissions to 20 degrees
will increase these engine start emissions by a factor of 11, if a multiplicative factor were to be
used. This is a much larger factor than suggested by the default MOBILE6 temperature
correction factors. The factors for the Tier 1 and LEVs are smaller, which is consistent with the
observation that the reduction in HC emissions at lower temperatures (primarily engine start
emissions) has not been proportional to the decrease in overall vehicle HC emissions at 75
degrees due to the new lower HC emission standards.
Results
An analysis was performed using the new estimates for the effects of cold temperatures
on HC emissions by substituting the new estimates for the default temperature adjustments in the
MOBILE6 highway emission factor model. MOBILE6 estimates emission rates by starting with
a base emission rate measured at base conditions, fuels and operation and then uses correction
factors to adjust the base emission rate to better reflect the conditions, fuels and vehicle operation
designated by the user. The calculation of the default temperature and soak time adjustments to
HC engine start emissions in MOBILE6 is very complex. The default MOBILE6 adjustment
factors are skipped and the new adjustments are applied whenever the temperature to be modeled
falls below 75 degrees.
The additional engine start emissions for a given temperature is the value linearly
interpolated between 50 and 75 degrees for temperatures in this range, linearly interpolated
between 20 and 50 degrees for temperatures in this range or linearly interpolated between zero
and 20 degrees for temperatures in this range. MOBILE6 will not model temperatures below
zero degrees Fahrenheit. Emission estimates for temperatures above 75 degrees are not affected.
No changes were made in the method used to adjust the engine start emissions to reflect differing
engine soak times.
To summarize, the effects of these changes in MOBILE6:
Affects only HC emissions.
Affects only gasoline fueled engines.
Affects only light duty cars and trucks.
Only affects engine start emissions at temperatures less than 75 degrees Fahrenheit.
Affects only vehicles certified to Tier 1, LEV and Tier 2 standards (1994 and later
model years).
Varies by model year for Tier 2 vehicles.
Affects only the Tier 2 high emitting vehicles.
The following chart shows the impact of these new estimates for the effect of cold
temperatures on HC engine start emissions on the fleet wide emissions of gasoline fueled
passenger cars by calendar year. The default MOBILE6 adjustment for temperature already
shows that exhaust emissions (running and engine start combined) at 20 degrees Fahrenheit are
-------
much higher than at 75 degrees. Once the new estimates are applied beginning with the 1994
model year, the average exhaust emissions of the fleet (all model years) begins to diverge from
the default values, and show less improvement in emissions due to the introduction of new
exhaust emission standards. By calendar year 2010, the fleet emissions at 20 degrees are much
less improved than predicted by MOBILE6.
Effects of New HC Engine Start LDGV Emission Rates at Cold Temperatures on
Composite Exhaust Engine Start and Running HC Emissions
\
9
\
- - New Rates 20 Degrees
Standard M6 20
Degrees
^Standard M6 75
Degrees
o
o
1990
1992
1994
1996
1998
2000
Calendar Year
2002
2004
2006
2008
2010
This chart includes only passenger cars. If we include diesel fueled vehicles and heavy
duty vehicles and assume that their engine start emission estimates will not need an adjustment,
the overall impact of these changes on the fleet wide inventory, including all highway vehicle
types, will be less. However, since gasoline fueled, light duty vehicles are such a large part of
the fleet, the impact on inventory estimates at low temperatures will still be large. This impact is
mitigated by the fact that the additional HC emissions we have estimated occurs at low
temperatures which are not as important for ozone formation as HC increases during hot days
and the fact that many areas do not experience extremely cold temperatures, even in winter. A
more detailed analysis of these changes in the estimates for the effects of cold temperatures on
exhaust HC emissions would need to account for the variation in temperatures across the nation
and include all sources of HC, such as other vehicle classes and evaporative HC. Any increases
in HC emission inventory estimates, even at just at low temperatures, will still be of concern for
other health related reasons.
-9-
-------
References
1. Certification and Fuel Economy Information System (CFEIS).
http://www.epa.gov/omswww/cfeis.htm
2. "VOC/PM Cold Temperature Characterization and Interior Climate Control
Emissions/Fuel Economy Impact," Draft Final Report Volume I, EPA Contract 68-C-05-
018, Work Assignment No. 0-4, SwRI Project No. 03.11382.04, Alan P. Stanard, October
2005.
3. "Characterization of Emissions from Malfunctioning Vehicles Fueled with Oxygenated
Gasoline," (Parts I, II & III) Fred Stump, Silvestre Tejada, and David Dropkin, National
Exposure Research Laboratory, U. S. Environmental Protection Agency and Colleen
Loomis, Clean Air Vehicle Technology Center, Inc.
4. AP-42: Compilation of Air Pollutant Emission Factors, Appendix H: Highway Mobile
Source Emission Factor Tables, http://www.epa.gov/otaq/ap42.htm
5. "Exhaust Emission Temperature Correction Factors for MOBILE6: Engine Start and
Running LA4 Emissions for Gasoline Vehicles," M6.STE.004 (EPA420-R-01-029, April
2001). http://www.epa.gov/otaq/models/mobile6/m6tech.htm
6. "Determination of NOx and HC Basic Emission Rates, OBD and I/M Effects for Tier 1
and Later LDVs and LDTs," M6.EXH.007 (EPA420-R-01-056, November 2001).
http://www.epa.gov/otaq/models/mobile6/m6tech.htm
- 10-
------- |