Use of Data from "Development of
Emission Rates for the MOVES Model,"
Sierra Research, March 3, 2010
&EPA
United States
Environmental Protection
Agency
-------
Use of Data from "Development of
Emission Rates for the MOVES Model,"
Sierra Research, March 3, 2010
Assessment and Standards Division
Office of Transportation and Air Quality
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
that 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.
&EPA
United States
Environmental Protection
Agency
EPA-420-R-12-022
August 2012
-------
10* "Gilt" SO; C? m f%"i^ 1S I it f J> f% 1 S ^ % ^ 1 ^si \» 1 4 % J™l ^S I
icji i i 11 I*H rv <
-------
MOVES2010b, nitrous acid (HONO) was added to the model. This required a change in the
NO2 and NO fractions. The HONO fraction was set to 0.008 for all records.l The HONO
fraction was subtracted from the MOVES2010a NO2 fraction to create a new NO2 fraction such
that the fractions still sum to one across the three pollutants.
The Sierra Research motorcycle emission rates were used to populate the exhaust and
evaporative emission factor tables in MOVES2010, MOVES2010a and MOVES2010b.
The Sierra Research data was used to populate the EmissionRateByAge and the
CrankcaseEmission Ratio tables for motorcycle exhaust emissions of total hydrocarbons (THC),
carbon monoxide (CO), oxides of nitrogen (NOx), elemental carbon (EC) and organic carbon
(OC). While motorcycle exhaust emissions are in the EmissionRateByAge table, they are
constant across all ages. For evaporative emissions, the Sierra data was used to populate the
EmissionRateByAge and the CumTVVCoeff table. The vapor venting rates do vary with age for
some modelyear groups.
The Sierra report did not address emissions of methane (CH4), nitrous oxide (N2O),
tirewear, brakewear or energy consumption rates for motorcycles. These rates were populated
in MOVES based on other data sources.
Motorcycle default population data is stored in the MOVES sourcetypeYear table. This
table also includes sales growth rates which are used to grow the population from the base year
to the analysis year, and an unused field for "migration rates." In MOVES2010, MOVES2010a
and MOVES2010b, the base year motorcycle populations (1990 and 1999) come from FHWA
registration data.2 In MOVES2010a and MOVES2010b, the MOVES default database
incorporates the salesgrowth factors indicated in the Sierra Report, including their forecasts for
future motorcycle sales growth rates. b
Sierra also provided motorcycle migration rates. However, the migrationrate field in
MOVES is not used. The Sierra-provided rate was stored in the MOVES2010 default database;
however, for clarity, it was replaced with values of'T'in the MOVES2010a and MOVES2010b
databases.
records. This created erroneous NO and NO2 results in the MOVES2010 output. As
documented in the MOVES2010 Errata/Information Sheet, the MOVES2010 error was corrected
in MOVES2010a and in the intermediate 20100512 MOVES database.
b In MOVES2010, the sales growth factors were inadvertently unchanged from Draft MOVES2009, where
they had been set equal to passenger car data.
a-2
-------
In MOVES, the survival fraction is multiplied by the population of vehicles of a given
age to calculate the population at the next age, thus it indicates the fraction of vehicles that
remain from one year to the next. However, the motorcycle survival rate documented in the
Sierra report and provided to EPA was the fraction of the initial population that remains at the
end of each year.
In MOVES2010, the Sierra survival values were input directly. This overestimated
scrappage and led to smaller motorcycle populations than expected. For MOVES2010a (and
201 Ob), we revised these rates. We calculated the revised scrappage rate in the format needed
for MOVES using following equation: New scrappage rate (age i) = Original scrappage rate
(age i)/Original scrappage rate (age i-1). Because the initial value was developed from a fitted
curve, the revised value was a constant of 0.940 for all ages 2 and greater.
In MOVES, the relative mileage accumulation rate is a normalized measure of the
average number of miles driven by vehicles of each age. The relativeMARS for motorcycles
provided by Sierra were used in MOVES2010, MOVES2010a and MOVES2010b.
Sierra calculated a motorcycle age distribution based on data for calendar year 2008. For
MOVES2010, MOVES2010a and MOVES2010b, this was used as the 1999 base-year age
distribution in the SourceTypeAgeDistribution table.
MOVES2010, MOVES2010a and MOVES2010b used the Sierra-calculated base year
VMT and VMT growth factors for motorcycles for years through 2008. The 2009-and-later
VMT growth rates were based on AEO2006 projections of VMT growth for cars.
MOVES2010, MOVES2010a and MOVES2010b use the Sierra-provided motorcycle
information to populate the tables SampleVehicleDay and SampleVehicleTrip. Note the
spreadsheet MOVES_MC_SampleVehicleDay.csv lacks seven vehicleids that are in MOVES
and are listed in MOVES_MC_SampleVehicleTrip.csv.
The text of the Sierra report discusses differences between weekend and weekday activity
for motorcycles, as well as seasonal differences. This information was used to update the daily
trip pattern information (SampleVehicleTrip and SampleVehicleDay), but it was not used to
update the values in MonthVMTFraction or DayVMTFraction. The MOVES VMT fractions by
month and daytype are the same for all sourcetypes3.
As recommended in the Sierra report, and in absence of better data, MOVES uses driving
cycles for passenger cars to represent driving cycles for motorcycles.
a-3
-------
MOVES2010 and MOVES2010a estimate emissions for a short list of mobile source air
toxics. MOVES2010b models many more "hazardous air pollutants" (HAPs). Most of these
emission rates were developed by EPA staff from available data. 4 The Sierra report was used
for the following emission rates for all vehicle types:
1) Exhaust gaseous HAPs for pre-Tier 2 vehicles running on gasoline (E0) and gasohol
(Eio).
2) Gas and particle phase polycyclic aromatic hydrocarbon (PAH) allocation factors (but
not actual emission rates).
3) Toxics ratios for pre-2007 diesels for 2,2,4-trimethylpentane, hexane,
propionaldehyde, and toluene (but not other gaseous HAPs).
1 Kurtenbach, R., Becker, K. H., Gomes, J.A.G., Kleffmann, I, Lorzer, J. C., Spittler, M, Wiesen, P.,
Ackermann, R., Geyer, A., Platt, U., 2001. Investigations of emissions and heterogeneous formation of HONO in a
road traffic tunnel. Atmospheric Environment 35, 3385-3394.
2U.S. EPA, 2010, "MOVES2010 Highway Vehicle Population and Activity Data," EPA-420-R-10-026,
November 2010, http://www.epa.gov/otaq/models/moves/420r 10026.pdf
3 EPA-420-R-10-026
4 U.S. EPA, 2012, "MOVES 2010b: Additional Toxics Added to MOVES," EPA-420-B-12-029a, May
2012, http://www.epa.gov/otaq/models/moves/documents/420bl2029a.pdf
a-4
-------
Development of Emission Rates
for the MOVES Model
prepared for:
U.S. Environmental Protection Agency
March 3, 2010
prepared by:
Sierra Research, Inc.
1801 J Street
Sacramento, California 95811
(916)444-6666
-------
DEVELOPMENT OF EMISSION RATES FOR THE MOVES
MODEL
prepared for:
U.S. Environmental Protection Agency
March 3, 2010
Principal authors:
Thomas R. Carlson
Thomas C. Austin
Dennis McClement
Sierra Research, Inc.
1801 J Street
Sacramento, C A 95811
(916)444-6666
-------
DEVELOPMENT OF EMISSION RATES
FOR THE MOVES MODEL
Table of Contents
1. INTRODUCTION 1
1.1 Background 1
1.2 Project Summary 2
1.3 Organization of the Report 4
2. NITROGEN-AND SULFUR-BASED EMISSION RATES 5
2.1 Ammonia Emission Rates 5
2.2 Sulfur-Based Emission Rates 14
2.3 Nitrogen Oxide and Nitrogen Dioxide Ratios 17
3. HIGHWAY MOTORCYCLE EMISSION RATES AND ACTIVITY FACTORS .21
3.1 Highway Motorcycle Characteristics 21
3.2 Uncontrolled Emissions Characteristics 24
3.3 Controlled Emissions 34
3.4 Inspection and Maintenance Effect 47
3.5 Motorcycle Populations and Activity 49
3.6 Incorporation into MOVES Tables 60
4. AIR TOXIC SPECIATION RATIOS 72
4.1 MOVES Speciation Ratios 72
4.2 Light-Duty Gasoline-Fueled Ratios 75
4.3 Heavy-Duty Diesel-Fueled Ratios 76
5. REFERENCES 82
-i-
-------
List of Figures
Figure Page
Figure 2-1 Ammonia Emissions (g/hr) vs. Operating Mode by Model Year Group 9
Figure 2-2 Exhaust SO4 vs. Fuel S- 1981 and Newer 15
Figure 2-3 Exhaust SO4 vs. Fuel S- 1980 and Older 16
Figure 3-1 Exhaust Hydrocarbons vs. Odometer Pre-1978 4-Stroke Models 26
Figure 3-2 Exhaust Carbon Monoxide vs. Odometer Pre-1978 4-Stroke Models 26
Figure 3-3 Exhaust Oxides of Nitrogen vs. Odometer Pre-1978 4-Stroke Models 27
Figure 3-4 Exhaust Hydrocarbons vs. Odometer Pre-1978 2-Stroke Models 28
Figure 3-5 Exhaust Carbon Monoxide vs. Odometer Pre-1978 2-Stroke Models 28
Figure 3-6 Exhaust Oxides of Nitrogen vs. Odometer Pre-1978 2-Stroke Models 29
Figure 3-7 Exhaust Hydrocarbons vs. Odometer 1978-1985 4-Stroke Models 36
Figure 3-8 Exhaust Carbon Monoxide vs. Odometer 1978-1985 4-Stroke Models 37
Figure 3-9 Exhaust Oxides of Nitrogen vs. Odometer 1978-1985 4-Stroke Models 37
Figure 3-10 Exhaust Hydrocarbons vs. Odometer 1986-1999 Models 38
Figure 3-11 Exhaust Carbon Monoxide vs. Odometer 1986-1999 Models 39
Figure 3-12 Exhaust Oxides of Nitrogen vs. Odometer 1986-1999 Models 39
Figure 3-13 Idle HC Versus Motorcycle Age Arizona I/M Data - July to
September 1997 40
Figure 3-14 Idle CO Versus Motorcycle Age Arizona I/M Data - July to
September 1997 41
Figure 3-15 Motorcycle Survival Fraction vs. Age Based on Analysis of
Registration and Retail Sales Data 52
Figure 3-16 Average Annual Lifetime Mileage vs. Age 53
Figure 3-17 World Motorcycle Test Driving Cycle SAFD (%) 58
Figure 3-18 California Urban & Rural Weighted (CalWtd) Driving Cycle SAFD
...59
-11-
-------
List of Tables
Table Page
Table 2-1 CE-CERT Ammonia Test Program Vehicle Counts 6
Table 2-2 VSP Scaling Step 1 8
Table 2-3 VSP Scaling MY 1996-2001 8
Table 2-4 VSP Scaling MY 1992-1995 8
Table 2-5 VSP Scaling MY 1981-1991 9
Table 2-6 Fleet Average Ammonia Emissions (mg/mi) 11
Table 2-7 Extended Fleet Average Ammonia Emissions (mg/mi) 12
Table 2-8 Ammonia Emissions for Diesel and Heavy Duty Vehicles (mg/mi) 13
Table 2-9 Gasoline NONO2 Ratios 18
Table 2-10 Diesel Without DPF trap NO/NO2 Ratios 19
Table 2-11 Diesel With DPF trap NONO2 Ratios 20
Table 2-12 Diesel With DPF trap and Extended IdleNONO2 Ratios 20
Table 3-1 Uncontrolled Highway Motorcycle Evaporative Emissions 31
Table 3-2 Uncontrolled Highway Motorcycle Emissions (g/mi) 33
Table 3-3 Motorcycle Exhaust Emission Standards (grams per kilometer, g/km) 35
Table 3-4 Controlled Highway Motorcycle Evaporative Emissions 44
Table 3-5 Controlled vs. Uncontrolled Emissions (g/mi) 47
Table 3-6 On-Highway Motorcycle Population and Survival Estimated from
Registration and Retail Sales Data 51
Table 3-7 Motorcycle Annual VMT vs. Age (years) 55
Table 3-8 Portion of Annual Operation by Season 56
Table 3-9 Normalized Motorcycle Running Exhaust Emission Rates (g/hr per
gram Hot FTP) by Operating Mode Bin 64
Table 3-10 MOVES Evaporative Processes by Pollutant/Process and OpMode 67
-in-
-------
Table 4-1 List of New MOVES HAPs and Numerical Codes 73
Table 4-2 Tier 1 and Earlier Light-Duty Gasoline Vehicle Speciation Ratios 75
Table 4-3 Tier 2 Light-Duty Gasoline Vehicle Speciation Ratios 76
Table 4-4 Study Data Represented in CRC-75 Speciation Database 77
Table 4-5 Mult-Phase PAH Allocation Factors Used 80
Table 4-6 Heavy-Duty Diesel Speciation Ratios and Basic Statistics 81
-IV-
-------
1. INTRODUCTION
This introduction provides a review of the background behind the effort, a brief summary
of the key issues and findings, and the organization of the remainder of the report.
1.1 Background
The successor to MOBILE6, MOtor Vehicle Emissions .Simulator (MOVES), was
released in April 2009 in draft form without certain emission rates included. Emission
rates that need to be added include certain nitrogen compounds and toxic air pollutants
from cars and trucks and all pollutants and activity rates from highway motorcycles. For
some of the missing emissions rates, EPA has data that needed to be analyzed and
processed into a form which is appropriate for MOVES. In other areas, additional data
needed to be identified. The purpose of this effort was to identify, analyze, and process
available data to create estimated emission rates for the above-described pollutants
currently missing from MOVES.
Highest priority was assigned to the development of sulfur dioxide (802), ammonia
(NHs), nitrogen dioxide (NO2), and nitrogen oxide (NO) emissions (orNOx fractions) for
all vehicle categories. Currently, the NMIM and/or MOBILE6.2 / PARTS models
contain estimates for NO, NO2 and SO2 pollutants and processes such as running and
start, and in some cases, crank-case and extended idle. Unfortunately, most of these
estimates were out of date (some date back to the 1970s), and needed to be updated or at
least reviewed.
Second priority was assigned to the development of motorcycle emission and fleet
activity rates. Rather than simply carrying forward motorcycle emissions and activity
data from MOBILE6 (that were originally developed in MOBILES), EPA left
-placeholders" in the Draft MOVES2009 model and underlying database structure until
more up-to-date emissions and activity data could be analyzed and incorporated into
MOVES.
Lower priority was assigned to the development of revised emission factors for an
expanded set of mobile source air toxics compounds. Draft MOVES2009 relied on
algorithms carried over from MOBILE6.2 and implemented these air toxic emission
factors as speciation ratios relative to VOC. These algorithms were derived from U. S.
EPA's Complex Model for Reformulated Gasoline, developed in the early 1990s for the
-1-
-------
Federal Reformulated Gasoline Rule. Also, Draft MOVES2009 included emission rates
for only the following seven air toxic pollutants:
• Benzene,
• 1,3-Butadiene,
• Formaldehyde,
• Acetaldehyde,
• Acrolein,
• Naphthalene and
• Ethanol.
This included the addition of toxic speciation ratios for vehicles running on E-85 fuel.
None of the long list of other toxics included in EPA's National Mobile Inventory Model
(NMEVI)—designated as Additional Hazardous Air Pollutants (HAPS)* throughout the
remainder of this report—were inserted into Draft MOVES2009. Thus, assembly of data
and development of toxic speciation ratios for these Additional HAPS was the focus of
this element of the study.
1.2 Project Summary
Under Contract No. EP-C-05-037, Work Assignments 3-03 and 4-03, Sierra Research,
Inc. (Sierra) was contracted to assist EPA in the assembly and development of these
emission rates and activity data that had not been included in Draft MOVES2009 but
would be implemented in a future release of the model.
The effort was divided into three analytical tasks as follows:
1. Nitrogenous and Sulfur Pollutant Rates;
2. Highway Motorcycle Emission and Activity Rates; and
3. Additional Air Toxic Speciation Ratios.
Data Sources - EPA provided Sierra with a list of existing studies and associated
databases that the agency had assembled to support the development of MOVES
emission rates and air toxic speciation ratios in each of these three task areas. In addition
to these sources, Sierra performed additional literature searches. Key additional sources
(which are cited later in the report sections where they were used) included the following:
• Unpublished emission test data (FTP, idle and 50 mph cruise measurements)
obtained from the California Air Resources Board (CARB) for a small sample of
* Table A-l from the Work Assignment contains the complete list of Additional HAPS to be considered
under this task. (The exceptions on the list that are not considered Additional HAPS for the purposes of
this effort are HC, CO, NOx, SO2, CO2, SOA, PM10, PM25, NH3, acetaldehyde, acrolein, benzene,
formaldehyde, 1,3 butadiene, andMTBE.)
-2-
-------
newer model heavy-duty Diesel vehicles with selective catalyst reduction (SCR)
and Diesel particulate filter (DPF) after-treatment systems;
Modal (second-by-second) motorcycle exhaust emissions test data from a 2000
testing program sponsored by CARB;
Nationwide on-road motorcycle registrations for calendar year 2008 compiled by
R.L. Polk and obtained through the Motorcycle Industry Council (MIC); and
Current highway motorcycle activity and usage data obtained from a recently
completed motorcycle owner survey.
Key Issues/Findings - During the course of this effort, several key issues arose in
evaluating data sources and assembling the required MOVES emission rates. These key
issues and findings are summarized below and described in greater detail in the body of
the report.
• Factors Affecting Ammonia Emission Rates - Prior studies and test measurements
show a clear inverse relationship between ammonia emission rates and NOx
emission controls. Catalytic NOx reduction implemented on vehicles beginning
in the early 1980s led to increases in ammonia emission rates up until phase in of
NLEV and Tier 2 emission standards.
• Highway Motorcycle Tampering Effects - In-use emission rates for highway
motorcycles are significantly affected by the assumed or estimated level of
tampering. Survey and test data indicate that OEM catalyst removal and use of
aftermarket exhaust systems are most significant in affecting in-use motorcycle
emissions. These types of tampering often occur soon after a motorcycle is sold.
Thus, all of the tampering/deterioration effects for motorcycles were assumed to
occur within the newest vehicle age group (0-3 years) defined in MOVES.
• Limited Motorcycle Evaporative Data - Although exhaust emission data were
available for roughly 170 individual motorcycle tests, evaporative emission
measurements were more limited. One-hour SHED measurements were available
for about 35 vehicles and only a handful of real-time 24-hour evaporative tests
were available. This necessitated conversion of the one-hour SHED data to an
equivalent 24-hour estimate in order to translate the test results into the new
evaporative emission processes defined and employed in MOVES.
• Processing Required for Air Toxic Speciation Data - Speciation ratios for Diesel-
fueled vehicles were developed from a CRC database for which a number of data
processing and validation steps were necessary (e.g., matching emission tests for
both the toxic compound and the -base" criteria pollutant upon which its ratio was
based).
-3-
-------
Hole-Filling Required to Generate MOVES Rates - Given the highly
disaggregated design of most of the MOVES emission rate tables, a fairly
extensive amount of extrapolation of -hole-filling" was necessary to populate the
data tables, by model year, age group, and operating mode. The report clearly
identifies the extent to which hole-filling was performed and the assumptions that
were made in doing so.
Difficulty in Quantifying Uncertainty - In the MOVES emission rate tables
developed under this effort, Sierra provided estimates of both mean emission rates
as well as calculated uncertainty, expressed as the coefficient of variation (COV).
However, sample sizes were often small and variance within the available data
large so that COVs entered into the MOVES tables were often capped at an upper
limit of 0.5 per EPA's direction.
1.3 Organization of the Report
Following this introductory section, the remainder of the report is organized in a manner
consistent with the three separate analytical tasks in the effort (Tasks 2 through 4).
Section 2 describes the sources considered and assembly, validation, and processing of
various emission test data to develop MOVES emission rate tables for nitrogen- and
sulfur-based pollutants not already represented in the model. Separate sub-sections in
Section 2 present methods for development of ammonia emission rates, sulfate and sulfur
dioxide emission rates, and nitrogen oxide/nitrogen dioxide emission rate ratios.
Section 3 describes the data sources and discusses the step-by-step methodologies used to
develop MOVES exhaust and evaporative emission rates (for all criteria pollutants) and
fleet activity data for motorcycles. It also includes a separate discussion of how these
emission and activity data were tabulated and integrated into the specific table structures
employed in the MOVES database.
Section 4 discusses the data sources, data assembly, and validation methods used to
develop speciation ratios for an additional list of mobile source air toxic compounds
identified by EPA for subsequent implementation in MOVES.
Section 5 contains a list of references cited in the body of the report.
###
-4-
-------
2. NITROGEN- AND SULFUR-BASED EMISSION RATES
In Draft MOVES2009, priority was placed on developing complete sets of underlying
emissions rates for those criteria pollutants and precursors that have significant emissions
from on-road vehicles: hydrocarbons (HC), carbon monoxide (CO), oxides of nitrogen
(NOx) and particulate matter (PM). The first analytical task under this effort consisted of
developing emission rates for three other nitrogenous and sulfur-based pollutants that had
not yet been built into MOVES:
1 . Ammonia
2. Sulfur dioxide and sulfates; and
3. Nitric oxide (NO) and nitrogen dioxide (NO2) ratios.
The data sources and methods used to develop these new emission rates and structure
them into MOVES-ready database tables are presented in the following three sub-
sections.
2.1 Ammonia Emission Rates
Early testing revealed trace ammonia emissions from both gasoline and Diesel-powered
Internal Combustion Engines (ICE). Studies in the 1970s and 1980s concluded that the
levels of ammonia emissions produced by then-current mobile sources were not a cause
for concern, or reason to trigger limiting regulations, but that levels should be monitored
in the future.
The early 1980s introduction of vehicles equipped with three-way catalytic converters
quickly raised concerns regarding ammonia emissions. Many studies revealed increased
levels of ammonia from early examples of such vehicles. Detailed studies revealed an
inverse relationship with NOx control: catalytic reduction of NOx in the exhaust stream
resulted in elevated NH^ levels.
As emissions standards continued to drop and vehicle emission control technology
continued to improve, ammonia emissions fell to nearly the levels observed prior to the
introduction of the three-way catalyst. In addition, as the early 1980s vehicles aged, NOx
emissions rose with catalyst deterioration, and ammonia levels dropped proportionately.
-5-
-------
The results of several programs were combined to calculate emission rates for the range
of VSP bins used in MOVES.
The most important source of continuous ammonia emissions data was a program
performed by the Center for Environmental Research and Technology (CE-CERT) of the
University of California, under a Cooperative Agreement with the US EPA (CX827692-
01-0). In the program, 39 modern production vehicles received standard Federal Test
Procedure (FTP) emission tests that included second-by-second ammonia measurements.
The test fleet was designed to represent the in-use fleet in 2001, including such factors as
manufacturer, age, engine/emission control technology, and car/truck split. Five of the
newest vehicles received additional tests using cycles at lower and higher speeds than the
FTP. Results from 35 of the vehicles were retrieved from EPA's Mobile Source
Observation Database (MSOD) with enough information to allow transformation of the
second-by-second ammonia emissions results into the VSP bins used in the MOVES
emissionsRateByAge table. In MOVES, results are stratified by model year and age at
time of test.
Table 2-1 summarizes the distribution of the CE-CERT test fleet with respect to model
year and age. The MOVES table schema stratifies results into seven age groups, as
shown below.
While it is possible to enter individual model year groups into MOVES, the results of this
limited program were pooled into model year groups roughly corresponding to the
stringency of the emission standards and resulting engine/emission control technology
included in the years represented. This averaging was performed to prevent single
vehicles from biasing the emission factors for small subgroups. The model year groups
Table 2-1
CE-CERT Ammonia Test Program Vehicle Counts
MY
Group
00-01
98-99
96-97
94-95
91-93
87-90
81-87
<81
Age Group
m
i
o
3
10
ir>
*f
1
t-~
VO
2
o\
oo
8
^r
o
4
o\
i
»/">
1
o
(N
II
A
§.3
3 (L>
fc >
20
10
5
0
-6-
-------
selected were 1981-1990, 1991-1995, and 1996-2001. There were five vehicles in the
oldest group, 10 vehicles in the middle group, and 20 vehicles in the newest group.
