*** Office of Transportation EPA420-R-05-024
and Air Quality December 2005
United States
Environmental Protection
Agency
EPA's National Inventory
Model (NMIM), A Consolidated
Emissions Modeling System for
MOBILE6 and NONROAD
-------
EPA420-R-05-024
December 2005
A for
and
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
-------
TABLE OF CONTENTS
ABSTRACT 1
INTRODUCTION AND OVERVIEW 1
HOW NMIM WORKS 2
Temporal and Spatial Approximations for Efficient Running of National Inventories . . 3
NMIM Master and NMIM Worker: Distributed Processing 3
HOW NMIM RUNS MOBILE6 4
Particulate Matter (PM) 7
Evaluation Year and Month 7
HOW NMIM RUNS NONROAD 8
POLLUTANTS FOR WHICH INVENTORIES ARE PRODUCED BY NMIM 9
SOURCE CATEGORIES FOR WHICH INVENTORIES ARE PRODUCED BY NMIM .... 12
THE NMIM COUNTY DATABASE (NCD) 14
Temperature and Humidity 16
Fuel Properties 17
Vehicle Miles Traveled (VMT) 18
Database Improvement 18
ESTIMATION OF HAZARDOUS AIR POLLUTANT (HAP) EMISSIONS 19
Gaseous HAPS 19
Polyaromatic Hydrocarbons (PAHs) 20
Metals, Dioxins, and Furans 20
RETROFIT AND FLEET MODELING 21
SUMMARY 22
REFERENCES 23
-------
ABSTRACT
This document describes EPA's National Mobile Inventory Model (NMIM) and explains
how it works. NMIM is a consolidated emissions modeling system for EPA's MOBILE6 and
NONROAD models. It generates county inventories using MOBILE6 and NONROAD at scales
ranging from individual counties to the nation. Its primary improvements over MOBILE6 and
NONROAD are 1) the inclusion of all the required county data for the nation in a single
database, 2) a graphical user interface, 3) "shortcuts" for generating national inventories, 4) tools
for aggregation and post-processing, 5) estimation of 33 HAPS and 17 dioxin/furan congeners,
6) modeling user-specified retrofit programs and fleets, and 7) distributed processing capability
to enhance performance. Previously, when national inventories have been constructed from
MOBILE6 and NONROAD, the necessary input data have been widely scattered in disparate
formats and have required additional specialized software to convert this data into input files for
MOBILE6 and NONROAD, to run the models, to integrate the results into a final inventory, and
to post-process the results into forms suitable for the national inventories. NMIM accomplishes
all of these tasks in a single package. NMIM's initial version in 2003 was intended to
approximate the procedures previously used for calculating nationwide emissions for EPA's
National Emissions Inventory. EPA subsequently has added the capability of incorporating
much more county-specific information into the database, including monthly average hourly
temperature and humidity, fleet distributions, speed distributions, and diesel fractions for
MOBILE6 and county-specific allocation and activity files for NONROAD.
INTRODUCTION AND OVERVIEW
This document describes EPA's National Mobile Inventory Model (NMIM) and explains
how it works.1 EPA's National Mobile Inventory Model (NMIM) is a consolidated emissions
modeling system for EPA's MOBILE6 and NONROAD models. It was developed to produce, in
a consistent and automated way, national, county-level mobile source emissions inventories for
the National Emissions Inventory (NEI) and for EPA rule making. Previously, when national
inventories have been constructed from MOBILE6 and NONROAD, the necessary input data
have been widely scattered in disparate formats and have required additional specialized
software to convert this data into input files for MOBILE6 and NONROAD, to run the models,
to integrate the results into a final inventory, and to post-process the results into forms suitable
for the national inventories. NMEVI accomplishes all of these tasks in a single package.
NMEVI comprises a Java framework, graphical and command line user interfaces, the
MOBILE6 and NONROAD models, a national county database, and postprocessing and
aggregation capabilities. NMIM's primary improvements over MOBILE6 and NONROAD are
1) the inclusion of all the required county data for the nation in a single database, 2) a graphical
user interface, 3) "shortcuts" for generating national inventories, 4) tools for aggregation and
post-processing, 5) estimation of 33 hazardous air pollutants (HAPS) and 17 dioxin/furan
congeners by ratio to various MOBILE6 and NONROAD output parameters, 6) modeling user-
specified retrofit programs and fleets, and 7) distributed processing capability to enhance
performance. NMIM extends MOBILE6's capabilities by producing inventories rather than just
1
-------
emissions factors. NMIM provides consistency across both models and all pollutants by using a
single input database for MOBILE6 and NONROAD and for criteria pollutants and HAPS.
However, NMEVI is not a complete GUI front end for MOBILE6 and NONROAD, nor
can it replace these models for all applications. Because it was designed primarily to generate
national inventories, NMIM has features intended to speed its operation and minimize its output
database size that also limit its applicability. For example, it will not produce hourly or by-
model-year output from MOBILE6. In addition NMIM, like NONROAD, does not produce
emissions inventories for aircraft, locomotives, and commercial marine vessels.2
This documentation begins with an overall explanation of how NMEVI works, followed
by the details of how it runs MOBILE6 and NONROAD. Next, it discusses the pollutant and
source category inventories available from running NMIM. Then, it describes the NMIM
County Database (NCD) and procedures for updating and improving it through the National
Emission Inventory (NET) process. Next is an explanation of how NMIM estimates various
HAPS that are not direct outputs of MOBILE6 and NONROAD. Finally, it describes how user-
specified retrofit programs and fleets are modeled by NMIM.
HOW NMIM WORKS
The NMIM user specifies a set of years and months, a geographic region (the whole US,
any combination of whole states, or any combination of particular counties, including Puerto
Rico and the Virgin Islands), a set of pollutants, and categories of onroad vehicles and nonroad
equipment. This collection of user requests is called a "run specification" or RunSpec, and can
be saved in a file for later execution or for text editing. RunSpecs can be produced by the
NMIM GUI or by using a text editor. NMIM RunSpecs can be executed from the GUI or from
the command line.
Based on the RunSpec and information in the NMIM county database (NCD), NMIM
writes input files for the MOBILE6 and NONROAD models. NMIM then runs these models,
reads their output files, performs additional processing if necessary, and puts the inventories into
an output database. Additional processing includes multiplying MOBILE6 emission factors by
vehicle miles traveled (VMT) and estimating emissions of some other pollutants (see below) as
ratios to pollutant inventories generated by MOBILE6 and NONROAD.
NMEVI has post-processing capability that can be applied after the inventory is generated.
This includes aggregation over months, roadway types, vehicle types, and equipment types.
Output can go into a database table or into a tab-delimited text file. The format can be
normalized (a pollutant on each line) or wide (a pollutant in each column.) NMIM can also
process the output database into annual, monthly, or ozone season day (OSD) inventories and
convert them into NEI Input Format Version 3 (NTF3), which is required for NEI submissions.
