MOVES2010 Highway Vehicle
Population and Activity Data
&EPA
United States
Environmental Protection
Agency
-------
MOVES2010 Highway Vehicle
Population and Activity Data
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-10-026
November 2010
-------
Table of Contents
1. Introduction 8
1.1. Default Inputs and Fleet and Activity Generators 8
1.2. MOVES SourceTypes 11
2. Data Sources 13
2.1. VIUS(andTIUS) 13
2.2. Polk NVPP® and TIP® 13
2.3. FHWAHighway Statistics 13
2.4. FTA National Transit Database 13
2.5. School Bus Fleet Fact Book 14
2.6. MOBILE6 14
2.7. Annual Energy Outlook & National Energy Modeling System 14
2.8. Transportation Energy Data Book 14
2.9. Oak Ridge National Laboratory Light-duty Vehicle Database 14
3. Vehicle Population Data by Calendar Year 15
3.1. 1999 SourceTypePopulation 15
3.2. 1990 SourceTypePopulation 19
3.3. SalesGrowthFactor 23
3.4. MigrationRate 27
4. Emission-Related Vehicle Characteristics (Source Bins) 27
4.1. Motorcycles 29
4.2. Passenger Cars 30
4.3. General Trucks 31
4.4. Buses 43
4.5. Refuse Trucks 50
4.6. Motor Homes 52
5. Age Distributions 56
5.1. Motorcycles 56
5.2. Passenger Cars 57
5.3. Trucks 58
5.4. Intercity Buses 60
5.5. School Buses and Motor Homes 60
5.6. Transit Buses 60
6. Vehicle Characteristics that Vary by Age 62
6.1. SurvivalRate 62
6.2. Relative Mileage Accumulation Rate 64
7. Vehicle-Specific-Power Characteristics by SourceType 67
7.1. SourceMass and Fixed Mass Factor 68
7.2. Road Load Coefficients 69
8. VMT by Year and Vehicle Type 71
S.l.HPMSBaseYearVMT 71
8.2. BaseYearOffNetVMT 72
8.3. VMTGrowthFactor 72
9. Roadtypes, VMT Distribution among Roadtypes, and Mappings to SCC 76
9.1. RoadTypeVMTFraction 77
-------
9.2. SCCRoadTypeDistribution 78
10. Average Speed Distributions 79
11. Driving Schedule Tables 82
12. Temporal Distributions of VMT and Hourly Extended Idle Activity 86
12.1. MonthVMTFraction 86
12.2. DayVMTFraction 87
12.3.HourVMTFraction 88
12.4. Extended Idle Activity by Hour 88
13. Vehicle Starts and Parking Activity 90
14. Geographical Allocation of Activity 91
14.1. SHOAllocFactor 92
14.2. StartAllocFactor and SHPAllocFactor 92
14.3. IdleAllocFactor 93
15. Air Conditioning Activity Inputs 94
15.1. ACPenetrationFraction 94
15.2. FunctioningACFraction 95
15.3. ACActivityTerms 96
16. Conclusion 98
17. References 100
Appendix A. Response to Peer Review Comments (A) 107
Appendix B. Response to Peer Review Comments (B) 121
-------
List of Tables and Figures
Table 1-1. MOVES Database Elements Covered in This Report 10
Table 1-2. MOVES2010 SourceTypes 12
Table 3-1. Vehicle Population Comparisons 1999 16
Table 3-2. Adjusted Vehicle Populations 16
Table 3-3. VIUS 1997 Codes Used for Distinguishing Truck SourceTypes 17
Table 3-4. VIUS 2002 Codes Used for Distinguishing Truck SourceTypes 17
Table 3-5. 1999 Truck SourceType Distribution and Populations 18
Table 3-6. 1999 Bus Population Comparisons 18
Table 3-7. 1999 SourceType Populations in MOVES2010 19
Table 3-8. 1990 Vehicle Population Comparisons 20
Table 3-9. TIUS92 Codes Used for Distinguishing Truck SourceTypes 21
Table 3-10. 1990 Truck SourceType Distribution and Populations 21
Table 3-11. 1990 Bus Population Comparisons 22
Table 3-12. 1990 SourceType Populations in MOVES2010 22
Table 3-13. SalesGrowthFactor by Calendar Year and Source Type 25
Table 4-1. Data Tables Used to Allocate SourceType to SourceBin 28
Table 4-2. Motorcycle Engine Size and Average Weight Distributions for Selected Model Years 29
Figure 4-1. Diesel Fractions for Light Trucks 32
Figure 4-2. Diesel Fractions for Single Unit Trucks 33
Table 4-3. Diesel Fractions for Trucks 34
Table 4-4. Mapping VIUS Engine Size Categories to MOVES EngSizelD 35
Table 4-5. Mapping VIUS Average Weight to MOVES WeightClassID 38
Table 4-6. Regulatory Classes in MOVES 39
Table 4-7. Light Truck Class 2 Weight Distribution 40
Table 4-8. Passenger & Light Commercial Truck Regulatory Class Fractions 41
Table 4-9. Fraction of Medium Heavy-Duty Trucks among Diesel-fueled Single-Unit and Combination
Trucks 42
Table 4-10. Fraction of Medium Heavy-Duty Trucks among Gasoline-fueled Single-Unit and
Combination Trucks 43
Table 4-11. Mapping National Transit Database Fuel Types to MOVES Fuel Types 44
Table 4-12. National Transit Database Implied Fuel Fractions for Transit Buses 45
Table 4-13. Transit Bus Fuel Fractions in MOVES2010 46
Table 4-14. School Bus Fuel Fractions in MOVES2010 47
Table 4-15. FTA Estimate of Bus Weights 47
Table 4-16. California School Buses 48
Table 4-17. Weight Distributions for Buses by Fuel Type 49
Table 4-18. SCC Mappings for Buses 49
Table 4-19. Refuse Truck Size Weight Fractions by Fuel Type 51
Table 4.20. SCC Mappings for Diesel Refuse Trucks 52
Table 4-21. Diesel Fractions for Motor Homes 53
Table 4-22. Weight Fractions for Diesel Motor Homes by Model Year 54
Table 4.23. Weight Fractions for Gasoline Motor Homes by Model Year 55
Table 4-24. SCCVtype Distributions for Diesel Motor Homes by Model Year 56
Figure 5-1 1999 Age Distributions for Passenger Cars 58
Figure 5.2 1999 Age Distributions for Passenger and Light Commercial Trucks 59
Table 5-1. 1999 Age Fractions for MOVES Source Types 61
Table 6-1. SurvivalRate by Age and SourceType 64
Figure 6-1. Relative Mileage Accumulation Rates in MOVES2010 66
-------
Table 7-1. MOVES Weight Classes 69
Table 7-2. Road Load Coefficients for Heavy-Duty Trucks, Buses, and Motor Homes 70
Table 7-3. SourceUseType Characteristics 71
Table 8-1. 1999 VMT by HPMS Vehicle Class 72
Table 8-2. VMT Growth Factors in MOVES2010 73
Table 9-1. Road Type Codes in MOVES 76
Table 9-2. Sourcetype VMT distribution among Road Types 77
Table 9-3. SCC RoadTypes 78
Table 10-1. MOVES Speed Bin Categories 80
Figure 10-1 Example Speed Distribution by Roadtype 81
Table 11-1. Driving Cycles for Motorcycles, Cars, Passenger Cars and Light Commercial Trucks 83
Table 11-2. Driving Cycles for Intercity Buses, Single-Unit Trucks and Motor Homes 83
Table 11-3. Driving Cycles for Combination Trucks 84
Table 11-4. Driving Cycles for Transit and School Buses 84
Table 11-5. Driving Cycles for RefuseTrucks 85
Table 12-1. MonthVMTFraction 87
Table 12-2. DayVMTFractions 87
Figure 12-1 Hourly VMT Fractions in MOVES2010 88
Figure 12-2 Extended Idle Activity Ratio 89
Table 13-1. Source Data for Sample Vehicle Trip Information 90
Table 13-2. Synthesis of Sample Vehicles for Source Types Lacking Data 91
Table 13-3. Starts per Day by SourceType 91
Table 15.1. AC Penetration Fractions in MOVES2010 95
Table 15-2. FunctioningACFraction by Age (All Use Types Except Motorcycles) 96
Table 15-3. Air Conditioning Activity Coefficients 97
Figure 15-1: Air Conditioning Activity Demand as a Function of Heat Index 97
-------
7
-------
1. Introduction
The Environmental Protection Agency's MOVES (Motor Vehicle Emission Simulator)
is a new set of modeling tools for estimating emissions produced by on-road (cars, trucks,
motorcycles, etc.) and, eventually, nonroad (backhoes, lawnmowers, etc) mobile sources.
MOVES2010 estimates greenhouse gases (GHG), criteria pollutants and selected air toxics from
highway vehicles. MOVES2010 replaces MOBILE6.2 as the model for use in official state
implementation plan (SIP) submissions to EPA and for transportation conformity analyses
outside of California
MOVES calculates emissions for running exhaust, start exhaust, a number of evaporative
processes and several other emission processes. In general, MOVES calculates these emissions
by multiplying emission rates by emission activity and applying correction factors as needed.
The emission rates and activity in MOVES are distinguished at much finer level than in
MOBILE6. For example, most running emissions are categorized into one of 25 operating
modes, depending on vehicle speed and vehicle specific power (VSP). Start emissions are
distinguished based on the time a vehicle has been idle prior to start, and evaporative emissions
modes are defined based on whether the vehicle is operating or has recently been operating.
Vehicles are categorized into narrow subtypes or "source bins" with similar fuels, engine sizes
and other emission-related characteristics.
MOVES is distributed with a default database of MOVES input data. The "domain" for
the default database is the entire United States. MOVES users may create other domains for the
model by supplying replacement data. In particular, EPA has issued "Technical Guidance on the
Use of MOVES2010 for Emission Inventory Preparation in State Implementation Plans and
Transportation Conformity1" for information on developing appropriate local inputs for SIP and
conformity MOVES runs.
This report describes the default database information on vehicle population and vehicle
activity as distributed in MOVES2010 and MOVES2010a (MOVESDB20091221 and
MOVESDB20100830). Generally, the fleet & activity values in the MOVES2010a database are
identical to those in MOVES2010. Where this is not true, the differences are explained in the
text of this report. Emission rates and correction factors values in the default databases are
described in other MOVES technical reports.2
1.1. Default Inputs and Fleet and Activity Generators
Much of the fleet & activity data used in the MOVES "core model" are calculated from
inputs that are in format that is condensed or more readily available. MOVES uses "generators"
to populate Core Model Input Tables (CMITS) from user inputs and MOVES defaults.
The Total Activity Generators (TAGs) estimates activity hours by taking base-year
vehicle miles travelled (VMT) estimates, growing the VMT to the analysis year, and using speed
information to transform VMT into source hours operating (SHO). The default database for
MOVES2010 has two base years: 1990 and 1999. Other types of vehicle activity are generated
by growing vehicle populations and applying appropriate conversions. For national inventory
runs, annual national activity is distributed in time and geography using distribution factors. A
separate version of the TAG creates inputs for emission rate runs.
8
-------
The Source Bin Distribution Generator (SBDG) uses information on gasoline/diesel
fractions, weightclass distributions and similar information to estimate the number of vehicles
belonging to each narrow sourcebin as a function of sourcetype and vehicle model year.
The Operating Mode Distribution Generators use information on speed distributions and
driving patterns to develop operating mode fractions for each sourcetype, roadtype and time of
day.
The details of each these generators and other MOVES2010 algorithms are described in
the MOVES Software Design and Reference Manual.3
This paper documents the sources and calculations used to produce the default population
and activity data in the MOVES2010 database used to compute national level emissions based on
defaults for individual counties, months, daytypes and hours of the day. In particular, this paper
will describe the data used to fill the tables and fields listed in Table 1-1.
-------
Table 1-1. MOVES Database Elements Covered in This Report
Database Table Name
AvgSpeedDistribution
DayVMTFraction
DriveSchedule
DriveScheduleAssociation
DriveScheduleSecond
FuelEngFraction
HourVMTFraction
HPMSVtypeYear
MonthGroupHour
MonthVMTFraction
PollutantProcessModelYear
RegClassFraction
RoadTypeDistribution
Sample VehicleDay
Sample VehiclePopulation
Fields
avgSpeedFraction
dayVMTFraction
average Speed
sourceTypelD
roadTypelD
drive SchedulelD
isRamp
speed
fuelEngFraction* *
hourVMTFraction
HPMSBaseYearVMT
baseYearOffNetVMT
VMTGrowthFactor
AC Activity Terms (A,
B&C)
monthVMTFraction
modelYearGroupID
regClassFraction* *
roadTypeVMTFraction
daylD
sourceTypelD
stmyFuelEngFraction
stmyFraction
Content*
Distribution of time among average speed
bins
Distribution of VMT between weekdays
and weekend days
Average speed of each drive schedule
Mapping of which drive schedules are
used for each combination of sourcetype
and roadtype.
Speed for each second of each drive
schedule.
Joint distribution of vehicles with a given
fuel type and engine technology. Sums to
one for each sourcetype & model year
Distribution of VMT among hours of the
day
Base Year VMT by HPMS vehicle types
and annual VMT growth factors.
Coefficients to calculate air conditioning
demand as a function of heat index.
Distribution of annual VMT among
months.
Assigns model years to appropriate model
year groups. These vary with
pollutant/process.
Fraction of vehicles in a given
"Regulatory Class." Sums to one for each
sourceType, modelYear and fuel/engtech
combination.
Distribution of VMT among roadtypes
Identifies vehicles in Sample VehicleTrip
Incorporates the fractions found in the
FuelEngFraction, RegClassFraction,
SizeWeightFraction and
SCCVTypeDistribution tables, but also
expected fractions for vehicles that do not
exist in the existing fleet. The expected
values are used with the Alternative
Vehicle Fuel & Technology Strategy
inputs to generate alternate future vehicle
fleet source bins.
Report Sections
Section 10
Section 12
Section 1 1
Section 1 1
Section 1 1
Section 4
Section 12
Section 8
Section 15
Section 12
Section 4
Section 4
Section 9
Section 13
Section 4
10
-------
Database Table Name
Sample VehicleTrip
SCCVTypeDistribution
SizeWeightFraction
SourceBinDistribution
SourceTypeAge
SourceTypeAgeDistribution
SourceTypeHour
SourceTypeModelYear
SourceTypePolProcess
SourceTypeYear
SourceUseType
Zone
ZoneRoadType
Fields
priorTripID
keyontime
keyOffTime
SCCVTypeFraction
SizeWeightFraction* *
sourceBinActivityFracti
on
survivalRate
relativeMAR
functioningACFraction
ageFraction
idleSHOFactor
ACPenetrationFraction
isSizeWeightReqd
isRegClassReqd
isMYGroupReqd
sourceTypePopulation
salesGrowthFactor
migrationRate
rollingTerm
rotatingTerm
dragTerm
sourceMass
idleAllocFactor
startAllocFactor
SHPAllocFactor
SHOAllocFactor
Content*
Trip start and end times; used to determine
vehicle start and soak times.
Distribution of sourcetypes to EPA Source
Classification Codes
Joint distribution of engine size and
weight. Sums to one for each sourceType,
modelYear and fuel/engtech combination.
Distribution of population among different
vehicle sub-types (sourcebins)
Rate of survival to subsequent age;
relative mileage accumulation rates and
fraction of air conditioning equipment that
is functioning
Fraction of vehicle population at each age.
Ratio of extended idle time to driving
time, by hour.
Prevalence of air conditioning equipment
Indicates which pollutant-processes the
source bin distributions may be applied to
and indicates which discriminators are
relevant for each source type and
pollutant/process.
Vehicle counts and growth factors
Road load coefficients for each
SourceType, used to calculate Vehicle
Specific Power.
Allocation of activity to zone (county).
Allocation of driving time to zone
(county) and roadtype.
Report Sections
Section 13
Section 4
Section 4
Section 4
Section 6 &
Section 15
Section 5
Section 14
Section 15
Section 4
Section 3
Section 7
Section 14
Section 14
* These summary descriptions are not intended to fully describe the input for each field. See the associated section
for a full description.
** These tables are used outside MOVES to generate the SampleVehiclePopulation table, but they are not used by
the MOVES2010 model and are included in the default database only for reference.
1.2. MOVES SourceTypes
The primary vehicle classification in MOVES is "SourceType." (Also sometimes called
"SourceUseType"). This name was selected because when MOVES eventually incorporates
nonroad equipment, the sourcetypes will include many emission sources that are not vehicles.
11
-------
sourcetypes are derived from DOT's HPMS vehicle classes and are intended to be groups of
vehicles with similar activity patterns. The MOVES2010 sourcetypes are listed in Table 1-2,
along with the associated DOT Highway Performance Monitoring System (HPMS) vehicle
classes.
Table 1-2. MOVES2010 SourceTypes
Source Type ID
11
21
31
32
41
42
43
51
52
53
54
61
62
SourceType
Motorcycles
Passenger Cars
Passenger Trucks (primarily personal use)
Light Commercial Trucks (other use)
Intercity Buses (non-school, non-transit)
Transit Buses
School Buses
Refuse Trucks
Single Unit Short-haul Trucks
Single Unit Long-haul Trucks
Motor Homes
Combination Short-haul Trucks
Combination Long-haul Trucks
HPMS Vehicle Class
Motorcycles
Passenger Cars
Other Two-Axle/Four Tire, Single Unit
Other Two-Axle/Four Tire, Single Unit
Buses
Buses
Buses
Single Unit
Single Unit
Single Unit
Single Unit
Combination
Combination
In MOVES, "long-haul" trucks are defined as trucks for which most trips are 200 miles
or more.
12
-------
2. Data Sources
A number of organizations collect data relevant to this report. The most important
sources used to populate the vehicle population and activity portions of MOVES database are
described here. These sources are referred to throughout this document by the abbreviated name
given in this description, but the reference citation is only given here.
2.1. VIUS(and TIUS)
Until 2002, the U.S. Census Bureau conducted the Vehicle Inventory and Use Survey
(VIUS)4 to collect data on the physical characteristics and activity of U.S. trucks every five
years. The survey is a sample of private and commercial trucks that were registered in the U.S.
as of July of the survey year. The survey excludes automobiles, motorcycles, government-owed
vehicles, ambulances, buses, motor homes and nonroad equipment.
For MOVES, VIUS provides information to characterize trucks by sourcetype and to
estimate age distributions, diesel fractions, and regulatory class distributions. MOVES2010 uses
data from both the 1997 and 20025 surveys. Before 1997, VIUS was known as TIUS (Truck
Inventory and Use Survey). To populate the 1990 base year, we used data from the 1992 TIUS.6.
While the survey includes a large number of vehicles and was designed to be representative of
the U.S. fleet, information on model year is not available for many of the older trucks. Thus, the
distribution data for many older model years is sparse and sometimes erratic.
Note that Census Bureau has discontinued the VIUS survey. We are investigating
alternate data sources and approaches for determining truck populations in the future.
2.2. Polk NVPP® and TIP®
R.L. Polk & Co. is a private company providing automotive information services. The
company maintains two databases relevant for MOVES: the National Vehicle Population Profile
(NVPP®)7 and the Trucking Industry Profile (TIP®Net) Vehicles in Operation database.8 The
first focuses on light-duty cars and trucks, the second focuses on medium and heavy-duty trucks.
Both compile data from state vehicle registration lists.
For MOVES2010, EPA used the 1999 NVPP® and TIP®._Polk data was used in
determining vehicle populations, diesel fractions, engine size fractions and vehicle weight class
fractions.
2.3. FHWA Highway Statistics
Each year the U.S. DOT Federal Highway Administration's (FHWA) Office of Highway
Policy Information publishes Highway Statistics. This volume summarizes a vast amount of
roadway and vehicle data from the states and other sources.
For MOVES210, we used data on vehicle populations (registrations) and vehicle miles
traveled (VMT), as summarized in four tables.910 u 12 Hereafter, references will be to FHWA
MV-1, MV-10, VM-1, and VM-2. For the 1999 base year, we used the 1999 statistics; for the
1990 base year, we used 1990 numbers.
2.4. FTA National Transit Database
The U.S. DOT, Federal Transit Administration (FTA) summarizes financial and
operating data from U.S. mass transit agencies in the National Transit Database (NTD).13
13
-------
For MOVES2010, we used 1999 data from the report, "Age Distribution of Active
Revenue Vehicle Inventory: Details by Transit Agency," to determine age distributions and
diesel fractions for transit buses.
2.5. School Bus Fleet Fact Book
The School Bus Fleet 1999 Fact Book includes estimates, by state, of number of school
buses and total miles traveled.14 The Fact Book is published by Bobit Publications.
Information from the 1999 and 1990 School Bus Fleet Fact Book was used in estimating
school bus vehicle populations. School bus mileage accumulation rates came from the 1997 Fact
Book by way of MOBILE6.
2.6. MOBILE6
MOBILE6 was a precursor to MOVES used to estimate highway vehicle emissions. In
some cases, we have used data from MOBILE6 modelwith only minor adaptation. In particular,
we used MOBILE6 data for mileage accumulations, air conditioning rates, school bus diesel
fractions, urban speed distributions, and many driving schedules.
The MOBILE6 data is documented in technical reports, particularly M6.FLT.002
"Update of Fleet Characterization Data for Use in MOBILE6 - Final Report."15 Additional
MOBILE6 documentation is available on the web at http://www.epa.gov/otaq/m6.htm
2.7. Annual Energy Outlook & National Energy Modeling System
The Annual Energy Outlook (AEO) 16'17 describes Department of Energy forecasts for
future energy consumption. The National Energy Modeling System (NEMS) is used to generate
these projections based on economic and demographic projections. For MOVES2010 we used
AEO2006 to forecast VMT growth and vehicle sales growth for most vehicles, but updated the
passenger car, passenger truck and light commercial truck with information from AEO2009. For
MOVES2010a we used AEO2010 to update VMT and sales growth estimates for heavy and
medium duty trucks.
2.8. Transportation Energy Data Book
Each year, Oak Ridge National Laboratory produces the DOE Transportation Energy
Data Book (TEDB). This book summarizes transportation and energy data from a variety of
sources.
For MOVES we used TEDB information in estimating vehicle population, sales, and
survival fractions. Beginning with MOVES2004, we relied on Edition 22, published in
September 200218 and Edition 23, published in October 2003.19 We later updated 1990 values
using Edition 13, published in 1993. For MOVES2010 we updated sales growth based on
Edition 27, published in 2008.20 MOVES2010a includes heavy- and medium-duty sales growth
updates from Edition 28, published in 2009.21
2.9. Oak Ridge National Laboratory Light-duty Vehicle Database
Oak Ridge National Laboratory Center for Transportation Analysis has compiled a
database of light-duty vehicle information which combines EPA Test vehicle data and Ward's
Automotive Inc. data spanning 1976 - 2001.22 We used this database to determine weight
distributions for light trucks by model year.
14
-------
3. Vehicle Population Data by Calendar Year
The SourceTypeYear table stores three data fields—SourceTypePopulation,
SalesGrowthFactor, and Migration Rate. SourceTypePopulation stores the total population of
vehicles by sourcetype for MOVES base years. SalesGrowthFactor field stores a multiplicative
factor indicating the change in sales by sourcetype for calendar years after the base year.
Migration Rate is not used in MOVES2010. Each field is described below in terms of what
information it contains, the sources of the data used for the field, and, when applicable, certain
data points used in determining the field parameters.
3.1. 1999 SourceTypePopulation
In the MOVES default database, the SourceTypePopulation field stores the total
population of vehicles in entire United States in 1990 and 1999 by sourcetype. Some of the
values are taken directly from the indicated sources; other values needed to be derived from the
available data.
SourceTypePopulation values are used for base years and provide the basis for Total
Activity Generator calculation of populations in calendar years after the base year. These
populations are, in turn, used to generate travel fractions by age and sourcetype and to allow
allocation of VMT by age.
The primary sources for calendar year 1999 vehicle population data are the FHWA
Highway Statistics Tables MV-1 and MV-10 and the Polk NVPP® and TIP® databases. The
Transportation Energy Data Book (TEDB) explains three factors that account for differences
between the two sources:
1. Polk data includes only vehicles that were registered on July 1 of 1999. FHWA data
includes all vehicles that have been registered at any time throughout the year and
thus may include vehicles that were retired during the year or may double count
vehicles registered in two or more states.
2. Polk and FHWA may differ in how they classify some minivans and SUVs as trucks
or automobiles. (This difference is less important since 1990).
3. FHWA includes all non-military Federal vehicles. Polk data includes only those
Federal vehicles that are registered within a state.
Also, FHWA data is available for Puerto Rico, but Puerto Rico does not appear to be
included in our Polk data set. MOVES will cover Puerto Rico and the Virgin Islands. In
addition, Polk collects data on Gross Vehicle Weight (GVW) class 3 vehicles in both the
NVPP® and TIP® databases, but the values are not the same. Polk staff recommended using the
01
TIP® values. Finally, our 1999 Polk data set includes school buses and motor homes (which
can be counted separately), but does not include "non-school buses."
Motorcycle population estimates were available from both FHWA registration data and
from the Motorcycle Industry Council. The MIC estimate is based on 1998 sales estimates,
adjusted to subtract noped sales (nopeds are similar to mopeds, but lack pedals) and to account
for scrappage.
The Department of Transportation's National Household Transportation Survey (NHTS)
combines the previous National Personal Transportation Survey and the American Travel Survey
15
-------
to collect data on personal travel patterns and includes data on motorcycles, personal trucks and
automobiles.24 Data from the 2001 survey is included in Table 3-1, but was not used in MOVES
because it is two years newer than the base year, and it excludes non-household vehicles. Values
from the five data sources are compared in Table 3-1.
Table 3-1. Vehicle Population Comparisons 1999
Data Source
FHWAMV-l(w Puerto
Rico and publically
ownedvehicles)
FHWAMV-l(w/o
Puerto Rico and
publically owned
vehicles)
Polk NVPP®& TIP®
NHTS (2001)
MIC (1998)25
Motorcycles
4,173,869
4,951,747
4,605,439
Automobiles
134,480,432
131,076,551
126,868,744
115,914,908
Trucks (total)
83,178,092
81,060,369a
80,323,528*
80,499,939
Buses (total)
732,189
Motor
Homes
902,949
1,446,469
* Excluding motor homes and NVPP® GVW3 trucks.
For automobiles and trucks, it is possible to do a direct comparison of Polk and FHWA
data. To estimate the MOVES population, we adjust the FFIWA data to account for double-
counting by multiplying the total FFIWA population by the ratio of the Polk population to the
FFIWA population without public vehicles and Puerto Rican vehicles.
Adjusted Population = FHWA w public & PR * (Polk/FHWA w/0 public & PR)
This leads to the values in Table 3-2.
Table 3-2. Adjusted Vehicle Populations
Automobiles
Trucks (total)
Population
130,163,354
83,007,993
For MOVES, total trucks are sub-classified into seven sourcetypes. The proportion of
total trucks in each sourceType was estimated using VIUS responses for Axle Arrangement,
Primary Area of Operation, Body Type and Major Use as detailed in Table 3-3 and Table 3-4.
With these definitions and with vehicles that lack AREAOP codes assigned
proportionally to the corresponding sourcetypes, we computed the distributions in Table 3-5.
aln our peer review, we learned that this number was recorded incorrectly. The correct number is 81,090,659, but
this was not remedied prior to MOVES release. This causes a discrepancy of less than 0.04% in the national total
truck population and will have no impact on runs using local population inputs.
16
-------
These distributions were multiplied by the total truck population from Table 3-2 to derive
population values for MOVES.
Table 3-3. VIUS 1997 Codes Used for Distinguishing Truck SourceTypes.
SourceType
Passenger Trucks
Light Commercial
Trucks
Refuse Trucks
Single Unit Short-
haul Trucks
Single Unit Long-
haul Trucks
Combination Short-
haul Trucks
Combination
Long-haul
Trucks
Axle Arrangement
2 axle/4 tire (AXLRE=
1,5,6,7)
2 axle/4 tire (AXLRE=
1,5,6,7)
Single Unit (AXLRE =
2-4, 8-16)
Single Unit (AXLRE =
2-4, 8-16)
Single Unit (AXLRE =
2-4, 8-16)
Combination (AXLRE
>=17)
Combination (AXLRE
>=17)
Primary Area of
Operation
any
any
off-road, local or short-
range (AREAOP <=4)
off-road, local or short-
range (AREAOP <=4)
long-range (AREAOP
>=5)
off-road, local or
medium (AREAOP <=4)
long-range (AREAOP
>=5)
Body Type
any
any
garbage hauler
(BODTYP=30)
any except
garbage hauler
any
any
any
Major Use
personal
transportation
(MAJUSE=20)
any but personal
transportation
Any
Any
Any
Any
Any
Table 3-4. VIUS 2002 Codes Used for Distinguishing Truck SourceTypes.
SourceType
Passenger Trucks
Light Commerical
Trucks
Refuse Trucks
Single Unit Short-
Haul Trucks
Single Unit Long-
Haul Trucks
Combination Short-
Haul Trucks
Combination
Long_Haul Trucks
Axle Arrangement
axle config in
(1,6,7,8)
axle config in
(1,6,7,8)
axle config in
(2,3,4,5,9,10,11,12,13,
14,15,16,17,18,19,20)
axle config in
(2,3,4,5,9,10,11,12,13,
14,15,16,17,18,19,20)
axle config in
(2,3,4,5,9,10,11,12,13,
14,15,16,17,18,19,20)
axle config>=21
axle_config>=21
Primary Area of
Operation
any
any
trip_primary in (1,2,3,4)
trip_primary in (1,2,3,4)
trip_primary in (5,6)
Long Range
trip_primary in (1,2,3,4)
trip_primary in (5,6)
Long Range
Body Type
any
any
bodytype=21
Any except
bodytype=21
any
sample_strata=5
Combination
Trucks
sample_strata=5
Combination
Trucks
Operator
Classifica
tion
opclass=5
opclass<>5
any
any
any
any
any
17
-------
Table 3-5. 1999 Truck SourceType Distribution and Populations
SourceType
Passenger Trucks
Light Commercial Trucks
Refuse Trucks
Single Unit Short-haul Trucks
Single Unit Long-haul Trucks
Combination Short-haul
Trucks
Combination Long-haul Trucks
Total
Percent
68.90%
23.02%
0.11%
5.39%
0.32%
1.31%
.97%
100.00%
Population
57,190,192
19,106,257
88,607
4,470,798
264,435
1,084,366
803,337
83,007,993
For buses, we needed to distribute the total buses from FHWA to the three MOVES
classes. Additional information on bus numbers was available from the FTA NTD, Polk TIP®,
and the School Bus Fleet Fact Book, the American Public Transit Association, and the American
Bus Association "Motorcoach Census 2000".26 The FTA NTD provides population numbers for
a variety of transit options. To determine the number of transit buses, we summed their counts
for Articulated Motor Buses, Motor Bus Class A, B & C, and Double Decked buses.
Table 3-6. 1999 Bus Population Comparisons
Data Source
FHWAMV-1
FHWA MV- 10
(excludes PR)
FHWA adjusted for
PRb
FTA NTD
APTA27***
Polk TIP®
School Bus Fleet Fact
Book
Motorcoach
Census**
Total Buses
732,189
728,777
Intercity Buses
44,200
Transit Buses
55,706
75,087
School Buses
592,029*
594,800
460,178
429,086
* Includes some church & industrial buses.
** Includes Canada.
*** Includes trolleybuses.
As Table 3-6 shows, estimates of bus populations vary. We chose to use the FFTWA
value because it includes church and industrial buses that we believe have activity patterns more
similar to school buses than to intercity buses. To calculate the number of buses for the
categories needed for MOVES, we used the FHWA school bus value and the FTA transit bus
b Peer review suggested adjusting the MV-10 values to account for Puerto Rico. This approach should be
considered for future databases. In 1999, this change would slightly increase the number of school buses and, more
importantly, would decrease the number of intercity buses to 81, 683, a change of about three percent.
