* — \
*1 PROt^
Documentation for the Onroad National
Emissions Inventory (NEI) for Base Years
(1970-2002)
-------
-------
EP A-454/B-20-018
January 2004
Documentation for the Onroad National Emissions Inventory (NEI) for Base Years (1970-
2002)
Prepared by:
E.H. Pechan & Associates, Inc.
5528-B Hempstead Way
Springfield, VA 22151
Prepared for:
Office of Air Quality Planning and Standards
Emission Factor and Inventory Group
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Quality Assessment Division
Research Triangle Park, NC
-------
[This page intentionally left blank.]
-------
CONTENTS
Page
A. INTRODUCTION 1
1. What Is the National Emissions Inventory? 1
2. What Is the Purpose of This Document? 1
3. Which Sources Does EPA Include in the On-road Vehicle Category? 1
B. WHAT IS EPA'S CURRENT METHODOLOGY FOR DEVELOPING CRITERIA POLLUTANT
AND HAP EMISSION ESTIMATES FOR ON-ROAD VEHICLES FOR THE YEARS 1970
THROUGH 2002? 1
1. What Data Does EPA Use in Estimating VMT? 3
a. How Does EPA Estimate Vehicle Miles Traveled (VMT)? 3
b. How Does EPA Develop Projected 2002 VMT? 6
c. How Were State VMT Estimates Incorporated into the NEI? 7
2. How Does EPA Develop On-road Criteria Pollutant Emission Factors? 8
a. What Temperature Data Does EPA Input to the MOBILE Model? 8
b. How Does EPA Calculate the Monthly RVP Inputs? 9
c. What Diesel Fuel Sulfur Inputs Does EPA Model? 11
d. What Header and Run Information Does EPA Include in the MOBILE6 Input Files? 11
e. What Speed Inputs and Facility Types Does EPA Use? 12
f What Altitude Inputs Does EPA Use? 13
g. What Registration Distributions by Vehicle Age Does EPA Use? 13
h. How Does EPA Model On-road Control Programs? 13
3. How Were Criteria Emissions Data Supplied by the States Incorporated? 15
4. How Were Criteria Emissions for Puerto Rico and the US Virgin Islands Calculated? 16
a. How Does EPA Calculate VMT for Puerto Rico and the USVT? 16
b. What Inputs were Used to Produce MOBILE6 Input Files for Puerto Rico and the USVI?
17
5. How Does EPA Calculate Emissions for Hazardous Air Pollutants for the On-road Vehicle
Category? 17
a. How Were the 1990, 1996, and 1999 MOBILE6 HAP Input Files Developed? 19
b. How Were the 2002 MOBILE6 HAP Input Files Developed? 21
6. How Were HAP Emissions Data Supplied by the States Incorporated? 22
7. How Were Criteria Pollutant Emissions for Interim Years Calculated? 23
8. What Caveats Should be Considered when Comparing Emissions Across Years 23
C. REFERENCES 24
Attachment A
iii
-------
County-Specific Fuel Parameters for 1990, 1996, and 1999 Toxic Emissions Modeling (Preparation for
MOBILE6.2 Model Runs)
Attachment B
EPA NTI 1999 On-Road Hazard Air Pollution Emissions Estimation Calculation Methodology MOBILE6.2
HAP Input File Creation
Attachment C
1999 Onroad Dioxin/Furan Emission Estimating Methodology
Attachment D
Potential Approaches for Developing a Mercury Inventory for Mobile Sources
Tables
Table 1. Methods Used to Develop Emission Estimates for Onroad Vehicle Sources 26
Table 2. Allocation of VMT from HPMS Vehicle Categories to MOBILE6 Vehicle Types for 1999 28
Table 3. VMT Seasonal and Monthly Temporal Allocation Factors 28
Table 4. Fractions for Converting VMT from 8 MOBILE5 Vehicle Types to 28 MOBILE6 Vehicle Types 29
Table 5. Cities Used for 1999 Temperature Data Modeling 30
Table 6. Surrogate City Assignment 31
Table 7. Substitute Survey City Assignment 35
Table 8. Average Speeds by Road Type and Vehicle Type
(mph) 36
Table 10. Oxygenated Fuel Modeling Parameters 40
Table 11. Number of Vehicles in Puerto Rico by Vehicle Type and Percent of Total 41
Table 13. Population Estimates in St. Thomas, St. John, and St Croix 41
Table 14. Percentage of Total USVI VMT by Roadway Type and Island 41
Table 15. 25 Year Trend of Vehicle Registrations and New Sales in Puerto Rico 42
Table 16. Average Temperature Data for Puerto Rico and USVI (°F) 42
iv
-------
A. INTRODUCTION
1. What Is the National Emissions Inventory?
The National Emissions Inventory (NEI) is a comprehensive inventory covering all criteria pollutants and
hazardous air pollutants (HAPs) for the 50 United States, Washington DC, Puerto Rico, and US Virgin Islands.
The NEI was created by the U.S. Environmental Protection Agency's (EPA's) Emission Factor and Inventory
Group (EFIG) in Research Triangle Park, North Carolina.
The NEI will be used to support air quality modeling, rule development, international reporting, air quality
trends analysis, and other activities. To this end, we, the EPA, established a goal to compile comprehensive
emissions data in the NEI for criteria and HAPs for nonroad mobile, onroad mobile, point, and nonpoint
sources.
2. What Is the Purpose of This Document?
This report summarizes the procedures we used to estimate emissions for the onroad mobile source category
component of EPA's NEI. Criteria pollutant emission estimates for onroad mobile sources are described in this
report for the years 1970, 1975, and 1978 through 2002. HAP emission estimates for onroad mobile sources
are described here for the years 1990, 1996, 1999, and 2002. However, the focus of this documentation is on
defining the methodologies and data used in version 3 of the 1999 NEI for criteria pollutants and HAPs, as well
as the draft 2002 NEI for onroad sources. Table 1 summarizes the methods applied and the pollutants for
which emissions were estimated for all onroad sources. More information about EPA's NEI and plans for
making revisions to these inventories can be found at
http://www.epa.gov/ttn/chief/net/index.html
3. Which Sources Does EPA Include in the On-road Vehicle Category?
The "on-road vehicles" category includes motorized vehicles that are normally operated on public roadways.
This includes passenger cars, motorcycles, minivans, sport-utility vehicles, light-duty trucks, heavy-duty trucks,
and buses.
B. WHAT IS EPA'S CURRENT METHODOLOGY FOR DEVELOPING
CRITERIA POLLUTANT AND HAP EMISSION ESTIMATES FOR ON-
ROAD VEHICLES FOR THE YEARS 1970 THROUGH 2002?
EPA calculated on-road vehicle criteria pollutant emissions for the years 1978, 1987, 1990, 1996, 1999,
2000, 2001, and 2002 using the final version of EPA's MOBILE6.2 mobile source emission factor model. The
onroad criteria pollutant emission estimates for 1970 and 1975 and the remaining years from 1979 through
2002 were calculated by interpolation. The MOBILE6.2 model was also used to calculate HAP emission
factors for 1990, 1996, 1999, and 2002. The criteria pollutant emissions for the years through 2001 were
calculated using a methodology that is consistent with that used in previous versions of EPA's Trends report1.
Note that due to the timing of the availability of the MOBILE6 model modules and supplemental data used for
1
-------
these estimates, onroad criteria and HAP emissions were not developed in tandem and therefore are not in
complete agreement. The draft 2002 NEI is the first that calculates criteria pollutant and HAP emission factors
on a consistent basis. MOBILE6.2 requires more detailed fuel parameter data that was not available for use in
earlier work to estimate criteria emissions.
For the years that MOBILE6.2 was used to calculate the onroad emissin factors, the on-road emissions
inventories for all criteria pollutants (CO, NOx, VOC, PM-10, PM-2.5, S02, and NH3) are calculated by
multiplying an appropriate MOBILE6.2 emission factor in grams per mile by the corresponding VMT in millions
of miles, and then converting the product to units of tons of emissions. Emission estimates include calculations
by month, county, road type, and vehicle type, with VOC broken down by exhaust and evaporative emissions
and PM-10 and PM-2.5 broken down by exhaust, brake wear, and tire wear emissions. The HAP emissions
were calculated in a similar manner, but emission factors for all years except 2002 were calculated by season
rather than month. For 2002, the HAP emissions were calculated on a monthly level, in a manner consistent
with the calculation of the criteria pollutant emission factors. The MOBILE6.2 model used to calculate
emission factors for both criteria pollutants and HAPs is the publicly available version from EPA's Office of Air
Transportation and Quality's (OTAQ) website (top://www.epa.gov/otaq/m6.htrh). This model incorporates
MOBILE6.0,2 which is used to estimate emission factors of VOC, CO, and NOx, MOBILE6.1,3 which is used
to calculate emission factors of PM-10, PM-2.5, S02, and NH3, and MOBILE6.2, which is used to calculate
onroad emission factors for HAPs. Prior to the release of MOBILE6, the particulate and S02 emission factors
were previously calculated using EPA's PART5 model4 and emission factors for NH3 were calculated at the
national level by vehicle type. The term MOBILE6 is used in this document to refer to the combined
MOBILE6.0, MOBILE6.1, and MOBILE6.2 model. The October 2002 version of the model was used for all
criteria pollutant and HAP emission factor model runs made for the NEI during 2003.
EPA does not calculate emission factors separately for every county. To determine the emission factor sets to
be modeled in each State, EPA prepared a county-level database that includes information on non-default
inputs to be modeled. For each county, the control programs applicable in the year to be modeled were
indicated in this database. Next, EPA determined for each State all unique combinations of control programs
and other non-default inputs in that year. MOBILE6 model runs were then made modeling each of these
unique combinations. Each combination was identified using the county code of one of the counties with this
combination of controls and inputs. To apply the emission factors to the appropriate counties, EPA developed
a county correspondence file which mapped all counties with the same unique set of input data and control
programs to the MOBILE6 emission factors modeled for the county representing that unique combination of
inputs and control programs. For some States, EPA applied a single set of emission factors to all counties in
the State, while for other States, EPA calculated a separate set of emission factors for each county. Most
States, though, had several sets of emission factors calculated for the State, with each set applying to one or
more counties within the State. These emission factors were then multiplied by the corresponding activity data,
vehicle miles traveled (VMT) at the county, monthly (or seasonal for the pre-2002 HAPs), roadway type, 28-
vehicle type level of detail. This process and the methods used for developing the necessary data inputs and
VMT are discussed below.
EPA gave states the ability to provide VMT or emissions to be included in these calculations. The data that
were accepted and used are discussed towards the end of this document. Onroad emissions for Puerto Rico
2
-------
and the US Virgin Islands were calculated only for 1999 and 2002. The procedures used to calculate VMT
and emission factors for these States are discussed separately.
1. What Data Does EPA Use in Estimating VMT?
To develop VMT for the NEI, EPA relies on data supplied by the Federal Highway Administration (FHWA)
and publicly available data from FHWA's Highway Statistics5 series. The procedures discussed here focus
on 1999, but the same procedures were applied in all the years that MOBILE6 was used to calculate emission
factors. Only the procedures for calculating the 2002 VMT, which is considered a projection year since actual
2002 VMT at the level of detail used in the other years, differ from those discussed here. The calculation of the
2002 VMT is discussed separately at the end of this section. From Highway Statistics 1999, EPA uses
Table VM-2 "Functional System Travel - 1999; Annual Vehicle-Miles
rhttp://www.fhwa.dot.gov/ohim/hs99/tables/vm2.pdf). This table contains state-level summaries of miles of
annual travel in each State by functional system and by rural and urban areas. Rural VMT is provided on a
state level for the following six roadway types: interstate, other principal arterial, minor arterial, major collector,
minor collector, and local. Urban VMT is provided on a state level for the following six roadway types:
interstate, other freeways and expressways, other principal arterial, minor arterial, collector, and local. EPA
also uses Table VM-1 "Annual Vehicle Distance Traveled in Miles and Related Data - 1999; by Highway
Category and Vehicle Type" from Highway Statistics 1999. This table provides annual VMT separated by
rural and urban areas broken down into the following vehicle categories: passenger cars, motorcycles, buses,
other 2-axle 4-tire vehicles, single-unit 2-axle 6-tire or more trucks, and combination trucks. In addition to
these publicly available tables, FHWA provides EPA with daily VMT by urban area (areas with a population of
50,000 or more) in each of the six urban roadway categories as listed for Table VM-2, broken down by urban
area and state. This data is similar to that in Table HM-71 from Highway Statistics 1999 with the exception
that Table HM-71 does not break down multi-state urban areas into the portion in each state. Finally, FHWA
provides EPA with a data file containing roadway mileage by county and each of the 12 roadway classes listed
above.
In addition to the FHWA data, EPA uses 1990 population data in developing the VMT data base. The EPA
relies upon two tables in the Bureau of the Census 1990 Number of Inhabitants (CNOI) documents6 as the
source for population data for the years 1999. The first is "Table 3: Population of Counties by Urban and
Rural Residence." This table lists the urban population living inside census-defined urban areas, the urban
population living outside census-defined urban areas, and the rural population for each county. The other is
"Table 13: Population of Urban Areas." This table divides an urban area's population among the counties that
contain portions of that urban area.
a. How Does EPA Estimate Vehicle Miles Traveled (VMT) ?
Vehicle miles traveled (VMT) is the activity factor EPA uses to estimate on-road vehicle emissions;
therefore, the development of a VMT database is critical to the estimation process. Starting with State VMT
totals for each year, EPA allocates VMT by county, roadway type, and vehicle type. There are four basic
steps in this process: (1) allocate state-level rural VMT by roadway type to county/roadway type level; (2)
allocate large urban area VMT by roadway type to the county/roadway type level; (3) allocate remaining state-
level small urban VMT by roadway type to the county/roadway type level; and (4) allocate county/roadway
3
-------
type level VMT to each of the 28 MOBILE6 vehicle classes for each county and roadway type combination.
Each of these steps is described in more detail in the following sections.
i. How Does EPA Estimate 1999 Rural VMT?
To calculate rural VMT by county for 1999, EPA first calculates each county's fraction of the state's total rural
interstate roadway mileage. Next, EPA calculates each county's rural interstate VMT by multiplying the
county's rural interstate roadway mileage fraction by the state's 1999 rural interstate VMT from Table VM-2.
Equation 1 shows this calculation.
ML
VMT„~ = VMT„~ x
** MIL
ia,c
(Eq. 1)
where: VMTrjc =
Rural
vmtris =
Rural
MILri,c =
Rural
MILri,s =
Rural
EPA calculates VMT for the remaining five rural roadway types in a similar manner. However, rural county
population data from CNOI Table 3 is the primary surrogate for distributing VMT by county for these roadway
types. In addition, VMT for a specific roadway type is distributed only to counties with nonzero roadway
mileage of the specified roadway type, based on the roadway mileage file provided by FHWA. Thus, rural
population within a state is totaled individually for each of the rural roadway types, including only population
from counties with nonzero roadway mileage of the specified roadway type. For the local roadway category,
VMT is distributed strictly by rural population, assuming that all counties with rural populations have mileage in
the rural local roadway category. Equation 2 shows the equation used to calculate county-level VMT on rural
roadway types other than rural interstates.
POP,
= VMTjzj. x
IXfi
(Eq. 2)
JXfi
where: VMTRX c = VMT on rural roadtype X in county C (calculated)
VMTRX s = VMT on rural roadtype X, State total (Highway Statistics Table VM-2)
POPr,c = Rural population in county C with nonzero mileage from rural roadway type X
(CNOI) (0 if zero mileage from rural roadway type X in county C)
POPR S = Rural population, State total of all counties with nonzero mileage from rural roadway
type X (CNOI)
4
-------
ii. How Does EPA Estimate 1999 Urban Area VMT?
To allocate daily VMT totals by road type for each individual urban area to the corresponding counties, EPA
uses data from CNOI Table 13 to calculate the fraction of population in each county containing a portion of a
given urban. As shown in Equation 3, EPA then calculates each county's share of an urban area's VMT by
distributing urban area VMT from FHWA's urban area VMT data base based on the percentage of the urban
area's population in each county. As with the rural VMT allocations, VMT for a specific roadway type is
distributed only to counties with nonzero roadway mileage of the specified roadway type, based on the
roadway mileage file provided by FHWA. Thus, urban population within a state is totaled individually for each
of the rural roadway types, including only population from counties with nonzero roadway mileage of the
specified roadway type. For the local roadway category, VMT is distributed strictly by urban population,
assuming that all counties with rural populations have mileage in the rural local roadway category.
POPnrj.
' VMTau « (Eq 3)
where: VMTUX c = Urban area's VMT on roadway type X in county C (calculated)
VMT[;x,s = Urban area's VMT on roadway type X for total urban area A contained in state
(FHWA)
POPux c = Urban area's population in county C with nonzero mileage from urban roadway
type X (CNOI)
POPux A = Urban area's population for total urban area A contained in state totaled for all
counties with nonzero mileage from urban roadway type X (CNOI)
As the urban area VMT provided by FHWA is reported in terms of daily VMT, the final urban area VMT by
county and roadway type is converted to millions of miles of annual VMT by multiplying the daily VMT by 365
and dividing by 1,000,000.
iii. How Does EPA Estimate 1999 Small Urban VMT?
The procedure for calculating each county's small urban VMT in 1999, is similar to that described above for
calculating each county's rural VMT, with one additional step. In this case, the resultant average annual VMT
for urban areas, calculated as discussed above, and totaled by state and roadway type, is subtracted from the
total urban VMT by state and roadway type that is reported in Table VM-2 of Highway Statistics 1999.5
This calculation results in small urban VMT by state and roadway type. Next, EPA uses data from CNOI
Table 3 on the urban population living outside census-defined urban areas to calculate the percentage of the
state's small urban population living in each county. Finally, as with the rural VMT, VMT for a specific
roadway type is distributed to counties with nonzero roadway mileage of the specified roadway type, based on
the roadway mileage file provided by FHWA. For the local roadway category, VMT is distributed strictly by
population, assuming that all counties with small urban populations have mileage in the urban local roadway
category. Equation 4 shows the equation used to calculate county-level VMT on small urban roadway types.
5
-------
POP,
VMT„e = VhfTgrr *
«;e aj: ™
Ofi
(Eq. 4)
jayr
where: VMTSX c = VMT on small urban roadtype X in county C (calculated)
VMTSX s = VMT on small urban roadtype X, State total (obtained by subtracting large urban
VMT from total urban VMT from Highway Statistics Table VM-2)
POPSX( = Small urban population in county C with nonzero mileage from urban roadway type X
(CNOI) (0 if zero mileage from urban roadway type X in county C)
POPsxs = Small urban population, State total of all counties with nonzero mileage from urban
roadway type X (CNOI)
iv. How Does EPA Determine 1999 VMT by Vehicle Type?
To calculate 1999 VMT at the county/roadway type/vehicle type level, the VMT totals by county and roadway
type need to be allocated among the 28 MOBILE6 vehicle types. This was done based on the distribution of
the 1999 rural and urban VMT among the six HPMS vehicle types found in Table VM-1 ("Annual Vehicle
Distance Traveled in Mies and Related Data - 1999 - by Highway Category and Vehicle Type") of FHWA's
Highway Statistics 1999i ('http://www.fhwa.dot.aov/ohim/hs99/tables/vin 1 .pdf) and a mapping of these
HPMS vehicle categories to the 28 M0BILE6 vehicle types, provided by OTAQ. First, the VMT totals for
each of the six HPMS vehicle categories were calculated as a fraction of the total VMT. This calculation was
performed separately for the rural VMT and the urban VMT. The resulting 1999 VMT fractions for rural
VMT and urban VMT are shown in Table 2. Next, EPA assigned each of the 28 M0BILE6 vehicle types to
one of the 6 HPMS vehicle categories, as shown in Table 2. Using the default M0BILE6 VMT fractions for
1999, the M0BILE6 VMT fractions were renormalized among all M0BILE6 vehicle types mapped to a given
HPMS vehicle category. Then the HPMS VMT fractions for rural and urban roads were separately multiplied
by the renormalized M0BILE6 VMT fractions for all M0BILE6 vehicle types included within a given HPMS
vehicle category. For example, Table 2 shows that the HPMS Passenger Car vehicle category includes the
M0BILE6 LDGV and LDDV vehicle types. Therefore, the default 1999 M0BILE6 VMT fraction for
LDGVs was divided by the sum of the LDGV and LDDV default 1999 M0BILE6 VMT fractions. This
number was then multiplied by the HPMS VMT fraction for Passenger Cars (0.5499 for rural roads and
0.6048 for urban roads). This resulted in a 1999 LDGV VMT fraction on rural roads of 0.5483 and 0.6030
on urban roads. Table 2 lists the resulting rural and urban VMT fractions for 1999 for each of the MOBILE6
vehicle types. Finally, each of the VMT records in the 1999 VMT data base, at the state/county/roadway type
level of detail was then multiplied by the fraction of VMT in each of the corresponding MOBILE6 vehicle type
categories to obtain total annual VMT at the state/county/roadway type/MOBILE6 vehicle type level.
The resulting annual county-level, vehicle, and roadway type-specific VMT data were temporally allocated to
months during the emission calculations. EPA used seasonal 1985 National Acid Precipitation Assessment
Program (NAPAP) temporal allocation factors7 to apportion the VMT to the four seasons. Monthly VMT
data were obtained using a ratio between the number of days in a month and the number of days in the
corresponding season. These temporal factors are shown in Table 3.
6
-------
b. How Does EPA Develop Projected 2002 VMT?
VMT data for 2002 are estimated by applying national VMT growth factors at the roadway type level to the
2001 VMT database. The 2001 VMT database was developed following the same procedures as discussed
above for 1999. The FHWA's Traffic Volume Trends reports
(http://www.fhwa.dot.gov/ohim/tvtw/tvtpage.htm) provide data at the roadway type level of detail comparing
national VMT from 2001 with preliminary data for 2002. EPA divided the VMT data for 2002 by the 2001
data for the same roadway type to estimate a VMT growth factor from 2001 to 2002 for that roadway type.
These growth factors were then multiplied by all VMT in the detailed (by state, county, vehicle type, and
roadway type) 2001 VMT database with the corresponding roadway type. These growth factors ranged from
3.5 percent on rural interstates to 1.1 percent on rural collectors and local roads, and urban freeways and
expressways and arterial roadways.
c. How Were State VMT Estimates Incorporated into the NEI?
For Version 3 of the 1999 NEI, 12 State or local agencies submitted 1999 VMT data that was accepted by
EPA for incorporation into the NEI. VMT data were submitted for all counties in the following States for
1999: Alabama, California, Colorado, Maine, Massachusetts, Mississippi, Utah, Oregon, Virginia, and West
Virginia. Additionally, VMT data submitted for Maricopa County, Arizona and Hamilton County, Tennessee
were submitted and accepted by EPA for incorporation into the 1999 NEI. California also submitted VMT
data for 2000 and 2001. Of the 1999 VMT submittals, the VMT data were submitted at the 8 vehicle type
and 12 roadway type level of detail in all cases except for California and Oregon. The VMT data provided by
California for 1999, 2000, and 2001 and the VMT data submitted by Oregon were at the 8 vehicle type level
of detail, but did not break the VMT out at the roadway type level of detail. The procedures followed to
expand the VMT for all of these States to the 28 vehicle type level and 12 roadway type level is discussed
below. The VMT data for each state was first converted to units of million miles, where necessary.
All of these State and local agency VMT submittals needed to be expanded from 8 vehicle types to 28 vehicle
types to be consistent with the VMT calculated based on the HPMS data for the remaining States. First, each
of the 28 MOBILE6 vehicle types was mapped to one of the 8 MOBILE5 vehicle types (LDGV, LDGT1,
LDGT2, HDGV, LDDV, LDDT, HDDV, and MC) based on vehicle weight. This was a straightforward
process since the MOBILE6 vehicle types are subsets of the MOBILE5 vehicle types. Table 4 shows which
MOBILE5 vehicle category corresponds to each of the MOBILE6 vehicle categories. The default MOBILE6
VMT fractions for 1999 were summed for each of the vehicles within a MOBILE5 vehicle type category and
then each of the MOBILE6 VMT fractions within this group was renormalized by dividing the default
MOBILE6 VMT fraction by the sum of the MOBILE6 VMT fractions of all MOBILE6 vehicle types included
in the MOBILE5 vehicle category. Table 4 shows the mapping of the MOBILE6 to the MOBILE5 vehicle
categories, the default 1999 MOBILE6 VMT fractions, and the resulting VMT fractions for all MOBILE6
vehicle types within each MOBILE5 vehicle category. Then, a new VMT database was created at the 28
vehicle type level by multiplying the State-supplied VMT for a given MOBILE5 vehicle type by each of the
fractions for that MOBILE5 vehicle type in Table 4 to create VMT records for each of the 28 MOBILE6
vehicle types. For example, if County C in State S had 10.0 million VMT on rural interstates for HDDVs
(SCC=2230070110), that VMT record would be replaced by 10 new VMT records all in County C, State S,
on rural interstates as follows: HDDV2b 1.194 million miles, HDDV3 0.347 million miles, HDDV4 0.268
7
-------
million miles, HDDV5 0.113 million miles, HDDV6 0.682 million miles, HDDV7 1.042 million miles,
HDDV8A 1.334 million miles, HDDV8B 4.753 million miles, HDDBT 0.114 million miles, and HDDBS 0.154
million miles. The California 2000 and 2001 VMT data were expanded to the 28 vehicle type level in the same
manner, except the default MOBILE6 VMT fractions for 2000 and 2001 replaced those for 1999.