The user's post-processing selections can be made from the GUI and saved as a text
-------
"aggregation specification" file or AgSpec, which can be reloaded into the GUI, edited from the
GUI or in a text editor, and run from the GUI or in acommand line.
Temporal and Spatial Approximations for Efficient Running of National Inventories
NMIM employs two main techniques, adopted from previous NEIs, to make the
production of national inventories tractable. The first is to assume that monthly time resolution
is adequate for both meteorology and source activity and therefore to do twelve monthly runs
instead of 365 daily runs. NMIM is designed to do only monthly runs and produces annual
inventories by adding the twelve monthly inventories.
The second technique is to group similar counties, allowing NMIM to do a single
MOBILE6 or NONROAD run for the entire group. The county groups are defined in the
CountyMap table of the NCD. One of the counties in each group is referred to as the
representing county, since it represents the others, which are called represented counties. It
should be noted that county grouping is not mandatory; NMIM allows the user to employ county
groups or to run all counties individually. If time is available, running individual counties is
superior, since it uses the most individualized data for each county. In either case, the output
contains inventories for all the individual counties in the selected geographic region.
As an example of county grouping, for some national runs that EPA has performed to
date, counties were placed in the same county group for a particular year if they were in the same
state and had the same fuel properties and control programs. This grouping was adapted from
past NEIs. It assumed that if counties are in the same state, their temperatures are similar
(California had two temperature groupings). The representing county was chosen as the one
with the highest VMT, so that the temperature and relative humidity data is from a county with a
significant share of mobile source activity. This scheme resulted in about 100 county groups for
NONROAD and about 200 for MOBILE6 in 2002, representing the 50 states plus the District of
Columbia, a total of more than 3000 counties.
EPA does not recommend or supply any particular county grouping. If they choose to
employ county groups, users must tailor them to suit the purposes of their inventories. As more
county-specific data becomes incorporated into the NCD, the problem of choosing county groups
becomes more difficult. On the other hand, as computing power increases, it becomes easier to
run every county.
NMIM Master and NMIM Worker: Distributed Processing
As a way of further improving performance, NMIM may be run in a distributed-
processing mode, employing multiple computers. NMIM comprises two programs, Master and
Worker. Master and Worker(s) each have a simple text configuration file which specifies the
path to a shared folder through which they communicate. In standalone mode, one Master, one
Worker, and the shared folder are on the same computer. In distributed mode, the Master is on
one computer, multiple workers are on separate computers, and the shared folder, which they all
-------
have access to, is on yet another computer (although it could be on the same computer as the
Master or one of the Workers). Therefore, to execute NMIM in distributed mode requires only
an ordinary local area network (LAN) where all the NMIM computers have access to a single
shared folder.
In either mode, to execute a RunSpec requires a Master, one or more Workers, and a
shared folder. When a RunSpec is executed, the Master 1) reads the RunSpec (which defines the
inventory in terms of geography, time, pollutants and sources), 2) writes appropriate MOBILE6
and NONROAD input files, 3) bundles these files together with any external data files needed,
and 4) posts them as "TODO" files on the shared folder. A TODO file always contains the data
for a single MOBILE6 or NONROAD run for one county for one month. A Worker, seeing a
TODO file on the shared folder, copies it to its local work space, executes MOBILE6 or
NONROAD as appropriate, and writes the results to a "DONE" file in the shared folder. The
Master, seeing the DONE file, downloads it from the shared folder to its local workspace,
extracts the MOBILE6 or NONROAD output file, processes it further as necessary, and inserts it
into the output database specified in the RunSpec.
HOW NMIM RUNS MOBILE6
NMEVI writes a MOBILE6 input file using data obtained from the NCD and executes
MOBILE6 once for each month for each county (if the user chose the Geographic
Representation option "County") or once for each representing county (if the user chose the
Geographic Representation option "County Group"). To produce an onroad inventory, NMIM
multiplies the resulting emission factors by the VMT for each requested county, regardless of
which Geographic Representation option was chosen.
The MOBILE6 input files constructed by NMIM are designed to accommodate detailed
user input and to use a consistent set of commands. In order to use consistent fuels data for all
pollutants, the AIR TOXICS command is always used, even if no air toxics are requested by the
user. To enable the AIR TOXICS command, MOBILE6 requires that at least one of the six air
toxics that it models internally be specified in its input file. To fulfill this requirement, NMIM
always inserts the command to model acrolein in the MOBILE6 input file, whether the user
requests it or not. However, if the user does not request acrolein, it will not appear in the output
table. (Note that acrolein was an arbitrary choice; any of the six air toxics would have served the
same function.) The ALTITUDE, POLLUTANTS and EVALUATION MONTH commands are
always explicitly used, rather than depending on MOBILE6 default settings and always appear in
the MOBILE6 input files written by NMIM. The input files always use the HOURLY
TEMPERATURES command, rather than MIN/MAX TEMPERATURE command. The hourly
RELATIVE HUMIDITY command is always used, rather than the ABSOLUTE HUMIDITY
command. The BAROMETRIC PRES command is always used, since this value interacts with
the relative humidity values. Average speed distributions are always specified using the SPEED
VMT command, rather than the AVERAGE SPEED command.
-------
Gasoline fuel parameters require the maximum amount of detail to properly model toxic
emissions. The OXYGENATE command is required when modeling air toxics, and supersedes
the OXYGENATED FUELS command used to model winter oxygenate programs in MOBILE6.
The OXYGENATE command requires that oxygen content be expressed as volume percent
instead of weight percent, as required by the OXYGENATED FUELS command. GAS
AROMATIC%, GAS OLEFIN% , GAS BENZENE%, E200 and E300 must always be specified.
RVP OXY WAIVER command is always set to 1 (no waiver), because Reid vapor pressure
(RVP) values from the fuel surveys are assumed to already account for any RVP effect from
oxygenated fuels. The FUEL RVP command is always required. The GASOLINE SULFUR
command and FUEL PROGRAM command Option 4 are always used to explicitly set the sulfur
content of gasoline. The same gasoline sulfur content is used for both commands and for all
years in the FUEL PROGRAM command, although it would normally have different sulfur
values for different years. Since each MOBILE6 run covers only a single month in a particular
calendar year, NMIM sets all possible sulfur values the same to avoid programming logic to
determine which of the possible years to change.
Some counties have local emission control programs. The basic information for these
programs is stored in the NCD and is used to create the appropriate commands for the input file
when needed. Inspection and maintenance (I/M) programs for counties are stored in external
data files and accessed using the I/M DESC FILE command.
In addition to the basic required information, NMEVI can also include county specific
data that is normally provided to MOBILE6 using external data files. Nearly any of the valid
MOBILE6 commands can be used, including commands used to model local Low Emission
Vehicle (LEV) phase-in programs and local natural gas vehicle fractions. Diesel sales fractions
are stored in an external file and used to create the appropriate input command.