In MOVES, 23 Operating Mode Bins are defined for running emissions processes. Bins
0 and 1 apply to braking and idle operation, 11-16 to operation at speeds between 0-25
mph, 21-30 to speeds between 25-50 mph, and bins 33-40 to speeds above 50 mph. The
bins are defined in terms of vehicle specific power (VSP). VSP was calculated using the
procedures described EPA's Draft MOVES2009 documentation.1*
Essentially, Track Road Load Horsepower (TRLHP) for each vehicle was allocated to A,
B, and C coefficients using default fractions of 0.35, 0.10, and 0.55, respectively. Actual
speed and dynamometer inertia weight were then used to compute acceleration and VSP
for each second of operation. A specific bin number, as defined in the EPA report, was
assigned to each individual second. The grams of ammonia measured during individual
seconds were then summed by VSP bin. The overall average ammonia emissions for
each bin within the three model year groups was calculated by summing the observed
emissions and dividing by the number of seconds included in the sum. Results were then
converted from milligrams per second to the grams per hour scale used in MOVES.
The driving schedule used for the FTP test does not include the speed and accelerations
required to enter higher VSP bins. Additional testing was performed, however, on five of
the vehicles in the 1996-2001 model year group. The US06 test, a component of the
Supplemental Federal Test Procedure (SFTP), includes operation at high speeds and rates
of acceleration. Results from these US06 tests were used to fill the missing bins (referred
to as -h-ole-filling" in other EPA documentation).
The US06 results of the five vehicle fleet were scaled to conform to the average FTP
results in the 20-vehicle fleet in two steps. First, the US06 results were scaled to match
to the corresponding FTP results in the five vehicle subset. Table 2-2 displays the five
vehicle FTP and US06 averages for Bins 25-40. An offset was noted between the tests.
As only a few seconds of data were collected in bin 28 of the FTP, bin 27 was selected as
the -pivot point" to fill Bins 28, 29, and 30. The ratio between the FTP results for bin 27
and the US06 results for Bin 27 was calculated (12.139/4.487 = 2.705). The US06 results
for Bins 28, 29, and 30 were calculated by dividing the original US06 results by 2.705
(12.628/2.705=4.668). Similarly, bin 35 was selected to scale the US06 results for Bins
37, 38, 39, and 40. The results for the five-car fleet were then merged as shown, using
the original FTP results for Bins 1-27 and 33-35 and the scaled US06 results for Bins 28-
30 and 37-40.
Next the merged results for the five-car fleet were scaled to the averages observed in the
twenty-car fleet. The same pivot points were selected (Bins 27 and 35), and the merged
five-car results were scaled to the 20-car results. The initial and final transformed results
are shown in Table 2-3. The final merged results were used as the average ammonia
emission factors for vehicles manufactured between 1996 and 2001 at between 0 and 5
years of age.
Numeric superscripts denote references provided in Section 5.
-7-
-------
Table 2-2
VSP Scaling Step 1
VSP Bin
FTP
US06
Scaling
Merged
25
2.126
1.680
2.126
27
4.487
12.139
2.705
4.487
28
2.627
12.628
4.668
4.668
29
13.840
5.116
5.116
30
29.497
10.903
10.903
33
1.684
8.259
1.684
35
2.081
10.485
5.039
2.081
37
20.700
4.108
4.108
38
28.760
5.708
5.708
39
24.776
4.917
4.917
40
36.148
7.174
7.174
Table 2-3
VSP Scaling MY 1996-2001
VSP Bin
Merged
FTP 20
Scaling
20 Merge
25
2.126
1.782
1.782
27
4.487
2.668
1.682
2.668
28
4.668
2.775
2.775
29
5.116
3.042
3.042
30
10.903
6.482
6.482
33
1.684
1.673
1.673
35
2.081
1.941
1.072
1.941
37
4.108
3.831
3.831
38
5.708
5.323
5.323
39
4.917
4.586
4.586
40
7.174
6.690
6.690
The next group of vehicles included model years 1992 through 1995. Only FTP results
were available for this group. Lacking any other source of data, the US06 ammonia
results obtained with the first group were scaled to the FTP results obtained with the
older cars using the same approach as described above. Results are shown in Table 2-4.
The final group of vehicles in this program included model years 1981 through 1991.
The same approach was again applied, with results displayed in Table 2-5.
Table 2-4
VSP Scaling MY 1992-1995
VSP Bin
Merged
FTP mid
Scaling
MidMerge
25
2.126
2.929
2.929
27
4.487
3.248
1.381
3.248
28
4.668
2.690
3.379
3.379
29
5.116
3.704
3.704
30
10.903
7.893
7.893
33
1.684
3.236
3.236
35
2.081
3.527
0.590
3.527
37
4.108
6.964
6.964
38
5.708
9.675
9.675
39
4.917
8.335
8.335
40
7.174
12.161
12.161
-------
Table 2-5
VSP Scaling MY 1981-1991
VSP Bin
Merged
FTP old
Scaling
OldMerge
25
2.126
1.338
1.338
27
4.487
0.823
5.450
0.823
28
4.668
0.309
0.856
0.856
29
5.116
0.939
0.939
30
10.903
2.000
2.000
33
1.684
2.028
2.028
35
2.081
1.524
1.365
1.524
37
4.108
3.009
3.009
38
5.708
4.181
4.181
39
4.917
3.602
3.602
40
7.174
5.255
5.255
The results from this scaling are summarized in Figure 2-1. The pattern observed is
similar to that obtained by EPA for other emissions, with factors within a speed group
(<25 mph, 25-50 mph, and >50mph) rising exponentially in the highest VSP bins. The
1992-1995 MY group generally had higher ammonia emissions, while the 1981-1991
group in general had the lowest. Bin 39 for all three groups had an unusual drop, which
may be related to the US06 cycle that was used to generate these results.
Figure 2-1
Ammonia Emissions (g/hr) vs. Operating Mode by
Model Year Group
10 20 30
VSP Bin (MOVES Operating Mode)
40
-9-
-------
Similar EPA analyses generally included thousands of tests. The patterns observed
between subgroups were generally much smoother; reflecting the effect of the relatively
small sample size used for this analysis.
Other test programs were used to estimate emission factors for earlier (1980 and older)
and more modern (2002 and later) vehicles.
CRC program E-602 revealed significant reductions in average fleet emissions for current
technology vehicles (LEV and better). The CRC program also tested artificially aged
catalysts equivalent to 120,000 road miles. The deterioration rate observed in the E-60
fleet was used to project deterioration from the rates observed at low mileage MOVE bins
derived in the first CE-CERT program to the aged bins for the complete MOVES
database.
A final significant study combining the results of a number of additional studies was
reported in SAE paper 830987.3 This literature survey reports the average and range of
ammonia emissions found in several previous testing programs. This 1983 paper
includes results from light-duty vehicles without catalyst, with oxidation-only catalysts,
and three way catalysts. Also reported are average emissions observed with heavy-duty
gasoline trucks and light- and heavy-duty Diesel trucks.
The E-60 and 1983 SAE paper did not, however, report the second-by-second emission
results required to develop MOVES database tables corresponding to the vehicles tested.
The average emission levels were, however, used to scale the MOVES bins developed in
the CE-CERT study to earlier and later vehicles.
The limited number of vehicles included in the primary studies could not directly
represent all possible MOVES bins combinations. While it was possible to estimate
future and past composite emissions, the second-by-second data required to fill all
possible MOVES strata were not available. Simple ratios were used to scale the VSP bin
results developed with the CE-CERT data to -4*ole fill" the remainder of the MOVES
emissionRateByAge table.
The most prevalent unit found in the literature to express average emission rate is mg/mi.
Table 2-6 displays the average rates drawn from the three programs described above.
Scaling factors for vehicle deterioration were estimated using the E-60 results. Tests in
this program were performed with as-received catalysts and with catalysts that had been
artificially aged the equivalent of 120,000 miles. The average as-received emissions for
the fleet were 14 mg/mi while the results with the aged catalysts were 21 mg/mi. A
factor of 21/14 = 1.5X was used to estimate the effect of aging full useful life. Linear
interpolation was used to assign an aging factor of 1.2X for half useful life deterioration.
No additional deterioration was assumed for the 20+ year age group.
-10-
-------
Table 2-6
Fleet Average Ammonia Emissions (mg/mi)
Program:
MY
Range:
Age/
0-5
6-9
10-19
20+
E-60
2002+
14
21
CE-CERT
96-01
43
92-95
83
81-91
37
SAE 830987
81-83
101
75-80
15
15
<1975
11
11
11
These deterioration factors were applied to the CE-CERT program to arrive at emission
factors for the 1996-2001 model year groups of 43 mg/mi, 1.2X43 for the 6-9 year age
group, and 1.5X43 for the 10-20+ year age groups. Assuming some deterioration had
already taken place with the 1992-1995 vehicles, an emission factor of 1.1X83 was
assigned to the 10-19 year old vehicles and 1.2X83 to the 20+ year old vehicles. No
additional deterioration was assigned to the 1981-1991 vehicles in the 10-19 year old age
group, leaving the average emission factor of 37 mg/mi for the 20+ year old group.
These average emission rates were then used to scale the VSP bins previously calculated
from the CE-CERT results. The 0-5 year VSP averages for the 1996-2001 vehicles
remained the same. Each bin in the 0-5 year group was multiplied by 1.2 to arrive at VSP
bins for the 6-9 year group. Each bin in the 0-5 year group was multiplied by 1.5 to
estimate VSP bins for the 10-19 and 20+ year groups of this model year group. The
1992-1995 model year VSP bin results were multiplied by 1.1 to represent the 10-19 year
old results and 1.2 to represent the 20+ year old vehicles. For the 1981-1991 vehicles,
the 10-19 year old bins were assigned without correction to the 20+ year old age group.
No second-by-second data were found for the newer vehicles. After reviewing results
obtained with large samples for HC, CO, and NOx, it appeared reasonable to scale the
newest CE-CERT VSP bin results (1996-2001) to the 2002+ model years, in proportion
to the fleet average composite results. Thus the 2002+ model year VSP bins were
assigned a value equal to 14/43 times the individual bin results for the 0-5 year old
newest vehicles. The deterioration rates were applied to the 0-5 year old bins by again
multiplying by 1.2 and 1.5.
The least amount of data was available for the earliest model year vehicles. The literature
review SAE paper provided averages and ranges for three groups of vehicles:
not-catalyst-equipped; equipped with oxidation-only catalysts; and equipped with
three-way catalysts. It was apparent that the oxidation catalyst and non-catalyst vehicles
-11-
-------
presented substantially lower ammonia emission rates than the newer technology
vehicles. Oxidation catalysts came into widespread use in 1975 following nationwide
installations of unleaded fuel dispensers. Three-way catalysts were introduced in federal
vehicles in 1981 to meet technology-forcing standards (requiring significant NOx
reduction). The non-catalyst results were therefore assigned to all model years prior to
1975, and the oxidation catalyst results were assigned to model years 1975-1980. No
deterioration results were available for either group. Recognizing the relatively small
remaining population of these older vehicles, a single set of VSP emission factors was
calculated for the 1975-1980 group by applying the ratio of that group to the 1981-1991
results (15/37 = .405) times the individual 1981-1991 VSP bins. These bins were used
for all age groups in the 1975-1980 model year range. Similarly, a factor of 11/37=0.297
was used to scale the 1981-1991 bins to all model years prior to 1975.
One significant group of unfilled strata remained: those representing vehicles at a
younger age than were available when the vehicles were tested in 2000-2001. It was not
possible to test a MY1985 vehicle at 0-10 years of age in CY2000. This does not present
major problems for a run performed in CY2009: all 1985 vehicles are 20+ years old.
While it is acknowledged that certain baseline runs representing earlier calendar years are
affected by these vehicles, again because of the lack of data, the results collected in the
CE-CERT program were applied to the younger vehicle age strata without modification.
Table 2-7 repeats the base results from Table 2-6 and summarizes the extrapolation and
interpolation used to complete this group of vehicles.
Table 2-7
Extended Fleet Average Ammonia Emissions (mg/mi)
Program:
MY
Range:
Age/
0-5
6-9
10-19
20+
E-60
2002+
14
17
1.2 X
21
1.5X
21
1.5X
CE-CERT
96-01
43
52
65
65
92-95
83
83
91
91
81-91
37
37
37
37
SAE 830987
81-83
101
75-80
15
15
0.405X
81-91
<1975
11
11
11
0.297X
81-91
-12-
-------
The MOVES EmissionRateByAge table requires identification of the fuel type and the
regulatory class of each entry. To this point only gasoline-fueled vehicles have been
discussed (fuel type 1). The regulatory classes of the vehicles included are 20 - Light
Duty Vehicles and 30 - Light Duty Trucks. Continuing the practice used in other
MOVES tables, the factors for gasoline are used when no data are available for ethanol.
Ethanol is fuel type 5.
The 1983 SAE paper reported tests of light- and heavy-duty Diesel trucks, and heavy-
duty gasoline-powered trucks. While Diesel-powered light and heavy-duty vehicles have
been subjected to increasingly stringent emission standards, emission control technology
for this class of vehicles has not included closed loop fuel mixture control and three-way
catalyst converters. For this effort, the approach used for oxidation and non-catalyst
vehicles was continued by applying the ratio of the reported average fleet emissions to
the emission bins computed using the 1981-1991 CE-CERT fleet.
Light-Duty Diesel Trucks (LDDT) were reported to have an average ammonia emission
rate of 7 mg/mi, which is 7/37 = 0.189 of the 37 mg/mi reported in the CE-CERT
program for 1981-1991 vehicles. This factor was used to develop binned VSP factors for
all group 20 and 30 Diesel-powered vehicles. Heavy-duty Diesel trucks (HDDT) were
reported to produce an average of 27 mg/mi, a ratio of 0.730 of the CE-CERT factor.
These results were applied to all group 41, 42, 46, 47, and 48 Diesel-powered vehicles.
These results are summarized below in Table 2-8.
Table 2-8
Ammonia Emissions for Diesel and Heavy Duty
Vehicles (mg/mi)
Class
LDDT
HDDT
HDGT
mg/km
4.2
16.8
28
mg/mi
7
27
45
1981-
1991
37
37
37
Ratio
0.189
0.730
1.216
Significant emission control rule changes have been enacted for Diesel powered vehicles
in recent years. Most manufacturers have elected to employ Selective Catalyst Reduction
(SCR) emission control technologies including urea injection to achieve the new
emission standards. A by-product of SCR is the release of excess ammonia (slip) from
the SCR device. Commonly reported SCR systems for on-road vehicles include a final
ammonia slip catalyst to minimize this release. These devices are reported to reduce
ammonia release to —ngligible" levels, with one system reporting 1-3 ppm in the exhaust
stream. That level in a large Diesel engine would represent between 1 and 2 grams/hour
of ammonia. It would be appropriate to review in-use performance as production SCR
-13-
-------
systems are placed into service to determine if the ammonia emission factors
recommended here should be updated.
The last type of vehicle for which ammonia emission factors were developed was
motorcycles (regulatory class 10). No test program including measurement of ammonia
with motorcycles was identified. The results of the CE-CERT program were therefore
used, with consideration of differing emission standards and emission control
technologies applied to the two different types of vehicle. As catalytic converters were
not common on motorcycles before 1999, the non catalyst (pre-1975) light-duty vehicle
rates were applied to this group. Catalytic converters and feedback fuel control systems
became increasing common on motorcycles between 2000 and 2005, so the 1975-1980
light-duty vehicle rates were applied. New standards have been phased in for
motorcycles, and three-way catalysts with feedback systems have become the norm. The
1981-1991 light-duty vehicle results were used for 2006 and newer motorcycles. As no
significant number of Diesel motorcycles are expected, no ammonia emission factors
were developed.
The amount of second-by-second test data, and information regarding changes in
ammonia formation with vehicle age, is extremely limited. No information regarding
ammonia emissions with the use of ethanol blends was found. The newest technology
(LEV) vehicles, on the other hand, were found to produce extremely low levels of
ammonia. One remote sensing study4 attributes more than 20% of total daily ammonia
emissions in the South Coast Air Basin to on-road vehicles. Kean reports5 that heavy-
duty Diesel emissions are difficult to measure in a tunnel study because of interference
from relatively higher light-duty ammonia emission production, but he does consider the
potential impact of ammonia slip. Future studies should be monitored to determine if
additional data collection for the purpose of updating the recommended factors is
appropriate.
2.2 Sulfur-Based Emission Rates
The Draft2009 MOVES model includes the sulfateEmissionRate table. The table was
populated using early results from the Kansas City program. A draft report6 detailing t
analytical procedure used was provided with the Work Assignment.
The sample size used in the original analysis was limited because QC review had not
been completed and each point required both the exhaust sulfate level and fuel sulfur
level. Thirteen points were included in the original analysis. An updated MSOD
database including results from Kansas City was provided with the Work Assignment.
Forty additional tests with both exhaust sulfate and fuel sulfur were identified in the
database. The analytical approach described in the draft report was applied to the
combined results.
-14-
-------
Figure 2-2 displays the results obtained with 1981 and later vehicles. The solid square
points reflect the data initially used, while the -X" points represent the new data. A
regression was performed to determine the slope of the equation:
SC>4 (g/mi) = slope x fuel S (ppm) (forcing zero intercept)
Figure 2-2
Exhaust SO4 vs. Fuel S - 1981 and Newer
Onno*^
n nnon -
E
-^ n nm^ -
S
w
to
3 n nnin -
.c
X
LU
n nnn^ -
Onnnn -
•
X
• x New Data
^S — — Linear (Original)
*'''
• " ' *>*
m^> ' x ^ — "*"
'&J**j($~i3SF* "xxx
0 100 200 300 400 500
Fuel S (ppm)
The newly developed combined slope was slightly more than half of that developed with
the original limited sample:
Original
Update
0.000002952
0.000001597
Only four of the more modern vehicles had SO4 emissions results above 0.0005 g/mi,
with little response to varying sulfur level in the fuel.
The average SO4 emission rate for this group of vehicles was 0.000344.
-15-
-------
The results obtained with 1980 and older vehicles were similar and are displayed in
Figure 2-3. The original analysis included only three vehicles. The slope was
determined by averaging the three Exhaust 864 measurements and dividing by the
average fuel S for those tests. Several additional tests were found, and a slope was
determined using linear regression.
Original
Update
0.00002892
0.00001907
Figure 2-3
Exhaust SO4 vs. Fuel S - 1980 and Older
n ni4n -,
n n-ion .
— n m nn -
D)
i
- n nn«n .
O
CO
"5 n nnfin .
OJ
x
MI n nciACi -
Onnon
Onnnn
c
•
X
^'
^.-- **"
_^^"
v_— ^ * XH
) 50 100 150 200
FuelS
• Original
x New Data
Linear (Original)
Linear (New Data)
The slope developed with the combined data set for older vehicles is more than 10X the
slope obtained with the 1981 and newer vehicles.
No additional results were obtained for Diesel vehicles. The same factors were used for
ethanol fuel, with the assumption that the sulfur level used as an input will be the level
measured in the diluted fuel + ethanol blend used on the road, not just in the blend fuel.
Lacking other information, the same factors were used for starting, running, and extended
idle.
-16-
-------
2.3 Nitrogen Oxide and Nitrogen Dioxide Ratios
NO and NC>2 ratios are to be expressed in MOVES as a fraction of total NOx. The
NONO2Ratio table will include the MOVES variables:
meanBaseNONO2ratio
NO to NOx ratio for entries including polProcessID 32xx
NO2 to NOx ratio for entries with polProcessID 33xx
meanBaseRateCV
Coefficient of Variation (COV) = standard deviation / mean
Results were additionally classified in terms of MOVES variables sourceTypelD (types
11-62), polProcessID (with pollutantID 32 for NO and pollutanflD 33 forNO2),
fuelTypelD (initially gasoline, Diesel, and ethanol) and modelYearGroupID (including
vehicle model years 1960-2050). The -Batasource" field for entries in the table initially
refer to this report with an ID of 8101. A -chained" calculator will retrieve total NOx
and the coefficients in the NONO2Ratio table to compute individual NO and NO2 results.
A wide range of NO2 to NOx ratios is reported in the literature. Because NO oxidizes to
NO2 over time, a significant problem is defining when to measure and report results.
Values of 1-3% are frequently reported for gasoline-powered vehicles. One study,
utilizing nearly instantaneous real-time exhaust emission data, reports an average value of
less than 1%. The study continues with the conclusion that CVS bag measurements are
questionable because -Ast oxidation of NO to NO2 occurs when diluted exhaust is
sampled in bags." Other studies report NO2 fractions as high as 30% with certain types
of emission control equipment and operating conditions.8'9'10
A study including continuous NO and NO2 results from a representative cross-section of
modern gasoline powered vehicles was identified—the CE-CERT study previously used
in the development of ammonia emissions factors.11 EPA's MSOD database contains
second-by-second NO and NO2 results for 20 of the vehicles, providing a representative
cross-section of results for vehicles ranging from Tier 0 (1981-1989) through NLEV and
ULEV (2001) light-duty cars and trucks. The data allow calculation of both average
emission levels and variation of results within groups. No studies reporting individual
1980 and earlier vehicle results were identified. This includes vehicles produced prior to
the introduction of catalytic converter (generally 1960-1974), and to vehicles produced
with oxidation catalysts (1975-1980). The -widely accepted assumption" of 2.5% NO2 to
NOX ratio was used to populate the MOVES tables for 1960 through 1974 vehicles.
Results for the remaining groups of gasoline-powered light-duty vehicles were computed
using the CE-CERT data, and are summarized in Table 2-9.
-17-
-------
Table 2-9
Gasoline NONO2 Ratios
Model
Years
1960-1980
1981-1990
1991-1995
1996+
Average
NOx Level
-
0.764
0.824
0.107
Running
NC>2 ratio
0.025
0.068
0.046
0.164
cov
-
0.094
0.054
0.099
Start
NC>2 ratio
0.025
0.039
0.013
0.049
COV
-
0.040
0.003
0.020
For the 1981-1996+ groups, the second-by-second NO and NO2 readings were summed
for each bag of each test in the respective groups, and then the ratios of NO and NO2 to
the total (NO+NO2) were calculated. The average and standard deviation of the ratios
within each model year group were then determined. The average NO2 ratio is reported
in the table. (The NO ratio must be 1 minus the NO2 ratio.) The Coefficient of Variation
(COV) is the standard deviation of the individual values divided by their average.
The limited sample sizes yielded a slightly higher average NOx g/mi value from the
1991-1995 than the 1981-1990 vehicle groups. This was not caused by a single outlier
vehicle, but rather normal variation found within small samples (six tests versus seven
tests). Of greater interest is the consistency found in the NO and NO2 ratios to the total in
their groups—all less than 10%. The most modern cars, with very low average total
NOx, resulted in the highest NO2 ratio (16%); the remaining tests displayed results
ranging between 1 and 7%, consistent with the general results reported in the literature.
The model year groups for sourceTypelD's 21,31, and 32 (light-duty cars and trucks)
were assigned the coefficients displayed in Table 2-9. The model year grouping for
sourceTypelD 11 (motorcycles) was modified to account for the differences in emission
standards and control technologies employed, with 2010+ motorcycles receiving the
coefficients assigned to 1991-1995 automobiles. Similarly, assignments to vehicles in
sourceTypelDs ranging between 41-62 (heavy-duty gasoline powered trucks) were
aligned with emission standards and technology changes.
No single study of NO/NOx ratios was identified for Diesel powered vehicles. Several
SAE papers summarizing testing on individual vehicles were reviewed, and provided
reasonably consistent results within technology groups. Earlier technologies had little
reported impact on NO/NO2 ratios, but advanced technologies potentially can have a
major influence.
1 9
Lanni included measurement of both NO2 and NOx total emissions, citing research that
indicated that retrofit of CRDPFs (Continuously Regenerating Diesel Particulate Filters)
had resulted in an increase in NO2 emissions. The CRDPF in the vehicles studied uses an
-18-
-------
upstream catalyst that partially oxidizes NO to NO2, as well as oxidizing HC, CO, and
SOF. A downstream particulate filter traps soot particles. The participate trap is
continuously regenerated by combustion of the soot particles with the NO2 remaining in
the exhaust stream. Lanni reports his unmodified test vehicles -agrees well with the 8%
average NO2/NOx volume ratio reported in the literature" and continues that total NOx
emissions were not affected by the DPF, but elevated NO2 emissions resulted in elevated
NO2/NOX ratios, approaching 50%.
Ayala13 reported similar results on Diesel-powered buses before and after installation of a
retrofit DPF, using parallel analysis of NO and NOx during several test cycles.
Unpublished results provided by Dr. Ayala on a Diesel-powered Class 8 tractor was also
considered.
Table 2-10 displays the results used to determine average ratios without DPF controls.