One additional step was needed to fully process the VMT data provided by California and Oregon. The VMT
data supplied by these States were not broken down by roadway type. These VMT data were expanded to
the 12 HPMs roadway types using the VMT database developed as discussed above using the HPMS VMT
data. The HPMS-based VMT data were totaled by State, county, and vehicle type. The State-supplied
California and Oregon VMT databases, expanded to 28 vehicle types, were matched by State/county/vehicle
type to the HPMS-based database and each record in the State VMT database was replaced by up to 12 new
VMT records by multiplying the State supplied VMT at the State/county/vehicle type level of detail by the
HPMS-based VMT at the State/county/vehicle type/roadway type level of detail and then dividing by the
HPMS-based VMT at the State/county/vehicle type level of detail.
For 2001 and 2002, all of the State-based 1999 VMT data were projected forward to replace the HPMS-
based VMT data for these years. Again, a ratio approach was applied using the growth in the HPMS-based
VMT data to develop the growth factors. Each record in the 1999 State-based VMT database at the county,
28 vehicle type, and 12 roadway type level of detail was multiplied by the corresponding ratio of the 2001
VMT or 2002 VMT to the 1999 VMT from the HPMS-based VMT database also by county, 28 vehicle
types and 12 roadway types. These projected State-based VMT data then replaced the corresponding
HPMS-based VMT in the 2001 and 2002 VMT databases. California supplied actual 2001 VMT, so in that
case, the actual 2001 California VMT replaced the 2001 HPMS-based VMT, rather than the projected State-
based VMT. Onroad emissions for 2000 were not recalculated during 2003, so the 2000 VMT are the
HPMS-based VMT with no State data incorporated.
2. How Does EPA Develop On-road Criteria Pollutant Emission Factors?
EPA used the MOBILE6 model2'3 to calculate 1978, 1987, 1990, 1996, 1999, 2000, 2001, and 2002 criteria
pollutant emission factors for on-road sources. More specifically, EPA modeled exhaust VOC, evaporative
VOC (which includes resting loss, running loss, and evaporative emissions), exhaust NOx, exhaust CO, exhaust
S02, exhaust PM-10, PM-10 from brake wear, PM-10 from tire wear, exhaust PM-2.5, PM-2.5 from brake
wear, PM-2.5 from tire wear, and exhaust NH3. VOC emissions include aldehydes and hydrocarbons
measured by Flame Ionization Detector (FID) testing. The MOBILE6 criteria pollutant emission factors are all
expressed as grams of pollutant per vehicle mile traveled (VMT). The MOBILE6 model takes into
consideration a number of parameters in tailoring emission factor calculations. A discussion of how EPA
develops these parameters follows. Note that due to the timing of the availability of the MOBILE model
modules and supplemental data used for these estimates and resource constraints, onroad criteria and HAP
emissions were not developed in tandem and therefore are not in complete agreement. For example,
MOBILE6 requires more detailed fuel parameter data that was not available for use in earlier work to estimate
criteria emissions. Starting with 2002, however, a consistent approach was applied to the MOBILE6 modeling
of criteria pollutants and HAPs. Specifics of the HAP emission factor modeling are discussed separately
below.
8
-------
a. What Temperature Data Does EPA Input to the MOBILE Model?
The temperature data inputs to the MOBILE model include monthly average daily maximum and minimum
temperature for each State for each year modeled. These data were obtained from The National Climatic Data
Center.8 EPA selected one city from each State to represent that particular State's temperature conditions.
Each chosen city is thought to be the most representative of the average conditions within the State. Generally
this means either centrally located cities or, in States with a majority of VMT clustered in one area, the most
populous cities. Due to the great temperature variation and the wide VMT distribution throughout California,
EPA divides California into two geographic regions, with Los Angeles representing the southern and interior
portions of the State and San Francisco representing the northern coastal region of the State. Table 5 lists the
cities used to represent each State's temperature conditions in 1999. In most cases, temperature data from
these same cities were used in all of the years modeled. However, in some instances, these sites were not used
in other years, or a complete set of data could not be obtained. In these cases, data from a nearby site with a
complete data set were substituted.
In cases where temperature data is missing for a month or more, EPA relies on 30-year average monthly
maximum and minimum temperature values reported by the Department of Commerce's Statistical Abstracts.9
The temperature range for input to the MOBILE6 model is 0°F to 100°F for the minimum daily temperatures
and 10°F to 120°F for the maximum daily temperatures. In the few cases where temperatures fall outside of
these ranges, EPA substitutes the endpoint of the range for the actual temperatures.
b. How Does EPA Calculate the Monthly RVP Inputs?
Allocating monthly Reid vapor pressure (RVP) values for each State is an important part of the MOBILE
modeling process. To determine these values, EPA uses RVP survey data to apply RVP values to each state
during the non-ozone season months. EPA then uses data on reformulated gasoline programs, low RVP
programs, and Federal Phase II RVP limits to determine RVP values for the months from May through
September. The procedures described here apply only to the years prior to 2002. For the 2002 modeling, the
RVP inputs for the criteria pollutant modeling were the same as those used in the HAP modeling. Thus, the
development of the RVP values used in 2002 is discussed in the sections of this document detailing the
development of the HAP fuel parameters.
i. How Does EPA Estimate RVP Values for the Months Outside of the Ozone Season?
The procedure for assigning RVP values by state to the months outside of the ozone season are based on
historic RVP survey data provided by OTAQ. This historic data includes the average January and July RVP
values weighted by the market share of each type of gasoline (regular unleaded, intermediate unleaded,
premium unleaded, etc.) from each of the cities included in the Alliance of Automobile Manufacturers (AAM)
fuel surveys.10 The OTAQ also provided a listing that matches each nonattainment area and many Metropolitan
Statistical Areas (MSAs) throughout the United States with the corresponding AAM survey city with which the
RVP should be used to represent that nonattainment areas. Using these data, EPA assigns January and July
RVP values to each State. These assignments were based on pipeline distribution maps and are shown in
Table 6. EPA then assigns the corresponding January and July weighted RVP values to each of the
nonattainment areas. EPA averages the January or July RVP values for a given year for all nonattainment areas
9
-------
and listed MS As within a State to estimate a single statewide January or July RVP value. For those States that
had no nonattainment areas or MSAs included in the OTAQ cross reference listing, OTAQ assigned survey
cities to these States based on a combination of location and pipeline maps. These assignments are as follows:
State Survey City
Idaho
Billings, MT and Seattle, WA
Iowa
Minneapolis, MN
Nebraska
Kansas City, MO and Minneapolis, MN
North Dakota
Minneapolis, MN
South Dakota
Minneapolis, MN
Wyoming
Billings, MT and Denver, CO
For States with two or more survey cities assigned to its nonattainment areas and MSAs, EPA averaged the
RVP values assigned to each of the nonattainment areas or MSAs within that State. Hawaii was not matched
with any survey city; instead, it was assigned winter and summer RVP values based on guidance from OTAQ.
Based on this guidance, Hawaii received a winter RVP value of 10.0 psi and a summer RVP value of 9.5 psi.
The next step in the process of allocating RVP values is to estimate statewide RVP values for the remaining
months outside of the ozone season based on the survey city January and July RVP values. The American
Society for Testing and Materials (ASTM) schedule of seasonal and geographical volatility classes provides the
basis for the RVP allocation by month.11 This schedule assigns one or two volatility classes to each State for
each month of the year. Volatility classes are designated by a letter (A through E), with A being the least
volatile. The ASTM schedule divides several States into two or more regions, with each region having its own
set of volatility class guidelines. The MOBILE4 User's Guide12 provides guidance on which ASTM class to
assign to each State for each month when more than one region is included for a State, or when two ASTM
classes are listed for a given State in a given month. EPA followed this guidance to select a single ASTM class
for each State and month. The MOBILE4 User's Guide also lists RVP limits that correspond to each ASTM
class. These RVP limits are as follows:
ASTM class A = 9.0 psi
ASTM class B = 10.0 psi
ASTM class C = 11.5 psi
ASTM class D = 13.5 psi
ASTM class E = 15.0 psi
EPA assigns the January ASTM class designation to the calculated January RVP value for each State and the
July ASTM class designation to the calculated July RVP value for each State. Those months with the same
ASTM class designation as either January or July are assigned the January or July RVP value for that State.
The RVP values for months with intermediate ASTM class designations are calculated by interpolation using the
January and July RVP values and the ASTM class RVP limits. This interpolation uses Equation 5.
IM = [(IA - SA) x (WM - SM) (WA - &4)] + SM (Eq. 5)
where: IM = Intermediate month's (not January or July) RVP value
10
-------
WM = Winter (January) RVP value
SM = Summer (My) RVP value
IA = Intermediate month's (not-January or July) ASTM RVP limit
WA = Winter (January) ASTM RVP limit
S A = Summer (July) ASTM RVP limit
Many of the AAM survey cities sold reformulated or low RVP gasoline, rather than conventional gasoline
during 1999 and the later years modeled. The July RVP of reformulated gasoline is almost always lower than
the July RVP of conventional gasoline would be for that same geographic area. As a result, using an RFG
survey city to represent RVP values for areas receiving regular gasoline results in inappropriately low RVP
values for these areas. To correct this situation, OTAQ provided a substitute survey city to use in place of each
of the AAM survey cities receiving reformulated gasoline or low RVP gasoline in 1999 or later years to use
when calculating the July RVP values of areas without reformulated gasoline or low RVP fuel.13 This substitute
survey city assignment is shown in Table 7. Again, this procedure and these survey city July RVP values were
only used for calculating the monthly RVP values for the months outside of the ozone season using equation 5.
ii. What RVP Values Does EPA Use for the Ozone Season Months?
The procedure discussed above was NOT applied to the ozone season months (May through September)
because most of the cities in the RVP surveys were implementing either a low RVP program or reformulated
gasoline. Therefore, the RVP values from these cities would not be applicable to a majority of the remaining
areas in the United States. Instead, RVP data for the ozone season months was based on data from OTAQ
showing RVP throughout the ozone season by State, or county if a particular county's RVP varied from the
remainder of the State's RVP. This information can be found at:
http://www.epa.gov/oms/regs/fuels/rfg/sumrvp4.pdf. The July RVP value from this table was applied in all five
of the ozone season months for a given county. These data were then superceded by actual July RVP survey
data for areas included in the AAM fuel survey.10 RVP values for the remaining months were calculated at the
State level, based on the AAM 1999 January RVP survey data. To estimate RVP values for the remaining
months, EPA first assigned a weighted January RVP value for each year to each State as discussed above for
the earlier years. However, the July RVP value used in this procedure for estimating the values for the non-
ozone season months, was the area's Phase II RVP limit (with 8.7 psi used to represent the 9.0 psi limit in most
areas to account for the typical margin of safety used by most refiners) rather than the July values from the RVP
survey data.
c. What Diesel Fuel Sulfur Inputs Does EPA Model?
The sulfur content of diesel fuel changed in 1993, due to the Clean Air Act Amendments of 1990. For the
years modeled prior to 1993 ( 1978, 1987, and 1990), in the MOBILE6 input files for all states the "DIESEL
SULFUR" command was used and set to a value of 2000 ppm in all scenarios. For the years modeled after
1993, the "DIESEL SULFUR" command was used and set to a value of 500 ppm in all scenarios.
11
-------
(L What Header and Run Information Does EPA Include in the M0BILE6 Input Files?
In order to get the necessary emission factor breakdown between exhaust VOC and evaporative VOC and to
obtain both PM-10 emission factors and PM-2.5 emission factors, two MOBILE6 input files were created for
every county modeled. The first input file was used to obtain the exhaust VOC emission factors and PM-10
(as well as NOx, CO, S02, and NH3 emission factors). The commands used for one of these sample
MOBILE6 input files is shown below:
MOBILE6 INPUT FILE :
> HEADER 01 0011999 - EXHAUST - PM 10.0
REPORT FILE
DATAEASE OUTPUT
WITH FIELDNAMES
NO DESC OUTPUT
DAILY OUTPUT
DATAEASE EMISSIONS
PARTICULATES
AGGREGATED OUTPUT
EMISSIONS TABLE
Trends99/Output99/N0100110.TXT REPLACE
2211 1111
S04 OCARBON ECARBON GASPM LEAD S02 NH3 BRAKE TIRE
Trends99/TB1 99/N0100110.TBI REPLACE
RUN DATA
>
EXPRESS HC AS VOC
NO REFUELING
EXPAND BUS EFS
EXPAND HDDV EFS
EXPAND HDGV EFS
EXPAND LDT EFS
The corresponding commands used to obtain the evaporative VOC emission factors and PM-2.5 emission
factors are shown below:
MOBILE6 INPUT FILE :
> HEADER 01 0011999 - EVAPORATIVE - PM 2.50
REPORT FILE
DATABASE OUTPUT
WITH FIELDNAMES
NO DESC OUTPUT
DAILY OUTPUT
DATABASE EMISSIONS
POLLUTANTS
PARTICULATES
AGGREGATED OUTPUT
EMISSIONS TABLE
Trends99/Output99/N0100125.TXT REPLACE
1122 2222
HC
ECARBON S04 OCARBON GASPM LEAD BRAKE TIRE
Trends99/TB1 99/N0100125.TBI REPLACE
RUN DATA :
>
EXPRESS HC AS VOC
NO REFUELING :
EXPAND BUS EFS :
12
-------
EXPAND HDDV EFS :
EXPAND HDGV EFS :
EXPAND LDT EFS :
e. What Speed Inputs and Facility Types Does EPA Use?
Speed is another input used in the M0BILE6 emission factor calculations. For MOBILE6, speed and facility
(roadway) type can be input with the "AVERAGE SPEED" command. In this analysis, EPA continued to
represent the speeds that had been modeled nationally in prior years of the Trends analysis for consistency in
comparison. These average speeds by roadway type and vehicle type are shown in Table 8. Within
MOBILE6, emission factor adjustments by speed also depend on the MOBILE6 roadway type being
modeled. There are four MOBILE6 roadway types: freeways, arterials, locals, and freeway ramps. The
twelve roadway types shown in Table 8 were assigned to one of these MOBILE6 roadway types based on
EPA guidance. The MOBILE6 freeway roadway type was assigned to rural interstates, urban interstates, and
urban other freeways and expressways. The MOBILE6 arterial roadway type was assigned to rural other
principal arterials, rural minor arterials, rural major collectors, rural minor collectors, rural locals, urban other
principal arterials, urban minor arterials, and urban collectors. The MOBILE6 local roadway type was
assigned to urban locals.
To model the urban local roadways, the "VMT BY FACILITY" distribution command was used along with an
external data file with 100 percent of travel modeled on local roadways. The "AVERAGE SPEED" command
was not used with this roadway type because emission factors modeled on the MOBILE6 local roadway type
do not vary by speed.
It should be noted that after the MOBILE6 runs were completed, EPA discovered a bug in the way that the
"AVERAGE SPEED" command works when combined with the freeway facility type. As a result, the
emissions computed here for rural interstates, urban interstates, and urban other freeways and expressways (the
three HPMS roadway types modeled with the MOBILE6 freeway facility type) will be incorrect. Preliminary
analyses performed by EPA indicated that the magnitude of this error on overall national emissions is on the
order of a half of a percent for VOC and even smaller for CO and NOx.
f. What Altitude Inputs Does EPA Use?
The States of Colorado, Nevada, New Mexico, and Utah were all modeled as high altitude areas; all other
States are treated as low altitude areas in the MOBILE6 modeling.
g. What Registration Distributions by Vehicle Age Does EPA Use?
The MOBILE6 model runs all included the default MOBILE6 registration distribution. Additionally, the
"EVALUATION MONTH" input was set to "1" to for scenarios modeling months from January through June
and to "7" to for scenarios modeling months from July through December. The setting of "7" prompts
MOBILE6 to adjust the default registration distribution to reflect an additional half-year of fleet turnover.
13
-------
h. How Does EPA Model On-road Control Programs?
The M0BILE6 model allows for the modeling of several area-specific on-road control programs, such as
reformulated gasoline (RFG), inspection and maintenance (I/M) programs, oxygenated fuels, and the national
low emission vehicle program (NLEV). Control measures that are applied nationally, such as the Tier 1
emission standards, are modeled as defaults with no user input needed. This section describes only those
control programs that are area-specific and require additional inputs to MOBILE6,
i. How Does EPA Account for the Reformulated Gasoline Program?
Phase I of the federal RFG program began on January 1, 1995. Phase I RFG provides year-round toxic
emission reductions and additional VOC emission reductions during the ozone season (May through
September). The Clean Air Act Amendments of 1990 (CAAA) mandate that RFG be used in the nine most
severe ozone nonattainment areas and allow additional nonattainment areas to opt in to the program. OTAQ
provided a list of areas that participated in this program. This list can be found at: http://www.epa.gov/
oms/regs/fuels/rfg/rfgarea.pdf. Table 9 shows the counties modeled with Federal RFG in 1999.
RFG was modeled in the appropriate MOBILE6 input files by including the "FUEL PROGRAM" command
with the value set to "2" to indicate reformulated gasoline, and either an "N" to model northern RFG parameters
or "S" to model southern RFG parameters, as shown in Table 9. In addition to the "FUEL PROGRAM'
command, the "SEASON" command was included in each scenario for the input files modeling RFG. Without
this command, MOBILE6 would apply the winter RFG rules to the scenarios modeled with the
"EVALUATION MONTH" command set to "1" (January) and summer RFG rules to the scenarios modeled
with the "EVALUATION MONTH" command set to "2" (July). In actuality, the summer RFG rules should be
applied in the months from May through September, so the "SEASON" command is used and set to "1" during
these months and to "2" during all remaining months.
ii. How Does EPA Model Inspection and Maintenance (I/M) Programs and Anti-Tamper ing
Programs (ATPs)?
Modeling an Inspection and Maintenance (I/M) program and an anti-tampering program (ATP) in MOBILE
require the most complex set of inputs of any highway vehicle control program. The sources used for
developing the necessary I/M and ATP program inputs include a summary prepared by OTAQ showing the
basic characteristics of I/M and ATP programs planned by the States15 and inputs prepared for previous
Trends inventories.
For States that had an I/M or ATP program in place in one or more counties in the year being modeled, EPA
created at least one additional MOBILE input file to model the characteristics of the I/M program in that State.
All other inputs (such as temperature, RVP, speeds, etc.) are identical to the input file without I/M modeled for
the State in the year being analyzed. The determination of whether or not a county has an I/M program in place
in a given year is based on a series of I/M program summaries released by OTAQ. I/M program
characteristics are also included in the I/M program summaries. These program characteristics vary by State
and in some cases by nonattainment area or county within a particular State. In general, the MOBILE6 I/M
program inputs were developed for 1978, 1987, 1990, and 1996 by converting the MOBILE5-based I/M
14
-------
program inputs developed previously for EPA's Trends report to MOBILE6-based inputs using ERG's
ROUTE5614 program. The 1999, 2000, 2001 and 2002 I/M inputs were developed by directly coding the
information from OTAQ's I/M program summary into MOBILE6 input format. However, due to the many
varieties of State I/M programs from State to State, as well as within a State in many cases, EPA does not
expect that these I/M program inputs exactly model the State programs.
iii. How Does EPA Account for Oxygenated Fuels in Developing Criteria Emissions Estimates?
The oxygenated fuel requirements of the 1990 CAAA began to take effect in late 1992. Therefore,
oxygenated fuel was modeled in the areas indicated by OTAQ, using the oxygenated fuel flag and the
oxygenated fuel market share and oxygen content inputs in MOBILE6 for the 1996, 1999, 2000, and 2001
NEI. OTAQ provided a listing of areas participating in the oxygenated fuel program,16 the months that each
area used oxygenated fuel, and market share data indicating the percentage of ether blends versus alcohol
blends in each oxygenated fuel area. EPA assumed the average oxygen content of ether blend fuels for all
areas, to be 2.7 percent while alcohol blend fuels were assumed to have an oxygen content of 3.5 percent.
Table 10 lists the areas modeled with oxygenated fuels and the corresponding inputs used for these areas. As
is the case with the RVP inputs, the 2002 criteria pollutant MOBILE6 input files use the oxygenated fuel inputs
developed for use in the HAP runs. Therefore, the 2002 oxygenated fuel inputs are documented in the HAPs
section of this report.
iv. How Does EPA Account for the National Low Emission Vehicle (NLEV) Program?
On March 2, 1998, EPA's voluntary National Low Emission Vehicle (NLEV) program came into effect. This
program was modeled as starting in the Northeast Ozone Transport Commission (OTC) States in 1999. States
in the OTC that had already adopted a LEV program on their own were modeled with the characteristics of
their own program. These States included Massachusetts, New York, Vermont, Maine, and Connecticut.
The implementation schedule of the NLEV program is shown below.
Model Year
Federal Tier I
Standards
Transitional LEV
Standards
LEV Standards
1999
30%
40%
30%
2000
40%
60%
2001 and later
100%
These LEV implementation schedules differ from the MOBILE6 default LEV implementation schedule. For the
model to access the implementation schedule of these other LEV programs, the command line "94+ LDG
IMP" was added to the input files representing areas with a LEV program in place in the year being modeled.
The appropriate external LEV implementation file was also referenced in the command line.
3. How Were Criteria Emissions Data Supplied by the States Incorporated?
California provided EPA with emissions for all criteria pollutants except NH; for the years 1999, 2000, and
2001. Annual on-road emissions of VOC, NOx, CO, S02, PM-10, and PM-2.5 were reported at the county
15
-------
level, for each of the 8 MOBILE5 vehicle types. These emissions were calculated using emission factors
derived from California's EMFAC model. VOC emissions were broken down into evaporative and exhaust
emissions and the PM emissions were separated by exhaust, tire wear, and brake wear. Since the 2000 NEI
was not recalculated during 2003, the 2000 NEI does not include the emissions supplied by California, and are
instead MOBILE6-based emissions. EPA allocated the California emissions to the SCC level of detail (12
vehicle types and 12 roadway types) using the VMT databases developed as discussed above for California.
Emissions were broken down the the desired level of detail in proportion to the ratio of VMT at the SCC level
to the VMT at the 8 vehicle type level of detail for each county. In this process, the additional detail regarding
the breakdown of the VOC emissions by evaporative and exhaust component and of the PM emissions by
exhaust, brake wear, and tire wear component was inadvertently lost. Therefore, the VOC emissions reported
for CA in the NEI include the evaporative plus exhaust components added together and the PM emissions
include the exhaust, brake wear, and tire wear components added together. EPA calculated NH3 emissions for
California by multiplying the VMT provided by California by the national average NH3 emission factors at the 8
vehicle type level, based on the NH3 emissions from the remaining States. California did not provide ozone
season day emissions. Therefore, the 1999 and 2001 NEI include OSD emissions for California calculated
with MOBILE6 emission factors as discussed above.
Colorado provided emission values for PM-10 exhaust, in addition to VMT. Thus, the PM-10 exhaust
emissions provided by Colorado replaced MOBILE6-based PM-10 exhaust emission factors. All other
criteria pollutant emissions, and PM-10 brake wear and tire wear emissions were calculated based on the
VMT provided by Colorado. The PM-10 exhaust emission values provided by Colorado were provided at
the same level of detail as the provided VMT. Thus, these emissions were allocated to the 28 vehicle type level
using the same factors that were used to expand the VMT from 8 vehicle types to 28 vehicle types. These
annual PM-10 emissions were then apportioned to the monthly emission level by multiplying the annual PM-10
exhaust emissions by national monthly temporal factors discussed above.
Oregon provided annual 1999 emissions for all seven criteria pollutants. These emission data were at the 12
vehicle type level of detail (matching the SCC vehicle type level of detail), but without roadway type or
emission component (i.e., exhaust, evaporative, brake wear, tire wear) identified. EPA incorporated these
emissions into the 1999 NEI using the same procedures as discussed above for incorporating California
emissions. The emissions were split out to the 12 roadway types. No emission component identifying
information was added. Emissions for a 1999 ozone season day were added using MOBILE6-based
emissions calculated as in the discussion above.
4. How Were Criteria Emissions for Puerto Rico and the US Virgin Islands Calculated?
EPA included criteria pollutant onroad emissions for Puerto Rico and the US Virgin Islands (USVI) in the 1999
and 2002 NEI. The procedure to develop calculate onroad emissions for Puerto Rico and the U.S. Virgin
Islands (USVI) required a different approach than that for the US, as not all of the data used to develop
emissions for the 50 states was available for Puerto Rico and the USVI. The procedures for calculating VMT
for each of these two areas also varied, based on the data available for each. The MOBILE6 emission factor
calculations for these two areas were similar. The development of VMT and MOBILE6 emission factors are
discussed below.
16
-------
a. How Does EPA Calculate VMTfor Puerto Rico and the USVI?