Not all MOBILE6 commands are used by NMIM. The VMT FRACTIONS and VMT
BY FACILITY commands are not needed, since these commands are only needed to create
composite emission rates. NMIM converts all emission rate results from MOBILE6 to tons
using the county specific VMT for each vehicle class and roadway type. Since all gasolines are
explicitly defined, both the SEASON command and FUEL PROGRAM command Option 2 are
never used. Options such as NO CLEAN AIR ACT can be specified as "global" NMIM inputs
that affect all counties and all months.
Only weekdays are modeled by NMEVI. Commands that apply to weekend variations are
not used. This simplification makes sense because most weekend differences in MOBILE6 are
temporal distributions, so MOBILE6's emission factors at the day level are little affected by
these differences. (Parameters that can differ between weekends and weekdays are hot soak
duration distribution, start distribution, starts/day, soak distribution, and trip length distribution.)
The major difference between weekdays and weekends is VMT, which is provided in the NCD
by month, vehicle type, and roadway type.
MOBILE6 has only 4 facility or roadway types: freeways, arterials, ramps, and locals.
-------
Ramp speed is fixed at 34.6 MPH and local speed at 12.9 MPH. Distributions of average speeds
are specified separately for freeways and arterials using the SPEED VMT command. If only a
single MOBILE6 scenario were run, the same distribution of average speeds would be applied to
all vehicle types. Separate specification of average speed distributions for all combinations of
the twelve roadway types and all twenty-eight MOBILE6 vehicle types would require running
168 MOBILE6 scenarios.
In order to avoid running so many MOBILE6 scenarios, while retaining reasonable
flexibility, NMIM groups vehicle class-roadway type combinations into 18 groups, shown in
Table 1. These groups are those that have been used in past NEIs, and provide flexibility in
assigning average speeds while limiting the number of MOBILE6 runs necessary to generate an
inventory. Since nine of these combinations use the MOBILE6 freeway facility type and nine
use the arterial facility type, a total of nine MOBILE6 scenarios are needed to model the
eighteen vehicle class/roadway type combinations. Average speed distributions for each of these
eighteen vehicle class/roadway type combinations can be specified for each county.
Table 1. The 18 vehicle class-roadway type combinations in NMIM.
M6Vtypes Road Types M6 Ftype
LDV Rural Interstate
LOT Rural Interstate
HDV Rural Interstate
LDV Urban Interstate
LDT Urban Interstate
HDV Urban Interstate
LDV Urban Freeways & Expressways
LDT Urban Freeways & Expressways
HDV Urban Freeways & Expressways
LDV,LDT Rural Principal Arterial
LDV,LDT Rural Minor Arterial
HDV Rural Principal Arterial
LDV,LDT Rural Major Collector
LDV,LDT Rural Minor Collector, Rural Local
HDV Rural Minor Arterial
Freeway
Freeway
Freeway
Freeway
Freeway
Freeway
Freeway
Freeway
Freeway
Arterial
Arterial
Arterial
Arterial
Arterial
Arterial
LDV,LDT Urban Principal Arterial, Urban Minor Arterial, Urban Collector Arterial
HDV Rural Major Collector, Rural Minor Collector, Rural Local Arterial
HDV Urban Principal Arterial, Urban Minor Arterial, Urban Collector Arterial
* Reference MOBILE6.2 User Guide, Appendix B
LDV = MOBILE6 Vehicle Types 1 and 16.
LDT = MOBILE6 Vehicle Types 2-5.
HDV = MOBILE6 Vehicle Types 6-15.
-------
Particulate Matter (PM)
Because MOBILE6 can model only one particle size at a time, if both PM10 and PM2.5
are desired, NMIM runs a tenth MOBILE6 scenario to obtain emission factors for the extra
particle size. MOBILE6 separates exhaust particulates into SO4, organic carbon (OC),
elemental carbon (EC), lead, tire wear, and brake wear. Of these, only SO4 depends on speed.
If either PM10 or PM2.5 is requested, the results are obtained from the standard nine scenarios.
If both PM10 and PM2.5 are requested, the nine scenarios are run for PM10. Since all SO4 is
PM2.5, if both PM10 and PM2.5 are requested, the SO4 emission factor for both is taken from
the nine scenarios that are sensitive to speed, and the tenth scenario is used to obtain the
emission factors for all the other PM2.5 components.
Evaluation Year and Month
MOBILE6 specifies a calendar year and an evaluation month of either January or July.
These two parameters determine the fleet composition for which emission factors are generated.
For each month of a given inventory year, NMIM writes the MOBILE6 input file using the
combination of calendar year and evaluation month shown in Table 2 below.
Table 2. The MOBILE6 calendar years and evaluation months that are used by NMIM for each
month of inventory year, Y.
NMIM Month of MOBILE6 calendar year MOBILE6 evaluation month
Inventory Year Y
1
2
3
4
5
6
7
8
9
10
11
12
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y+l
Y+l
Y+l
1
1
1
7
7
7
7
7
7
1
1
1
The reasoning behind this scheme is that the fleet composition in October, November, and
December of year Y is more like that of January of year Y+l than it is like July of year Y. This
scheme does not cause a problem with fuel properties, because NMIM always looks up the fuel
properties in the NCD for the inventory year and month being modeled. Control programs in
MOBILE6 are always assumed to begin on January 1, but MOBILE6 assumes that these
programs have no effect on that day, since the program has had no time to get started. Hence
-------
NMIM is not erroneously introducing next year's control programs by modeling October,
November, and December as January 1 of the following year.
HOW NMIM RUNS NONROAD
NONROAD estimates monthly fuel consumption and emissions of total hydrocarbons
(THC), carbon monoxide (CO), nitrogen oxides (NOX), sulfur dioxide (SO2), and particulate
matter (PM). NMIM then processes the monthly results as needed to produce annual and ozone
season day emissions. Additional pollutants are produced by NMIM as ratios to some of these
outputs.
The NONROAD Model reads a set of text instructions, known as an "opt file" (for
options). NMIM creates this file from data in the NCD. As employed in NMIM, the opt file is
limited to one state and specifies month and year, fuel properties, temperature, and the counties
for which to calculate emissions, which may be all or a subset of the counties in the state.
NONROAD internally produces emissions for the whole state and then allocates the emissions
for each SCC to the requested counties. Output is produced only for the county or counties
selected in the NMIM RunSpec.
The NONROAD Model includes a group of files that specify equipment populations,
emission factors, deterioration rates, activities, and allocations from the state to the county.
County-specific allocation, population, seasonality, and activity files can be specified in the
NCD. These files will override NONROAD's default files.