Table 2-10
Diesel Without DPF trap NO/NO2 Ratios
Pollutant
or Ratio
NO2
NO
NOx
NO2/NOx
NO/NOx
SAE 2003 -01 -03 00
CBD
OEM
2.1
24.5
26.6
0.079
0.921
CBD
OEM
1.6
22.5
24.1
0.066
0.934
NYBDC
4.6
65.70
70.30
0.065
0.935
SAE 2002-0 1-1 722
CBD
OEM
0.92
29.28
30.20
0.030
0.970
SS
OEM
2.14
22.49
24.63
0.087
0.913
Ayala II
Baseline
1.26
19.87
21.13
0.060
0.940
The average NO2/NOx ratio observed was 0.065, with a standard deviation of 0.019 and a
COV of 0.301.
Table 2-11 displays average results obtained in three programs for Diesel vehicles with
DPF traps installed. The average NO2/NOx ratio for this sample is 0.594, with a standard
deviation of 0.406 and COV of 0.193.
-19-
-------
Table 2-11
Diesel With DPF trap NONO2 Ratios
Pollutant
or Ratio
NOx
NO
NO2
NO/NOx
NO2/NOx
SAE 2003 -01 -03 00
HDDS
25.9
12.8
13.1
0.494
0.506
HDDS
22.1
11.8
10.3
0.534
0.466
SAE 2002-01-1722
CBD
34.4
17.7
16.7
0.515
0.485
SS
26.5
13.8
12.7
0.521
0.479
NYBC
52.1
28.3
23.8
0.543
0.457
HDDS
23.1
14.1
9.0
0.610
0.390
Ayala II
CRT
18.8
12.9
5.8
0.688
0.312
VSCRT
5.2
3.9
1.3
0.747
0.253
ZSCRT
5.7
4.2
1.5
0.731
0.269
DPX
10.58
7.75
2.83
0.733
0.267
CCRT
8.0
3.3
4.7
0.412
0.588
Table 2-12 displays results obtained following extended idle. The average NO2/NOx
ratio observed was 0.108, with a standard deviation of 0.070 and COV of 0.645.
Table 2-12
Diesel With DPF trap and Extended Idle NONO2 Ratios
Pollutant
or Ratio
NOX
NO
NO2
NO/NOx
NO2/NOx
Ayala II
CRT
VSCRT
90.50
86.53
3.97
0.956
0.044
ZSCRT
77.2
69.6
7.6
0.901
0.099
DPX
43.05
35.18
7.86
0.817
0.183
CCRT
###
-20-
-------
3. HIGHWAY MOTORCYCLE EMISSION RATES
AND ACTIVITY FACTORS
Task 3 of this study required the development of emission and activity factors for
highway motorcycles. Subtasks included a review and analysis of available emissions
data and consultation with the Motorcycle Industry Council (MIC) to obtain retail sales
data, registration data, and the results of an owners survey.
This section of the report identifies the data sources for the highway motorcycle emission
rates and activity data being developed for MOVES and describes the methods used to
process and validate these data and assemble them into MOVES-ready database tables. It
is organized into sub-sections that provide an overview of key highway motorcycle
characteristics, describe uncontrolled and controlled emission rates, discuss Inspection
and Maintenance program effects on emissions, present latest estimates on motorcycle
populations and activity levels and finally, describe how the emission rate and activity
data were organized into specific MOVES-ready tables.
3.1 Highway Motorcycle Characteristics
The U.S. highway motorcycle fleet encompasses an extremely wide range of vehicles.
The variation in available models is actually greater than the variation in passenger cars
in terms of engine size, engine power, vehicle weight, and vehicle performance. At one
end of the spectrum are small scooters weighing less than 200 pounds and equipped with
<50 cc single cylinder engines and CVT transmissions that have inadequate performance
for freeway operation. Historically, many of the <50 cc scooters have used 2-stroke
engines. At the other end of the performance spectrum are 400-500 pound sport models
equipped with 1000-1300 cc, 4-cylinder, 4-stroke engines and 6-speed manual
transmissions with performance capabilities exceeding those of high-performance
passenger cars. At the other end of the size spectrum are 3-wheeled -trike" configuration
vehicles weighing close to 1,500 pounds with automotive V-8 engines exceeding 6 litres
displacement.
Excluding the relatively low volume of 3-wheeled vehicles classified as motorcycles,
highway motorcycle models are commonly divided into six separate categories:
-scooter," -standard," -d-ual-sport," -sport," -touring," and -eruiser." Some
manufacturers use variations on these categories. The term —supesport" is sometimes
used to distinguish ultra-high-performance models from lower-powered models of the
same basic design. -Sport touring" and -kixury touring" are terms used to describe
-21-
-------
subcategories of touring motorcycles. (There are also a number of -crossover" models
that do not fit neatly into any of the six categories.)
-Scooters" might be considered the motorcycle equivalent of an economy car. They are
motorcycles designed primarily for providing low-cost transportation in an easy-to-ride
configuration with a -step-through" frame, a relatively low seat, an upright riding
position, and an automatic transmission. Most scooters use some form of automatic
transmission, making them easier for inexperienced riders to use. They usually have a
short wheelbase to facilitate maneuverability in congested urban areas. Almost all
scooters have a partial fairing and windscreen to provide wind and weather protection.
Scooters have smaller diameter wheels and tires than other motorcycles. There is an
extremely wide range of performance within the scooter category, ranging from models
that are too slow for freeway driving to more powerful models that have the performance
of typical passenger cars.
-Standard" models are conventional motorcycles with an upright riding position and no
fairing. They are sometimes referred to as —nked" bikes; however, small windscreens
and partial fairings may be standard on some models and available as an option on others.
They are seldom equipped with saddlebags, but may have bags available as optional or
aftermarket equipment. Standard motorcycles are usually intended to be general-purpose
vehicles for all-around riding. Standard models are available in a wide range of
performance levels, ranging from barely adequate for freeway travel to ultra-high
performance. Representative models include the BMW R1200R and the Suzuki Bandit
1250S.
-T-ouring" models are the motorcycle equivalent of luxury sedans and sports sedans.
They have the upright riding position of a -standard," which is the most comfortable for
long distance riding, in combination with a fairing, windscreen, and saddlebags. The
distinction between -luxury touring" and -sport touring" models is primarily related to
size and performance. Sport touring models tend to be smaller, more nimble, and higher
performance. Representative models include the Yamaha FJR1300 and the Kawasaki
Concours 14. Pure touring models are larger and designed to provide more room and
wind protection for a rider and a passenger. Representative models include the Honda
Gold Wing and the BMW K1200LT. The minimum level of performance from touring
motorcycles is adequate performance for freeway travel. At the other end of the
spectrum, some sport touring models have a level of performance comparable to a typical
-sport" bike.
-Sport" motorcycles are designed primarily for exceptional handling and performance.
They are the motorcycle equivalent of a -sports car." They generally have a relatively
short wheelbase, and narrow, forward-located handlebars that put the rider in a leaning
forward position. The footpegs are often set high and to the rear of the motorcycle to
provide increased cornering clearance. Most are equipped with full fairings to reduce
aerodynamic drag at high speeds, but the fairings are narrow and the windscreens are too
short to take the wind off of a rider's head and shoulders unless the rider is in a -tacked"
position that most riders would find comfortable only for short periods of time. Most
-22-
-------
sport models have relatively high-output engines that provide a greater level of
performance than is available from most high performance passenger cars.
Representative models include the Suzuki GSX750R, the Honda CBR600RR, the
Yamaha YZF-R1, and the Kawasaki ZX10R.
-Bual-sport" models are the motorcycle equivalent of a sport-utility vehicle or a -Jeep."
They share the upright riding position of a -standard" model but they tend to be -taller"
motorcycles with relative long suspension travel to make them suitable for off-road
operation, or at least operation off of paved roads. It is particularly important for the
footpegs on a dual-sport bike to be located directly under the rider so that they are easy to
stand on while riding in rough terrain. Representative models include the BMW
R1200GS, the Suzuki V-Strom, and the Kawasaki KLR650.
The design criteria for the -eruiser" category are primarily focused on preservation of a
more classic style and sound than on high performance. Most cruisers are equipped with
relatively large displacement, V-twin engines that generate maximum horsepower at a
relatively low engine speed. These engines produce a distinctly different sound than the
high speed, 4-cylinder engines used in most other large motorcycles. Cruiser category
motorcycles are also designed to provide a fundamentally different rider position, with
higher handlebars and forward" foot controls (often with floorboards rather than
footpegs) and relative low seat height. Cruiser models typically have a relatively long
wheelbase and steering geometry that provides increased stability in a straight line at the
expense of cornering performance. To preserve a classic visual appearance, cruiser
models are equipped with either air-cooled engines or water-cooled engines that are
designed to resemble an air-cooled engine. Minimal use of fairings or other body work
preserves the visual prominence of the engine, an important aspect of cruiser design.
Representative models include the Harley-Davidson Softail, the Honda VTX1800, the
Suzuki Boulevard C90, and the Kawasaki Vulcan 900 Classic.
While the differences between a -scooter" and a -touring" bike are extreme, the
variations within an individual category are almost as extreme. For example, there are
greater than 10:1 variations in engine power among -scooter" models.
Under current EPA regulations, highway motorcycles are divided into three classes based
on engine displacement. Class I covers 0-169 cc; Class II covers 170-279 cc; and
Class III covers >279 cc. Prior to model year 2006, Class I covered 50-169 cc and
motorcycles with <50 cc engines were exempt from emission standards. Motorcycles
categorized as -scooters" fall into all three displacement classes. Motorcycles
categorized as -dual sport," and -standard" are primarily in Classes II and III.
Motorcycles categorized as —spdr" "cruiser," and -touring" are primarily in Class III.
Prior to calendar year 2000, over 90% of all highway motorcycles were in Class III. Due
to a recent increase in the sales of small scooters, Class III motorcycles are currently
estimated to be approximately 85% of the highway-legal fleet (as is discussed in more
detail below).
-23-
-------
3.2 Uncontrolled Emissions Characteristics
Since engine displacement and weight are correlated, Class I motorcycles are the lightest
weight, which contributes to inherently low NOx emission levels. Class I motorcycles
also have relatively small fuel systems, which contributes to relatively low evaporative
emissions. However, the small displacement engines have relatively high surface-to-
volume ratios, which contributes to higher HC exhaust emissions. Prior to model year
2006, the lack of emission standards for motorcycles <50 cc allowed for the use of 2-
stroke engines, which have much higher HC exhaust emission levels than 4-stroke
engines. Prior to 1978, high-emission 2-stroke engines were used on many models with
>50 cc engines and control of crankcase vent emissions was not required from
motorcycles with 4-stroke engines.
Crankcase Emissions - The imperfect sealing of the combustion chamber in a premixed
charge, reciprocating engine results in —blowb" of unburned fuel past the piston rings
and into the crankcase. To prevent blowby from pressurizing the crankcase (causing oil
leaks) and to prevent unburned gasoline from building up in the crankcase, a crankcase
ventilation system is incorporated in 4-stroke engines. (Two-stroke engines use a
pressurized crankcase and do not have a crankcase vent.)
Prior to the imposition of emissions control requirements, crankcase ventilation was
routinely provided by a road draft tube. Air flow under the vehicle created a slight
vacuum which assisted in the removal of blowby gases from the crankcase while also
drawing some fresh air through the crankcase breather/vent. Based on block data
statements in MOBILES, uncontrolled crankcase emissions were approximately 4 g/mi of
unburned hydrocarbon, nearly half as much as uncontrolled exhaust hydrocarbon
emissions. Subsequent versions of EPA's vehicle emissions model assume crankcase
emissions are proportional to exhaust emissions, with pre-controlled vehicles having
crankcase emissions that are 33% of running" exhaust.
For vehicles with crankcase controls, estimated crankcase emissions drop to 1.3% of
running exhaust. Actual testing by Sierra Research determined that typical crankcase
emissions from passenger cars without Positive Crankcase Ventilation (PCV) systems
installed are approximately 2.5 grams per mile over the LA4 driving cycle.14 Second-by-
second data collected by Sierra indicated that crankcase emissions increase with
acceleration rate. As a result, actual emissions in customer service are likely to be higher
than 2.5 g/mi for uncontrolled cars.
Measurement of uncontrolled motorcycle crankcase emissions was conducted during a
study conducted by Southwest Research Institute for EPA.15 Measurement of blowby
emissions was limited to idle, 20 mph road load, and 40 mph road load operation of a
1972 model year, 220 pound, 100 cc, 4-stroke motorcycle. Average emissions were
calculated to be 0.421 g/mi, assuming 20% idle. Assuming average crankcase emissions
are proportional to fuel consumption, the measured crankcase emissions rate would be
increased by 60% to account for the difference between the 80 mpg measured for the
100 cc motorcycle over the LA4 driving cycle and the 50 mpg of the average motorcycle.
-24-
-------
It would also be appropriate to account for the expected increase in crankcase emissions
during vehicle operation more representative of typical driving. However, analysis of the
second-by-second data collected by Sierra Research during its testing of passenger car
crankcase emissions indicates that crankcase emissions during a mixture of idle, 20 mph,
and 40 mph operation are very close to average emissions over the LA4 driving cycle.
Our best estimate of uncontrolled crankcase emissions for the average motorcycle is
therefore 0.7 g/mi HC. Assuming crankcase emissions are proportional to fuel
consumption, this is consistent with the 2.5 g/mi crankcase emissions rate measured on a
full-size passenger car.
Exhaust Emissions - Although physically much smaller than passenger cars and light-
duty trucks, the uncontrolled exhaust emissions from highway motorcycles were nearly
as high as those of cars and light trucks. This was due to the fact that a significant
fraction of the uncontrolled motorcycle fleet was powered by 2-stroke engines with
relatively high HC emissions. Based on information available to EPA in 1992, the
exhaust emissions of uncontrolled 2-stroke motorcycles were estimated at 15.4 g/mi HC
and 27.0 g/mi CO.16 The same document estimates the emissions of uncontrolled 4-
stroke motorcycles at 2.9 g/mi HC and 42.0 g/mi CO.
Regulated Pollutants - Sierra has independently estimated uncontrolled exhaust
emissions from motorcycles by analyzing data collected during the previously referenced
SwRI study and CARB surveillance testing programs.17'18'19 Figures 3-1, 3-2, and 3-3
present the FTP results for HC, CO, and NOx as a function of odometer for 35 pre-1978,
4-stroke motorcycles ranging from 90 cc to 1340 cc displacement (average 573 cc).
Three of the vehicles were tested only on the LA92 driving cycle rather than the LA4.
FTP emission levels were estimated based on the correlation between LA92 and LA4 for
29 motorcycles tested on both cycles during CARB's third (1998) surveillance testing
program. As shown in the figures, there is wide variation in the test results. HC
emissions range from about 1 g/km to 7 g/km. CO emissions range from just under
6 g/km to 60 g/km. NOx emissions range from 0.04 g/km to 1.9 g/km. Since emissions-
related defects were identified in very few of the vehicles, the variation appears to be
related to differences between the various models tested. Examination of the detailed
information available for each tested vehicle also indicates no significant relationship
between emissions and engine displacement. As evidenced by the low coefficient of
determination (r2) values, the relationship between emissions and odometer is not
statistically significant.
Average emissions for pre-1978 4-stroke models were 2.92 g/km HC, 31.2 g/km CO, and
0.27 g/km NOx. Given the variability in the available test results, these averages should
be applied to all three displacement classes—I, II, and III. The available data do not
support any specific deterioration rate and, as discussed below, data on later model
vehicles indicate an insignificant increase in emissions as mileage is accumulated.
-25-
-------
Figure 3-1
Exhaust Hydrocarbons vs. Odometer
Pre-1978 4-Stroke Models
y = 0.00002397X + 2.50087761
R2= 0.13086401
20,000 40,000 60,000 80,000 100,000 120,000
Odometer Miles
Figure 3-2
Exhaust Carbon Monoxide vs. Odometer
Pre-1978 4-Stroke Models
70
60
50
40
*»* *
»» *
O
o
30
20
10
y = 0.00028179X + 26.27756540
R2= 0.20167872
20,000 40,000 60,000 80,000 100,000 120,000
Odometer Miles
-26-
-------
Figure 3-3
Exhaust Oxides of Nitrogen vs. Odometer
Pre-1978 4-Stroke Models
2.0
1.5
D)
1.0
0.5
0.0
y = -0.00000245X + 0.31161056
R2= 0.02401888
20,000 40,000 60,000 80,000
Odometer Miles
100,000 120,000
Figures 3-4, 3-5, and 3-6 present the more limited data available for motorcycles with 2-
stroke engines. Data were available for only nine vehicles, ranging from 125 cc to 500 cc
engine displacement. HC emissions range from about 6 g/km to 24 g/km. CO emissions
range from about 5 g/km to 38 g/km. NOx emissions range from 0.0 to 0.1 g/km. As
was the case with the 4-stroke models, the variation appears to be related to differences
between the various models tested, rather than displacement or odometer.
Average FTP emissions for pre-1978 2-stroke models were 11.94 g/km HC, 19.7 g/km
CO, and 0.04 g/km NOx. As in the case of the 4-stroke models, the available data do not
support any specific deterioration rate.
-27-
-------
30
25
Figure 3-4
Exhaust Hydrocarbons vs. Odometer
Pre-1978 2-Stroke Models
= -0.0006x+20.247
R2= 0.0872
!
o
20
15
10
5,000 10,000
Odometer Miles
15,000
20,000
40
30
Figure 3-5
Exhaust Carbon Monoxide vs. Odometer
Pre-1978 2-Stroke Models
-0.0005x+28.177
R2 = 0.023
10
5,000 10,000
Odometer Miles
15,000
20,000
-28-
-------
0.12
0.10
0.08
O 0.06
0.04
0.02
0.00
Figure 3-6
Exhaust Oxides of Nitrogen vs. Odometer
Pre-1978 2-Stroke Models
y=-0.00000x+ 0.06299
R2= 0.02524
5,000 10,000
Odometer Miles
15,000
20,000
Paniculate Emissions - Very limited data are available for particulate emissions from
pre-1978 model motorcycles. The above-referenced SAE paper by Hare, et al. provides
parti culate emissions estimates based on the tests of four 4-stroke and three 2-stroke
engines tested on leaded gasoline under steady-state conditions. Translating the steady-
state test results to grams per mile, the paper estimates average particulate emissions at
0.048 g/mi for 4-stroke motorcycles and 0.36 g/mi for 2-stroke motorcycles. (However,
it should be noted that these estimates are based on the use of fuel containing 3 grams per
gallon of tetra-ethyl lead.)
Particulate emissions data more representative of current fuels are available from a 2003
SAE paper by Ricardo.20 Ten 4-stroke and two 2-stroke motorcycles were tested over
three different driving cycles: two versions of the European test cycles and a preliminary
version of the transient World Motorcycle Test Cycle (WMTC), which our analysis
indicates is reasonably representative of light-duty vehicle operation in the U.S. The 2-
stroke models were both 50 cc; the 4-stroke models ranged from 150 to 1300 cc. All of
the testing was done on unleaded gasoline. Particulate emissions were measured using
the technique specified for Diesel particulate measurement in Europe.
Except for one of the 2-stroke models, particulate emissions were reported to be in the
range of 0.001 g/km to 0.013 g/km. As best we can determine from the poor quality
-29-
-------
graphs in the paper, the average particulate emission rate appears to be approximately
0.007 g/km for the 4-stroke motorcycles tested on the WMTC.
One of the 2-stroke models was a direct injection engine that produced particulate
emissions comparable to those of 4-strokes. The more conventional carbureted 2-stroke
produced particulate emissions approximately ten times higher than the 4-strokes, i.e.,
approximately 0.07 g/km.
At the time they were originally sold, the substantially higher particulate emissions
associated with the use of leaded fuel were applicable to pre-1978 model motorcycles.
However, when pre-1978 models are running on unleaded fuel, the measurements
reported by Ricardo are considered to be representative of vehicles currently in operation.
Evaporative Emissions - Uncontrolled evaporative emissions were estimated from three
primary sources: (1) surveillance testing of uncontrolled motorcycles by CARB;
(2) testing of uncontrolled off-highway motorcycles by ATL; and (3) permeation
emission estimates for plastic fuel tanks from EPA's Regulatory Support Document for
91
the regulations for 2006 and subsequent model year highway motorcycles.
Test results for uncontrolled motorcycles tested by ATL* and CARB are shown in
Table 3-1. All tests were run on California specification fuels at the time of the testing,
which ranged from 7 to 9 psi Reid Vapor Pressure. The resting loss portion of the diurnal
emissions measured by ATL was calculated from the total emissions that occurred during
the period of declining temperatures during a real-time, 24-hour diurnal emissions test.
Since the CARB data are available only for the accelerated diurnal test, they are assumed
to contain only a small fraction of actual resting loss emissions. Likewise, the CARB test
results do not include running loss emissions.
As shown in the first row of Table 1, evaporative emissions measured by ATL averaged
6.83 grams hot soak, 9.47 grams during the heat build portion of the diurnal, 4.00 grams
during the declining temperature portion of the diurnal (nesting loss"), and 1.08 g/mi for
running loss. The only two vehicles to receive a full complement of evaporative tests
were relatively small (200-250 cc) -eff-highway" motorcycles (a Honda XR200R and a
Yamaha WR250F). Although it would have been preferable to have had test results for
highway models, these were the only tests available with ^eal time" evaporative testing
(i.e., a 24-hour diurnal test) and running loss measurements.
The resting loss emissions were calculated from the continuous ATL emissions results.
Diurnal emissions occurring during the portion of the test when temperatures were
decreasing were considered nesting loss." The calculated resting loss emissions are
subtracted from total diurnal emissions to determine the emissions occurring during the
-heat build" portion of the diurnal. It should also be noted that the uncontrolled running
loss emissions used in the calculations are based on very limited data. The project
manager for the testing program reports that high levels of tank vibration occurred during
the testing as the result of the test vehicles being equipped with knobby tires.22 This
See http://www.arb.ca.gov/research/apr/past/00-315.pdf.
-30-
-------
Table 3-1
Uncontrolled Highway Motorcycle Evaporative Emissions
1.
2.
3.
4.
5.
6.
7.
8.
9.
ATL, Average of 2 Vehicles
CARB, Average of 35 Vehicles
Resting Loss Calculation
For Vehicle with Plastic Tank
Resting Loss Estimate
For Vehicle with Metal Tank
Resting Loss Calculation
For 90% Metal Tanks
Best Estimate of Uncontrolled
Evaporative Emissions
Grams/mile, assuming
8.5 mi/day, 2 hot soaks/day,
Full diurnal emissions ea. day
Grams/mile assuming
8.5 mi/day, 4 hot soaks/week
Diurnals reduced by 14%
Grams/mile assuming
30 mi/day, 2 hot soaks and a full
diurnal each day
Hot
Soak
6.83 g
8.69 g
-
-
-
8.69 g
2.04
0.58
0.58
Diurnal/
Heat Build
9.47 g
8.93 g
-
-
-
8.93 g
1.05
0.90
0.30
Resting
Loss
4.00 g
n.a.
8.6 g
2.0 g
2.66
2.66 g
0.31
0.31
0.09
Running
Loss
1.08g/mi
n.a.
-
-
-
1.08g/mi
1.08
1.08
1.08
TOTAL
-
-
-
-
-
-
4.48
2.87
2.05
vibration would be expected to contribute to vapor growth. Anecdotal evidence indicates
that many motorcycles actually have a net inflow of air to the gasoline tank during
normal operation because the effect of decreasing fuel level more than offsets the vapor
growth caused by relatively modest fuel heating associated with tanks that are not located
in close proximity to exhaust systems.
As shown in the second row of Table 3-1, the average hot soak and diurnal emissions
reported by CARB for 35 uncontrolled highway motorcycles were similar to the
emissions reported for the two off-highway motorcycles tested by ATL. The individual
hot soak emission results contributing to the 8.69 gram average ranged from just under
4 grams to just over 17 grams per test. Individual diurnal emissions results contributing
to the 8.93 gram average ranged from 1 to 20 grams per test.
Also shown in Table 3-1 is an estimate of resting loss emissions estimated using the
methodology described in EPA's Regulatory Support Document (RSD) for the 2008
model year permeation standards applicable to highway motorcycles. This calculation
was done because of the limited amount of resting loss data and the fact that the resting
-31-
-------
loss emissions measured by ATL may be higher than the highway motorcycle average
because both of the vehicles were equipped with plastic tanks.