Table VM-2 of Highway Statistics 19995 includes VMT data by roadway type for Puerto Rico, just as it does
for the 50 states. These VMT data were then distributed to the county (municipality) level based on the ratio of
each county's population to the total Puerto Rico population, as provided by the Bureau of the Census.17 A
breakdown of the number of registered vehicles by vehicle type in Puerto Rico was obtained from Puerto
Rico's Highway and Transportation Authority.18 This included data for 21 vehicle types. These vehicle types
were aggregated to the HPMS vehicle classes, as shown in Table 11. The data at this level are in a format
comparable to the U.S. data by vehicle type that are used from Highway Statistics 1999.5 These numbers
represent registered vehicles rather than VMT, but these data were used to perform the allocations by vehicle
type since VMT data by vehicle type were not available for Puerto Rico. From this point, the conversion to the
28 MOBILE6 vehicle types was performed in the same manner as was done for the U.S., thus producing 1999
VMT by county, roadway type, and vehicle type. To estimate 2002 VMT for Puerto Rico, growth rates by
roadway type were developed based on 2002 VMT data by roadway type for Puerto Rico from Table VM-2
of Highway Statistics 200219. Each record from the from the 1999 VMT database for Puerto Rico at the
county, 28 vehicle type, 12 roadway type level of detail was then multiplied by the ratio of the 2002 Puerto
Rico VMT for the specified roadway type to the 1999 Puerto Rico VMT for the specified roadway type.
The procedure to develop VMT for the USVI was more difficult and more uncertain than that for Puerto Rico,
as vehicle registration and total VMT data were not available. Due to similarities in island geography and
roadway network, Kauai County, Hawaii, was selected as a surrogate to develop an estimate of total VMT
and VMT fractions by roadway type in the USVI. The total VMT in Kauai County was divided by the total
number of vehicles on this island. This ratio of annual VMT accumulated per vehicle was then multiplied by the
total number of vehicles in the USVI.20 This gives a rough estimate of the total VMT in the USVI. To estimate
the allocation of this total USVI VMT to the roadway type level, the distribution of roadway mileage from
Kauai County was applied to the total USVI roadway mileage21 to estimate roadway mileage on each island, as
shown in Table 12. Since data showing the mix of vehicles or VMT by vehicle type was unavailable for the
USVI, the VMT mix by vehicle type estimated for Puerto Rico was applied to the USVI. This total VMT by
roadway type was then distributed to the three islands within the USVI based on the population of each island.
The USVI has a large tourist population year round.22'23 It was necessary to include the impact of this group on
total resident populations in order to more accurately allocate VMT to each island. The percentage of the
population of the Virgin Islands on each island is shown in Table 13. This VMT distribution was then combined
with the roadway mileage data to determine the fraction of total USVI VMT to apply by roadway type and
island, as shown in Table 14. To estimate 2002 VMT for the USVI, the overall growth rate in VMT from
Puerto Rico was applied to the 1999 USVI VMT database. Based on the Puerto Rico VMT totals in Table
VM-2 of the 1999 and 2002 Highway Statistics5-'y, VMT in Puerto Rico grew by a factor of 1.0657 from
1999 to 2002. Therefore, all of the 1999 USVI VMT values were multiplied by 1.0657 to estimate 2002
VMT in the USVI.
b. What Inputs were Used to Produce MOBILE6 Input Files for Puerto Rico and the USVI?
To determine whether the default MOBILE6 registration distribution would be appropriate to apply in Puerto
Rico and the USVI, data available from Puerto Rico listing new vehicle sales and the total number of vehicle
registrations, both by model year, was examined and compared to national trends in the U.S. Table 15 lists the
17
-------
25-year trend of vehicle sales and registrations in Puerto Rico.19 Based on comparisons made between this list
and the national trend, and without more specific data, it was determined that the default MOBILE6 registration
distribution would sufficiently represent Puerto Rico and the Virgin Islands. RVP data for both Puerto Rico and
the USVI were modeled based on the RVP data modeled for Collier County, Florida. This county was
selected by OTAQ as an appropriate surrogate for fuel properties for Puerto Rico and the USVI. The
temperature data modeled for Puerto Rico and the USVI, representing 20-year average temperatures, are
shown in Table 16.24 These temperatures were applied in both 1999 and 2002. The speed data developed for
the U.S. was also applied in Puerto Rico and the USVI. No onroad control programs were modeled in Puerto
Rico or the USVI.
5. How Does EPA Calculate Emissions for Hazardous Air Pollutants for the On-road Vehicle
Category?
EPA calculated annual emissions from on-road vehicles for a total of 33 hazardous air pollutants (HAPs) for the
years 1990, 1996, 1999, and 2002. The emissions for 1990, 1996 and 1999 were calculated using seasonal
emission factors generated by using EPA's MOBILE63 model and the same VMT database described above.
HAP emission factors for 2002 were generated at the monthly level using inputs consistent with the criteria
pollutant MOBILE6 inputs. While criteria pollutants and HAPs can be modeled concurrently in MOBILE6,
HAP estimates for the 1990, 1996, and 1999 NEI were calculated subsequent to the model runs done for
Version 2.0 for criteria pollutants. Therefore, in addition to the seasonal versus monthly differences between
the criteria pollutant and HAP MOBILEy input files, the HAP inputs for these years also include some
differences in winter and summer fuel parameters from what was modeled for the criteria pollutants, due to
revised methods and additional fuel parameters needed for modeling HAPs in MOBILE6.
Emissions were calculated for the HAPs listed below.
| Pollulanl Name |
1,3-Butadiene
2.2.4-Tri methyl pen tane
2,3,7,8-TCDD TEQ
Acenaphthene
Acenaphthylene
Acetaldehyde
Acrolein
Anthracene
Benz[a] Anthracene
Benzene
Benzo[a]Pyrene
Benzo[b]FluorantheneBenzo[e]Pyrene
Benzo[g,h,i,]Perylene
Benzo[k]Fluoranthene
Chromium (VI)
Chromium 111
Chrysene
18
-------
Dibenzo[a,h] Anthracene
Ethyl Benzene
Fluoranthene
Fluorene
Formaldehyde
n-Hexane
lndeno[ 1,2,3-c,d]Pyrene
Manganese & Compounds
Methyl Tert-Butyl Ether
Naphthalene
Nickel & Compounds
Phenanthrene
Propionaldehyde
Pyrene
Styrene
Toluene
Xylenes (Mixture of o, m, and p Isomers)
Within the MOBILE6 model, six HAPs (benzene, formaldehyde, acetaldehyde, 1,3 butadiene, acrolein, and
methyl tertiary butyl ether [MTBE]) can be calculated directly by including detailed fuel parameters within the
MOBILE6 scenario descriptions. These fuel parameters are: sulfur content, olefins content, aromatics content,
benzene content, E200 value, E300 value, oxygenate content by type, and oxygenate sales fraction by type.
Since these fuel parameters are area-specific, EPA developed county-level inputs for each of these parameters
for summer and winter gasoline. Attachment A describes the development of these parameters for 1990,
1996, and 1999. Fuel parameters for 2002 were developed in the same manner as described in Attachment A
for the other years, but using survey data from 2000. EPA used 2000 fuel parameters in the 2002 modeling as
there were no significant changes expected in fuel properties from 2000 to 2002. The fuel parameter data for
each year are posted at ftp://ftp.epa.gov/EmisInventory/finalnei99ver3/haps/datafiles/onroad/auxiliary/.
MOBILE6 also has a command (ADDITIONAL HAPS) which allows the user to enter emission factors or air
toxic ratios for additional air toxic pollutants. Emission factors for an additional 27 HAPs were calculated by
MOBILE6 through the use of external data files specifying emission factors for these pollutants in one of three
ways: as fractions of VOC, fractions of PM , or by supplying the basic emission factors (primarily used for
metals and metal compounds). The ratios must be expressed as milligrams of HAP per gram of VOC or PM.
Attachment B describes the development of these speciation factors and emission factors and also lists the
HAPs for which emission factors were calculated in this manner. This attachment indicates that the emission
factors were developed for 1999. However, these same factors were applied in 1990, 1996, 1999, and 2002.
The final HAP for which on-road emissions were calculated was dioxin. The dioxin emissions were calculated
without the use of MOBILE6, by multiplying a dioxin emission factor by the corresponding VMT. One factor
was used nationally for gasoline-powered vehicles and another was used for diesel-powered vehicles. The
dioxin emission factors are discussed in more detail in Attachment C. Dioxin is not included in the 2002 onroad
HAPs emission inventory at this time. When EPA calculates onroad dioxin later in 2004 using the National
Mobile Inventory Model (NMDVI) model, congener-specific dioxin emissions will be added to the 2002 onroad
HAPs emission inventory.
19
-------
It should be noted that EPA has temporarily suspended the calculation of mercury and arsenic emissions from
onroad vehicles. Although Attachment B includes an estimate of mercury and arsenic emission factors from
onroad vehicles, EPA has determined that these data are not currently adequate to develop credible emission
estimates. Thus, the current NEI onroad HAP emission inventories exclude mercury and arsenic emissions.
EPA is currently evaluating alternative approaches to developing more credible emission factors for mercury
and arsenic from onroad vehicles. Attachment D discusses EPA's plans for the development of these emission
factors.
a. How Were the 1990,1996, and 1999 MOBILE6 HAP Input Files Developed?
Although the 1990, 1996, and 1999 fuel parameter data were prepared for only two seasons (summer and
winter), four seasonal scenarios were developed. The months corresponding to each season were selected to
best coincide with seasonal fuel requirements. The summer season included the months from May through
September. These months correspond with the summer reformulated gasoline season and the months of the
Phase IIRVP requirements. The fall season included only October. The winter season included the months
from November through February. These are the months that most frequently correspond with the winter
oxygenated fuel season. Finally, the spring season included the months of March and April. Summer fuel
parameters were applied in the fall scenarios and winter fuel parameters were applied in the spring scenarios.
The fuel parameters used are representative of fuel conditions in either January (winter) or July (summer). No
averaging was applied to fuel parameters, such as RVP, because the independent averaging of the various fuel
parameters could lead to inappropriate fuel descriptions.
The maximum and minimum temperature inputs for each of the seasonal scenarios were developed as the
average maximum and minimum daily temperatures from all of the months included in a given season for the
state being modeled. The HAP emission factor calculations were based on the same temperature inputs used
for the criteria pollutant emission factor calculations, but with the necessary seasonal averaging. For example,
the maximum temperature for the summer scenarios was calculated as the average of the May through
September maximum temperatures for a given State and the minimum temperature for the summer scenarios
was calculated as the average of the May through September minimum temperatures for a given State.
For each of the speed and road type combinations modeled in the criteria pollutant MOBILE6 runs, four
seasonal MOBILE6 scenarios were developed for estimating HAP emission factors. Other than the fuel
parameters and seasonal temperatures, inputs to the MOBILE6 files were the same as those in the MOBILE6
criteria pollutant emission factor runs for control programs (such as I/M and NLEV). However, the command
indicating the presence of a reformulated gasoline program cannot be used in combination with the MOBILE6
fuel inputs. The appropriate combination of RVP, fuel sulfur content, and oxygen content in the MOBILE6
input files give almost the same emission factor results as the use of the "FUEL PROGRAM' command in
MOBILE6 to indicate the use of reformulated gasoline. It should be pointed out that RVP is also a parameter
used in estimating criteria pollutant emission factors. The RVP values used in the criteria pollutant runs were
obtained solely from surveys done by the Alliance of Automobile Manufacturers (AAM), as described earlier in
the document. However, these survey data were not adequate to characterize across the U.S. the full range of
fuel parameters required by MOBILE6. Thus AAM data were supplemented by data from other surveys, as
described in Attachment A, the model HAPs. As a result, RVP levels specified in HAP runs are not
necessarily identical to RVP levels used in criteria pollutant runs.
20
-------
It should also be noted that the MOBILE6 runs were completed after the time that EPA discovered the bug in
the way that the "AVERAGE SPEED" command works when combined with the freeway facility type. Thus,
the MOBILE6 emission factors computed here for rural interstates, urban interstates, and urban other freeways
and expressways (the three HPMS roadway types modeled with the MOBILE6 freeway facility type) followed
OTAQ's guidance for correcting for this bug. This was done by specifying a mix of 92 percent of VMT on
freeways, 0 percent of VMT on arterials, 0 percent of VMT on local roads, and 8 percent of VMT on freeway
ramps at the end of each command line whenever the "AVERAGE SPEED" command was used with the
"Freeway" facility type.
The unique set of MOBILE6 input files needed for each State was determined using the same methodology as
the MOBILE6 runs for the criteria pollutants. This included determining the unique combinations of control
programs for the criteria pollutants. In addition to these inputs, for the MOBILE6 HAPs runs, the winter and
summer fuel parameters were also included in determining the minimum number of input files needed for each
State.
In some counties, particularly in the Mdwest, EPA determined that both alcohol blend fuels and non-alcohol
fuel blends were present in significant amounts. Thus, for counties where this occurred, EPA determined the
market share of non-alcohol blend fuels and the market share of alcohol blend fuels. Then, two sets of
MOBILE6 input files were developed for each of these counties with one set of files using the non-alcohol
(MTBE-based) fuel profiles and the other using the ethanol-based profiles, as appropriate. VMT values for
these counties were multiplied by the non-alcohol blend market share. This portion of the VMT was then
mapped to the emission factors calculated with the MBTE-based profiles. VMT values for these counties were
also multiplied by the alcohol-blend market shares and then mapped to the emission factors calculated with the
ethanol-based profiles. Thus, the emission factors were weighted according to the market share of these two
fuel types. The resulting emissions from both fuel types for a given county were then added together to
determine the county's total on-road HAP emissions.
As described in Attachment B, five sets of external emission factor speciation files or profiles were developed
for calculating 27 of the HAPs emission factors with MOBILE6. These included a profile for baseline fuel, a
profile for reformulated gasoline with MTBE (with 2.0 percent MTBE by weight), a profile for winter
oxygenated fuel with MTBE (with 2.7 percent MTBE by weight), a profile for reformulated gasoline with
ethanol, and a profile for winter oxygenated fuel with ethanol (the two ethanol profiles are identical, both for a
gasoline with 3.5% ethanol by weight). Files with these data are posted at
ftp://ftp.epa.gov/EmisInventory/finalnei99ver3/haps/datafiles/onroad/auxiliary/. The appropriate profile to
be used with each M0BILE6 scenario was determined in the following manner.
The winter oxygenated fuel MTBE profile was applied to areas with an ether-based oxygenate with an oxygen
content of 12 percent or greater by volume and with a market share of the ether-based oxygenated fuel greater
than the market share of alcohol-based oxygenated fuel. The reformulated gasoline MBTE profile was applied
to areas with an ether-based oxygenate with an oxygen content of at least 5 percent by volume and less than 12
percent by volume, along with a market share of the ether-based oxygenated fuel greater than the market share
of alcohol-based oxygenated fuel. The ethanol oxygenated fuel profile was applied in areas with an alcohol-
based oxygenate with an oxygen content of 5 percent or more by volume and a market share of the alcohol-
based oxygenated fuel greater than the market share of ether-based oxygenated fuel. The baseline profile was
21
-------
applied in all other areas. It should be noted that the profile applied could vary by season for a given area,
depending upon the seasonal fuel parameters for the area.
b. How Were the 2002 MOBILE6 HAP Input Files Developed?
The 2002 MOBILE6 HAP input files were developed differently than the HAP input files for the earlier years.
The primary difference between the 2002 MOBILE6 HAP files and those for the earlier years is the change
from the seasonal scenarios used in the 1990, 1996, and 1999 HAP MOBILE6 files to monthly scenarios used
in the 2002 MOBILE6 HAP input files. Thus, the 2002 HAP MOBILE6 input files are fully consistent with the
2002 MOBILE6 criteria pollutant input files. The only reason for developing separate MOBILE6 files for the
criteria pollutants and HAPs was for ease in post-processing the emission factors and calculating emissions. In
order to support this change from seasonal to monthly scenarios, the fuel data needed to be developed at the
monthly level. The January and July fuel parameter data were developed by EPA from fuel survey data, in the
same manner as the fuel data for the earlier years. The actual year of the survey data used for 2002 was 2000.
EPA does not expect any significant differences between actual 2000 fuel parameters and 2002 fuel
parameters.
After the January and July fuel parameters were allocated by county in the manner discussed in Attachment A,
the fuel parameters were then distributed by month using the interpolation method developed by OTAQ for use
in preparing a national fuel parameter database to populate its National Mobile Inventory Model (NMDVI)25.
This procedure is similar to that discussed above, and shown in Equation 5. First, a monthly interpolation factor
is calculated based on the RVP ASTM classes assigned to a county in a given month, in summer (July), and in
winter (January). The monthly interpolation factor is calculated using Equation 6.
MIF = (IA - SA) / (WA - SA) (Eq. 6)
where: MIF = Monthly Interpolation Factor (unitless)
LA = Intermediate month's (not-January or July) ASTM RVP limit
WA = Winter (January) ASTM RVP limit
S A = Summer (July) ASTM RVP limit
As discussed above, the ASTM RVP limits used in this equation are taken from the MOBILE4 User's Guide
list of the RVP limits that correspond to each ASTM class. These RVP limits are as follows:
• ASTM class A = 9.0 psi
• ASTM class B= 10.0 psi
• ASTM class C= 11.5 psi
• ASTM class D= 13.5 psi
• ASTM class E= 15.0 psi
Once the monthly interpolation factor is calculated for each month, all of the necessary fuel parameters are
interpolated using this monthly interpolation factor along with the winter and summer values for that parameter in
that county using Equation 7.
MFP = SFP + MIF * (WFP - SFP) (Eq. 7)
22
-------
where: MFP
Monthly Fuel Parameter (e.g., RVP, sulfur content, etc.)
Summer (July) fuel parameter
Monthly Interpolation Factor (as calculated in Equation 6)
Winter (January) fuel paramter
SFP
MIF
WFP
The process of assigning external emission factor speciation profiles for calculating the HAP emission factors
with MOBILE6 for each county and month was the same as discussed above for the 1990, 1996, and 1999
HAP emission inventories. In the case of 2002, however, this assignment was made on a monthly, rather than
seasonal, basis.
Once the fuel parameters were estimated for all twelve months and for all counties, MOBILE6 input files were
created. Twelve monthly scenarios were created at each of the thirteen speed and roadway type combinations,
as modeled for the criteria pollutants. The same monthly temperatures that were used in the 2002 MOBILE6
criteria pollutant input files were used in the 2002 MOBILE6 HAP input files. The additional fuel parameter
data needed for the HAP input files, as discussed above, were included in these input files at the monthly level
of detail. All other MOBILE6 input parameters, such as I/M program parameters, were identical to the input
parameters modeled for the criteria pollutants.
6. How Were HAP Emissions Data Supplied by the States Incorporated?
California provided its own estimates of HAP emissions for 1999. These emissions replaced the emissions
calculated by EPA for California. California's emissions estimates for arsenic and mercury were removed
before incorporation into the MF database, as EPA had determined that mercury and arsenic emissions should
not currently be included in the NEI until these emission factors can be developed with greater certainty. As
with the criteria pollutants, the HAP emissions provided by California did not include a breakdown by roadway
type. Thus, the HAP emissions provided by California were allocated to the 12 HPMS roadway classes in the
same manner as discussed above for criteria pollutants. It should also be noted that there is a small difference
in the labeling of pollutant codes by California and EPA. For example, California separately reports p-xylenes,
m-xylenes, and o-xylenes, whereas EPA includes all xylenes (mixture of o, m, and p isomers) as one pollutant.
7. How Were Criteria Pollutant Emissions for Interim Years Calculated?
EPA calculated on-road vehicle criteria pollutant emissions for the years from 1979 through 1998, excluding
1987, 1990, and 1996 by interpolation. Emissions for the years from 1979 through 1986 were calculated by
interpolation between the SCC-level emission files for the years 1978 and 1987. Emissions for all criteria
pollutants were estimated at the SCC level using linear interpolation between these years to develop new SCC-
level emission estimates for each of these intervening years. Emissions for 1988 and 1989 were calculated by
linear interpolation between the years 1988 and 1989. Emissions for the years from 1991 through 1995 were
calculated by linear interpolation between the years 1990 and 1996.
The 1997 and 1998 emissions were calculated using a 2-step interpolation procedure between the years 1996
and 1999. The 1999 emissions database used in the interpolation for 1997 and 1998 was Version 2 of the
1999 NEI, calculated in 2002, as these interpolations were made prior to the calculation of Version 3 of the
23
-------
1999 NEI. This 2-step interpolation applied trends in the previous MOBILE5-based onroad emission
calculations from 1996 through 1999 and applying these trends to the linearly interpolated emission calculations.
Emissions for 1970 and 1975 were calculated at the national level only, by vehicle type, through extrapolation.
These emissions were extrapolated by first calculating the overall change in emissions by vehicle type for the
period from 1978 through 1987 and dividing by the number of years in this period to estimate the average
change in emissions per year for each vehicle type, assuming linear growth in emissions. Eight years of
emissions change were then applied to the 1978 emissions to estimate 1970 emissions and three years of
emissions change were applied to the 1978 emissions to estimate 1975 emissions by vehicle type.
8. What Caveats Should be Considered when Comparing Emissions Across Years
It should be noted that criteria pollutant emissions for years other than 2002 that were prepared in the NEI
Input Format (NIF) are reported in tons for annual emissions and pounds for ozone season day emissions, both
with two decimal places included. HAP emissions for years other than 2002 are reported in pounds with two
decimal places included. For the 2002 HAP and criteria pollutant emission files, emissions are reported in
scientific notation with up to 6 decimal places allowed for the criteria pollutant emissions and 12 decimal places
for the HAP emissions. Thus, for the earlier years, emissions would be underreported in cases where the
emissions would have been rounded off to 0 with the limited number of decimal places. Due to the magnitude
of VOC, NOx, and CO emissions, this rounding has almost no effect on these emissions. However, for the
S02, PM, NH3, and many of the HAPs, this rounding would have eliminated emissions from a number of
counties and SCCs.
24
-------
C. REFERENCES
1. "National Air Pollutant Emission Trends, Procedures Document, 1900-1996," EPA-454/R-98-008,
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, NC, May 1998.
2. "User's Guide to MOBILE6.0: Mobile Source Emission Factor Model," EPA420-R-02-001, U.S.
Environmental Protection Agency, Office of Transportation and Air Quality, January 2002.
3. "User's Guide to MOBILE6.1 and MOBILE6.2: Mobile Source Emission Factor Model," EPA420-
R-02-028, U.S. Environmental Protection Agency, Office of Transportation and Air Quality, October
2002.
4. "Draft User's Guide to PART 5: A Program For Calculating Particle Emissions From Motor
Vehicles," EPA-AA-AQAB-94-2, U.S. Environmental Protection Agency, Office of Mobile Sources,
Ann Arbor, MI, July 1994.
5. Highway Statistics 1999. Federal Highway Administration, U.S. Department of Transportation,
Washington, DC, 2000.
6. "1990 Census of Population, Volume I Characteristics of Population, Chapter B Number of
Inhabitants," Bureau of the Census, U.S. Department of Commerce, Washington, DC, July 1992.
7. "The 1985 NAPAP Emissions Inventory: Development of Temporal Allocation Factors," EPA-600/7-
89-0lOd, Air & Energy Engineering Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC, April 1990.
8. National Climatic Center, data files to E.H. Pechan & Associates, Inc., Asheville, NC, 2000.
9. "National Data Book and Guide to Sources, Statistical Abstract of the United States - 1993," U.S.
Department of Commerce, Bureau of the Census, Washington, DC. 1994.
10. "Fuel Volatility Survey 1999," Alliance of Automobile Manufacturers, Washington, DC, 1999.
11. "1988 Annual Book of ASTM Standards," American Society for Testing and Materials, (Section 5:
Petroleum Products, Lubricants, and Fossil Fuels; Volume 05.01: Petroleum Products and Lubricants
(I): D 56 - D 1947), Philadelphia, PA, 1988.
12. "User's Guide to MOBILE4 (Mobile Source Emission Factor Model)," EPA-AA-TEB-89-01, U.S.
Environmental Protection Agency, Office of Mobile Sources, Ann Arbor, MI, February 1989.
13. Table provided by Greg Janssen, Office of Mobile Sources, U.S. Environmental Protection Agency, to
E.H. Pechan & Associates, Inc., May 11, 1996.
25
-------
14. Route56 Translator, Version 0.9.10 (Beta 10), Eastern Research Group Inc., downloaded from
www.erg.com/Route56 , April 17, 2001.
15. "Major Modeling Elements for Operating I/M Programs," EPA420-B-99-008, U.S. Environmental
Protection Agency, Office of Transportation and Air Quality, Ann Arbor, MI, December 1999.
16. "State Winter Oxygenated Fuel Program Requirements for Attainment or Maintenance of CO
NAAQS," U.S. Environmental Protection Agency, Office of Transportation and Air Quality, Ann
Arbor, MI, October 2001, available at http://www.epa.gov/otaq/regs/fuels/oxy-area.pdf.
17. Estimates of the Population of Puerto Rico Municipios, U.S. Census Bureau, 1999.
18. Registro De Vehicilos De Moter Por Munipiois Y Por Categorias, Highway and Transportation
Authority of Puerto Rico, Annual, 1998.
19. Highway Statistics 2002. Federal Highway Administration, U.S. Department of Transportation,
Washington, DC, 2002.
20. Wards Facts and Figures 2000, Total Motor Vehicle Registration by Country -1998,
21. World Fact Book - United States Virgin Islands - Total Roadway, CIA, Annual,1998.
22. Tourist arrivals, USVI Bureau of Economic Research, monthly, 2000.
23. USVI resident population numbers by island, USVI government fact sheet, 1990.
24. 20-year Average Temperature Data, National Weather Service, the International Station
Meteorological Climate Summary, Version 4.0, and the Global Historical Climatology Network,
versions 1 andb2, annual, 1998.