The fuel properties required by the NONROAD Model are not the same as those in the
NCD. The NONROAD Model requires "Oxygen Weight %" in its opt file. The conversion
from NCD fuel properties to oxygen weight percent is performed by NMIM as follows:
oxywtpct =
etohvolume * 0.3448* etohmktshare
+ mtbevolume*0.1786*mtbemktshare
+ tamevolume*0.1636*tamemktshare
+ etbevolume*0.1533*etbemktshare
These conversion factors are detailed under the OXYGENATE command in the MOBILE6
User's Guide.3
How NMIM converts from Total Hydrocarbons (THC) to other hydrocarbon (HC) species
THC is the NONROAD Model's native output. The other HC species that can be
requested from NMIM are listed and described in Table 3 below. The conversion from THC to
the other HC species differs between exhaust and evaporative emissions. For NMIM's
emissiontype classifications (exhaust, evaporative, refueling), NONROAD crankcase emissions
are classified as exhaust.
-------
NMIM uses factors in the SCC table to convert NONROAD exhaust THC to the other
hydrocarbon outputs (VOC, NMHC, TOG, and NMOG). For evaporative emissions except for
FuelType=CNG, no conversion is necessary (i.e., VOC = NMHC = TOG = NMOG = THC). For
evaporative CNG emissions, TOG = THC, and NMOG = NMHC = VOC = 0.
POLLUTANTS FOR WHICH INVENTORIES ARE PRODUCED BY NMIM
Hydrocarbons may be expressed in one of five forms, listed in Table 3 below. The
conversion factors in the SCC table are those used in the MOBILE6 and NONROAD models and
depend on fuel and engine type.
Table 3. Hydrocarbon forms available from NMIM. (MOBILE6 User Guide).
Hydrocarbon Form Includes Includes Includes Includes
FIDHC Methane Ethane Aldehydes
Total Hydrocarbons
Non-methane Hydrocarbons
Volatile Organic Compounds
Total Organic Gases
Non-methane Organic Gases
(THC)
(NMHC)
(VOC)
(TOG)
(NMOG)
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
No
Yes
Yes
Partially
Partially
Yes
Yes
Yes
Table 4 lists all pollutants for which NMIM produces inventories. The pollutant codes
are those specified by the NIF3.4 Numeric codes are Chemical Abstracts Service Registry
Numbers5 (CASRN) with the hyphens removed.
In Table 4, a non-blank "Ratio to" column (MB for MOBILE6, NR for NONROAD)
indicates that the pollutant is calculated by NMIM, after the MOBILE6 or NONROAD model is
run, by ratio to the pollutant listed in the column. A blank "Ratio to" column indicates that the
pollutant is calculated inside MOBILE6 or NONROAD. The ratio depends on source type,
expressed as a source classification code (SCC), and fuel characteristics. The complete list of
these ratios may be found in the SCC and SCCToxics tables of the NCD. Ratio units are
g/gallon, g/mile, and g/g of PM or VOC. For onroad vehicles, naphthalene is ratioed to exhaust
PM and to evaporative VOC. For nonroad equipment, it is ratioed to exhaust PM10 only. SOA
is present to provide input to REMSAD (Regional Modeling System for Aerosols and
Deposition6).
The "Six HAPS" category represents the first HAPS studied for mobile sources. They
are selected individually in the NMIM RunSpec. The 27 "Add'l. HAPS" (additional HAPS) are
selected as a group in the NMIM RunSpec. The 17 dioxin/furan congeners are also selected as a
group. Pollutants in Table 4 without a category listed may be selected individually in the NMIM
RunSpec. All pollutants are output separately, even if they are selected as a group.
-------
Table 4. List of pollutants for which inventories are produced by NMIM. (See text.)
Pollutant
Code
CO
HC
NOX
SO2
PM10-PRI
PM25-PRI
NH3
75070
107028
71432
106990
50000
1634044
100414
100425
108883
110543
120127
123386
129000
1330207
16065831
18540299
191242
193395
205992
206440
207089
208968
218019
50328
53703
540841
56553
7439965
7440020
83329
85018
86737
91203
1746016
PollutantName
Carbon Monoxide
Hydrocarbons (choice of five forms)
Nitrogen Oxides
Sulfur Dioxide
Primary PM10 (Filterables and Condensibles)
Primary PM2.5 (Filterables and Condensibles)
Ammonia
Acetaldehyde
Acrolein
Benzene
1,3 -Butadiene
Formaldehyde
MTBE
Ethyl Benzene
Styrene
Toluene
Hexane
Anthracene
Propionaldehyde
Pyrene
Xylene
Chromium (Cr3+)
Chromium (Cr6+)
Benzo(g,h,i)perylene
Indeno(l,2,3,c,d)pyrene
Benzo(b)fluoranthene
Fluoranthene
B enzo(k)fluoranthene
Acenaphthylene
Chrysene
Benzo(a)pyrene
Dibenzo(a,h)anthracene
2,2,4-Trimethylpentane
Benz(a)anthracene
Manganese
Nickel
Acenaphthene
Phenanthrene
Fluorene
Naphthalene
2,3,7,8-Tetrachlorodibenzo-p-Dioxin
Category
Six HAPS
Six HAPS
Six HAPS
Six HAPS
Six HAPS
Six HAPS
Add' . HAPS
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
Add'
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
.HAPS
Dioxin/furan
Ratio to
MB NR
voc
voc
voc
voc
PM10
voc
PM10
VOC
Mile
Mile
PM10
PM10
PM10
PM10
PM10
PM10
PM10
PM10
PM10
VOC
PM10
Mile
Mile
PM10
PM10
PM10
PMVOC
Mile
PM10*
Gal*
VOC
voc
voc
voc
voc
voc
voc
voc
voc
voc
PM10
voc
PM10
voc
Gal
Gal
PM10
PM10
PM10
PM10
PM10
PM10
PM10
PM10
PM10
VOC
PM10
Gal
Gal
PM10
PM10
PM10
PM10
Gal
10
-------
Pollutant
Code
19408743
3268879
35822469
39001020
39227286
40321764
51207319
55673897
57117314
57117416
57117449
57653857
60851345
67562394
70648269
72918219
CO2
SOA
PollutantName
1,2,3, 7,8, 9-Hexachlorodibenzo-p-Dioxin
Octachlorodibenzo-p-dioxin
1,2,3,4,6,7,8-Heptachlorodibenzo-p-Dioxin
Octachlorodibenzofuran
1,2,3,4,7,8-Hexachlorodibenzo-p-Dioxin
1,2,3, 7,8-Pentachlorodibenzo-p-Dioxin
2,3,7,8-Tetrachlorodibenzofuran
1, 2,3,4,7,8, 9-Heptachlorodibenzofuran
2,3,4,7,8-Pentachlorodibenzofuran
1,2,3,7,8-Pentachlorodibenzofuran
1,2,3,6,7,8-Hexachlorodibenzofuran
1,2,3, 6,7,8-Hexachlorodibenzo-p-Dioxin
2,3,4,6,7,8-Hexachlorodibenzofuran
1,2,3,4,6,7,8-Heptachlorodibenzofuran
1,2,3,4,7,8-Hexachlorodibenzofuran
1,2,3, 7,8, 9-Hexachlorodibenzofuran
Carbon Dioxide
Secondary Organic Aerosol
Categorv
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Dioxin/furan
Ratio
MB
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
Mile
voc*
to
NR
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
Gal
VOC*
* Ratios for the pollutants are in the NCD SCC table. All others are in the SCCToxics table.