The RSD estimates permeation emissions from plastic tanks at 1.32 g/gal/day, which is
3.7 grams for the average 2.8 gallon tanks on the vehicles tested by ATL. This is just
under the total resting loss emissions estimated from the ATL test results. For the
average 5 gallon tank size assumed by EPA, the permeation rate for uncontrolled plastic
tanks would be 6.6 grams per day (g/day).
9
The RSD estimates emissions from "R7"-specification fuel line at 873 g/m /day at 29° C
and assumes 1.5 feet of hose is used on the average motorcycle. Assuming 5/16" inner
diameter fuel lines, this translates to 9.95 grams per day, which is more than double the
resting loss emissions calculated from the ATL test data. If 3.7 grams of the resting loss
measured by ATL were from plastic tank permeation, the estimate for fuel line
permeation is high by a factor of over 30 times. It is also significant to note that CARB
evaporative emissions test results on several 1990s vintage motorcycles with evaporative
emissions control systems have diurnal emissions in the range of 0.1 grams measured
during a 1-hour test. If 100% of the diurnal emissions were from fuel line permeation,
daily emissions would be only 2.4 grams.
The available ATL and CARB data make it clear that permeation emissions from
motorcycle fuel lines are significantly lower than the estimates based on EPA's assumed
emission rate for R7 fuel line. This could be due to the fact that the fuel hose permeation
rate assumed by EPA is the upper limit allowed under the SAE R7 standard. Actual
emissions for typical fuel hoses, whether R7-spec or not, are obviously much lower.
Tests of two commercial R7-spec hoses presented in a 1988 SAE paper indicated
emissions as low as 70% below the maximum allowable.23
Based on the ATL and CARB test results, it appears that the maximum fuel hose
permeation rate for vehicles not subject to the 2006 and later model year permeation
standards is at or below 2 g/day. We have assumed this rate applies to pre-2006 models.
Row 3 of Table 3-1 shows that resting losses are estimated at 8.6 g/day for uncontrolled
vehicles with plastic tanks. This value is the sum of the 6.6 g/day estimate for a 5 gallon
tank and a 2.0 g/day estimate for fuel hose permeation. Row 4 shows the resting loss
estimate for uncontrolled motorcycles with metal fuel tanks, which includes only the
estimated emissions for fuel hoses. Row 5 shows the resting loss emissions for a fleet of
uncontrolled motorcycles with 10% plastic tanks.
Rows 7, 8, and 9 of Table 3-1 show the estimated fleet average evaporative emissions
translated into grams per mile. Rows 7 and 8 assume an average daily VMT of 8.5 miles,
which is equivalent to about 3,100 miles per year (an estimate supported by activity data
described below).
Row 7 shows the results for the assumption that the average motorcycle, like the average
car, experiences two hot soaks and one full diurnal per day. As described in the
-32-
-------
subsequent discussion of motorcycle usage, two hot starts per day is not a reasonable
assumption for the average motorcycle. Unlike passenger cars, most motorcycles are not
used on a daily basis. Because they are usually stored in garages when not being used,
the emissions associated with the full diurnal test also do not occur on a daily basis.
Row 8 shows the results based on the assumption that there are only 4 hot soaks per week
(which results in the same number of miles per hot soak as with passenger cars) and that
diurnal emissions are reduced by 20% on days when the vehicle is not driven. Table 5.1-
6 of EPA's Final Regulatory Impact Analysis done for the Marine SI and Small SI
Engines, Vessels, and Equipment rule indicates that diurnal emissions are reduced by
20% when a vehicle is stored in a garage.24 Survey data obtained by MIC indicate that
83% of motorcycles are stored in garages on days they are not driven.25 For vehicles
driven two days per week, a 20% reduction on the non-drive days results in a 14%
reduction in average diurnal emissions.
Row 9 shows the results for those motorcycles that are used like cars and driven about 30
miles per day with 2 hot soaks daily and one full diurnal. By comparing the last column
of row 7 with the last column of row 9, it is apparent that how the vehicle is used causes
evaporative emissions to vary by more than 100% on a g/mi basis.
Summary of Uncontrolled Emissions - Table 3-2 summarizes the analysis described
above with all emissions translated into grams per mile. Due to the extremely high
exhaust HC emissions of 2-stroke models, average exhaust HC for uncontrolled
motorcycles is about 24% higher than the estimate for uncontrolled passenger cars. In
contrast, uncontrolled CO and NOx emissions are significantly lower. Evaporative
emissions from motorcycles are about 30% lower than from uncontrolled passenger cars.
Uncontrolled crankcase emissions are 90% lower.
Table 3-2
Uncontrolled Highway Motorcycle Emissions (g/mi)
(FTP test conditions, 8.5 mi/day, 4 hot soaks per week, 5 reduced diurnal s/week)
Vehicle
2-Stroke
Motorcycle
4-Stroke
Motorcycle
Average
Motorcycle
(45% 2-stroke)
Passenger
Car
Exhaust
HC
19.2
4.7
11.2
9.0
CO
31.7
50.2
41.9
90
NOx
0.06
0.43
0.26
4.0
PM
0.11
0.01
0.06
-
HC+NOx
19.26
5.13
11.46
13.0
Evap
2.87
2.87
2.87
4.0
Crankcase
0
0.7
0.4
4.0
Total
HC+NOx
22.13
8.70
14.73
21.0
Note: Passenger car evaporative emissions based on 30 mi/day, 2 hot soaks per day. PM emissions assume
the use of unleaded fuel.
-33-
-------
3.3 Controlled Emissions
New highway motorcycles >50 cc have been subject to exhaust and crankcase emissions
standards since model year 1978, ten years after the first federal standards were applied
to passenger cars and light-duty trucks. Conventional evaporative emissions standards
have not yet been required for highway motorcycles under federal regulations; however,
2006 and subsequent model year motorcycles are required to use low permeation fuel
system components. In addition, evaporative emissions controls required by the State of
California since the 1984 model year have been voluntarily applied to some motorcycles
sold in other states.
Crankcase Emissions - Beginning with model year 1978, all highway motorcycles >50 cc
have been required to be designed so that —noxankcase emissions shall be discharged
into the ambient atmosphere." Although it is likely that some crankcase emissions are
emitted as the result of deterioration of breather hoses in customer service, no data were
identified on which a non-zero emission rate could be based.
Exhaust Emissions - Table 3-3 summarizes the motorcycle exhaust emissions standards
adopted by EPA and CARB. (Unlike the standards applicable to cars and trucks, the
motorcycle standards may be met based on each manufacturer's sales-weighted average
emissions of the various models sold within each class.) The California standards are
relevant to the emission factors used in MOVES for two reasons. First, as discussed in
more detail below, many motorcycles sold in all 50 states were voluntarily certified to the
California standards to avoid the cost associated with distributing two different types of
motorcycles. Second, even in cases where a manufacturer produces a California and -49-
state" version, the California-certified model is sometimes sold in other western states.
As shown in the first three rows of Table 3-3, EPA and CARB standards were identical
for all 1978 and 1979 models. Class I vehicles were required to meet a 5 g/km (8.05
g/mi) HC standard and a 17 g/km (27.4 g/mi) CO standard. The HC standard for larger
displacement models increased in proportion to engine displacement to a maximum of
14 g/km (22.5 g/mi). These could be considered transitional standards intended to initiate
a phase-out of conventional 2-stroke engines.
For model years 1980 and 1981, a 5 g/km HC standard applied to all displacements,
which eliminated all but a few 2-stroke models. The federal CO standard was reduced to
12 g/km (19.3 g/mi) for all displacements.
From 1980 through 2005, the federal standards remained unchanged; however, the
California standards become more stringent starting with the 1982 model year when
Class I and Class II vehicles become subject to a 1.0 g/km (1.61 g/mi) HC standard and
Class III vehicles become subject to a 2.5 g/km (4.0 g/mi) HC standard. For 1986 and
1987, the HC standard for Class III California motorcycles is reduced to 1.4 g/km
(2.25 g/mi).
-34-
-------
Table 3-3
Motorcycle Exhaust Emission Standards (grams per kilometer, g/km)
Model
Year
1978-1979
1978-1979
1978-1979
1980-1981
1982-1985
1982-1985b
1986-1987
1986-1987c'd
1988-2003
1988-2003
2004-20056
2004-2005
2006-2007
2006-2007
2008-2009
2008-2009
>2010
>2010
Engine
Size (cc)
50-169
170-749
>750
>50
50-279
>280
50-279
>280
50-699
>700
50-279
>280
0-279
>280
0-279
>280
0-279
>280
Federal Standards
HC
5.0
5-14a
14
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
1.0
n.a.
1.0
n.a.
1.0
n.a.
CO
17
17
17
12
12
12
12
12
12
12
12
12
12
12
12
12
12
12
HC+NOx
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
1.4
n.a.
1.4
n.a.
0.8
California Standards
HC
5.0
5-14a
14
5.0
1.0
2.5
1.0
1.4
1.0
1.4
1.0
n.a.
1.0
n.a.
1.0
n.a.
1.0
n.a.
CO
17
17
17
17
12
12
12
12
12
12
12
12
12
12
12
12
12
12
HC+NOx
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
1.4
n.a.
1.4
n.a.
0.8
n.a.
0.8
a 5.0 + 0.0155 (D-170), where -9" is engine displacement in cubic centimeters
b Note: California standards apply to vehicles produced prior to March 1, 1985
0 Note: California standards apply to vehicles produced after February 28, 1985
d Compliance is based on corporate average after February 28, 1985 for California and 2006 federally.
e Note: Class I and II motorcycles subject to the 1.0 g/km HC standard, beginning in 2004 for California
models and 2006 for federal models, may be optionally certified to a 1.4 g/km HC+NOx standard.
For 1988 to 2003, CARS extended the 1.0 g/km HC standard to all motorcycles up to
699 cc displacement. Motorcycles >700 cc remained subject to a 1.4 g/km HC standard.
Beginning with model year 2004, CARB revised its standards for Class III motorcycles to
incorporate an HC+NOx standard and to delete the separate HC standard. Because
uncontrolled NOx emissions are about 0.3 g/km, the new 1.4 g/km HC+NOx standard
requires motorcycles >700 cc to have HC+NOx emissions about 18% lower than under
the previous HC-only standard. In 2008, CARB's HC+NOx standard for Class III
motorcycles is reduced from 1.4 g/km to 0.8 g/km (1.29 g/mi). Beginning in 2004,
Class I and II motorcycles subject to the 1.0 g/km HC standard may optionally be
certified to a 1.4 g/km HC+NOx standard.
The federal 2006 and 2010 model year standards were set to be equivalent to the 2004
and 2008 California standards. The federal and California exhaust emission standards are
therefore identical for 2010 and subsequent model years.
-35-
-------
By comparing the standards in Table 3-3 to the uncontrolled emission levels described
above, it is apparent that nearly a 50% reduction in CO emissions was required for
4-stroke motorcycles beginning in model year 1978. Because there were no NOx
standards, the most cost-effective way for manufacturers to meet the CO standards
involved enleanment of the air-fuel ratio. This also reduced hydrocarbon emissions by an
amount that was generally sufficient to comply with the CARB HC standard through
model year 1985. As a result, 1978 through 1985 model 4-stroke motorcycles have
similar emissions characteristics.
Regulated Pollutants - The available data on 1978-1985 models from SwRI and CARB
are displayed in Figures 3-7, 3-8, and 3-9. As was the case with pre-1978 models, there
is still a wide variation in test results from model to model. It is also apparent that there
is little trend in emissions vs. mileage accumulation. The average of tests of 55
Figure 3-7
Exhaust Hydrocarbons vs. Odometer
1978-1985 4-Stroke Models
12
10
y = 0.00000165X + 2.50546828
R2= 0.00018883
* *
20,000 40,000
Odometer Miles
60,000
80,000
-36-
-------
Figure 3-8
Exhaust Carbon Monoxide vs. Odometer
1978-1985 4-Stroke Models
!
o
o
CL
20,000 40,000
Odometer Miles
60,000
80,000
0.8
Figure 3-9
Exhaust Oxides of Nitrogen vs. Odometer
1978-1985 4-Stroke Models
y - -0. 00000086X + 0. 44384230
* R2= 0.00667984
20,000 40,000
Odometer Miles
60,000
80,000
-37-
-------
individual 1978-1985 model year vehicles indicated FTP emissions of 2.534 g/km HC,
18.041 g/km CO, and 0.429 g/km NOx.
Figures 3-10, 3-11, and 3-12 present the available test results for 1986-1999 model year
motorcycles from CARB's 1998 surveillance testing program. The California standards
applicable to these model years ranged from 1.0 to 1.4 g/km HC depending on the
displacement class.
As with earlier models, there is a wide variation in test results from model to model.
Models with relative high emissions continue to be allowed because compliance with the
California standards is based on corporate average emissions beginning with vehicles
produced after February 1985. The average of tests of 59 individual 1986-1999 model
year vehicles indicated FTP emissions of 1.34 g/km HC, 13.31 g/km CO, and 0.36 g/km
NOx. Since only one Class I vehicle and one Class II vehicle were tested, there are
insufficient data to distinguish between the three classes. The use of certification data to
distinguish between the classes is frustrated by the lack of NOx data.
Although the federal exhaust emission standards remained at 5 g/km HC from 1986
through 2005, the majority of models certified for sale actually met the California exhaust
emissions standards. Thus, the emission levels depicted in Figures 3-10 through 3-12 are
representative of 49-state motorcycles.
Figure 3-10
Exhaust Hydrocarbons vs. Odometer
1986-1999 Models
o
y = 0.00001013x+ 1.16188026
R2= 0.04526414
20,000
40,000 60,000 80,000 100,000
Odometer Miles
-38-
-------
Figure 3-11
Exhaust Carbon Monoxide vs. Odometer
1986-1999 Models
y = 0.00002083X + 12.94940988
R2= 0.00228756
20,000
40,000 60,000
Odometer Miles
80,000
100,000
Figure 3-12
Exhaust Oxides of Nitrogen vs. Odometer
1986-1999 Models
y = 0.00000168x + 0.32708159
R2= 0.02395603
0.0
20,000 40,000 60,000
Odometer Miles
80,000
100,000
-39-
-------
There appears to be no significant trend in emissions vs. vehicle age for motorcycles of
this vintage based on idle emissions data collected in the Arizona I/M program. Figures
3-13 and 3-14 illustrate the relationship between emissions and vehicle age using data
from the late 1990s.26
Beginning with model year 2004 in California and 2006 federally, the HC-only standard
is eliminated for Class III motorcycles and large-volume manufacturers are required to
meet a 1.4 g/km HC+NOx standard. The CO standard remains unchanged at 12 g/km.
Beginning with model year 2008 in California and 2010 federally, the HC+NOx standard
for Class III motorcycles is tightened to 0.8 g/km (1.29 g/mi). Only very limited
emissions data are available for motorcycles recruited from customer service that are
subject to the HC+NOx standards. As a result, an alternative technique is required in
order to estimate emissions in customer service.
Analysis of California certification data for model year 2009 Class III motorcycles
indicates that the average HC+NOx emission rate reflects a 30% compliance margin for
the applicable 0.8 g/km standard. Average HC+CO emissions for Class III models were
0.56 g/km. (This represents a 67% reduction from the average 1.70 g/km HC+NOx
emissions of 1986-1999 models.) Average CO emissions were 4.0 g/km. Average
emissions for the 84% of the 2009 Class III catalyst-equipped models were 0.45 g/km
HC+NOx and 3.57 g/km CO. The non-catalyst models averaged 1.14 g/km HC+NOx
and 6.16 g/km CO. Based on an analysis of 2008 model year certification data, HC
exhaust emissions are about 60% of HC+NOx for both catalyst and non-catalyst models.
Idle HC Veriii^f^Brcvcle Age
Idlf HC. Versus Motorcycle Age.
?
Q.
a.
O
I
0)
2
1 ,UUU
900
800
700
600
500
400
300
200
100
0
C
*
+
*
•» A A
V * *
* 4> +
*
*
+
i i i i i i i i
) 2 4 6 8 10 12 14 16 1
Age (years)
-40-
-------
Idle CO VeiiGW-Mbtfcrcycle Age
Arizona
Ariz
rizona
- Tiny to September 199
2.5
8
0.5
+
— * * • <
* *
^> ,
* * * * * •
1 1
D 2 4
l l l l l l
6 8 10 12 14 16 1
Age (years)
Because corporate averaging was allowed, compliance with the 1.4 g/km HC+NOx
standard required for model years 2004-2007 in California was generally achieved by
phasing in vehicles designed to meet the 0.8 g/km standard that became effective in 2008.
As a result, model years subject to the 1.4 g/km -^Fier 1" standard can be represented by a
fleet with a smaller fraction of Tier 2 compliant vehicles. Only 25% of the Tier 2
vehicles are required to achieve the Tier 1 standard assuming the emissions of the non-
catalyst vehicles are the same as 1986-1999 models. However, maintenance of a 30%
compliance margin required about a 60/40 split of Tier 2 catalyst vehicles and vehicles
with the average emissions of the 1986-1999 models. This yields a fleet average of
0.95 g/km HC+NOx and 7.5 g/km CO.
Based on analysis of 2008 model year certification test results, the average HC+NOx and
CO exhaust emissions for Class I and Class II motorcycles are only slightly higher than
the emissions from Class III models. NOx data were not available for all models (since
compliance with the optional HC+NOx standard is voluntary); however, based on the
available data, the HC+NOx emissions are 0.85 g/km for Class la, 0.91 g/km for Class Ib,
and 0.95 g/km for Class II. CO emissions were 5.2 g/km for Class la, 6.5 g/km for
Class Ib, and 7.1 g/km for Class II. Although catalytic converters are used on some
Class I and Class II models, an insignificant catalyst removal rate is anticipated in
customer service.
-41-
-------
Unlike for non-catalyst motorcycles, catalyst-equipped motorcycles are not expected to
maintain the same emissions in customer service that are demonstrated during
certification testing. Survey data collected by the Motorcycle Industry Council indicate
that replacement of OEM exhaust systems is a common occurrence on highway
motorcycles. Although CARB has recently adopted regulations that provide for the
certification of catalyst-equipped aftermarket exhaust systems, such systems will be more
expensive than other aftermarket systems and generally quieter (because of the muffling
effect of the catalyst). Given the current lack of any effective enforcement mechanism, it
does not appear likely that the availability of catalyst-equipped aftermarket systems will
have much effect on the emissions increase associated with the removal of OEM systems.
Surveys conducted over the last 10 years indicate an OEM replacement rate of 38%.
Analysis of the 2008 model year CARB certification data indicates that catalysts are used
on 84% of Class III motorcycles. Replacement of 38% of the OEM exhaust system will
therefore eliminate the emission reduction accomplished by the catalyst on 32% of the
vehicles.
To account for the effect of exhaust system replacement, the in-use emissions from
vehicles certified to meet the 2008 California and 2010 federal standards can be
estimated based on the following assumptions:
1. 52% of the vehicles will have emissions equivalent to the certification levels
achieved by catalyst equipped motorcycles;
2. 32% of the vehicles will exhibit emissions similar to 1986-1999 models; and
3. 16% of the vehicles will have emissions equivalent to the certification levels
achieved by non-catalyst motorcycles.
The assumption regarding emissions from motorcycles that have had their catalyst
removed is consistent with data recently collected by CARB.* Based on these
assumptions, fleet average emissions will be 0.96 g/km (1.54 g/mi) HC+NOx and
7.1g/km(11.4g/mi)CO.
After accounting for catalyst tampering, the estimated emissions for Tier 2 certified
motorcycles are almost identical to the certification levels for Class I and Class II
motorcycles. Since tampering is not expected to be significant for Class I and Class II
models, the emission rates estimated for Tier 2 certified models can be applied to Class I
and Class II models as well.
To estimate the effect of tampering on the fleet subject to Tier 1 standards, we have
assumed that 37% of the vehicles emit at levels of catalyst-equipped Tier 2 vehicles
* Unpublished data provided by CARB from Project 2R0814. CARB tested four late-model motorcycles
originally equipped with catalytic converters. With catalysts removed, the HC+NOx emissions averaged
1.59 g/km and CO emissions averaged 8.2 g/km.
-42-
-------
(reflecting a 38% tampering rate on 60% of the vehicles) and the remainder emit like the
average of the 1986-1999 models. This yields in-use emissions from vehicles certified to
meet the 2004 California and 2006 federal standards of 1.24 g/km HC+NOx and
9.7 g/km CO.
Paniculate Emissions - In the absence of data on U.S. specification motorcycles, the
Ricardo estimates for 4-stroke motorcycles running on unleaded gasoline described above
are assumed to be applicable to all controlled motorcycles.
Evaporative Emissions - California motorcycles have been subject to evaporative
emissions standards since a phase-in that began in 1983 and was completed in 1984. The
applicable test procedure is the one-hour SHED test that only partially captures resting
loss emissions and does not require either real time or multiday diurnal tests. Federal
standards require the use of low permeation fuel tanks and fuel hoses beginning in model
year 2008. Table 3-4 summarizes our estimates of evaporative emissions from
motorcycles with some form of evaporative controls.
Hot soak and diurnal emissions test results for six California-certified motorcycles
equipped with evaporative emissions controls are shown in the first row of Table 3-4.
Because a simple 1-hour SHED test was used, permeation and running loss emissions are
not accounted for. As shown in the second row of the table, running loss emissions are
assumed to be eliminated from a motorcycle equipped with a canister and purge system.
Rows 3, 4, and 5 of the table repeat the resting and running loss estimates presented
earlier in the discussion of uncontrolled evaporative emissions. These estimates apply to
California-certified motorcycles prior to the requirement for low permeation fuel tanks
and fuel hoses. The third row of Table 3-4 shows the resting loss calculation for a
motorcycle with a plastic tank. The fourth row of the table shows the resting loss
estimate for just the fuel hoses that was described previously. Row 5 of the table shows
the resting loss estimate for a fleet of motorcycles with 10% plastic tanks.
Row 6 combines the estimates from rows 1, 2, and 5, and provides our estimates of each
category of evaporative emissions from 1984-2007 model year California-certified
motorcycles. Row 7 translates the evaporative emissions shown in row 6 into grams per
mile based on a daily average VMT of 8.5 miles and 4 hot soaks per week. Diurnal
emissions are estimated based on the assumption that 2 diurnal events per week produce
emissions equal to the SHED test results and, on the other 5 days, diurnal emissions are
estimated to be 32% of the uncontrolled diurnal emissions of a motorcycle parked
outside.
* It is theoretically possible for a canister-equipped motorcycle to have sufficient vapor generation during
operation to cause breakthrough, but such systems would have difficulty during the hot soak test because
the canister would already be saturated.
-43-
-------
Table 3-4
Controlled Highway Motorcycle Evaporative Emissions
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Tests of 6 CA-certified
Vehicles
Estimated Running Loss for
CA-certified Vehicles
Pre-2008 Resting Loss
Estimate For Vehicle with
Plastic Tank
Pre-2008 Resting Loss
Estimate For Vehicle with
Metal Tank
Pre-2008 Resting Loss For
90% Metal Tanks
Best Estimate of Controlled
Emissions, 1984-2007 CA
Grams/mile, 1984-2007 CA,
assuming 8.5 mi/day, 4 hot
soaks/week, 71% of diurnals
reduced by 68%
As Above, Plus 15%
Tampering
Grams/mile, 2008 CA,
assuming 8.5 mi/day, 4 hot
soaks/week, 71% of diurnals
reduced by 68%
As Above, Plus 15%
Tampering
Grams/mile, 2008 federal,
assuming 8.5 mi/day, 4 hot
soaks/week, uncontrolled
diurnals reduced by 14%
Hot
Soak
0.71 g
-
-
-
-
0.71 g
0.05
0.13
0.05
0.13
0.58
Diurnal/
Heat Build
0.48 g
-
-
-
-
0.48 g
0.26
0.36
0.26
0.36
0.90
Resting
Loss
n.a.
-
8.6 g
2.0 g
2.66
2.66 g
0.31
0.31
0.03
0.03
0.03
Running
Loss
n.a.
0 g/mi
-
-
-
0 g/mi
0
0.16
0
0.16
1.08
TOTAL
-
-
-
-
-
-
0.62
0.96
0.34
0.68
2.59
-44-
-------
The diurnal emissions during the 5 days per week that the average California-certified
motorcycle is not operating were calculated based on EPA's estimate from the earlier
referenced Final Regulatory Impact Analysis that canisters reduce diurnal emissions by
60% when they are not purged between repeated diurnal cycles. The RIA summarizes
the rationale for this estimate as follows:
...we have collected information showing that, during cooling periods, the
canister is purged sufficiently enough so that it can be used effectively to reduce
diurnal emissions. When the fuel in the tank cools, fresh air is drawn back
through the canister into the fuel tank. This fresh air will partially purge the
canister and return hydrocarbons back to the fuel tank. Therefore, the canister
will have open sites available to collect vapor during the next heating event. Test
data presented below show that a canister that starts empty is more than 90
percent effective at capturing hydrocarbons until it reaches saturation. Once the
canister reaches saturation, it is still capable of achieving more than a 60 percent
reduction in diurnal emissions due to passive purging. Passive purging occurs as
a result of fresh air that is pulled through the canister during fuel tank cooling
periods.