18
-------
value. We assigned the remaining total FHWA buses (732,189-592,029-55,706 = 84,454) to the
intercity category.
For motorcycles we used the 1999 FHWA value from table MV-1. For comparison,
Table 3-1 also shows the 1998 population as estimated by the Motorcycle Industry Council based
on sales and estimated scrappage rates, and the 2001 population estimated by the 2001 NHTS.
The FHWA population estimates are noticeably lower than the other estimates.
For motor homes we used the population from the Polk TIP® database. In Table 3-1, this
value is compared to the estimate from the 2001 NHTS. As for motorcycles, the FHWA
registration count is noticeably lower than the household survey estimate. This could reflect
population growth in the years between the estimates, or it may reflect difference in the way
motor homes are defined in the two studies, or be an artifact of the method used to extrapolate
from the NHTS sample to the national population estimate. If time and resources allow, EPA
may investigate this further for future versions of the MOVES model.
Table 3-7 summarizes the 1999 vehicle populations used in MOVES2010.
Table 3-7. '
1999 Sou rceTy
SourceType ID
11
21
31
32
41
42
43
51
52
53
54
61
62
pe Populations in MOVES2010
SourceType
Motorcycles
Passenger Cars
Passenger Trucks
Light Commercial Trucks
Intercity Buses
Transit Buses
School Buses
Refuse Trucks
Single Unit Short-haul Trucks
Single Unit Long-haul Trucks
Motor Homes
Combination Short-haul Trucks
Combination Long-haul Trucks
1999 Population
4,173,870
130,163,000
57,190,200
19,106,300
84,454
55,706
592,029
88,607
4,470,800
264,435
902,949
1,084,370
803,337
3.2. 1990 SourceTypePopulation
Because some State Implmentation Plans require estimates of 1990 emissions, the
MOVES database includes a 1990 base year. The SourceTypePopulation inputs for 1990 were
developed using methods and data similar to those used for 1999.
The primary sources for calendar year 1990 vehicle population data are the FHWA
Highway Statistics Tables MV-200, VM- 201 A, MV-10 and the Polk NVPP® databases. As
in 1999, the FHWA and Polk data differ in how vehicles are counted. (See previous section.)
Additionally, the 1990 Polk data does not include buses and motor homes. The National
Personal Transportation Survey includes data on personal trucks, automobiles and motorcycles.
-------
Data on motorcycles were also obtained from the Motorcycle Statistical Annual published by the
Motorcycle Industry Council. Values from all four sources are compared in Table 3-1.
Registration data on vehicles registered in Puerto Rico for year 1990 was obtained from
FHWA's Highway Statistics 1990.
Table 3-8. 1990 Vehicle Population Comparisons
Data Source
FHWA(w/ Puerto
Rico and Publicly
owned vehicles)1
FHWA (w/o Puerto
Rico and w/ Publicly
owned vehicles)2
Polk NVPP®
NPTS (1990)4
Motorcycle Industry
Council5
Motorcycles
4,278,286
4,259,461
na
2,089,523
4,310,000
Automobiles
135,022,124
133,700,497
123,276,600
120,712,000
na
Trucks (total)
54,673,458
54,470,430
56,023,0003
37,110,000
na
Buses (total)
629,943
626,987
na
na
na
Motor
Homes
na
na
na
821,000
na
1 Data on Puerto Rico was obtained from Highway Statistics 1990, published by the FHWA.
2 For these numbers, we used data from FHWA Highway Statistics TableVM-201A, April 1997 and Table MV-
200 (state motor vehicle registrations, by years 1990-1995).
3 As published in TEDB edition 23. Does not include Puerto Rico and publicly -owned vehicles.
41990 NPTS special report on travel modes- Chapter 3, the Demography of the US Vehicle Fleet. The motorcycle
number is calculated using the appendix table and the proportion of MCs from Table 20 of the 2001 NHTS
Summary of Travel Trends.
5 The Motorcycle number was obtained as a sum of on-highway and dual motorcycles for year
1990 as published in the 1999 Motorcycle Statistical Annual.
For MOVES, total trucks are sub-classified into seven sourcetypes. The proportion of
total trucks in each subtype was estimated using TIUS92 responses for Axle Arrangement,
Primary Area of Operation, Body Type and Major Use as detailed in Table 3-9.
With these definitions and with vehicles that lack AREAOP codes assigned
proportionally to the corresponding sourcetypes, we computed the distributions in Table 3-10.
These distributions were multiplied by the Polk total truck population in Table 3-8 to derive
population values for MOVES.
20
-------
Table 3-9. TIUS92 Codes Used for Distinguishing Truck SourceTypes.
SourceType
Passenger Trucks
Light Commercial
Trucks
Refuse Trucks
Single Unit Short-
haul Trucks
Single Unit Long-
haul Trucks
Combination Short-
haul Trucks
Combination Long-
haul Trucks
Axle Arrangement
2 axle/4 tire (AXLRE=
1,5,6,7)
2 axle/4 tire (AXLRE=
1,5,6,7)
Single Unit (AXLRE =
2-4, 8-16)
Single Unit (AXLRE =
2-4, 8-16)
Single Unit (AXLRE =
2-4, 8-16)
Combination (AXLRE
>=17)
Combination (AXLRE
>=17)
Primary Area of
Operation
any
any
off-road, local or short-
range (AREAOP <=4)
off-road, local or short-
range (AREAOP <=4)
long-range (AREAOP
>=5)
off-road, local or
medium (AREAOP <=4)
long-range (AREAOP
>=5)
Body Type
Any
Any
garbage hauler
(BODTYP=30)
any except
garbage hauler
any
Any
Any
Major Use
personal
transportation
(MAJUSE=20)
any but personal
transportation
any
any
any
any
any
Table 3-10. 1990 Truck SourceType Distribution and Populations
SourceType
Passenger Trucks
Light Commercial Trucks
Refuse Trucks
Single Unit Short-haul Trucks
Single Unit Long-haul Trucks
Combination Short-haul Trucks
Combination Long-haul Trucks
Total
Percent
67.32%
24.07%
0.11%
6.12%
0.23%
1.35%
0.81%
100.00%
Population
37,713,840
13,483,198
59,037
3,426,459
128,776
758,091
453,599
56,023,000
For buses, we needed to distribute the total buses from FHWA to the three MOVES
classes. Additional information on bus numbers was available from the American Public Transit
Association (APTA) Fact Book, the School Bus Fleet Fact Book, and the Transportation Energy
Data Book.
21
-------
Table 3-11. 1990 Bus Population Comparisons
Data Source
FHWA
(w/o PR and with
Publicly-owned
Vehicles)*
FHWA (w/o PR
and w/o Publicly-
owned Vehicles)
APTA Historical
Tables28
TEDB**
School Bus Fleet
Fact Book* * *
Total Buses
626,9871
275,4931
Intercity Buses
20,6802
58,141
Transit Buses
58,714
59,753
School Buses
545,7223
508,261
391,714
FHWA Highway Statistics, Summary to 1995, Table MV-200
** Transportation Energy Data Book : Edition 13, March 1993, Table 3.29. 1990 buses. "Intercity Buses" is sum
of "Intercity Bus" and "Other;" "School Buses" includes other non-revenue buses.
*** Average of school years 1989-90 and 1990 -91, School Bus Fleet Fact Books 1990 and 1991.
Table 3-12 summarizes the 1990 vehicle populations used in MOVES2010. For motor
homes we used the only available data from NPTS. We used the TEDB data for buses. For
trucks the TIUS data was used; the remaining values were based on FHWA data.
Table 3-12. 1990 SourceType Populations in MOVES2010
SourceType ID
11
21
31
32
41
42
43
51
52
53
54
61
62
SourceType
Motorcycles
Passenger Cars
Passenger Trucks
Light Commercial Trucks
Intercity Buses
Transit Buses
School Buses
Refuse Trucks
Single Unit Short-haul Trucks
Single Unit Long-haul Trucks
Motor Homes
Combination Short-haul Trucks
Combination Long-haul Trucks
1990 Population
4,278,286
135,022,124
37,713,840
13,483,198
58,141
59,753
508,261
59,037
3,426,460
128,776
821,000
758,091
453,599
22
-------
3.3. SalesGrowthFactor
The SalesGrowthFactor field stores a multiplicative factor indicating changes in sales by
sourcetype for calendar years after the base year. It determines the number of new vehicles
added to the vehicle population each year, and is expressed relative to the previous year's sales.
For example, " 1" means no change from previous year sales levels, " 1.02" means a two percent
increase in sales, and "0.98" means a two percent decrease in sales. SalesGrowthFactor is used
in the Total Activity Generator calculation of source type populations for calendar years after the
base year- in MOVES2010, years 2000 through 2050.
Note that the sales growth factors are not used in the calculation of county-level or
project level emissions, where users must input local vehicle populations for each year that is
modeled.
The data sources and methodologies by source type are described below:
• Passenger Cars, Passenger Trucks and Light Duty Commercial Trucks:
SalesGrowthFactors for calendar year 2000 through 2007 were derived from total
sales numbers reported in the TEDB28 Table 4.5 and 4.6. Factors for calendar years
2008 through 2030 were derived from new vehicle sales estimates presented in
AEO2009Supplemental Table 57, generated by NEMSC. A constant annual growth
rate of 0.76% was used for years 2031-2050. 0.76% is consistent with the value used
in Draft MOVES2009 for passenger cars, and significantly lower than the value used
for trucks in Draft MOVES2009. This decrease in future sales is consistent with the
overall decrease in truck sales predicted in AEO2009. Note that the growth factor in
each year is relative to the preceding year's sales. 1999 sales are calculated from the
1999 population after applying the Age 0 age fraction and survival fraction. They are
slightly different that the sales numbers shown in the TEDB. Also, with no data to
distinguish sales of passenger trucks and light duty commercial trucks, EPA assumed
the same sales growth rates for both.
• Motorcycles: SalesGrowthF actors for calendar year 2000 and 2008 were computed
from information from the Motorcycle Industry Council and from Polk registration
data from 2008. More details are available in a contractor report on the analysis.29 In
MOVES2010, SalesGrowthF actors for years 2008 through 2050 were inadvertently
unchanged from Draft MOVES2009, where they had been set equal to passenger car
growth factors (ie, based on AEO2006). In MOVES2010a, the sales growth was
updated for years 2009 and later as indicated in the contractor report.
• Buses, Single Unit Trucks & Motor Homes: For MOVES2010 these estimates were
unchanged from previous versions of the model. Calendar years 2000-2001 were
based on sales as reported in TEDB23 Table 5.3 (gross weight range 10,000-33,000
Ibs). Years 2004 through 2030 were calculated from medium-duty truck sales
projections fromy4JE'02006Supplernental Table 55. For MOVES2010a, calendar
0 AEO2009 predates the Model Year 2011 CAFE rule, and the related redefinitions of cars and trucks, so it is
consistent with the definitions used in MOVES2010.
-------
years 2000-2007 were based on sales reported in TEDB28 and calendar years 2008
and later were based on values in AEO2010 Table 67.
• Combination Trucks, Refuse Trucks: For MOVES2010,these estimates were
unchanged from previous versions of the model. Calendar years 2000-2001 were
based on sales as reported in TEDB23 Table 5.3 (gross weight range 33,001 and
greater pounds). Years 2004 through 2030 were calculated from heavy-duty truck
sales projections found in AEO2006 Supplemental Table 55. For MOVES2010a,
calendar years 2000-2007 were based on sales reported in TEDB28 and calendar
years 2008 and later were based on values in AEO2010 Table 67.
The resulting SalesGrowthFactors by source type are shown in Table 3-13 and Table 3-
14.
24
-------
Table 3-13. MOVES2010 SalesGrowthFactor by Calendar Year and Source Type
Calendar
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031+
Motorcycles
1.244
1.178
1.109
1.055
1.111
1.108
1.073
0.992
1.040
0.985
0.980
1.005
0.996
0.991
0.989
0.994
1.001
1.002
1.005
1.004
1.007
1.007
1.009
1.009
1.009
1.008
1.010
1.008
1.007
1.008
1.008
1.008
Passenger
Cars
1.049
0.952
0.962
0.939
0.991
1.023
1.013
0.974
0.895
0.892
1.192
1.137
1.091
1.065
1.055
1.101
1.000
0.995
1.026
1.024
1.021
1.000
1.004
1.014
1.018
1.020
1.015
1.007
1.008
1.011
1.012
1.008
Passenger
and Light
Commercial
Trucks
1.087
1.037
1.001
1.026
1.047
0.991
0.936
0.975
0.671
0.792
1.419
1.089
1.025
1.037
1.007
0.986
0.970
0.979
0.978
0.983
0.980
0.986
0.990
1.005
1.013
1.019
1.008
1.005
0.998
0.992
1.015
1.008
Buses,
Single
Unit
Trucks &
Motor
Homes
0.963
0.850
0.882
1.067
1.170
1.082
1.001
1.001
1.003
1.026
0.992
0.997
0.986
1.000
1.029
1.035
1.025
1.015
1.010
0.995
0.997
1.006
1.012
1.015
1.018
1.018
1.016
1.012
1.006
1.010
1.013
1.013
Combination
Trucks
0.809
0.660
0.923
1.042
1.310
1.130
1.010
0.940
0.990
1.000
1.000
0.990
1.000
1.010
1.020
1.020
1.020
1.020
1.000
0.980
0.980
1.000
1.010
1.010
1.020
1.020
1.020
1.010
1.000
1.010
1.010
1.010
25
-------
Te
ible 3-14. MOVES2010a Sal esGrowth Factor by Calendar Year and Source T}
Calendar
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035+
Motorcycles
1.244
1.178
1.109
1.055
1.111
1.108
1.073
0.992
1.040
0.500
1.340
1.240
1.200
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
Passenger
Cars
1.049
0.952
0.962
0.939
0.991
1.023
1.013
0.974
0.895
0.892
1.192
1.137
1.091
1.065
1.055
1.101
1.000
0.995
1.026
1.024
1.021
1.000
1.004
1.014
1.018
1.020
1.015
1.007
1.008
1.011
1.012
1.008
1.008
1.008
1.008
1.008
Passenger
and Light
Commercial
Trucks
1.087
1.037
1.001
1.026
1.047
0.991
0.936
0.975
0.671
0.792
1.419
1.089
1.025
1.037
1.007
0.986
0.970
0.979
0.978
0.983
0.980
0.986
0.990
1.005
1.013
1.019
1.008
1.005
0.998
0.992
1.015
1.008
1.008
1.008
1.008
1.008
Buses,
Single
Unit
Trucks &
Motor
Homes
0.980
0.902
0.850
1.128
1.232
1.238
0.991
0.991
0.967
0.908
1.041
1.086
1.071
1.048
1.036
1.036
1.036
1.034
1.032
1.030
1.029
1.024
1.022
1.026
1.028
1.028
1.028
1.028
1.027
1.028
1.028
1.027
1.027
1.028
1.028
1.028
Combination
Trucks
0.980
0.902
0.850
1.128
1.232
1.238
0.991
0.991
0.967
0.908
1.041
1.086
1.071
1.048
1.036
1.036
1.036
1.034
1.032
1.030
1.029
1.024
1.022
1.026
1.028
1.028
1.028
1.028
1.027
1.028
1.028
1.027
1.027
1.028
1.028
1.028
/pe
26
-------
3.4. MigrationRate
The MigrationRate field stores a yearly multiplicative factor that could be used to
estimate how many vehicles join or leave the population of a sourcetype in the given domain in a
given year. When MOVES was initially designed, we expected this field would be useful when
modeling emissions on relatively small geographic scale where vehicle populations might
change due to factors other than sales and scrappage. This field is currently not used and is
populated with a migration rate of 1, indicating no migration of vehicles.d
4. Emission-Related Vehicle Characteristics (Source Bins)
The sourcetypes in MOVES are defined based on large scale, easily observable characteristics
such as number of axles, and activity characteristics (such as long-haul vs. short haul). But to
estimate emissions, MOVES must also know the emission-related characteristics of the vehicle
such as the type of fuel that it uses and the emission standards it is subject to. Thus, in MOVES,
we group vehicles into SourceBins that classify a vehicle by discriminators relevant for
emissions and energy calculations: fuel and engine technology, average vehicle weight and
engine displacement, model year group, and regulatory class.
SourceBin information in MOVES is stored in the SampleVehiclePopulation Table. This table
also stores information to link each sourcebin to a Source Classification Code (SCC) for use if a
user requests output by SCC.
The MOVES Source Bin Generator code determines which discriminators are relevant
for a given pollutant/process combination and multiplies the relevant fractions from the tables
listed above to determine the detailed SourceBinDistribution for each combination of Pollutant,
Process, sourcetype, and Model Year. In general, fueltype and model year group are relevant for
all emission calculations. Regulatory class is relevant for most pollutants and processes, except
some energy calculations in MOVES2010, which rely on engine size and vehicle weight.
MOVES2010a does not use engine size and vehicle weight classifications in any calculation.
For some uses, particularly the preparation of national inventories, modelers will need to
produce output aggregated by EPA's Source Classification Codes (SCC). The EPA's highway
vehicle SCC were derived from MOBILES and MOBILE6 and do not directly correspond to the
MOVES sourcetypes. For example, depending on its fuel and Gross Vehicle Weight (GVW)
limits, a vehicle in the MOVES Passenger Truck category may be coded with one of eight
SCCs—including the SCC for a Light-Duty Gasoline Truck 1, a Light-Duty Gasoline Truck 2, a
Heavy-Duty Gasoline Truck, a Light-Duty Diesel Truck, or one of the four codes for Heavy-
Duty Diesel Vehicle.
d For motorcycles, the migration rate was mistakenly populated for some years in MOVES2010. We fixed this in
MOVES2010a.
27
-------
Table 4-1. Data Tables Used to Allocate SourceType to SourceBin
Generator Table Name
SourceTypePolProcess
FuelEngFraction*
SizeWeightFraction*
RegClassFraction*
PollutantProcessModelYear
SCCVtypeDistribution*
Sample VehiclePopulation
Key Fields**
SourceTypelD
PolProcessID
SourceTypelD
ModelYearlD
FuelTypelD
EngTecMD
SourceTypelD
ModelYearlD
FuelTypelD
EngTecMD
WeightClassID
EngSizelD
SourceTypelD
ModelYearlD
FuelTypelD
EngTecMD
RegClassID
PolProcessID
ModelYearlD
SourceTypelD
ModelYearlD
FuelTypelD
SCCVtypelD
SourceTypelD
ModelYearlD
FuelTypelD
EngTecHD
RegClassID
WeightClassID
EngSizelD
SCCVTypelD
Additional Fields
isSizeWeightReqd
isRegClassReqd
isMYGroupReqd
fuelEngFraction
SizeWeightFraction
regClassFraction
modelYearGroupID
SCCVtypeFraction
stmyFuelEngFraction
stmyFraction
Notes
Indicates which pollutant-processes the
source bin distributions may be applied
to and indicates which discriminators
are relevant for each sourceType and
polProcess (pollutant/process
combination)
Joint distribution of vehicles with a
given fuel type and engine technology.
Sums to one for each combination of
sourceType & modelYear
Joint distribution of engine size and
weight. Sums to one for each
sourceType, modelYear and fuel/engine
technology combination.
Fraction of vehicles in a given
"Regulatory Class." Sums to one for
each sourceType, modelYear and
fuel/engine technology combination.
Assigns model years to appropriate
model year groups.
Fraction of vehicles in a given SCC
vehicle class. Sums to one for each
sourceType, modelYear and fuelType
combination.
Includes the fractions found in the
FuelEngFraction, RegClassFraction,
SizeWeightFraction and
SCCVTypeDistribution tables, but also
for combinations that do not exist in the
existing fleet. This table is also used
with the Alternative Vehicle Fuel &
Technology Strategy inputs to generate
alternate future vehicle fleet source
bins.
* These tables are used outside the model to generate the Sample VehiclePopulation table, but they are not used by
the MOVES2010 model and are included in the default database only for reference.
** In these tables, the SourceTypelD and ModelYearlD are combined into a single SourceTypeModelYearlD.
The MOVES model is designed to aggregate emissions to the user's choice of sourcetype
or SCC using information from the SCCVTypeDistribution table. For each combination of
sourcetype, Model Year and FuelType, the SCCVTypeDistribution table lists IDs for the
possible SCC and the fraction of vehicles assigned to each SCC. The full SCC also includes a
suffix that indicates roadway type. This is a mapping from the MOVES roadtype on which the
28
-------
emissions occur to the HPMS Facility Type used in the SCC codes. This mapping is captured in
the SCCRoadTypeDistribution table described in Section 9.2.
The fractions described here are intended to represent national averages. Because the
distribution of vehicle characteristics varies geographically, local inputs should be used for local
runs when available.
More detailed descriptions of the SourceBin Distribution and SCC inputs for each
sourcetype follow.
4.1. Motorcycles
Motorcycle characteristics were assigned based on information from EPA motorcycle
experts and from the Motorcycle Industry Council.
4.1.1. FuelEngFraction
We assume all motorcycles are powered by conventional gasoline engines.
4.1.2. SizeWeightFraction
The SizeWeightFraction is used for calculating energy consumption in MOVES2010, but
not in MOVES2010a. The Motorcycle Industry Council "Statistical Annual" provides
information on displacement distributions for highway motorcycles for model years 1990 and
1998. These were mapped to MOVES engine displacement categories. Additional EPA
certification data was used to establish displacement distributions for model year 2000. We
assumed that displacement distributions were the same in 1969 as in 1990, and interpolated
between the established values to determine displacement distributions for all model years from
1990 to 1997 and for 1999. Values for 2000-and-later model years are based on model year
2000 certification data.
We then applied weight distributions for each displacement category as suggested by
EPA motorcycle experts. The average weight estimate includes fuel and rider. The weight
distributions depended on engine displacement but were otherwise independent of model year.
This information is summarized in Table 4-2.
Table 4-2. Motorcycle Engine Size and Average Weight Distributions for
Selected Model Years
Displacement
Category
0-169 cc(l)
170-279 cc (2)
280+ cc (9)
1969 MY
distribution
(assumed)
0.118
0.09
0.792
1990 MY
distribution
(MIC)
0.118
0.09
0.792
1998 MY
distribution
(MIC)
0.042
0.05
0.908
2000 MY
distribution
(certification
data)
0.029
0.043
0.928
Weight distribution
(EPA staff)
100%: <=5001bs
50%: <= 500 Ibs
50%: 5001bs -7001bs
30%: 500 lbs-700 Ibs
70%: > 7001bs
29
-------
4.1.3. RegClassFraction
All Motorcycles were assigned to the "Motorcycle" (MC) regulatory class.
4.1.4. SCCVtypeDistribution
All Motorcycles were assigned to the "Motorcycle" SCC (prefix 2201080).
4.2. Passenger Cars
For base year 1999, passenger car distributions were derived from the 1999 Polk
NVPP®. The national files for domestic and imported cars were consolidated into a single file.
7.2.1. FuelEngFraction
The FuelEngFraction table assigns fractions of each source type and model year to all relevant
combinations of fuel type bin and engine technology bin. For MOVES2010 defaults, the only
engine technology used was "conventional internal combustion." Fuel fractions were computed
from the Polk data car counts and fuel classifications. The fractions were edited to remove the
small fractions of non-diesel, non-gasoline fuels and renormalized. For 2000-and-later a diesel
fraction of 0.38% was used for each model year. This is an average of the diesel fractions
reported in Ward's Automotive Yearbook for model years 1998-200730.
4.2.2. SizeWeightFraction
The SizeWeightFraction is used for calculating energy consumption in MOVES2010, but
not in MOVES2010a. The Polk cubic displacement values were converted to liters and assigned
to the MOVES engine size bins. The weight ID was assigned by adding 300 Ibs to the Polk curb
weight and grouping into MOVES weight bins. For each fuel type, model year, engine size, and
weight bin, the number of cars was summed and fractions were computed. In general, entries for
which data was missing were omitted from the calculations. Also, analysis indicated a likely
error in the Polk data (an entry for 1997 gasoline-powered Bentleys with engine size 5099 and
weight class 20). This fraction was removed and the 1997 values were renormalized. 1999
model year values were used for all 2000-and-later model years.
4.2.3. RegClassFraction
All Passenger Cars were assigned to the "Light-Duty Vehicle" (LDV) regulatory class.
4.2.4. SCCVtypeFraction
All gasoline Passenger Cars were assigned to the Light Duty Gas Vehicle (LDGV) SCC
(prefix 2201001). All other Passenger Cars were assigned to the Light Duty Diesel Vehicle SCC
(prefix 2230001). If the Alternative Vehicle Fuels and Technologies control strategy is used to
assign vehicles to the "Electric" fueltype, those vehicles are mapped to "LDGV" because there is
no SCC for electric vehicles.
-------
4.3. General Trucks
This section describes how default Source Bin information was compiled for Passenger
Trucks, Light Commercial Trucks, Single-Unit Short-haul and Long-haul Trucks, and
Combination Short-haul and Long-haul Trucks. Source Bin information for Buses, Refuse
Trucks, and Motor Homes are described in separate sections following.
The Vehicle Inventory and Use Survey (VIUS) conducted by the Census Bureau was the
primary source for information on truck distributions. Information from the 1997 and 2002
VIUS was supplemented with information from MOBILE6 and from the Oak Ridge National
Laboratory Light Duty Vehicle database.
VIUS records were assigned to sourcetypes as described previously in Table 3-3 and 3-4.
Not all sourcetypes had data for all model years, and no data was available beyond model year
2002. For years where no vehicles or only a few vehicles were surveyed by VIUS, we
duplicated fractions from the nearest available model year. The 2002 VIUS was used for 1986
and later model years and 1997 VIUS information was only used for the older model years not
surveyed in the 2002 VIUS. In the MOVES2010 release, the oldest model year observed diesel
fractions were applied to the older model years for combination trucks only. These older model
years for the other truck categories were assumed to have no diesel trucks.
4.3.1. FuelEngFraction
Most trucks were assigned to conventional internal combustion engines. In
MOVES2010, some passenger trucks and light commercial trucks were assigned to "advanced
internal combustion" in order for the model to correct phase-in improvements in energy
consumption. The model-year engine technology fractions were back-calculated to match mile-
per-gallon values. These fractions do not impact emissions of other pollutants. In
MOVES2010a, the energy consumption rates for all light duty vehicles were revised and all
trucks were assigned to the "conventional internal combustion" class.
VIUS data was our primary source for fuel information, though we also used AEO2009
data for future years. The VIUS ENGTYP field was converted to the MOVES FuelTypelD. For
MOVES2010, all non-gasoline trucks were assigned to diesel fuel, so that the default fleet
contains only gasoline and diesel fuel trucks.It was not possible to identify the fuel used for the
VIUS category "Other." Vehicles in this category were omitted from the analysis and model
year results were renormalized.
As noted in peer review, the original truck diesel fractions were quite erratic, so the
MOVES2010 values were smoothed to reduce year-to-year variability.
For MOVES2010, we smoothed the diesel fractions for the passenger trucks and
recalculated the diesel fractions for the light commercial trucks by adding a new source of
information: Ward's sales data from 1980 through 2007.31 Assuming that the Ward's data was
correct for the total of passenger trucks and light commercial trucks, we back-calculated the
diesel fractions for the light commercial trucks. Specifically, we recalculated each passenger
truck fraction as a 3-year weighted rolling average of the VIUS passenger truck results for model
years 1980-1999 and used the 1999 result of 2.3% for all post-1999 model years. To avoid an
extreme value for the light commercial trucks, we also changed the 1982 value for passenger
trucks from 1.9% to 4%. We then assumed a 3:1 ratio of passenger trucks to light commercial
trucks (as was true in 1999) and used the formula (Ward's percent-0.75*Avg VIUS)/0.25 to
estimate the diesel sales fractions for light commercial trucks. For model years beyond 2007 we
31
-------
assumed the average Ward's result of 4.37%. For model years prior to 1979, the 1979 diesel
fractions are used. For model years after 2007, the 2007 fractions are used.
The inputs and results of this analysis are illustrated in Figure 4-1. While the diesel
fractions continue to exhibit year-to-year variability, they no longer have the unrealistic extreme
values found in the original data.
Figure 4-1. Diesel Fractions for Light Trucks
Diesel Fractions for Light Trucks
n Vt -,
n ?
n 7*1 -
c
o
"o n 9 -
C5
i_
"5 n 1^
8) U. ID
o
b 01 -
On^
A
m
j
A 7?
• \ / \/
! , A L,
E WVfit .
vV^S^C^v ^
— VIUS31S
— VIUS32S
— Wards
final 31
final 32
1970 1975 1980 1985 1990 1995 2000 2005 2010
Model year
Diesel fraction values were also smoothed for the other trucks. For the single-unit short-
haul trucks (52s), the minimum recorded fraction (0.01) was used for model years 1970 and
earlier. The VIUS value of 0.6 was used for 1990 and and the AEO2009 value of 0.70 was used
for 2003-and-later. We used linear interpolation to establish values in the years between these
established years, and assigned the same fractions to the less abundant single-unit long-haul
trucks (53s). This is illustrated in Figure 4-2.
-------
Figure 4-2. Diesel Fractions for Single Unit Trucks
Diesel Fractions for Single Unit Trucks
o
1960
1970 1980 1990 2000 2010 2020
Model year
Draft 52s • Draft
. 52 & 53 final
Combination long-haul trucks are virtually all diesel, so we set the fraction to 1 for all
model years. Combination short-haul trucks were also mostly diesel, except in model years prior
to 1990 when a small fraction of gasoline trucks appeared in the VIUS data. The VIUS values
are used for model years 1984+, but given the small sample size for 1983-and-earlier, we
assigned the 1984 fraction to these model years. Table 4-3 summarizes the diesel fractions for
MOVES general truck categories by model year. The gasoline fractions can be calculated as one
minus the diesel fractions listed here.
33
-------
Table 4-3. Diesel Fractions for Trucks
Source
Type
Model
Year
1979 and
earlier
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007+
Passenger
Trucks
31
0.0139
0.0124
0.0178
0.0400
0.0209
0.0145
0.0172
0.0222
0.0105
0.0049
0.0076
0.0134
0.0200
0.0207
0.0212
0.0180
0.0149
0.0198
0.0187
0.0100
0.0232
0.0232
0.0130
0.0130
0.0130
0.0130
0.0113
0.0137
0.0099
Light
Commercial
Trucks
32
0.0419
0.1069
0.0706
0.2200
0.2053
0.1484
0.1003
0.0814
0.0606
0.0773
0.0931
0.0838
0.0680
0.0698
0.0846
0.1021
0.1192
0.0887
0.1360
0.0380
0.1665
0.1225
0.1730
0.1570
0.1330
0.1810
0.1011
0.0609
0.0592
Single-Unit
Short-haul
Trucks
52
0.2655
0.2950
0.3245
0.3540
0.3835
0.4130
0.4425
0.4720
0.5015
0.5310
0.5605
0.6000
0.6077
0.6154
0.6231
0.6308
0.6385
0.6462
0.6538
0.6615
0.6692
0.6769
0.6846
0.6923
0.7000
0.7000
0.7000
0.7000
0.7000
Single-Unit
Long-haul
Trucks
53
0.2655
0.2950
0.3245
0.3540
0.3835
0.4130
0.4425
0.4720
0.5015
0.5310
0.5605
0.6000
0.6077
0.6154
0.6231
0.6308
0.6385
0.6462
0.6538
0.6615
0.6692
0.6769
0.6846
0.6923
0.7000
0.7000
0.7000
0.7000
0.7000
Combination
Short-haul
Trucks
61
0.9146
0.9146
0.9146
0.9146
0.9146
0.9146
0.8985
0.9628
0.9940
0.9855
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
Combination
Long-haul
Trucks
62
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
1.0000
4.3.2. SizeWeightFraction
The SizeWeightFraction is used for calculating energy consumption in MOVES2010, but
not in MOVES2010a. Engine size distributions for trucks were determined using the VIUS 1997
and 2002 database. The VIUS database categorizes engine size by fuel type and the categories
do not exactly match the MOVES categories. We mapped from the VIUS engine size categories
to the MOVES engine size categories as described in Table 4-4. For comparison, the engine size
ranges for both the VIUS and MOVES categories are listed in cubic inches displacement.