SOURCE CATEGORIES FOR WHICH INVENTORIES ARE PRODUCED BY NMIM
NMIM's output is always in terms of source classification codes (SCC), which are
described in the SCC table. For onroad output, NMIM also distinguishes five emission types
(exhaust, evaporation, refueling, brake wear, and tire wear). For nonroad output, NMIM
distinguishes three emission types (exhaust, evaporation, and refueling) and also reports the
NONROAD power classes, which subdivide a given SCC by horsepower range.
The VMT in the BaseYearVMT table is by the 28 MOBILE6 vehicle classes. In NMEVI
output, however, these 28 vehicle classes are aggregated into the 12 vehicle classes that
correspond to SCC codes. These 12 vehicle classes are shown in Table 5. The M6VClass table
defines the correspondence between these two sets of vehicle classes, which is shown in Table
5a.
11
-------
Table 5. The twelve vehicle classes that correspond to SCCs.
Class Description
Class Description
LDGV Light duty gasoline vehicles
LDGT1 Light duty gasoline truck 1
LDGT2 Light duty gasoline truck 2
HDGV Heavy duty gasoline vehicles, including
buses
MC Motorcycles
LDDV Light duty diesel vehicles
LDDT Light duty diesel trucks
2BHDD Class 2b heavy duty diesel
V vehicles
LHDDV Light heavy-duty diesel vehicles
MHDDV Medium heavy-duty diesel
vehicles
HHDDV Heavy heavy-duty diesel
vehicles
BUSES Diesel buses
12
-------
Table 5a. The 28 MOBILE6 vehicle classes and the 12 vehicle classes corresponding to SCCs
that are output by NMIM.
M6
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
28 M6
LDGV
LDGT1
LDGT2
LDGT3
LDGT4
HDGV2B
HDGV3
HDGV4
HDGV5
HDGV6
HDGV7
12 SCC
LDGV
LDGT1
LDGT1
LDGT2
LDGT2
HDGV
HDGV
HDGV
HDGV
HDGV
HDGV
HDGV8AHDGV
HDGV8B
LDDV
LDDT12
HDDV2B
HDDV3
HDDV4
HDDV5
HDDV6
HDDV7
HDGV
LDDV
LDDT
2BHDD
V
LHDDV
LHDDV
LHDDV
MHDDV
MHDDV
HDDV8AHHDDV
HDDV8B
MC
HDGB
HDDBT
HDDBS
LDDT34
HHDDV
MC
HDGV
BUSES
BUSES
LDDT
Description
Light-Duty Gasoline Vehicles (Passenger Cars)
Light-Duty Gasoline Trucks 1 (0-6,000 Ibs. GVWR, 0-3750 Ibs.
LVW)
Light-Duty Gasoline Trucks 2 (0-6,000 Ibs. GVWR, 3751-5750 Ibs.
LVW)
Light-Duty Gasoline Trucks 3 (6,001-8,500 Ibs. GVWR, 0-5750 Ibs.
ALVW)
Light-Duty Gasoline Trucks 4 (6,001-8,500 Ibs. GVWR, 5751 Ibs.
and greater ALVW)
Class 2b Heavy-Duty Gasoline Vehicles (8501-10,000 Ibs. GVWR)
Class 3 Heavy-Duty Gasoline Vehicles (10,001-14,000 Ibs. GVWR)
Class 4 Heavy-Duty Gasoline Vehicles (14,001-16,000 Ibs. GVWR)
Class 5 Heavy-Duty Gasoline Vehicles (16,001-19,500 Ibs. GVWR)
Class 6 Heavy-Duty Gasoline Vehicles (19,501-26,000 Ibs. GVWR)
Class 7 Heavy-Duty Gasoline Vehicles (26,001-33,000 Ibs. GVWR)
Class 8a Heavy-Duty Gasoline Vehicles (33,001-60,000 Ibs. GVWR)
Class 8b Heavy-Duty Gasoline Vehicles (>60,000 Ibs. GVWR)
Light-Duty Diesel Vehicles (Passenger Cars)
Light-Duty Diesel Trucks 1 and 2 (0-6,000 Ibs. GVWR)
Class 2b Heavy-Duty Diesel Vehicles (8501-10,000 Ibs. GVWR)
Class 3 Heavy-Duty Diesel Vehicles (10,001-14,000 Ibs. GVWR)
Class 4 Heavy-Duty Diesel Vehicles (14,001-16,000 Ibs. GVWR)
Class 5 Heavy-Duty Diesel Vehicles (16,001-19,500 Ibs. GVWR)
Class 6 Heavy-Duty Diesel Vehicles (19,501-26,000 Ibs. GVWR)
Class 7 Heavy-Duty Diesel Vehicles (26,001-33,000 Ibs. GVWR)
Class 8a Heavy-Duty Diesel Vehicles (33,001-60,000 Ibs. GVWR)
Class 8b Heavy-Duty Diesel Vehicles (>60,000 Ibs. GVWR)
Motorcycles (Gasoline)
Gasoline Buses (School, Transit and Urban)
Diesel Transit and Urban Buses
Diesel School Buses
Light-Duty Diesel Trucks 3 and 4 (6,001-8,500 Ibs. GVWR)
SCC output also distinguishes twelve roadway types, listed in Table 6 below. The twelve
roadway types are those used by the Federal Highway Administration's (FHWA) Highway
Performance Monitoring System (HPMS).7 These roadway types, in combination with the
twelve vehicle types, result in 144 SCCs for onroad mobile sources.
13
-------
Table 6. Twelve Roadway Types
Interstate: Rural Interstate: Urban
Other Principal Arterial: Rural Other Freeways and Expressways: Urban
Minor Arterial: Rural Other Principal Arterial: Urban
Major Collector: Rural Minor Arterial: Urban
Minor Collector: Rural Collector: Urban
Local: Rural Local: Urban
Emissions are estimated by the NONROAD Model for 214 SCCs. SCCs distinguish
between equipment types, fuels (gasoline, diesel, liquified petroleum gas (LPG), and compressed
natural gas (CNG)) and between two stroke and four stroke gasoline engines. In addition,
NONROAD produces horsepower categories, and NMIM retains these in its output.
Within NMIM and the NONROAD Model, the nonroad SCCs are grouped into 12
segments, listed in Table 7.
Table 7. NONROAD Model equipment segments.