As shown in the last column of row 7, the total evaporative emissions for a 1984-2007
model year California-certified motorcycle are estimated to be 0.62 g/mi, which
represents a 78% reduction from the uncontrolled evaporative emission rate of 2.87 g/mi.
Row 8 of the table accounts for tampering with evaporative emission control systems on
motorcycles in customer service. Sierra was unable to identify any existing survey data
regarding evaporative system tampering, so a preliminary survey was conducted of the
members of an Internet forum of motorcycle enthusiasts (www.ldriders.com). A
question was posed to members of the forum by one of the forum administrators who
requested private responses and guaranteed that the respondents would remain
anonymous. Owners of motorcycles that were originally equipped with evaporative
canisters (i.e., 1984 and later models that were California or 50-state certified) were
asked to report whether the canister was still in place. The make and model year of the
motorcycle were also reported.
Based on 78 responses to the above-described survey, canisters had been removed from
28% of the motorcycles owned by the forum members. The removal rate was higher for
Unlike forums focused on a particular make and model of motorcycle, this forum includes members
owning a broad range of makes and models. The focus of the forum is long distance riding and many of
the forum topics address modifications to motorcycles to make them more suitable for long distance riding,
such as auxiliary lighting, aftermarket saddles, GPS systems, and auxiliary fuel systems. The members are
therefore frequently involved in modifying their motorcycles and are probably more likely to have
tampered with evaporative emissions control system than the average owner.
t Responses usually were rejected if the owner did not know whether the motorcycle was originally
equipped with a canister or did not know where the canister was located. The exception was in cases where
the make and model was known to be 50-state certified and the original owners reported that, although they
didn't know how to identify the canister, they had not removed any components.
-45-
-------
the older models, many of which were near the end of their life. The removal rate for
models up to five years old, the vehicles that have the highest annual average mileage
accumulation, was approximately 19%. Based on this survey, 15% has been used as a
conservative estimate of the canister removal rate for two reasons. First, motorcycle
owners in the survey population are more likely to have tampered with their emissions
control system than the average owner. Second, there is a likely reporting bias associated
with the fact that owners who don't know whether their motorcycle was originally
equipped with a canister would be unlikely to respond (and certainly wouldn't have
removed the canister).
Row 8 of Table 3-4 shows the effects of the assumed 15% canister removal rate for
1984-2007 model year California-certified motorcycles. For hot soak, diurnal, and
running loss, 15% of the fleet is assumed to have the uncontrolled evaporative emissions
reported earlier in Table 1 and 85% of the fleet is assumed to have the controlled
emissions estimates from row 8 of Table 3-4. The net effect is that total evaporative
emissions are 0.96 g/mi, which is 67% below the uncontrolled estimate.
Row 9 of the table shows the effect of reducing resting losses by 90% to account for the
use of gasoline tanks and fuel hoses meeting the permeation standards required by EPA
regulations for 2008 and subsequent models. Ignoring the effect of tampering, the total
evaporative emissions of motorcycles certified to meet the California evaporative
emissions standards are reduced to 0.34 g/mi, which is 88% below uncontrolled levels.
Row 10 of the table shows the estimated evaporative emissions for 2008 and subsequent
model year California-certified motorcycles when tampering is taken into account. The
0.68 g/mi emission rate is 76% below uncontrolled emissions.
The final row of Table 3-4 shows our estimate for the evaporative emissions of 2008 and
subsequent model year federally certified motorcycles. The 2.59 g/mi composite
emission rate is only 10% below uncontrolled emissions.
Summary of Controlled Emissions - Table 3-5 summarizes the analysis described above
with all emissions translated into grams per mile. Due to the combined effects of
crankcase, exhaust, and permeation standards, 2010 and later model federally certified
motorcycles have 72% lower HC+NOx emissions than uncontrolled, pre-1978 models.
Due to the evaporative emissions standards, the HC+NOx emissions of California-
certified models are 85% lower than uncontrolled. Both California and federal
motorcycles have CO emissions that are 73% lower than uncontrolled.
Because of the relatively high fraction of models meeting California exhaust emissions
standards that are sold in all 50 states, relatively little error is introduced by using the
exhaust emissions estimates for 2004 and later California models for all 50 states.
However, California evaporative emissions systems are not routinely used on models
produced for sale outside of California.
-46-
-------
Table 3-5
Controlled vs. Uncontrolled Emissions (g/nii)
(FTP test conditions, 8.5 mi/day, 4 hot soaks per week,
2 regular diurnals per week, 5 multiday diurnals per week)
Vehicle
Pre-1978
Motorcycles
1978-1985
Motorcycle
1986-2003
Motorcycle
2004-2007
California
2006-2009
Federal
>2008
California
>2010
Federal
Exhaust
HC
11.2
4.07
2.16
1.51
1.51
1.09
1.09
CO
41.9
29.0
21.4
15.6
15.6
11.4
11.4
NOx
0.26
0.69
0.58
0.47
0.47
0.45
0.45
PM
0.06
0.01
0.01
0.01
0.01
0.01
0.01
HC+NOx
11.46
4.76
2.74
1.98
1.98
1.54
1.54
Evap
2.87
2.87
2.87
0.96
2.87
0.68
2.59
Crankcase
0.4
0
0
0
0
0
0
Total
HC+NOx
14.73
7.63
5.61
2.94
4.85
2.22
4.13
MOVES model year groups 70 (1960-1970) and 51 (1971-1977) can be used for
pre-1978 models. Groups 6 and 7 can be used for 2011 and subsequent models. Most
intermediate model years need to be in model year specific groups for greatest accuracy;
however, Groups 61, 62, 63, 64, and 65 could also be used to minimize the number of
groups. We recommend that California emission factors, except for evaporative
emissions, be used for all 2004 and subsequent model years.
3.4 Inspection and Maintenance Effect
Arizona is the only state in the nation that includes motorcycles in its Inspection and
Maintenance (I/M) program. Our review of the available documentation27 indicates that
no mass emissions data have ever been collected that can be used to provide a meaningful
estimate of the benefits associated with subjecting motorcycles to I/M. The estimates of
emission reductions made by the State of Arizona are based on the assumption that
motorcycles will experience benefits proportional to those predicted by MOBILE6 for
light-duty vehicles, with certain adjustments made to account for known differences in
starts per day and VMT. In addition, an adjustment was made to account for the
-47-
-------
observed idle emissions test results for passing and failing vehicles in the I/M lanes. (The
difference between the idle emissions of failing vehicles and the idle emissions standard
was apparently assumed to be proportional to excess emissions.) Based on the
information presented in the above-referenced report, the State of Arizona's 2004
calendar year estimate of the benefits of subjecting motorcycles to I/M was a 13.5%
reduction in exhaust HC and a 20.1% reduction in exhaust CO.
Undoubtedly the Arizona I/M program required some motorcycles to have their idle air-
fuel ratio adjusted in order to pass the test. However, the net effect on emissions in
customer service is unclear for several reasons. First, there are no data correlating
changes in the concentration of HC and CO emissions at idle to mass emissions changes
in customer service. Second, there is no evidence that the air-fuel ratio adjustments
required to pass the idle test were permanent. Third, the previously presented data
showing the lack of correlation between failure rate and vehicle age raise questions about
the appropriateness of the idle emissions standards being used.
The lack of correlation between the idle emissions failure rate and vehicle age described
above is consistent with the lack of a significant trend between emissions and odometer
in the surveillance testing data available from CARB. It therefore appears that the idle
standards being used by Arizona were merely identifying that fraction of the motorcycle
fleet at one end of a fairly broad distribution of idle emission levels. Under these
circumstances, it does not appear reasonable to estimate I/M benefits based on a
calculation that assumes excess emissions from failing vehicles are proportional to the
extent to which their idle concentrations exceed the I/M standards.
The potential benefits of I/M may be fundamentally different for later model motorcycles
originally equipped with catalytic converters. Unlike for non-catalyst motorcycles, there
is fairly solid evidence of significant emissions deterioration in customer service
associated with the replacement of OEM exhaust systems. However, there is great
uncertainty regarding how effective I/M will be as a deterrent to catalyst removal. Unlike
with light-duty vehicles, the exhaust systems on motorcycles can be changed rather
quickly, typically in less than one hour. Re-installation of an OEM system for one day to
pass an I/M test may be the approach used by many owners who prefer to use an
aftermarket exhaust system on a routine basis.
In addition to the likelihood of repeat tampering, there are other practical problems
associated with using a conventional I/M program as a deterrent to catalyst removal.
There is no available analysis indicating that emission testing (dynamometer or idle) will
be effective in identifying tampered vehicles. The existence of small-volume
manufacturer standards and corporate averaging affects the feasibility of making an
accurate pass/fail decision based on emissions test results. The HC+NOx standards that
apply to recently certified motorcycle models range from 0.3 to 2.5 g/km HC + NOx.
This is substantially greater than the difference in emissions associated with catalyst
removal. It is also likely that visual inspections will be ineffective because OEM
catalysts are often hidden. Checking for the presence of the EPA noise certification
-48-
-------
stamp may also be ineffective because the stamp is not placed on the catalyst itself; the
stamp can remain when the catalyst has been removed.
For the reasons described above, the effective deterrence of catalyst tampering is likely to
require something other than a conventional annual or biennial I/M program. There is
greater potential for deterring tampering through more routine enforcement of noise
standards. A substantial increase in noise level is typically associated with the use of
aftermarket exhaust systems, especially when the system change involves removal of the
OEM catalyst because catalysts are effective mufflers.
Until data are available demonstrating the actual effectiveness of a particular type of
inspection program, we do not believe there is a reasonable basis for assigning a specific
emissions benefit to the inclusion of motorcycles in a conventional I/M program.
3.5 Motorcycle Populations and Activity
Population - The population of highway motorcycles in customer service is difficult to
accurately estimate and forecast for two reasons. First, registration data collected by
individual states often fail to distinguish between highway motorcycles, off-highway
motorcycles, and all-terrain vehicles. Second, the annual sales volume of new highway
motorcycles is much more variable than annual sales of other highway vehicles because
motorcycles are primarily used for recreational purposes and sales are extremely sensitive
to the overall level of economic prosperity. During economic recessions, purchases of
recreational vehicles decline much more than the purchase of other vehicles.
In model year 1980, the nationwide sales volume of highway motorcycles (including
scooters and dual sport models) was 757,000 units. By 1990, the annual sales volume
had declined by 70% to 226,000 units.28 Sales stayed below 300,000 per year until 1998
9Q
and then started increasing rapidly, reaching 920,463 for model year 2008. Projections
for model year 2009 are that sales may decline by 40-50%.30 How long this recent
decline in sales will continue is impossible to accurately forecast.
In addition to the volatility in total highway motorcycle sales, several trends are apparent.
The scooter population has expanded substantially in recent years. Just 10 years ago (in
model year 1999), the sales of new scooters was 7,500 units per year, which was less than
2% of total highway motorcycle sales. Sales increased dramatically each year through
model year 2008 when they reached 222,155 and accounted for 24% of total highway
motorcycle sales. In addition to increasing sales, the annual average travel per scooter
has increased by 38% in the last 10 years, to 1,331 miles per year.31 The expanding use
of scooters is correlated with the increased use of highway motorcycles by women. As
recently as 1990, only 6% of motorcycle riders were female. In 2008, 12% of all
motorcycle riders were female and 27% of scooter riders were female.31
Despite the increasing fraction of scooters, the average size of all highway motorcycles is
increasing. In 1998, only 27% of highway motorcycles were >1100 cc. By 2008, the
-49-
-------
percentage of motorcycles >1100 cc had grown to 38%.31 Average engine displacement
increased from 690 cc to 833 cc during this same period.
In addition to becoming larger, highway motorcycles are lasting longer. The average age
of motorcycles in customer service has gone down slightly in recent years, but this is due
to the rapid increase in the sale of new motorcycles. When sales and registration data are
jointly analyzed, it is apparent that the average life of each new motorcycle is increasing.
To develop the population data required by MOVES, Sierra performed a detailed analysis
of the highway motorcycle population and scrappage rates using a combination of
registration data and new vehicle sales data provided by MIC. Because the registration
data were proprietary and confidential, Sierra agreed not to retain the data at the
completion of the analysis.
The registration data were compiled by R.L. Polk for the 2008 calendar year. The data
base included the total number of U.S. registrations for 9,156 unique combinations of
make, model, and model year. Each entry was categorized as either -0n Highway," -Off
Highway," -On/Off," -Scooter," Sloped," -XIV," or -Unknown" -Segment," -^ody
Style," -Cylinder" count, -Displacement" (in cubic centimeters), and -Stroke" (2-stroke
or 4-stroke) were also identified.
The level of detail in the registration data should have been sufficient to readily identify
the total number of registered vehicles legal for highway use. Unfortunately, spot
checking of the data uncovered numerous problems. Motorcycle models that were
clearly legal for highway use were sometimes categorized as -Off Highway."
Motorcycles that were clearly illegal for highway use were sometimes categorized as -0n
Highway" or -On/Off." Some models labeled as ATVs were actually motorcycles.
Sierra made hundreds of edits to correct errors in the registration data. The primary
editing technique involved independently documenting the type of vehicle from the
alpha-numeric model name.
Another problem with the Polk database is that only 36 specific manufacturers were
identified. Other manufacturers were lumped together under the category -Other
Motorcycle" and no detail was available for models produced by -Other" manufacturers
with the exception of model year. While it might appear that detailed information
regarding 36 specific manufacturers would be sufficient to cover over 99% of sales, that
is not the case due to the relatively large volumes of scooters produced by a variety of
Chinese manufacturers. Our analysis of the Polk data is based on the assumption that all
of the models produced by -Other" manufacturers are Class I scooters.
Table 3-6 summarizes our analysis of the retail sales and registration data after removing
non-highway motorcycles from the registration data. All sales data in the table were
provided by MIC. The last column of the table shows the -survival fraction," which is
calculated by dividing each model year's population estimate by the original sales for that
model year. (The population for 2008 has been set equal to the sales estimate because the
registration data for 2008 were incomplete.)
-50-
-------
Table 3-6
On-Highway Motorcycle Population and Survival
Estimated from Registration and Retail Sales Data
Model Year
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
1981
Pre-1981
TOTAL
Age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
n.a.
Population
920,463
818,017
782,641
685,914
516,665
594,481
484,488
405,379
320,950
260,149
192,851
163,237
161,308
134,027
112,809
100,802
73,690
57,066
58,631
64,862
61,991
78,866
121,352
107,104
84,732
104,797
141,210
109,761
516,761
7,729,882
Original Sales
920,463
885,000
892,000
831,000
750,000
675,300
640,000
577,000
490,000
394,000
311,000
260,000
242,000
230,000
228,000
217,000
203,000
206,000
226,000
245,000
335,000
465,000
470,000
565,000
605,000
605,000
575,000
695,000
n.a.
Survival Fraction
Raw
1.00
0.92
0.88
0.83
0.69
0.88
0.76
0.70
0.66
0.66
0.62
0.63
0.67
0.58
0.49
0.46
0.36
0.28
0.26
0.26
0.19
0.17
0.26
0.19
0.14
0.17
0.25
0.16
n.a.
Smoothed
1.00
0.98
0.92
0.86
0.81
0.76
0.72
0.67
0.63
0.59
0.56
0.53
0.49
0.46
0.44
0.41
0.38
0.36
0.34
0.32
0.30
0.28
0.26
0.25
0.23
0.22
0.21
0.19
n.a.
As shown in the next to last column of Table 3-6, the survival fraction does not smoothly
and continuously decrease with age. The source of this unexpected result is not clear;
-51-
-------
however, it should be noted that highway motorcycles are primarily recreational vehicles
and it is common for owners to temporarily let the registration lapse on recreational
vehicles after the first several years of use. The last column of the table shows the
recommended survival fraction for use in MOVES.
Figure 3-15 shows the curve fit to the age vs. survival fraction data from Table 3-6 that
was used to developed the smoothed estimate. The equation represented by the curve is
Survival Fraction = C0 * e"(C1*A), where C0 is 1.1089 and Ci is 0.0623. The coefficient of
r\
determination (r ) is 0.92.
Figure 3-15
Motorcycle Survival Fraction vs. Age
ffcn Analogic ffcf Pfurigj-yqi-iffcn onri T?<*iqil W
1.0 i *-^s
0.9 •
0.8
0.6
- 0.5
| 0.4
(/>
0.3
0.2
0.1
0.0
10 15 20 25 30
Age (years)
Due to the uncertainty associated with the category of vehicles produced by -Other"
manufacturers in the Polk data, we are recommending the same survival curve for all
classes of highway motorcycles, including scooters. Although the lifetimes indicated by
the survival curve may seem excessive, this is due to the fact that many owners retain
motorcycles that they are no longer riding.* This is supported by the discussion below
regarding mileage accumulation rates as a function of age.
As noted previously, sales of highway motorcycles are down significantly for model year
2009. Our recommendation is that 2009 model sales be set at 50% of 2008 model year
sales. For 2010 and 2011, we are recommending that sales be estimated at 67% and 83%
* Motorcycles are more easily stored for extended periods of time than passenger cars or light-duty trucks
because of their relatively small size.
-52-
-------
of 2008, respectively. For 2012 and later model years, we are recommending that sales
be set at 100% of 2008 values.
Activity - Although MIC provided VMT as a function of age from its most recent survey,
sample sizes for older vehicles were insufficient. In addition, the questions asked of
survey respondents did not address total accumulated mileage. Based on the responses
received, a question regarding mileage accumulated during the previous 12 months may
have been misinterpreted by some respondents. Occasionally, responses from owners of
vehicles more than 10 years old were representative of what would have been anticipated
from the vehicle's odometer reading. Because of concerns with the MIC survey data,
Sierra used odometer readings listed in advertisements for motorcycles offered for sale at
www.cycletrader.com. Advertisements were sorted in order of their publication date for
each of 25 different model years ranging from 1975 through 2007. Initial results
indicated that the average of the odometer readings tended to stabilize at a sample size of
approximately 60. Except for the oldest model years, where insufficient data were
available, the most recent 60 ads with odometer values were recorded for each model
year. To avoid a manufacturer-specific bias, ads for Harley-Davidson motorcycles were
limited to one-third of the records for each model year. A total of approximately 1,500
ads were included in the sample. The results are presented in Figure 3-16.
Figure 3-16
Average Annual Lifetime Mileage vs. Age
7,000
6,000
5,000
£ 4,000
oi 3,000
2,000
1,000
10
15 20
Age (years)
25
30
35
A modified Weibull curve fit to the data produces a coefficient of determination (r )
value of 0.92. The equation is of the following form:
-53-
-------
Miles/Year = C0 * CiA(-((A/C2)AC3))
where: A = age
Co = 872,845
Ci = 202.79
C2 = 2.3397
C3 = 0.0874
The modified Weibull curve-fitted average annual lifetime mileages were then translated
into cumulative odometer levels with age by simply multiplying the age by the curve-
fitted average miles per year. These curve-fit based cumulative odometer levels were
then converted to annual mileage accumulation rates as a function of age by simply
differentiating cumulative mileages at successive ages. The MARCalcs sheet in the
MC_Activity_All.xls file on the accompanying delivery CD contains these translations.
While the modified Weibull curve fits the data well up to about 25 years of vehicle age,
the available data indicate that the average annual lifetime mileage for older vehicles
drops below the curve. Our review of the available advertisements for motorcycles more
than 25 years of age indicates that most of them are collector's items that are
infrequently, if ever, ridden.
In addition, we observed that the actual and Weibull curve-fitted VMT at Age 1 (model
year 2008) was significantly higher than EPA assumed for Age 1 vehicles in MOBILE6
based on an earlier MIC survey. Actual annual VMT at Age 1 estimated from
CycleTrader was over 6,800 miles, compared to Age 1 mileage from MOBILE6 of 4,786.
We believe this higher recent Age 1 mileage, though real, is likely a short-term
phenomenon that occurred in response to the run-up in retail gasoline prices during much
of 2008, with the effect that motorcycles were driven more in households where drivers
could choose between a motorcycle and a less fuel-efficient light-duty vehicle.
We therefore are recommending the annual VMT distribution shown in Table 3-7, which
is based on a -piecewise" fit using the modified Weibull curve through Age 23, and a
power curve-based fit beyond Age 23 that matches the available data for the pre-1985
model year vehicles. To deal with the short-term effect of higher motorcycle VMT
accumulation in 2008 likely resulting from high gasoline prices, we determined Age 1
mileage as the average of that predicted by the Weibull fit at Age 1 (6,296) and the
MOBILE6 Age 1 mileage (4,786), or 5,541.
In conjunction with the scrappage curve described above, the VMT data in Table 3-7
produce a lifetime average total mileage accumulation value of 35,378 miles. For the
first 13 years (at which point approximately half of the vehicles have been retired from
-54-
-------
Table 3-7
Motorcycle Annual VMT vs. Age (years)
Age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Annual VMT
5,541
2,958
2,237
1,852
1,604
1,426
1,292
1,185
1,097
1,024
962
909
862
820
783
750
720
692
667
644
623
603
585
551
518
Age
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Annual VMT
485
452
418
392
366
339
313
287
260
234
208
182
155
129
103
76
50
24
0
0
0
0
0
0
0
service), the annual average VMT is approximately 1,700 miles. (The annual average
VMT drops as age increases and some of the vehicles are periodically out of service.)
As would be expected, the available data indicate that the mileage accumulation rate for
scooters and Class I motorcycles is about one-third of the average for the entire highway
motorcycle fleet. The most recent MIC owner survey indicates an average of 3,195 miles
per year for all motorcycles vs. 1,331 miles/year for scooters and 941 miles/year for all
motorcycles and scooters under 125 cc displacement. (It should be noted that the annual
average VMT from the MIC survey does not include motorcycles -^tot in running
condition" at the time of the survey. The lower annual average VMT from our analysis is
due in part to the fact that it accounts for vehicles being periodically not in running
condition and accumulating 0 miles per year.) Since, as described previously, we are not
-55-
-------
differentiating between the emission levels of Class I, II, and III motorcycles, there is no
need to incorporate class-specific VMT within MOVES.
Non-Uniformity of Activity - Most highway vehicles exhibit a relatively uniform pattern
of activity and are used on almost a daily basis. In contrast, motorcycles are primarily
recreational vehicles and are not practical for use in freezing temperatures. Data
available from the MIC owner surveys and other sources indicate both seasonal and daily
variations in use that are significant.
Table 3-8 presents data from the most recent MIC owners survey for the distribution of
motorcycle use by season. Not surprisingly, owners in the Midwest region of the country
report that only 1% of their annual operation of highway motorcycles occurs in the
winter. This is not an unexpected result of the cold and snowy conditions associated with
Midwestern winters. Perhaps more surprising is that owners in every region of the
country use their motorcycles less in the winter. This is in part due to the fact that wind
chill makes the use of motorcycles below 50°F uncomfortable without the use of
electrically heated liners or heavily insulated riding suits. Significant highway
motorcycle activity occurs only when temperatures are above 50°F. The daily average
VMT in the summer is approximately twice what would be expected from the annual
average VMT. In the Spring and Summer months, daily average VMT is approximately
the annual average divided by 365.
Table 3-8
Portion of Annual Operation by Season
Season
Spring
Summer
Fall
Winter
East
25%
53%
19%
3%
Midwest
22%
58%
19%
1%
South
27%
41%
22%
10%
West
26%
44%
22%
8%
U.S.
25%
48%
21%
6%
Available data make it clear that highway motorcycle operation is more frequent on
weekends than on weekdays. According to a study by the U.S. Department of
Transportation,32 -ftjhere were 1.5 times as many two-vehicle motorcycle crashes
involving passenger vehicles in 2005 during weekends than during weekdays."
Assuming accidents are proportional to VMT, 60% of MC activity is on the two weekend
days. Since weekend days are only 28.6% of all days, this indicates that VMT on the
average weekend day is over twice as high as on the average week day.
* Electrically heated riding gear is available to extend the range of comfortable riding temperature to
approximately freezing; however, very few motorcyclists own heated gear and many motorcycles lack the
electrical system capacity necessary to operate such gear.