25. "Draft National Mobile Inventory Model (NMDVI) Base and Future Year County Database
Documentation and Quality Assurance Procedures," prepared for U.S. Environmental Protection
Agency, Office of Transportation and Air Quality by Eastern Research Group, Inc., Chantilly, VA,
May 30, 2003.
26
-------
Table 1. Methods Used to Develop Emission Estimates for Onroad Vehicle Sources
Base
Year(s)
Pollutant(s)
Geographic Area
Emission Estimation Method
1970, 1975
1978, 1987,
1990, 1996,
2000
All Criteria
All Criteria
US
US
Linear extrapolation at national vehicle type level based on 1978 and 1987
national data
Calculated at State/county/SCC level by month using MOBILE6, no State data
incorporated
1979-1986
1988-1989
1991-1995
1990, 1996
1997-1998
1999
1999
1999
1999
1999
1999
1999
1999
All Criteria
All Criteria
All Criteria
HAPs
All Criteria
All Criteria
VOC, NOx, CO, S02, PM-
10, PM-2.5
nh3
PM-10 Exhaust
VOC, NOx, CO, S02, PM-
10 brake and tire wear,
PM-2.5, NH3
All Criteria
All Criteria
HAPs
US
US
US
US
US
AL; ME; MA; MS; UT; VA; WV;
Maricopa County, AZ; Hamilton
County, TN
California
California
Colorado
Colorado
Oregon
Rest of US, Puerto Rico, and US
Virgin Islands
California
Linear interpolation at State/count/SCC level based on 1978 and 1987
State/count/SCC level data
Linear interpolation at State/count/SCC level based on 1987 and 1990
State/count/SCC level data
Linear interpolation at State/count/SCC level based on 1990 and 1996
State/count/SCC level data
MOBILE6 emission factors calculated at State/county/SCC level by season;
applied to FHWA-based VMT
2-step linear interpolation at State/count/SCC level based on 1996 and 1999
State/count/SCC level data
Calculated at State/county/SCC level by month using MOBILE6; State-
provided VMT data used
Emissions and VMT provided by California at county/vehicle type level; State-
provided emissions expanded to county/SCC level by EPA
Calculated at State/county/SCC level by month using MOBILE6 emission
factors with State-provided VMT data
PM-10 emissions and VMT provided by State
Calculated at State/county/SCC level by month using MOBILE6; State-
provided VMT data used
Emissions and VMT provided by Oregon at county/vehicle type level; State-
provided emissions expanded to county/SCC level by EPA
Calculated at State/county/SCC level by month using MOBILE6 and FHWA-
based VMT
HAP emissions and VMT provided by California at county/vehicle type level;
emissions allocated to SCC level by EPA
27
-------
Table 1. Methods Used to Develop Emission Estimates for Onroad Vehicle Sources
1999 HAPs
2001 VOC, NOx, CO, S02, PM-
10, PM-2.5
2001 NH3
2001 All Criteria
2001 All Criteria
2002 All Criteria, HAPs
2002 All Criteria, HAPs
Rest of US, Puerto Rico, and US
Virgin Islands
California
California
AL; CO; ME; MA; MS; OR; UT; VA;
WV; Maricopa County, AZ;
Hamilton County, TN
Rest of US
AL; CA; CO; ME; MA; MS; OR; UT;
VA; WV; Maricopa County, AZ;
Hamilton County, TN
Rest of US
MOBILE6 emission factors calculated at State/county/SCC level by season;
applied to FHWA-based VMT
Emissions and VMT provided by California at county/vehicle type level; State-
provided emissions expanded to county/SCC level by EPA
Calculated at State/county/SCC level by month using MOBILE6 emission
factors with State-provided VMT data
State-provided VMT grown to 2001; emissions calculated by EPA using
MOBILE6 emission factors
Calculated at State/county/SCC level by month using MOBILE6 and FHWA-
based VMT
State-provided VMT grown to 2002; emissions calculated by EPA using
MOBILE6 emission factors at State/county/SCC level by month
Calculated at State/county/SCC level by month using MOBILE6 and FHWA-
based VMT grown to 2002
28
-------
Table 2. Allocation of VMT from HPMS Vehicle Categories to MOBILE6 Vehicle Types for
1999
HPMS Vehicle
HPMS 1999 VMT Fractions
MOBILE6
1999 VMT Fractions
Category
Rural
Urban
Vehicle Type
Rural
Urban
Passenger Cars
0.5499
0.6048
LDGV
0.5483
0.6030
LDDV
0.0016
0.0017
Motorcycles
0.0042
0.0038
MC
0.0042
0.0038
Other 2-Axle 4-
0.3307
0.3375
LDGT1
0.0513
0.0524
Tire Vehicles
LDGT2
0.1708
0.1744
LDGT3
0.0520
0.0531
LDGT4
0.0239
0.0244
LDDT12
0.0004
0.0004
LDDT34
0.0010
0.0010
HDGV2B
0.0232
0.0237
HDDV2B
0.0079
0.0081
Single-Unit 2-
0.0339
0.0211
HDGV3
0.0013
0.0008
Axle 6-Tire or
HDGV4
0.0008
0.0005
More Trucks
HDGV5
0.0016
0.0010
HDGV6
0.0034
0.0021
HDGV7
0.0017
0.0010
HDDV3
0.0036
0.0022
HDDV4
0.0028
0.0017
HDDV5
0.0012
0.0007
HDDV6
0.0070
0.0043
HDDV7
0.0107
0.0066
Combination
0.0770
0.0310
HDGV8A
0.0000
0.0000
Trucks
HDGV8B
0.0000
0.0000
HDDV8A
0.0169
0.0068
HDDV8B
0.0602
0.0242
Buses
0.0044
0.0018
HDGB
0.0010
0.0004
HDDBT
0.0014
0.0006
HDDBS
0.0019
0.0008
Total
1.0000
1.0000
Total
1.0000
1.0000
Table 3. VMT Seasonal and Monthly Temporal Allocation
Factors
Roadway
Seasonal VMT Factors
Vehicle Type
Type
Winter
Spring
Summer
Fall
LDV, LDT, MC
Rural
0.2160
0.2390
0.2890
0.2560
LDV, LDT, MC
Urban
0.2340
0.2550
0.2650
0.2450
HDV
All
0.2500
0.2500
0.2500
0.2500
Monthly VMT Factors
Roadway
Vehicle Type
Type
Jan
Feb
Mar
Apr
May Jun
Jul
Aug
Sep
Oct
Nov
Dec
LDV, LDT, MC
Rural
0.0744
0.0672
0.0805
0.0779
0.0805 0.0942
0.0974
0.0974
0.0844
0.0872
0.0844
0.0744
LDV, LDT, MC
Urban
0.0806
0.0728
0.0859
0.0832
0.0859 0.0864
0.0893
0.0893
0.0808
0.0835
0.0808
0.0806
HDV
All
0.0861
0.0778
0.0842
0.0815
0.0842 0.0815
0.0842
0.0842
0.0824
0.0852
0.0824
0.0861
29
-------
Table 4. Fractions for Converting VMT from 8 M0BILE5 Vehicle Types to 28 M0BILE6
Vehicle Types
7-Digit SCC
Default 1999
Fraction of VMT by
Code for
MOBILE6
MOBILE6
MOBILE6
MOBILE6 Vehicle Type
MOBILES Vehicle
MOBILES
Vehicle
Vehicle
VMT
for Each MOBILES
Type
Vehicle Type
Type
Type Code
Fraction
Vehicle Type
LDGV
2201001
LDGV
1
0.5138
1.0000
LDGT1
2201020
LDGT1
2
0.0621
0.2310
LDGT2
3
0.2066
0.7690
LDGT2
2201040
LDGT3
4
0.0630
0.6850
LDGT4
5
0.0290
0.3150
HDGV
2201070
HDGV2B
6
0.0281
0.7891
HDGV3
7
0.0010
0.0282
HDGV4
8
0.0006
0.0182
HDGV5
9
0.0012
0.0347
HDGV6
10
0.0027
0.0746
HDGV7
11
0.0013
0.0368
HDGV8A
12
0.0000
0.0001
HDGV8B
13
0.0000
0.0000
HDGB
25
0.0006
0.0182
MC
2201080
MC
24
0.0064
1.0000
LDDV
2230001
LDDV
14
0.0015
1.0000
LDDT
2230060
LDDT12
15
0.0005
0.2913
LDDT34
28
0.0012
0.7087
HDDV
2230070
HDDV2B
16
0.0096
0.1194
HDDV3
17
0.0028
0.0347
HDDV4
18
0.0022
0.0268
HDDV5
19
0.0009
0.0113
HDDV6
20
0.0055
0.0682
HDDV7
21
0.0084
0.1042
HDDV8A
22
0.0107
0.1334
HDDV8B
23
0.0382
0.4753
HDDBT
26
0.0009
0.0114
HDDBS
27
0.0012
0.0154
30
-------
Table 5. Cities Used for 1999 Temperature Data Modeling
State
City
Alabama
Birmingham
Alaska
Anchorage
Arizona
Phoenix
Arkansas
Little Rock
California
Los Angeles
California
San Francisco
Colorado
Colorado Springs
Connecticut
Hartford
Delaware
Dover
District of Columbia
Washington
Florida
Orlando
Georgia
Atlanta
Hawaii
Honolulu
Idaho
Boise
Illinois
Springfield
Indiana
Indianapolis
Iowa
Des Moines
Kansas
Topeka
Kentucky
Louisville
Louisiana
Baton Rouge
Maine
Portland
Maryland
Baltimore
Massachusetts
Boston
Michigan
Detroit
Minnesota
Minneapolis
Mississippi
Jackson
Missouri
Springfield
Montana
Billings
Nebraska
Lincoln
Nevada
Las Vegas
New Hampshire
Concord
New Jersey
Newark
New Mexico
Albuquerque
New York
New York City
North Carolina
Greensboro
North Dakota
Bismarck
Ohio
Columbus
Oklahoma
Oklahoma City
Oregon
Eugene
Pennsylvania
Middletown
Rhode Island
Providence
South Carolina
Columbia
South Dakota
Pierre
Tennessee
Nashville
Texas
Dallas/Fort Worth
Utah
Salt Lake City
Vermont
Montpelier
Virginia
Richmond
Washington
Seattle
West Virginia
Charleston
Wisconsin
Milwaukee
Wyoming
Casper
31
-------
Table 6. Surrogate City Assignment
Nonattainment Area/MSA
State
Survey City
Albany-Schenectady-Troy, NY MSA
NY
New York City
Albuquerque, NM MSA
NM
Albuquerque
Allentown-Bethlehem, PA-NJ MSA
PA-NJ
Philadelphia
Altoona, PA MSA
PA
Philadelphia
Anchorage, AK MSA
AK
Cleveland
Anderson, SC MSA
SC
Atlanta
Appleton-Oshkosh-Neenah, Wl MSA
Wl
Chicago
Atlanta
GA
Atlanta
Atlantic City, NJ MSA
NJ
Philadelphia
Bakersfield, CA MSA
CA
San Francisco
Baltimore, MD MSA
MD
Washington, DC
Baton Rouge
LA
New Orleans
Beaumont-Port Arthur, TX MSA
TX
Dallas
Bennington Co., VT
VT
Boston
Birmingham, AL MSA
AL
Atlanta
Boston Metropolitan Area
MA
Boston
Boston Metropolitan Area
MA-NH
Boston
Bowling Green, KY
KY
Chicago
Buffalo-Niagara Falls, NY CMSA
NY
New York City
Canton, OH MSA
OH
Cleveland
Charleston, WV MSA
WV
Washington, DC
Charlotte-Gastonia-Rock Hill, NC-SC MSA
NC
Atlanta
Chattanooga, TN-GA MSA
GA-TN
Atlanta
Cherokee Co., SC
SC
Atlanta
Chester Co., SC
SC
Atlanta
Chicago-Gary-Lake County, IL-IN-WI CMSA
IL-IN-WI
Chicago
Chico, CA MSA
CA
San Francisco
Cincinnati-Hamilton, OH-KY-IN CMSA
OH-KY-IN
Cleveland
Cleveland Metropolitan Area
OH
Cleveland
Clinton Co., OH
OH
Cleveland
Colorado Springs, CO MSA
CO
Denver
Columbia, SC MSA
SC
Atlanta
Columbus, OH MSA
OH
Cleveland
Dallas-Ft. Worth, TXCMSA
TX
Dallas
Dayton-Springfield, OH MSA
OH
Cleveland
Denver-Boulder, CO CMSA
CO
Denver
Detroit-Ann Arbor, Ml CMSA
Ml
Detroit
Door Co., Wl
Wl
Chicago
Duluth, MN-WI MSA
MN
Minneapolis
Edmonson Co., KY
KY
Chicago
El Paso, TX MSA
TX
Albuquerque
Erie, PA MSA
PA
Cleveland
Essex Co., NY
NY
New York City
Evansville, IN-KY MSA
IN-KY
Chicago
Fairbanks, AK
AK
Cleveland
Fayetteville, NC MSA
NC
Atlanta
Flint, Ml MSA
Ml
Detroit
32
-------
Table 6 (continued)
Nonattainment Area/MSA
State
Survey City
Fort Collins-Loveland, CO MSA
CO
Denver
Fresno, CA MSA
CA
San Francisco
Glens Falls, NY MSA
NY
New York City
Grand Rapids, Ml MSA
Ml
Chicago
Great Falls, MT MSA
MT
Billings
Greater Connecticut Metropolitan Area
CT
Boston
Greeley, CO MSA
CO
Denver
Greenbrier Co., WV
WV
Washington, DC
Greensboro-Winston-Salem-High Point PMSA
NC
Atlanta
Greenville-Spartanburg, SC MSA
SC
Atlanta
Hancock Co., ME
ME
Boston
Harrisburg-Lebanon-Carlisle, PA MSA
PA
Philadelphia
Hartford-New Britain-Middletown, CT
CT
Boston
Houston-Galveston-Brazoria, TX CMSA
TX
Dallas
Huntington-Ashland, WV-KY-OH MSA
WV-KY-OH
Washington, DC
Huntsville, AL MSA
AL
Chicago
Indianapolis, IN MSA
IN
Chicago
Jacksonville, FL MSA
FL
Miami
Janesville-Beloit, Wl MSA
Wl
Chicago
Jefferson Co., NY
NY
Philadelphia
Jersey Co., IL
IL
Chicago
Johnson City-Kingsport-Bristol, TN-VA MSA
TN
Atlanta
Johnstown, PA MSA
PA
Philadelphia
Josephine Co., OR
OR
Seattle
Kansas City, MO-KS MSA
MO
Kansas City
Kent and Queen Anne's Cos., MD
MD
Philadelphia
Kewaunee Co., Wl
Wl
Chicago
Kings Co., CA
CA
San Francisco
Klamath Co., OR
OR
San Francisco
Knox Co., ME
ME
Boston
Knoxville, TN MSA
TN
Atlanta
Lafayette-West Lafayette, IN MSA
IN
Chicago
Lake Charles, LA MSA
LA
New Orleans
Lake Tahoe South Shore, CA
CA
San Francisco
Lancaster, PA MSA
PA
Philadelphia
Las Vegas, NV MSA
NV
Las Vegas
Lawrence Co., PA
PA
Cleveland
Lewiston, ME
ME
Boston
Lexington-Fayette, KY MSA
KY
Chicago
Lincoln Co., ME
ME
Boston
Livingston Co., KY
KY
St. Louis
Longmont, CO
CO
Denver
Longview-Marshall, TX MSA
TX
Dallas
Los Angeles-Anaheim-Riverside, CA CMSA
CA
Los Angeles
Los Angeles-South Coast Air Basin, CA
CA
Los Angeles
Louisville, KY-IN MSA
KY-IN
Chicago
Manchester, NH MSA
NH
Boston
Manitowoc Co., Wl
Wl
Chicago
33
-------
Table 6 (continued)
Nonattainment Area/MSA
State
Survey City
Medford, OR MSA
OR
San Francisco
Memphis, TN-AR-MS MSA
TN-AR-MS
St. Louis
Miami-Fort Lauderdale, FL CMSA
FL
Miami
Milwaukee Metropolitan Area
Wl
Chicago
Minneapolis-St. Paul, MN-WI MSA
MN-WI
Minneapolis
Missoula, MT
MT
Billings
Mobile, AL MSA
AL
New Orleans
Modesto, CA MSA
CA
San Francisco
Montgomery, AL MSA
AL
Atlanta
Muskegon, Ml MSA
Ml
Chicago
Nashville, TN MSA
TN
Atlanta
New Orleans, LA MSA
LA
New Orleans
New York-Northern New Jersey-Long Island CMSA
NY-NJ-CT
New York City
Norfolk-Virginia Beach-Newport News, VA MSA
VA
Washington, DC
Northampton Co., VA
VA
Washington, DC
Oklahoma City, OK MSA
OK
Dallas
Owensboro, KY MSA
KY
Atlanta
Paducah, KY
KY
Chicago
Parkersburg, WV
WV
Cleveland
Parkersburg-Marietta, WV-OH MSA
OH-WV
Cleveland
Philadelphia Metropolitan Area
PA-NJ-DE-MD
Philadelphia
Phoenix, AZ MSA
AZ
Phoenix
Pittsburgh-Beaver Valley, PA CMSA
PA
Philadelphia
Portland, ME
ME
Boston
Portland-Vancouver, OR-WA CMSA
OR-WA
Seattle
Portsmouth-Dover-Rochester, NH-ME MSA
ME-NH
Boston
Poughkeepsie, NY MSA
NY
New York City
Providence-Pawtucket-Fall River, RI-MACMSA
MA-RI
Boston
Provo-Orem, UT MSA
UT
Denver
Raleigh-Durham, NC MSA
NC
Atlanta
Reading, PA MSA
PA
Philadelphia
Reno, NV MSA
NV
San Francisco
Richmond-Petersburg
VA
Washington, DC
Rochester, NY MSA
NY
Philadelphia
Sacramento, CA MSA
CA
San Francisco
Salt Lake City-Ogden, UT MSA
UT
Denver
San Antonio, TX MSA
TX
San Antonio
San Diego, CA MSA
CA
Los Angeles
San Francisco-Oakland-San Jose, CACMSA
CA
San Francisco
San Joaquin Valley, CA
CA
San Francisco
Santa Barbara-Santa Maria-Lompoc, CA MSA
CA
Los Angeles
Scranton-Wilkes-Barre, PA MSA
PA
Philadelphia
Seattle-Tacoma, WA
WA
Seattle
Sheboygan, Wl MSA
Wl
Chicago
Smyth Co., VA
VA
Washington, DC
South Bend-Elkhart, IN
IN
Chicago
South Bend-Mishawaka, IN MSA
IN
Chicago
Southeast Desert Modified AQMA, CA
CA
Los Angeles
34
-------
Table 6 (continued)
Nonattainment Area/MSA
State
Survey City
Spokane, WA MSA
WA
Seattle
Springfield, MA MSA
MA
Boston
St. Louis, MO-IL MSA
MO-IL
St. Louis
Steubenville-Weirton, OH-WV MSA
OH-WV
Cleveland
Stockton, CA MSA
CA
San Francisco
Sussex Co., DE
DE
Philadelphia
Syracuse, NY MSA
NY
New York City
Tampa-St. Petersburg-Clearwater, MSA
FL
Miami
Toledo, OH MSA
OH
Detroit
Tulsa, OK MSA
OK
Kansas City
Ventura Co., CA
CA
Los Angeles
Visalia-Tulare-Porterville, CAMSA
CA
San Francisco
Waldo Co., ME
ME
Boston
Walworth Co., Wl
Wl
Chicago
Washington, DC-MD-VA MSA
DC-MD-VA
Washington, DC
Wheeling, WV-OH MSA
WV-OH
Cleveland
Winnebago Co., Wl
Wl
Chicago
Winston-Salem, NC
NC
Atlanta
Worcester, MA MSA
MA
Boston
Yakima, WA MSA
WA
Seattle
York, PA MSA
PA
Philadelphia
Youngstown-Warren, OH MSA
OH
Cleveland
Yuba City, CA MSA
CA
San Francisco
35
-------
Table 7. Substitute Survey City Assignment
Nonattainment Area/MSA
State
Original Survey City
New Survey City
Albany-Schenectady-Troy, NY MSA
NY
New York City
Cleveland
Allentown-Bethlehem, PA-NJ MSA
PA-NJ
Philadelphia
Cleveland
Altoona, PA MSA
PA
Philadelphia
Cleveland
Appleton-Oshkosh-Neenah, Wl MSA
Wl
Chicago
Minneapolis
Beaumont-Port Arthur, TX MSA
TX
Dallas
New Orleans
Bennington Co., VT
vr
Boston
Minneapolis
Bowling Green, KY
KY
Chicago
Cleveland
Buffalo-Niagara Falls, NY CMSA
NY
New York City
Cleveland
Charleston, WV MSA
WV
Washington, DC
Cleveland
Door Co., Wl
Wl
Chicago
Minneapolis
Edmonson Co., KY
KY
Chicago
Cleveland
Essex Co., NY
NY
New York City
Cleveland
Evansville, IN-KY MSA
IN-KY
Chicago
Cleveland
Glens Falls, NY MSA
NY
New York City
Cleveland
Grand Rapids, Ml MSA
Ml
Chicago
Detroit
Greenbrier Co., WV
WV
Washington, DC
Cleveland
Harrisburg-Lebanon-Carlisle, PA MSA
PA
Philadelphia
Cleveland
Huntington-Ashland, WV-KY-OH MSA
WV-KY-OH
Washington, DC
Cleveland
Huntsville, AL MSA
AL
Chicago
Atlanta
Indianapolis, IN MSA
IN
Chicago
Cleveland
Jefferson Co., NY
NY
Philadelphia
Cleveland
Jersey Co., IL
IL
Chicago
Cleveland
Johnstown, PA MSA
PA
Philadelphia
Cleveland
Kewaunee Co., Wl
Wl
Chicago
Minneapolis
Lafayette-West Lafayette, IN MSA
IN
Chicago
Cleveland
Lancaster, PA MSA
PA
Philadelphia
Cleveland
Longview-Marshall, TX MSA
TX
Dallas
New Orleans
Louisville, KY-IN MSA
KY-IN
Chicago
Cleveland
Manitowoc Co., Wl
Wl
Chicago
Minneapolis
Muskegon, Ml MSA
Ml
Chicago
Detroit
Northampton Co., VA
VA
Washington, DC
Atlanta
Oklahoma City, OK MSA
OK
Dallas
St. Louis
Paducah, KY
KY
Chicago
Cleveland
Pittsburgh-Beaver Valley, PA CMSA
PA
Philadelphia
Cleveland
Reading, PA MSA
PA
Philadelphia
Cleveland
Rochester, NY MSA
NY
Philadelphia
Cleveland
Sheboygan, Wl MSA
Wl
Chicago
Minneapolis
Smyth Co., VA
VA
Washington, DC
Atlanta
South Bend-Elkhart, IN
IN
Chicago
Cleveland
South Bend-Mishawaka, IN MSA
IN
Chicago
Cleveland
Syracuse, NY MSA
NY
New York City
Cleveland
Waldo Co., ME
ME
Boston
Minneapolis
Walworth Co., Wl
Wl
Chicago
Minneapolis
York, PA MSA
PA
Philadelphia
Cleveland
36
-------
Table 8. Average Speeds by Road Type and Vehicle Type
(mph)
Rural
Interstate
Principal
Arterial
Minor
Arterial
Major
Collector
Minor
Collector
Local
LDV
60
45
40
35
30
30
LDT
55
45
40
35
30
30
HDV
40
35
30
25
25
25
Urban
Interstate
Other Freeways
& Expressways
Principal
Arterial
Minor
Arterial
Collector
Local
LDV
45
45
20
20
20
20
LDT
45
45
20
20
20
20
HDV
35
35
15
15
15
15
37
-------
Table 9. Counties Modeled with Federal Reformulated Gasoline
State (Northern or Southern
RFG-N or S)/
Nonattainment Area
County
State (Northern or Southern
RFG-N or S)/
Nonattainment Area
County
Connecticut (N)
Greater Connecticut
Hartford Co
Litchfield Co
Middlesex Co
New Haven Co
New London Co
Tolland Co
Windham Co
New York-Northern New Jersey-Long Island
Fairfield Co
District of Columbia (S)
Washington DC
Washington
Delaware (N)
Philadelphia-Wilmington-Trenton
Kent Co
New Castle Co
Sussex County
Sussex Co
Illinois (N)
Chicago-Gary-Lake County
Cook Co
Du Page Co
Grundy Co
Kane Co
Kendall Co
Lake Co
McHenry Co
Will Co
Indiana (N)
Chicago-Gary-Lake County
Lake Co
Porter Co
Kentucky (N)
Cincinnati-Hamilton
Boone Co
Campbell Co
Kenton Co
Maine (N)
Knox & Lincoln Counties
Knox Co
Lincoln Co
Lewiston-Auburn
Androscoggin Co
Kennebec Co
Portland
Cumberland Co
Sagadahoc Co
York Co
Maryland (S)
Baltimore
Anne Arundel Co
Baltimore
Baltimore Co
Carroll Co
Harford Co
Howard Co
Kent & Queen Annes Counties
Kent Co
Queen Annes Co
Philadelphia-Wilmington-Trenton
Cecil Co
Washington DC
Calvert Co
Charles Co
Frederick Co
Montgomery Co
Prince Georges Co
Massachusetts (N)
Boston-Lawrence-Worcester-Eastern MA
Barnstable Co
Bristol Co
Dukes Co
Essex Co
Middlesex Co
Nantucket Co
Norfolk Co
Plymouth Co
Suffolk Co
Worcester Co
38
-------
Table 9. Counties Modeled with Federal Reformulated Gasoline
State (Northern or Southern
RFG-N or S)/
Nonattainment Area
County
State (Northern or Southern
RFG-N or S)/
Nonattainment Area
County
Louisville
Bullitt Co
Jefferson Co
Oldham Co
New Hampshire (N)
Manchester
Hillsborough Co
Merrimack Co
Portsmouth-Dover-Rochester
Rockingham Co
Strafford Co
New Jersey (N)
Allentown-Bethlehem-Easton
Warren Co
Atlantic City
Atlantic Co
Cape May Co
New York-Northern New Jersey-Long Island
Bergen Co
Essex Co
Hudson Co
Hunterdon Co
Middlesex Co
Monmouth Co
Morris Co
Ocean Co
Passaic Co
Somerset Co
Sussex Co
Union Co
Philadelphia-Wilmington-Trenton
Burlington Co
Camden Co
Cumberland Co
Gloucester Co
Mercer Co
Salem Co
New York (N)
New York-Northern New Jersey-Long Island
Bronx Co
Kings Co
Nassau Co
Springfield/Pittsfield-Western MA
Berkshire Co
Franklin Co
Hampden Co
Hampshire Co
New York (N)
Poughkeepsie
Dutchess Co
Putnam Co
Pennsylvania (N)
Philadelphia-Wilmington-Trenton
Bucks Co
Chester Co
Delaware Co
Montgomery Co
Philadelphia Co
Rhode Island (N)
Providence
Bristol Co
Kent Co
Newport Co
Providence Co
Washington Co
Texas(S)
Dallas-Fort Worth
Collin Co
Dallas Co
Denton Co
Tarrant Co
Houston-Galveston-Brazoria
Brazoria Co
Chambers Co
Fort Bend Co
Galveston Co
Harris Co
Liberty Co
Montgomery Co
Waller Co
Virginia (S)
Norfolk-Virginia Beach-Newport News
Chesapeake
Hampton
James City Co
39
-------
Table 9. Counties Modeled with Federal Reformulated Gasoline
State (Northern or Southern
State (Northern or Southern
RFG-N or S)/
RFG-N or S)/
Nonattainment Area
County
Nonattainment Area
County
New York Co
Newport News
Orange Co
Norfolk
Queens Co
Poquoson
Richmond Co
Portsmouth
Rockland Co
Suffolk
Suffolk Co
Virginia Beach
Westchester Co
Williamsburg
York Co
Virginia (S)
Richmond-Petersburg
Charles City Co
Chesterfield Co
Colonial Heights
Hanover Co
Henrico Co
Hopewell
Richmond
Washington DC
Alexandria
Arlington Co
Fairfax
Fairfax Co
Falls Church
Loudoun Co
Manassas
Manassas Park
Prince William Co
Stafford Co
Wisconsin (N)
Milwaukee-Racine
Kenosha Co
Milwaukee Co
Ozaukee Co
Racine Co
Washington Co
Waukesha Co
40
-------
Table 10. Oxygenated Fuel Modeling Parameters
Market Shares (%)
Oxvaen Content (%)
Oxygenated
State
County
MTBE
Alcohol Blends
MTBE
Alcohol Blends
Fuel Season
Alaska
Anchorage Ed
0
100
2.7
3.5
NOV-FEB
Arizona
Maricopa Co
80
20
2.7
3.5
OCT - FEB
Colorado
Adams Co
75
25
2.7
3.5
NOV-FEB
Colorado
Arapahoe Co
75
25
2.7
3.5
NOV-FEB
Colorado
Boulder Co
75
25
2.7
3.5
NOV-FEB
Colorado
Douglas Co
75
25
2.7
3.5
NOV-FEB
Colorado
Jefferson Co
75
25
2.7
3.5
NOV-FEB
Colorado
Denver Co
75
25
2.7
3.5
NOV-FEB
Colorado
El Paso Co
75
25
2.7
3.5
NOV-FEB
Colorado
Larimer Co
75
25
2.7
3.5
NOV-FEB
Connecticut
Fairfield Co
90
10
2.7
3.5
NOV-FEB
Minnesota
Anoka Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Carver Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Dakota Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Hennepin Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Ramsey Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Scott Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Washington Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Wright Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Chisago Co
10
90
2.7
3.5
OCT - JAN
Minnesota
Isanti Co
10
90
2.7
3.5
OCT - JAN
Montana
Missoula Co
0
100
2.7
3.5
NOV-FEB
Nevada
Clark Co
0
100
2.7
3.5
OCT - MAR
Nevada
Washoe Co
95
5
2.7
3.5
OCT - JAN
New Jersey
Bergen Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Essex Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Hudson Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Hunterdon Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Middlesex Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Monmouth Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Morris Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Ocean Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Passaic Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Somerset Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Sussex Co
95
5
2.7
3.5
NOV-FEB
New Jersey
Union Co
95
5
2.7
3.5
NOV-FEB
New York
Bronx Co
95
5
2.7
3.5
NOV-FEB
New York
Kings Co
95
5
2.7
3.5
NOV-FEB
New York
Nassau Co
95
5
2.7
3.5
NOV-FEB
New York
New York Co
95
5
2.7
3.5
NOV-FEB
New York
Queens Co
95
5
2.7
3.5
NOV-FEB
New York
Richmond Co
95
5
2.7
3.5
NOV-FEB
New York
Rockland Co
95
5
2.7
3.5
NOV-FEB
New York
Suffolk Co
95
5
2.7
3.5
NOV-FEB
New York
Westchester Co
95
5
2.7
3.5
NOV-FEB
New York
Orange Co
95
5
2.7
3.5
NOV-FEB
New York
Putnam Co
95
5
2.7
3.5
NOV-FEB
Oregon
Clackamas Co
1
99
2.7
3.5
NOV-FEB
Oregon
Jackson Co
1
99
2.7
3.5
NOV-FEB
Oregon
Multnomah Co
1
99
2.7
3.5
NOV-FEB
Oregon
Washington Co
1
99
2.7
3.5
NOV-FEB
Oregon
Josephine Co
1
99
2.7
3.5
NOV-FEB
Oregon
Klamath Co
1
99
2.7
3.5
NOV-FEB
Oregon
Yamhill Co
1
99
2.7
3.5
NOV-FEB
Texas
El Paso Co
15
85
2.7
3.5
NOV-FEB
Utah
Utah Co
20
80
2.7
3.5
NOV-FEB
Washington
Clark Co
1
99
2.7
3.5
NOV-FEB
Washington
Spokane Co
1
99
2.7
3.5
SEP - FEB
Wisconsin
St. Croix Co
10
90
2.7
3.5
OCT - JAN
41
-------
Table 11. Number of Vehicles in Puerto Rico by Vehicle Type and Percent of Total
Vehicle Type Number Percent
Passenger Car
1,847
,980
82.800%
Motorcycle
32,
,030
1.435%
2-axle/4-tire
277,
,360
12.427%
Other Single Unit
32,
,378
1.451%
Combination Trucks
38,
,600
1.729%
Buses
3.