34
-------
Table 4-4. Mapping VIUS Engine Size Categories to MOVES EngSizelD
Fuel Type
Gasoline
Gasoline
Gasoline
Gasoline
Gasoline
Gasoline
Gasoline
Diesel
Diesel
Diesel
Propane
Alcohol
Alcohol
Alcohol
Alcohol
Other
Other
Other
Other
Other
Other
Fuel Not
Reported
Vehicle Not
In Use
All
VIUS
Fuel_CID
code
1,2
3,4
5,6
7,8
9,10
11,12
13-18
20
21
22-36
38-41
43
44
45
46
48
49
50
51
52
53-56
58-61
63-66
19,37,42,47,5
7,62,67
VIUS CID
Range
1-129
130-149
150-179
180-209
210-239
240-299
300 & Up
1-249
250-299
300 & Up
All
1-229
230-269
270-339
340 & Up
1-99
100-149
150-199
200-249
250-299
300 & Up
All
All
Unknown
MOVES
EngSizelD
Code
20
2025
2530
3035
3540
4050
5099
3540
4050
5099
5099
3035
3540
4050
5099
20
2025
2530
3540
4050
5099
5099
5099
0
MOVES CID
Range
1-122
122-153
153-183
183-214
214-244
244-305
305 & Up
214-244
244-305
305 & Up
305 & Up
183-214
214-244
244-305
305 & Up
1-122
122-153
153-183
214-244
244-305
305 & Up
305 & Up
305 & Up
Unknown
Determining weight categories for light trucks was fairly complicated. The VIUS 1997
data combines information from two different survey forms. The first form was administered
for VIUS "Strata" 1 and 2 trucks: pickup trucks, panel trucks, vans (including mini-vans), utility
type vehicles (including jeeps) and station wagons on truck chassis. The second form was
administered for all other trucks. While both surveys requested information on engine size, only
the second form requested detailed information on vehicle weight. Thus for Strata 1 and 2
trucks, VIUS classifies the trucks only by broad average weight category (AVGCK): 6,000 Ibs or
less, 6,001-10,000 Ibs, 10,001-14,0001bs, etc. To determine a more detailed average engine size
and weight distribution for these vehicles, we used the Oak Ridge National Laboratory (ORNL)
light-duty vehicle database to correlate engine size with vehicle weight distributions by model
year.
In particular, for sourceTypes 31 and 32 (Passenger Trucks and Light Commercial
Trucks):
35
-------
• VIUS 1997 trucks of the sourcetype in Strata 3, 4, and 5 were assigned to the
appropriate MOVES weight class based on VIUS detailed average weight
information.
• VIUS 1997 trucks of the sourcetype in Strata 1 and 2 were identified by enginesizelD
and broad average weight category.
• Strata 1 and 2 trucks in the heavier (10,001-14,000 Ibs, etc) VIUS 1997 broad
categories were matched one-to-one with the MOVES weight classes.
• For trucks in the lower broad categories (6,000 Ibs or less and 6001-10,000 Ibs), we
used VIUS 1997 to determine the fraction of trucks by model year and fuel type that
fell into each engine size/broad weight class combination (the "VIUS fraction")
• We assigned trucks in the ORNL light duty vehicle database to a weightclassID by
adding SOOlbs to the recorded curb weight and determining the appropriate MOVES
weight class.
• For the trucks with a VIUS 1997 average weight of 6,000 Ibs or less, we multiplied
the VIUS 1997 fraction by the fraction of trucks with a given weightclassID among
the trucks in the ORNL database that had the given engine size and an average weight
of 6,000 Ibs or less. Note, the ORNL database did not provide information on fuel
type, so the same distributions were used for all fuels.
• Because the ORNL database included only vehicles with a GVW up to 8500 Ibs, we
did not use it to distribute the trucks with a VIUS 1997 average weight of 6,001-
10,000 Ibs. Instead these were distributed equally among the MOVES
WeightClassIDs 70, 80, 90 and 100.
Source Types 52 and 53 (Long- and Short-haul Single Unit Trucks) also included some
trucks in VIUS 1997 Strata 1 and 2, thus a similar algorithm was applied.
• VIUS 1997 trucks of the SourceType in Strata 3, 4, and 5 were assigned to the
appropriate MOVES weight class based on VIUS 1997 detailed average weight
information.
• VIUS 1997 trucks of the SourceType in Strata 1 and 2 were identified by
enginesizelD and broad average weight category.
• Strata 1 and 2 trucks in the heavier (10,001-14,000 Ibs, etc) VIUS 1997 broad
categories were matched one-to-one with the MOVES weight classes.
• For trucks in the lower broad categories (6,000 Ibs-or-less and 6001-10,000 Ibs), we
used VIUS 1997 to determine the fraction of trucks by model year and fuel type that
fell into each engine size/broad weight class combination (the "VIUS fraction")
• We did not believe the ORNL light duty vehicle database adequately represented
single unit trucks. Thus, the trucks with a VIUS 1997 average weight of 6,000 Ibs or
less and an engine size less than 5 liters were distributed equally among the MOVES
weight classes 20, 25, 30, 35, 40, 45, 50, and 60. Because no evidence existed of
very light trucks among the vehicles with larger engines (5 liter or larger), these were
equally distributed among MOVES weight classes 40, 45, 50 and 60.
-------
• The trucks with a VIUS 1997 average weight of 6,001-10,000 Ibs were distributed
equally among the MOVES weight classes 70, 80, 90 and 100.
Sourcetypes 61 and 62 (Long- and Short-haul combination trucks) did not include any
vehicles of VIUS 1997 Strata 1 or 2. Thus we used the detailed VIUS 1997 average weight
information and engine size information to assign engine size and weight classes for all of these
trucks.
When VIUS2002 became available, we updated values that had been based on
VIUS 1997. The VIUS 2002 contains an estimate of the average weight (vehicle weight plus
cargo weight) of 1998-2002 model year vehicle or vehicle/trailer combination as it was most
often operated when carrying a typical payload during 2002. These estimates were used to
determine the MOVES weightClassID categories for these trucks. Table 4-5 shows the weight
ranges used for each weightClassID. Any vehicles with a zero or missing value for the average
weight and without a weight classification in the WeightAvgCK field were excluded from the
analysis for determining the average weight distributions.
Since there is a smaller number of gasoline trucks among the single unit and refuse
trucks, all model years (1998-2002) were combined to determine a single weight distribution to
use for these model years.
The VIUS 1997 based estimates were retained for light duty trucks (sourceTypelD = 31,
32) and for all model years prior to 1998.
In cases where distributions were missing (no survey information), distributions from a
nearby model year with the same source type was used. Weight distributions for all 2003 and
newer model years were set to be the same as for the 2002 model year for each source type.
37
-------
Table 4-5. Mapping VIUS Average Weight to MOVES WeightClassID
Where Weight Avg is not zero:
weightClassID
20
25
30
35
40
45
50
60
70
80
90
100
140
160
195
260
330
400
500
600
800
1000
1300
9999
WeightAvg Range
1-2000
2000-2499
2500-2999
3000-3499
3500-3999
4000-4499
4500-4999
5000-5999
6000-6999
7000-7999
8000-8999
9000-9999
10000-13999
14000-15999
16000-19499
19500-25999
26000-32999
33000-39999
40000-49999
50000-59999
60000-79999
80000-99999
100000-129999
130000 & Up
Where Weight Avg is zero:
weightClassID
140
160
195
WeightAvgCK
4 (10000-14000)
5 (14000-16000)
6 (16000-19500)
4.3.3. RegClassFraction
Regulatory classes are used to group vehicles subject to similar emission standards. The
regulatory classes used in MOVES are summarized in Table 4-6 below. The "Doesn't Matter"
regulatory class is used internally in the model if the emission rates for a given pollutant and
process are independent of regulatory class. The Motorcycle and Light Duty Vehicle regulatory
classes have a one-to-one correspondence with sourcetype. Other sourcetypes are allocated
between regulatory classes based on gross vehicle weight rating (GVWR) (the maximum weight
that a truck is designed to carry), the regulatory definition of urban buses, and an internal
MOVES rule that only passenger trucks and light commercial trucks may be assigned to
regulatory classes 41 and 42, while only buses, single unit and combination trucks are assigned
to regulatory classes 46 and higher.6
e This final condition is necessary because of a change in the way Scaled Tractive Power was calculated for heavy
trucks. See Section 7 for more information.
-------
Table 4-6. Regulatory Classes in MOVES
Reg Class ID
0
10
20
30
41
42
46
47
48
Reg Class Name
Doesn't Matter
MC
LDV
LOT
LHD<=14k
LHD45
MHD
HMD
Urban Bus
Reg Class Description
Doesn't Matter
Motorcycles
Light Duty Vehicles
Light Duty Trucks
Light Heavy Duty (8500
lbs 33K Ibs)
Urban Bus (see CFR
Sec. 86.091_2)
In particular, we used the VIUS response "PKGVW" in VIUS 1997 and ADM_GVW in
VIUS 2002 and the Davis & Truit report on Class 2b Trucks32 to determine GVW fractions by
fuel type. The VIUS fields are intended to identify the Polk weight class. Work for MOBILE6
using the VIUS precursor, TIUS 1992 indicated that the PKGVW measure in VIUS is
problematic. TIUS PKGVW is taken from the truck VEST, but is not always consistent with the
indicated average and maximum weight. (For example, the reported "maximum weight" often
exceeded the PKGVW.) These problems were also seen in VIUS. However, "maximum
weight" was not available for smaller trucks, and the other measures of weight reported in VIUS
were not consistent with the need for an indicator of the relevant emission standards. When the
PKGVW led to unusual results, for example, particularly high fraction of LDT among
combination trucks, we checked additional VIUS fields to determine if the PKGVW was
mistaken. In some cases, the PKGVW was manually revised to a higher value and fractions
were recomputed. In other cases, the PKGVW was consistent with the other fields, and the
difference reflected the fact that our sourcetype categories are based on axle counts and trailer
configurations rather than weight. For example, a 6-tire ("dually") pickup that regularly pulls a
trailer is classified as a "Combination Truck," although, by weight, it would be in the LDT
regulatory class. Some model years had relatively high fractions of such trucks. It is likely these
high values indicate a problem with small sample size for the model year. For MOVES2010, all
the light heavy duty (<195,000 Ibs) combination and short haul trucks were assigned to the
medium heavy duty regulatory class.
Also, because the split between the LDT and LHD<=14K regulatory class is at 8500 Ibs,
it was necessary to split the Polk GVW Class 2 into class 2a (6001-8500 Ibs) and class 2b (8501-
10,000 Ibs). Davis & Truitt33 report that, on average, 23.3 percent of Class 2 trucks are in Class
2b; 97.4 percent of Class 2a trucks are powered by gasoline, and 76 percent of Class 2b trucks
are powered by gasoline. From this information, we estimate that 19.2 percent of gasoline-
powered Class 2 trucks are Class 2b and that 73.7 percent of diesel-powered class 2 trucks are
Class 2b.
39
-------
Table 4-7. Light Truck Class 2 Weight Distribution
Fuel Type
Gasoline
Diesel
Any
Class 2a
6001-8500 Ibs.
GVWR
74.7%
2.0%
76.7%
Class 2b
8501-10000 Ibs. GVWR
17.7%
5.6%
23.3%
Class 2b Fraction
19.2%
73.7%
The regulatory class fractions for trucks are listed below in Tables 4-8, 4-9 and 4-10. All
1986 and newer model year data was obtained from VIUS 2002. The pre-1986 model year
values are from VIUS 1997.
40
-------
Table 4-8. Passenger & Light Commercial Truck Regulatory Class Percents
Mod
el
Year
1966
and
earlier
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002+
Passenger Trucks
Gasoline
LDT
81%
90%
88%
100%
99%
96%
96%
95%
95%
97%
95%
89%
85%
87%
90%
96%
94%
95%
94%
94%
93%
95%
95%
95%
95%
96%
95%
95%
95%
95%
95%
95%
95%
93%
94%
94%
95%
LHD
<=14
K
19%
10%
12%
0%
1%
3%
4%
5%
5%
3%
5%
11%
15%
13%
10%
4%
6%
5%
6%
6%
7%
5%
5%
5%
5%
4%
5%
5%
5%
5%
5%
5%
5%
7%
6%
6%
5%
Diesel
LD
T
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
38%
26%
26%
26%
26%
26%
30%
31%
29%
26%
26%
26%
36%
23%
28%
LHD
<=14
K
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
74%
74%
74%
74%
74%
70%
68%
70%
74%
74%
73%
63%
76%
72%
LHD
>14K
*
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
33%
1%
1%
0%
1%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Light Commercial Trucks
Gasoline
LDT
24%
72%
67%
91%
80%
94%
75%
59%
65%
72%
88%
79%
81%
78%
74%
89%
72%
90%
87%
87%
82%
90%
89%
89%
91%
89%
91%
91%
87%
88%
86%
88%
89%
87%
88%
88%
89%
LHD
<=14
K
6%
17%
1%
0%
12%
4%
5%
9%
9%
17%
8%
13%
16%
9%
17%
5%
12%
6%
9%
12%
11%
10%
9%
10%
7%
10%
9%
8%
12%
11%
13%
11%
10%
12%
11%
12%
10%
LHD
>14K
71%
11%
32%
9%
9%
2%
20%
32%
26%
10%
4%
7%
3%
13%
9%
6%
16%
4%
4%
1%
7%
0%
2%
1%
2%
2%
1%
1%
1%
1%
1%
1%
1%
1%
0%
0%
0%
Diesel
LDT
7%
7%
7%
7%
7%
7%
7%
7%
7%
7%
7%
7%
7%
7%
40%
12%
27%
23%
24%
23%
35%
9%
21%
14%
6%
18%
22%
15%
16%
20%
21%
36%
14%
21%
22%
34%
26%
LHD
<=14
j^**
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
42%
49%
63%
46%
27%
52%
63%
47%
50%
56%
48%
52%
44%
58%
50%
54%
62%
LH
D>1
4K
93%
93%
93%
93%
93%
93%
93%
93%
93%
93%
93%
93%
93%
93%
60%
88%
73%
77%
76%
77%
23%
42%
16%
39%
67%
30%
15%
38%
34%
24%
31%
12%
42%
21%
29%
11%
12%
*Note, the relatively high fraction of 42s for pre-1989 diesel passenger trucks is an error, but this has a very small
impact on emissions because these are a small portion of the fleet, and for most pollutants, the emission rates for
regulatory classes 41s and 42s are identical.
**In the future, the 1985-and-earlier diesel light commercial trucks could be split between regulatory classes 41 and
42, but the impact of this change would be negligible for most pollutants.
-------
Table 4-9. Percentage of Medium Heavy-Duty Trucks among Diesel-fueled Single-
Unit and Combination Trucks*
Model
year
1972 and
earlier
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002+
Refuse
Trucks
51
100%
100%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
37%
30%
19%
29%
26%
42%
21%
48%
30%
25%
25%
19%
15%
13%
13%
16%
22%
3%
3%
Single Unit
Trucks
52&S3
0%
3%
6%
14%
44%
43%
36%
34%
58%
47%
66%
90%
59%
65%
51%
64%
62%
68%
72%
78%
66%
74%
73%
73%
74%
77%
70%
77%
76%
78%
73%
Motor
Homes
54
100%
100%
100%
100%
100%
100%
100%
100%
100%
97%
95%
96%
98%
98%
98%
99%
99%
99%
98%
98%
96%
96%
97%
97%
97%
97%
96%
97%
97%
97%
97%
Short-haul
Comb. Trucks
61
0%
8%
30%
3%
13%
31%
18%
16%
29%
31%
14%
28%
56%
36%
16%
25%
18%
20%
27%
19%
25%
15%
20%
17%
21%
12%
18%
17%
15%
11%
22%
Long-haul Comb.
Trucks
62
0%
0%
0%
0%
0%
0%
0%
0%
5%
6%
0%
17%
63%
30%
5%
3%
4%
6%
21%
8%
8%
17%
7%
7%
9%
6%
7%
4%
2%
4%
5%
* Among these sourcetypes, all remaining trucks are in the heavy-heavy-duty regulatory class.
42
-------
Table 4-10. Percentage of Medium Heavy-Duty Trucks among Gasoline-fueled
Single-Unit and Combination Trucks*
Model
year
1985 and
earlier
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002+
Refuse
Trucks &
Motor
Homes
51
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
Single Unit
Trucks
52&S3
100%
95%
100%
100%
99%
100%
99%
99%
99%
96%
98%
96%
96%
98%
94%
96%
92%
96%
Short-haul
Comb. Trucks
61
100%
79%
79%
79%
#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
#N/A
* Among these sourcetypes, all remaining trucks are in the heavy-heavy-duty regulatory class.
4.3.4. SCCVtypeFraction
Trucks span a wide range of GVWs and, thus, a wide range of SCCs. We used VIUS
values for GVW to determine the truck SCC fractions by model year. To separate Light-Duty
Trucks 1 and Light-Duty Trucks 2, which are distinguished by Loaded Vehicle Weights, we used
information from the Oak Ridge National Laboratory Light-Duty Vehicle database. And to
separate Class 2a and 2b trucks, we used information from Davis and Truitt.34
The resulting truck mappings are too complex to summarize here, but are available in the
MOVES database. If the Alternative Vehicle Fuels and Technologies control strategy is used to
assign vehicles to the "Electric" fueltype, those vehicles are mapped to "LDGV" because there is
no SCC for electric vehicles.
4.4. Buses
Because buses are not included in VIUS and because the Polk data we had for school
buses was incomplete, the source bin fractions for buses is based on a variety of data sources and
assumptions. Values for transit buses, school buses, and intercity buses were calculated
separately.
43
-------
7.4.1. FuelEngFraction
All buses were assigned to EngTechID "1" (conventional internal combustion).
We followed the Energy Information Administration (EIA) in assigning all intercity
buses to conventional diesel engines (AEO2006, Supplemental Table 34).
The National Transit Database (NTD) responses to form 408 (Revenue Vehicle
Information Form) included information classifying transit buses to a variety of fuel types by
model year. The mapping from NTD fuel types to MOVES fuel types is summarized in Table
4-11. The associated fractions by model year are summarized in Table 4-12.
Table 4-11.
Types
Mapping National Transit Database Fuel Types to MOVES Fuel
NTD code
BF
CN
DF
DU
EB
EP
ET
GA
GR
KE
LN
LP
MT
OR
NTD description
Bunker fuel
Compressed natural gas
Diesel fuel
Dual fuel
Electric battery
Electric propulsion
Ethanol
Gasoline
Grain additive
Kerosene
Liquefied natural gas
Liquefied petroleum gas
Methanol
Other
MOVES
Fuel ID
na
3
2
2
9
9
5
1
na
na
3
4
6
na
MOVES Fuel
Description
CNG
diesel
diesel
electric
electric
ethanol
gasoline
CNG
LPG
methanol
44
-------
Table 4-12. National Transit Database Implied Fuel Fractions for Transit Buses
Model
Year
1978-and
earlier
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
Gasoline
0
0.033981
0
0.002088
0.001894
0
0.001603
0
0.00079
0.001402
0.002377
0.00113
0.002941
0.003134
0.010769
0.003061
0.010711
0.009555
0.017963
0.012702
0.012003
0.005998
Diesel
1
0.966019
1
0.997912
0.992424
1
0.998397
0.999565
0.996447
0.998598
0.997623
0.998306
0.990271
0.978064
0.933903
0.918707
0.900625
0.835108
0.881825
0.810162
0.838409
0.878041
CNG
0
0
0
0
0
0
0
0.000435
0.002764
0
0
0
0.006787
0.018106
0.046417
0.07551
0.084796
0.153153
0.097613
0.174365
0.1487
0.113296
LPG
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.000743
0.00068
0.000893
0
0.000709
0.000462
0
0
Ethanol
0
0
0
0
0
0
0
0
0
0
0
0.000565
0
0
0
0.001361
0
0
0
0
0
0
Methanol
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.005941
0
0
0
0
0
0
0
Electric
0
0
0
0
0.005682
0
0
0
0
0
0
0
0
0.000696
0.002228
0.00068
0.002975
0.002184
0.001891
0.002309
0.000889
0.002666
For MOVES2010, most alternative fuels were removed from the model. However,
because the number of compressed natural gas (CNG) transit buses was high for some model
years, CNG was retained as an option for transit buses, and the default database includes a CNG
for some model years, as summarized in Table 4-13 below. For each model year, one percent of
the transit bus fleet was assigned to gasoline engines and the remaining, (non-gasoline, non-
CNG) fraction was assigned to diesel.
45
-------
Table 4-13. Transit Bus Fuel Fractions in MOVES2010
Model
Year
1989-and-
earlier
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000+
Gasoline
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
1.0%
Diesel
99.0%
98.3%
97.2%
94.4%
91.4%
90.5%
83.7%
89.2%
81.6%
84.1%
87.7%
93.0%
CNG
0%
0.7%
1.8%
4.6%
7.6%
8.5%
15.3%
9.8%
17.4%
14.9%
11.3%
6.0%
The available Polk data excluded fuel information on school buses and we were unable to
locate any other source for bus fuel fractions. (The Union of Concerned Scientists estimated that
about one percent of school buses are fueled by either CNG or propane, but does not provide
estimates by model year.35) Thus we used the diesel fractions from MOBILE6, which were
derived from Polk 1996 and 1997 data. We assigned non-diesel buses to gasoline. These
fractions are summarized in Table 4-14. In the future it would be desirable to obtain up-to-date,
detailed fuel information for school buses from Polk or some other source.
-------
Table 4-14. School Bus Fuel Fractions in MOVES2010
Model Year
1975-and-earlier
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996+
Gasoline
0.991272
0.99145
0.976028
0.970936
0.95401
0.94061
0.736056
0.674035
0.676196
0.615484
0.484507
0.326706
0.265547
0.249771
0.229041
0.124036
0.089541
0.010041
0.120539
0.147479
0.114279
0.041539
Diesel
0.008728
0.00855
0.023972
0.029064
0.04599
0.05939
0.263944
0.325965
0.323804
0.384516
0.515493
0.673294
0.734453
0.750229
0.770959
0.875964
0.910459
0.989959
0.879461
0.852521
0.885721
0.958461
4.4.2. SizeWeightFraction
The SizeWeightFraction is used for calculating energy consumption in MOVES2010, but
not in MOVES2010a. While the vast majority of buses of all types have engine displacement
larger than five liters (EngSizeID=5099), it was difficult to find detailed information on average
bus weight.
For intercity buses, we used information from Table II-7 of the FT A 2003 Report to
Congress36 that specified the number of buses in various weight categories. This information is
summarized in below in Table 4-15. Note the FT A uses the term "over-the-road bus" to refer to
the class of buses roughly equivalent to the MOVES "intercity bus" category. The FTA weight
categories were mapped to the equivalent MOVES weight classes.
Ta
ble 4-15. FTA Estimate of Bus Weights
Weight (Ibs)
0-20,000
20,000-30,000
30,000-40,000
40,000-50,000
total
MOVES
Weight
ClassID
400
500
MOVES
Weight Range
(Ibs)
33,000-40,000
40,000-50,000
Number buses
(2000)
173,536
392,345
120,721
67,905
754,509
Bus type
school & transit
school & transit
school & transit & intercity
intercity
47
-------
Using our 1999 bus population estimates (in Table3-6), we were able to estimate the
fraction of all buses that were intercity buses and then to estimate the fraction of intercity buses
in each weight bin. In particular:
Estimated number of intercity buses in 2000:
754,509 * (84,4547(84,454+55,706+592,029)) = 87,028
Estimated number of intercity buses 30,000-40,000 Ibs:
Estimated intercity bus weight distribution:
This distribution was used for all model years.
87,028-67,905 = 19,123
Class 400 = 19,123/87,028 = 22%
Class 500 = 67,905/87,028 = 78%
For transit buses, we took average curb weights from Figure II-6 of the FT A Report to
Congress37and added additional weight to account for passengers and alternative fuels. The
resulting in-use weights were all in the range from 33,850 to 40,850. Thus all transit buses were
assigned to the weight class "400" (33,000 - 40,000 Ibs) for all model years. This estimate could
be improved if more detailed weight information for transit buses becomes available.
For school buses, we used information from a survey of California school buses. While
this data is older and may not be representative of the national average distribution, it was the
best data source available. The California data38 provided information on number of vehicles by
gross vehicle weight class and fuel as detailed in Table 4-16.
Table 4-16. California School Buses
LHDV
MHDV
HHDV
Total
Gas
2740
467
892
4099
Diesel
4567
2065
11639
18271
Other
8
2
147
157
Total
7315
2534
12678
To estimate the distribution of average weights among the MOVES weight classes, we
assumed that the Light Heavy-Duty (LHDV) school buses were evenly distributed among
weightClassIDs 70, 80, 90, 100, and 140. Similarly, we assumed the Medium Heavy-Duty
(MHDV) school buses were evenly distributed among weightClassIDs 140, 160, 195, 260, and
330 and the Heavy Heavy-Duty (HHDV) school buses were evenly distributed among
weightClassIDs 195, 260, 330, and 440.
The final default weight distributions for buses are summarized in Table 4-17.
-------
Table 4-17. Weight Distributions for Buses by Fuel Type
Weight Class
70
80
90
100
140
160
195
260
330
400
500
Intercity
Buses
(41)
Diesel
0.2197
0.7800
Transit Buses
(42)
Diesel & Gas
1.0000
School Buses (43)
Diesel
0.0500
0.0500
0.0500
0.0500
0.0726
0.0226
0.1819
0.1819
0.1819
0.1593
Gas
0.1337
0.1337
0.1337
0.1337
0.1565
0.0228
0.0772
0.0772
0.0772
0.0544
4.4.3. RegClassFraction
For buses, the same regulatory class fractions were used for all model years. All gasoline
buses were assigned to the medium-heavy-duty regulatory class. Diesel intercity buses were
assigned to the heavy-heavy-duty class. Diesel transit buses were assigned to the urban bus
class. Diesel school buses were split using the California survey data and MOVES assignment
rules, with 36 percent assigned to the medium-heavy-duty class and 64 percent assigned to the
heavy-heavy-duty class.
4.4.4. SCCVtypeFraction
For most buses, the mapping to SCCVtype was straightforward. These mappings are
summarized in Table 4-18.
Table 4-18. SCC Mappings for Buses
Source
Type ID
41
41
42
42
43
43
SourceType
Intercity Bus
Intercity Bus
Transit Bus
Transit Bus
School Bus
School Bus
Fuel Type
gasoline
other
gasoline
other
gasoline
other
SCC-ID
4
12
4
12
4
12
SCC
prefix
2201070
2230075
2201070
2230075
2201070
2230075
Abbreviated
Description
HDGV&B
HDDB
HDGV&B
HDDB
HDGV&B
HDDB
If the Alternative Vehicle Fuels and Technologies control strategy is used to assign
buses to the "Electric" fueltype, those vehicles are mapped to "FtDGV&B" because there is no
SCC for electric vehicles.
49
-------
4.5. Refuse Trucks
Values for Refuse Trucks (Source Type 51) were computed from information in VIUS.
4.5.1. FuelEngFraction
All Refuse Trucks were assumed to have conventional internal combustion engines.
Because the VIUS sample was small and the fuel fractions by model year were quite erratic, we
calculated an average gasoline fraction (4.0%) and applied it in all model years.
4.5.2. SizeWeightFraction
The SizeWeightFraction is used for calculating energy consumption in MOVES2010, but
not in MOVES2010a. Because the sample of Refuse Trucks in VIUS was small, the SizeWeight
distributions were calculated for model year groups rather than individual model years. As for
other trucks, the EngineSize group was determined from the VIUS engine size categories and the
WeightClass was determined from the VIUS reported average weight.
-------
Table 4-19. Refuse Truck SizeWeight Fractions by Fuel Type
Gasoline
Engine Size
3-3. 5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
Sum
Diesel
Engine Size
3.5-4L
4-5L
4-5L
4-5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
>5L
Sum
Weight (Ibs.)
5000-6000
7000-8000
9000-10000
10000-14000
14000-16000
16000-19500
19500-26000
26000-33000
33000-40000
50000-60000
Weight (Ibs.)
10000-14000
10000-14000
14000-16000
16000-19500
9000-10000
10000-14000
14000-16000
16000-19500
19500-26000
26000-33000
33000-40000
40000-50000
50000-60000
60000-80000
80000-100000
100000-130000
Pre-1997
0.009074
0.148826
0.070720
0.135759
0.199961
0.055085
0.205341
0.022105
0.153129
0
1.000000
Pre-1998
0.007758
0
0
0
0.006867
0.011727
0.022960
0.063128
0.099782
0.102077
0.237485
0
0.336484
0.111730
0
0
1.000000
1997 and
Newer
0
0
0
0.324438
0.593328
0
0
0
0
0.082234
1.000000
1998
0
0
0
0
0.009593
0
0
0
0.035378
0.019625
0.103922
0.283642
0.338511
0.196424
0
0.012904
1.000000
1999
0
0
0
0
0
0
0
0.011367
0.026212
0.067419
0.088975
0.275467
0.326902
0.193238
0.010420
0
1.000000
2000
0
0
0.015505
0
0
0
0
0.047200
0.052132
0.072106
0.085991
0.165624
0.384612
0.176831
0
0
1.000000
2001
0
0
0
0.011670
0
0.019438
0
0
0.018329
0.043877
0.042678
0.266357
0.315133
0.282517
0
0
1.000000
2002 and Newer
0
0.006614
0
0
0
0
0
0
0.026079
0
0.046966
0.194716
0.474469
0.224995
0.013081
0.013081
1.000000
4.5.3. RegClassFraction
Using the VIUS data on gross vehicle weight, diesel Refuse Trucks were classified as
Heavy-Heavy-Duty Trucks or Medium-Heavy-Duty trucks, as detailed in the truck Table 4-9.
above. Using VIUS data and MOVES regulatory class assignment rules, gasoline Refuse Trucks
were all assigned to the Medium-Heavy-Duty class.
4.5.4. SCCVtypeFraction
We used VIUS data on gross vehicle weight to determine fractions for diesel Refuse
Trucks. They were classified as Heavy-Heavy-Duty Diesel Vehicles or Medium-Heavy-Duty
Diesel Vehicles, as detailed in Table 4-20, below. All gasoline Refuse Trucks were all assigned
to the Heavy-Duty Gasoline Vehicle and Bus class.
51
-------
Table 4.20. SCC Mappings for Diesel Refuse Trucks
Model
Year
1973-and-
earlier
1974-1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002+
MHHDDV
1
0
0.3740
0.2963
0.1850
0.2861
0.2563
0.4164
0.2109
0.4799
0.3034
0.2543
0.2536
0.1868
0.1496
0.1256
0.1331
0.1565
0.218
0.0324
0.0298
HHDDV
0
1
0.6259
0.7036
0.8150
0.7139
0.7437
0.5836
0.7891
0.5201
0.6966
0.7457
0.7464
0.8132
0.8504
0.8744
0.8669
0.8435
0.782
0.9676
0.9702
4.6. Motor Homes
Determining source bin distribution for Motor Homes required a number of assumptions
and interpolation due to the lack of detailed information. For each field, the following describes
the information available, assumptions made, and how data points were determined.
4.6.1. FuelEngFraction
Detailed information on motor home fuel distribution was not available. Staff of the
Recreational Vehicle Industry Association (RVIA) told us that the fraction of diesel motor
homes had been relatively constant at 10 to 20 percent for many years.39 This fraction began to
increase steadily and was about 50% in 200940. Based on this information, we interpolated to
determine the diesel fractions listed in Table 4-21. The remaining 1999-and-earlier motor homes
are assumed to be gasoline-fueled. We assigned all motor homes to the conventional internal
combustion engine type.
-------
Table 4-21. Diesel Fractions for Motor Homes.