-56-
-------
Data from the most recent MIC owner survey are consistent with the DOT accident data,
indicating 56% weekend use.
The disproportionate use of motorcycles on weekends is consistent with owner responses
to questions in the MIC owner survey regarding the -type of riding" done. -Casual
Pleasure Riding" was reported to be the primary type of riding in all regions of the
country.
Driving Patterns - In the absence of motorcycle-specific data, it is not unreasonable to
assume that highway motorcycle driving patterns in customer service are similar to those
for light-duty vehicles. Although motorcycles typically have higher performance than
passenger cars and light trucks, previous surveys conducted by Sierra demonstrate that
high-performance passenger cars have a similar driving pattern to other vehicles because
they are constrained by traffic. Although lane sharing/splitting is allowed in California,
motorcycles are similarly constrained by the flow of traffic in most areas.
Unlike the wealth of data that exists for light-duty vehicles, Sierra was not able to obtain
detailed, second-by-second data for highway motorcycle operation in customer service.
Efforts to obtain the data used to develop the World Motorcycle Test Cycle (WMTC)
were unsuccessful. A representative of Harley-Davidson, the company that originally
collected the U.S. data, reported that the data are no longer available at Harley.
Discussions with representatives of Harley-Davidson led us to the conclusion that the
database from which the WMTC was developed was not demonstrated to be truly
representative of highway motorcycle operation in the U.S. The selection of the road
routes was somewhat arbitrary and only professional riders were used. It is not clear how
these riders' patterns of operation compare to other riders.
Notwithstanding the concerns about the data used to develop the WMTC, Figures 3-17
and 3-18 compare the Speed-Acceleration Frequency Distributions (SAFDs) and relevant
driving statistics for the WMTC to Sierra's best estimate of passenger car operation in
customer service based on chase-car data collected in California subsequent to the
elimination of the national 55 mph speed limit.
In Figure 3-17, the joint frequency distributions of vehicle speed (horizontal axis) and
acceleration (vertical axis) for the WMTC are in the SAFD table for the WMTC. Below
the SAFD table, a series of key statistics for the WMTC are shown, including MOVES
VSP bin distributions.
Figure 3-18 presents similar distributions and key driving statistics for a driving cycle
referred to as the -GalWtd" cycle. It was developed from light-duty vehicle driving data
collected in California under a series of chase car-based driving studies33'34 conducted in
2000 that were jointly sponsored by the California Department of Transportation
(Caltrans) and the California Air Resources Board (CARB) encompassing four study
areas:
-57-
-------
Figure 3-17
World Motorcycle Test Driving Cycle SAFD (%)
World Motorcycle Test Driving Cycle SAFD (%)
ACCEL
BIN
(mph/s)
-20
-19
-18
-17
-16
-15
-14
-13
-12
-11
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Totals
SPEED BIN
(mphl
0
8.889
0.444
0.111
9.444
5
0.056
0.444
0.722
0.778
0.056
0.500
0.333
0.611
0.167
0.222
3.889
10
0.333
0.389
0.333
0.167
0.556
0.611
0.333
0.167
0.056
0.167
0.111
3.222
15
0.278
0.333
0.444
1.222
3.667
1.500
0.333
0.278
0.222
0.056
8.333
20
0.333
0.222
0.667
1.556
4.500
1.222
0.444
0.333
0.222
0.056
9.556
25
0.222
0.222
0.556
0.889
2.278
1.056
0.500
0.444
0.056
0.056
6.278
30
0.222
0.389
0.333
0.778
3.000
0.833
0.889
0.056
0.056
6.556
35
0.056
0.444
0.778
1.444
3.833
2.167
0.500
0.056
9.278
40
0.222
0.444
1.000
2.500
1.556
0.278
0.056
6.056
45
0.278
0.444
0.556
2.944
1.056
0.222
5.500
50
0.056
0.389
0.667
3.222
1.000
0.278
5.611
55
0.056
0.278
0.222
2.444
0.722
0.111
3.833
60
0.222
0.278
2.556
0.556
3.611
65
0.222
0.333
0.667
0.722
1.944
70
0.333
0.389
7.278
0.444
8.444
75
0.167
0.556
6.778
0.278
7.778
80
0.667
0.667
85
90
95
100
TOTALS
1.500
3.056
6.333
10.833
55.833
14.667
4.333
1.944
0.556
0.667
0.278
100.000
Summary Statistics
Avg Speed (mph)
Min Speed (mph)
Max Speed (mph)
Avg Non-Idle Speed (mph)
Avg Cruise Speed (mph)
Avg Acceleration (mph/sec)
Max Acceleration (mph/sec)
Avg Deceleration (mph/sec)
Max Deceleration (mph/sec)
Bin 1 1 % (0-25 mph, <0 kw/tonne)
Bin 12 % (0-25 mph, 0-3 kw/tonne)
Bin 13 % (0-25 mph, 3-6 kw/tonne)
Bin 14 % (0-25 mph, 6-9 kw/tonne)
Bin 15 % (0-25 mph, 9-12 kw/tonne)
mp ,>-
35.93
0.00
77.90
39.68
41.95
0.86
6.03
-1.10
-4.47
5883.523
8562.099
3778.853
861.000
861.000
956 657
Time at Idle (%)
Time at Cruise (%)
Time in Accel (%)
Time in Decel (%)
Avg Trip Length (miles)
Avg Trip Time (min)
# Stops per Mile
# of 1-Sec Observations
# of Trips
EPA VSP (Vehicle Specific
EPA Avg Veh Spec Pwr, VSP (kW/tonne)
Bin 0 % (braking)
Bin 1 % (idle)
Bin 21 % (25-50 mph, <0 kw/tonne)
Bin 22 % (25-50 mph, 0-3 kw/tonne)
Bin 23 % (25-50 mph, 3-6 kw/tonne)
Bin 24 % (25-50 mph, 6-9 kw/tonne)
Bin 25 % (25-50 mph, 9-12 kw/tonne)
Bin 27 % (25-50 mph, 12-18 kw/tonne)
Bin 28 % (25-50 mph, 18-24 kw/tonne)
Bin 29 % (25-50 mph, 24-30 kw/tonne)
Bin 30 % (25-50 mph, >=30 kw/tonne)
9.444
68.444
9.667
12.444
17.967
30.000
0.612
1,800
1
Power)
4633.103
6122.715
8657.902
5166.000
3970.147
5835.704
3204.803
3396.136
0 000
3300.482
861.000
191.335
143.500
Avg Specific PKE (hp-sec/lb)
Avg Total Specific Power (hp/lb)
Avg Non-Zero Specific Power (hp/lb)
Max Specific Power (hp/lb)
Spec Pwr Freq (%): 0 hp/lb
Spec Pwr Freq (%): >0-0.1 hp/lb
Spec Pwr Freq (%): >0. 1-0.2 hp/lb
Spec Pwr Freq (%): >0.2-0.3 hp/lb
Spec Pwr Freq (%): >0.3 hp/lb
Statistics
Bin 33 % (>50 mph, <6 kw/tonne)
Bin 35 % (>50 mph, 6-12 kw/tonne)
Bin 37 % (>50 mph, 12-18 kw/tonne)
Bin 38 % (>50 mph, 18-24 kw/tonne)
Bin 39 % (>50 mph, 24-30 kw/tonne)
Bin 40 % (>50 mph, >=30 kw/tonne)
4.68
0.0467
0.0977
0.6480
52.22
31.44
8.72
5.28
2.33
3109.148
4687.619
0 000
8131.599
7079.294
1195.846
143.500
-58-
-------
Figure 3-18
California Urban & Rural Weighted (CalWtd)
Driving Cycle SAFD (%)
California Urban & Rural Weighted (CalWtd) Driving Cycle SAFD (%)
ACCEL
BIN
(mph/s)
-20
-19
-18
-17
-16
-15
-14
-13
-12
-11
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Totals
SPEED BIN
(mph)
0
9.180
0.245
0.100
0.100
9.625
5
0.015
0.100
0.398
0.873
1.324
1.616
1.172
0.758
0.559
0.413
0.299
0.100
7.627
10
0.100
0.100
0.214
0.329
0.429
0.444
0.115
0.329
0.299
0.758
0.758
0.314
0.214
4.404
15
0.214
0.146
0.513
0.314
0.429
0.398
0.612
0.697
0.812
0.751
0.214
0.115
5.215
20
0.100
0.115
0.146
0.429
0.597
0.597
0.398
1.455
1.011
0.842
0.590
0.230
6.509
25
0.015
0.146
0.329
0.314
0.498
0.697
2.167
1.324
1.103
0.376
0.115
7.083
30
0.031
0.031
0.345
0.214
0.513
0.796
2.711
1.470
0.804
0.230
0.214
7.360
35
0.015
0.115
0.031
0.115
0.245
0.498
1.141
2.971
1.731
0.720
0.261
7.843
40
0.031
0.100
0.115
0.046
0.827
0.842
2.995
1.379
0.422
0.214
6.971
45
0.031
0.314
0.130
0.161
0.698
2.636
1.005
0.276
0.115
5.366
50
0.015
0.015
0.115
0.261
0.790
2.731
0.814
0.376
5.118
55
0.015
0.130
0.622
3.187
0.837
0.146
4.937
60
0.130
0.936
4.169
0.936
0.130
6.302
65
0.100
0.331
4.633
0.598
0.130
5.792
70
0.015
0.744
4.381
0.897
6.037
75
0.100
0.100
0.254
2.288
0.223
0.015
2.979
80
0.108
0.493
0.062
0.663
85
0.015
0.154
0.170
90
95
100
TOTALS
0.100
0.214
0.621
0.874
2.604
2.819
5.575
10.210
48.709
14.699
7.393
3.854
1.501
0.727
0.100
100.000
Summary Statistics
wg Speed (mph)
/lin Speed (mph)
/lax Speed (mph)
wg Non-Idle Speed (mph)
wg Cruise Speed (mph)
wg Acceleration (mph/sec)
/lax Acceleration (mph/sec)
wg Deceleration (mph/sec)
/lax Deceleration (mph/sec)
lin 11 %(0-25 mph, <0 kw/tonne)
lin 12 % (0-25 mph, 0-3 kw/tonne)
lin 13 %(0-25 mph, 3-6 kw/tonne)
lin 14 % (0-25 mph, 6-9 kw/tonne)
lin 15 %(0-25 mph, 9-12 kw/tonne)
lin 16% (0-25 mph >=1 2 kw/tonne)
34.73
0.00
83.20
38.14
42.06
1.22
5.81
-1.50
-9.05
4475.686
4279.244
3969.049
2406.305
1556.606
1273 095
Time at Idle (%)
Time at Cruise (%)
Time in Accel (%)
Time in Decel (%)
Avg Trip Length (miles)
Avg Trip Time (min)
# Stops per Mile
# of 1-Sec Observations
# of Trips
EPA VSP (Vehicle Specific
EPA Avg Veh Spec Pwr, VSP (kW/tonne)
Bin 0 % (braking)
Bin 1 % (idle)
Bin 21 % (25-50 mph, <0 kw/tonne)
Bin 22 % (25-50 mph, 0-3 kw/tonne)
Bin 23 % (25-50 mph, 3-6 kw/tonne)
Bin 24 % (25-50 mph, 6-9 kw/tonne)
Bin 25 % (25-50 mph, 9-12 kw/tonne)
Bin 27 % (25-50 mph, 12-18 kw/tonne)
Bin 28 % (25-50 mph, 18-24 kw/tonne)
Bin 29 % (25-50 mph, 24-30 kw/tonne)
Bin 30 % (25-50 mph, >=30 kw/tonne)
9.625
59.088
16.092
15.195
8.888
14.789
0.721
2,003
2
Power)
4209.310
8072.835
8762.598
5249.241
4570.025
3521.147
4886.247
3852.850
0 000
3115.659
964.072
494.750
112.219
Avg Specific PKE (hp-sec/lb)
Avg Total Specific Power (hp/lb)
Avg Non-Zero Specific Power (hp/lb)
Max Specific Power (hp/lb)
Spec Pwr Freq (%): 0 hp/lb
Spec Pwr Freq (%): >0-0.1 hp/lb
Spec Pwr Freq (%): >0.1-0.2 hp/lb
Spec Pwr Freq (%): >0.2-0.3 hp/lb
Spec Pwr Freq (%): >0.3 hp/lb
Statistics
Bin 33 % (>50 mph, <6 kw/tonne)
Bin 35 % (>50 mph, 6-12 kw/tonne)
Bin 37 % (>50 mph, 12-18 kw/tonne)
Bin 38 % (>50 mph, 18-24 kw/tonne)
Bin 39 % (>50 mph, 24-30 kw/tonne)
Bin 40 % (>50 mph, >=30 kw/tonne)
6.29
0.0586
0.1319
0.6104
55.54
20.44
15.02
5.87
3.14
5471.770
5388.292
0 000
6726.784
4665.725
1704.819
581.568
-59-
-------
1. Sacramento metropolitan area;
2. San Francisco Bay Area;
3. Stanislaus County (including the Modesto metropolitan area); and
4. South Coast (i.e., Los Angeles) combined metropolitan area.
These datasets contained driving in both urban and rural areas. In the post-processing
that was performed under each of these studies, the type of roadway the vehicle was
traveling on during each second was also recorded in the output dataset and categorized
using the HPMS (Highway Performance Monitoring System) Functional Class scheme.
These driving data were binned into urban and rural Functional Class groups and re-
weighted using California VMT weightings by functional class obtained from HPMS.
Driving cycles for urban and rural light-duty operation in California were then developed
from the data in these re-weighted groups. The second-by-second speed traces for these
individual urban and rural cycles, called Cal-Urban and Cal-Rural, were then weighted
together using a VMT-based split of 81.8% urban vs. 18.8% rural operation for California
from HPMS. Figure 3-18 presents the SAFD and key statistics for this combined
urban/rural driving cycle called -GalWtd."
The tabulations for WMTC and CalWtd cycles contained in Figures 3-17 and 3-18 show
that they are fairly similar. Average speeds are 35.9 mph and 34.7 mph, respectively.
Top speeds are 77.9 mph for the WMTC and 83.2 for the CalWtd. Acceleration and
specific power statistics are also generally comparable between the two cycles. Average
VSP (based on EPA's VSP definition) is 5.38 kW/tonne for the WMTC and 4.89
kW/tonne for the CalWtd cycle. And the VSP distributions for both cycles are in
relatively close agreement.
Despite the aforementioned concerns regarding the origins and methods used to develop
the WMTC, the statistical comparisons reflected in Figures 3-17 and 3-18 indicate that
the WMTC appears to closely resemble overall driving patterns of light-duty vehicles.
Thus, until additional motorcycle-specific driving data are collected, we believe
motorcycle driving patterns can be reasonably represented with those of light-duty
vehicles.
3.6 Incorporation into MOVES Tables
The final element in the development of motorcycle emission and activity rates for
MOVES consisted of translating the FTP-based exhaust and evaporative emission rates
and activity data presented in the preceding sub-sections into the specific data tables used
within the MySQL database underlying the MOVES model. These steps are described
below, first for the emission rates, then the activity data.
http://www.fhwa.dot.gov/policy/ohim/hs02/vm2.htm
-60-
-------
Motorcycle Emission Rate Tables - For the emission rate table, two key elements were
performed to incorporate the basic exhaust and evaporative emission rates presented
earlier in Table 3-5 into the specific tables and their data structures required by MOVES:
1. Translation of FTP-based exhaust rates into emission rates by operating mode bin;
and
2. Conversion of ^node"-based evaporative emission rates into new process-based
evaporative emission categories.
Translation of Running Exhaust Rates to Operating Mode Bins - A spreadsheet called
MOVES_MC_EmissionRates_FT01.xls was developed to perform these detailed
translations. (This spreadsheet was also provided to EPA as a deliverable on CD.)
First, FTP-based running exhaust emission rates by model year range for HC, CO and
NOx were translated into emission rates by individual VSP/speed bin based on an
analysis of modal, second-by-second emission tests collected under an in-use motorcycle
emissions surveillance study35 sponsored by CARB. Under that study, over 100
motorcycles were randomly procured and FTP and Unified Cycle (UC) bag and
composite emissions were measured. Second-by-second FTP and UC cycle emission
measurements were collected for a 15 vehicle subset, although modal data for one UC
test were missing.
This modal emissions sample of 15 motorcycles ranged from model years 1966 through
1996. Only one of these motorcycles was catalyst-equipped. Thus, given the size of the
sample and available resources, no effort was made to divide the test data by age or
model year range to reflect their effects on emission distributions by VSP bin. In the
supporting analysis spreadsheets, separate relationships for non-catalyst vs. catalyst
groups were developed and examined, but the resulting catalyst group bin allocations
contained anomalies caused by the fact that the data were based on a single motorcycle
(run over both FTP and UC cycles). Thus, given the sample of available modal data for
motorcycles, a single set of emission rates by operating mode bin was developed simply
from averaging of the entire 15-vehicle sample.
Sierra performed fairly extensive data validation and cleaning on these modal emission
test results prior to their use under this analysis. Second-by-second dilute concentration
measurements were converted to a mass basis and properly time-aligned with the
measured driving traces. This time alignment was carefully performed and applied on a
test-by-test basis using comparisons of the second-by-second speed traces and dilute
measurements.
The time-aligned second-by-second mass exhaust emissions were then binned into
MOVES-based operating mode/VSP bins by calculating second-by-second VSP using the
speed traces, dynamometer coast-down data and vehicle weights recorded for each test
from the following equation:
-61-
-------
vsp ^
m
where
VSPt is vehicle specific power (kW/tonne) at time t;
vt is vehicle speed (m/sec) at time t;
at is vehicle acceleration (m/sec ) at time t;
m is vehicle mass (tonne); and
A, B and C are track road load coefficients, representing rolling resistance,
rotational resistance and aerodynamic drag, in units of kW-sec/m, kW-sec2/m2
and kW-sec3/m3, respectively.
(In the original modal motorcycle emissions testing study, track road load horsepower at
50 mph and mass was measured and recorded for each vehicle . The A, B, and C values
were then calculated from these measurements using 0.35, 0.10 and 0.55 weightings,
respectively.)
Note that the second-by-second VSP and mass emissions were binned only for the
warmed-up portions (Bags 2 and 3) of the FTP and UC cycles. Modal measurements
during Bag 1 were not used for this element of the analysis, which focused on
development of stabilized/running exhaust emission allocations by VSP bin.
Binned VSP and exhaust emission measurements over the stabilized portions of the FTP
and UC tests were assembled in a separate spreadsheet called MC_OpModeBinned.xls
In the OpBinEmis sheet within this spreadsheet, a series of curve fits of HC, CO and NOx
emissions vs. VSP were developed to address — hbe-filling" issues in bins where there
were little or no data. From this 15-vehicle sample of modal FTP and UC cycle
measurements, most of the operating mode bins contained at least 50 single-second
observations. However, Bins 30, 39, and 40 — which reflect high VSP ranges at the edge
of, or beyond that, in the FTP or UC cycles — contained fewer observations. Upon
examination, the 23 and 1 1 data points in Bins 30 and 39, respectively, were not found to
substantially affect the resulting curve fits, so data in these bins were used. However, the
few data points in Bin 40 were not included in the curve fitting process.
These curve fits were developed using the following exponential equation:
where
EJ represents predicted emissions (g/hr) of pollutant / (either HC, CO or NOx) for
operating mode biny;
VSP, is the measured mean VSP (kW/tonne) within each biny; and
Co and Cj are least-squares derived coefficients determined separate for each
pollutant and operating mode bin range (i.e., Bins 1 1-16, 21-30 and 3 1-40 using
EPA's numbering scheme as applied in MOVES).
-62-
-------
Within the MC_OpModeBinned.xls analysis spreadsheet, the Co and Ci coefficients were
determined using Excel's built-in —Soler" utility. As noted above, separate curve fits
were generated for each pollutant and set of operating mode bins within a speed range.
Emissions for Bins 0 (deceleration/braking) and 1 (idle) were not based on curve fits but
simple means of measured values within each of those bins. As shown in detail in
highlighted cells inRows 91-207 and Columns N, U, and AC of the OpBinEmis sheet in
MC_OpModeBinned.xls, these exponential curve fits generally produced correlations
(expressed as R2) to the actual binned measurements in excess of 0.9 for each of the bin
groups and pollutants. However, correlations for the high speed (> 50 mph) group
covering Bins 31-40 were generally lower than those for the other speed groups—for this
speed range, the correlation for NOx was only R2=0.30.
The resulting curve-fitted emission rates by operating mode bin (in grams/hr) reflect the
emission levels of the 15-vehicle test sample upon which they were based, not the
broader fleet of all on-highway motorcycles. Adjustments were made to translate the
emission rates by bin from the test sample to hot-FTP —laocation factors" that could be
applied to the FTP-based fleet exhaust emission factors presented earlier in Table 3-5.
The first step involved calculating composite emissions (in grams) from hot FTP
weighting of the curve-fitted emission rates (in g/hr) each bin. In the OpBinEmis sheet,
these composite emission calculations are performed in Column AD and Rows 218-236
where the bin weightings are simply the number of seconds of observation in each bin
over an FTP test.
The second and final step consisted of normalizing the curve-fitted bin emission rates
from the measured study sample by their composite hot FTP emissions calculated above
to yield operating mode bin emission rates in g/hr per gram of hot FTP emissions. This
step is reflected in the table within the OpBinEmis sheet in Rows 259-271 and Columns E
through AC.
These normalized bin rates could then be applied directly to the FTP-based fleet emission
factors to generate running (hot) exhaust emission rates by operating mode bin for
incorporation into MOVES.
The modal motorcycle emission measurement study only measured HC, CO, NOx and
CO2 exhaust emissions. PM exhaust measurements were not collected. Thus, the
distribution of PM running exhaust by operating mode bin for motorcycles was assumed
to be identical to that of gasoline-fueled passenger cars. Passenger car PM running
exhaust emission rates were -harvested" from the EmissionRatebyAge table in the Draft
MOVES2009 database and loaded into the PMRatesOC and PMRatesEC sheets in the
MC_OpModeBinnecLxls file, for the organic carbon (OC) and elemental carbon (EC)
components of exhaust PM, respectively. (Rates for both the OC and EC components
were loaded from MOVES to confirm that the emission distributions by bin were the
same for each, and they were.) A series of Excel pivot tables were generated in the
PMPvts sheet to tabulate and aggregate these MOVES-based PM OC and EC rates across
age and model year. These rates were then loaded into the OpBinEmis sheet and
-63-
-------
normalized to a gram/hour per gram of hot-FTP basis across the operating mode bins as
was done for the HC, CO, and NOx data from the modal study.
The complete set of normalized running exhaust rates by operating mode bin for HC, CO,
NOx, and PM (the latter separately as PM-OC and PM-EC) are presented below in
Table 3-9.
Table 3-9
Normalized Motorcycle Running Exhaust Emission Rates (g/hr per gram Hot FTP) by
Operating Mode Bin
Poll
HC
CO
NOx
PM-OC
PM-EC
Poll
ID
1
2
3
111
112
Running Exhaust Operating Mode Bin
0
1 7
1 7
1 3
1.5
1.5
1
1 5
1 ?
06
1 8
1 8
11
94
1 9
08
90
90
12
98
99
1 1
94
94
13
3 1
94
1 5
9S
9S
14
34
96
1 9
36
36
15
38
9Q
95
54
54
16
45
33
40
177
177
21
97
9Q
33
99
99
22
3 1
34
38
1 8
1 8
23
34
37
49
? 1
? 1
34
37
40
46
9Q
9Q
25
41
44
50
58
58
27
46
50
57
85
85
28
5 5
60
69
44.4
44.4
29
66
71
83
128
128
30
79
87
10 1
365
365
33
38
76
137
4.6
4.6
35
43
85
144
4.1
4.1
37
47
91
148
5.7
5.7
38
5 1
98
153
9.7
9.7
39
56
106
158
24.7
24.7
40
6.5
12.0
16.7
27.1
27.1
The normalized rates were then loaded from the MC_OpModeBinned.xh file into the
-^nain" spreadsheet used for translation to specific MOVES data tables and record
structures, MOVES_MC_EmissionRates_FT01.xls, into a sheet called BinEmis.
Within the main MOVES_MC_EmissionRates_FT01.xls spreadsheet, motorcycle
running exhaust emission rates by pollutant, model year, age, and operating mode were
then calculated and stored within the MOVES_MCEmissionRateByAge sheet, which is
similar in field structure to the EmissionRateByAge data table in the MOVES database.