Recreational
Construction
Industrial
Lawn/Garden
Agriculture
Commercial
Logging
Airport Support
Underground Mining
Oil Field
Pleasure Craft
Railroad
Any single SCC always falls under only one of these segments, corresponding to its most typical
application, although it may be used in other segments as well. For example, skid steer loaders
are in the construction segment, although they are also used in agriculture. Fuels are gasoline,
diesel, LPG, and CNG. NMIM users must choose a segment and fuel; individual SCCs may not
be selected. Output, however, is by individual SCC.
THE NMIM COUNTY DATABASE (NCD)
The NMIM County Database (NCD) contains all the county-specific information needed
to run MOBILE6 and NONROAD. It also contains the list of pollutants and the ratios of HAPS,
dioxins/furans, and some metals to various NONROAD and MOBILE6 outputs that are used to
estimate inventories of these non-standard pollutants. This database is in MySQL, an open
source data base management system that is available from wwioig^qLcom. The tables in the
database are listed in Table 8.
14
-------
Table 8. Tables in the NMIM County Database.
Table
BaseYearVMT
County
CountyMap
CountyMonth
CountyMonthHour
County NRFile
County VMTMonth
Allocation
CountyYear
County YearMonth
County YearMonth
Hour
DataSource
Diesel
EmissionType
FileType
Gasoline
Hour
HPMSRoadType
M6VClass
M6VType
NaturalGas
PollutantCode
sec
SCCToxics
State
VMTGrowth
VMTMonth
Allocation
Contents
VMT bv vear. county. M6VClass. and HPMSRoadType.
For each county, Federal Information Processing System (FIPS) codes for the
county and state, altitude, beginning and end of ozone season, Stage 2
information. Natural Gas Vehicle (NGV) fraction file name
The representing county for each county, one for NONROAD and one for
MOBILE6.
Defines the set of possible county-month combinations.
Monthly average hourly temperature and humidity table used if "Use yearly
weather data" is not selected or there is no data for the requested year in the
County YearMonthHour table.
References to external NONROAD files pertaining to a county.
Mileage allocation factors for the twelve months of the year, by county.
Stage2 percent input to the NR model, plus external file references for
MOBILE6 and NR.
Gasoline, diesel, and natural gas fuel IDs for each county for each year and
month.
Historical hourly temperature and relative humidity.
Defines datasource identifiers used in other tables.
Diesel sulfur content associated with each diesel ID.
Associates emission types (exhaust, evap, brake, tire) with EmissionTypelD
used in other tables
Defines the set of valid external files and their 3-character extensions.
Detailed fuel properties associated with each gasoline ID.
Defines the hour identifiers.
Defines the twelve HPMS road type identifiers.
Defines the 28 vehicle classes used in MOBILE6. These are the valid
combinations of M6Vtype and fuel.
Defines the fuel-independent vehicle types used in MOBILE6.
Natural gas sulfur content associated with each natural gas ID.
Associates NIF pollutant codes and pollutant names with PollutantCodelD used
in other tables.
Associates with each SCCID an SCC code and description, and ratios for NH3,
PM25, and for converting between HC forms.
SCC and fuel property-dependent ratios for calculating HAPS, dioxin/furans,
and metals.
Associates state names and abbreviations with state FIPS codes used in other
tables.
The annual VMT growth rate for a M6VClass by county and year.
Factors for allocating annual VMT to the 12 months, by M6VType and
HPMSRoadType, used if there are no county-specific values in
County VMTMonthAllocation .
The main emphasis in putting together the database has been on historical years, for
which data are available. The best year as of this writing is 2002, since states and local
15
-------
organizations submitted corrected information as part of the NEI process. However, future years
are of interest for rulemaking and other purposes. For future years, refinery modeling of the
impacts of anticipated fuel changes, such as phase-in of low sulfur fuels, is used to develop
projected fuel properties. Items such as age and speed distributions are assumed to remain
unchanged from the most recent historical year, unless future estimates for such data are
submitted by states. NMEVI uses 20-year average temperature and humidity data from the
CountyMonthHour table for future years. Data for all years from 1999 to 2050 is present in the
NCD, except for VMT. Users wishing to generate a future year onroad inventory may have to
provide their own VMT data and insert it into the BaseYearVMT table of the NCD. For future
year nonroad inventories, the necessary estimates (e.g., of population and activity) are contained
in the data files associated with the NONROAD Model.
Temperature and Humidity
The NCD contains monthly average hourly temperature and humidity for historical years
in the County YearMonthHour table. It contains 20-year averages (1981 to 2000) of monthly
average hourly temperature and humidity in the CountyMonthHour table. The user can select
whether to use the historical or 20-year average data. When historical data are not present in the
database, the program automatically uses the 20-year average.
Because minimum and maximum temperatures occur at different hours each day, the
minimum of the hourly averages will be higher than the average of the daily minima, and the
maximum of the hourly averages will be lower than the average of the daily maxima. To avoid
this narrowing of the daily temperature range, the monthly average of hourly temperatures is
assumed to capture the daily temporal pattern and is mathematically stretched so that the low
temperature equals the monthly average of the daily minima and the high temperature equals the
monthly average of the daily maxima, producing a set of "adjusted monthly average hourly
temperatures." The same procedure is applied to dewpoint. An adjusted monthly average hourly
relative humidity is then calculated from the adjusted monthly average hourly temperatures and
dewpoints. It is these adjusted hourly temperatures and relative humidities that are found in the
County YearMonthHour table of the NCD. A similar procedure was used to generate the 20-year
average temperatures and humidities in the CountyMonthHour table.
The historical and 20-year average hourly temperatures and humidities were developed
by Air Improvement Resources (AIR), Inc. for EPA under contract, from data obtained from the
National Climatic Data Center. County temperatures were determined by weighting nearby
temperature stations by their distance from the population-based centroid of each county. For
Puerto Rico and the Virgin Islands, averages were calculated using the five calendar years 1999
through 2003.
MOBILE6 uses hourly temperature and humidity directly from the
County YearMonthHour table if the data are available and if the user has selected "Use Yearly
Weather Data." Otherwise, 30-year average data are used from the CountyMonthHour table.
NONROAD does not use hourly temperatures, but rather daily minimum, maximum and average
16
-------
temperatures. For NONROAD, NMIM selects the minimum and maximum from all 24 hourly
temperatures, but calculates the average from the hourly temperatures from 6 AM to 9 PM
(hourly intervals 7 to 21). Humidity is not an input for NONROAD.
Fuel Properties
The County YearMonth table contains gasoline, diesel, and natural gas fuel identifiers for
each county for each month of each calendar year. Onroad vehicles and nonroad equipment
have separate fuel identifiers. The identifiers are associated with fuel properties in the Gasoline,
Diesel, and NaturalGas tables.