Model Year
1993-and-earlier
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010+
Percent Diesel
15%
18%
21%
23%
26%
29%
32%
34%
37%
40%
41%
43%
44%
46%
47%
49%
50%
50%
4.6.2. SizeWeightFraction
The SizeWeightFraction is used for calculating energy consumption in MOVES2010, but
not in MOVES2010a. No detailed information was available on average engine size and weight
distributions for motor homes. We assumed all motor home engines were 5 L or larger. As a
surrogate for average weight, we used information on gross vehicle weight provided in the Polk
TIP® 1999 database by model year and mapped the Polk GVW Class to the MOVES weight
bins. These values are likely to overestimate average weight. The Polk TIP® information did
not specify fuel type, so we assumed that the heaviest vehicles in the Polk database were diesel-
powered and the remainder were powered by gasoline. This led to the weight distributions in
Table 4-22 and Table 4-23.
53
-------
Table 4-22. Weight Fractions for Diesel Motor Homes by Model Year
Polk GVW
bin
MOVES
weight class
Model Year
1975-and-
earlier
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999+
3
140
4
160
5
195
6
260
7
330
8
400
Diesel
0.171431
0.637989
0.68944
0.423524
0.096922
0.462916
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.792112
0.340639
0.292308
0.574539
0.899344
0.537084
0.941973
0.868333
0.912762
0.932659
0.881042
0.855457
0.791731
0.72799
0.73298
0.173248
0
0
0
0
0
0
0
0
0
0.029828
0.018755
0.012168
0
0
0
0
0
0.000203
0.000835
0.001474
0.013381
0.085493
0.148917
0.128665
0.614798
0.619344
0.551548
0.345775
0.45546
0.635861
0.553807
0.666905
0.267
0
0
0.000436
0.000277
0.000387
0.001067
0
0.030174
0.049
0.014845
0.009183
0.010761
0.022962
0.022498
0.015469
0.043052
0.043628
0.063712
0.01901
0.471873
0.354386
0.163195
0.229529
0.193167
0.335069
0.736656
0.006629
0.002181
0.005531
0.00155
0.002667
0
0
0.03
0.030096
0.036732
0.083285
0.089534
0.087164
0.093335
0.082792
0.149939
0.296399
0.385085
0.144844
0.159622
0.17468
0.184208
0.111299
0.357508
0.233886
0
0
0.000277
0
0
0
0.027853
0.052667
0.042094
0.020592
0.023438
0.018667
0.013113
0.014289
0.012511
0.018387
0.020545
0.044356
0.037509
0.030531
0.026264
0.032456
0.028628
0.040423
0.029458
54
-------
Table 4.23. Weight Fractions for Gasoline Motor Homes by Model Year
Polk GVW
bin
MOVES
weight class
Model Year
1975-and-
earlier
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999+
3
140
4
160
5
195
6
260
7
330
8
400
Gasoline
1
1
1
1
1
1
0.747723
0.732235
0.714552
0.641577
0.692314
0.720248
0.606635
0.459429
0.551601
0.543354
0.612025
0.54464
0.583788
0.481099
0.52997
0.435959
0.221675
0.288222
0.170133
0
0
0
0
0
0
0.252277
0.267765
0.285448
0.358423
0.307686
0.279752
0.393365
0.540571
0.448399
0.456646
0.322022
0.373999
0.361277
0.361146
0.198479
0.289453
0.433334
0.581599
0.392451
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.065952
0.081361
0.054935
0.157755
0.271551
0.274588
0.344991
0.13018
0.288411
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.149004
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
4.6.3. RegClassFraction
Based on Polk data and MOVES regulatory class assignment rules, we assigned all
gasoline motor homes and most diesel-powered motorhomes to the medium-heavy-duty
regulatory class. A small fraction of diesel-powered motorhomes were assigned to the heavy-
heavy-duty class as detailed in Table 4-9 above.
4.6.4. SCCVtypeDistribution
We assigned all gasoline motor homes to the HDGV class. Based on Polk data, we
assigned most diesel-powered motorhomes to the medium-heavy-duty diesel class, as detailed in
Table 4-24 below.
55
-------
Table 4-24. SCCVtype Distributions for Diesel Motor Homes by Model Year
Model Year
1980-and-
earlier
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999+
MHHDDV
1.000
0.972
0.947
0.958
0.979
0.977
0.981
0.987
0.986
0.987
0.982
0.979
0.956
0.962
0.969
0.974
0.968
0.971
0.96
0.971
HHDDV
0.000
0.028
0.053
0.042
0.021
0.023
0.019
0.013
0.014
0.013
0.018
0.021
0.044
0.038
0.031
0.026
0.032
0.029
0.04
0.029
5. Age Distributions
The age distribution for each sourcetype is stored in the SourceTypeAgeDistribution
table. Because sales are not constant, these distributions vary by calendar year. MOVES uses
age distributions for the base year combined with sales and scrappage information to compute
the age distribution in the calendar year selected for analysis.
This section describes how the age distributions were determined for the primary default
base year of 1999, and the 1990 base year. Age distributions for the 1999 base year are
summarized in Table 5-1. Age distributions for the 1990 base year are available in the
MOVES2010 default database SourceTypeAgeDistribution table.
5.1. Motorcycles
To determine the 1999 age fractions for motorcycles, a contractor analyzed Polk
registration data from 2008. These were normalized and input as age distributions for 1999. 41,
The 1990 fractions were determined earlier and were not updated. For these, we began
with Motorcycle Industry Council estimates of the number of motorcycles in use, by model year,
in 1990. However, data for individual model years starting from 1978 and earlier were not
available. A logarithmic regression curve ( R2 value = 0.82) was fitted to available data, which
was then used to extrapolate age fractions for earlier years beginning in 1978.
-------
5.2. Passenger Cars
We considered three approaches to determine 1999 age fractions for passenger cars.
Our original approach (used for MOVES2004 and MOVES Demo) began with Polk
NVPP® 1999 data on car registration by model year. This data presents a snapshot of
registrations on July 1, 1999, and we needed age fractions as of December 31, 1999. To adjust
the values, we used monthly data from the Polk new car database to estimate the number of new
cars registered in the months July through December 1999. Model Year 1998 cars were added to
the previous estimate of "Age 1" cars and Model Year 1999 and 2000 cars were added to the
"Age 0" cars. We then computed fractions by age. However, because this method counts both
Model Year 1999 and Model Year 2000 as "Age 0", the Age 0 age fraction is inflated. When the
MOVES Total Activity Generator applies growth factors, the number of cars in future years is
inflated, and the fraction of passenger cars compared to other source types is skewed. Thus, we
rejected this approach.
A second approach was similar to the first, but with only Model Year 1999 vehicles
counted as "Age 0" in 1999.
Our third approach used passenger car sales data from Table 4.5 of the TEDB42 and
applied the NHTSA survival fractions, extrapolated to age 30 and shifted such that NHTSA age
n = MOVES age n+1. Survival fractions for MOVES age 0 and 1 were interpolated as described
in Section 5.1.
Not surprisingly, the age distributions resulting from the three approaches are very
similar, as illustrated in Figure 5-1. All show a fairly flat age distribution in the first eleven years
followed by a steep decline and a leveling off. The third approach provides a slightly more
generic age distribution than the second approach because the direct Polk data approach is for a
single year and the NHTSA survival fractions were derived by regression through many years of
data. For the MOVES2010 default database, we selected the age distributions generated with the
third approach. For future versions of MOVES, we are considering updating these values to
better account for more recent data.
57
-------
Figure 5-1.1999 Age Distributions for Passenger Cars
Passenger Car Age Distribution
0.09
0.08
-5K Original Polk
-«--Fblk2
-A- • • Cars—Sales & Scrappage
To determine the 1990 age fractions for passenger cars, we began with Polk NVPP®
1990 data on car registration by model year. However, this data presents a snapshot of
registrations on July 1, 1990, and we needed age fractions as of December 31, 1990. To adjust
the values, we used monthly data from the Polk new car database to estimate the number of new
cars registered in the months July through December 1990. Model Year 1989 cars were added to
the previous estimate of "Age 1" cars and Model Year 1990 and 1991 cars were added to the
"Age 0" cars.
Also the 1990 data did not detail model year for ages 15+. Hence, regression estimates
were used to extrapolate the age fractions for individual ages 15+ based on an exponential curve
(R2 value =0.67) fitted to available data.
5.3. Trucks
To determine 1999 age fractions for refuse trucks, short-haul and long-haul single unit
trucks and short-haul and long-haul combination trucks, we used data from the VIUS database.
Vehicles in the VIUS database were assigned to MOVES source types as summarized in Table
3-3 and Table 3-4.
VIUS does not include a model year field and records ages as 0 through 10 and 11-and-
greater. Because we needed greater detail on the older vehicles, we followed the practice used
for MOBILE6 and determined the model year for some of the older vehicles by using the
responses to the VIUS questions "How did you obtain this vehicle?" (VIUS field "OBTAIN" in
VIUS 1997 or "ACQUIREHOW" in VIUS 2002) and "When did you obtain this vehicle?"
-------
(VIUS field "ACQYR" in VIUS 1997 or "ACQUIREYEAR" in VIUS 2002) to derive the model
year of the vehicles that were obtained new. These derived model years also were used for
much of the source bin distribution work described elsewhere in this report.
To calculate age fractions, it was important to account for the inconsistent methodologies
used for the older and newer vehicles. Thus, for each source type, we adjusted the age 11-and-
older vehicle counts by dividing the original count by model year by the fraction of the older
vehicles that were coded as "obtained new." This created an array of adjusted vehicle counts by
model year for calendar year 1997. This 1997 array may overestimate the fraction of mid-aged
vehicles since the fraction of vehicles purchased new likely declines with time; however, we
believe the procedure is reasonable given the limited data available.
We then used the sales growth for 1997 and 1998 from TEDB22 Tables 7.6 and 8.3 and
the scrappage rates from TEDB22 Tables 6.10 and 6.11 to grow the population to the 1999 base
year and then we calculated age fractions.
Initially, we determined the 1999 age fractions for passenger trucks and commercial
trucks in the same way as other trucks. However, when the NHTSA survival rates for light duty
trucks became available in 2006, we reexamined this approach. We compared (1) our original
approach with VIUS data for 1997 and the TEDB scrappage rates, (2) a similar approach using
VIUS data and NHTSA survival rates, and (3) a "sales and scrappage" approach similar to that
used for passenger cars, combining passenger trucks and commercial light trucks and using
TEDB sales data. The results of the three approaches are illustrated in Figure 5-2.
Figure 5-2.1999 Age Distributions for Passenger and Light Commercial Trucks
Passenger and Commercial Light Truck Age Distributiions
0.12
•31 -Original VIUS
•32-OriginalVIUS
•31-VIUS & NHTSA
32-VIUS & NHTSA
•Trucks-Sales & Scrappage
59
-------
Use of the original VIUS data leads to low values for age 0-3 passenger trucks that is not
reflected by vehicle sales data. The other approaches all create similar trends of fairly steep
declines in age fractions until about age 7, a brief leveling off, another steep decline from about
age 12 to 17 and a final leveling off. For the MOVES default database, we selected the age
distribution generated with the "Sales and Scrappage" approach, which will be applied to both
passenger trucks and light commercial trucks.
For the 1990 age fractions for passenger trucks, light commercial trucks, refuse trucks,
short-haul and long-haul single unit trucks and short-haul and long-haul combination trucks, we
used data from the TIUS92 (1992 Truck Inventory and Use Survey) database. Vehicles in the
TIUS92 database were assigned to MOVES source types as summarized in Table 3-3.
Like VIUS97, TIUS92 does not include a model year field and records ages as 0 through
10 and 11-and-greater. Because we needed greater detail on the older vehicles, we followed the
practice used for the 1999 fractions and determined the model year for some of the older vehicles
by using the responses to the questions "How was the vehicle obtained?" (TIUS field
"OBTAIN") and "When did you obtain this vehicle?" (TIUS field "ACQYR") and we adjusted
the age-11-and-older vehicle counts by dividing the original count by model year by the fraction
of the older vehicles that were coded as "obtained new."
5.4. Intercity Buses
For 1990 and 1999 we were not able to identify a data source for estimating age
distributions of intercity buses. Because the purchase and retirement of these buses is likely to
be driven by general economic forces rather than trends in government spending, we will use the
1999 and 1990 age distributions that were derived for short-haul combination trucks, as
described above.
5.5. School Buses and Motor Homes
To determine the age fractions of School Buses and Motor Homes, we used information
from the Polk TIP® 1999 database. School Bus and Motor Home counts were available by
model year. Unlike the Polk data for passenger cars, these counts reflect registration at the end
of the calendar year and, thus, did not require adjustment. We converted model year to age and
calculated age fractions. Because we did not have access to 1990 data, these fractions were also
used for 1990
5.6. Transit Buses
To determine the 1999 age fractions for Transit Buses, we used data from the Federal
Transit Administration database. In particular, we used responses to 1999 Form 408, which
included counts of in-use vehicles by year of manufacture.
To properly account for the fraction of Age 0 vehicles at the end of 1999, it was
necessary to adjust the counts for model-year-1999 vehicles to account for the different reporting
periods of the various transit organizations. The counts were adjusted proportionally depending
on the month in which the fiscal year ended. The adjusted counts were used to calculate the age
fractions.
For 1990 Transit Bus age distributions, we used the MOBILE 6 age fractions since 1990
data on transit buses was not available from the Federal Transit Administration database.
60
-------
Table 5-1. 1999 Age Fractions for MOVES Source Types
source
type
age
0
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
29
30
11
0.1118
0.0993
0.0950
0.0833
0.0627
0.0722
0.0588
0.0492
0.0390
0.0316
0.0234
0.0198
0.0196
0.0163
0.0137
0.0122
0.0089
0.0069
0.0071
0.0079
0.0075
0.0096
0.0147
0.0130
0.0103
0.0127
0.0171
0.0133
0.0152
0.0152
0.0323
21
0.0646
0.0602
0.0610
0.0624
0.0626
0.0642
0.0597
0.0562
0.0543
0.0596
0.0608
0.0622
0.0549
0.0522
0.0419
0.0320
0.0226
0.0155
0.0129
0.0105
0.0080
0.0060
0.0045
0.0034
0.0026
0.0019
0.0014
0.0008
0.0006
0.0005
0.0000
31&32
0.1011
0.0906
0.0837
0.0791
0.0720
0.0700
0.0603
0.0502
0.0429
0.0450
0.0431
0.0422
0.0379
0.0351
0.0311
0.0244
0.0170
0.0127
0.0100
0.0100
0.0081
0.0066
0.0053
0.0041
0.0032
0.0031
0.0030
0.0029
0.0027
0.0026
0.0000
42
0.0624
0.0771
0.0742
0.0727
0.0627
0.0576
0.0504
0.0461
0.0492
0.0759
0.0609
0.0506
0.0489
0.0434
0.0394
0.0320
0.0321
0.0181
0.0082
0.0231
0.0071
0.0032
0.0007
0.0013
0.0009
0.0009
0.0002
0.0004
0.0003
0.0001
0.0002
43
0.0794
0.0660
0.0647
0.0594
0.0798
0.0406
0.0511
0.0435
0.0585
0.0696
0.0419
0.0526
0.0556
0.0512
0.0464
0.0374
0.0144
0.0111
0.0136
0.0138
0.0118
0.0104
0.0107
0.0073
0.0092
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
51
0.0498
0.0398
0.0340
0.0767
0.0926
0.0604
0.0544
0.0243
0.0696
0.0625
0.0514
0.0730
0.0610
0.0796
0.0442
0.0479
0.0145
0.0169
0.0156
0.0040
0.0043
0.0043
0.0000
0.0092
0.0027
0.0070
0.0001
0.0000
0.0000
0.0000
0.0000
52
0.0622
0.0520
0.0412
0.0466
0.0559
0.0572
0.0434
0.0344
0.0351
0.0435
0.0578
0.0531
0.0460
0.0580
0.0430
0.0251
0.0409
0.0220
0.0219
0.0239
0.0190
0.0225
0.0088
0.0112
0.0115
0.0125
0.0130
0.0265
0.0059
0.0032
0.0026
53
0.1697
0.1419
0.1124
0.0585
0.0609
0.1017
0.0783
0.0185
0.0138
0.0686
0.0748
0.0517
0.0129
0.0031
0.0064
0.0067
0.0000
0.0032
0.0024
0.0000
0.0002
0.0101
0.0006
0.0011
0.0005
0.0000
0.0021
0.0000
0.0000
0.0000
0.0000
54
0.0737
0.0456
0.0739
0.0487
0.0605
0.0608
0.0441
0.0408
0.0320
0.0442
0.0602
0.0563
0.0574
0.0447
0.0501
0.0531
0.0363
0.0221
0.0127
0.0017
0.0138
0.0191
0.0267
0.0169
0.0045
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
61&41
0.0843
0.0672
0.0576
0.0506
0.0693
0.0562
0.0488
0.0379
0.0453
0.0535
0.0560
0.0550
0.0597
0.0528
0.0487
0.0400
0.0167
0.0147
0.0133
0.0180
0.0112
0.0090
0.0099
0.0038
0.0048
0.0048
0.0040
0.0036
0.0026
0.0006
0.0000
62
0.1668
0.1331
0.1140
0.1140
0.1186
0.0804
0.0643
0.0403
0.0304
0.0315
0.0320
0.0290
0.0080
0.0087
0.0115
0.0062
0.0013
0.0011
0.0035
0.0012
0.0010
0.0006
0.0010
0.0000
0.0009
0.0003
0.0003
0.0000
0.0002
0.0000
0.0000
61
-------
6. Vehicle Characteristics that Vary by Age
Three fields comprise the SourceTypeAge table in MOVES2010: SurvivalRate,
Relative MAR, and FunctioningACFraction. The first two are described below, including data
sources and some relevant data points used in the model. The third is described in Section 15
with other air conditioning inputs.
6.1. SurvivalRate
The SurvivalRate field describes the fraction of vehicles of a given sourcetype and Age
that remain on the road one year to the next. SurvivalRate is used in the Total Activity
Generator in the calculation of source type populations by age in calendar years after the base
year. In MOVES, a separate SurvivalRate is applied to each age in each sourcetype fleet. The
SurvivalRates in MOVES are used for all model years in a sourcetype in all calendar years.
SurvivalRates for Motorcycles were calculated based on a smoothed curve of retail sales
f 41
and 2008 registration data as described in a contractor report.'
Survival rates for Passenger Cars, Passenger Trucks and Light Commercial Trucks came
from NHTSA's survivability Table 3 and Table 4.44 These survival rates are based on a detailed
analysis of Polk vehicle registration data from 1977 to 2002. We modified these rates to fit
them into the MOVES format:
• NHTSA rates for Light Trucks were used for both MOVES Passenger Trucks
and MOVES Light Commercial Trucks.
• MOVES calculates emissions to age 30 for both cars and trucks, but NHSTA car
rates were available only to age 25, so we extrapolated car rates to age 30 using
the estimated survival rate equation in section 3.1 of the NHTSA report. When
converted to MOVES format, this caused a striking discontinuity at age 26
which we removed by interpolating between ages 25 and 27.
• According to the NHTSA methodology, NHTSA "age= 1" corresponds to
MOVES "ageid =2," so the survival fractions were shifted accordingly.
• Because MOVES requires survival rates for MOVES ages < 2, the survival rates
for age 0 and age 1 were interpolated using a linear interpolation and assuming
that the survival rate prior to age 0 is 1.
• NHTSA defines survival rate as the ratio of the number of vehicles remaining in
the fleet at a given year as compared to a base-line year. MOVES calculations
f For motorcycles, the survival rates in MOVES2010 were calculated relative to initial sales rather than previous
year population. This causes very aggressive scrappage and significantly reduces the MOVES2010 motorcycle
population for calendar years after the base year. This error does not impact county-level runs where the analysis
year and base year are the same. We fixed this error in MOVES20lOa.
62
-------
require a value that is the ratio of a given year to the previous year, so we
transformed the NHTSA rates to MOVES rates using this ratio.
Because MOVES ageid= 30 is intended to represent all ages 30-and-greater, the
survival rate for ageid=30 was set to 0.3. The MOVES algorithm eventually
transfers all vehicles to this age group and requires a low survival rate to assure that
the population of very old vehicles does not grow excessively. The actual survival
rates of these age 30+ vehicles is unknown.
• Quantitatively the formula used to derive the MOVES Survival rates was:
MOVES Survival Rate (ageid =0) = 1 - (1-NHTSA Survival Rate (age =2)/3)
MOVES Survival Rate (ageid =!) = !- (1- 2* NHTSA Survival Rate (age =2)/3)
MOVES Survival Rate (age = 2 through 29) =
NHTSA Survival Rate (age = n-1)/ NHTSA Survival Rate (age = n-2)
MOVES Survival Rate (age = 30) = 0.3
The data for all other sourcetypes came from the Transportation Energy Data Book We
used the Heavy-Duty rates for the 1980 model year (TEDB22, Table 6.11, same as TEDB26
Table 3.10). The 1990 model year rates were not used because they were significantly higher
than the other model years in the analysis (e.g. 45 percent survival rate for 30 year-old trucks),
and seemed unrealistically high. While limited data exists to confirm this judgment, a snapshot
of 5-year survival rates can be derived from VIUS 1992 and 1997 results for comparison.
According to VIUS, the average survival rate for model years 1988-1991 between the 1992 and
1997 surveys was 88 percent. The comparable survival rate for 1990 model year Heavy-Duty
vehicles from TEDB was 96 percent, while the rate for 1980 model year trucks was 91 percent.
This comparison lends credence to the decision that the 1980 model year survival rates are more
in line with available data.
TEDB does not include scrappage rates for GVWR 10,000-26,000 vehicles, so it was
necessary to apply the Heavy-Duty rates to predominantly Medium-Duty use types.
The TEDB survival rates were transformed into MOVES format in the same way as the
NHTSA rates. Survival rates for all "age 30" sourcetypes8 were set to 0.3. This is set to keep the
fraction of oldest vehicles from growing excessively.
SurvivalRates used in MOVES2010 are shown in Table 6-1.
8Except motorcycles, where in MOVES2010 we used the rates developed by our contractor. The 0.3 value was used
inMOVES2010a.
-------
Table 6-1. Survival Rate by Age and SourceType
Age
0
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
29
30
Motorcycles
MOVES2010
1.000
0.979
0.920
0.864
0.812
0.763
0.717
0.674
0.633
0.595
0.559
0.525
0.493
0.464
0.436
0.409
0.385
0.361
0.340
0.319
0.300
0.282
0.265
0.249
0.234
0.220
0.206
0.194
0.182
0.171
0.161
Motorcycles
MOVES2010a
1.000
0.979
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.940
0.300
Passenger
Cars
0.997
0.997
0.997
0.993
0.990
0.986
0.981
0.976
0.971
0.965
0.959
0.953
0.912
0.854
0.832
0.813
0.799
0.787
0.779
0.772
0.767
0.763
0.760
0.757
0.757
0.754
0.754
0.567
0.752
0.752
0.300
Passenger Trucks
Light Comm.
Trucks
0.991
0.991
0.991
0.986
0.981
0.976
0.970
0.964
0.958
0.952
0.946
0.940
0.935
0.929
0.913
0.908
0.903
0.898
0.894
0.891
0.888
0.885
0.883
0.880
0.879
0.877
0.875
0.875
0.873
0.872
0.300
All Other
SourceTypes
1.000
1.000
1.000
1.000
0.990
0.980
0.980
0.970
0.970
0.970
0.960
0.960
0.950
0.950
0.950
0.940
0.940
0.930
0.930
0.920
0.920
0.920
0.910
0.910
0.910
0.900
0.900
0.900
0.890
0.890
0.300
6.2. Relative Mileage Accumulation Rate
The Relative Mileage Accumulation Rate (Relative MAR) is listed for each MOVES
sourcetype and Age. The Relative MAR is computed as the annual MAR divided by the highest
MAR within the HPMS vehicle class. This allows MOVES to maintain a constant MAR ratio
between ages and between the sourcetypes that make up each HPMS vehicle type even as
vehicle populations and the total VMT for an HPMS vehicle class changes over time. Table 1-2
(previous) lists the groupings of the MOVES sourcetypes within the six HPMS Vehicle Classes.
The following discussion refers to the Source Type ID numbers found in this table.
For many sourcetypes, the annual MARs were derived from the MARs developed for
MOBILE6. These were mapped from the MOBILE6 Vehicle Classes to the MOVES
sourcetypes. We then used regression to smooth these initial MARs and to extend the MARs
from 25 to 30 ages.
64
-------
The MAR values described below were then used to calculate the "relative MARs" by
computing the ratio of the value for each sourcetype and age to the highest value within the
HPMS class. For example, all of the bus values are relative to each other. The relative MARs
for all sourcetypes are illustrated in Figure 6-1
6.2.1. Motorcycles
The MARs for motorcycles (category 11) were updated by a contractor based on the
model years and odometer readings listed in motorcycle advertisements. A stratified sample of
about 1,500 ads were examined. A modified Weibull curve was fit to the data to develop the
relative mileage accumulation rates used in MOVES.45
6.2.2. Passenger Cars, Passenger Trucks and Light Commerical Trucks
The MARs for passenger cars, passenger trucks and light commercial trucks (categories
21, 31 & 32) were taken from the NHTSA report on survivability and mileage schedules.46 In
the NHTSA analysis, annual mileage by age was determined for cars and for trucks using data
from the National Household Travel Survey. In this NHTSA analysis, vehicles that were less
than one year old at the time of the survey were classified as "age 1", etc. NHTSA used cubic
regression to smooth the VMT by age estimates.
We used NHTSA's regression coefficients to extrapolate mileage to ages not covered by
the report. We divided each age's mileage by the NHTSA "age 1" mileage to determine a
relative MAR. For consistency with MOVES age categories, we then shifted the relative MARs
such that the NHTSA "agel" ratio was used for MOVES age 0, etc. We used NHTSA's light
truck VMT to determine relative MARS for both passenger trucks and light commercial trucks.
6.2.3. Other Trucks
The initial MARs for truck categories 51, 52, 53, 61, and 62 in MOVES were calculated
based on weighting fractions assigned to the MOBILE6 truck classes. We used VIUS 1997
values for Gross Vehicle Weight (PKGVW) to determine weighting fractions by model year.
To separate Light-Duty Trucks 1 and Light-Duty Trucks 2, which are distinguished by Loaded
Vehicle Weights, we used information from the Oak Ridge National Lab Light Duty Vehicle
database. To separate Class 2a and 2b trucks, we used information from the Oak Ridge National
Laboratory Report by Davis and Truitt.47 The initial MARs for the MOVES truck categories
were then calculated as the product of the weighting fractions and the MARs from MOBILE6.
In order to smooth the data and to extend the MARs from the 25 ages in MOBILE6 to the 30
ages in MOVES, we used statistical regression to determine the curves that best fit the data for
years starting in 1997 and going back to 1973 (ages 1 to 25).
6.2.4. Buses
For the School Buses (category 43) the initial MARs were taken from the MOBILE6
value for diesel school buses (HDDBS). As in MOBILE6, the same annual MAR was used for
each age. The MOBILE6 value of 9,939 miles per year came from the 1997 School Bus Fleet
Fact Book.
65
-------
For Transit Buses (category 42), the initial MARs were taken from the MOBILE6 values
for diesel transit buses (HDDBT). This mileage data was obtained from the 1994 Federal
Transportation Administration survey of transit agencies. 48 The MOBILE6 equation was also
applied to ages 26 through 30.
For Intercity Buses (category 41), the initial MARs were taken from Motorcoach Census
2000.49 The data did not distinguish vehicle age, so the same MAR was used for each age. This
MAR is high compared to transit and school buses.
6.2.5. Motor Homes
For motor homes (category 54), the initial MARs were taken from an independent
research study50 conducted in October 2000 among members of the Good Sam Club. The
members are active recreation vehicle (RV) enthusiasts who own motor homes, trailers and
trucks. The average annual mileage was estimated to be 4,566 miles. The data did not distinguish
vehicle age, so the same MAR was used for each age.
Figure 6-1. Relative Mileage Accumulation Rates in MOVES2010
Relative Mileage Accumulation Rates bySourceType
1 XIKIKXIKXXIKXXIKXXXXXXXXXXXXXXXIKXXIK
-11
-21
31
32
-41
-42
-43
-51
-52
53
54
61
-62
66
-------
7. Vehicle-Specific-Power Characteristics by SourceType
The MOVES model calculates emissions by calculating a weighted average of emisisons
by operating mode. For running exhaust emissions, the operating modes are defined by Vehicle
Specific Power (VSP) or the related concept, Scaled Tractive Power (STP). Both VSP and STP
are calculated based on a vehicle's speed and acceleration. They differ in how they are scaled.
The VSP equation is used for light duty vehicles (sourcetypes 11-32) and the STP equation is
used for heavy-duty vehicles (sourcetypes 41-62).
The SourceUseType table describes the vehicle characteristics needed for the VSP and
STP calculations. In particular, this table lists average vehicle mass, fixed mass factor and three
average road load coefficients for each SourceType. These are averaged over all model years
and ages. The mass is listed in metric tons. The road load coefficients are a rolling term "A," a
rotatating term "B," and a drag term "C."
MOVES uses these coefficients to calculate VSP and STP for each source type according
to the equation:
•V.
where A, B, and C are the road load coefficients in units of (kiloWatt second)/(meter), (kiloWatt
second2)/(meter2, and (kiloWatt second3)/(meter3), respectively. The detenominator term, m is
the fixed mass factor for the sourcetype in metric tons, g is the acceleration due to gravity (9.8
meter/second2), v is the vehicle speed in meter/second, a is the vehicle acceleration in
meter/second2, and sin 6 is the (fractional) road grade.
The values in the SourceUseType table were averaged from values in the Mobile Source
Observation Database (MSOD). The values were weighted using the age and sourcebin
distributions described elsewhere in this report. In particular, the average values were computed
using the equation:
weightedvalue = -
I •
i =1 , total # of ages
A-
^\cCj -unweightedvalue |
7=1, total # of sourcebins
IX
j =1, total # of sourcebins j
i =1, total # of ages
where the "unweighted value" was either the vehicle mid-point mass or one of the three different
road load coefficients determined from the road load-vehicle mass relations described below: Oj
were the sourceBinActivityFractions in the MOVES database and (3; were the ageFractions in the
MOVES database. Age fractions were matched to model years for calendar year 1999 (i.e.,
Model Year 1999 corresponds to vehicle agelD of 0; Model Year 1969 corresponds to agelD of
30.) Only sourcebins and ages with vehicles in the MSOD were used in these weightings. Thus,
the "total number of sourcebins" in the MSOD and "total number of ages" in the MSOD were
used to normalize the results.
-------
7.1. SourceMass and Fixed Mass Factor
MOVES2010 includes both a SourceMass and a fixedMassFactor. The SourceMass represents
the average weight of a given sourcetype. One can model changes in average weight of a
sourcetype by changing this factor. The fixedMassFactor is the value that was used to calculate
the relevant power measure used to define operating modes for running emissions, that is, VSP
or STP. Note for motorcycles, cars, and light trucks, the default database is populated with a
fixedMassFactor that was calculated as the mean for that sourcetype. This differs from the factor
used in the actual calculation of the emission rates which was the measured weight for each
vehicle. For other sourcetypes, the fixedMassFactor represents a scaling factor to bring the
numerical range of tractive power into the same numerical range as the VSP values when
assigning operating modes, hence scaled tractive power (STP). The fixedMassFactor of 17.1 is
roughly equivalent to the average running weight (in metric tons) of all heavy-duty vehicles. It
was also used in the development of the heavy-duty emission rates.
The SourceMass was computed as the weighted average of the "mid-point" mass for the
Weight Class associated with each sourcebin. Sourcebins not represented in the MSOD were
excluded.