(This sheet also housed the calculations of motorcycle emission rates for the other
evaporative and crankcase emission processes required by MOVES explained later in this
section.) The FTP-based fleet emission rates by model year range presented earlier in
Table 3-5 were loaded into a sheet called FTPExh. Within this sheet, FTP emission
factors were translated into —Ht FTP/LA4" and Cold Start Increment" components using
FTP bag measurements from the 15-vehicle modal motorcycle study sample discussed
above and contained in the neighboring ColdHotFactors sheet in the
MOVES MCEmissionRatesFT01.xls workbook.
Beginning in Row 22 of the FtpExh sheet, the FTP-based emission factors, now separated
into Running Exhaust and start Increment component, were asssigned to the specific set
of model years or model year ranges mandated by EPA in MOVES for light-duty
vehicles (and motorcycles). For the MOVES ModelYearIDGroup=19601980 (1980 and
older models), this involved the development of separate allocation factors for the pre-
1978 and 1978-1985 model year groups for which motorcycle FTP rates were developed
and shown earlier in Table 3-5. These allocation factors are contained in a sheet called
MYGAlloc within the workbook. The splits for the 1980 and older range were based on
default motorcycle data in MOBILE6.
-64-
-------
In Column V (labeled RevRun) of the MOVES_MCEmissionRateByAge sheet, the FTP-
based running exhaust emission factors by pollutant and MOVES model year group (in
g/mi) from the FTPExh sheet were then combined with the normalized running exhaust
operating mode emission rates discussed earlier using a series of table lookups. The
calculation was performed as follows:
RunExhm,p,y (g/hr) = HotFTPp,y (g/mi) x 7.5 mi x OpModeERm,p (g/hr per g FTP)
where HotFTP is the Hot FTP (Bag 3 + Bag 2) emission rate; OPModeER is the
normalized distribution of FTP-based running exhaust emissions by operating mode; and
indices m = operating mode,/? = pollutant, and_y = model year group.
As explained earlier in Section 3.3, highway motorcycle in-use emissions deterioration is
believed to largely result from tampering that occurs in the first few years of ownership.
All of the assumed deterioration was modeled to occur in the first MOVES Age Group
(0-3 years). Thus, Age Group was not incorporated in the above equation since
motorcycle emissions were not assumed to vary by age after accounting for deterioration
within this first 0-3 year old Age Group.
Translation of Starting Exhaust Rates to Operating Mode Bins - Translation of FTP -
based starting exhaust rates into the MOVES-required structures was much simpler than
running exhaust. MOVES uses eight discrete operating mode bins to represent the
starting emissions reflecting different soak time intervals ranging from <6 minutes
(OpModeID=101) to > 12 hours (OpModeID=108). Unlike with running exhaust, there
were no motorcycle-specific emission test data available from which to develop start
increment vs. soak time relationships by MOVES operating mode bin.
In the absence of data, the distribution of starting exhaust emissions by MOVES
operating mode bin for motorcycles was developed using starting exhaust rates extracted
from MOVES for gasoline-fueled passenger cars. These extracted MOVES starting
exhaust rates are contained in the PCStartRates sheet. A series of Excel Pivot tables
were constructed from these passenger car starting exhaust rates in the StartPvts sheet,
tabulating by model year group and operating mode bin for each pollutant. Separate
tabulations were performed for two age groups (0-3 years) and (20+ years) at each end of
the Age Group categories employed within MOVES to determine whether EPA has
modeled age dependence into the distributions of starting exhaust rates, not their absolute
levels, which are age dependent. It was confirmed that the normalized distributions
(normalized to the >12 hour soak bin) are neither age nor model year dependent.
These normalized starting exhaust emission distributions for each pollutant were then
loaded into Rows 38-42 of the ColdHotFactors sheet.
MOVES starting exhaust emission rates (in g/start) were then calculated in Column W
(labeled RevStart) of the MOVES_MCEmissionRateByAge sheet using lookup formulas
-65-
-------
for these normalized distributions and the FTP Cold Start Increment rates by model year
group (reflecting a >12 hour soak) contained in the FTPExh sheet as follows:
StartExhm^y (g/start) = ColdFTPIncrp>y (g/start) x OpModeESm,p
Where ColdFTPIncr is the FTP cold start increment (Bag 1 - Bag 3); OPModeES is the
normalized starting emission rate for a specific soak bin; and indices m = starting
operating mode bin (soak bin),/? = pollutant, and_y = model year group.
Translation of Crankcase Rates - In the MOVES database, crankcase emissions are
represented in a separate data table called CrankcaseEmissionRatio. Motorcycle
crankcase emissions were thus entered into a separate sheet in the main workbook called
MOVES'_MC Crankcase that resembled the field structure of the corresponding MOVES
table. MOVES requires that the emission rates in this table be stored as ratios of
crankcase-to-running exhaust emissions.
As discussed earlier in Sections 3.2 and 3.3, crankcase emissions on highway
motorcycles were controlled starting in model year 1978 and no data were available to
support an assumption of non-zero crankcase emissions from in-use deterioration for
1978 and later motorcycles. Thus, crankcase ratios for all 1978 and later models were set
to zero in the MOVES_MCCrankcase sheet.
As shown earlier in Table 3-5, crankcase emission rates for pre-1978 motorcycles were
estimated to be 0.4 g/mi, reflecting an assumed 45%/55% split between two- and four-
stroke engines for those models. (Two-stroke engines use a pressurized crankcase and do
not have a crankcase vent; thus they have no crankcase emissions.) For pre-1978 models,
the crankcase ratio for motorcycles was then simply calculated as:
CC Ratio = 0.4 g/mi + 10.6 g/mi (running exhaust FTP) = 0.0363
and stored in the MOVES_MCCrankcase sheet for HC and for these model years.
Evaporative Emission Rate Conversions - As described earlier, the evaporative emission
data compiled under this study consisted of one-hour SHED (hot soak and diurnal)
measurements from roughly 20 highway motorcycles, plus nesting loss" (i.e., non-heat
build portion) and running loss estimates developed from real-time 24-hour evaporative
tests of two off-highway motorcycles. These older tests had to be translated to represent
the three evaporative processes defined in MOVES:
1. Permeation - defined as the migration of hydrocarbons through elastomers in a
vehicle's fuel system;
2. Tank Vapor Venting (TW) - expulsion of fuel vapor generated from fuel
evaporative within the fuel system; and
-66-
-------
3. Liquid Leaks (LLs) - fuel in liquid form leaking from the fuel tank or fuel system
which ultimately evaporates into the atmosphere.
In addition, the ambient conditions and diurnal temperature changes reflected in the
California evaporative test data had to be adjusted to reflect the reference conditions EPA
established for these processes in MOVES: base permeation at 72°F and vapor venting
based on a 72-96°F diurnal temperature change.
Two different database tables are used in MOVES to store evaporative emission rate
data:
1. EmissionRateByAge - contains permeation, liquid leaker and -hot soak" TVV
rates; and
2. CumTWCoeff - contains -eold soak" TVV coefficients used during MOVES
execution to dynamically generated cold soak TVV emissions from hourly
ambient temperature and tank fuel temperature changes.
In the EmissionRateByAge table, MOVES uses a combination of the OpModelD and
PolProcID fields to -map" and store process-specific evaporative rates as shown below in
Table 3-10.
Table 3-10
MOVES Evaporative Processes by Pollutant/Process and OpMode
PolProcID
1 1 1 - Permeation
112 -TVV
113-LL
OpModelD
150 -Hot Soak
-
Hot Soak TVV
-
151 -Cold Soak
-
-
Cold Soak LLs
300 - All Running
Base Permeation
Running Loss
-
The first step in translating the one-hour SHED test data and running loss estimates for
motorcycles into this structure consisted of assembling basic rates for the old evaporative
process definitions (hot soak, diurnal, resting and running losses) by model year range.
For federally certified motorcycles, three model year ranges were employed: (1) pre-
2004; (2) 2004-2007; and (3) 2008 and later. These rates, based on motorcycle
evaporative estimates described in detail earlier in Sections 3.2 and 3.3, were loaded into
the MOVES_MC_EmissionRates_FT01.xls workbook at the top of the FTPEvap sheet.
MOVES base permeation rates (in g/hr) were calculated from the resting loss (non-heat
build) rates (in g/day) by dividing by 24 and adjusting these rates to reflect the difference
between the ambient temperature during which these resting loss emissions were
measured, best estimated to be 70.5°F, and the 72°F reference temperature used by EPA
from the following equation:
-67-
-------
p _
p
_ _ _
base 0.0385 (Tadj-Tbase)
Base permeation rates were calculated to be 0.117 g/hr for pre-2008 models and
0.012 g/hr for 2008 and later models. As shown in Table 3-10, these base permeation
emission rates were coded into the records in the EmissionRateByAge sheet for
combinations of OpModeID=300 and PolProcID=l 11.
Running loss rates were converted from grams per mile to grams per hour. A value of
17.81 g/hr was applied to represent running loss emissions for federally certified
motorcycles across all model year. Federal evaporative emissions standards for 2008 and
later model years apply only to permeation; thus, running loss rates for all model years
were estimated to be the same. In the EmissionRateByAge sheet, these running loss rates
were applied for OpModeID=300 and PolProcID=l 12 records as noted in Table 3-10.
As described in EPA's MOVES Evaporative Emission Calculation Methodology,36 the
remaining" evaporative emissions from the hot-soak and diurnal portions of SHED test
measurements are accounted for under the Tank Vapor Venting process in MOVES.
First, hot soak TVV emission rates were calculated by subtracting out base effects from
the hot soak SHED measurements using a five-step procedure detailed in Section 3.3.2 of
the MOVES Evaporative Emission Calculation Methodology that accounts for tank
temperature rise from the LA-4 cycle run prior to hot soak SHED measurement as well as
temperature adjustments to the base permeation rates (at the 72°F reference temperature)
to reflect the average temperature during the hot soak test. These calculations were
performed in the TankTempCalc sheet and Rows 153 through 186 of the FTPEvap sheet
in the MOVES_MC_EmissionRates_FT01.xls workbook. The resulting hot soak TVV
rates by model year range were then loaded into the EmissionRateByAge sheet for records
with OpModeID=150 and PolProcID=l 12 as indicated in Table 3-10.
Second, diurnal SHED measurements were translated into cold soak TVV emission rates
following these steps contained in Section 3.3.1 of the MOVES Evaporative Emission
Methodology. For cold soak TW emissions, MOVES first finds the amount of tank
vapor generated (TVG) as a function of fuel tank temperature and RVP during the
temperature rise portion of a 24-hour diurnal temperature profile based on the Reddy
Equation.37
To perform these calculations in a manner consistent with the MOVES approach, the
1-hour diurnal SHED measurements had to be translated to a real-time, 24-hour basis.
These translations were based on regression equations developed from 1-hour and
24-hour testing performed by Automotive Testing Laboratories (ATL) for CARB in 1994
and cited in Section 5.3 of CARB's EMFAC emission factor model technical support
-68-
-------
documentation. A linear regression of 1-hour to 24-hour diurnal emissions developed
from 19 vehicles over a diurnal range of 60-84 °F were used to convert the 1-hour diurnal
measurements to an equivalent 24-hour basis as follows:
24-Hour Diurnal = 1-Hour Diurnal x 1.303 + 4.902 (R2 = 0.46)
(The ATL testing also included measurements over a 65-105 °F range. However, only
the 60-84 °F data were used since they more closely resembled the 68-86 °F temperature
cycle used in the actual 1-hour diurnal measurements from the CARB Motorcycle
Surveillance Testing study.19)
Once these 1-hour to 24-hour diurnal adjustments were made, hourly permeation rates
were calculated over the 68-86°F temperature cycle upon which the diurnal
measurements were based. These permeation rates were then summed over the entire
24-hour period and subtracted from the adjusted 24-hour diurnal emissions to calculate
Total 24-Hour TVV emission rates by model year range.
Within Rows 33-149 of the FTPEvap sheet, hourly Tank Vapor Generation rates (in
g/gal) were then calculated over this temperature cycle as a function of the hourly
temperature and the fuel RVP based on the Reddy Equation. Once the hourly TVG rates
were known and summed over the entire 24-hour period, cold-soak TVV rates were then
calculated by apportioning Total TVV based on the hourly TVG rates.
In EPA's MOVES Evapororative Emissions Methodology documentation,36 TVV rates
for light-duty vehicles were determined based on polynomial curve fits of hourly
measurements over a 24-hour diurnal test as a function of TVG using the following form:
Total TW = Ao + Aj TVG + A2 TVG 2
For motorcycles, these curve fits were not necessary since real-time diurnal
measurements were not available. Thus, the second-order A2 coefficients described in
this documentation were not applicable and were set to zero for the motorcycle cold-soak
TVV rates.
In the FTPEvap sheet, the non-zero Aj coefficient was then calculated simply as:
Aj = Total TW ^Total TVG
In addition, since MOVES is designed to perform the TVV emission rate calculation
dynamically during execution based on a user-selected geographic modeling area and
associated diurnal temperature profiles, these cold-soak TVV rates themselves were not
stored in the MOVES database. Instead, the calculated coefficients' are stored within a
http://www.arb. ca.gov/msei/onroad/doctable_test.htm
t In the MOVES Evaporative Emission Calculation documentation these coefficients are labeled Aj and A2.
Within the MOVES database, these TW coefficients are referred to as —B and -G." (The intercept
coefficient A0 or A is set to zero, based on EPA's constraint that the TW curves pass through the origin.)
-69-
-------
separate table in the database called CumTVVCoeff. A separate sheet named
MOVES MCCumTWCoeff within the MOVES_MC_EmissionRates FTOl.xh
workbook was created based on the field structure of the corresponding MOVES
database table and used to store the calculated cold-soak TVV coefficients.
Finally, Liquid Leak (LL) emission rates for motorcycles were assumed to be zero. This
assumption was made because of the absence of liquid leak measurements specifically for
motorcycles coupled with the inherent uncertainty in extrapolating LL rates from light-
duty vehicles. This assumption also reflects the belief that, unlike cars and trucks, liquid
fuel leaks from motorcycles are much more visible (i.e., puddles or drips are not hidden
by the body/chassis) and that when detected, motorcycle owners would have their
vehicles repaired in short order.
All of the MOVES evaporative process emission rates as a function of OpModelD,
PolProcID, and Model YearGroupID were then assembled into a table in Rows 193-270
of the FTPEvap sheet and used to populate the evaporative process-specific records
within the EmissionRateByAge sheet in the MOVES_MC_EmissionRates_FT01.xls
workbook.
Motorcycle Activity Data Tables - The updated motorcycle activity data described earlier
in Section 3.5 were also converted into a series of Excel spreadsheets that conformed to
the data structures of corresponding activity data tables in the MOVES database. The
workbook MC_Activity_All.xls (provided on the delivery CD) was used to perform these
conversions. Updated motorcycle activity data were provided for the six separate
MOVES database tables listed below.
1. SourceTypeAge - normalized survival rates and mileage rates with age
2. SourceTypeAgeDistribution - normalized populations (registrations) with age
3. SourceTypeYear
4. HPMSVTypeYear
5. SampleVehicleDay
6. SampleVehicleTrip
First, the survival fractions and mileage accumulation rates presented earlier in Tables 3-
6 and 3-7, respectively, are contained in the sheets SurvivalCurve andMARCalcs within
the MC_Activity_AH.xls workbook. The SourceTypeAge table of the MOVES database
requires survival rates and annual mileage rates (normalized to a value of 1 at Age 0) for
ages 0 through 30 years. The properly normalized survival rates and mileage rates with
age were computed in the SurvivalCurve and MARCalcs sheets and loaded into a sheet
within the MC_Activity_AH.xls workbook called MOVES_MCSourceTypeAge with the
same field structure as the corresponding table in the MOVES database.
Second, normalized registration distributions with age were computed from the
population data by model year presented in Table 3-6 within the PopnAge sheet. These
-70-
-------
normalized registration distributions were then loaded into the sheet named
MOVES_MCSourceTypeAgeDistribut in the MC_Activity_Allxls workbook.
Next, updated motorcycle sales growth and migration rates for calendar years 2000
through 2008 were calculated in the GrowthCalcs sheet and loaded into the
MOVES_MCSourceTypeYear sheet in the MC_Activity_Allxls workbook. It was
necessary to calculate updated sales growth and migrates rates over this calendar year
range (2000-2008) to account for differences between the default 2008 population
projections contained in the MOVES database (projected from 1999 base year
populations) and those obtained directly from MIC under this effort.
VMT growth rates between 2000 and 2008 were similarly updated based on calculations
performed in the PopnAge sheet and loaded into the MOVES_MCHPMSVTypeYear
sheet. The updated VMT growth factors (expressed in MOVES as the change in VMT
relative to the previous calendar year) between 2000 and 2008 were based on annual U.S.
motorcycle VMT estimates by calendar year obtained from FHWA.*
Finally,the motorcycle-specific records in the SampleVehicleDay and SampleVehicleTrip
database tables were updated using a FORTRAN program called SynTables09 to
incorporate constraints on weekday (WD) vs. weekend (WE) trip ratios and activity. As
noted earlier in Section 3.5, U.S. DOT crash data32 suggests the weekday/weekend
activity split for motorcycles is about 40% weekdays and 60% weekends. (If motorcycle
trip/activity rates were constant across all seven days, the weekend fraction would only
be about 29%.)
SynTables09 is a modified version of an earlier program called SynTables developed by
Sierra under a 2007 MOVES-related EPA study. 8 The earlier program was modified to
ensure engine on-off trips by day of week matched this 40%/60% WD/WE split in the
SampleVehicleDay and SampleVehicleTrip tables for motorcycles.
MySQL Table Loading -The sheets in the MOVES_MC_EmissionRates_FT01.xls and
MC_Activity_All.xls Excel workbooks containing the emission rate and activity data in
structure identical to their corresponding MOVES database tables were exported to a
series of comma-delimited (CSV) files.
Two short MySQL batch scripts called Make_MOVESTable_MC_092809.txt and
Load_MOVESTables_4-03_092809.txt were then written and executed to load the new
emission rate data and updated activity data for motorcycles into the appropriate MOVES
database tables in MySQL format.
###
FHWA Highway Statistics VM-1 Table VMT by Year
-71-
-------
4. AIR TOXIC SPECIATION RATIOS
Draft MOVES2009 currently relies on algorithms carried over from MOBILE6.2 that are
used to calculate air toxic emissions from on-road vehicles. These algorithms cover the
following seven species: benzene, 1,3-butadiene, formaldehyde, acetaldehyde, acrolein,
naphthalene, and ethanol. The final analytical task under this effort envisioned
development of new speciation ratios for a broader list of Hazardous Air Pollutants
(HAPs) included in EPA's National Mobile Inventory Model (NMEVI) that were not
included in Draft MOVES2009.
This section of the report describes the work performed under this air toxics task. It
begins with an explanation of the form of air toxic emission rates defined and used in
MOVES and identification of the list of additional HAPs contained in NMEVI that EPA is
seeking to integrate into MOVES. It then presents separate discussions on the data
assembled and emission rates for individual air toxic species estimated for light-duty
gasoline- and heavy-duty Diesel-fueled vehicles, respectively.*
4.1 MOVES Speciation Ratios
Table 4-1 lists the additional HAP species that EPA is integrating into MOVES. Along
with the MOVES PollutantID and NEI pollutant codes that uniquely identify each
compound, it also lists the PollutantDisplayGroupID for each species. This value is used
internally within MOVES to group pollutants/compounds into logical groups for which
specific calculations or processing options are performed within the model. For the
HAPs listed in Table 4-1, values of-?," -8," and -9" refer to Additional Air Toxics,"
Dioxins/Furans," and -Metallic Air Toxics," respectively.
The rightmost column in Table 4-1 (and the row shading) identifies whether each
compound was included in the speciation ratios developed from the data sources
employed under this effort. Values of -G" and -0" refer to those species represented in
the light-duty gasoline and heavy-duty Diesel study datasets, respectively. Species for
which no data were available from the datasets examined under this effort are noted with
* EPA originally envisioned development of MOVES air toxic emission rates for other vehicle/fuel types
including heavy-duty gasoline vehicles, heavy-duty Diesel using bioDiesel, light and heavy-duty CNG and
LPG vehicles, as well as non-road engines. During the study, resources were re-prioritized into other tasks
such that air toxic emissions for these other vehicle/fuel types were not considered. EPA plans to pursue
development of MOVES air toxics rates for these vehicles under a separate study.
-72-
-------
Table 4-1
List of New MOVES HAPs and Numerical Codes
PollutantName
2 ,2,4 -Trimethy Ipentane
Acenaphthene
Acenaphthylene
Anthracene
Benz(a)anthracene
Benzo(a)pyrene
Benzo(b)fluoranthene
Benzo(g,h,i)perylene
Benzo(k)fluoranthene
Chrysene
Dibenzo(a,h)anthracene
Ethyl Benzene
Fluoranthene
Fluorene
Hexane
lndeno( 1 ,2 ,3 ,c ,d)pyrene
Phenanthrene
Propionaldehyde
Pyrene
Styrene
Toluene
Xylene
2,3,7,8-Tetrachlorodibenzo-p-Dioxin
1,2,3,7,8-Pentachlorodibenzo-p-Dioxin
1,2,3,4,7,8-Hexachlorodibenzo-p-Dioxin
1, 2,3,6 ,7,8-Hexachlorodibenzo-p-Dioxin
1,2,3,7,8,9-Hexachlorodibenzo-p-Dioxin
1,2,3,4,6,7,8-Heptachlorodibenzo-p-Dioxin
Octachlorodibenzo-p-Dioxin
2,3,7,8-Tetrachlorodibenzofuran
1,2,3,4,6,7,8-Heptachlorodibenzofuran
1,2,3,4,7,8,9-Heptachlorodibenzofuran
1,2,3,4,7,8-Hexachlorodibenzofuran
1,2,3,6,7,8-Hexachlorodibenzofuran
1,2,3,7,8,9-Hexachlorodibenzofuran
1 ,2,3,7,8-Pentachlorodibenzofuran
2,3,4,6,7,8-Hexachlorodibenzofuran
2,3,4,7,8-Pentachtorodibenzofuran
Octachlorodibenzofuran
MOVES
PollutantlD
40
41
42
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
NEIPollutant
Code
540841
83329
208968
120127
56553
50328
205992
191242
207089
218019
53703
100414
206440
86737
110543
193395
85018
123386
129000
100425
108883
1330207
1746016
40321764
39227286
57653857
19408743
35822469
3268879
51207319
67562394
55673897
70648269
57117449
72918219
57117416
60851345
57117314
39001020
Pollutant
DisplayGrouplD
7
T
T
T
T
T
T
T
T
T
T
1
T
T
1
T
T
1
T
1
1
1
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
Included
in Study
G,D
D
D
D
D
D
D
D
D
D
D
G
D
D
G,D
D
D
G,D
D
G
G,D
G
-
-
-
-
-
D
D
D
-
D
-
-
-
D
D
-
D
-73-
-------
Table 4-1
List of New MOVES HAPs and Numerical Codes
PollutantName
Mercury (elemental gaseous)
Mercury (divalent gaseous)
Mercury (particulate)
Arsenic & compounds
Chromium (Cr3+)
Chromium (Cr6+)
Manganese
Nickel
MOVES
PollutantlD
201
202
203
204
205
206
207
208
NEIPollutant
Code
200
201
202
93
16065831
18540299
7439965
7440020
Pollutant
DisplayGrouplD
9
9
9
9
9
9
9
9
Included
in Study
-
-
D
D
-
-
D
D
a Pollutants within the 16-compund Polycyclic Aromatic Hydrocarbon (PAH) group (that also includes
naphthalene) defined by EPA that occur in both gaseous and particle phases.
a dash (—') symbol. As indicated in this column, measured test data were found for
most, but not all, of the additional HAPs identified for inclusion in MOVES.
Within MOVES, air toxic emission rates are calculated based on toxic -speciation ratios"
that represent the mass emission fraction of an individual species relative to VOC
(volatile organic compounds) or PM (particulate matter) of the following form:
Speciation Ratio = Species Mass Emission Rate / VOC Mass Emission Rate
or
= Species Mass Emission Rate /PMMass Emission Rate
Speciation ratios for those compounds listed Table 4-1 with on-road measurement data
were calculated by dividing the species emission rate by the VOC emission rate for the
gaseous compounds (those with PollutantDisplayGroupID=7 or 8) or by the PM emission
rate for metal-based compounds (PollutantDisplayGroupID=9). Within MOVES, these
speciation ratios are multiplied by either VOC or PM emission rates as appropriate to
produce air toxic emission rates (in grams per hour) by vehicle type, model year, etc. and
generate composite rates for the vehicle fleet.