For gasoline, the fuel properties were determined by the Eastern Research Group, Inc.,
(ERG) under contract to EPA using gasoline survey data.8 The fuel properties stored in the
Gasoline table are:
o Average Reid Vapor Pressure (RVP)
o Average sulfur content
o Maximum sulfur content
o Ethanol (ETOH) percent by volume of ethanol blended gasolines
o Ethanol blend market share
o MTBE (methyl tertiary butyl ether) percent by volume of ether blended gasolines
o MTBE blend market share
o ETBE (ethyl tertiary butyl ether) percent by volume of ether blended gasolines
o ETBE blend market share
o TAME (tertiary amyl methyl ether) percent by volume of ether blended gasolines
o TAME blend market share
o Aromatic content
o Olefm content
o Benzene content
o E200 (vapor percentage of gasoline at 200 degrees Fahrenheit)
o E300 (vapor percentage of gasoline at 300 degrees Fahrenheit)
The gasoline properties were derived from several surveys:
o EPA's Reformulated Gasoline survey, 20009
o EPA's Oxygenated fuel program summary, 200110
o TRW (previously NIPER) fuel survey, 199911
o Alliance of Automobile Manufacturers of America (AAMA) survey, 199912
Market share for oxygenated gasolines was obtained from the EPA Oxygenate Type
Analysis Tables13 and the Federal Highway Administration website.14
If the methyl tertiary butyl ether (MTBE) percent volume content was less than 0.1
percent, MTBE content was set to be zero with a zero percent market share. Similarly, if the
17
-------
ethanol percent volume content was less than 0.1 percent, ethanol content was set to be zero with
a zero percent market share. For any area which reported non-zero values for tertiary amyl
methyl ether (TAME) or ethyl tertiary butyl ether (ETBE), the entire market share was set to be
MTBE, since it was not possible to distinguish the market share between these specific
oxygenates.
All gasoline properties are area-wide averages, except for oxygenates, which are allowed
to have market shares. Three fuels (winter, summer, and spring/fall) are determined for each
county and assigned to months by season. Months representing seasons vary by location.
Spring/fall gasoline properties are derived from summer and winter fuels by interpolation.
Gasolines in Puerto Rico and the Virgin Islands were assumed to be similar to gasolines in
Collier County, Florida. Details of how the fuel survey data were applied to individual states
and counties are described in the ERG report to EPA.15 The NCD allows counties to have
different NR and onroad gasolines, but none do as of this writing, because data have not been
available to make this distinction.
For diesel fuel and natural gas, the only fuel property stored is sulfur content, in the
Diesel and NaturalGas tables, respectively. A sulfur content of 30 ppm for natural gas is used
for all counties in all states for all years and months. Nonroad and onroad diesel sulfur values
tend to differ in the same county because they are subject to different regulations. In addition,
there are two kinds of nonroad diesel fuel: that used for pleasure craft and that used for all other
nonroad segments.
Vehicle Miles Traveled (VMT)
VMT is the critical activity parameter for converting MOBILE6's emission factors into
inventories. Historical VMT is derived from data from the Federal Highway Administration
(FHWA). Unfortunately, these data are not gathered at the spatial and temporal scales needed by
NMEVI, nor for the same vehicle classes required by MOBILE6. The procedures for gathering
these data and processing them into the required scales and vehicle classes are described in detail
in the NEI documentation.16 States and local organizations may submit corrections to the VMT
data.
Database Improvement
As part of the NEI process, States and regional and local organizations are asked to
review the NCD and provide more accurate data, if available. This review includes the county
data and also the external files that can be referenced by MOBILE6 or NONROAD, such as fleet
distributions and equipment allocations. The development of this database improvement system
promises to extend NMIM's advantages beyond convenience to include the production of better
mobile source inventories. Procedures for submitting improved data for inclusion in the NCD
are contained in a technical memorandum posted with the Draft 2002 NEI.17 This document is
also useful as a comprehensive description of what county-specific data can go into the NCD.
Considerable data has been received as part of the 2002 NEI process.
18
-------
ESTIMATION OF HAZARDOUS AIR POLLUTANT (HAP) EMISSIONS
The hazardous air pollutants for which inventories are produced by NMIM are listed in
Table 4 above. The "six HAPS" are produced internally by MOBILE6. In all other cases, as
indicated in the "Pollutants" section above, HAP inventories are generated by ratios to various
MOBILE6 and NONROAD outputs.
HAPS are estimated using data sources and methods developed for the 1999 NEI for
HAPS, version 3,18'19 with some modifications, described below. NMIM does not estimate HAP
emissions for compressed natural gas (CNG) engines.
HAPS are estimated in NMIM using one of three approaches:
1) Gaseous HAPS - Apply toxic to VOC ratios to VOC estimates.
2) Poly-Aromatic Hydrocarbons (PAHs) - Apply toxic to PM10 ratios to PM10 estimates.
3) Metals, Dioxins and Furans - For NONROAD, multiply HAP gram per gallon emission
factors by county level fuel consumption estimates. For MOBILE6, multiply HAP gram
per mile emission factors by county level VMT estimates.
The NCD SCCToxics table provides a complete listing of toxic ratios and emission factors for all
SCCs and fuel combinations. The above approaches are described in more detail in the
following sections.
Gaseous HAPS
NMEVI uses the toxic to VOC ratios described in the documentation for the 1999 NEI for
HAPS, version 3, and summarized in Volume 1, Appendix D, Table 1 of the documentation.
Separate ratios are used for evaporative and exhaust emissions for each of the following four
categories of gasoline blends:
1. Baseline Gasoline. All cases that do not fall into categories 2-4 below. Ratios are in
variables "ExhBaseGas" and "EvapBaseGas" in the SCCToxics table.
2. WO (Winter Oxygenate) Gasoline / Ethanol or ETBE - Used where the fuel contains
ethanol which is greater than or equal to 5% by volume or ETBE greater than or equal to
5% by volume. Ratios are in variables "ExhEthGas" and "EvapEthGas" in the
SCCToxics table.
3. WO (Winter Oxygenate) Gasoline / MTBE / TAME - Used where the fuel contains
MTBE which is greater than or equal to 12% by volume or TAME greater than or equal
to 13% by volume. Ratios are in variables "ExhMTBEGas" and "EvapMTBEGas" in the
SCCToxics table.
19
-------
4. RFG/MTBE/TAME - Used where the fuel is RFG and where the fuel contains oxygenate
greater than 5% by volume and where the fuel contains MTBE which is less than 12% by
volume or TAME less than 13% by volume. Ratios are in variables "ExhRFGGas" and
"EvapRFGGas" in the SCCToxics table.
In many cases, HAP profiles for specific nonroad equipment and engine type
combinations are available. However, for some equipment/engine type combinations, no
speciation data are available. In such instances, default values for 2-stroke gasoline engines, 4-
stroke gasoline engines, and diesel engines are used. These default values represent an average
fraction for various equipment types within an engine category.
Polyaromatic Hydrocarbons (PAHs)
All PAHs emitted in exhaust are estimated as fractions of PM10, although the data used
to calculate mass ratios include both gas and particle phase PAH emissions. The data used to
develop the PAH fractions are described in the documentation for the 1999 NEI for HAPS.