68
-------
Table 7-1. MOVES Weight Classes
Weight
ClassID
0
20
25
30
35
40
45
50
60
70
80
90
100
140
160
195
260
330
400
500
600
800
1000
1300
9999
5
7
9
Weight Class Name
Doesn't Matter
weight < 2000 pounds
2000 pounds <= weight < 2500 pounds
2500 pounds <= weight < 3000 pounds
3000 pounds <= weight < 3500 pounds
3500 pounds <= weight < 4000 pounds
4000 pounds <= weight < 4500 pounds
4500 pounds <= weight < 5000 pounds
5000 pounds <= weight < 6000 pounds
6000 pounds <= weight < 7000 pounds
7000 pounds <= weight < 8000 pounds
8000 pounds <= weight < 9000 pounds
9000 pounds <= weight < 10000 pounds
10000 pounds <= weight < 14000 pounds
14000 pounds <= weight < 16000 pounds
16000 pounds <= weight < 19500 pounds
19500 pounds <= weight < 26000 pounds
26000 pounds <= weight < 33000 pounds
33000 pounds <= weight < 40000 pounds
40000 pounds <= weight < 50000 pounds
50000 pounds <= weight < 60000 pounds
60000 pounds <= weight < 80000 pounds
80000 pounds <= weight < 100000 pounds
100000 pounds <= weight < 130000 pounds
130000 pounds <= weight
weight < 500 pounds (for MCs)
500 pounds <= weight < 700 pounds (for MCs)
700 pounds <= weight (for MCs)
Midpoint
Weight
[NULL]
1000
2250
2750
3250
3750
4250
4750
5500
6500
7500
8500
9500
12000
15000
17750
22750
29500
36500
45000
55000
70000
90000
115000
130000
350
600
700
7.2. Road Load Coefficients
The information available on road load coefficients varied by regulatory class.
Motorcycle road load coefficients are typically parameterized with mass dependent A and
C terms which take into account rolling resistance and aerodynamic drag. Parameters adopted
here are from the United Nations report51'52:
A = 0.088M and
C=0.26+ 1.94xlO-4M
where M is the inertial mass of the motorcycle (SourceMass) and driver and has units of
metric tons.
For vehicles with a weight of 8500 Ibs or less, the road load coefficients were derived
from the track road load horspower (TRLHP@50mph) recorded in the MSOD.53 The calculations
applied the following empirical equations:54
69
-------
A = 0.7457*(0.35/50*0.447)
B = 0.7457*(0.10/(50*0.447)2)
C = 0.7457*(0.55/(50*0.447)3)
* TRLHP@50mph
* TRLHP@50mph
* TRLHP@50mph
Where 0.447 is a conversion from mile per hour to meters per second.
For the heavier vehicles, no road load parameters were available in the MSOD. Instead
EPA used the relationships of road load coefficent to vehicle mass from a study done by V.A.
Petrushov,55 as shown in Table 7-2. The mid-point mass for the sourcebin was used as the
vehicle mass.
Table 7-2. Road Load Coefficients for Heavy-Duty Trucks, Buses, and Motor
Homes
A(kW*s/m)/
M(metric ton)
B(kW*s2/m2)/
M( metric ton)
C(kW*s3/m3)
/M(metric ton)
8500 to 14000 Ibs
(3.855 to 6.350
metric ton)
0.0996
0
3.40 x 10'4
(mass is the average
mass of the weight
category)
1.47 c
+ j.22 *-±U
mass(kg)
14000 to 33000 Ibs
(6.350 to 14.968
metric ton)
0.0875
0
1.97 xlO'4
(mass is the average
mass of the weight
category)
1.93 _5
+ 5.90 <-40
mass(kg)
>33000 Ibs
(>14.968 metric
ton)
0.0661
0
1.79 xlO'4
(mass is the
average mass of the
weight category)
2.89 _ 5
+ 4.21 <-40
mass(kg)
Buses and
Motor Homes
0.0643
0
3.22 _5
mass {kg)
In both cases, values of A, B, and C were computed for each SourceBin-associated
vehicle in the MSOD and a weighted average was computed as described above. The final
SourceMass, FixedMassFactor and road load coefficients for all sourcetypes are listed in Table
7-3.
70
-------
Table 7-3. SourceUseType Characteristics
Source
TypelD
11
21
31
32
41
42
43
51
52
53
54
61
62
HPMS
Vtype ID
10
20
30
30
40
40
40
50
50
50
50
60
60
SourceType
Name
Motorcycle
Passenger Car
Passenger Truck
Light
Commercial
Truck
Intercity Bus
Transit Bus
School Bus
Refuse Truck
Single Unit Short-
haul Truck
Single Unit Long-
haul Truck
Motor Home
Combination
Short-haul Truck
Combination
Long-haul Truck
Rolling
TermA
(kW-s/m)
0.0251
0.156461
0.22112
0.235008
1.29515
1.0944
0.746718
1.41705
0.561933
0.498699
0.617371
1.96354
2.08126
Rotating
TermB
(kW-s2/m2)
0
0.002002
0.002838
0.003039
0
0
0
0
0
0
0
0
0
Drag
TermC
(kW-s3/m3)
0.000315
0.000493
0.000698
0.000748
0.003715
0.003587
0.002176
0.003572
0.001603
0.001474
0.002105
0.004031
0.004188
Source
Mass (metric
tons)
0.285
1.4788
1.86686
2.05979
19.5937
16.556
9.06989
20.6845
7.64159
6.25047
6.73483
29.3275
31.4038
FixedMass
Factor
(metric
tons)
0.285
1.4788
1.86686
2.05979
17.1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
17.1
8. VMT by Year and Vehicle Type
For national level calculations, MOVES uses national VMT by vehicle type to determine
source operating hours. The model's Total Activity Generator takes a default VMT for a base
year and uses growth factors to estimate VMT in later analysis years. Three fields comprise
HPMSVTypeYear in MOVES2010: HPMSBaseYearVMT, BaseYearOffNetVMT, and
VMTGrowthF actor.
8.1. HPMSBaseYearVMT
The HPMSBaseYearVMT field stores the base year VMT for each HPMS Vehicle Type.
This VMT was calculated from the FHWA VM-1 and VM-2 tables by summing over HPMS
Vehicle Class.
The resulting values for 1999 and 1990 by HPMS Vehicle Class are listed in Table 8-1.
-------
Table 8-1. 1999 VMT by HPMS Vehicle Class
HPMS Vehicle Class
Motorcycles
Passenger Cars
Other 2 axle - 4 tire vehicles
Buses
Single unit trucks
Combination trucks
1990 VMT
9,557,000,000
1,408,270,000,000
574,571,000,000
5,726,000,000
51,901,000,000
94,341,000,000
1999 VMT
10,579,600,000
1,568,640,000,000
900,735,000,000
7,657,000,000
70,273,700,000
132,358,000,000
8.2. BaseYearOffNetVMT
Off Network VMT refers to the portion of activity that is not included in travel demand
model networks or any VMT that is not otherwise reflected in the other four roadtypes. This
field is provided in case it is useful for modeling local areas. However, the reported HPMS
VMT values, used to calculate the national averages discussed here, are intended to include all
VMT. Thus, for MOVES2010 national defaults, the BaseYearOffNetVMT is zero for all vehicle
types.
8.3. VMTGrowthFactor
The VMTGrowthFactor field stores a multiplicative factor indicating changes in total
vehicle miles for calendar years after the base year. Total VMT data are reported according the
HPMS vehicle classes discussed previously, i.e. passenger car, other 2-axle / 4-tire vehicle,
single-unit truck, combination truck, bus and motorcycle. VMTGrowthFactor is expressed
relative to the previous year's VMT; for example, 1 means no change from previous year VMT,
1.02 means a two percent increase in VMT, and 0.98 means a two percent decrease in VMT.
VMTGrowthFactor is used in the Total Activity Generator calculation of VMT for
calendar years after the base year, meaning calendar years 2000 through 2050 in MOVES2010.
It is important to note that VMTGrowthFactor is a key component for estimates of future activity
in MOVES, because the level of total activity in future years for many emission processes is
derived from projections of total VMT. For these processes, projections of future populations
based on sales growth, survival rates, etc. are only used to allocate total VMT.
For motorcycles, default growth factors for years 2000 through 2008 were derived from
Highway Statistics Table VM-1. Growth factors for years 2009-and-later were borrowed from a
previous (AEO2006-based) estimate for passenger cars.
For passenger cars, passenger trucks and light commercial trucks, growth factors for
historical years 2000 through 2007 were derived from estimates of total VMT data as reported in
the Transportation Energy Data Book. For these years the growth factors are simply total VMT
for the applicable vehicle class for the calendar year divided by total VMT from the previous
year. For 2008-2020, we used values from AEO2009. Unlike TEDB, AEO does not break VMT
out by cars and trucks. Consequently, EPA developed a formula to apportion the projected AEO
light duty VMT between cars and trucks.
72
-------
VMT = Previous Year VMT x (1+percent change in class) x per-vehicle growth
rate
Where
Previous year VMT =
Percent change in Class =
Per vehicle growth rate =
the VMT of the previous year, starting with TEDB in
2007
the percent change in car or truck population relative
to the previous year, derived by applying scrappage
and sales to the previous year fleet.
a constant growth rate that is used to reflect the
increase in per-vehicle annual VMT that is
commonly observed.
The per vehicle growth rate was kept identical and constant between cars and trucks during the
years 2008-2030. The per-vehicle growth rate was raised in the years 2030 to 2050 so that the
total annual growth in light duty VMT was consistent with the average for the time period (1.7%
annual growth in total LD VMT). 0.3% annual growth in per vehicle VMT was assumed in
2008-2030, while 1% change in per-vehicle growth was assumed in the years 2030-2050
In MOVES2010, the default VMTGrowthFactor estimates for other sourcetypes were
taken from FHWA Highway Statistics Table VM-1 for 2000 through 2004, and from AEO2006
for years 2005-and-later. VMT projections are provided for total Medium-Duty and total Heavy-
Duty in AEO2004 Supplemental Table 55. The growth factors derived from the AEO2006
Medium-Duty VMT estimates were applied to the single-unit truck and bus HPMS vehicle
classes. The growth factors derived from the AEO2006 Heavy-Duty VMT estimates were
applied to the combination truck vehicle class.
In MOVES2010a, the default VMTGrowthF actor estimates for the heavy sourcetypes
were taken from VM-1 for 2000 through 2008 and values from AEO2009 Table 67 were used for
years 2009-and-later.
Table 8-2. VMT Growth Factors in MOVES2010
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Motorcycles
0.990
0.909
1.002
0.999
1.061
1.064
1.119
1.130
1.131
1.012
1.012
Passenger
Cars
.020
.018
.019
.008
.017
.005
0.990
0.988
0.988
0.983
0.993
Passenger
& Light
Comm.
Trucks
1.025
1.022
1.024
1.019
1.044
1.014
1.040
1.027
1.006
0.993
1.008
Buses
0.992
0.920
0.968
0.991
0.979
0.998
1.007
1.016
1.013
1.018
1.021
Single
Unit
Trucks
1.004
1.025
1.048
1.025
1.043
0.998
1.007
1.016
1.013
1.018
1.021
Combination
Trucks
.021
.003
.015
.010
.037
.022
.034
.033
.025
.025
.026
73
-------
Year
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
Motorcycles
1.007
1.006
1.006
1.007
1.007
1.007
1.008
1.009
1.009
1.008
1.008
1.009
.009
.009
.009
.010
.010
.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
.010
.010
.010
.010
.010
.010
.010
.010
1.010
1.010
Passenger
Cars
.002
.009
.014
.019
.028
.028
.027
.029
.031
.032
.031
1.031
1.030
1.030
1.029
1.027
1.025
1.023
.021
.020
.026
.025
.023
.023
.022
.021
.021
.020
.020
1.019
1.019
1.019
1.018
1.018
1.018
1.018
1.018
1.018
1.018
1.018
Passenger
& Light
Comm.
Trucks
1.011
1.011
1.012
1.010
1.008
1.005
1.003
1.001
1.000
0.998
0.997
0.997
0.997
0.997
0.998
0.999
0.999
0.999
0.999
1.000
1.007
1.008
1.009
1.010
1.010
1.011
1.012
1.014
.016
.015
.015
.015
.015
.016
.016
.016
.017
1.017
1.017
1.017
Buses
1.025
1.023
1.022
1.023
1.024
1.025
1.026
1.026
1.025
.025
.026
.027
.027
.027
.027
.028
.027
.027
1.027
1.026
1.026
1.026
1.026
1.026
1.026
1.026
.026
.026
.026
.026
.026
.026
.026
.026
.026
.026
.026
1.026
1.026
1.026
Single
Unit
Trucks
1.025
1.023
1.022
1.023
1.024
1.025
1.026
1.026
.025
.025
.026
.027
.027
.027
.027
.028
.027
.027
1.027
1.026
1.026
1.026
1.026
1.026
1.026
.026
.026
.026
.026
.026
.026
.026
.026
.026
.026
.026
.026
1.026
1.026
1.026
Combination
Trucks
.025
.023
.022
.022
.023
.023
.025
.025
.021
.020
.020
.021
.021
.022
.023
.024
.024
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
.023
74
-------
Table 8-3. VMT Growth Factors in MOVES2010a
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Motorcycles
0.990
0.909
1.002
0.999
1.061
1.064
1.119
1.130
1.131
1.012
1.012
1.007
1.006
1.006
1.007
1.007
1.007
1.008
1.009
1.009
1.008
1.008
1.009
1.009
1.009
1.009
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
Passenger
Cars
1.020
1.018
1.019
1.008
1.017
1.005
0.990
0.988
0.988
0.983
0.993
1.002
1.009
1.014
1.019
1.028
1.028
1.027
1.029
1.031
1.032
1.031
1.031
1.030
1.030
1.029
1.027
1.025
1.023
1.021
1.020
1.026
1.025
1.023
1.023
1.022
1.021
1.021
1.020
1.020
1.019
Passenger
& Light
Comm.
Trucks
1.025
1.022
1.024
1.019
1.044
1.014
1.040
1.027
1.006
0.993
1.008
1.011
1.011
1.012
1.010
1.008
1.005
1.003
1.001
1.000
0.998
0.997
0.997
0.997
0.997
0.998
0.999
0.999
0.999
0.999
1.000
1.007
1.008
1.009
1.010
1.010
1.011
1.012
1.014
1.016
1.015
Buses
0.992
0.920
0.968
0.991
0.979
1.026
1.002
0.998
1.019
0.908
1.041
1.086
1.071
1.048
1.036
1.036
1.036
1.034
1.032
1.030
1.029
1.024
1.022
1.026
1.028
1.028
1.028
1.028
1.027
1.028
1.028
1.027
1.027
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
Single
Unit
Trucks
1.004
1.025
1.048
1.025
1.043
1.001
1.023
1.021
1.024
0.908
1.041
1.086
1.071
1.048
1.036
1.036
1.036
1.034
1.032
1.030
1.029
1.024
1.022
1.026
1.028
1.028
1.028
1.028
1.027
1.028
1.028
1.027
1.027
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
Combination
Trucks
1.021
1.003
1.015
1.010
1.037
1.012
0.991
1.016
0.989
0.891
0.998
1.043
1.046
1.034
1.024
1.018
1.017
1.018
1.022
1.023
1.020
1.013
1.011
1.014
1.017
1.017
1.017
1.015
1.013
1.012
1.013
1.012
1.011
1.013
1.013
1.015
1.015
1.015
1.015
1.015
1.015
75
-------
Year
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
Motorcycles
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
1.010
Passenger
Cars
1.019
1.019
1.018
1.018
1.018
1.018
1.018
1.018
1.018
1.018
Passenger
& Light
Comm.
Trucks
1.015
1.015
1.015
1.016
1.016
1.016
1.017
1.017
1.017
1.017
Buses
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
Single
Unit
Trucks
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
1.028
Combination
Trucks
1.015
1.015
1.015
1.015
1.015
1.015
1.015
1.015
1.015
1.015
Note that MOVES uses a single national growth factor by vehicle class, thus it does not
capture variations in growth across roadtypes and counties. Therefore, for local calculations,
locally available data will often better represent local VMT.
9. Roadtypes, VMT Distribution among Roadtypes, and
Mappings to SCC
MOVES will calculate emissions separately for each road type and for "off-network"
activity. The road type codes used in MOVES are listed in Table 9-1. The MOVES road types
are aggregations of the HPMS functional facility types that are also used for reporting in EPA
Source Classification Codes (SCCs).
Table 9-1. Road Type Codes in MOVES
RoadTypelD
1
2
o
5
4
5
Description
Off Network
Rural Restricted
Access
Rural Unrestricted
Access
Urban Restricted
Access
Urban Unrestricted
Access
HPMS functional Types
Off Network
Rural Interstate
Rural Principal Arterial, Minor
Arterial, Major Collector, Minor
Collector & Local
Urban Interstate & Urban
Freeway/Expressway
Urban Principal Arterial, Minor
Arterial, Collector & Local
SCCRoadTypelD
1
11
13, 15, 17, 19,21
23,25
27,29,31,33
The number of default roadtypes in MOVES was limited to reduce database size and to
help improve model performance. The urban/rural distinction is used primarily for national level
calculations. It allows different default speed distributions in urban and rural settings. Of
course, finer distinctions are possible. Users with more detailed information on speeds and
-------
acceleration patterns may choose to create their own additional roadtypes or may run MOVES at
project level where emissions can be calculated for individual links.
9.1. RoadTypeVMTFraction
For each sourcetype, the RoadTypeVMTFraction field stores the fraction of total VMT
for each vehicle class that is traveled on each of the 5 roadway types. For MOVES2010, we
used data from 1999 FHWA Highway Statistics, Tables VM-1 and VM-2. VM-1 provides detail
on VMT by vehicle type; VM-2 provides detail by HPMS functional type. At the time of the
analysis, VM-1 (October 2000) had not been updated, but VM-2 was updated in January 2002.
We used the total values from VM-2 to distribute VMT by HPMS facility type and allocated this
VMT to vehicle class in proportion to the values in VM-1. We then calculated facility type
VMT fractions for each HPMS Vehicle Type. We later aggregated the values to the five
MOVES road types and mapped from HPMS Vehicle Type to MOVES Sourcetype.
The FHWA Highway Statistics is currently considered the best available source for
national information regarding vehicle miles traveled. However, there are problems and
constraints associated with using the (mostly) state-reported data in Highway Statistics. In many
cases, locally derived VMT data may be more accurate when modeling local areas.
The VMT distributions in Table 9-2 assume that all VMT reported by HPMS is
accumulated on one of the 12 HPMS roadway types and thus one of the four "on-network"
MOVES roadtypes. No VMT is currently assigned to the "off-network" category in the national
defaults. See the discussion of BaseYearOffNetVMT in Section 8.2.
Table 9-2. Sourcetype VMT distribution among Road Types
RoadType ID
1
2
3
4
5
Total
Road type
Description
Off Network
Rural Restricted
Access
Rural Unrestricted
Access
Urban Restricted
Access
Urban
Unrestricted
Access
Motorcycles
0.0000
0.1040
0.3161
0.2177
0.3623
1.0000
Passenger
Cars
0.0000
0.0834
0.2891
0.2097
0.4178
1.0000
Other 2axle
- 4tire
vehicles
0.0000
0.0846
0.3055
0.2031
0.4068
1.0000
Buses
0.0000
0.1268
0.4821
0.1385
0.2526
1.0000
Single
unit
trucks
0.0000
0.1149
0.3972
0.1715
0.3165
1.0000
Combination
trucks
0.0000
0.3247
0.2941
0.2075
0.1737
1.0000
We are currently assuming identical VMT distributions for all sourcetypes within an
HPMS Vehicle Type. However the MOVES model is designed to allow roadway type allocation
by sourcetype and one would expect the different sourcetypes to have different roadway type
allocations. For example, the long-haul trucks generally would have a greater fraction of travel
on rural restricted access roadways than the short-haul trucks. While national data to quantify
77
-------
these distinctions is not currently available, users may find information available at the local
level to apply different distibutions at the source type level.
9.2. SCCRoadTypeDistribution
Each SCC includes a suffix that indicates the HPMS Facility Class on which the
emissions occur. Because MOVES calculations are done for MOVES roadtypes, the
SCCRoadTypeFraction is needed to allocate emissions on each MOVES roadtype to the
appropriate SCCRoadTypes.
Table 9-3. SCC RoadTypes
SCCRoadTypelD
11
13
15
17
19
21
23
25
27
29
31
33
1
SCCRoadTypeDesc
Rural Interstate
Rural Principal Arterial
Rural Minor Arterial
Rural Major Collector
Rural Minor Collector
Rural Local
Urban Interstate
Urban Freeway/Expressway
Urban Principal Arterial
Urban Minor Arterial
Urban Collector
Urban Local
Off-Network
Because roadtype distributions vary geographically, the mapping of MOVES roadTypes
to SCCRoadTypes varies by zone (in this case, county). For SCCRoadTypeDistribution we
determined the proportion of hours of operation on a given MOVES roadtype within a county
that occurred on each SCCRoadType. Hours of operation were estimated by dividing the 1999
National Emission Inventory (NET) VMT by the 1999 NEI average speed. Both measures were
documented by Pechan & Associates.56 The NEI VMT estimates are based on the Highway
Performance Monitoring System (HPMS) data collected by the Federal Highway
Administration57 for use in transportation planning and vehicle type breakdowns from the EPA
MOBILE6 Emission Factor model.58 The VMT estimates were obtained from the NMEVI
database for each county and HPMS facility type. The average speed estimates are taken from
Table 8 of the NEI documentation.
The SCCRoadType fractions were calculated using the following formula, where i refers
to the county, j refers to the MOVES roadtype, k refers to the SCCRoadType within a MOVES
road type, and m refers to the VMT for each source type.
SCCRoadTypeFraction(i,j,k) = Sum(j,j,k)( VMT(k,m)/ Average Speed(k,m)) /
Sum(i,j)((VMT(k,m)/AverageSpeed(k,m))
-------
In cases where a county had no VMT for a given roadtype, the average values were used.
The SCCRoadTypeFraction for OffNetwork travel was set to 1 (mapping all "off-network"
emissions to this new roadtype. The SCCRoadType fractions for each roadway type will sum to
one for each county. Although the data is from 1999 calendar year estimates, the same
allocations will be used for all calendar years.
10. Average Speed Distributions
Average speed is used in MOVES to convert VMT inputs into the Source Hours
Operating (SHO) units that MOVES uses for internal calculations. Also, MOVES2010 uses
average speeds to select appropriate driving cycles, which are then used to calculate exhaust
running operating mode distributions at the national, county (and sometimes project) level.
Instead of using a single average speed in these tasks, MOVES2010 uses a distribution of
average speeds. The AvgSpeedDistribution table lists the default fraction of driving time for
each sourcetype, Road Type, Day, Hour, in each average speed bin. The fractions sum to one for
each combination of sourcetype, Road Type, Day, and Hour.
For MOVES2010, the average speed fractions for urban roads were derived from the
default speed distributions (SVMT) in MOBILE659. These fractions do not vary by vehicle type.
The MOBILE6 speed fractions were adapted to MOVES by converting the fraction of miles
travelled to the fraction of time used, and by mapping from the MOBILE6 road types to the
MOVESroad types, with the MOBILE6 "freeway" values mapped to the MOVES "urban
restricted" roadtype and the MOBILE6 "arterial" values mapped to the MOVES "urban
unrestricted" roadtype. The time fractions were normalized to sum to one for each hour of the
day over all 14 speed bins. The values for the off-network roadway type were set to null. The
detailed distributions are available in the MOVES default database.
For rural road average speed distributions, we relied on light-duty driving data collected
in California under studies performed for the California Department of Transportation (Caltrans).
Under these Caltrans driving studies, instrumented "chase cars" were equipped with laser
rangefmders mounted behind the front grill of each chase car. The studies were performed in the
Sacramento area, the San Francisco Bay area and the San Joaquin Valley. Another driving study
was also conducted in the South Coast (i.e., Los Angeles Basin), but was conducted entirely in
urbanized areas. Thus, this data was not used for the rural area analysis. A contractor report
describes the analysis done to develop speed distributions from the Caltrans datasets.60 In post-
processing, the driving data was grouped by HPMS functional class. The urban area travel in
these datasets was discarded for this analysis. The average speed was calculated over each one-
way driving traverse of a roadway link. Once the average speed was calculated for each link
traverse, the VMT was allocated into one of sixteen speed bins defined by EPA for the purpose
of calculating speed distributions for use in MOVES. The MOVES speed bins are shown in
Table 10-1.
79
-------
Table 10-1. MOVES Speed Bin Categories.
Bin
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Average Speed (mph)
2.5
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
Average Speed Range (mph)
speed < 2.5 mph
2.5 mph <= speed < 7.5 mph
7.5 mph <= speed < 12.5 mph
12.5 mph <= speed < 17.5 mph
17.5 mph <= speed < 22.5 mph
22.5 mph <= speed < 27.5 mph
27.5 mph <= speed < 32.5 mph
32.5 mph <= speed < 37.5 mph
37.5 mph <= speed < 42.5 mph
42.5 mph <= speed < 47.5 mph
47.5 mph <= speed < 52.5 mph
52.5 mph <= speed < 57.5 mph
57.5 mph <= speed < 62.5 mph
62.5 mph <= speed < 67.5 mph
67.5 mph <= speed < 72.5 mph
72.5 mph <= speed
To import this information into MOVES, we started with the contractor-provided values of
"Time-weighted Distributions (% of time) of California Rural Chase Car Driving Data by
Average Link Speed for each HPMS Functional Class."61 These values were used directly for
the rural restricted access roadtype (2). For the MOVES rural unrestricted access roadtype, the
calculation required consolidating values on the five HPMS functional road classes to the single
MOVES roadtype. This was done separately for each HPMS Vehicle Class. For each vehicle
class, we used the roadtype VMT distribution (see preceding section) to calculate the fraction of
VMT on each road class. We then changed to a time-basis by calculating the average speed on
each road class, dividing by the average speed and re-nomalizing. We then computed a sum-
product of the speed bin fractions and the road class distributions to calculate the weighted-
average speed bin distribution for each vehicle class and assigned this distribution to each
sourcetype in the HPMS vehicle class.
Our calculations of default average speed distributions required a number of assumptions
and extrapolations. For both urban and rural road types, the same speed data was used for all
sourcetypes.h Also the existing data from the rural studies used in this analysis were collected
entirely in California. Using these California results to represent national rural speed
distributions implicitly assumes that average speeds on rural roadways, within each HPMS
functional class, do not significantly vary across the U.S. And the same rural speed distributions
were used for all hours of the day. Because of these extrapolations, local data on speed
h While the underlying speed data used in MOVES2010 does not vary by sourcetype, the speed distributions in
MOVES2010 do. This is because they were originally calculated on twelve roadtypes. When the roadtypes were
combined to four, the road type weighting used to calculate the new speed distributions varied by sourcetype,
leading, in some cases, to small variations in the associated speed distributions.
80
-------
distributions often will be more accurate than our national defaults.
National default speed distributions are available in the default database for each
roadtype, sourcetype and hourday, and are not provided here. However, for illustration, Figure
10-1 shows the speed distributions on different roadtypes for passenger cars for the time period
11 am. to noon on weekdays.
Figure 10-1 Example Speed Distribution by Roadtype
Speed Distribution by Roadtype Passenger Cars, Ham-Noon,
Weekdays
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Average Speed Bin
I Rural Restricted n Rural Unrestricted • Urban Restricted n Urban Unrestricted
8l
-------
11. Driving Schedule Tables
DriveSchedule refers to a second-by-second vehicle speed trajectory. A drive schedule
typically includes both driving and idling time. Drive schedules are used in MOVES to
determine the operating mode distribution for most MOVES running process emissions and for
energy consumption.
A key feature of MOVES is the capability to accommodate any number of drive
schedules to represent driving patterns across source type, roadway type and average speed. For
the national default case, MOVES2010 employs 47 drive schedules, mapped to specific source
types and roadway types. In brief, the average speed of a driving schedule is used to determine
the weighting of that schedule for a given roadtype and sourcetype, based on the average speed
distribution. For each speed bin in the speed distribution, the MOVES model selects the two
associated driving cycles with average speeds that bracket the speed bin's average speed. The
Vehicle Specific Power (VSP) distributions determined for each bracketing driving schedule are
averaged together, weighted by the proximity of the speed bin average speed to the driving
schedule average speeds. In this way, the VSP distribution of any roadtype's speed distribution
is determined from the available driving schedules. For more details, see the Operating Mode
Distribution Generator sections in the MOVES Software and Design Reference Manual.62 This
approach is, of course, imprecise. Users with more information about driving activity may
choose to model at the project level where users can enter specific driving cycles or operating
mode distributions.
MOVES stores drive schedule information in four database tables. DriveSchedule
provides the drive schedule name, identification number, and the average speed of the drive
schedule. DriveScheduleAssoc defines the set of schedules which are available for each
combination of source use type and road type. DriveScheduleSecond contains the second-by-
second vehicle trajectories for each schedule. In some cases the vehicle trajectories are not
contiguous; as detailed below, they may be formed from several unconnected microtrips.
RoadOpModeDistribution lists operating mode distributions used for ramps for each source
use type, road type and speed bin.
Tables 11-1 through 11-5 list the driving schedules used for different sourcetypes. The
freeway and non-freeway driving cycles are intended to cover most of the driving on these
respective roadtypes. However, some speed distributions for non-freeway roadtypes will include
average speeds faster than the fastest non-freeway cycles. The reverse will be true for some
freeway speed distributions. In these cases, the model will use appropriate average speed drive
schedules from a different roadtype. This mapping is appropriate since, when the average speed
is very low or very high, the roadtype has little impact on the driving pattern.
The driving schedule tables also include light-duty, medium-duty and heavy-duty ramp
driving schedules, but these are not used directly in MOVES2010. Instead, for inventory
calculations1 the ramp schedules were transformed into a set of driving cycles consistent with
1 When MOVES2010 is used to calculate "Emission Rates," ramps are not included in the rates, but "Inventory"
calculations in MOVES2010 use the ramp operating mode distribution that matches the average speed as calculated
from the average speed distribution. The ramp methodology for both inventory and rate calculations was revised in
MOVES2010a such that the emission rate calculations include ramp operating modes appropriate for each identified
speed bin, and the inventory calculations use a weighted average of the ramp operating mode distributions for all the
speeds in the average speed distribution for that sourcetype, roadtype, day and hour. See the MOVES Software
Design and Reference Manual for more information.
82
-------
connecting to and from a freeway with the given average speed. The cycles were then converted
to operating mode distributions, which are stored in RoadOpModeDistribution.