,515
0.157%
Table 12. Estimation of Road Length by Road Type in USVI
(Kauai Island, pop. 51,177 / Area 549 sq mi)
Roadtype Roadlength Percent of Total Estimated Roadway Mileage (miles)
Kauai
Kauai (mi)
Road Mileage
USVI Total
St Croix
St Thomas
St John
2
8.31
2.02%
10.73
5.05
5.23
0.45
6
47.81
11.60%
61.74
29.08
30.09
2.56
7
79.13
19.21%
102.18
48.13
49.81
4.24
8
2.9
0.70%
3.74
1.76
1.83
0.16
9
186.84
45.35%
241.27
113.65
117.61
10.01
14
8.81
2.14%
11.38
5.36
5.55
0.47
16
7.09
1.72%
9.16
4.31
4.46
0.38
17
19.35
4.70%
24.99
11.77
12.18
1.04
19
51.75
12.56%
66.82
31.48
32.57
2.77
Total
411.99
100.00%
532
250.6
259.33
22.07
Table 13. Population Estimates in St. Thomas, St. John, and St Croix
St. Thomas
St. John
St. Croix
Average tourist population on any given day
5,558
473
1,039
Resident population as of 1999
56,831
4,837
59,249
Overall population of each island on a given day
62,389
5,310
60,288
Percentage oftotal USVI population
48.75%
4.15%
47.10%
42
-------
Table 14. Percentage of Total USVIVMT by Roadway Type and Island
Road Type
St Croix
St Thomas
St John
2
0.95%
0.98%
0.08%
6
5.47%
5.66%
0.48%
7
9.05%
9.36%
0.80%
8
0.33%
0.34%
0.03%
9
21.36%
22.11%
1.88%
14
1.01%
1.04%
0.09%
16
0.81%
0.84%
0.07%
17
2.21%
2.29%
0.19%
19
5.92%
6.12%
0.52%
Total
47.10%
48.75%
4.15%
Table 15. 25 Year Trend of Vehicle Registrations and New Sales in Puerto Rico
Year
New Vehicle Sales
Total Vehicle Registrations
1973
138,108
681,596
1974
66,738
738,485
1975
73,388
773,742
1976
83,505
814,373
1977
110,393
830,373
1978
101,254
980,200
1979
103,859
1,035,200
1980
88,000
1,120,312
1981
98,193
1,201,774
1982
66,158
1,228,405
1983
60,987
1,259,111
1984
92,974
1,245,000
1985
116,431
1,353,670
1986
141,219
1,451,281
1987
118,048
1,560,308
1988
131,958
1,551,415
1989
148,459
1,567,319
1990
125,577
1,582,081
1991
116,386
1,516,102
1992
113,682
1,650,709
1993
141,550
1,740,371
1994
146,951
1,872,361
1995
160,394
2,014,207
1996
147,605
2,166,697
1997
180,027
2,272,643
43
-------
Table 16. Average Temperature Data for Puerto Rico and USVI (°F)
City Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Charlotte Amalie, USVI, 84 84 85 86 87 88 89 89 89 88 87 85
Avg Daily High
Charlotte Amalie, USVI, 73 73 74 76 78 79 80 79 79 78 76 74
Avg Daily Low
San Juan, PR, 83 84 85 86 87 89 88 89 89 88 86 84
Avg Daily High
San Juan, PR, 70 70 71 73 74 76 76 76 76 75 74 72
Avg Daily Low
44
-------
Attachment A
County-Specific Fuel Parameters for 1990, 1996, and 1999 Toxic
Emissions Modeling (Preparation for MOBILE6.2 Model Runs)
Prepared for:
Emission Factor and Inventory Group (D205-01)
Emissions, Monitoring and Analysis Division
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Prepared by:
Eastern Research Group, Inc.
1600 Perimeter Park Drive
Morrisville, North Carolina 27560
October 9, 2002
45
-------
TABLE OF CONTENTS
Section Page
1.0 Purpose 1
2.0 Overview of Methodology 1
3.0 Data Requirements 1
4.0 Available Data 2
5.0 Mapping Methodology 6
6.0 Outputs 8
Figure Page
1 Decision Making Process for 1999 Fuel Assignments 7
Appendix A - Mapping Assignments by Analysis Year - 1999
Appendix B - Mapping Assignments by Analysis Year - 1996
Appendix C - Mapping Assignments by Analysis Year - 1990
0136.02.004.003/MOBILE6.wpd
1
-------
1.0 PURPOSE
In this exercise, fuel parameter data were collected and processed to estimate toxic emissions estimates for on-
road gasoline vehicles, for calendar years 1990, 1996, and 1999. Resulting data were used as inputs into
EPA's draft MOBILE6.2 emission factor model to update existing National Toxics Inventory (NTI) estimates
for the 6 hazardous air pollutants (HAPs) modeled by the new MOBILE model1:
• Benzene
• Formaldehyde
• Acetaldehyde
• 1,3 Butadiene
• Acrolein
• Methyl Tertiary Butyl Ether (MTBE)
Using MOBILE6.2 will provide more accurate emission rate estimates compared to previous efforts, which
utilized a combination of earlier models, (i.e., MOBILE5 and MOBTOX).
2.0 OVERVIEW OF METHODOLOGY
Fuel parameters were collected for winter and summer seasons using a number of different data sources. The
seasonal data were "mapped" to the county level for all 50 states, for the 1990, 1996, and 1999 calendar
years. The data contained all of the required fields for use in the new MOBILE6.2 model, and was organized
in a standardized flat ASCII file and provided to E.H. Pechan for further processing. E.H. Pechan
subsequently generated input files for four seasons to account for temperature variations (summer fuel
parameters used for the fall, and winter for spring), ran MOBILE6.2 for each county/year/season combination,
and combined the resulting emission factors with county level VMT estimates to generate mass emissions for
each of the 6 HAPs listed above.
3.0 DATA REQUIREMENTS
The new MOBILE6.2 model requires highly detailed fuel parameter information in order to generate HAP
estimates. These new inputs include:
Input Parameter
Sulfur
Olefins
Aromatics
Benzene
E200
E300
Oxygenate Content by Type
Oxygenate Sales Fraction by Type
Description
ppm
% by volume
% by volume
% by volume
% vapor at 200 degrees Fahrenheit
% vapor at 300 degrees Fahrenheit
MTBE/Ethanol/ETBE/TAME % by volume
Sales fraction for each oxygenate in %
1 Additional HAPs are being modeled using MOBILE6.2 under a separate EPA work assignment.
0136.02.004.003/MOBILE6.wpd 1
-------
Note, that only one oxygenate can be modeled at a time by MOBILE6.2. Blends of more than one oxygenate
cannot be modeled; however, different single oxygenate fuels can be modeled separately, with the final results
weighted by relative sales fractions for each oxygenate type. The "Oxygenate Sales Fraction" data listed above
provides the necessary weightings for this operation.
Also note, that none of these data were required for previous versions of the MOBILE model, and had to be
compiled in many cases for the first time.
4.0 AVAILABLE DATA
Local and regional winter and summer fuel sampling data containing the required input parameters were
available from three sources.2
EPA Reformulated Gasoline Surveys - These surveys have been performed in most major RFG areas since
the mid-1990s. All of the required MOBILE6.2 fuel parameters are included in the survey results, with the
exception of the California area surveys which only provide oxygenate information. (For this reason an
alternative data source - the Alliance Surveys, see below - were used for the California counties.) Samples
were averaged across fuel grades (regular, mid, and premium) prior to reporting. Unlike the Alliance and TRW
surveys discussed below, oxygenate sales fraction data were provided as well. No RFG area surveys were
performed in 1990, since RFG programs had not yet been developed or implemented at that time.
In addition, some minor adjustments were necessary to develop fuel parameter estimates in the units required
by MOBILE6.2. These adjustments included:
Conversion of oxygenate from %weight to be %volume, using guidance as presented in User's Guide to
Mobile6.1 and Mobile6.2 section 2.8.10.7f. (The following equations assume there is only a single oxygenate
in the fuel.)
Volume percent MTBE= Weight Percent Oxygen / .1786
Volume percent ETBE= Weight Percent Oxygen / . 1533
Volume percent TAME= Weight Percent Oxygen / . 1636
Volume percent ETOH= Weight Percent Oxygen / .3448
The following lists the RFG survey areas by calendar year of interest.
RFG Survey Area
Required RFGOpt-In (Voluntary)
1995
1995
1995
1995
RFG
Y
Y
Y
Y
Y
Y
Y
Atlantic City, NJ
Baltimore, MD
Boston-Worchester, MA
Chicago-Lake Co., IL, Gary, IN
Covington, KY
Dallas-Fort Worth, TX
Hartford, CT
1995
1995
1995
With the exception of the RFG area surveys, the data sets did not provide information on oxygenate sales
fractions. This information was provided supplementally using Federal Highway Administration data, as
discussed below.
0136.02.004.003/MOBILE6.wpd
2
-------
Houston-Galveston, TX
Los Angeles, CA
Louisville, KY
Manchester, NH
Milwaukee-Racine, WI
NY-NJ-Long Is.-CT
Norfolk-Virginia Beach, VA
Phila.-Wilm, DE-Trenton, NJ
Portsmouth-Dover, NH
Poughkeepsie, NY
Rhode Island
Richmond, VA
Sacramento Metro, CA
San Diego, CA
Springfield, MA
St Louis,MO
Sussex County, DE
Warren County, NJ
Washington, D.C.-area
Phoenix, AZ
Knox, Lewiston and Portland, ME
1995
1995
1994
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1999
1995
1995
1995
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
1997-1998
until 1999
The Alliance of Automobile Manufacturers (AAM) North American Gasoline and Diesel Fuel
Survey - These surveys have been conducted since prior to 1990 for a number of different areas across the
United States, Mexico, and Canada. While there is some overlap with the RFG Survey areas, many areas lie
outside RFG program regions. Separate results were provided for each fuel grade. Oxygenate fuel sales
fractions were not provided.
This information is copyrighted and must be purchased from the Alliance for use.
The following lists the different Alliance area surveys used in this analysis (i.e., those not overlapping with the
RFG Survey areas).
Albuquerque, NM
Atlanta, GA
Billings, MT
Boston, MAC
Chicago, ILC
Cleveland, OH
Dallas, TXC
Denver, CO
Detroit, MI
a Data collection initiated in 1994.
b Data collection initiated in 2000.
c AAMA data were used only for 1996
Fairbanks, AKb
Kansas City, MO
Las Vegas, NV
Los Angeles, CA
Miami, FL
Minneapolis/ St. Paul, MN
New Orleans, LA
New York City, NY"
and 1990 fuel parameters.
Philadelphia, PAC
Phoenix, AZ
Pittsburgh, PAa
St. Louis, MO
San Antonio, TX
San Francisco, CA
Seattle, WA
Washington, DCC
In addition, some minor adjustments were necessary to develop fuel parameter estimates in the units required
by MOBILE6.2. These adjustments included:
1. Multiply sulfur content by 10,000 to convert from percent concentration to parts per million.
0136.02.004.003/MOBILE6.wpd
3
-------
2. For the 1996 and 1990 data sets, E200 and E300 were extracted from the raw survey data, with
interpolation between points from the distillation curve.
TRW Petroleum Technologies Survey - Formerly known as the National Institute for Petroleum and Energy
Research (NIPER), these surveys are available for all three target years. Data were collected and presented
for 15 Districts across the United States (not including Alaska and Hawaii). Survey data were reported
separately for each grade of gasoline, as well as reformulated versus conventional fuel, and fuel with and
without alcohols.
The following lists the TRW/NIPER survey districts.
District
1 (Northeast)
2 (Mid-Atlantic Coast)
3 (Southeast)
4 (Florida)
5 (North Central)
6 (Ohio Valley)
7 (Central and Upper Plains)
8 (Oklahoma and East Texas)
9 (North Mountain States)
10 (Central Mountain States)
11 (New Mexico, West Texas)
12 (West Southwest)
13 (Pacific Northwest)
14 (North California and North Nevada)
15 (South California)
States
Connecticut, Massachusetts, New Jersey,
New York, Pennsylvania, Rhode Island
Washington, DC, Maryland, Virginia
Alabama, Arkansas, Georgia, Louisiana,
North Carolina, South Carolina, Tennessee
Northern Illinois, Michigan, Minnesota,
Wisconsin
Indiana, Kentucky, West Virginia, Ohio
North Dakota, South Dakota, Nebraska,
Kansas, Iowa, Missouri, Southern Illinois
Montana, Wyoming, Eastern Washington,
Eastern Oregon
Colorado, Utah
Arizona, Southern Nevada, Southeastern
CA
Western Washington, Western Oregon
Note, that each survey District does not necessarily report the full slate of possible fuel type combinations. For
example, District 3 (Southeast) does not report any values for fuels containing alcohols. In many cases, this is
consistent with independent data sources such as the Federal Highway Administration data on statewide
alcohol fuel sales. However, in certain cases discrepancies may appear, and are noted accordingly (see
Mapping Methodology below for details).
Also note, that the 1990 NIPER survey data reported results for leaded fuel in addition to the other fuel types
noted above. However, given that less than 1% of total 1990 fuel sales in the U.S. were leaded, these data
were not included in the analysis.
In addition, some minor adjustments were necessary to develop fuel parameter estimates in the units required
by MOBILE6.2. These adjustments included:
0136.02.004.003/MOBILE6.wpd
4
-------
1. Multiply sulfur content by 10,000 to convert from percent concentration to parts per million.
2. E200 and E300 were extracted from the raw survey data with interpolation between points from the
distillation curve.
Additional Data Sources - As noted above, only the RFG Survey data contained information on the relative
fraction of different oxygenates in their fuel samples. Therefore additional data from the Federal Highway
Administration (FHWA) were used to supplement the Alliance and TRW/NIPER data.3
The FHWA data contained estimates of percent alcohol on an annual basis, at the state level. These data were
rather aggregated compared to the seasonal, county-level detail required by the model. Therefore descriptions
of Oxy-Fuel Program areas were obtained from EPA to further refine alcohol market share estimates in these
cases. Oxy Fuel Program descriptions included program start and end dates, as well as ethanol/MTBE splits,
and overall implementation and opt-out dates as appropriate for each area. Specific use of this supplemental
information is noted in the Mapping Methodology section below.
Finally, as noted above, most data sources reported sample data separately by fuel grade. Therefore grade-
specific sales data were needed to developing weighting factors for composite fuel parameter estimates. The
required data was obtained at the State level
from the Petroleum Marketing Annual reports (Energy Information Administration (EIA), Office of Oil and Gas,
Department of Energy)4
5.0 MAPPING METHODOLOGY
After survey results were composited across the different fuel grades using the EIA data, rules were developed
for assigning specific surveys to specific counties. As a first step, "de minimis" criteria were established to
simplify modeling when very low, inconsequential levels of certain fuel parameters appeared in the data.
Specifically, if measured oxygenate percent volume content was less than 0.1 percent, oxygenate content was
set to zero, as was the corresponding market share. This was true for MTBE, ethanol, ETBE, and TAME.
Next, a hierarchical approach was adopted, according to the following priorities:
1. If appropriate RFG Survey data were available for a given county (County designations
corresponded to RFG area definitions). If not, then
2. Alliance survey data were used if available for a given county (County assignments were made
according to MSA designations). Otherwise
3. Appropriate district level data from TRW/NIPER were used.
3 http://www.fhwa.dot.gov/ohim/1999/index.html
4 http://www.eia.doe.gov/oil_gas/petroleum/data_publications/petroleum_marketing_annual/pma_
historical.html
0136.02.004.003/MOBILE6.wpd
5
-------
While the overall decision criteria followed the above prescription, data limitations and other concerns (e.g.,
information on Oxy-Fuel Program areas) sometimes necessitated alternative assignments. Most importantly, in
1990 AAM and NIPER surveys had to be used in place of the missing RFG survey data.)
A detailed schematic of the decision process for the 1999 calendar year is provided below (see Figure 1). The
specific assignments and decisions for each state and surveyed area, for each analysis year, are provided in the
Appendix.
Lfcl or All
Counties in US
By FIPS
Assigned Cairiv to Each
Stiri/ey Area as Lislsd arr
EPA Wabsite
Is this county sxfel in
RFG data ? ,
No
Assigned County h Each
Survey Area as Lisisdw V\
Census Websiie Sor
t*/lslri>plilan Aiss sndMuiitpi\ ¦
wilh SaJe Fraction iof Each
Graded G"as aline from DOl'
WebSite
1SS6MM
Survey Data
(35 areas)
Is this not my -:xb! try
AAV data ?
Add tyl&ket Store
wt/i Multiple D/s)r/ef mil use mi's /Kf ,v
guanine la assigned survey sres
No
Note
EPA Eii vimnmental Protectant Agency
DOE Qepartmeitt! or Energy
RFG Reformulate Gasoline
AAM Alliance o( Automobile Manufacturers
TRW TRW Petroleum Technologist
fs this county lisl survt
area as city or regian1
Figure 1. Decision Making Process for 1999 Fuel Assignments
0136.02.004.003/MOBILE6.wpd
6
-------
Mapping assignments for the "Corn Belt" states were particularly problematic, since many of these areas have
significant volumes of both ethanol and MTBE in their fuel. While the RFG and Alliance Surveys report their
results composited across oxygenate types, the TRW/NIPER surveys report fuel parameter data separately for
fuels with and without alcohols. Since it was not possible to combine the alcohol and non-alcohol fuel
parameters into one table, an alternative strategy was developed.5 Specifically, for those counties with
non-trivial fractions of both alcohol and ether oxygenates, separate county entries were provided for ethers and
alcohols, each with 100% "market share" listed for MOBILE6. This requires MOBILE6 to be run twice for
each of these counties, with the model outputs combined with the VMT weighted by relative oxygenate sales
fractions from the FHWA.
Finally, as a simplifying assumption, for TRW/NIPER surveys significant measured volumes of MTBE and
another ether (ETBE or TAME), we assigned 100% of the sales fraction to MTBE. This decision was made in
the absence of any other data on relative sales fractions across different ether species. (Note, that the
MOBILE model is more sensitive to relative changes in alcohol versus ether content in general, and less
sensitive to the specific ether involved.)