There was one exception to this basic approach. Polycyclic aromatic hydrocarbons
(PAHs) identified as footnoted in the PollutantDisplayGroupID column in Table 4-1 are
found in multiple phases: (1) gas phase, (2) particle phase, and (3) as semi-volatile
organic compounds. As explained later in Section 5.3, multi-phase apportionment of
PAH species measurements were performed based on factors from EPA's SPECIATE
database.
-74-
-------
Thus, the basic approach to this task consisted of reviewing literature and previous testing
study datasets in which either individual air toxic emission rates were measured in
conjunction with measurements of denominator pollutant emission rates or speciation
ratios were developed from these measurements. EPA requested development of separate
speciation ratios by vehicle type (i.e., SourceType), model year range or certification
standards group (e.g., Tier 0, 1, 2), and fuel type where sufficient data were available.
Given this overall approach, the remaining two sub-sections provide further detail on the
assembly of data and development of speciation ratios for light-duty gasoline and heavy-
duty Diesel-fueled vehicles, respectively.
4.2 Light-Duty Gasoline-Fueled Ratios
Development of light-duty gasoline-fueled toxic speciation ratios was relatively
straightforward. Of the nearly 50 additional toxic species listed earlier in Table 4-1,
measurement data and ratios were available for seven species for light-duty gasoline
vehicles.
Ratios for Tier 1 and earlier (model year 2003 and earlier) vehicles were assembled from
SPECIATE model profiles39 developed for EPA by ENVIRON. Separate speciation
ratios were developed for exhaust and evaporative emission processes and the following
fuels/blends: gasoline (i.e., EO), E10 (10% ethanol, by volume), and E85 (85% ethanol).
Table 4-2 presents these Tier 1 and earlier speciation ratios assembled from the
ENVIRON study.
Table 4-2
Tier 1 and Earlier Light-Duty Gasoline Vehicle Speciation Ratios
Pollutant
Name
2,2,4-Trimethylpentane
Ethyl Benzene
Hexane
Propionaldehyde
Styrene
Toluene
Xylene a
Exhaust Emission Ratio
Gasoline
0.01823
0.02147
0.01570
0.00086
0.00108
0.09619
0.07814
E10
0.01849
0.01932
0.01593
0.00087
0.00097
0.08657
0.07032
E85
0.00898
0.00222
0.00213
0.00019
0.00022
0.00813
0.00699
Evaporative Emission Ratio
Gasoline
0.01984
0.02521
0.02217
n/a
0.00000
0.09643
0.07999
E10
0.03354
0.01721
0.02536
n/a
0.00000
0.14336
0.06423
E85
0.01949
0.00651
0.00606
n/a
0.00043
0.05193
0.02205
n/a - not available
a Xylene is reported as the sum of O-xylene and M- and p-xylene.
-75-
-------
Toxic speciation ratios for Tier 2 and later vehicles (model year 2004 and newer) were
based on SPECIATE model profiles developed by EPA.40 There were only exhaust
emission data available for Gasoline and E10. No evaporative emission profiles or E85
profiles were available. Table 4-3 lists these available Tier 2 speciation ratios.
Table 4-3
Tier 2 Light-Duty Gasoline Vehicle Speciation Ratios
Pollutant
Name
2,2,4-Trimethylpentane
Ethyl Benzene
Hexane
Propionaldehyde
Styrene
Toluene
Xylenea
Exhaust Emission Ratio
Gasoline
0.0319
0.0168
0.0028
0.0012
0.0009
0.0754
0.0613
E10
0.0123
0.0166
0.0291
0.0005
0.0008
0.0744
0.0605
1 Xylene is reported as the sum of O-xylene and M- and p-xylene.
4.3 Heavy-Duty Diesel-Fueled Ratios
Diesel-fueled speciation ratios were calculated from emission measurements that were
assembled into a comprehensive spreadsheet database that was assembled under the
Coordinating Research Council (CRC) E-75 study.41 This database organized speciated
exhaust emission measurements from on-road medium- and heavy-duty Diesel vehicles
from several earlier studies that are listed in Table 4-4.
The E-75 database is a large spreadsheet with information from the prior testing studies
organized into individual sheets as summarized below.
• Engine Data - Contains descriptive data of the vehicle/engine that was tested,
such as make/manufacturer, vehicle type, engine application, model, model year,
engine size, horsepower, etc.
• Fuel Data - Lists basic name or type of fuel used for each test (e.g., conventional
Diesel, bioDiesel, low-sulfur Diesel, etc.) and properties (cetane number, sulfur,
oxygen, aromatic content, etc.) of that fuel.
-76-
-------
Table 4-4
Study Data Represented in CRC-75 Speciation Database
Study Name
Southwest Research Institute (SwRI) School Bus Study
SwRI study for the Ad Hoc Diesel Fuel Test Program
New York City Clean Diesel Demonstration Program
California Institute of Technology
BP Southern California ultra-low sulfur Diesel
(ULSD)/DPF/CNG Heavy-duty study
SwRI Fischer-Tropsch study
CRC Mass vehicle tests
CRC AVFL-lOa & lOb
CE-CERT
Gasoline-Diesel paniculate matter (PM) Split Study
CRC E55/59
Desert Research Institute
Environment Canada
Vehicles Tested
2 school buses
1LDDV
2 Diesel transit buses
2 MHDDT, 2 HHDDT
3HDDT
1 transit bus 1 school bus
2 HDD engines
16 LDDVs
4 LDDV/T,
-110L/HDDV
14HDDT, 28 LDDV/T
30 HDDT 2 buses
1 MHDDT 8 HHDDT
4 LDDV 4 Diesel trucks
1LDDT 2 urban buses
Model Years
Covered
2001
N/A
1999
1995 1987
1996-1999
1998 1998
1999, 2000
1977-1994
2004 1978-2000
1996-2000 1983-
1999
1982-2001 1982,
1992
1997 1985-2003
1991, 1998, 1999,
2000 N/A
1998 1989, 1998
• Pollutant List - Provides a cross-referenced list of nearly 1,000 individual
pollutant species represented in the assembled test data that included primary and
alternate pollutant names, Chemical Abstracts Service (CAS) numbers, and
pollutant ID codes.
• Test Cycle - Provides a description for each of the test cycle codes used in the
database that identified the driving cycle over which emission measurements for
each test were collected.
• Emissions - The emissions data were stored in several individual sheets broken
into the following compound groups: regulated pollutants (e.g., THC, CO, NOx,
PM, etc.), carbonyls, dioxins/furans, PAHs, speciated hydrocarbons, SVOCs and
elements/inorganic compounds (i.e., metals).
The data records in each of the individual sheets in the E-75 database could be related or
linked to one another using Project ID, Study ID, and Test ID fields contained in each
sheet that uniquely identified data for a specific emission test.
-77-
-------
Data Processing and Validation - Despite the organization of the E-75 spreadsheet,
development of toxic speciation ratios for heavy-duty Diesel vehicles was much more
involved than for the gasoline-fueled vehicles because a number of quality assurance
checks, data linkage processes, and record culling steps had to be performed. A modified
version of the original E-75 database ca\\edMOVES_AT_Speciations_Diesel_E-75.xls
was created in which these steps were performed. (Both the original and modified E-75
spreadsheet databases are included on the delivery CD.) These processing steps are
summarized below.
• Removal of Invalid Data - Data that had —hd," -N/A," -ad," or —99" entries in
either the Emission or Emission Unit columns were excluded.
• Assembly of Paired Measurements - Paired measurements for the same vehicle
and test that have both toxic emission measurements (numerator) and regulated
emission measurements (denominator) were located in different sheets within the
database. They were combined into a single sheet in the spreadsheet called
AllPollutcmts by lookup formulas based on ProjectTD, Study ID, and TestID fields
in each individual sheet in order to calculate speciation ratios. In a number of
cases, paired measurements of both the numerator and denominator measurements
were not found using this lookup approach. Further investigation of these cases
indicated that some of the mismatches occurred from incorrect entries (e.g., typos)
in one of these three lookup fields (usually the TestID field) as originally stored in
the database. Where it was obvious that this type of error had been made,
corrective edits were made to values in the lookup fields to properly retrieve both
the toxic species emission measurement and the paired -denominator" or base
pollutant measurement for the same vehicle and test.
• Unit Conversions - In a number of instances, the emission units for the paired
measurements were different. In these instances, the denominator emission unit
was converted to a numerator emission unit to properly calculate the toxic
speciation ratio. In some cases, incompatible emission units were observed—for
example, if the numerator emission unit was —mgnile" and the denominator unit
was -g/mode." (The testing studies included both transient cycles and steady-
state modes.) In these cases where incompatible units were observed, the paired
measurements were excluded.
• Removal of Non-Conventional Fuel Tests - The E-75 database included test
measurements from a number of research studies for which a variety of Diesel
blends and other alternative fuels were used. For this analysis, EPA directed
Sierra to exclude all tests for fuels other than those that reflect conventional or
low-sulfur Diesel blends currently or recently in-use within the U.S. As a result,
tests for all other types of fuels contained in the E-75 database were excluded or
removed from further analysis. These non-conventional blends included Fischer-
Tropsch, bioDiesel, ethanol-Diesel blends, emulsified fuel, European blends, and
other obvious research fuels.
-78-
-------
MOVES Source Type Categorizations - Based on vehicle description, engine
application, make, and model entries in the Engine_Data sheet, classifications of
vehicles into the MOVES SourceType scheme were performed when possible. In
many cases, such for vehicles marked as —R&se Hauler" or -School Bus," it was
simple to assign the appropriate SourceType category. For others, internet
searches performed based on manufacturer, model, and engine size clearly
revealed the type of heavy-duty vehicle application (e.g., single- vs. combination-
unit or delivery vehicle). In instances were there was ambiguity, a missing code
was simply entered for the SourceTypelD field. The purpose behind these
categorizations was to enable development of separate speciation ratios by
SourceType for use in MOVES. However, once the categorizations were made
and sample sizes were tabulated by SourceType, it became apparent that
insufficient data were available for most of the SourceType categories containing
Diesel-fueled vehicles. Given these small sample sizes, subsequent tabulations of
Diesel speciation ratios under this effort were not SourceType-specific, but based
on averages across all valid SourceTypes. (However, these SourceType
categorizations were preserved for later use when these data are combined with
other datasets for which MOVES SourceType categorizations are needed.)
Removal ofNon Heavy-Duty Vehicle Tests - When performing the SourceType
categorizations, a small number of tests were found to have been included in the
E-75 database for light-duty Diesel vehicles (e.g., Volkswagen Jetta and Toyota
Corolla, Ford F-250). Because of the decision described above to simply
calculate average speciation ratios across all valid vehicle types due to insufficient
sample sizes), it was jointly agreed with EPA to remove these few light-duty
vehicle tests from the validated E-75 sample. Thus, the resulting speciation ratios
were solely representative of heavy-duty Diesel vehicles.
Outlier Treatment - A simple screening threshold of 0.5 was used to filter and
remove data with implausible speciation ratios. Ratios greater than or equal to 0.5
are considered as outliers and the data for these test records were removed from
the final, validated sample contained in the AllPollutants sheet of the
MOVES_AT_Speciations_Diesel_E-75.xlsfi\e.
Treatment of Multi-Phase PAH Species - In addition to the data processing and
validation steps described above, additional processing was performed for the PAH
compounds identified earlier in Table 4-1. This additional process was performed to
account for the fact that PAHs occur in multiple phases: gas phase, particle phase, and as
SVOCs.
Multi-phase allocation factors were developed for the 15 individual PAH species
contained in Table 4-1 using estimates from EPA's SPECIATE4.2 database42 where
available. For species where allocations were not found in the SPECIATE database,
emissions were estimated to occur either in the gas or particle phase based on each PAH
-79-
-------
compound's physical properties. Those with lower molecular weights tend to be volatile,
and were allocated to the gas phase; —bavy" higher weight compounds with higher
molecular weights were allocated to the particle phase.
Table 4-5 summarizes the multi-phase allocation factors used to apportion emission rates
for individual PAH compounds to the gas and particle phases. In the E-75 database, the
speciation data were organized in a manner that identified whether the test measurement
reflected the gas phase, particle phase, or both phases. If a test record was marked as
-Gas Phase," the speciation ratio calculated from the specie (numerator) and base
(denominator) pollutant emission rates was adjusted by the Gas Phase fraction listed in
Table 4-5. This adjustment was applied similarly for those records identified as —P&icle
Phase" tests. When the E-75 Phase field indicated the measurements were collected in
both phases, no PAH allocation adjustments from Table 4-5 were used. PAH records for
which no information was recorded in the Phase field were assumed to be collected in
particle (PM) phase.
Table 4-5
Mult-Phase PAH Allocation Factors Used
PAH Species
Acenapthene
Acenaphthylene
Anthracene
Benz(a)anthracene
Benzo(a)pyrene
Benzo(b)fluoranthene
Benzo(g,h,i)perylene
Benzo(k)fluoranthene
Chrysene
Dibenzo(a,h)anthracene
Fluoranthene
Fluorene
Indeno( 1 ,2,3 -cd)pyrene
Phenanthrene
Pyrene
Molecular
Weight
154
152
178
228
252
252
276
252
228
278
202
166
276
178
202
Allocation Fraction
PM Phase
0
0
0.466
0.723
1
1
0.773
1
0.823
1
0.516
0.215
1
0.335
0.552
Gaseous Phase
1
1
0.534
0.277
0
0
0.227
0
0.177
0
0.484
0.785
0
0.665
0.448
Reference
EPA/SPECIATE
EPA/SPECIATE
EPA/SPECIATE
EPA/SPECIATE
Sierra Research
Sierra Research
EPA/SPECIATE
Sierra Research
EPA/SPECIATE
Sierra Research
EPA/SPECIATE
EPA/SPECIATE
Sierra Research
EPA/SPECIATE
EPA/SPECIATE
Heavy-Duty Diesel Speciation Ratios - Once the processing, validation, and PAH
allocations were completed, average speciation ratios for each available toxic pollutant
were then calculated across all vehicle types and model years represented in the validated
-80-
-------
sample. Table 4-6 below shows these mean speciation ratios, along with the sample sizes
(number of vehicle tests) and standard deviations (to provide a sense of the variation of
the individual observations).
Table 4-6
Heavy-Duty Diesel Speciation Ratios and Basic Statistics
Pollutant Name
1,2,3,4,6,7,8-HpCDD
1,2,3,4,7,8,9-HpCDF
1,2,3,7,8-PeCDD
2,2,4-trimethylpentane
2,3,4,6,7,8-HxCDF
2,3,7,8-TCDF
Acenaphthene
Acenaphthylene
Anthracene
benz[a]anthracene
benzo[a]pyrene
benzo [b]fluoranthene
benzo [ghi]perylene
benzo [k]fluoranthene
Chrysene
dibenz[ah]anthracene
Fluoranthene
Fluorine
Hexane
indeno [123 -cd]pyrene
OCDD
OCDF
Phenanthrene
Propionaldehyde
Pyrene
Toluene
As
Hg
Mn
Ni
Pollutant Group
DioxinFuran
DioxinFuran
DioxinFuran
TOG-VOC-NMOG
DioxinFuran
DioxinFuran
PAH
PAH
PAH
PAH
PAH
PAH
PAH
PAH
PAH
PAH
PAH
PAH
TOG-VOC-NMOG
PAH
DioxinFuran
DioxinFuran
PAH
carbonyl
PAH
TOG-VOC-NMOG
Pm
Pm
Pm
Pm
Count
1
1
1
43
1
1
105
104
102
115
101
15
89
13
115
2
123
108
38
88
1
1
106
60
123
49
95
98
98
97
Average
1.95E-10
2.49E-11
5.14E-12
1.90E-03
1.98E-11
4.11E-11
2.46E-04
7.49E-04
1.44E-03
2.49E-04
9.44E-04
1.17E-05
1.08E-05
6.52E-06
7.87E-05
2.12E-06
1.72E-03
2.20E-03
2.07E-03
8.40E-06
7.71E-11
6.94E-11
2.80E-03
4.91E-03
1.95E-03
4.54E-03
1.67E-05
4.87E-05
3.57E-05
2.65E-05
Std Dev
n/a
n/a
n/a
5.19E-03
n/a
n/a
5.75E-04
1.63E-03
6.70E-03
9.87E-04
7.48E-03
1.07E-05
3.97E-05
1.09E-05
1.39E-04
1.80E-06
7.85E-03
7.21E-03
7.21E-03
2.47E-05
n/a
n/a
9.31E-03
7.34E-03
8.75E-03
5.17E-03
4.63E-05
1.33E-04
8.00E-05
4.69E-05
###
-81-
-------
5. REFERENCES
1. -Development of Emissions Rates for Light-Duty Vehicles in the Motor Vehicle
Emissions Simulator (MOVES2009) - Draft Report," EPA-420-P-09-002, August
2009.
2. T. Durbin et al., —le Effect of Fuel Sulfur on NH3 and Other Emissions from
2000-2001 Model Year Vehicles," Final Report submitted to program sponsors CRC
and CONCAWE, May 2003.
3. C.A. Harvey et al., -A Study of the Potential Impact of Some Unregulated Motor
Vehicle Emissions," SAE 830987 presented at the Passenger Car Meeting, June 1983.
4. M.M. Baum et al., —Multcomponent Remote Sensing of Vehicle Exhaust by
Dispersive Absorption Spectroscopy. 2. Direct On-Road Ammonia Measurements,"
Environ. Sci. Technol. 2001, 35, 3735-3741.
5. AJ. Kean, et al., -Trends in on-road vehicle emissions of ammonia," Atmospheric
Environment 43 (2009) 1565-1570.
6. E.L. Glover, -Development of Gasoline and Diesel Vehicle Sulfate Emissions for the
MOVES Model," undated DRAFT.
7. N.V. Heeb et al., -Trends of NO-, NO2-, and NHa-emissions from gasoline-fueled
Euro-3- to Euro-4-passenger cars," Atmospheric Environment, 42 (2008), 2543-2554.
8. M. Lenner et al., —le NO2/NOx Ratio in Emissions from Gasoline-Powered Cars:
High NO2 percentage in idle engine measurements," Atmospheric Environment
(1967), Volume 17, Issue 8, 1983, Pages 1395-1398.
9. M. Lenner, -Nitrogen Dioxide in Exhaust Emission Measurements from Motor
Vehicles," Atmospheric Environment (1967), Volume 21, issue 1, pages 37-43.
10. J.P. Shi and R.M. Harrison, -Rapid NO2 Formation in Diluted Petrol-Fuelled Exhaust
- a Source of NO2 in Winter Smog Episodes," Atmospheric Environment (1997)
Volume 31, No 23, pp.38573866.
11. T.D. Durbin et al., -Emissions of Ammonia from Light-Duty Vehicles," Cooperative
Agreement with the US EPA (CX827692-01-0).
-82-
-------
12. T. Lanni et al., —Pdbrmance and Emissions Evaluation of Compressed Natural Gas
and Clean Diesel Buses at New York City's Metropolitan Transit Authority," SAE
paper 2003-01-0300.
13. A. Ayala et al., —Efcsel and CNG Heavy-Duty Transit Bus Emissions over Multiple
Driving Schedules: Regulated Pollutants and Project Overview," SAE paper 2002-01-
1722, May 2002.
14. F. DiGenova, et al., ^fhe Potential Significance of Motor Vehicle Crankcase
Emissions in the Mobile Source Emissions Inventory," Sierra Research, Inc., 6th
CRC On-Road Vehicle Emissions Workshop, March 1996.
15. C.T. Hare, et al., —Motorycle Emissions, Their Impact, and Possible Control
Techniques," Southwest Research Institute and U.S. Environmental Protection
Agency, SAE Paper No. 740627, 1974.
16. -Mobile Source Emission Standards Summary," U.S. Environmental Protection
Agency, July 16, 1992.
17. FJ. Mearing, -Final Report of the Motorcycle Surveillance Test Program, First
Series, CVS (MCSTP-I)," Air Resources Board, In-Use Vehicle Surveillance Section,
ARE Report No. M.S. 80-005, February 1980.
18. -Test Report of the Motorcycle Surveillance Test Program - Series 2 (MCSTP-II),"
Air Resources Board, Mobile Source Control Division, Report No. M.S. 81-006,
March 13, 1981.
19. J. Jones, —fiial Report for Surveillance Testing of Motorcycles for Emissions,"
Project No. 97-600, California Environmental Engineering, February 29, 2000.
20. J.D. Anderson, et al., -DfT Motorcycle Emissions Measurement Programmes:
Regulated Emissions Results," SAE Paper No. 2003-01-1897, 2003.
21. -Final Regulatory Support Document: Control of Emissions from Highway
Motorcycles," Assessment and Standards Division, Office of Transportation and Air
Quality, U.S. Environmental Protection Agency, Report No. EPA420-R-03-015,
December 2003.
22. Personal communication with Dennis McClement, February 2, 2009.
23. G. F. Baltz, -Effects of Alcohol Extended Fuels on the Rate of Fuel Hose
Permeation," SAE Paper No. 880709, 1988.
24. -Control of Emissions from Marine SI and Small SI Engines, Vessels, and
Equipment, Final Regulatory Impact Analysis, Chapter 5, Feasibility of Evaporative
Emission Control," Assessment and Standards Division, Office of Transportation and
-83-
-------
Air Quality, U.S. Environmental Protection Agency, Report No. EPA420-R-08-014,
September 2008.
25. Unpublished survey results provided by Pamela Amette, Motorcycle Industry
Council.
26. -Motorcycle Industry Council Proposal for Revised Exhaust Emissions Standards for
On-Road Motorcycles," Motorcycle Industry Council, October 7, 1998.
27. -Report on Potential Exemptions from Vehicle Emissions Testing for Motorcycles,
Collectible Vehicles and Vehicles 25 Model Years Old and Older," Arizona
Department of Environmental Quality, Air Quality Division, December 2004.
28. -Statistical Annual 2008," Motorcycle Industry Council, 2009.
29. Unpublished retail sales estimates for 2008 provided by Pamela Amette, Vice-
President, Motorcycle Industry Council, June 25, 2009.
30. Unpublished retail sales estimates for first six months of 2009 provided by Pamela
Amette, Vice-President, Motorcycle Industry Council, July 24, 2009.
31. Draft MIC Motorcycle/ATV Owner Survey results provided by Pamela Amette,
Vice-President, Motorcycle Industry Council, June 30, 2009.
32. Anders Longthorne, Cherian Varghese, and Umesh Shankar, -Fatal Two-Vehicle
Motorcycle Crashes," DOT HS 810 834, September 2007.
33. Sierra Research, Inc., ^fask Order No. 2 SCF Improvement - Field Data Collection,"
prepared for California Department of Transportation, Report No. SR02-07-04,
July 18,2002.
34. T. Carlson, et al., -Task Order No.7 SCF Improvement - Driving Data Collection,
South Coast Air Basin," prepared for California Department of Transportation,
Report No. SR02-07-05, July 31, 2002.
35. J. Jones, —fiial Report for Surveillance Testing of Motorcycle Emissions #97-600,"
prepared for California Air Resources Board and California Environmental Protection
Agency, California Environmental Engineering, February 29, 2000.
36. -Development of Evaporative Emissions Calculations for the Motor Vehicle
Emissions Simulator (Draft MOVES2009), Draft Report," U.S. Environmental
Protection Agency, Report No. EPA-420-P-09-006, August 2009.
37. R. Reddy, -Prediction of Fuel Vapour Generation from a Vehicle Fuel Tank as a
Function of Fuel RVP and Temperature," SAE892089, 1989.
-84-
-------
38. T. Carlson, et al., -Development of Trip and Soak Activity Defaults for Passenger
Cars and Trucks in MOVES," Report No. SR2007-06-01, June 29, 2007.
39. C.E. Lindhjem, Emission Profiles for EPA SPECIATE Database, EPA Contract No.
EP-C-06-094, Work Assignment No. 1-7, ENVIRON International Corporation,
January 31,2008.
40. EPA, Exhaust Emission Profiles for EPA SPECIATE Database: Energy Policy Act
(EPAct) Low-Level Ethanol Fuel Blends and Tier 2 Light-Duty Vehicles, EPA-420-
R-09-002, Assessment and Standards Division Office of Transportation and Air
Quality U.S. EPA, June 2008.
41. Coordinating Research Council (CRC), Compilation of Diesel Emissions Speciation
Data Final Report, CRC Contract NO. E-75, October 2007.
42. U.S. Environmental Protection Agency, SPECIATE 4.2 Speciation Database
Development Documentation, Office of Research and Development U.S. EPA,
EPA/600-R-09/038, June 2009.
-85-
------- |