Table 11-1. Driving Cycles for Motorcycles, Cars, Passenger Cars and Light
Commercial Trucks
ID
101
1033
1043
1041
1021
1030
153
1029
1026
1020
1011
1025
1019
1024
1018
1017
1009
158
Cycle Name
LD Low Speed 1
Final FC14LOSF
Final FC19LOSAC
Final FC17LOSD
Final FC11LOSF
Final FC14LOSC
LD LOS E Freeway
Final FC14LOSB
Final FC12LOSD
Final FC11LOSE
Final FC02LOSDF
Final FC12LOSE
Final FC11LOSD
Final FC12LOSC
Final FC11LOSC
Final FC11LOSB
Final FC01LOSAF
LD High Speed Freeway 3
Average
Speed
2.5
8.7
15.7
18.6
20.6
25.4
30.5
31.0
20.6
46.1
49.1
46.1
58.8
63.7
64.4
66.4
73.8
76.0
Non-Freeway
Rural
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
Table 11-2,
Homes
Driving Cycles for Intercity Buses, Single-Unit Trucks and Motor
ID
201
202
203
204
205
206
251
252
253
254
255
Cycle Name
MD 5 mph Non-Freeway
MD lOmph Non-Freeway
MD 15mph Non-Freeway
MD 20mph Non-Freeway
MD 25mph Non-Freeway
MD 3 Omph Non-Freeway
MD 3 Omph Freeway
MD 40mph Freeway
MD 50mph Freeway
MD 60mph Freeway
MD High Speed Freeway
Average
Speed
4.6
10.7
15.6
20.8
24.5
31.5
34.4
44.5
55.4
60.4
72.8
Non-Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
83
-------
Table 11-3. Driving Cycles for Combination Trucks
ID
301
302
303
304
305
306
351
352
353
354
355
Cycle Name
HD 5 mph Non-Freeway
HD lOmph Non-Freeway
HD 15mph Non-Freeway
HD 20mph Non-Freeway
HD 25mph Non-Freeway
HD 3 Omph Non-Freeway
HD 3 Omph Freeway
HD 40mph Freeway
HD 50mph Freeway
HD 60mph Freeway
HD High Speed Freeway
Average
Speed
5.8
11.2
15.6
19.4
25.6
32.5
34.3
47.1
54.2
59.4
71.7
Non-Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
Table 11-4. Driving Cycles for Transit and School Buses
ID
201
202
401
203
204
205
402
206
251
252
403
253
254
255
Cycle Name
MD 5 mph Non-Freeway
MD 1 Omph Non-Freeway
Bus Low Speed Urban*
MD 15mph Non-Freeway
MD 20mph Non-Freeway
MD 25mph Non-Freeway
Bus 30 mph Flow*
MD 3 Omph Non-Freeway
MD 3 Omph Freeway
MD 40mph Freeway
Bus 45 mph Flow*
MD 50mph Freeway
MD 60mph Freeway
MD High Speed Freeway
Average
Speed
4.6
10.7
15
15.6
20.8
24.5
30
31.5
34.4
44.5
45
55.4
60.4
72.8
Non-Freeway
Rural
X
X
X
X
X
X
Urban
X
X
X
X
X
X
Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
* This speed represents average of traffic the bus is traveling in, not the average speed of the bus, which is
lower due to stops.
-------
Table 11-5. Driving Cycles for RefuseTrucks
ID
501
301
302
303
304
305
306
351
352
353
354
355
Cycle Name
Refuse Truck Urban
HD 5 mph Non-Freeway
HD lOmph Non-Freeway
HD 15mph Non-Freeway
HD 20mph Non-Freeway
HD 25mph Non-Freeway
HD 3 Omph Non-Freeway
HD 3 Omph Freeway
HD 40mph Freeway
HD 50mph Freeway
HD 60mph Freeway
HD High Speed Freeway
Average
Speed
2.2
5.8
11.2
15.6
19.4
25.6
32.5
34.3
47.1
54.2
59.4
71.7
Non-Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
Freeway
Rural
X
X
X
X
X
X
X
X
X
X
X
Urban
X
X
X
X
X
X
X
X
X
X
X
The default drive schedules listed in the tables above were developed from several
sources. "LD LOS E Freeway" and "High Speed 1" were retained from MOBILE6 and are
documented in report M6.SPD.001.63 "LD Low Speed 1" is a historic cycle used in the
development of speed corrections for MOBILES and is meant to represent extreme stop-and-go
"creep" driving. "High Speed 3" was developed for MOVES to represent very high speed
freeway driving. It is a 580-second segment of freeway driving from an in-use vehicle
instrumented as part of EPA's On-Board Emission Measurement "Shootout" program,64 with an
average speed of 76 mph and a maximum speed of 90 mph. In MOVES2010, other historical
cycles have been removed and replaced with 15 new light duty cycles developed by a contractor
based on urban and rural data collected in California in 2000 and 2004.65 The new cycles were
selected to best cover the range of roadtypes and average speeds that need to be modeled in
MOVES.
Medium-Duty and Heavy-Duty schedules were developed specifically for MOVES,
based on work performed for EPA by Eastern Research Group (ERG), Inc. and documented in
the report "Roadway-Specific Driving Schedules for Heavy-Duty Vehicles."66 ERG analyzed
data from 150 medium and heavy-duty vehicles instrumented to gather instantaneous speed and
GPS measurements. ERG segregated the driving into freeway and non-freeway driving for
medium and heavy-duty vehicles, and then further stratified vehicles trips according the pre-
defined ranges of average speed covering the range of vehicle operation. ERG characterized
representative driving within each speed range, using distributions of vehicle specific power
(VSP), speed and acceleration. Driving schedules were then developed for each speed bin by
creating combinations of idle-to-idle "microtrips" until the representative target metrics were
achieved. The schedules developed by ERG are, thus, not contiguous schedules which would be
run on a chassis dynamometer, but are made up of non-continguous "snippets" of driving meant
to represent target distributions. For use in MOVES, the highway heavy-duty schedules
developed by ERG were modified to isolate operation on freeway ramps. The segments of
freeway microtrips identified by ERG as taking place on on-and off-ramps were extracted and
used to create medium-duty and heavy-duty ramp schedules (299 and 399). Thus, the schedules
which represent on-freeway driving do not contain ramp operation. Another minor modification
-------
to the schedules for use in MOVES was made to the time field in order to signify, within a drive
schedule, when one microtrip ended and one began. The time field increments two seconds
instead of one when each new microtrip begins. This two second increment signifies that these
should not be regarded by the model as contiguous operation.
The two higher-speed transit bus driving cycles were developed based on Ann Arbor
Transit Authority buses instrumented in Ann Arbor.67 Non-contiguous snippets of driving were
used to develop cycles with the desired average speeds. The "Low Speed Urban" bus cycle is
the last 450 seconds of the standard New York Bus Driving Cycle. The Refuse Truck cycle
represents refuse truck driving with many stops and a maximum speed of 20 mph.
12. Temporal Distributions of VMT and Hourly Extended
Idle Activity
MOVES can estimate emissions for every hour of every day of the year. For national
scale runs ("macroscale") annual VMT estimates and extended idle time need to be allocated to
months, days, and hours.
A 1996 report from the Office of Highway Information Management (OHTM)68 describes
analysis of a sample of 5,000 continuous traffic counters distributed through the United States.
EPA obtained the data used in the report and used it to generate VMTinputs in the form needed
forMOVES2010.
The report does not specify VMT by sourcetype or Vehicle Type. Thus, we currently use
the same value for all sourcetypes. We hope to update this in future versions of MOVES,
perhaps using data from the U.S. Vehicle Travel Information System (VTRIS).
In MOVES, Extended Idle activity is calculated as proportional Source Hours Operating
(SHO) and thus is derived in the model from the VMT and speed distributions. However, the
proportions used in MOVES vary by hour of day, as decribed below.
The temporal distribution of start and evaporative emissions is described in Section 13.
12.1. MonthVMTFraction
For Month VMTFraction, we use the data from the OHEVI report, Figure 2.2.1 "Travel by
Month, 1970-1995," but modified to fit MOVES specifications.
The figure shows VMT/day, normalized to January=l. For MOVES, we need the
fraction of total VMT per month, with different values for leap year and non-leap year. We
computed the fractions using the report values and the number of days in each month.
86
-------
Table 12-1. MonthVMTFraction
Month
January
February
March
April
May
June
July
August
September
October
November
December
Normalized
VMT/day
1.0000
1.0560
1.1183
1.1636
1.1973
1.2480
1.2632
1.2784
1.1973
1.1838
1.1343
1.0975
MOVES
not Leap
Year
0.0731
0.0697
0.0817
0.0823
0.0875
0.0883
0.0923
0.0934
0.0847
0.0865
0.0802
0.0802
MOVES
Leap Year
0.0729
0.0720
0.0815
0.0821
0.0873
0.0881
0.0921
0.0932
0.0845
0.0863
0.0800
0.0800
12.2. DayVMTFraction
The OHIM report provides VMT percentage values for each day and hour of a typical
week for urban and rural roadway types for various regions of the United States for both 1992
and 1995. The data obtained from the OHIM report is not disaggregated by month or
sourcetype. The same values will be used for every month and sourcetype. We used 1995 data
(which is very similar to 1992) as it is displayed in Figure 2.3.2 of the OHIM report.
For the DayVMTFraction needed for MOVES, we first summed the reported percentages
for each day of the week and converted to fractions. Note, the report explains that data for
"Sam" refers to data collected from Sam to 4am. Thus data labeled "midnight" belongs to the
upcoming day. Because MOVES2009 classifies days into two types of days, "weekdays" and
"weekend," we then summed the daily fractions to compute fractions for each type of day.
Table 12-2. DayVMTFractions
Weekday
Weekend
Rural
0.2788
0.7212
Urban
0.2376
07624
We assigned the "Rural" fractions to the rural Roadtypes and the "Urban" fractions to the
urban Roadtypes. The correct distribution for "Off network" VMT is unknown. Since the
majority of U.S. travel is urban, the default DayVMTFraction for "Off network" will be assigned
the urban fractions. Note the MOVES2009 default VMT on "Off-network" roadtypes is zero.
-------
12.3. HourVMTFraction
For HourVMTFraction we used the same data as for DayVMTFraction. We converted
the OHIM report data to percent of day by dividing by the DayVMTFraction.
There are separate sets of HourVMTFractions for "Urban" and "Rural" road types. Road
types were assigned as for DayVMTFraction. All sourcetypes use the same HourVMTFraction
distributions. The Off-Network roadtype was assigned the "Urban" fractions. Figure 12-1
graphs the hourly VMT fractions.
Figure 12-1 Hourly VMT Fractions in MOVES2010
n DQ
n OR
1- 0 07 -
> 0 Ofi -
= 0.05 -
es
Q n 04
0 n no _
o 0 0? -
"o n 01
£ U.U I
iii o
C
Hourly VMT
&G
m »•'-'—>"•». \
h i / 5
/y \
r %
\ ;/ h>
f Tft /<
**&*-
) 10 20 3
Hour of Day
0
— • — Off-Network and
Urban Roadtypes-
Weekend
— • — Off-Network and
Urban Roadtypes-
Weekday
Rural Roadtypes-
Weekend
x Rural Roadtypes-
Weekday
12.4. Extended Idle Activity by Hour
Extended idling, also referred to as "hoteling," is defined as any long period of
discretionary idling that occurs during long distance deliveries by heavy-duty trucks. While
MOVES includes short-term idling (such as at stop-lights) in the default driving cycles, the
emissions from extended idling are modeled separately. In MOVES2010, only the long haul
combination truck sourcetype is assumed to have hoteling activity.
The IdleSHOFactor field in the SourceHour table is the number used to determine the
number of hours of extended idling for each hour of the day. All source use types other than
long haul combination trucks have hoteling activity fractions set to zero.
Federal law limits the number of hours which long haul truck drivers can operate each
day. These regulations are described in the Federal Register.69 Using the distribution of truck
hoteling duration times (shown in Figure 1 of the Lutsey, et al. paper70) and assuming that long
haul truck drivers travel an average of 10 hours a day when engaged in hoteling behavior, we can
estimate the average duration of hoteling as 5.9 hours for every 10 hours of long-haul truck
driving.
However, for MOVES we need the fraction of hours spent hoteling versus hours of
vehicle operation by time of day. This value can be derived from the known truck activity. In
particular, the report, "Roadway-Specific Driving Schedules for Heavy-Duty Vehicles,"71
88
-------
combines data from several instrumented truck studies. The data contains detailed information
about truck driver behavior; however, none of the trucks in any of the studies was involved in
long haul, interstate activity. We assumed that all long haul truck trips have the same hourly
truck trip distribution as the heavy heavy-duty trucks in the instrumented studies and that all long
haul trips are 10 hours long, and thus deduced an hourly distribution of long haul trip ends. The
distribution of hoteling durations from the Lutsey report was applied to these trip-end
distributions. From these calculations, we estimated the number of hours of truck operation and
hours of truck hoteling. For MOVES, we then calculated the ratio of hoteling hours to truck
operation hours for each hour of the day. Weekday data was used for both weekday and
weekend fractions.
Figure 12-2 Extended Idle Activity Ratio
Extended Idle Activity
Ratio of Extended Idle Time to Driving Time by Hour
0 04
o OTS
n m
2 n 09^
CB
U- o 09
2 n ni 5
co
o n 01
— n flO^
n
^•"•"V ' '
*^*^ *--*^
^ s
^ /
^ s
^^-^ ^s
^^*— * A— +~^
0 4 8 12 16 20 24
Hour
Note that the MOVES2010 defaults assume no anti-idling measures or truck-stop-
electrification efforts. MOVES2010a includes a "generic importer" intended to make it easier
for users to modify the inputs of extended idling behavior to account for new or locally available
data on such activity.
-------
13. Vehicle Starts and Parking Activity
To estimate start and evaporative emissions, it is important to estimate the number of
starts by time of day, and the duration of time between vehicle trips. (This between-trip duration
is often called "soak time.") To determine typical patterns of trip starts and ends, MOVES uses
information from instrumented vehicles. This data is stored in two tables: SampleVehicleDay
and SampleVehicleTrip.
The first table, SampleVehicleDay, lists a "sample population" of vehicles, each with an
identifier (vehID), an indication of vehicle type (sourceTypelD), and a "dayID" that indicates
whether the vehicle is part of the weekend or weekday vehicle population.
The second table, SampleVehicleTrip, lists the trips made by each of these vehicles. It
records the vehID, day ID, a trip number (tripID), the hour of the trip (hourlD), the trip number of
the prior trip (priorTripID), and the times at which the engine was turned on and off for the trip
(keyOnTime and keyOffTime, each recorded in minutes since midnight of the day of the trip).
To account for overnight soaks, many first trips reference a prior trip with a null value for
keyOnTime and a negative value for keyOffTime. And, to account for vehicles that sit for one
or more days without driving, the SampleVehicleDay table includes some vehicles that have no
trips in the SampleVehicleTrip table.
The data and processing algorithms used to populate these tables are detailed in two
contractor reports.72'73 The data comes from a variety of instrumented vehicle studies,
summarized in Table 13-1. This data was cleaned, adjusted, sampled and weighted to develop a
distribution intended to represent average urban activity across the U.S. For vehicle classes that
were not represented in the available data, the contractor synthesized trips using trip-per-
operating hour information from MOBILE6 and soak time and time-of-day information from
sourcetypes that did have data. The application of synthetic trips is summarized in Table 13.2.
The resulting trip per day estimates are summarized and compared to MOBILE6 in Table 13.3.
Note, for some sourcetypes, there are hours with no recorded trip starts.
Table 13-1. Source Data for Sample Vehicle Trip Information
Study
3 -City
Minneapolis
Knoxville
Las Vegas
Battelle
TxDOT
Study Area
Atlanta, GA;
Baltimore, MD;
Spokane, WA
Minneapolis/St.
Paul, MN
Knoxville, TN
Las Vegas, NV
California,
statewide
Houston, TX
Study Years
1992
2004-2005
2000-2001
2004-2005
1997-1998
2002
Vehicle Types
Passenger cars &
trucks
Passenger cars &
trucks
Passenger cars &
trucks
Passenger cars &
trucks
Heavy duty trucks
Heavy, heavy duty
diesel dump trucks
Number of
Vehicles
321
133
377
350
120
4
90
-------
Table 13-2. Synthesis of Sample Vehicles for Source Types Lacking Data
SourceType
Motorcycles
Passenger Cars
Passenger Trucks
Light Commercial Trucks
Intercity Buses
Transit Buses
School Buses
Refuse Trucks
Single-unit short-haul trucks
Single-unit long-haul trucks
Motor homes
Combination short-haul trucks
Combination long-haul trucks
Based on Direct
Data?
No
Yes
Yes
No
No
No
No
No
Yes
No
No
Yes
Yes
Synthesized From
Passenger Cars
n/a
n/a
Passenger Trucks
Combination long-haul trucks
Single-unit short-haul trucks
Single-unit short-haul trucks
Combination short-haul trucks
n/a
Combination long-haul trucks
Passenger Cars
n/a
n/a
Table 13-3. Starts per Day by SourceType
SourceType
Motorcycles
Passenger Cars
Passenger Trucks
Light Commercial Trucks
Intercity Buses
Transit Buses
School Buses
Refuse Trucks
Single-unit short-haul trucks
Single-unit long-haul trucks
Motor homes
Combination short-haul trucks
Combination long-haul trucks
MOVES2010
Weekday
0.78
5.89
5.80
6.05
2.77
4.58
5.75
3.75
6.99
4.29
0.57
5.93
4.29
MOVES2010
Weekend
0.79
5.30
5.06
5.47
0.88
3.46
1.26
0.92
1.28
1.29
0.57
1.16
1.29
MOBILE6*
1.35
6.75
7.38
7.38
6.88
6.88
6.88
6.88
6.88
6.88
6.88
6.88
6.88
Note, MOBILE6 distinguished "starts" and "trips." MOVES does not, but MOVES does include some
very short "trips."
14. Geographical Allocation of Activity
MOVES is designed to model activity at a "domain" level and then to allocate that
activity to "zones." The MOVES2010 default database is populated for a domain of the entire
United States (including Puerto Rico and the Virgin Islands), and the default zones correspond to
individual counties. While geographic allocations clearly change over time, the MOVES
defaults were developed for 1999 and are used for all years. If users doing national-level runs
have geographical information by year, this can be handled by doing each year as a separate run,
with different, user-input, allocations. County- and Project-level calculations do not use the
default geographical allocation factors. Instead, they require that the user input local total
activity for each individual year being modeled.
-------
The MOVES geographic allocation factors are stored in two tables, Zone and
ZoneRoadType.
14.1. SHOAllocFactor
The SHOAllocFactor field in the ZoneRoadType table is used to determine the hours of
vehicle operation in each zone on each of the MOVES roadway types.
The national source hours of operation (SHO) are calculated from estimates of VMT and
speed.
The estimate for the VMT by county comes from the 1999 National Emission Inventory
(NET) analysis documented by Pechan & Associates.74 These estimates are based on the
Highway Performance Monitoring System (HPMS) data collected by the Federal Highway
Administration75 for use in transportation planning and vehicle type breakdowns from the EPA
MOBILE6 Emission Factor model.76 The NEI VMT estimates were incorporated into the
National Mobile Inventory Model (NMIM) county database.
To calculate default inputs for MOVES2010, the 1999 NEI VMT estimates were
obtained from the NMEVI database for each county and HPMS facility type. The average speed
estimates were taken directly from Table 8 of the NEI documentation. VMT estimates for each
MOVES road type(i) were determined for each county(j) in the nation and the allocation was
calculated using the following formula, where k refers to the HPMS facility types within a
MOVES road type, and m refers to the VMT for each source type.
County Allocation(ij) = (Sum(j)(( County VMT (i,j,k,m)/Average Speed(k,m))) /
(Sum(ij)((CountyVMT(i,j,k,m)/AverageSpeed(k,m)))
The county allocation values for each roadway type sum to one for the nation. Although
the data is from 1999 calendar year estimates, the same allocations are used for all calendar
years.
14.2. StartAllocFactor and SHPAllocFactor
The StartAllocFactor in the Zone table distributes the domain-wide estimates of the
number of trip starts to the zones. In the default database for MOVES2010, the domain is the
nation and the zones are counties. There is no national data on the number of trip starts by
county, so for MOVES2010, we have used VMT to determine this allocation.
The estimate for the VMT by county comes from the 1999 National Emission Inventory
(NEI) analysis.77 The NEI estimates are based on the Highway Performance Monitoring System
(HPMS) data collected by the Federal Highway Administration78 for use in transportation
planning and vehicle type breakdowns from the EPA MOBILE6 Emission Factor model.79 The
NEI VMT estimates have been incorporated into the National Mobile Inventory Model county
database.
The VMT estimates were obtained from the NMEVI database. VMT estimates for each
county in each state and the allocation calculated using the following formula, where "i"
represents each individual county.
92
-------
CountyAllocation(i) = ( CountyVMT(i) / Sum(CountyVMT(i) )
The county allocation values sum to one for the nation. Although the data is from 1999
estimates, the same allocations will be used for all calendar years.
The SHPAllocFactor in the same table, distributes to the zones the domain-wide
estimates of the number of hours that vehicles are parked, No national data is available, so for
MOVES2010, this estimate was set to equal the StartAllocFactor.
For national level runs, where starts and parking must be allocated to all 3222 counties,
we believe that VMT is an adequate surrogate for start and parking distributions and one of the
few measures that is readily available on a national basis for every county and that includes both
household and non-household vehicles. To test how well this approach compared to other
methods, we computed fractions of vehicles for each county using information from the
U.S.Census Bureau's 2005-2007 American Community Survey80, three-year estimates of
aggregate number of household vehicles available by county. While the survey was lacking data
for more than 1000 counties (this elevates fractions),and came from different years than the
MOVES data, the aggregate household vehicle based estimates for the counties available
correlated well with the VMT-based estimates (a simple regression of MOVES defaults to ACS
values had a linear coefficient of 1.03 and an R2 of 0.96). Some counties where the VMT
approach greatly exceeded the census approach were rural counties with heavy freeway traffic
(for example, Caroline County, Virginia and St. Francis County, Arkansas). Counties where the
household vehicle approach estimates greatly exceeded the VMT-based estimates included some
Chicago suburbs and a large number of counties in Puerto Rico.
14.3. IdleAllocFactor
The IdleAllocFactor field in the Zone table stores the factor used to determine the hours
of extended idling in each zone in each calendar year.
No sources exist that directly measure extended idling in order to geographically allocate
the hours of extended idling estimated for heavy-duty trucks. However, extended idling (or
hoteling) occurs primarily on long-haul trips across multiple states, which suggests that travel on
rural and urban interstates would best represent long-haul trips. Extended idling mainly occurs
among the largest (Class 8) trucks, which are now almost exclusively diesel. Since we have
estimates for the amount of rural and urban interstate VMT by Class 8 heavy-duty diesel trucks
in each county of the nation, we can use this estimate to create a national allocation factor for
extended idling hours.
We did this calculation in two steps. First, the actual total demand for overnight parking
by trucks has been estimated by the Federal Highway Administration on a state by state basis.81
These estimates were used to determine the allocation to each State(i) using the following
formula:
StateAllocation(i) = StateParkingDemand(i) / Sum( StateParkingDemand(i))
93
-------
The State allocation values will sum to one for the entire United States. This method
results in no idling in Washington, D.C., Hawaii, Virgin Islands, or Puerto Rico, which make
sense, since none of these areas have VMT associated with rural or urban interstates.
We then allocated the state values to county. The estimate for the VMT from Class 8
heavy-duty diesel trucks by county comes from the 1999 National Emission Inventory (NEI)
analysis.82 The NEI estimates are based on total VMT from the Highway Performance
Monitoring System (HPMS) data collected by the Federal Highway Administration83 for use in
transportation planning and proportions by vehicle type from the EPA MOBILE6 Emission
Factor model.84 The NEI VMT estimates have been incorporated into the National Mobile
Inventory Model (NMIM) county database.
The VMT estimates were obtained from the NMEVI database. VMT estimates for Class
8 heavy-duty diesel trucks on rural and urban interstates were determined for each county in each
state and the allocation calculated using the following formula where "j" refers to the counties in
each particular state.
IdleAllocFactor(i) = StateAllocation(i) * (CountyVMT(j) / Sum(CountyVMT(j))
The county allocation values will sum to one for the entire United States. The sum of the
county allocations for a given state will equal the state allocation for that state.
15. Air Conditioning Activity Inputs
This report describes three inputs used in determining the impact of air conditioning on
emissions. The ACPenetrationFraction is the fraction of vehicles equipped with air
conditioning. FunctioningACFraction describes the fraction of these vehicles in which the
airconditioning sytem is working correctly. The ACActivityTerms relate air conditioning use to
local heat and humidity. More information on air conditioning effects is provided in the
MOVES technical report on adjustment factors.85
15.1. ACPenetrationFraction
The ACPenetrationFraction is a field in the SourceTypeModelYear table. Default values,
by sourcetype and model year were taken from MOBILE6. 86 Market penetration data by model
year were gathered from Ward's Automotive Handbook for light-duty vehicles and light-duty
trucks for model years 1972 through the 1995 for cars and 1975-1995 for light trucks. Rates in
the first few years of available data are quite variable, so values for early model years were
estimated by applying the 1972 and 1975 rates for cars and trucks, respectively. Projections
beyond 1995 were developed by calculating the average yearly rate of increase in the last five
years of data and applying this rate until a predetermined cap was reached. A cap of 98% was
placed on cars and 95% on trucks under the assumption that there will always be vehicles sold
94
-------
without air conditioning, more likely on trucks than cars. No data was available on heavy-duty
trucks. While VIUS asks if trucks are equipped with airconditioning, "no response" was coded
the same as "no," making the data unusable for this purpose. For MOVES, the light-duty vehicle
rates were applied to passenger cars, and the light-duty truck rates were applied to all other
sourcetypes (except motorcycles, for which AC penetration is assumed to be zero).
Table 15-1. AC Penetration Fractions in MOVES2010
1972-and-earlier
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999+
Motorcycles
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Passenger Cars
0.592
0.726
0.616
0.631
0.671
0.720
0.719
0.694
0.624
0.667
0.699
0.737
0.776
0.796
0.800
0.755
0.793
0.762
0.862
0.869
0.882
0.897
0.922
0.934
0.948
0.963
0.977
0.980
All Trucks and Buses
0.287
0.287
0.287
0.287
0.311
0.351
0.385
0.366
0.348
0.390
0.449
0.464
0.521
0.532
0.544
0.588
0.640
0.719
0.764
0.771
0.811
0.837
0.848
0.882
0.906
0.929
0.950
0.950
15.2. FunctioningACFraction
The FunctioningACFraction field indicates the fraction of the air-conditioning equipped
fleet with fully functional A/C systems, by source type and vehicle age. A value of 1 means all
systems are functional. This is used in the calculation of total energy to account for vehicles
without functioning A/C systems. Default estimates were developed for all source types using
the "unrepaired malfunction" rates used for 1992-and-later model years in MOBILE6. The
MOBILE6 rates were based on the average rate of airconditioning system failure by age reported
in a consumer study and assumptions about repair frequency during and after the warranty
95
-------
period The MOBILE6 rates were applied to all source types except motorcycles, which were
assigned a value of zero for all years.
Table 15-2. FunctioningACFraction by Age (All Use Types Except Motorcycles)
Age
0
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
29
30
FunctioningAC
Fraction
1
1
1
1
0.99
0.99
0.99
0.99
0.98
0.98
0.98
0.98
0.98
0.96
0.96
0.96
0.96
0.96
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.95
15.3. ACActivityTerms
ACActivityTerms A, B and C are coefficients for a quadratic equation that calculates air
conditioning activity demand as a function of the heat index. These terms are applied in the
calculation of the A/C adjustment in the energy consumption calculator. The methodology and
the terms themselves were originally derived for MOBILE6 and are documented in the report
"Air Conditioning Activity Effects in MOBILE6."88 They are based on analysis of air
conditioning usage data collected in Phoenix, Arizona, in 1994.
In MOVES, ACActivityTerms are allowed to vary by monthGroup and Hour, in order to
provide the possibility of different A/C activity demand functions at a given heat index by season
and time of day (this accounts for differences in solar loading observed in the original data).
-------
However, for MOVES2010, the default data uses one set of coefficients for all MonthGroups and
Hours. These default coefficients represent an average A/C activity demand function over the
course of a full day. The coefficients are listed in Table 15-3.
Table 15-3. Air Conditioning Activity Coefficients
A
-3.63154
B
0.072465
C
-0.000276
The A/C activity demand function that results from these coefficients is shown in Figure 15-1. A
value of 1 means the A/C compressor is engaged 100 percent of the time; a value of 0 means no
A/C compressor engagement.
Figure 15-1. Air Conditioning Activity Demand as a Function of Heat Index
0.9
0.8
o
0.7
0.6
•o
(0
i
^ 0.4
0.5
o
0.3
0.2
0.1
70 75 80 85 90
Heat Index (F)
95
100
105
110
97
-------
16. Conclusion
Properly characterizing emissions from vehicles requires a detailed understanding of the
cars and trucks that make up the vehicle fleet, and their patterns of operation. The national
default information in MOVES2010 and MOVES2010a provide a reliable basis for estimating
national emissions. The most important of these inputs are well-established: base year VMT
and population estimates come from long-term, systematic national measurements. The emission
characteristics for the most prevalent vehicle classes are well-known; base year age distributions
are well-measured, and driving activity has been the subject of much study in recent years.
Still, the fleet and activity inputs do have significant limitations, and the uncertainties and
variability in this data can contribute significant uncertainty in resulting emission estimates. In
particular, when modellers estimate emissions for specific geographic locations, it is often
appropriate to replace many of the MOVES fleet and activity defaults with local data. This is
especially true for inputs that vary geographically and for inputs where local data is more
detailed or up-to-date than that provided in the MOVES defaults. EPA's Technical Guidance89
provides more information on customizing MOVES with local inputs.
The fleet and activity defaults also are limited by the necessity of forecasting future
emissions. The inputs for MOVES2010 and MOVES2010a were developed for a 1999 base
year, and much of the source data is from 1999-and-earlier. This information needs to be
updated to assure that the model defaults reflect available information on the U.S. fleet.
Updating the vehicle fleet data will be complicated by the fact that one of the primary
data sources for this document, the Census Bureau's Vehicle In-Use Survey, has been
discontinued. As the data gathered from the last survey (2002) ages, it will become more and
more important to find substitutes that can be used to provide age distributions, fuel distributions,
weight class distributions and other essential data. Without such a data source, future MOVES
calculations may need to be simplified.
A related complication is the cost of data. Collecting data on vehicle fleet and activity is
expensive, especially when the data is intended to accurately represent the entire United States.
Even when EPA does not generate data directly (for example, compilations of state vehicle
registration data) obtaining the information needed for MOVES can be costly and, thus,
dependent on large-scale budget choices.
In addition to these general limitations, there are also specific data elements in
MOVES2010 and MOVES2010a that could be improved with additional research. Such areas
include extended idle activity, vehicle-type distinctions in temporal activity, heavy truck and bus
daily trip activity patterns (particularly at night), characterization of refuse trucks and
motorhomes, classification of passenger and commercial trucks into "light" and "heavy"
regulatory classes, and information on the prevalence of alternative fueled vehicles.
Future updates to fleet and activity defaults will need to focus on the most critical
elements required for national fleet-wide estimates, namely gasoline light-duty cars and trucks
and diesel heavy duty trucks. Information collection on motorcycles, refuse trucks, motor
homes, diesel light duty vehicles and gasoline heavy-duty vehicles will be a lower priority.
In addition to updating the model defaults, we will need to consider whether the current
MOVES design continues to meet our modeling needs. Simplifications to the model to remove
categories (such as sourcetypes or roadtypes) might make noticeable improvements in run time.
-------
At the same time, the fundamental MOVES assumption that vehicle activity varies by sourcetype
and not by fueltype or other sourcebin characteristic may be challenged by the growing market
share of electric vehicles and other vehicles that may have distinct activity patterns.
As we progress with MOVES, development of fleet and activity inputs will continue to
be an essential area of research.
99
-------
17. References
1 U.S. EPA, "Technical Guidance on the Use of MOVES2010 for Emission Inventory
Preparation in State Implementation Plans and Transportation Conformity," EPA-420-B-
10-023, April 2010. http://www.epa.gov/otaq/models/moves/420bl0023.pdf
2 These reports are not yet finalized. Available draft reports are:
U.S. EPA, "Development of Emission Rates for Light-Duty Vehicles in the Motor Vehicle
Emissions Simulator (MOVES2009)," EPA-420-P-09-002, August 2009.
http://www.epa.gov/otaq/models/moves/techdocs/420p09002.pdf
U.S. EPA, "Development of Emission Rates for Heavy-Duty Vehicles in the Motor Vehicle
Emissions Simulator (Draft MOVES2009)," EPA-420-P-09-005, August 2009.
http://www.epa.gov/otaq/models/moves/techdocs/420p09005.pdf
U.S. EPA, "Development of Evaporative Emissions Calculations for the Motor Vehicle
Emissions Simulator (Draft MOVES2009)," EPA-420-P-09-006, August 2009.
http://www.epa.gov/otaq/models/moves/techdocs/420p09006.pdf
U.S. EPA, "Development of Gasoline Fuel Effects in the Motor Vehicle Emissions Simulator
(MOVES2009)," EPA-420-P-09-004, August 2009.
http://www.epa.gov/otaq/models/moves/techdocs/420p09004.pdf
U.S. EPA, "Draft MOVES2009 Highway Vehicle Temperature, Humidity, Air Conditioning, and
Inspection and Maintenance Adjustments," EPA-420-P-09-003, August 2009.
http://www.epa.gov/otaq/models/moves/techdocs/420p09003.pdf
3 This report is not yet finalized. A draft report is available:
U.S. EPA, "Draft Motor Vehicle Emission Simulator (MOVES) 2009 Software Design and
Reference Manual Software Design and Reference Manual," EPA-420-B-09-007, March
2009. http://www.epa.gov/otaq/models/moves/420b09007.pdf
4 U.S. Census Bureau, 1997 Vehicle Inventory and Use Survey, CD-EC97-VIUS. January 2000.
http ://www. census. gov/prod/www/ab s/vius-pdf. html
5 5 U.S. Census Bureau, 2002 Vehicle Inventory and Use Survey.
http ://www. census. gov/svsd/www/vius/2002. html
6 U.S. Census Bureau, 1992 Truck Inventory and Use Survey.
http://www.census.gov/svsd/www/92vehinv.html
7 R.L. Polk & Co., National Vehicle Population Profile.® Southfield, MI. 1999.
http ://usa. polk. com/Products/1 _nvpp .htm.