The specific mapping assignments for each area are described in the Appendix, for each analysis year.
6.0 OUTPUTS
Once mapping assignments were complete, flat ASCII files were generated for each calendar year of interest,
containing both winter and summer results. For most counties, there is one row entry for winter, and one for
summer for each year. If there is more than one oxygenate present in significant quantities (i.e., > 0.1%), then
county row entries were "duplicated", one for each oxygenate, with the oxygenate sales fraction provided in the
last column of the file.
The following lists the fields in order for each file.
• County FIPS code
• State
• County Name
• Season
• Fuel RVP6
• Sulfur
• Olefin content
• Aromatics content
• Benzene content
. E200
. E300
. MTBE (% vol)
• MTBE sales fraction
• Ethanol (% vol)
5 Simple averaging of fuel parameters, or linear weighting by sales fractions, would not take into account
the non-linear effects of blending, such as we would expect for E200/300, for example.
6 Note that the RVP values used in this analysis are different from those used in the most recent criteria
pollutant modeling. Therefore, there will be inconsistencies between these HAP figures and the previous
criteria estimates.
-------
• Ethanol sales fraction
. ETBE (% vol)
• ETBE sales fraction
And for user reference the following fields are also provided:
• Source of data
• RFG area (l=yes)
• AAM area (l=yes)
• Associated TRW District
• Local RFG survey area, if applicable
• Local Alliance survey area, if applicable
• TRW/NIPER Table # reference, if applicable
• Source of oxygenate market share information (RFG, AAM, FHWA)
• FHWA oxygenate fuel sales fraction
0136.02.004.003/MOBILE6.wpd
8
-------
Appendix A
Mapping Assignments by Analysis Year ~ 1999
-------
Mapping Assignments by Analysis Year - 1999
1. AL — 0.16% Ethanol sales market share as per FHWA, but no alcohol samples (table 10) for TRW
District 3. Therefore entire state was filled by data from TRW table 9 with 100% MTBE market share.
2. AK was filled by data from data from AAM survey (Fairbanks, AK) with FHWA market share
between MTBE and ETOH.
3. AZ
• Phoenix area was filled by data from data from AAM survey (Phoenix, AZ) with EtOH market
share for winter = 100% (as per EPA Oxy Fuel Program Summary, October 2001), and
100%) MTBE for summer. Compares with statewide annual average of-7.6% from FHWA.
• The rest of AZ was filled by data from TRW table 9 with 100% MTBE market share.
4. AR was filled by data from TRW table 9 with 100% MTBE market share.
5. CA
• San Francisco Bay area was filled by data from AAM survey (San Francisco, CA) with 50/50
split market share between MTBE and EtOH in summer and 100% MTBE market share in
winter. Corresponding to FHWA reported on annual average of 6.36% EtOH statewide.
• The rest of CA was filled by data from Los Angeles AAM survey, with 100% MTBE.
6. CO
• Denver area was filled by data from AAM survey (Denver, CO) with 100 % ETOH market
share all year round.
• The rest of CO was filled by data from TRW table 9 with 100% MTBE market share in
summer and TRW table 10 with 100% ETOH market share in winter to compare well with
27.27%) EtOH annual statewide average from FHWA.
7. CT was filled by data from RFG survey to all RFG areas in Connecticut with EPA RFG market share
survey. Except Fairfield County, which was assigned to the NY-NJ-Long Island RFG survey, with
market share from the EPA RFG Survey. Only a slight discrepancy with FHWA number of 221%
EtOH.
0136.02.004.003/MOBILE6.wpd
A-l
-------
Mapping Assignments by Analysis Year - 1999
8. DE
• All of the counties except Sussex were filled by data from RFG survey (Phila.-Wilm, DE-
Trenton, NJ) with RFG Survey market share between MTBE and ETOH. Slight contradiction
with FHWA value of 0% for the state.
• Sussex County was filled by data from RFG survey (Sussex County, DE) with 100% MTBE
market share from RFG survey as well.
9. DC was filled by RFG survey (Washington, DC) with RFG Survey market share between MTBE and
ETOH. Slight discrepancy with FHWA value of 0%.
10. FL
• Miami area was filled by data from AAM survey (Miami, FL) with 100% MTBE market share.
• The rest of FL was filled by data from TRW table 9 with 100% MTBE market share.
• Essentially same as FHWA value of 0% EtOH for the state.
11. GA
• Atlanta area was filled by data from AAM survey (Atlanta, GA) with 100% MTBE market
share.
• The rest of GA was filled by data from TRW table 9 with 100% MTBE market share.
• Consistent with FHWA value of 0%.
12. HI was filled by TRW table 9 (district 14 Northern California) with 100% MTBE market share.
13. IA
• Duplicate each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%
• If the data came from table 10 then ETOH market share is equal to 100%
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
14. ID was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA of 0.
15. IL
• Chicago area was filled by data from AAM survey (Chicago, IL) with 100% EtOH market
share.
• Counties in St Louis were filled by data from AAM survey (St. Louis, MO) with RFG survey
market share between MTBE and EtOH.
• The rest of the counties were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
0136.02.004.003/MOBILE6.wpd
A-2
-------
Mapping Assignments by Analysis Year - 1999
• Assumes St. Louis at 0% and Chicago at -95% EtOH "cancel out" with rest of the state at
FHWA level of 49.22% to meet statewide average.
16. IN
• Counties in Chicago area were filled by data from AAM survey (Chicago, IL) with 100%
EtOH market share.
• The rest of the counties were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
17. KS
• Duplicate each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
18. KY
• Counties in Covington, KY RFG area were filled by data from RFG survey with RFG market
share.
• Counties in Louisville RFG area were filled by data from Louisville RFG survey with RFG
market share.
• The rest of the state was filled by table 9 of TRW survey with 100% MTBE market share.
• The rest of the state was filled by TRW table 9 with 100% MTBE market share.
• May slightly overestimate EtOH when compared with FHWA statewide market share of
1.52% EtOH.
19. LA
• New Orleans area was filled by data from AAM survey (New Orleans, LA) with 100% MTBE
market share.
• The rest of LA was filled by data from TRW table 9 with 100% MTBE market share.
• Essentially same as FHWA value of 0.65% EtOH for the state.
20. MA
• Counties in Boston area were filled by RFG survey (Boston-Worchester, MA) with RFG
Survey market share between MTBE and ETOH.
• Counties in Springfield area were filled by RFG survey (Springfield, MA) with RFG Survey
market share between MTBE and ETOH.
21. MD
• Counties in Philadelphia area were filled by RFG survey (Phila.-Wilm, DE-Trenton, NJ) with
RFG Survey market share between MTBE and ETOH.
0136.02.004.003/MOBILE6.wpd
A-3
-------
Mapping Assignments by Analysis Year - 1999
Counties in Washington DC area were filled by RFG survey (Washington DC) with RFG
Survey market share between MTBE and ETOH.
Counties in Baltimore area were filled by RFG survey (Baltimore, MD) with RFG Survey
market share between MTBE and ETOH.
The rest of the state was filled by data from TRW table 9 with 100% MTBE.
22. ME
• ME Counties in RFG area were filled by TRW table 11 with 100% MTBE market share.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share.
Consistent with FHWA value of 0% EtOH.
23. MI
• Counties in Detroit area were filled by AAM survey (Detroit, MI) with 100% ETOH market
share in.
• The rest of the state were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
24. MN
• Counties in Minneapolis/St. Paul Detroit area were filled by AAM survey (Minneapolis/St.
Paul, MN) with 100% ETOH market share both winter and summer, based on low (0.1)
measured MTBE values.
• The rest of the state were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
25. MO
• Counties in St. Louis and Kansas City areas were filled by AAM survey (St. Louis, MO and
Kansas City, MO). St. Louis area was assigned percent MTBE and EtOH market share as
report in RFG Survey. Kansas City area was assigned 100% MTBE market share.
• The rest of the counties were filled by TRW table 9 with 100% market share.
• Note: FHWA data indicate a 5.3% EtOH sales fraction for the state. Prior years also
indicated non-trivial EtOH sales fractions and county entries were duplicated accordingly for
1996 and 1990. We recommend similar duplications for the 1999 calendar year in the future.
26. MS was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA value of 0%.
27. MT
• Yellow Stone County was filled by data from AAM survey (Billings, MT) with 100%) MTBE
market share in summer and winter.
0136.02.004.003/MOBILE6.wpd
A-4
-------
Mapping Assignments by Analysis Year - 1999
• The rest of the state was filled by data from TRW table 9 with 100% MTBE share.
• Except Missoula county that was filled by data from TRW table 10 with 100%. ETOH market
share in the Winter as per EPA's Oxy Fuel Program Summary.
28. NC was filled by TRW table 9 with 100% MTBE market share. Note that FHWA has reported
7.47%) of gasohol was used in this state, TRW survey did not have any survey collect on gasoline
containing alcohol in this area.
29. ND
• Duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%>.
• If the data came from table 10 then ETOH market share is equal to 100%>.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
30. NE
• Duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%>.
• If the data came from table 10 then ETOH market share is equal to 100%>.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
31. NH
• Counties located in Manchester, NH and Portsmouth-Dover, NH metropolitan area were filled
by RFG survey (Manchester, NH and Portsmouth-Dover, NH) with RFG Survey market share
between MTBE and ETOH.
• The rest of the state was filled by data from TRW table 9 with 100%> MTBE.
32. NJ
• Counties in NY-NJ-Long Island-CT RFG area were filled by RFG survey (NY-NJ-Long Is-
CT) with RFG Survey market share between MTBE and ETOH.
• Counties in Trenton area were filled by RFG survey (Phila.-Wilm, DE-Trenton, NJ) with RFG
Survey market share between MTBE and ETOH.
• Counties in Atlantic City area were filled by RFG survey (Atlantic City, NJ) with RFG Survey
market share.
• Somewhat underestimates FHWA estimate of 2.10% EtOH.
33. NM
• Counties in Albuquerque area were filled by RFG survey (Albuquerque, NM) with 100%)
ETOH market share in the winter, as per Oxy Fuel Program, description. 100%> EtOH share
was assumed during the summer as well due to the low measured levels of MTBE (0.1) vs.
EtOH (0.8).
• The rest of the state was filled by data from TRW table 9 with 100%> MTBE market share in
summer and TRW table 10 with 100%> ETOH market share in winter, as there were no data
0136.02.004.003/MOBILE6.wpd
A-5
-------
Mapping Assignments by Analysis Year - 1999
for summer alcohol fuels in NIPER District 11, and because winter MTBE levels were
measured as 0 in Table 9 here as well.
34. NV
• Counties in Las Vegas area were filled by RFG survey (Las Vegas, NV) with 100% ETOH
market share in the winter as per the Oxy Fuel Program description. Summer market share set
to 100% MTBE to be more consistent with FHWA estimate of 0%.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share to
be more consistent with FHWA estimate of 0% EtOH.
35. NY
• Counties in NY-NJ-Long Island-CT area were filled by RFG survey (NY-NJ-Long Island-
CT) with the RFG Survey market share.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share
(since there was no Table 10 in the TRW data set).
36. OH
• Counties in Cleveland area were filled by RFG survey (Cleveland, OH) with 100% ETOH
market share, based on low measured values of MTBE (-0.1).
• The rest of the state was duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
37. OK was filled by TRW table 9 with 100% MTBE market share, consistent with FHWA estimate of
0% EtOH.
38. OR
• All of the counties were filled by data from TRW table 9 with 100% MTBE market share.
• Except Clackamas, Columbia, Jackson, Josephine, Klamath, Multnomah, Washington, and
Yamhill counties using TRW table 10 with 100% ETOH market share in winter season, as per
Oxy Fuel Program descriptions.
• Assumes 100% EtOH fraction (for the excepted counties) accounts for statewide 7.3% fraction
from FHWA.
39. PA
• Counties in Philadelphia area were filled by RFG survey (Phila.-Wilm, DE-Trenton, NJ) with
RFG Survey market share.
• Counties in Pittsburgh area were filled by AAM survey (Pittsburgh, PA) with 100% MTBE
market share.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share,
since there were no alcohol containing samples in the NIPER District 1 surveys.
• This contradicts the FHWA estimate of 2.11% EtOH.
0136.02.004.003/MOBILE6.wpd
A-6
-------
Mapping Assignments by Analysis Year - 1999
40. RI
• All counties were filled by RFG survey (Rhode Island) with RFG market share as well.
41. SC was filled by data from TRW table 9 with 100% MTBE market share. Consistent with FHWA
estimate of 0% EtOH.
42. SD
• All counties were filled by Duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
43. TN was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA estimate of
0% EtOH.
44. TX
• Counties in Dallas-Fort Worth area were filled by RFG survey (Dallas-Forth Worth, TX) and
RFG survey for ETOH and MTBE market share.
• Counties in Houston-Galveston area were filled by RFG survey (Houston-Galveston, TX) and
RFG survey for ETOH and MTBE market share.
• Counties in San Antonio Metropolitan area were filled by AAM survey (San Antonio, TX)
with 100%) MTBE, based on low measured EtOH, ETBE, and TAME levels (-0.1).
• Counties in the eastern part of the state were filled by data from TRW District 8 table 9 with
100%) MTBE market share (since District 8 had no survey information for fuels with alcohols).
• Counties in the western part of the state was filled by data from TRW District 11 table 10 with
100% ETOH market share in winter (since measured MTBE levels in winter non-alcohol fuels
- Table 9 - were 0 in winter), and table 9 with 100% MTBE market share in summer (since
District 11 had no survey for fuels with alcohols in the summer). The winter EtOH fuels in the
Western counties in the wintertime may help account for the FHWA estimate of 4.95% EtOH.
45. UT
• All of the counties were filled by data from TRW table 9 with 100% MTBE market share,
except Utah and Weber Counties were filled by TRW table 10 with 100% ETOH market
share in winter season, as per Oxy Fuel Program description.
• Note: FHWA data indicate a 10.7% EtOH sales fraction for the state. We recommend
duplication of all counties in the state for the 1999 calendar year in the future.
46. VA
• Counties in Washington DC area were filled by RFG survey (Washington, DC) both fuel
parameters and market share.
• Counties in Richmond and Norfolk RFG were filled by RFG survey (Richmond, VA) both fuel
parameters and market share.
0136.02.004.003/MOBILE6.wpd
A-7
-------
Mapping Assignments by Analysis Year - 1999
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share.
• For FHWA survey in VA reported 8.61% gasohol, but there is no TRW survey on gasoline
contain alcohol for this area.
47. VT was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA estimate of
0% EtOH.
48. WA
• Counties in Seattle metropolitan area were filled by AAM survey (Seattle, WA) with 100%
ETOH market share in winter (based on 0.1 measured MTBE levels), and 100% MTBE
market share in summer.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share.
• Except Clark and Spokane counties were filled by data from TRW table 10 with %100 ETOH
market share in winter as per Oxy Fuel Program description.
• Assumptions probably over predict EtOH fraction compared to FHWA estimate of 9.93%.
49. WI
• Counties in Milwaukee-Racine RFG were filled by RFG survey for both fuel parameters and
market share.
• Assumes 100% EtOH fraction (from RFG survey) accounts for statewide 10.98%) fraction
from FHWA.
• The rest of the state was filled by TRW table 9 with 100% MTBE.
50. WV was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA estimate of
0.01% EtOH.
51. WY was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA estimate of
0% EtOH.
0136.02.004.003/MOBILE6.wpd
A-8
-------
Appendix B
Mapping Assignments by Analysis Year ~ 1996
-------
Mapping Assignments by Analysis Year - 1996
1. AL — 2.0% Ethanol sales market share as per FHWA, but no alcohol samples (table 10) for NIPER
District 3. Therefore entire state was filled by data from TRW table 9 with 100% MTBE market share.
2. AK was filled by data from data from AAM survey (Fairbanks, AK) with FHWA market share
between MTBE and ETOH.
3. AZ
• Phoenix area was filled by data from data from AAM survey (Phoenix, AZ) with EtOH market
share for winter = 100% (as per EPA Oxy Fuel Program Summary, October 2001), and
100%) MTBE for summer. Compares with statewide annual average of—16% from FHWA.
• The rest of AZ was filled by data from TRW table 9 with 100%> MTBE market share.
4. AR was filled by data from TRW table 9 with 100% MTBE market share (only 0.2% EtOH from
FHWA).
5. CA
• San Francisco Bay area was filled by data from AAM survey (San Francisco, CA) with RFG
market share between MTBE and ETOH.
• The rest of CA was filled by data from Los Angeles AAM survey, with 85/15 MTBE/EtOH
market share from Oxy Fuel Program Description (EPA, Oct 2001). If maintained year round
as assumed, compares well with FHWA annual average of 12.6% EtOH statewide.
6. CO
• Denver area was filled by data from AAM survey (Denver, CO) with 100 % ETOH market
share in winter, as per Oxy Fuel program description. There is a discrepancy here with the
AAM survey data which shows non-trivial levels of BOTH EtOH (8.461) and MTBE (1.5173)
in winter fuel.
• The rest of CO was filled by data from TRW table 9 with 100% MTBE market share in
summer and TRW table 10 with 100% ETOH market share in winter to compare well with
45%) EtOH annual statewide average from FHWA. Slight discrepancy with AAM Denver
summer survey showing EtOH levels at 1.44.
7. CT All counties duplicated with 100%> MTBE market share (for table 11) and 100%> EtOH market
share (for table 12). EPA RFG survey was used to assign percent between MTBE and EtOH in the
Oxygenate Fuel Sale Percentage column. Except Fairfield County, which was assigned to the NY-
NJ-Long Island AAM survey, with market share from the EPA RFG Survey. Only a slight discrepancy
with FHWA number of 2.7% EtOH.
8. DE
• All of the counties except Sussex were filled by data from AAM survey (Philadelphia, PA) with
RFG Survey market share between MTBE and ETOH. Slight contradiction with FHWA value
of 0%> for the state.
• Sussex County was filled by data from TRW table 11 with 100% MTBE market share.
9. DC was filled by AAM survey (Washington, DC) with RFG Survey market share between MTBE and
ETOH. Slight discrepancy with FHWA value of 0%.
0136.02.004.003/MOBILE6.wpd
B-l
-------
Mapping Assignments by Analysis Year - 1996
10. FL
• Miami area was filled by data from AAM survey (Miami, FL) with 100% MTBE market share.
• The rest of FL was filled by data from TRW table 9 with 100% MTBE market share.
• Essentially same as FHWA value of 0.13% EtOH for the state.
11. GA
• Atlanta area was filled by data from AAM survey (Atlanta, GA) with 100% MTBE market
share.
• The rest of GA was filled by data from TRW table 9 with 100% MTBE market share.
• Consistent with FHWA value of 0%.
12. HI was filled by TRW table 9 (district 14 Northern California) with 100% MTBE market share.
13. IA
• Duplicate each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
14. ID was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA of 0.
15. IL
• Chicago area was filled by data from AAM survey (Chicago, IL) using RFG Survey market
share.
• Counties in St Louis were filled by data from AAM survey (St. Louis, MO) with 100% MTBE
market share, based on very low EtOH measured levels.
• The rest of the counties were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
• Assumes St. Louis at 0% and Chicago at —95% EtOH "cancel out" with rest of the state at
FHWA level of 30% to meet statewide average.
16. IN
• Counties in Chicago area were filled by data from AAM survey (Chicago, IL) with RFG
Survey market share.
• The rest of the counties were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
0136.02.004.003/MOBILE6.wpd
B-2
-------
Mapping Assignments by Analysis Year - 1996
17. KS
• Duplicate each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
18. KY
• Counties in Cincinnati RFG area were filled by data from TRW table 11 with 100% MTBE
market share.
• Counties in Louisville RFG area were duplicated each county with TRW tablel 1 and table 12.
• If the data came from table 11 then MTBE market share is equal to 100%.
• If the data came from table 12 then ETOH market share is equal to 100%.
• Uses RFG Survey market share for Louisville area (-75/25) to assign percent of sale between
MTBE and ETOH in the Oxygenate Fuel Sale Percentage column.
• The rest of the state was filled by TRW table 9 with 100% MTBE market share.
• May slightly overestimate EtOH when compared with FHWA statewide market share of 3.3%
EtOH.
19. LA
• New Orleans area was filled by data from AAM survey (New Orleans, LA) with 100% MTBE
market share.
• The rest of LA was filled by data from TRW table 9 with 100% MTBE market share.
• Essentially same as FHWA value of 0.94% EtOH for the state.
20. MA
• Counties in Boston area were filled by AAM survey (Boston, MA) with RFG Survey market
share between MTBE and ETOH.
• Counties in Springfield area were duplicated each county with TRW table 11 and table 12.
• If the data came from table 11 then MTBE market share is equal to 100%.
• If the data came from table 12 then ETOH market share is equal to 100%.
• Uses RFG Springfield survey to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
21. MD
• Counties in Philadelphia area were filled by AAM survey (Philadelphia, PA) with RFG Survey
market share between MTBE and ETOH.
• Counties in Washington DC area were filled by AAM survey (Washington DC) with RFG
Survey market share between MTBE and ETOH.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE.
22. ME
• ME Counties in RFG area were filled by TRW table 11 with 100% MTBE market share.
0136.02.004.003/MOBILE6.wpd
B-3
-------
Mapping Assignments by Analysis Year - 1996
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share.
Consistent with FHWA value of 0% EtOH.
23. MI
• Counties in Detroit area were filled by AAM survey (Detroit, MI) with 100% ETOH market
share in winter and 100% MTBE market share in summer.
• The rest of the state were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%
• If the data came from table 10 then ETOH market share is equal to 100%
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
24. MN
• Counties in Minneapolis/St. Paul Detroit area were filled by AAM survey (Minneapolis/St.
Paul, MN) with 100% ETOH market share both winter and summer, based on low (0.1)
measured MTBE values.
• The rest of the state were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
25. MO
• Counties in St. Louis and Kansas City areas were filled by AAM survey (St. Louis, MO and
Kansas City, MO) and duplicated to use FHWA sales fractions.
• The rest of the state were duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
26. MS was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA value of 0%.
27. MT
• Yellow Stone County was filled by data from AAM survey (Billings, MT) with 100%) MTBE
market share in summer and winter.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE share.
• Except Missoula county that was filled by data from TRW table 10 with 100% ETOH market
share in the Winter as per EPA's Oxy Fuel Program Summary.
28. NC was filled by TRW table 9 with 100% MTBE market share. Note that FHWA has reported
9.84%) of gasohol was used in this state, TRW survey did not have any survey collect on gasoline
containing alcohol in this area.
0136.02.004.003/MOBILE6.wpd B-4
-------
Mapping Assignments by Analysis Year - 1996
29. ND
• Duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
30. NE
• Duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
31. NH
• Counties located in Boston-Worcester metropolitan area were filled by AAM survey (Boston,
MA) with RFG Survey market share between MTBE and ETOH. Note this is inconsistent
with the FHWA estimate of 0% EtOH.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE.
32. NJ
• Counties in NY-NJ-Long Island-CT RFG area were filled by AAM survey (New York, NY)
with RFG Survey market share between MTBE and ETOH.
• Counties in Philadelphia area were filled by AAM survey (Philadelphia, PA) with RFG Survey
market share between MTBE and ETOH.
• Counties in Atlantic City area were filled by data from TRW table 11 with RFG Survey market
share.
• Somewhat underestimates FHWA estimate of 3.9% EtOH.
33. NM
• Counties in Albuquerque area were filled by AAM survey (Albuquerque, NM) with 100%
ETOH market share in the winter, as per Oxy Fuel Program description. 100% EtOH share
was assumed during the summer as well due to the low measured levels of MTBE (0.1) vs.
EtOH (0.8).
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share in
summer and TRW table 10 with 100%) ETOH market share in winter, as there were no data
for summer alcohol fuels in NIPER District 11, and because winter MTBE levels were
measured as 0 in Table 9 here as well.
34. NV
• Counties in Las Vegas area were filled by AAM survey (Las Vegas, NV) with 100% ETOH
market share in the winter as per the Oxy Fuel Program description. Summer market share set
to 100% MTBE to be more consistent with FHWA estimate of 0%.
0136.02.004.003/MOBILE6.wpd
B-5
-------
Mapping Assignments by Analysis Year - 1996
• Washoe, Storey, Carson City, Douglas, Lyon counties were filled by data from TRW table 10
in winter with 100% ETOH market share, as per Oxy Fuel program description.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share to
be more consistent with FHWA estimate of 0% EtOH.
35. NY
• Counties in NY-NJ-Long Island-CT area were filled by AAM survey (New York, NY) with
the RFG Survey market share.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share
(since there was no Table 10 in the NIPER data set).
36. OH
• Counties in Cleveland area were filled by AAM survey (Cleveland, OH) with 100% ETOH
market share, based on low measured values of MTBE (-0.1).
• The rest of the state was duplicated each county with TRW table 9 and table 10, for summer
only (no Table 10 for winter in District 6).
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
37. OK was filled by TRW table 9 with 100% MTBE market share, consistent with FHWA estimate of
0% EtOH.
38. OR
• All of the counties were filled by data from TRW table 9 with 100% MTBE market share
except Clackamas, Columbia, Jackson, Josephine, Klamath, Multnomah, Washington, and
Yamhill counties using TRW table 10 with 100% ETOH market share in winter season, as per
Oxy Fuel Program descriptions.
• Not consistent with FHWA estimate of 0% EtOH.
39. PA
• Counties in Philadelphia area were filled by AAM survey (Philadelphia, PA) with RFG Survey
market share.
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share,
since there were no alcohol containing samples in the NIPER District 1 surveys.
• This contradicts the FHWA estimate of 13.1% EtOH.
40. RI
• Duplicated each county with TRW table 11 and table 12.
• If the data came from table 11 then MTBE market share is equal to 100%
• If the data came from table 12 then ETOH market share is equal to 100%
0136.02.004.003/MOBILE6.wpd
B-6
-------
Mapping Assignments by Analysis Year - 1996
• Uses RFG Survey market share for RI to assign percent of sale between MTBE and ETOH in
the OxygenateFuelSalePercentage column. This is slightly inconsistent with FHWA
estimate of 0% EtOH.
41. SC was filled by data from TRW table 9 with 100% MTBE market share. Consistent with FHWA
estimate of 0% EtOH.
42. SD
• Duplicated each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%.
• If the data came from table 10 then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
43. TN was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA estimate of
only 0.1% EtOH.
44. TX
• Counties in Dallas and San Antonio Metropolitan area were filled by AAM survey (Dallas, TX
and San Antonio, TX) with RFG survey for the Dallas area ETOH and MTBE market share,
and 100%) MTBE for the San Antonio area market share, based on low measured EtOH,
ETBE, and TAME levels (-0.1).
• Counties in Houston Metropolitan area were filled by data from TRW table 11 with RFG
survey for their ETOH and MTBE market share.
• Counties in the eastern part of the state were filled by data from TRW District 8 table 9 with
100%) MTBE market share (since District 8 had no survey information for fuels with alcohols).
• Counties in the western part of the state was filled by data from TRW District 11 table 10 with
100% ETOH market share in winter (since measured MTBE levels in winter non-alcohol fuels
- Table 9 - were 0 in winter), and table 9 with 100% MTBE market share in summer (since
District 11 had no survey for fuels with alcohols in the summer). The winter EtOH fuels in the
Western counties in the wintertime may help account for the FHWA estimate of 2% EtOH.
45. UT
• All of the counties were filled by data from TRW table 9 with 100%> MTBE market share,
except Utah and Weber Counties were filled by TRW table 10 with 100%> ETOH market
share in winter season, as per Oxy Fuel Program description. This may account for the 1.1%
EtOH fraction from the FHWA.
46. VA
• Counties in Washington DC area were filled by AAM survey (Washington, DC) with RFG
survey for market share.
• Counties in Richmond and Norfolk RFG were filled by TRW survey table 11 with 100%)
MTBE (same as RFG survey) for market share.
• The rest of the state was filled by data from TRW table 9 with 100%> MTBE market share.
0136.02.004.003/MOBILE6.wpd
B-7
-------
Mapping Assignments by Analysis Year - 1996
• For FHWA survey in VA reported 13.81% gasohol, but there is no TRW survey on gasoline
contain alcohol for this area.
47. VT was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA estimate of
0% EtOH.
48. WA
• Counties in Seattle metropolitan area were filled by AAM survey (Seattle, WA) with 100%
ETOH market share in winter (based on 0.1 measured MTBE levels), and 50% MTBE-
50%ETOH market share in summer (based on non-trivial (> 0.1) measured values of both
oxygenates).
• The rest of the state was filled by data from TRW table 9 with 100% MTBE market share.
• Except Clark and Spokane counties were filled by data from TRW table 10 with %100 ETOH
market share in winter as per Oxy Fuel Program description.
• Assumptions probably over predict EtOH fraction compared to FHWA estimate of 6.6%.
49. WI
• Counties in Milwaukee-Racine RFG area duplicate each county with TRW tables 11, 12
• If the data came from table 11 then MTBE market share is equal to 100%.
• If the data came from table 12 then ETOH market share is equal to 100%.
• Uses RFG survey to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
• Assumes 96% EtOH fraction (from RFG survey) accounts for statewide 26% fraction from
FHWA.
• The rest of the state was filled by TRW table 9 with 100% MTBE.
50. WV was filled by TRW table 9 with 100% MTBE market share. Consistent with FHWA estimate of
0.29% EtOH.
51. WY
• Duplicate each county with TRW table 9 and table 10.
• If the data came from table 9 then MTBE market share is equal to 100%
• If the data came from table 10 then ETOH market share is equal to 100% (Table 10 data
available only winter season, therefore 100% MTBE in summer is assumed).
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
0136.02.004.003/MOBILE6.wpd
B-8
-------
Appendix C
Mapping Assignments by Analysis Year ~ 1990
-------
Mapping Assignments by Analysis Year - 1990
1. AL - 9.64% Ethanol sales market share as per FHWA.
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
2. AK was filled by TRW table 3 (district 13 West Washington) with 100% MTBE market share in
winter. FHWA survey has reported zero percent Ethanol sale market share as well.
3. AZ
• Phoenix area was filled by data from data from AAMA survey (Phoenix, AZ) with FHWA
market for MTBE and EtOH. FHWA survey has reported zero percent Ethanol sale market
share for AZ.
• The rest of AZ was filled by data from TRW table 3 with 100% MTBE market share.
4. AR -5.15%) Ethanol sales market share as per FHWA.
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
5. CA
• San Francisco Bay area was filled by data from AAMA survey (San Francisco, CA) with
FHWA market share between MTBE and ETOH.
• The rest of CA was filled by data from Los Angeles AAMA survey, with FHWA market share
between MTBE and EtOH.
6. CO
• Denver area was filled by data from AAMA survey (Denver, CO) with FHWA market share
between MTBE and EtOH.
• The rest of CO was filled by data from TRW table 3 with 100% MTBE market share in winter.
There is no TRW survey on gasoline containing alcohol in this district (10). This may
underestimate EtOH market share since FHWA survey has reported 6.53% EtOH annual
statewide average.
7. CT was filled by data from TRW table 3 with 100% MTBE market share, with the exception of
Fairfield, which was assigned to the NY-NJ-Long Island AAMA survey, with market share from
FHWA Survey. This market share are consistent with FHWA number of 0% EtOH.
8. DE
• All of the counties except Sussex were filled by data from AAMA survey (Philadelphia, PA)
with FHWA value of 0% for the state.
0136.02.004.003/MOBILE6.wpd
C-l
-------
Mapping Assignments by Analysis Year - 1990
• Sussex County was filled by data from TRW table 3 with 100% MTBE market share.
9. DC was filled by AAMA survey (Washington, DC) with FHWA value of 0% EtOH.
10. FL
• Miami area was filled by data from AAMA survey (Miami, FL) with 100% MTBE market
share.
• The rest of FL was filled by data from TRW table 3 with 100% MTBE market share.
• Slightly discrepancy with FHWA number of 1.33% EtOH, since TRW did not report any
survey on gasoline containing alcohol in Florida (district 4).
11. GA
• Atlanta area was filled by data from AAMA survey (Atlanta, GA) with FHWA market share
• The rest of GA was duplicated by data from TRW table 3 and table 3 A with 100% MTBE
market share and 100% EtOH market share respectively.
• Consistent with FHWA value of 2.51%.
12. HI was filled by TRW table 3 (district 14 Northern California) with 100%) MTBE market share.
13. IA
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column (28.92% EtOH sale fraction statewide).
14. ID
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%
• If the data came from table 3 A then ETOH market share is equal to 100%
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column (15.1% EtOH sale fraction statewide).
15. IL
• Chicago area was filled by data from AAMA survey (Chicago, IL) using FHWA Survey
market share
• Counties in St Louis were filled by data from AAMA survey (St. Louis, MO) with FHWA
survey market share.
• The rest of the counties were duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
0136.02.004.003/MOBILE6.wpd
C-2
-------
Mapping Assignments by Analysis Year - 1990
16. IN
• Counties in Chicago area were filled by data from AAMA survey (Chicago, IL) with FHWA
Survey market share.
• The rest of the counties were duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
17. KS
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
18. KY
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
19. LA
• New Orleans area was filled by data from AAMA survey (New Oreleans, LA) with FHWA
survey market share.
• The rest of LA was duplicated by data from TRW table 3 and table 3Awith 100% MTBE
market share and 100% EtOH market share respectively.
• Essentially same as FHWA value of 2.16% EtOH for the state.
20. MA
• Counties in Boston area were filled by AAMA survey (Boston, MA) with FHWA Survey
market share between MTBE and ETOH.
• The rest of MA was filled by data from TRW table 3 with 100% MTBE market share.
• Essentially same as FHWA value of 0% EtOH for the state.
21. MD
• Counties in Philadelphia area were filled by AAMA survey (Philadelphia, PA) with FHWA
Survey market share between MTBE and ETOH.
• Counties in Washington DC area were filled by AAMA survey (Washington DC) with FHWA
Survey market share between MTBE and ETOH.
• The rest of the state was filled by data from TRW table 3 with 100% MTBE.
22. ME
0136.02.004.003/MOBILE6.wpd
C-3
-------
Mapping Assignments by Analysis Year - 1990
• ME Counties in RFG area were filled by TRW table 3 with 100% MTBE market share.
• The rest of the state was filled by data from TRW table 3 with 100% MTBE market share.
Consistent with FHWA value of 0% EtOH.
23. MI
• Counties in Detroit area were filled by AAMA survey (Detroit, MI) with FHWA survey market
share.
• The rest of the state were duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
24. MN
• Counties in Minneapolis/St. Paul Detroit area were filled by AAMA survey (Minneapolis/St.
Paul, MN) with FHWA survey market share.
• The rest of the state were duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
25. MO
• Counties in St. Louis and Kansas City areas were filled by AAMA survey (St. Louis, MO and
Kansas City, MO) and used FHWA sales fractions as market share.
• The rest of the state were duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
26. MS was filled by TRW table 3 with 100% MTBE market share. Consistent with FHWA value of 0%.
27. MT
• Yellow Stone County was filled by data from AAMA survey (Billings, MT) with 50-50 split
market share between MTBE and EtOH.
• The rest of the state was filled by data from TRW table 3 with 100% MTBE share.
28. NC was filled by TRW table 3 with 100% MTBE market share.
29. ND
• Duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
0136.02.004.003/MOBILE6.wpd
C-4
-------
Mapping Assignments by Analysis Year - 1990
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
30. NE
• Duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
31. NH
• Counties located in Boston-Worcester metropolitan area were filled by AAMA survey
(Boston, MA) with FHWA Survey market share between MTBE and ETOH.
• The rest of the state was filled by data from TRW table 3 with 100% MTBE.
32. NJ
• Counties in NY-NJ-Long Island-CT RFG area were filled by AAMA survey (New York,
NY) with FHWA Survey market share between MTBE and ETOH.
• Counties in Philadelphia area were filled by AAMA survey (Philadelphia, PA) with FHWA
Survey market share between MTBE and ETOH.
• Counties in Atlantic City area were filled by data from TRW table 3 with FHWA Survey
market share.
33. NM
• Counties in Albuquerque area were filled by AAMA survey (Albuquerque, NM) with FHWA
market share.
• The rest of the state was filled by data from TRW table 3 in summer and TRW table 3 A with
100%) ETOH market share in winter, as there were no data for summer alcohol fuels in TRW
District 11, and because winter MTBE levels were measured as 0 in Table 3 here as well.
34. NV
• Counties in Las Vegas area were filled by AAMA survey (Las Vegas, NV) with FHWA
survey market share.
• The northern part of the state was filled by data from TRW table 3 while the southern part of
the state was filled by TRW table3 A. to be more consistent with FHWA estimate of 7.84%>
EtOH.
35. NY
• Counties in NY-NJ-Long Island-CT area were filled by AAMA survey (New York, NY) with
the FHWA Survey market share.
• The rest of the state was filled by data from TRW table 3 with 100%> MTBE market share.
36. OH
0136.02.004.003/MOBILE6.wpd
C-5
-------
Mapping Assignments by Analysis Year - 1990
• Counties in Cleveland area were filled by AAMA survey (Cleveland, OH) with FHWA market
share.
• The rest of the state was duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%
• If the data came from table 3 A then ETOH market share is equal to 100%
• FHWA survey was used survey to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
37. OK was filled by TRW table 3 with 100% MTBE market share, consistent with FHWA estimate of
0% EtOH.
38. OR
• All of the counties were filled by data from TRW table 3 with 100% MTBE market share.
• Consistent with FHWA estimate of 0% EtOH.
39. PA
• Counties in Philadelphia area were filled by AAMA survey (Philadelphia, PA) with FHWA
Survey market share.
• The rest of the state was filled by data from TRW table 3 with 100% MTBE market share.
40. RI was filled by TRW table 3 with 100% MTBE market share, consistent with FHWA estimate of 0%
EtOH.
41. SC - 3.54% Ethanol sales market share as per FHWA.
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
42. SD
• Duplicated each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
43. TN - 10.20%) Ethanol sales market share as per FHWA.
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
0136.02.004.003/MOBILE6.wpd C-6
-------
Mapping Assignments by Analysis Year - 1990
44. TX
• Counties in Dallas and San Antonio Metropolitan area were filled by AAMA survey (Dallas,
TX and San Antonio, TX) with FHWA survey for the Dallas area ETOH and MTBE market
share.
• The rest of the counties were duplicated with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
45. UT
• All of the counties were filled by data from TRW table 3 with 100% MTBE market share
(since there was no Table 3 A in the NIPER data set). This may under estimate for the 0.01%
EtOH fraction from the FHWA.
46. VA
• Counties in Washington DC area were filled by AAMA survey (Washington, DC) with RFG
survey for market share.
• The rest of the state was filled by data from TRW table 3 with 100% MTBE market share.
• For FHWA survey in VA reported 5.49% gasohol, but there is no TRW survey on gasoline
contains alcohol for this area.
47. VT was filled by TRW table 3 with 100% MTBE market share. Consistent with FHWA estimate of
0% EtOH.
48. WA
• Counties in Seattle metropolitan area were filled by AAMA survey (Seattle, WA) with 100%
ETOH market share in winter (based on 0.1 measured MTBE levels), and 50% MTBE-
50%ETOH market share in summer (based on non-trivial (> 0.1) measured values of both
oxygenates).
• The rest of the state was filled by data from TRW table 3 with 100% MTBE market share.
• Assumptions probably underestimate EtOH fraction compared to FHWA estimate of 3.89%
since there is no TRW survey for gasoline contains alcohol in district 13 (west Washington and
west Oregon).
49. WI
• All counties were duplicated with TRW tables 3, 3A
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
Oxygenate Fuel Sale Percentage column.
50. WV was filled by TRW table 3 with 100% MTBE market share. Consistent with FHWA estimate of
0% EtOH.
0136.02.004.003/MOBILE6.wpd
C-7
-------
Mapping Assignments by Analysis Year - 1990
51. WY
• Duplicate each county with TRW table 3 and table 3 A.
• If the data came from table 3 then MTBE market share is equal to 100%.
• If the data came from table 3 A then ETOH market share is equal to 100%.
• FHWA survey was used to assign percent of sale between MTBE and ETOH in the
OxygenateFuelSalePercentage column.
0136.02.004.003/MOBILE6.wpd
C-8
-------
Attachment B
EPANTI
1999 On-Road Hazard Air Pollution Emissions
Estimation Calculation Methodology
MOBILE6.2 HAP Input File Creation
-------
I. General Methodology
The U.S. Environmental Protection Agency (EPA) has recently released MOBILE6.2, a modeling tool used to
estimate emission factors for on-road mobile sources. While this program has historically concentrated on
criteria emissions, the newest release also includes factors for six hazardous air pollutants (HAPs) and allows a
user to input factors for up to 50 different HAPs. The next phase of the U.S. EPA's National Toxic Inventory
will take advantage of this added functionality to allow for the calculation of toxic pollutants at the same time as
the criteria pollutants for on-road mobile sources.
Factors for 29 different HAPs have been taken from a variety of sources and formatted to match the
MOBILE6.2 HAP speciation input file format. Only minor conversions were performed on the factors as
needed in order to function properly with MOBILE6.2. The six HAPs built into MOBILE6.2 are shown in
Table 1-1. The 29 additional HAPs included in the speciation files are shown in Table 1-2. All of the factors
have been collected for each of the 28 different vehicle types used in MOBILE6.2 as shown in Table 1-3.
All factors that are ratios of either VOC or PM have been entered into the MOBILE6.2 HAP input file in units
of milligram of HAP over gram of criteria pollutant.
Table 1-1: HAPs built into MOBILE6.2
MOBILE6.2
Pollutant
Number
Pollutants
Contaminate Code
16
Benzene
71432
17
MTBE (Methyl Tert-Butyl Ether)
1634044
18
1,3-Butadiene
106990
19
Formaldehyde
50000
20
Acetaldehyde
75070
21
Acrolein
107028
-------
Table 1-2: Additional HAPs Specified
MOBILE6.2
Number
Contaminant
Code
Pollutant Label
Pollutants Name
Factor Type
60
83329
Acenaphthene
Acenaphthene
RATIOPM
61
208968
Acenaphthylene
Acenaphthylene
RATIOPM
62
120127
Anthracene
Anthracene
RATIOPM
63
56553
Benzo(a)anthracene
Benz[a]Anthracene
RATIOPM
65
205992
Benzo(b)fluoranthene
Benzo[b]Fluoranthene
RATIOPM
64
50328
Benzo(a)pyrene
Benzo[a]Pyrene
RATIOPM
66
191242
Benzo(ghi)perylene
Benzo[g,h,i,]Perylene
RATIOPM
67
207089
Benzo(k)fluoranthene
Benzo[k]Fluoranthene
RATIOPM
68
218019
Chrysene
Chrysene
RATIOPM
69
53703
Dibenz(ah)anthracene
Dibenzo[a,h]Anthracene
RATIOPM
70
206440
Fluoranthene
Fluoranthene
RATIOPM
71
86737
Fluorene
Fluorene
RATIOPM
72
193395
lndeno(123cd)pyrene
lndeno[1,2,3-c,d]Pyrene
RATIOPM
74
85018
Phenanthrene
Phenanthrene
RATIOPM
75
129000
Pyrene
Pyrene
RATIOPM
73
91203
Napthalene
Naphthalene
RATIOPM and RATIOVOC
76
100414
Ethylbenzene
Ethyl Benzene
RATIOVOC
77
110543
n-Hexane
n-Hexane
RATIOVOC
78
100425
Styrene
Styrene
RATIOVOC
79
108883
Toluene
Toluene
RATIOVOC
80
1330207
Xylene
Xylenes (mixture of o, m, and p isomers)
RATIOVOC
81
18540299
Chromim (Cr6+)
Chromium (CR6+)
BEF
82
16065831
Chromim (Cr3+)
Chromium (CR3+)
BEF
83
198
Manganese
Manganese & Compounds
BEF
84
226
Nickel
Nickel & Compounds
BEF
85
199
Mercury
Mercury & Compounds
BEF
86
93
Arsenic
Arsenic & Compounds (inorganic including arsine)
BEF
87
540841
224T rimethylpentane
2,2,4-T rimethylpentane
RATIOVOC
88
123386
Propionaldehyde
Propionaldehyde
RATIOVOC
-------
Table 1-3 Vehicle Types
MOBILE6.2
Vehicle
Types
AMS 7-digit
Current 7-digit DESC
Revised 7-digit DESC
LDGV
22-01-001
Light Duty Gasoline Vehicles (LDGV)
(same)
LDGT1
22-01-020
Light Duty Gasoline Trucks 1 (LDGT1)
Light Duty Gasoline Trucks 1&2(M6) = LDGT1(M5)
LDGT2
22-01-020
Light Duty Gasoline Trucks 1 (LDGT 1)
Light Duty Gasoline Trucks 1&2(M6) = LDGT1(M5)
LDGT3
22-01-040
Light Duty Gasoline Trucks 2 (LDGT2)
Light Duty Gasoline Trucks 3&4(M6) = LDGT2(M5)
LDGT4
22-01-040
Light Duty Gasoline Trucks 2 (LDGT2)
Light Duty Gasoline Trucks 3&4(M6) = LDGT2(M5)
HDGB
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV2B
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV3
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV4
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV5
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV6
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV7
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV8A
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
HDGV8B
22-01-070
Heavy Duty Gasoline Vehicles (HDGV)
Heavy Duty Gasoline Vehicles 2B-8B & Buses (HDGV)
MC
22-01-080
Motorcycles (MC)
(same)
LDDV
22-30-001
Light Duty Diesel Vehicles (LDDV)
(same)
LDDT12
22-30-060
Light Duty Diesel Trucks (LDDT)
Light Duty Diesel Trucks 1-4 (M6) (LDDT)
LDDT34
22-30-060
Light Duty Diesel Trucks (LDDT)
Light Duty Diesel Trucks 1-4 (M6) (LDDT)
HDDV2B
22-30-071
2B Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 2B
HDDV3
22-30-072
Light Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 3,4,&5
HDDV4
22-30-072
Light Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 3,4,&5
HDDV5
22-30-072
Light Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 3,4,&5
HDDV6
22-30-073
Medium Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 6&7
HDDV7
22-30-073
Medium Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 6&7
HDDV8A
22-30-074
Heavy Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 8A&8B
HDDV8B
22-30-074
Heavy Heavy Duty Diesel Vehicles
Heavy Duty Diesel Vehicles (HDDV) Class 8A&8B
HDDBS
22-30-075
Buses Heavy Duty Diesel Vehicles
Heavy Duty Diesel Buses (School & Transit)
HDDBT
22-30-075
Buses Heavv Dutv Diesel Vehicles
Heavv Dutv Diesel Buses (School & Transit)
-------
II. Polycyclic Aromatic Hydrocarbons (PAH)
Factors for PAH emissions were provided by the U.S. Environmental Protection Agency's (EPA) Office of
Transportation and Air Quality (OTAQ). The EPA recommended that the factors given as a fraction of PM10
be used instead of the VMT based factors (Cook, 2001).
PAH Exhaust Factors
Gasoline Vehicles
LDD
HDDV
Contam
Code
PAH
mg/mi
Fraction of
PM10
mg/mi
Fraction of
PM10
Fraction of
PM2.5
56553
Benzo(a)anthracene
0.008
0.00010
0.027
0.000048
0.000040
192972
Benzo(a)pyrene
0.008
0.00010
0.025
0.000045
0.000013
205992
Benzo(b)fluoranthene
0.010
0.00012
0.044
0.000078
0.000011
207089
Benzo(k)fluoranthene
0.010
0.00012
0.044
0.000078
0.000011
218019
Chrysene
0.008
0.00010
0.032
0.000057
0.000007
53703
Dibenz(a,h)anthracene
0.000
0.00000
0.001
0.000002
0.000000
193395
lndeno(1,2,3-cd)pyrene
0.006
0.00008
0.012
0.000021
0.000001
83329
Acenaphthene
0.057
0.00073
0.048
0.000086
0.000024
208968
Acenaphthalene
0.321
0.00412
0.545
0.000971
0.000037
120127
Anthracene
0.066
0.00085
0.102
0.000182
0.000037
191242
Benzo(ghi)perylene
0.020
0.00026
0.030
0.000053
0.000009
206440
Fluoranthene
0.071
0.00091
0.301
0.000536
0.000022
86737
Fluorene
0.118
0.00151
0.214
0.000381
0.000049
91203
Napthalene
7.074
0.09073
2.056
0.003663
0.001401
85018
Phenanthrene
0.198
0.00254
0.594
0.001058
0.000056
129000
Pvrene
0.097
0.00124
0.387
0.000689
0.000039
MOBILE6.2 is only able to calculate one cut off size for particulate matter per run. Due to this, the factors for
each of the above HAPs were converted to be a fraction of PM 10. The conversion factors were taken from
Table 3.4 of the MOBILE6.1 technical documentation (Cook, 2001 and EPA 2002). The above factors were
converted to be a ratio of milligrams of HAP to grams of PM when they were entered into the HAP input files.
The only PAH with substantial evaporative emissions is Napthalene from gasoline vehicles only. Factors for
evaporative emissions of Napthalene were taken from the SPECIATE program from EPA (EPA, 2001). The
data in SPECIATE is organized under a multitude of vehicle profiles. Guidance on which profiles to use where
given by Rich Cook of the EPA (Cook, 2001).
Napthalene Evaporative Emissions From SPECIATE
Profile #
Profile Description
% voc
1301
10% Ethanol Composite (Hot Soak + Diurnal) Evaporative
0.02
1305
Industry Average (circa 1990) Gasoline Composite (Hot Soak + Diurnal)
Evaporative
0.04
1309
11 % MTBE Composite (Hot Soak + Diurnal) Evaporative
0.06
Factor Used
Fraction of VOC
mg Nap /gm VOC
Gasoline Evaporative
0.0004
0.4
-------
References:
Cook, Rich. Memorandum entitled Revised Methodology and Emission Factors for Estimating
Mobile Source PAH Emissions in the National Toxics Inventory, Laurel Driver, Office of Air Quality
Planning and Standards. U.S. EPA Office of Transportation and Air Quality (OTAQ). Ann Arbor,
Ml. June, 8, 2001.
Cook, Rich. Personal Communications. U.S. Environmental Protection Agency Office of
Transportation and Air Quality (OTAQ). Ann Arbor, MI. Summer 2001.
U.S. Environmental Protection Agency. SPECIATE, Version 3.1. Office of Air Quality Planning and
Standards. Research Triangle Park, NC. 2001.
Cook, Rich. Personal Communications. U.S. Environmental Protection Agency Office of Transportation and
Air Quality (OTAQ) . Ann Arbor, MI. Summer 2002.