8 R.L. Polk & Co, Trucking Industry Profile TIP® Vehicles in Operation. Southfield, MI. 1999.
http://usa.polk.com/Products/14_tipnet.htm.
1OO
-------
9 U.S. Federal Highway Administration. Highway Statistics, 1999. Table MV-1, "State Motor
Vehicle Registrations," October 2000. http://www.fhwa.dot.gov/ohim/hs99/index.htm
10 U.S. Federal Highway Administration. Highway Statistics, 1999. Table MV-10, "Bus
Registrations,"October 2000. http://www.fhwa.dot.gov/ohim/hs99/index.htm
11 U.S. Federal Highway Administration. Highway Statistics, 1999. Table VM-1, "Annual
Vehicle Distance Travelled in Miles and Related Data by Highway Vehicle Category and
Vehicle Type," October 2000. http://www.fhwa.dot.gov/ohim/hs99/index.htm
12 U.S. Federal Highway Administration. Highway Statistics, 1999. Table VM-2, "Functional
System Travel," January 2002. http://www.fhwa.dot.gov/ohim/hs99/index.htm
13 U.S. Federal Transit Administration. National Transit Database 1999, Table 29. "Age
Distribution of Active Revenue Vehicle Inventory: Details by Transit Agency."
http ://www.ntdprogram. com
14 Bobit Publications, School Bus Fleet Fact Book. Torrance. CA, 1999.
http://www.schoolbusfleet.com
15 Browning, Louis, Michael Chan, Doug Coleman, and Charlotte Pera. ARCADIS Geraghty &
Miller Inc. "Update of Fleet Characterization Data for Use in MOBILE6 - Final Report."
M6.FLT.002, EPA420-P-98-016, June 1998.
http://www.epa.gov/otaq/models/mobile6/m6flt002.pdf
16 Energy Information Adminstration. Annual Energy Outlook 2003 (AEO2003\ Report #:
DOE/EIA-0383 (2003), released January 9, 2003.
http://www.eia.doe.gov/oiaf/archive/aeo03/index.html
17 Energy Information Administration, Supplemental Tables to the Annual Energy Outlook 2006,
Transportation Demand Sector, February 2006.
http://www.eia.doe.gov/oiaf/archive/aeo06/supplement/index.html
18 Davis, Stacy C. and Susan W. Diegel, Transportation Energy Data Book (TEDB), Edition 22.
Center for Transportation Analysis, Oak Ridge National Laboratory. ORNL-6967.
September 2002.
19 Davis, Stacy C. and Susan W. Diegel, Transportation Energy Data Book (TEDB), Edition 23.
Center for Transportation Analysis, Oak Ridge National Laboratory. ORNL-6967.
October 2003.
20 Davis, Stacy C., Susan W. Diegel and Robert G. Boundy, Transportation Energy Data Book
(TEDB), Edition 27. Center for Transportation Analysis, Oak Ridge National Laboratory.
ORNL-6991. 2008.
1O1
-------
21 Davis, Stacy C., Susan W. Diegel and Robert G. Boundy, Transportation Energy Data Book
(TEDB), Edition 28. Center for Transportation Analysis, Oak Ridge National Laboratory.
ORNL-6991. 2009.
22 Ward's Automotive Inc. http://www.wardsauto.com/
23
24
Hart, Larry. R.L. Polk & Company. Personal communication, June 16, 2003.
National Household Transportation Survey (NHTS). 2001 http://nhts.ornl.gov/download.shtml
25 Motorcycle Industry Council, 1998 Population. Available in EPA Docket A-2000-01, IIB-22
26 American Bus Association. "Motorcoach Census 2000," conducted by R. L. Banks and
Associates, Inc. July 2000. http://www.buses.org/industrv/ABA-RLBanksReport.pdf
27 American Public Transportation Association, 2007 Public Transportation Fact Book as cited in
Transportation Energy Data Book 27, Table 5.13 (page 5-20).
28 APTA, "2010 Public Transportation Fact Book, Appendix A, Historical Tables," Table 17,
"Revenue Vehicles by Mode," April 2010.
www.apta.com/resources/statistics/Documents/FactBook/2010 Fact Book Appendix A.pdf
29 Sierra Research, "Development of Emission Rates for the MOVES Model," Report No.
SR2010-07-01. Prepared for the U.S. Environmental Protection Agency, July 2, 2010.
30 Davis, Stacy C., Susan W. Diegel and Robert G. Boundy, Transportation Energy Data Book,
Edition 28. Table 4.5, Center for Transportation Analysis, Oak Ridge National
Laboratory. ORNL-6991. 2009.
31 Davis, Stacy C., Susan W. Diegel and Robert G. Boundy, Transportation Energy Data Book
(TEDB), Edition 28. Center for Transportation Analysis, Oak Ridge National Laboratory.
ORNL-6991. 2009. Table 4.6
32 Davis, Stacy C. and LorenaF. Truitt. "Investigation of Class 2b Trucks (Vehicles of 8,500 to
10,000 Ibs GVWR)," Oak Ridge National Laboratory. ORNL/TM-2002.49, March 2002.
33
Davis and Truitt, March 2002.
34 Davis and Truitt, March 2002.
35 Union of Concerned Scientists, http://www.ucsusa.org/
36 U.S. Federal Transit Administration, December 2003.
37 U.S. Federal Transit Administration, December 2003.
1O2
-------
38 Yuji Hone, Craig Tranby and Steven Sidawi, Valley Research Corporation. "On-Road Motor
Vehicle Activity Data: Volume I - Bus Population and Activity Pattern, Final Report."
Tables 3-9 & 2-2. Contract A132-182. Prepared for California Air Resources Board,
September 1994.
39 Brian, Mac. Recreational Vehicle Industry Association. Phone conversation, October 29,
2003.
40 Brian, Mac, Recreational Vehicle Association, personal communication, October 19, 2009.
41 Sierra Research, 2010
42 Davis and Diegel, 2007
43 Sierra Research, 2010
44 NHTSA. "Vehicle Survivability and Travel Mileage Schedules," DOT HS 809 952, January
2006. http://www-nrd.nhtsa.dot.gov/Pubs/809952.PDF
45 Sierra Research, 2010
46 NHTSA, 2006.
47 Davis and Truitt, March 2002.
48 U.S. Federal Transit Administration (FTA). "Study & Report to Congress: Applicability of
Maximum Axle Weight Limitations to Over-the-Road and Public Transit Buses,"
December 2003.
49
50
American Bus Association, July 2000.
Good Sam Club, "Highways Member Study 2000." TL Enterprises, Inc., Ventura, California.
(805) 667-4100.
51 USEPA Code of Federal Regulations. (CFR) 40 section 86.529-78
52 United Nations (UN) "Worldwide Harmonised Motorcycle Emissions Certification
Procedure", Informal document No. 15, 46th GRPE, 19-23 May 2003, agenda item number
3. http://www.epa.gov/epahome/cfr40.htm
53 USEPA. "IM240 and Evap Technical Guidance," EPA420-R-00-007, April 2000.
http://www.epa.gov/otaq/regs/im/r00007.pdf
54 Warila, J. "Derivation of Mean Energy Consumption Rates within the MOVES Modal
Framework", 14th Coordinating Research Council On-Road Vehicle Emissions Workshop
Poster Session, San Diego, California, March 29-31, 2004.
103
-------
55 Petrushov, V.A., "Coast Down Method in Time-Distance Variables," SAE 970408, February
24, 1997. http://www.sae.org/
56 Pechan, E.H. & Associates, Inc., January 2004.
57 U.S. Federal Highway Administration (FHA). Highway Performance Monitoring System
Field Manual. OMB No. 21250028, December, 2000.
http://www.fhwa.dot.gov/ohim/hpmsmanl/hpms.htm
58 Jackson, September 2001.
59 Systems Applications International, Inc. "Development of Methodology for Estimating VMT
Weighting by Facility Type," M6.SPD.003, EPA420-R-01-009, April 2001.
60 Sierra Research, Inc. Memo from Tom Carlson to John Koupal, "Analysis of Rural Average
Speed Distributions for MOVES," Purchase Order EP05B00129, December 1, 2004.
61 Sierra Research, Inc. Memo from Tom Carlson to John Koupal, "Analysis of Rural Average
Speed Distributions for MOVES," Purchase Order EP05B00129, December 1, 2004
62 U.S. EPA, March 2009.
63 Sierra Research, Inc. M6.SPD.001 "Development of Speed Correction Cycles." EPA Contract
No. 68-C4-0056, Work Assignment 2-01, June 26, 1997.
http://www.epa.gov/otaq/models/mobile6/m6spd001.pdf
64 Hart, Constance. "EPA's Onboard Analysis Shootout: Overview and Results." EPA420-R-02-
026, October 2002. Online at
http://www.epa. gov/otaq/ngm. htm
65 Sierra Research, "Development of Generic Link-Level Driving Cycles." SR2009-05-02 EPA
Contract EP-C-05-037, Work Assignment 3-02, May 5, 2009.
66
67
Eastern Research Group, Inc. (ERG), "Roadway-Specific Driving Schedules for Heavy-Duty
Vehicles." EPA Contract 68-C-OO-l 12, Work Assignment 3-07, August 15, 2003.
Sensors, Inc., "On-Road Emissions TEsting of 18 Tier 1 Passenger Cars and 17 Diesel
Powered Public Transport Buses," Reference# QT-MI-01-000659, October 22, 2002.
Festin, Scott. "Summary of National and Regional Travel Trends: 1970-1995," Office of
Highway Information Management, Dept. of Transportation, May 1996.
http://www.fhwa.dot.gov/ohim/bluebook.pdf
USEPA Combined Federal Register Vol. 65, No. 85, Tuesday, May 2, 2000. Proposed Rules,
49 CFR Parts 350, 390, 394, 395 and 398. http://www.fmcsa.dot.gov/Pdfs/050200p.pdf
104
-------
70 Lutsey, Nicholas, Christie-Joy Brodrick, Daniel Sperling, and Carollyn Oglesby. "Heavy-Duty
Truck Idling Characteristics - Results from a Nationwide Truck Survey." Annual Meeting
of the Transportation Research Board, January 2004.
71 Eastern Research Group, August 2003.
72 Sierra Research, "Development of Trip and Soak Activity Defaults for Passenger Cars
andTrucks in MOVES2006," SR2006-03-04, EPA Contract EP-C-05-037, Work
Assignment No. 0-01, March 27, 2006.
73 Sierra Research, "Development of Trip and Soak Activity Defaults for Passenger Cars and
Trucks in MOVES," SR2007-06-01, EPA Contract EP-C-05-037, Work Assignment No. 1-
01, June 29, 2007.
74 Pechan, E.H. & Associates, Inc. "Documentation for the Onroad National Emissions
Inventory (NET) For Base Years 1970-2002," prepared for EPA Office of Air Quality
Planning and Standards, January 2004.
ftp://ftp.epa.gov/EmisInv entory/2002fmalnei/documentation/mobile/onroad_nei_basel 970
_2002.pdf.
75 U.S. Federal Highway Administration (FHA). Highway Performance Monitoring System
Field Manual. OMB No. 21250028, December, 2000.
76 Jackson, September 2001.
77 Pechan & Associates, Inc. October 2002.
78
79
U.S. Federal Highway Administration, December 2000.
Jackson, September 2001.
80 U.S. Census Bureau, 2005-2007 American Community Survey 3-Year Estimates,
dc_acs_2007_3yr_gOO_datal.txt, field "B25046."
http://factfinder. census. gov/servlet/DatasetMainPageServlet?_program=ACS&_submenuId=&_l
ang=en&_ts=
81 Fleger, Stephen A., Robert P. Haas, Jeffrey W. Trombly, Rice H. Cross III, Juan E. Noltenius,
Kelley K. Pecheux, and Kathryn J. Chen. "Study of Adequacy of Commercial Truck
Parking Facilities." Table 7. U.S. Federal Highway Administration, FHWA-RD-01-158,
March 2002. http://www.tfhrc.gov/safety/pubs/01158/
82
Pechan & Associates, Inc. October 2002.
83 U.S. Federal Highway Administration, December 2000.
105
-------
84 Jackson, September 2001.
85 U.S. EPA, "Draft MOVES2009 Highway Vehicle Temperature, Humidity, Air Conditioning
and Inspection and Maintenance Adjustments," EPA-420-P-09-003, August 2009.
http://www.epa.gov/otaq/models/moves/techdocs/420p09003.pdf (Final report currently
under development.)
86 John Koupal. M6.ACE.001 "Air Conditioning Activity Effects in MOBILE6," EPA420-R-01-
054, November 2001. http://www.epa.gov/otaq/models/mobile6/r01054.pdf
87
88
Koupal, November 2001 .
Koupal, November 2001 .
89 U.S. EPA, "Technical Guidance on the Use of MOVES2010 for Emission Inventory
Preparation in State Implementation Plans and Transportation Conformity," EPA-420-B-
10-023, April 2010. http://www.epa. gov/otaq/models/moves/420b 1 0023 .pdf
106
-------
Appendix A. Response to Peer Review Comments (A)
Draft MOVES2009 Highway Vehicle Population and Activity Data
Peer Review Comments & EPA Response
Professor Lisa Aultman-Hall
As part of the MOVES2010 Peer Review process, EPA solicited comments from
Professor Lisa Aultman-Hall on the July 2009 draft of report Draft MOVES2009 Highway
Vehicle Population and Activity Data
Professor Aultman-Hall is the faculty director of the University of Vermont
Transportation Research Center. She is a full professor in the School of Engineering and her
three degrees are all in civil engineering specializing in transportation. Her research involves
travel data collection and statistical analysis of this travel data. The majority for funding for the
Transportation Research Center is derived from the US DOT. Her research has been funded by
the US DOT, several state DOTs andNSF. The Center includes a Transportation Air Quality
Lab where graduate students collect second-by-second vehicle activity and tailpipe emissions
data from light duty and hybrid vehicles. Staff in the Center work with travel demand
forecasting models that generate trip rates and length distributions related to the activity data in
MOVES. At the time of her review, she was Principal Investigator of a research project that
involves estimation of start and soak activity for typical personal travel from GPS data. While
her staff and graduate students have run MOBILE and MOVES, she does not. One previous
Ph.D. student and one current Ph.D. student study vehicle activity and tailpipe emissions using
real-world on-road data. Based on this combination of experience, she is familiar with the
concepts, models and approaches discussed in the document reviewed, but claims no explicit
detailed experience with the MOVES model itself.
Professor Aultman-Hall's comments are copied below, with EPA response in italics.
I must open by stating that the task of estimating vehicle population and activity for the nation,
or zones within the nation, is huge. I understand the authors have made, by necessity,
assumptions and used older data from limited sources. I know they would prefer more accurate
or timely measures but that these simply do not exist in many cases. Furthermore, the authors
have done an excellent job especially given their limited resources. My requests and suggestions
for changes are made while thoroughly understanding the project resources and timeline will not
allow all changes to be made. When I am particularly concerned about a modeling approach or
data source I have noted that extra concern in my comments below. I have divided my
comments according to specific requests in Mr. Koupal's correspondence. I have shown these
questions in bold below. Finally, there are edits and small wording changes that I have circled
on my copy of the report that I have provided in pdf.
Overall Comments:
107
-------
The authors have done the best they can with very limited data. However, given the level of
emissions, energy use and environmental damage stemming from mobile source tailpipe
emissions there is a critical need for more robust data on vehicle population and activity
including forecasts. The necessity of relying on relatively weak secondary data for this
important task is inappropriate. Support for dedicated data resources and forecasts is important.
I categorize the data sources in this report into my own three broad classes: 1) vehicle
population; 2) total travel (VMT) by roadway location and 3) operational patterns (operating
modes or driving schedules/cycles). In the later case, operating mode, data are understandably
limited as this approach is new and the methods to collect robust representative data limited. In
fact I was impressed by the efforts already undertaken to fill this data gap for MOVES. New
field studies and use of traffic simulation models will improve this data source in the 5 year time
horizon. VMT by location, my second broad class, is estimated and described in this report. The
methods and data sources are reasonable and might be improved by use of different
disaggregations (replacing the urban/rural types with design speed and volume to capacity ratios;
addition of holiday as a day type for example) but for the most part methods are good. However,
in the first class, vehicle population, as well as the total VMT per year by vehicle type, the data
sources available to the authors were of un-acceptable quality. Given that vehicles require
government licenses and that mileage on vehicles is tracked during inspection or upon sale,
resources to tabulate better data in the appropriate categories from at least a representative
sample of states should be made available and should be a national transportation/environment
data priority.
We agree that vehicle population and VMT data in MOVES are not ideal. We did not
have the time to update these values for MOVES2010, but we plan to update this data for future
versions of MOVES. Note that these defaults are most important for runs where national
defaults are used. For State Implementation Plans and conformity determinations, we expect
local values to be used in place of national defaults.
While it may seem easy to compile state registration data, differences between states in
the way vehicles and fuels are counted and classified make determining vehicle populations and
mileage accumulations with the level of detail needed for MOVES quite challenging. And the
discontinuation of the Census Bureau's Vehicle Inventory and Use Survey makes it more difficult
to estimate truck populations in the categories needed for MOVES. Unfortunately, EPA does not
have the resources to generate this data directly. For the future, we are budgeting funds and
resources to purchase and analyze more recent existing population and mileage accumulation
data, but we expect to use similar data sources to those used for Draft MOVES2009. We expect
the values will be more up-to-date, but otherwise of similar quality.
In terms of report style, it was developed assuming a level of knowledge of MOVES or
MOBILE and the associated modeling approaches. I believe, the introduction must include basic
information and context for the new analyst. It is my hope to use this manual and others in
teaching undergrads and grads whom we hope will join the ranks of an increasing number of new
professionals using MOVES. I think if the authors think of these graduate student readers as
their target reader it will help them add this background information. It is so clear that the
108
-------
authors "live and breath" MOBILE and MOVES. They know it inside out and with my
experience I can follow what they are talking about in most cases, but in many sections more
context would be very useful to inexperienced professional readers.
The introduction should start with a general description of how MOVES will calculate emissions
- very general for those who do not know yet. Things like MOVES combines second-by-second
vehicle specific power (VSP) for different sources or types of vehicles when they travel on
different roads types by time of day, day of week and month of year.
By midway through chapter 3, it is starting to be confusing regarding timelines - 1990, 1999 and
future years. This gets even more confusing for a new reader in Table 4-1 when AC data goes
back to 1972 and in chapter 7 when fuel fractions are introduced. These will seem like arbitrary
divisions for those not aware of regulatory history or requirements. Ideally, all these concepts
including both the data needed as well as the timelines associated with each type of data should
be included in the introduction. I also believe the subsections dealing with <1999 and >2000 can
be combined with single tables for each variable for all years.
A conclusion chapter should be written. This chapter can summarize the authors' knowledge of
data source limitations and also provide a ranking of the most critical data gaps that must be
filled. This would set a context for research.
We have substantially revised the introduction, reorganized material, and edited the text
to make the report clearer.
Are the data sources appropriate? Are important data sources missing? Are there
alternative data sources?
I assess that for the most part that appropriate assumptions and estimations have been made with
limited data in this report. However, the data sources while the best available are not always
optimal. As I suggest above, a dedicated vehicle population dataset with appropriate
classifications could be obtained from a subset of 6-10 state DMV offices. A similar approach is
used in highway safety in the Highway Safety Information System (HSIS). I recommend this as
a long-term improvement requiring policy support and program funding. Resources invested in
proper documentation of vehicle population and total VMT activity would more than likely
substantially increase MOVES accuracy.
We have begun investigating whether using detailed data from a sample of states would
be a worthwhile approach for MOVES purposes, perhaps in addition to our current data
sources.
The lack of alternative fuel vehicle (AFV) data and accurate projections for market penetration
and travel is concerning. I believe national policy makers may look to MOVES for important
alternative scenario evaluations for AFVs and the model data will limit the usefulness of results.
An effort to pursue these data should not delay release of MOVES2009 but should be a top
priority for data improvement.
109
-------
In the future, MOVES may become an important tool for analyzing emissions from AFVs.
However, because AFV modeling is not a primary concern for State Implementation Plans,
conformity determinations, EPA rulemakings or EPA inventories, we have removed most AFV
options from MOVES2010. Removing these fuels from the user interface makes it clearer that
modeling alternative fuels is currently outside the intended scope for MOVES.
That said, the algorithms for modeling alternative fuels still exist in the model and if EPA
priorities change, the model could easily be updated. Or an interested user could develop the
required inputs to do this analysis.
The data on intercity buses might be improved by using the American Bus Association or the
United Motorcoach Association (established in 1926 and 1971 respectively).
We used data from the American Bus Association for Draft MOVES2009 and
MOVES2010. We plan to request more recent data for MOVES updates, and will also request
information from the United Motorcoach Association.
Although I find the agreement of methods reassuring in Figures 6-1 and 6-2, the methods/data
especially for truck age distribution is weak. The dated TIUS and the importance of trucks to
emissions suggest a need to seek funding to repeat or replace this survey.
We agree that this data is essential, but implementing a full VIUS replacement is outside
the scope of work for our office. We are exploring other options for substitute data.
The California School bus data (Table 7-15) is dated. Although I am unaware of a source, I
cannot believe Department of Education or Energy sources do not have more up to date
information.
The data used was the best available at the time. We are investigating more up to date
sources of data on school bus weights andfueltypes.
When VMT is divided by day type, weekday and weekend are used. Many research studies have
found Saturday is different from Sunday and that holidays are unique. More than two categories
of day may be considered in future versions.
In designing MOVES we considered treating each day separately (because Fridays are
also quite different from other weekdays), or treating all days as the same. Choosing to use two
day types was a compromise intended to improve model run-time when compared to the seven-
day option and to optimize file size and local data entry requirements, while still accounting for
the substantial differences in traffic patterns that are found when comparing weekdays and
weekends. To model different day-type categories than those chosen inMOVES2010, a user
could re-populate the appropriate data tables describing daily traffic variation.
I note the important addition of higher speed freeway drive schedules as important. This may
have a definable impact when MOVES starts being used for project level analysis.
no
-------
The average speeds listed in Table 14.2 of the draft report indicate that MOVES2009
(andMOVES2010) use driveschedules with average speeds up to 76.0 mph. The draft report
does not detail that many of the driveschedules include significant speed variations, with
maximum instantaneous speeds up to 90 mph. The final version of the report will make this
clear. Of course, if users wish to model a project where average speeds are greater than 76
mph, the users would need to provide their own appropriate driving schedule or VSP
distribution.
Should the modeling or data for soak times be included in this report? Would it not flow directly
from the drive schedules for vehicles? Chapter 20 presents only start data for urban areas only.
The soak time data does not come from drive schedules, instead it comes from "key-off, "
"key-on "data from a collection of vehicle studies. The data and its analysis are detailed in a
contractor report that will be posted on the EPA website. It is true that most of the data is from
vehicles based in urban areas. It is possible that a study with a rural focus would see different
patterns.
Although you state that no default values for migration rate between zones are provided because
this version is estimating national emissions, it would be very straightforward to estimate vehicle
migration based on population increases and decreases which are available back further than the
timeline required here and at the county level.
The migration rate discussed here is at the national level, so it would refer to vehicles
entering the U.S. though mechanisms other than new vehicle sales. We will clarify. At the
national level, vehicle sales data are sufficient to estimate the national vehicle population.
The reviewer's actual concern is the distribution of population between zones, which is
handled by the various "AllocFactors" in the Zone and ZoneRoadType tables, described later in
the report. While these allocations clearly do change over time, MOVES2010 simplifies with a
single value for each zone. While entering allocations by year would vastly increase table size,
this is an area we are considering revising in future versions of the model. Note, however, that it
is not an issue for State Implementation Plan and conformity analysis because EPA 's guidance
requires users to enter specific county-level population for each calendar year.
Are the default values for vehicle population and activity appropriate for national use?
I believe these are the best estimates available. My most significant concern regarding use of the
model for national estimates is the inability to evaluate the impact of AFVs market penetration.
I am concerned that the report suggests that idling is only considered for heavy duty trucks.
Idling is a policy concern that may be evaluated at any study area level for any vehicle type. Is
idling perhaps indirectly incorporated for all vehicles in the drive schedules? If yes, this should
be explicitly mentioned in Chapter 15.
ill
-------
Yes, idling is incorporated in the drive schedules. We will clarify. The "Extended
Idling" emission process for heavy duty trucks differs from the in-cycle idling because it is
typically a high-speed idle designed to run accessories for several hours
Were assumptions and extrapolations appropriate? How could they be improved?
While the default data are likely appropriate for national use, the move to smaller zones will
require a future effort to develop methods to construct driving schedules from simulated traffic
data from microscopic traffic simulation models. One long term improvement I believe should
be pursued is the provision of speed distributions, not just as a function of road type and average
speed, but also by volume to capacity ratio (V/C). This might be a better approach than rural
versus urban which are not causal variables (acknowledged indirectly on page 85 when Table 14-
1 is discussed) but rather surrogates whose limitations are now widely recognized within
transportation. The authors might acknowledge this limitation when road classes are introduced.
We will revise the paper to be more explicit about the limitations ofroadtype in
determining appropriate drive schedules and/or operating mode distributions. Users who have
more detailed information on driving activity may instead choose to use the MOVES project-
level option where they can input this information directly.
On page 23, it is unclear why 0.3 was selected as a survival rate for age 30 years and older. The
modeled survival rate for 29 years is so much higher for all vehicle types. This suggests 0.3 is
too low.
The survival rate for age-30-and-older is not the survival rate for the year after age 29,
but an average for all cars 3 0-years an older. A simple one-year extrapolation of the survival
rate leads to a ballooning population of very old cars, not supported by data. Instead, we chose
a low survival rate that leads to reasonable age distributions. More research in this area could
certainly improve our value for this number, but is unlikely to have a significant impact on fleet-
wide emission results. Our revised report will explain this more clearly.
It is unclear on page 69 why the rolling resistance was multiplied by a factor of 5.
That statement was incorrect. The equation is correct as written. We will fix this in the
final report.
On page 70, it is unclear why if you are conducting sensitivity analysis by varying weights that
you should not change all vehicle weight terms in the equation.
The "Fixed Mass Factor " is not a true mass, but a value used in the emission rate
calculations to transform measured emissions into VSP-specific rates. We will revise this
explanation to clarify.
112
-------
The county allocation on page 91 section 17 could be tested by checking against population and
employment which are available by county (by year).
While we would expect some correspondence between VMT and population or
employment, we would not consider this a reliable check of the VMT estimates. Public transit
can reduce VMT per person in urban counties, while major freeways can increase VMT per
person in rural counties. Patterns in non-passenger vehicle traffic, especially long-haul
vehicles, are unlikely to match employment or population distributions.
In chapter 3, the differences between the distribution of truck types in 1990 is so similar to 1999
that I am not sure why both datasets are needed. The difference is well below what I expect the
error level is.
The MOVES activity generator assumes that base year vehicle populations are
independent, so the model expects a separate set of values for each base year.
Overall are the fleet and activity inputs described here adequate for calculating national
highway emissions inventories?
I have reasonable confidence in past and current emissions estimates. But future projections
(described in section 3.3, 7.7 and elsewhere) are very dependent on AOE estimates. At the very
least, this merits a more complete discussion of what their projections are based on, levels of
errors expected and what factors affect these estimates. I am concerned that the information
regarding different vehicle technologies, fuels and efficiencies into the future are not accurate
and that this will affect the accuracy of policy scenario evaluation. As an example, the aging
baby boomers and their retirement makes me doubt levels of increasing sales growth. How do
demographics fit into AOE projections?
We will provide better citations to AEO methods, including the NEMS Transportation
Demand Module. This module projects light-duty passenger vehicle sales based on income per
capital, fuel prices and average predicted vehicle prices.
As indicated above, I am concerned about the inclusion of alternatively fueled vehicles. It is
unclear to me in sections 7.2 and 7.7 if data exists to be able to model these and I suspect
MOVES will be called upon to help evaluate emissions benefits from AFVs. As an example on
page 55, CNG and LPG refuse trucks are coded as diesel. In section 7.7 AFVs are discussed but
then in 7.7.2 and subsection tables they are not included at all.
As explained in the response above, modeling alternative fuels is currently outside the
intended scope for MO VES.
I am concerned the data regarding AC discussed in chapter 4 and elsewhere are not fully
appropriate. For example, treating all trucks and buses the same in Table 4-1 and neglecting
regional differences especially in older models and for previous year estimates is a source of
113
-------
error. It is impossible for me to assess the level of error without running the model and assessing
the impact of AC on the ultimate emissions rates.
In our testing of the model, we saw a small impact of air conditioning on the fuel
consumption from heavy duty diesel trucks. In the warmest months, we saw a fleet average air
conditioning effect on fuel consumption of less than 10 per cent. The fuel consumption change
also results in a proportional increase in refueling hydrocarbon emissions and SO2 and SO4
emissions. Other pollutants were unaffected.
We have attempted to find other data sources on air conditioning in heavy trucks, but
have not found a reliable source. However, discussions with EPA trucking industry experts
suggest that the current fraction of trucks equipped with air conditioning is quite high. It is
certainly possible that we are overestimating or underestimating the number of older trucks with
air conditioning, but as these older vehicles are scrapped from the fleet, the potential impact of
errors here will decline.
In Table 7-4,1 cannot understand why the diesel fractions bounce up and down by so much for
individual truck types. The discrepancies are large enough to make me question the data or
procedures. Note the one non-1.0 number for source type 62.
While we do not have a better source for this data, we agree that some of the large
variations don't make sense. In some of these cases, we have smoothed the inputs used in
MOVES. The details will be described in the final report.
It is unclear how empty hauls are handled in the truck weight distributions. An explicit
discussion could be added, perhaps in section 7.3.2.
The average weight was intended to include a mix of full and empty trucks. We will
explain this in the final report.
Appropriateness and completeness of the literature discussed
It might be prudent in terms of building the case for better vehicle population and activity data in
the US to provide a review of what other countries do. Do they have better direct sources?
I am concerned about the web references as a long-term reference. In addition to the weblink I
prefer to have enough information to use a search engine to find the report or data if the specific
web link is dead. Some references are very brief and non-specific as circled in my pdf of the
report.
We will improve the references in the final report.
Integration of Information from multiple areas
This is very challenging and the report achieves very good integration for experienced modelers.
However, this report and MOVES will be used at the intersection of diverse professional
114
-------
disciplines. I recommend considering the new professional analysts or graduate students as the
target reader and defining all jargon for their consumption. Examples are as simple as "vehicle
source type" clarifying "source type" or requiring more complete discussion of a driving
schedule.
Our technical reports are intended primarily for specialists, but we will work to make
them clearer.
Clarity of Presentation
A list of all acronyms at the front or back of the report would be very helpful.
(page 6) A third column should be added to Table 1-2 that describes in lay words the brief
definition of the data table. For example, sourceusetype is Emission Model Variables or
Variables for Vehicle Specific Power Equation. Similarly, a fourth column could provide the
report section numbers where the data table is described.
(page 6) Define activity. Here and throughout, it would be easier for new analysts if the word
"vehicle" was inserted before "population" and "source".
(page 6) Change "the sources and calculations" to "the data sources, assumptions, and
calculation procedures"
Throughout the report the U.S. DOT and its Administrations and Offices are not referred to
consistently. I would recommend searching and replacing throughout with U.S. DOT and
providing office names in parenthesis.
(page 9 and 10) In some sections of chapter 2 the document indicates "what" the data are used
for and in others it does not. For example it does not say what Polk or FTA is used for.
(page 11) sourcetypeyear is not an appropriate section or chapter heading for the whole set of
target readers. Throughout it seems that variable names are used as section titles. I would
recommend changing chapter 3 to something like "The Vehicle Source Type and Age".
(page 11) bullet #2 - why are the differences less important since 1990. Can an example be
provided for bullet #3. I believe information like this that is known well to the authors can be
useful in educating diverse future users of MOVES.
(page 11) "Also, FHWA data is..." I think this point is actually a bullet #4.
(page 12) Table 3-1 - It is unclear the difference between a blank and an "na"
(page 12 and throughout but especially chapter 8) equation numbers would be very helpful.
(page 12 and 15) Table 3-2 There are several places where it is unclear whether this is the
documentation for DRAFTMOVES2009 or FinalMOVES2009. This table (and Table 3-4)
115
-------
suggested to me that the report was about final MOVES but later sections suggest otherwise.
Should Table 3-6 have final MOVES populations too. I think the number cited for refuse trucks
is for final MOVES not Draft. Table 3-6 could have a total row.
(page 12) Table 3-2 Why not include all source types in this table?
(page 13) Table 3-3a and 303b could have the same column widths which would help a reader
compare. Use of AXLRE, axle_config, and AREAOP has no meaning to the reader.
Table 3-8 could also have the same column widths.
(page 13) Table 3-5 Should APTA be added to the data sources in chapter 2.
(page 15) For both motorcycles and motor homes the lower option is selected without
justification.
(page 16) In Table 3-7 and others it might be helpful to very lightly shade the cells which contain
the value used.
(page 17) The text on the bottom of the page regarding bus population could be expanded with a
few more details including reference to Table 3-10. Table 3-10 includes some extra digits in
columns 1 and 2.
(page 19) Table 3-11 - the population numbers for the three types of buses are slightly different
from Table 3-10.
(page 21) Table 3-12 Motorcycles and cars are the same for 2000 and 2001 - is this correct?
(page 21) Section 3-3 - Clearly define migration rate.
(page 22) The data related to functioning air conditioning could be included here from 5.3.
(page 11-24) Sections 3 through 5.1 are all about vehicle populations. This could be grouped as
one chapter. Section 5.2 is about vehicle activity or travel and belongs back with those sections
in chapter 11 or 12?
(page 23 and 25) MAR needs an explicit definition up front for the new reader. It is still unclear
in section 5.2 whether this is mileage in year x as a function of year x-1 or as a function of
mileage in the first year. This needs to be made explicit for Figure 5-2 as well. On page 25 state
why it is desirable to keep MAR constant.
(page 26) Why was the MAR for motorcycles set equal to the MOBILE value. Simply say why
for the new reader.
(page 28) The reader needs source type names in Figure 5.1 to be able to interpret the graph.
116
-------
(page 30) Another example of the need for lay-word titles. I would propose Chapter 6 be called
"Vehicle Age Distributions for Base Year". But Chapter 6 is also another example of what
might be considered in re-organizing the report. This is no longer data source descriptions but
rather calculation with the data to get age distributions of the fleet. The general procedure could
be outlined at the beginning of this section as population by vehicle age is calculated as age
distribution for the base year plus sales and minus survival. Chapter 6 is also the only place a
rejected method is described in detail.
(page 34) Table 6-1 Use both sourcetype number and names.
(pages 35-36) The information from sections 6.7 through 6.12 could be easily, and more
effectively, addressed in sections 6.1-6.6. This would also allow the authors to explain the
context and need for different base years.
(page 37) Sourcebindistribution needs a name new analysts and policy makers will understand -
perhaps Vehicle Sub-types. It is also important that a discussion of context be added as this
chapter opens explaining how emissions vary by these different vehicle characteristics and that
MOVES will have different emissions rates for each type. This is implied but needs to be stated
explicitly especially for the new reader. There is also a need to explicitly describe in the first
section of 7.0 that the system of delineating source bins is different pre-2000 and post-2000.
(page 37) Post 2000 is all in section 7.7 yet pre-2000 is in sections 7.1 through 7.6. I would
suggest putting both <2000 and >2000 in a single section by vehicle type and producing a single
table for all years. The discussion could then explicitly discuss any discrepancies at the timeline
breakpoint.
(page 38) Table 7-1 Row 1 is unclear, rows 2 and 3 are very clear, row 4 is less clear and may
need an example, rows 5 and 6 do not seem to fit with the others. Some additional explanation
of what these variables are used for could be added to the second paragraph of Chapter 7 on page
37. (Same comments for paragraph that follows Table 7-1 - the variable name jargon will be
hard for new users to follow).
(page 40) Table 6-3 should read Table 7-3.
(page 41) Is Table 7-4 for final MOVES or DraftMOVES? Sentence above the table may
contribute to confusion.
(page 48) Is Table 7-8 referenced from the text.
(page 49 and 50) Tables 7-9 and 7-10 illustrate two options that could be considered for many of
the tables. First, minimizing decimal places would allow the reader to more easily see trends (it
also helps authors edit and see errors). Second, all tables could go from start year (1967 or 1970)
completely through to 2009 regardless of the source of the data or the estimation procedures
used. This would make each Table a comprehensive look up source. For several items, one
needs to look to different sections or different tables for the same data for different years. Why
are 2000 and 2001 mostly missing from both sets of tables?
117
-------
(page 51) Last sentence above Table 7-11 should read Table 7-12 not Table 7-10.
(page 54) When you have a very small section such as 7.4.3 or 7.6.3 could the information not be
included in the introduction to the main sub-section - in this case in section 7.4 or 7.6.
(page 56) Tables 7-17 and 7-18 are not referred to or discussed in the text.
(page 60) Will methanol and hydrogen be in final MOVES?
(page 61) Honda and Toyota - I would not mention specific manufacturers in this document.
(page 63) The Table naming changes from #-# to #.#
(page 67) Should chapter 8 be called Emissions Model Variables or Emissions Model Constants
(page 67) As chapter 8 opens VSP should be defined. The equation should be referenced. A, B
and C might better be called constants than terms. Take care to avoid confusion about the rolling
constant/term and rolling resistance - the constant verses the whole term in the VSP equation.
There may not be enough context for someone unfamiliar with VSP to follow this section.
(page 67) I think that the second equation in this section is the weight for a given vehicle class
and age in the whole fleet in a given year but I am unsure. I would recommend more
explanation in this section.
(page 68) Table 8-1 - You could add the mass M in kg that is actually used in the VSP equation
as a fourth column.
(page 68) Section 8.2 The "UN Report" is not referenced and not in section 2.0. The equation is
really two - one for A and one for C - place on different lines and use unique equation numbers.
It is unclear in the paragraph that opens "For vehicles with a weight..." which source categories
this applies to.
Chapter 8 - The constants could be subscripted with the source type and weight class.
Chapter 9 - Title could be Distribution of VMT by Road Type
(page 70) "fixed mass factor" This paragraph is very hard to accept. Why if you are changing
the mass of the vehicle fleet would some, but not all, of the mass terms be changed in the model.
Would use of an equation here help explain this situation?
(page 71) - The Highway Statistics are not self reported - are they perhaps state-reported?
(page 72) Table 9-2 - I do not believe these are road type distributions. This seems like VMT
Distributions by Road Type.
118
-------
(page 73) In the paragraph that starts "The average speed was..." it is very hard to follow the
logic. Links is introduced. This is fine for me a traffic engineer but I suspect this will get
confusing for others. The concept of driving schedules is also hard to follow. I took this to be
similar to driving cycles but different because the time sequence is not necessarily continuous or
complete.
(page 75) Chapter 11 title - VMT by Vehicle Type per Year? I would recommend the opening
sub-section contain a more complete context of why these data are needed and how they will be
used in lay language. Note that assuming VMT growth is constant by road type may be limiting
for some types of scenario evaluation.
(page 76) Third last line should be 11-2 not 11-3.
(page 79) Is this OHEVI the same as the HPMS data source described earlier in the report? Is
1996 the only data? Are more recent estimates not available? In section 12.1 I would explicitly
refer to Table 12-1 and state that total VMT by source for the year is allocated to months using
these data.
(page 81) Figure 12-1 is useful and would be more useful if bigger. The table might also be
presented for consistency with other data charts. It is unclear here and in other places whether
these data are for draft MOVES or final MOVES. The text requests input on whether hourly
fractions is important for inclusion. I believe it will be more appropriate for inclusion after
velocity distributions are provided by V/C ratio by road type.
(page 82) Chapter 13. The authors could be more explicit in distinguishing between vehicle
trajectories (the source data), vehicle schedules (the tabulated data) and driving cycles (a
traditional sequence of related items used for emissions data collection in labs). It is unclear
from the text whether trajectories are actually used or simply a distribution of speed patterns
("snippets") by road type. Are driving schedules dimensionless in time such that they map to
any mileage of a given road type?
(page 82) This section introduces for the first time in this report the important road type of
ramps. It appears that drive schedule distributions are generated for ramps but the VMT
information by vehicle sources is not provided that way. How can this discrepancy be
explained? Are a set proportion of ramps associated with a unit length of freeway?
(page 84) No new chapter - 14 can be included in 13. Some of the description of what drive
schedules are from this section could be more useful to reader if it came at the start of chapter 13.
(page 86) Table 14-2 is unclear and contains jargon. I was expecting a list of average speeds /
road class combinations without their own drive cycle and the road class and speed combination
used instead. I think that this information is there, but expressed as file names or data table
names instead of words.
(page 88) This chapter could be entitled Idling.
119
-------
(page 90-92) Chapter 16 and 17 should be reversed in order.
(page 91) The zone change in MOVES is important and should be included in the introduction
where the general form of the model is outlined.
(page 91) Should this chapter be called Vehicle Starts, Idling and Parks? Is this information only
provided for class 8 trucks? How are soak emissions estimated for all vehicle types? Soaks are a
function of activity and drive schedules and might be appropriately handled in this report.
(page 93) This chapter could be named "Creating Output by EPA's Source Category Codes"
(page 94) There are 2 Table 18-ls. The text indicates that Table 18-1 will contain the
proportions of vehicles by delineation but it does not.
(page 96) Chapter 19 could better follow chapter 13. I would suggest different names for
variables A B and C here to avoid confusion with VSP equation terms. Can an equation be
provided for air condition activity demand? It is unclear how Figure 19-1 is used in combination
with yearly distribution data. Does this vary by region with temperature?
(page 97) Should chapter 20 be called "Start and Soak Emissions". Table 20-1. Can references
to reports on these data be provided? If the number of days of data collection is known it would
be a good addition to Table 20-1.
(page 98) The big differences in Table 20.2 need to be addressed in the text.
These are helpful suggestions. We have tried to address them in the revised text. Note
the references listed in the comments refer to the original page numbers, chapters and tables.
12O
-------
Appendix B. Response to Peer Review Comments (B)
Draft MOVES2009 Highway Vehicle Population and Activity Data
Peer Review Comments & EPA Response
Dr. Kanok Boriboonsomsin
As part of the MOVES2010 Peer Review process, EPA solicited comments from Dr.
Kanok Boriboonsomsin on the July 2009 draft of report Draft MOVES2009 Highway Vehicle
Population and Activity Data
Dr. Boriboonsomsin is an Assistant Research Engineer (research faculty) at the College
of Engineering -Center for Environmental Research and Technology, University of California at
Riverside. He is also a registered Traffic Engineer in the state of California. His areas of
expertise include transportation planning, vehicle emissions modeling, vehicle activity analysis,
transportation conformity, traffic simulation, and GIS applications in transportation. He has
published extensively in these research areas, and has served as a Principal Investigator (PI)
or Co-Pi on a number of research projects funded by various air and transportation agencies.
He is also a member of the Transportation Research Board's Transportation and Air Quality
Standing Committeewe.
Dr. Boriboonsomsin's comments are copied below, with EPA response in italics.
Introduction
This is a review of the Draft MOVES2009 Highway Vehicle Population and Activity Data report prepared
by the EPA's MOVES team. The reviewer is charged with a set of questions designed to focus the review
on specific areas of the report. These questions are answered in the "Overall Comments" section. Also,
the reviewer is asked to comment on four elements of the report. Therefore, I organize the "Specific
Comments/Questions" section according to these elements. Finally, the reviewer is requested to give
special attention to whether there are alternate data sources or approaches that would better allow the
model to estimate national default values. In response to this request, the dedicated "Potential Data
Sources" section is provided at the end of this review.
Overall Comments
1. What are your recommendations for improving how we model fleet and activity in MOVES and
how we populate the associated national default input data?
In Draft MOVES2009, the national default values for highway vehicle population and activity data have
been updated and improved from the previous version in MOVES2004 in many areas. To name a few, the
SourceTypeAgeDistribution of passenger cars and trucks has been improved with the methodology that
combines sales and scrappage information. Also, the SalesGrowthFactor of future years beyond the
horizon year has been updated with a more realistic assumption. Instead of setting it to 1 (in
MOVES2004), indicating no growth in sales, it is now set to the value of the horizon year (in Draft
MOVES2009), indicating the same growth rate thereafter.
121
-------
To improve the national default values in Draft MOVES2009 further, several recommendations are made
with regards to the methodologies for better estimating the AvgSpeedDistribution and the
StartAllocFactor, the use of other data sources to reinforce the existing estimates, and the consideration of
other idle types besides extended idle, among others. These are discussed in more details in the "Specific
Comments/Questions" section.
2. Are the data sources used to populate the MOVES default values appropriate? Are we missing
any important data sources?
Almost all the critical data sources that exist have been gathered and used in preparing the national
default values. A few data sources that can be used to improve the estimate of extended idling activity are
suggested in the "Specific Comments/Questions" section. In addition, potential data sources for longer-
term consideration are listed at the end of this review.
3. Given your knowledge of U.S. vehicle populations and activity, are the default values
summarized in the report reasonable?
Most of the default values in the report are reasonable. There are some default values that could be based
on more updated data or data from multiple sources. These are discussed in more detail in the "Specific
Comments/Questions" section.
4. Data was not always available where it was needed. In these cases, were the assumptions and
extrapolations used to populate the model appropriate? How could they be improved?
The majority of the assumptions and extrapolations used are appropriate, although a few of them need a
validation or a better justification, especially the assumptions made when estimating the average speed
distributions. Detailed comments are provided in the "Specific Comments/Questions" section below.
5. Overall, are the fleet and activity inputs described here adequate for calculating national
highway vehicle emission inventories?
In their current form, the fleet and activity inputs described in the report are considered to be adequate for
calculating national highway vehicle emission inventories. Of course, these inputs should continue to be
updated with newer and better data once they become available. Note that the many updated tables in the
Draft MOVES2009 highway vehicle population and activity data not only provide better information
about vehicle fleet and activity in the U.S., but also reflect the benefits of the modular design of MOVES
architecture that allows for an easy update of any specific model inputs.
Specific Comments/Questions
1. Clarity of the Presentation
• Page 9: The term "base year" is abruptly introduced here. In Draft MOVES2009, there are now
two base years—1990 and 1999. A paragraph describing this in the Introduction section will be
helpful (such description is provided in the MOVES2004 report, but is excluded in this Draft
MOVES2009 report).
• Page 68: In Table 8-1, should the Midpoint Weight for the Weight ClassID 5 be 250 instead of
350?
122
-------
• Page 73: In Table 10-2 from a mathematical point of view, should the Average Speed for the
Average Speed Range of < 2.5 mph be 1.25 instead of 2.5?
• Page 75: What are the "other twelve categories" mentioned in Section 11.2?
• Page 76: The calculation described in the paragraph prior to the last is too complex to be
understood in plain text. Supplemental equations would be helpful.
• Page 88: It is not clear how the data from the Lutsey report was used to calculate the hourly
distribution of hoteling activity. Additional explanation or some equations will be helpful.
• There are some errors in the report, e.g.
o Page 20: In the Commercial Trucks bullet, "Factor for Calendar year 2002 through
2030..." should read "Factor for Calendar year 2006 through 2030..."
o Page 18: There are errors (likely to be typos) in the numbers of buses in Table 3-10.
Please check accordingly.
o Page 40: The last sentence in the paragraph above Section 7.2.2 is incomplete.
o Page 95: There are errors in the equations. Indexes /' and/ seem to be missing from the
variable VMT.
• There are a few occasions of incorrect call-out of table numbers, i.e.
o Table 1-2 on Page 25. It should be Table 1-1.
o Table 11-3 on Page 76. It should be Table 11-2.
o Also, Table 20-2 on Page 98 is listed twice.
These are helpful suggestion; we will improve the table references and clarify the text.
2. Integration of Information from Multiple Areas
• Page 12: As indicated in the footnote, the value of total trucks in Table 3-2 needs to be corrected.
In doing so, please be cautioned that the current value of the total number of trucks (w/o Puerto
Rico and publically owned vehicles) shown in Table 3-1 (i.e. 81,060,369) does not match up with
the value from the 1999 Highway Statistics (i.e. 81,090,659). Using the value from the 1999
Highway Statistics and the equation provided, the calculated total truck population is 82,391,214.
The reviewer is correct; we mistakenly subtracted the Puerto Rican truck population. This causes a
discrepancy of less than 0.04% in the national total truck population. This error was not fixed for
MOVES2010, but should be fixed when default national populations are updated in the future. Note, the
error should not impact SIP or conformity estimates which are required to use local population data.
• Page 14; The number of school buses could be adjusted to account for Puerto Rican school buses
using the ratio of School and Other Buses to Private and Commercial Buses in MV-10. Then,
multiply this ratio to the Puerto Rican Private and Commercial Buses in MV-1. As a result, the
calculated number of intercity buses will also be reduced.
The reviewer is correct; this approach would cause a small percentage increase in the number of school
buses and would reduce the national population of intercity buses by about 3 percent. This approach was
not usedfor MOVES2010, but should be considered when default national populations are updated in the
future.
123
-------
• Page 72: In the estimation of the AvgSpeedDistribution for urban driving, it is unclear why the
available data from other driving studies (e.g. South Coast, St. Louis, etc.) are not used in
conjunction with the MOBILE6 default? Also, why are the MOBILE6 "ramp" values not
included in the MOVES "urban restricted access" road type? Similarly, why are the MOBEILE6
"local" values not included in the MOVES "urban unrestricted access" road type?
MOBILE6 does not have speed distributions for ramps or local roads, instead these were modeled at a
single speed. Adding them to the distribution would have created artificial peaks at these two speeds.
Computing national average rural speed distributions, as was done for MOVES, is a relatively easy task
only because so little data is available. For urban areas, there is a wealth of information from chase cars
studies, traffic models and other sources, so the challenge is collecting the data, analyzing it, and
weighting it into a national average distribution appropriate for national-scale modeling (SIP and
conformity modelers are required to input local speed data). This was a massive undertaking for
MOBILE6, and it is an area where EPA should invest resources when next updating MOVES activity
inputs.
• Page 85: Care should be exercised when mixing drive schedules from different road types.
Although it may be true often than not that, as assumed in the report, when the average speed is
very low or very high, the road type has little impact on the driving pattern. However, there are
some cases that warrant investigation. For example, the Freeway LOS E schedule (average speed
of 30.5 mph), which involves driving speed of up to 60 mph plus frequent
acceleration/deceleration events, is not likely to be a good representative of the driving on an
urban street with a 30 mph average speed. Based on the available data in hand, statistical tests
could be performed to validate the assumption (see e.g. [Boriboonsomsin et al., 2009]).
Between draft MOVES2009 and final MOVES2010, we made substantial revisions to the drive schedule
mapping provided in the MOVES default to improve representativeness and continuity. The Freeway
LOS E schedule mentioned in the comment is now assigned only to the urban and rural restricted
roadtypes (ie freeways).
• Page 91: For the StartAllocFactor, VMT is probably not the best surrogate for vehicle start. Other
trip generation-related indicators such as the number of vehicles per household are better ones.
For local and regional runs, we expect county vehicle population will be used to estimate starts at the
county level. If users choose to allocate emissions to sub-counties, vehicle population (perhaps computed
as the product of number-of-households andvehicles-per-household) could be used to determine
StartAllocFactor.
For national level runs, where starts must be allocated to 3222 counties, we believe that VMT is an
adequate surrogate for start distributions and one of the few measures that is readily available on a
national basis for every county. To test how well this compares to other methods, we computed fractions
of vehicles for each county using information from the U.S.Census Bureau's 2005-2007 American
Community Survey, three-year estimates of aggregate number of vehicles available per housing unit by
county (B25046). While the survey was lacking data for more than 1000 counties (this elevates
fractions),and comes from different years, the estimates for the counties available correlated well with the
VMT-based estimates (a simple regression of MOVES defaults to ACS values had a linear coefficient of
1.03 and an R2 of 0.96). Some counties where the VMT approach greatly exceeded the census approach
were rural counties with heavy freeway traffic (for example, Caroline County, VA and St. Francis County,
AR). Counties where the household vehicle approach estimates greatly exceeded the VMT-based
estimates included Chicago suburbs and a very large number of counties in Puerto Rico.
124
-------
3. Appropriateness and Completeness of Literature Discussed
• Page 26: The mileage accumulation rate for school buses could be taken from the more updated
1999 School Bus Fleet Fact Book, as it is already available to the EPA.
• Page 79: Data from existing traffic monitoring systems such as the Freeway Performance
Measurement System or PeMS (https://pems.eecs.berkeley.edu/) in California could be useful in
estimating the temporal distributions of VMT.
• Page 85: The data used for developing the 45 light-duty drive schedules is biased toward
California driving. Incorporating additional data from other regions of the country would be
desirable.
• Page 88: There is another survey related truck idling conducted by the American Transportation
Research Institute [Tunnell and Dick, 2006]. A recently published paper [Frey et al., 2008] also
examines the idling activity of long-haul trucks. The EPA may consider incorporating the results
from these sources into the default estimate of truck idling activity. Furthermore, the National
Cooperative Freight Research Program is sponsoring a research project to characterize truck
idling at the regional and national level
(http://144.171.11.40/cmsfeed/TRBNetProiectDisplav.asp?ProiectID=2671). Once completed, it
could be another source of truck idling activity in MOVES.
• Page 89: The American Transportation Research Institute survey [Tunnell and Dick, 2006] also
has data on the anti-idling measures in use. The National Deployment Strategy for Truck Stop
Electrification (TSE) (http://tse.tamu.edu/) has data on existing TSE sites in the U.S.
• Page 92: The National Truck Stop Directory (http://www.truckstops.com/) lists truck stops by
city. Therefore, it should enable a direct determination of IdleAllocFactor by county.
These are good suggestions, but we did not have time to pursue them for MOVES2010 orMOVES2010a.
We will add them to the list of possible data sources for future fleet and activity updates.
4. Appropriateness of the Resulting Data for Use in National-Level Highway Vehicle Emissions
Modeling
• Page 73: In the estimation of the AvgSpeedDistribution for rural driving based on the chase car
data sets, it is described that for each link the average speed is first calculated, and then the
driving time is allocated to one of the speed bins defined in Table 10-2. This approach, although
sounds reasonable in a general sense, suffers from the lost of data variability. For instance, a
driving snippet on a rural unrestricted access link consists of 10 seconds at 30 mph and another
10 seconds at 0 mph (idle at a traffic light). Using the current approach, this 20-second driving
snippet will be allocated to Speed Bin 4 (12.5 mph <= speed < 17.5 mph). However, it is more
appropriate to allocate a half of the snippet to Speed Bin 1 and the other half to Speed Bin 7.
Since the data sets used in this analysis already contain HPMS Functional Class designation on a
second-by-second basis, it is recommended that the speed distributions be estimated directly from
the second-by-second data for each MOVES road type. Furthermore, if the data sets have time
stamp information, then the analysis can be performed for each hour of day, resulting in different
rural speed distributions by time of day.
This comment indicates that the report needs to explain the MOVES algorithm more clearly. The
reviewer is correct that, in his example, the snippet would be allocated to Speed Bin 4. However, that
does not mean that the driving associated with this snippet would be modeled as a constant 15 mph.
Instead, it would be modeled as a weighted average of the operating modes from two driving schedules;
in this case, a small fraction of driving schedule 101 (average speed of 2.5, minimum speed ofO, and
maximum speed of 10 mph) and a larger fraction of schedule 1041 (average speed of 18.5, minimum
125
-------
speed ofO and maximum speed of 50.3 mph). Thus, the average speed distribution is used to map to the
MOVES drive schedules, but does not restrict activity to the average speeds. Of course, to exactly model
a single drive schedule, the user would have to enter the desired drive schedule or operating mode
directly.
• Page 74: Using the speed distributions of light-duty vehicles to represent those of heavy-duty
vehicles is not appropriate, especially for the restricted access road types. In many states, the
freeway speed limits for trucks are different than those for cars (see
http://en.wikipedia.org/wiki/Speed limits in the United States). Therefore, it is recommended
that the speed distributions of heavy-duty vehicles be derived from available measured truck
activity data (e.g. [Battel, 1999]). In addition, the speed limits for cars also vary by state. Hence,
it is recommended that an adjustment be made to the speed distributions of light-duty vehicles
derived from the California data. This may be performed by synthesizing the speed distributions
for each state by shifting the California distributions according to the differences in speed limits.
Then, the national speed distributions can be calculated by weighting the speed distributions for
each state by VMT.
These are good suggestions, but we did not have time to pursue them for MOVES2010 orMOVES2010a.
We will add them to the list of recommendations for future fleet and activity updates.
Note that SIP and conformity users are expected to enter local speed distributions directly and can
account for state and local speed limits directly.
Also note that even when light and heavy duty vehicles are modeled with the same average speed
distribution, they do not have the same driving schedules. For example, the highest speed driveschedule
for heavy duty trucks has an average speed of 72, a maximum speed of 81 and a minimum speed of
63mph. For light duty vehicles, the highest speed driveschedule has an average speed of 76, a maximum
of 90, and a minimum of 66 mph.
• Page 81: I believe that the hourly VMT fractions for heavy-duty trucks, especially for long-haul
trucks, would be significantly different from those for cars. This could be important for air quality
modelers as different hourly VMT fractions for heavy-duty trucks would affect the diurnal NOX
and PM profiles.
We agree. We did not have time to incorporate this for MOVES2010, but we encourage local modelers to
use local data on differences in hourly VMT fractions by sourcetype and we hope to differentiate these
distributions in future MOVES updates.
• Page 84: In Table 14-1, there is only the Refuse Truck Urban schedule (average speed of 2.2
mph) for refuse trucks on unrestricted access road types. This may not be sufficient and Heavy
Heavy-Duty Non-Freeway schedules may be needed to help supplement the driving at higher
speeds.
777/5 is a good point. We have added additional cycles to account for higher speed refuse truck travel.
• Page 88: MOVES now only models extended idle of long-haul trucks. However, other types of
idle may also worth consideration, such as idle while loading/unloading (buses and commercial
trucks), idle while in operation (parcel trucks and refuse trucks), idle in parking lots or at drive-
thru restaurants (mostly light-duty vehicles), etc. For these types of idle, while each occurrence
may not be long in duration, their frequent occurrences could still add up to a significant amount
of idling time.
126
-------
This comment indicates that the report needs to explain the MOVES algorithm more clearly. Idling in
normal operation is included in the driving cycles. So, for example, the LD Low Speed 1 cycle (cycle 101)
has 602 seconds; 280 seconds at 0 mph. Modelers wishing to model specific operation with a high
percentage of idle time may use the project-level modeling approach and enter their own driveschedules
or operating mode distributions. Heavy-duty extended idle differs from this regular idle because it is
characterized by a higher engine speed, and thus higher emissions. We will attempt to explain this more
clearly in the final report.
Potential Data Sources
• International Registration Plan (http://www.irponline.org/) and International Fuel Tax Agreement
(http://www.iftach.org/): These are programs that facilitate commercial vehicle registration and
fuel tax reporting across multiple jurisdictions though a single system. Participating fleets are
required to log Individual Vehicle Distance Record (IDVR), which includes information
regarding distance traveled in each jurisdiction for each vehicle in the fleets. The IDVR could be
useful in estimating the age distribution, relative mileage accumulation rate, and possibly
SHOAllocFactor of intercity buses and long-haul trucks.
• Truck Engine Control Units (ECUs): The ECUs of heavy-duty diesel engines are capable of
monitoring and storing a variety of engine operation data such as fuel consumption, time at idle,
active fault codes (i.e. problems in the engine), the amount of time that a truck spent in various
speed bins, etc. The information is typically kept for the time period following the last ECU reset
and can span the life of the engine. Once downloaded, the information can be used for several
analyses such as the estimation of temporal distribution of VMT [Earth et al., 2009].
• Fleet Management Systems: Many trucking companies equip their truck fleets with a fleet
management system that has tracking and telematics capabilities so that engine operating
parameters (from ECU) as well as positioning information (from GPS) can be wirelessly
transmitted to a computer server on a periodic basis. The data from these trucking companies are
then compiled by data aggregators into very large databases such as the Highway Visibility
System (http://www.calmartelematics.com/hivis.php). If accessible, these databases along with
proper fleet characterization will be useful in developing many of the MOVES highway vehicle
activity data tables.
• Traffic Information Providers: In order to get real-time traffic info, traffic information providers
maintain a network of probe vehicles (i.e. GPS-enabled) traveling around the country, for
example, the Smart dust Network (http://www.inrix.com/techdustnetwork.asp). These probe
vehicles encompass not only passenger cars, but also other variety of vehicle types including
commercial fleet, delivery, and taxi vehicles. If accessible, these probe vehicle data along with
proper vehicle characterization will be useful in developing many of the MOVES highway
vehicle activity data tables.
• Traffic Monitoring Systems: Many transportation agencies have developed and maintain a traffic
monitoring system for their jurisdiction. Examples include California's Freeway Performance
Measurement System or PeMS (https://pems.eecs.berkeley.edu/) and Houston TranStar
(http://www.houstontranstar.org/). Although the spatial coverage of these systems is not as wide
as that of the HPMS, they provide data with better resolution (both spatially and temporally) than
the HPMS for the areas they cover. Thus, their data can be used to supplement the HPMS data.
127
-------
• Weigh-in-Motion: This data source was mentioned in the MOVES2004 peer review. It is worth
mentioning again that it could be useful for estimating heavy-duty truck activity on freeways.
These are good suggestions, but we did not have time to pursue them for MOVES2010. We will add them
to the list of recommendations for future fleet and activity update.
References
Barth, M., Boriboonsomsin, K., and Scora, G. (2009). Evaluation and validation of CO 2 estimates of
EMFAC and OFFROAD through vehicle activity analysis. Final report to the California Air
Resources Board, June.
Battelle. Heavy-Duty Truck Activity Data. Prepared for Federal Flighway Administration, April 30, 1999.
Boriboonsomsin, K., Barth, M., and Xu, H. (2009). Improvements to on-road mobile emissions modeling
of freeways with high-occupancy vehicle facilities. Proceedings of the 88th Annual Meeting of
the Transportation Research Board (DVD), Washington, DC, January 11-15.
Frey, H. C., P.-Y. Kuo, and C. Villa. (2008). Methodology for characterization of long-haul truck idling
activity under real-world conditions. Transportation Research Part D, 13, 516-523.
Tunnell, M. A. and Dick, V. (2006). Idle reduction technology: Fleet preferences survey. Prepared for
New York State Energy Research and Development Authority, February.
128
------- |