U.S. Environmental Protection Agency. MOBILE6.1 Particulate Emission Factor Model Technical
Description, DRAFT. EPA420-R-02-012, March 2002.
-------
III. Acrolein, Ethylbenzene, n-Hexane, Propionaldehyde, Styrene, Toluene, and
Xylene Emissions
The emission factors for these chemicals were originally created as fractions of Total Organic Gas (TOG).
Before these factors could be combined with the factors for the other HAPs, they had to be converted to
fractions of VOC. For the exhaust portion of these HAP emissions, the conversion factors were provided by
the U.S. Environmental Protection Agency's (EPA) Office of Mobile Sources (OMS) (Cook, 1997). For the
evaporative portion, it was determined that evaporative TOG and evaporative VOC emissions were so similar
that a conversion factor of 1 could be used (Cook, 2002).
VOC to TOG Exhaust Conversion Factor
Vehicle Type
TOG/VOC
Light-Duty Gasoline Vehicles and Motorcycles (LDGV)
1.216
Light-Duty Gasoline Trucks-1 and 2 (LDGT)
1.180
Heavy-Duty Gasoline Vehicles (HDGV)
1.086
Light-Duty Diesel Vehicles (LDDV)
1.056
Light-Duty Diesel Trucks (LDDT)
1.056
Heavy-Duty Diesel Vehicles (HDDV)
N fi
The exhaust and evaporative HAP fractions of TOG for LDGVs, LDGTs, and HDGVs for each gasoline fuel
type are listed below. Note that the HDGV acrolein speciation profile (0.0044 acrolein/TOG) and styrene
speciation profile (0 styrene/TOG) were provided by OMS (Cook, 1997). The factors presented below were
converted to be a ratio of milligrams of HAP to grams of VOC when they were entered into the HAP input
files.
Exhaust Factors for Gasoline Vehicles
Fuel Type
Acrolein/
TOG
Fraction
Ethylbenzene/
TOG Fraction
n-Hexane/
TOG
Fraction
Propionaldeh
yde/TOG
Fraction
Styrene/
TOG
Fraction
Toluene/
TOG
Fraction
Xylene/
TOG
Fraction
BaselineGasoline
0.0006
0.0147
0.007
0.0006
0.0034
0.104
0.0586
WO Gasoline/
MTBE/TAME
0.0006
0.0115
0.0073
0.0006
0.0027
0.0812
0.0457
WO Gasoline/
Ethanol
0.0006
0.0134
0.0069
0.0006
0.0031
0.0946
0.0533
RFG/MTBE/
TAME
0.0006
0.0122
0.0072
0.0006
0.0028
0.0863
0.0486
RFG/Ethanol
0.0006
0.0134
0.0069
0.0006
0.0031
0.0946
0.0533
Contam Code
107028
100414
110543
123386
100425
108883
1330207
1 EPA, 1995. Profiles 1313,1314, and 1315.
2 Lindhiem. 1992.
-------
Evaporative Factors for Gasoline Vehicles
(TOG Fraction assumed to equal VOC Fraction for Evaporative Emissions)
Fuel Type
Acrolein/
TOG
Fraction
Ethylbenzene/
TOG Fraction
n-Hexane/
TOG
Fraction
Propionaldeh
yde/TOG
Fraction
Styrene/
TOG
Fraction
Toluene/
TOG
Fraction
Xylene/
TOG
Fraction
BaselineGasoline
0
0.0077
0.0234
0
0
0.0413
0.0223
WO Gasoline/
MTBE/TAME
0
0.0063
0.0087
0
0
0.0276
0.0188
WO Gasoline/
Ethanol
0
0.0045
0.0096
0
0
0.0195
0.0119
RFG/MTBE/TAME
0
0.0063
0.0087
0
0
0.0276
0.0188
RFG/Ethanol
0
0.0045
0.0096
0
0
0.0195
0.0119
Contam Code
107028
100414
110543
123386
100425
108883
1330207
1 EPA, 1995. Profiles 1301,1305, and 1309.
The propionaldehyde exhaust speciation profile for LDDVs and LDDTs was provided in profile 1201 of the
TOC/PM Speciation Data System, Version 2.03 (EPA, 1995).
The acrolein, ethylbenzene, propionaldehyde, styrene, toluene, and xylene exhaust speciation profile for
HDDVs was derived from information provided in Evaluation of Factors That Affect Diesel Exhaust
Toxicity (Truex and Norbeck, 1998). Below is an example of how the HDDV exhaust fraction was calculated:
1.22 ethylbenzene weighted total (mg/Bhp-hr) / 604.91 (mg/Bhp-hr) VOC weighted total =
0.0020 ethylbenzene VOC fraction
The HDDV acrolein, ethylbenzene, n-hexane, styrene, toluene, and xylene exhaust speciation profiles were also
used as surrogates for LDDVs and LDDTs. The factors presented below were converted to be a ratio of
milligrams of HAP to grams of VOC when they were entered into the HAP input files.
Exhaust Factors for Diesel Vehicles
Vehicle Type
Acrolein/
VOC
Fraction
Ethylbenzene/
VOC Fraction
n-Hexane/
VOC
Fraction
Propionaldeh
yde/VOC
Fraction
Styrene/
VOC
Fraction
Toluene/
VOC
Fraction
Xylene/
VOC
Fraction
LDDV
0.0035
0.002
0.0055
0.0186912
0.0021
0.0032
0.0048
LDDT
0.0035
0.002
0.0055
0.0186912
0.0021
0.0032
0.0048
HDDV
0.0035
0.002
0.0055
0.0061
0.0021
0.0032
0.0048
Contam Code
107028
100414
110543
123386
100425
108883
1330207
Note: For ProDionaldehvde the TOG/VOC conversion is already included in the VOC factor.
Evaporative emissions from diesel vehicles is assumed to be zero.
References:
Cook, Rich. Memorandum entitled Guidance on Mobile Source Emission Estimates in the 1996 National
Toxics Inventory, to Laurel Driver and Anne Pope, U.S. EPA Office of Air Quality Planning and Standards.
U.S. EPA Office of Mobile Sources. Ann Arbor, MI. June 9, 1998.
Cook, Rich. Personal Communications. U.S. Environmental Protection Agency Office of Transportation and
Air Quality (OTAQ). Ann Arbor, MI. Summer 2002.
-------
Cook, Rich. Memorandum entitled Source Identification and Base Year 1990 Emission Inventory
Guidance for Mobile Source HAPs on the OAQPS List of 40 Priority HAPs, to Anne Pope U.S. EPA
Office of Air Quality Planning and Standards (OAQPS). U.S. EPA Office of Mobile Sources (OMS). Ann
Arbor, MI. June 11, 1997.
Lindhjem, Chris, Penny Carey, and Joe Sommers. Memorandum entitled Speciation for SAIRuns, to Project
File. U.S. EPA Office of Air and Radiation. Ann Arbor, MI. April 14, 1992.
Mullen, Maureen. E-mail entitled 1996 On-roadExhaust VOC from Trends, to Darcy Wilson, Eastern
Research Group, Inc. Durham, NC. September 8, 1998.
Truex, Dr. Timothy J. and Dr. Joseph M. Norbeck. Evaluation of Factors That Affect Diesel Exhaust
Toxicity. University of California-Riverside, Center for Environmental Research and Technology. Riverside,
CA. March 16, 1998.
U.S. Environmental Protection Agency. National Emission Trends Viewer, Version 2.0 (CD-ROM). Emission
Factor and Inventory Group. Research Triangle Park, NC. June 10, 1998.
U.S. Environmental Protection Agency. TOC/PM Speciation Data System, Version 2.03. Research Triangle
Park, NC. May 1995.
-------
IV. Metals
Gasoline Vehicles
The emission estimates of chromium, manganese, and nickel for gas vehicles were calculated using VMT based
emission factors from Emission Rates and Elemental Composition of Particles Collected From 1995 Ford
Vehicles Using the Urban Dynamometer Driving Schedule, the Highway Fuel Economy Test, and the
US06 Driving Cycle (Ball, 1997). The study provided two sets of emission factors representing two different
vehicle testing cycles, the Urban Dynamometer Driving Schedule (UDDS) and the US06 driving cycle.
Based on a recommendation from the U.S. EPA, Office of Mobile Sources (OMS), the emission factors were
weighted at 28% for the US06 cycle and 72% for the UDDS cycle to best reflect the contribution of actual
vehicle operations. The emission factors were based on testing data from two LDGVs; for the purposes of this
inventory, the factors were also applied to LDGTs. After calculating the weighted average emission factor for
both vehicles, a simple average was taken to represent all LDGV and LDGT vehicle types; this average can be
found on the table below. These factors where then applied to all gas vehicles. (Cook, 1997)
Chromium was speciated into 40% Cr6+ and 60% Cr3+. This was based on instructions and data provided
by EPA (Driver, 2001).
Metal Emissions for Gasoline Vehicles
Metal
ug/mile
tons/mile
Contam Code
Total Chromiun
4.95
5.46E-12
Chromim (Cr6+) 40%
1.98
2.18258E-12
18540299
Chromim (Cr3+) 60%
2.97
3.27387E-12
16065831
Manganese
1.66
1.83E-12
198
Nickel
3.6
3.97E-12
226
Arsenic and Mercury factors were set at of half the detection limit. This was based on instructions and data
provided by EPA (Cook, 2001).
Arsenic and Mercury Values, based on half detection limit.
Vehicle Type
As
Hg
As
Hg
milligrams/mile
milligrams/mile
tons/mile
tons/mile
LDDV
0.007894737
0.006578947
8.70246E-12
7.25E-12
HDDV
0.054110616
0.086576985
5.96468E-11
9.54E-11
LDGV
0.002875
0.000875
3.16915E-12
9.65E-13
HDGV
0.002756472
0.000838926
3.03849E-12
9.25E-13
Contam Code
93
199
93
199
Metal Emissions for Diesels
Factors for metal emissions from diesel vehicles were taken from the report Evaluation of Factor that Affect
Diesel Exhaust Toxicity from Center for Environmental Research and Technology, College of
Engineering, University of Califronia, 1998.
-------
Metal Emissions from Diesels
Pollutant
Pre-1993 Fuel
Low Aromatic Fuel
Reform. Diesel Fuel
Cold Start
Hot Start
Cold Start
Hot Start
Cold Start
Hot Start
Cr
0.81
0.72
-
0.06
-
0.03
Mn
1.04
0.36
1.39
0.02
1.37
-
Ni
2.23
1.34
-
-
2.96
1.11
All factors in ug/Bhp-hr (as per Rich Cook, multiply by 1.8 to convert to ug/mile)
Factors for Metal Emissions from Diesels
Factors
Cr6
Cr3
Mn
Ni
in ug/mile
0.527657143
0.791485714
0.822857143
2.640857
in Tons/mile
5.81643E-13
8.72464E-13
9.07045E-13
2.91E-12
Contam Code
18540299
16065831
198
226
References:
Ball, James C. Emission Rates and Elemental Composition of Particles Collected From 1995 Ford
Vehicles Using the Urban Dynamometer Driving Schedule, the Highway Fuel Economy Test, and the
US06 Driving Cycle. 97FL-376. Society of Automotive Engineers, Inc. 1997.
Cook, Rich. Memorandum entitled Source Identification and Base Year 1990 Emission Inventory
Guidance for Mobile Source HAPs on the OAQPS List of 40 Priority HAPs, to Anne Pope U.S. EPA
Office of Air Quality Planning and Standards (OAQPS). U.S. EPA Office of Mobile Sources (OMS). Ann
Arbor, MI. June 11, 1997.
Cook, Rich. Personal Communications. U.S. Environmental Protection Agency Office of Transportation and
Air Quality (OTAQ) . Ann Arbor, MI. Summer 2001.
Truex, Dr. Timothy J. and Dr. Joseph M. Norbeck. Evaluation of Factors That Affect Diesel Exhaust
Toxicity. University of California-Riverside, Center for Environmental Research and Technology. Riverside,
CA. March 16, 1998.
U.S. Department of Transportation (DOT). Highway Statistics 1996. FHWA-PL-91-003. Office of Highway
Information Management, Office of Policy Development, Federal Highway Administration. U.S. Government
Printing Office. Washington, D.C. Available at the following Internet site:
http://www.fhwa.dot.gov/ohim/1996/index.html. August 17, 1998.
U.S. Environmental Protection Agency. National Emission Trends Viewer, Version 2.0 (CD-ROM). Emission
Factor and Inventory Group. Research Triangle Park, NC. June 10, 1998.
U.S. Environmental Protection Agency. TOC/PM Speciation Data System, Version 2.03. Research Triangle
Park, NC. May 1995.
U.S. Environmental Protection Agency. Motor Vehicle-Related Air Toxics Study (MOBTOX).
EPA-420-R-93-005. U.S. EPA Office of Mobile Sources. Ann Arbor, MI. April 1993.
-------
V. 2,2,4 Trimethylpentane
Factors for 2,2,4 Trimethylpentane were taken from the SPECIATE database from EPA (EPA, 2001).
Guidance on which profiles to use where given by Rich Cook of the EPA (Cook, 2001). These factors are
given as a percentage of VOC emissions.
2,2,4 Trimethylpentane Percentages for Gasoline Vehicles
(Contam Code 540841)
PNO
Description
% VOC
1301
10% Ethanol Composite (Hot Soak + Diurnal) Evaporative
1.71
1305
Industry Average (circa 1990) Gasoline Composite (Hot Soak +
Diurnal) Evaporative
1.43
1309
11% MTBE Composite (Hot Soak + Diurnal) Evaporative
1.59
1313
Industry Average (circa 1990) Gasoline Exhaust
4.32
1314
10% Ethanol Exhaust
4.27
1315
11% MTBE Exhaust
4.39
Factors
Trimeth Fraction
mg Trimeth / gm VOC
Gasoline - Evap
0.0158
15.8
Gasoline - Exhaust
0.0433
43.3
Factors for diesel emissions were taken from documentation provided by Rich Cook of the EPA (Cook,
2001).
Diesel 2,2,4 Trimethylpentane Data
ENGINE
Recommend Value
for NTI
Surrogate Category
VOC
Fraction
mg TM 1
gm VOC
CUMMINS L10
Yes (49 state)
Non-road and On-road
0.00066
0.66
CUMMINS L10
Yes(California)
Non-road and On-road
0.00059392
0.59392
evap - speciate
Yes (50 state evap)
Heavy Duty Gasoline evap
0.01411538
14.11538
exhaust-speciate
Yes (50 state)
Heavy Duty Gasoline
0.02582083
25.82083
References:
Cook, Rich. Personal Communications. U.S. Environmental Protection Agency Office of Transportation and
Air Quality (OTAQ). Ann Arbor, MI. Summer 2001.
U.S. Environmental Protection Agency. SPECIATE, Version 3.1. Office of Air Quality Planning and
Standards. Research Triangle Park, NC. 2001.
-------
Attachment C
1999 Onroad Dioxin/Furan Emission Estimating Methodology
For dioxins/furans, emissions can be calculated in terms of toxic equivalents. The toxicity of 2,3,7,8-
tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) is considered the standard used to compare other dioxin/furan
cogeniers. These toxic equivalents (TEQs) a represent a single, aggregated measure of all dioxin and furan
congeners. For this reason, only 2,3,7,8-TCDD TEQs are used in this inventory to represent dioxins/furans.
Dioxin/furan emission factors for gasoline and diesel powered vehicles were provided in 1990 Emissions
Inventory of Section 112(c)6 Pollutants: Polycyclic Organic Matter (POM), 2,3,7,8-Tetrachlorodibenzo-
P-Dioxin (TCDD)/ 2,3,7,8-Tetrachlorodibenzofuran (TCDF), Polychlorinated Biphenyl Compounds
(PCBs), Hexachlorobenzene, Mercury, and Alkylated Lead, Final Report (EPA, 1997). These emission
factors were converted to tons TEQ/mile, as follows and are summarized in Table C-l.
Gasoline-Powered Vehicles
0.36xl0"12 g TEQ/km X lg/10 12 pg X 1 ton/ 907,184.70 g X 1.609344 km/ 1 mile
= 6.39x10 -19 tons TEQ/mile
Diesel Vehicles
0.50xl0"9 g TEQ/km X lg/ 10 9 ng X 1 ton/ 907,184.7 g X 1.609344 km/ 1 mile
= 8.87x10 -16 tons TEQ/mile
Table C-l Dioxin/Furans as 2,3,7,8-TCDD TEQ Emissions Factors
Fuel Type
Factor
Units
Gasoline-Powered Vehicles
6.39E-19
tons TEQ/mik
Diesel Vehicles
8.87E-16
tons TEQ/mik
To calculate 1999 emissions for dioxins/furans as 2,3,7,8-TCDD TEQ, emission factors were applied
to the 1999 vehicle miles traveled (VMT) for gasoline and diesel powered vehicles, as noted in the following
equation:
(VMT for gasoline vehicles ) x (Dioxin Emission Factor) = (Toxic Annual Emission)
Example: (20 x 106 miles) x (6.39 x 10"19 tons/mile) = 1.26xl0"n tpy of Dioxins/Furans
References:
U.S. Department of Transportation (DOT). Highway Statistics 1999. Office of Highway Information
Management, Office of Policy Development, Federal Highway Administration. U.S. Government Printing
-------
Office. Washington, D.C. Available at the following Internet site:
http://www.fhwa.dot.gov/ohim/1999/index.html. VMT data downloaded Summer of 2001.
U.S. Environmental Protection Agency (EPA). 1990 Emissions Inventory of Section 112(c)6 Pollutants:
Poly cyclic Organic Matter (POM), 2,3,7,8-Tetrachlorodibenzo-P-Dioxin (TCDD)/
2,3,7,8-Tetrachlorodibenzofuran (TCDF), Polychlorinated Biphenyl Compounds (PCBs),
Hexachlorobenzene, Mercury, and Alkylated Lead, Final Report. Research Triangle Park, NC.
June 1997.
-------
Attachment D
Potential Approaches for Developing a Mercury Inventory for Mobile
Sources
-------
MEMORANDUM
Potential Approaches for Developing a Mercury Inventory for Mobile Sources
Rich Cook
Air Toxics Center
Office of Transportation and Air Quality
Marion Hoyer
Air Toxics Center
Office of Transportation and Air Quality
Laurel Driver
Emission Factor and Inventory Group
Office of Air Quality Planning and Standards
Phil Lorang
Emission Factor and Inventory Group
Office of Air Quality Planning and Standards
EPA's 1997 Mercury Report to Congress did not include a mobile source emissions estimate, concluding that
currently available data were insufficient. The purpose of this memorandum is to describe potential approaches
for developing a nationwide mercury inventory for mobile sources, for use in the National Emissions Inventory
(NEI) and other Agency assessments. The available data which could be used are discussed, along with their
limitations. The memorandum also briefly describes testing being conducted under contract to the Office of
Transportation and Air Quality (OTAQ) which could potentially be used to develop a preliminary on-road
mercury inventory in the future.
We recently estimated mercury emissions from mobile sources for the draft 1999 National Emissions Inventory
for Hazardous Air Pollutants, Version 3. Data used for the estimate were particulate matter emissions testing
conducted on various vehicle classes for the Northern Front Range Air Quality Study. Emission factors were
derived from the uncertainty values reported for mercury from X-ray fluorescence (XRF) analysis of particulate
emissions.7'8 The emission factors were then applied to activity estimates. The result was a nationwide mobile
source inventory of about 28 tons. This approach has a number of serious limitations: X-ray fluorescence
(XRF) cannot be used to accurately quantify trace amounts of particulate mercury since the high energy X-rays
can cause dissociation of mercury from the particulate fraction into the gaseous phase, potentially
7Heavy-Duty Engines - Gertler, A. W., W. Coulombe, R. Troppe, J. A. Gillies, C. F. Rogers,
and W. R. Pierson. 2000. Mobile Source Issues Related to the Proposed PM2.5 Standard: Year 2.
Desert Research Institute, Reno, NV, June 1, 2000. See Table II, page III-169.
8Light-Duty Gasoline and Light-Duty Diesel Vehicles - Joseph Norbeck, Thomas Durbin, and
Timothy Truex. 1998. Measurment of Primary Particulate Matter Emission from Light-Duty Diesel
Motor Vehicles. CRC Project No. E-24-2, Appendix I, pages 1-1 through 1-8 and Appendix J, pages
J-l through J-4.
SUBJECT:
FROM:
D-l
-------
underestimating levels; the data account for particulate mercury only and the majority of emitted mercury is in
vapor phase; and the detection limit for mercury measurement was not used.
We have identified two potential alternative approaches for developing an inventory:
1) Estimate an inventory using the mercury emission factor from the 1977 Tuscorora Tunnel Study - A 1977
study of emissions in the Tuscorora Tunnel9 estimated a particulate mercury emission factor for motor vehicles
of 2.09 |j,g/mi. This study used neutron activation analysis (NAA) to measure particle phase mercury. NAA is
capable of quantifying low levels of mercury. However, vapor-phase mercury (elemental and reactive), which
comprise the majority of mercury emissions from anthropogenic sources was not measured. The emission factor
potentially includes the contribution of mercury from brake wear (which may be substantial) and other sources
as well. This study does not include separate emission factors for gasoline and diesel vehicles, which are likely
to be quite different. In addition, it is unlikely this emission factor is representative of the current fleet, since the
data were collected over 25 years ago, and predate the use of catalytic converters. The lack of separate
emission factors for gasoline versus diesel engines would make it impossible to develop a source-category
specific inventory. The highway vehicle portion of the mobile source inventory in 1999 would be about 6 tons
using the above emission factor.
2) Estimate an inventory based on the mercury concentration in gasoline and diesel fuel — A recent paper
summarized data on mercury levels in various petroleum products.10 The concentration of mercury in gasoline
and diesel fuel was estimated to be roughly 2 ng/g. Environment Canada recently estimated mercury emissions
for onroad vehicles using this concentration and obtained a mobile source mercury inventory for Canada of 30
kg.11 The U. S. mobile source inventory using this approach would be a little less than 0.2 tons, based on fuel
sales data from the Energy Information Administration (Table l).12 There are no measurements of the amount
of mercury in lube oil. Environment Canada used an estimate of 50 ng/g, which is the concentration of mercury
in heavy crude oil, and assumed 1% of the mercury in lube oil contributes to mercury emissions in exhaust. This
results in a very small emission factor, about 30 kg for all of Canada. It should be noted that estimates of
mercury in petroleum products may be biased low, because much of the mercury emitted can adhere to walls of
sample containers, and mercury can also be lost to volatilization. The most significant drawback to
implementing this approach is the lack of data representing currently available fuels in the U.S. as well as the
lack of published data regarding the concentration of mercury in lube oil. As discussed below, preliminary data
suggest that mercury measured in light-duty vehicle exhaust is greater than that which can be accounted for by
the fuel alone.
In order to confirm emissions of mercury from motor vehicles and to evaluate the potential level of emissions,
investigators at the University of Michigan Air Quality Laboratory recently completed a small scale test program
9Pierson, William R. and W. A. Brachaczek. 1983. Particulate Matter Associated with
Vehicles on the Road. Aerosol Science and Technology 2: 1- 40.
10Wilhelm, Mark S. 2001. Estimate of Mercury Emissions to the Atmosphere from Petroleum.
Env. Sci. Technol. 35: 4704 - 4710.
"Environment Canada. 2003. Emissions of Toxic Substances from On-Road Motor Vehicles
in Ontario - Draft. Toxics Prevention Division, Environment Canada - Ontario Region, Ontario,
Canada.
12Energy Information Administration. 2000. Petroleum Marketing Annual 1999.
D-2
-------
under contract to EPA Office of Transportation and Air Quality. This work only includes two light duty
gasoline vehicles and one heavy duty diesel engine. We are currently analyzing results of this test program. It
may be possible to use the pilot study results to make some interim inventory calculations. However, an
inventory using these data would still be highly uncertain due to the limited number of vehicles tested and could
underestimate total mercury since quantification of reactive gaseous mercury was not possible. Preliminary data
from the University of Michigan Air Quality Laboratory suggest that mercury levels in gasoline can be at least an
order of magnitude higher than the limited data in the literature, and mercury levels in lube oil were higher than
levels in gasoline.
A complete testing program to assess mercury emissions from motor vehicles is needed. EPA's OTAQ is
formulating a test program plan and will be soliciting assistance to characterize emissions from onroad and non-
road motor vehicles and their fuels and lubricating oils.
C C: Marion Hoyer
Joe Somers
Carl Scarbro
Richard Baldauf
Kathryn Sargeant
Chet France
John Bachmann (OAQPS)
Anne Pope (OAQPS)
Paul Almodovar (OAQPS)
Richard Wayland (OAQPS)
D-3
-------
Table 1. Estimate of Mobile Source Mercury Emissions in the U.S. from Trace Levels in
Fuel
Total Fuel Sales in the U.S.,
1999, Gallons (Source: Energy
Information Administration,
Petroleum Marketing Annual)
Specific Gravity
(g/gai)
Hg Level in Fuel
(ng/g)
Total Nationwide
Emissions from
Mercury in Fuel
(tons)
Gasoline
Diesel
22,630,000,000
7,701,500,000
2,780
3,207
0.15
0.05
D-4
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-454/B-20-018
Environmental Protection Air Quality Assessment Division January 2004
Agency Research Triangle Park, NC
------- |