Evaporative naphthalene emissions from onroad vehicles are estimated as a fraction of VOC.
NMEVI does not currently estimate evaporative naphthalene emissions for nonroad equipment.
Metals, Dioxins, and Furans
Onroad
NMEVI estimates onroad emissions using g/mile emission factors developed for the 1999
NEI for HAPS, version 3.
Nonroad
Metals
The approach used by NMIM to estimate nonroad county level metal emissions differs in
a number of respects from the approach used in the 1999 NEI for HAPs, version 3. In the 1999
NEI, nationwide metal emissions for gasoline engines were obtained by applying a mass per
gallon emission factor by nationwide gasoline consumption from the NONROAD model. For
diesel engines, a mass per brake-horsepower hour emission factor was multiplied by nationwide
energy output. The resultant nationwide emission estimates were then spatially allocated to
counties relative to the county proportion of PM10 emissions compared to the national PM10
emissions, as obtained from the NONROAD model.
In contrast, NMIM multiplies mass per gallon emission factors for both gasoline and
diesel engines by county level fuel consumption in gallons to obtain a county level inventory
estimate. For diesel engines, mass per brake-horsepower hour emission factors were converted
to mass per gallon emission factors using the following equation:
20
-------
grams per gallon = grams per brake-horsepower hour (g/bhphr) * average fuel density
(Ib/gallon) / fuel consumption per brake-horsepower hour (Ib/bhphr)
where:
average fuel density = 7.01 Ib/gal
fuel consumption per brake horsepower hour = 0.408 Ib for engines less than 100 hp
= 0.367 Ib for engines greater than 100 hp
The fuel consumption per brake horsepower hour estimates are from the NONROAD model.20
NMIM looks up the gram per gallon metal emission factors for gasoline and diesel engines in the
SCCToxics table of the NCD.
Dioxins/furans
Mass per gallon emission factors for dioxins and furans from nonroad engines were
calculated by multiplying the onroad vehicle emission factors in grams per mile by fleet average
fuel economy estimates. The assumed fuel economy for gasoline vehicles was 21.5 miles per
gallon; for diesel vehicles it was 7 miles per gallon. Resulting gram per gallon emission factors
are contained in the NCD SCCToxics table.
RETROFIT AND FLEET MODELING
NMEVI models user-specified retrofit programs by post-processing the MOBILE6 and/or
the NONROAD by-model-year output files. The user requests this function via the GUI or
RunSpec. The user supplies the parameters of the retrofit program in a comma-delimited
external text file. Multiple programs can be specified in a single file. Only one retrofit file can
be provided for each NMIM run. The retrofit programs defined in that file are applied to all
counties and months in the NMIM run. Modeling different retrofit programs for different
counties or groups of counties requires separate NMIM runs. For onroad retrofit programs, the
user specifies pollutant, vehicle class, start and end calendar years of the retrofit program, the
initial and final model years to which retrofits will be applied, the percentage of the fleet that is
retrofit in each year of the program, and the percentage effectiveness of the retrofit. For nonroad
retrofit programs, the user specifies pollutant, SCC, horsepower range, technology type
(optional), start and end calendar years of the retrofit program, the initial and final model years
to which retrofits will be applied, the percentage of the fleet that is retrofit in each year of the
program, and the percentage effectiveness of the retrofit. The retrofit algorithm reduces the
emissions of each specified model year by the cumulative percentage retrofit (capped at 100%)
times the percentage retrofit effectiveness. Detailed instructions for retrofit modeling are in the
"NMIM User's Guide," which is under the Help menu of the NMIM GUI.
NMIM models user-specified fleets by post-processing the MOBILE6 and/or the
NONROAD by-model-year output files. The user requests this function via the GUI or
21
-------
RunSpec. The user supplies the parameters of the specified fleet in a comma-delimited external
text file. Multiple fleets can be specified in a single file. All the fleets specified in a single fleet
parameter file are assumed to exist in a single county. The temperatures, fuels, I/M programs,
etc. for that county are obtained from the NCD and applied to the emissions calculation for the
fleet. For onroad fleets, the user specifies vehicle class, model year, number of vehicles, and
annual mileage. For nonroad fleets, the user specifies SCC, horsepower range, model year,
number of pieces of equipment, and optionally technology type, annual activity (hours of
operation per year), and monthly activity distribution. The fleet algorithm multiplies the
emission factors derived from MOBILE6 or NONROAD by the activity of the fleet. Detailed
instructions for fleet modeling are in the "NMIM User's Guide," which is under the Help menu
oftheNMIMGUI.
SUMMARY
NMEVI is a consolidated modeling system for MOBILE6 and NONROAD designed to
produce national inventories. A standard input configuration provides consistency between
MOBILE6 and NONROAD and between criteria pollutants and HAPS. NMIM's run
specification can be constructed from its GUI or by a text editor. It can be executed from the
GUI or from a command line. It is capable of standalone operation and can also take advantage
of multiple computers on a LAN through a simple distributed processing system. Post-
processing capabilities include the production of output in NIF3 format. NMIM can also model
emissions from user-designed retrofit programs and user-specified fleets. The inclusion of a
database improvement system in the NEI process promises to extend NMIM's advantages
beyond convenience to the production of better mobile source inventories.
22
-------
REFERENCES
1 . This document is an update of a paper that was presented at the International Emission
Inventory Conference (IEIC) at Clearwater, Florida in June 2004. At that conference, a training
was given using source code NMEVI200405 12 and database County20040508. Since that time,
bugs have been fixed, capabilities have been expanded, and the database has been improved.
2. For a detailed explanation of how emissions were calculated for aircraft, commercial marine
vessels, and locomotives for the Draft 2002 NEI, see
ftp://ftp.epa.gov/EmisInventory/draftnei2002/mobile/nonroad/documentation/
3 . User 's Guide to MOBILE6. 1 andMOBILE6. 2: Mobile Source Emission Factor Model., U. S.
Environmental Protection Agency, Ann Arbor, MI, August 2003; EPA420-R-03-010. May be
found as a link to http://www.epa.gov/otaq/m6.htm
4. Details of the NIF3.0 format may be found as links to
http ://www. epa. gov/ttn/chief/nif/index.html#ver3 .
5. See EPA's Substance Registry System: jM^Jlwww^ja^effv/SKJ and the CAS Registry
website: http://www.cas.org/EO/regsys.html.
6. Information on REMSAD may be found at http://remsad.saintl.com/overview.htm.
7. Information on the Federal Highway Administration's (FHWA) Highway Performance
Monitoring System (HPMS) is available at http://www.fhwa.dot.gov/policy/ohpi/hpms/.
8. "National Mobile Inventory Model (NMIM) Base and Future Year County Database
Documentation and Quality Assurance Procedures," prepared by Eastern Research Group, Inc.,
for U.S. Environmental Protection Agency, Office of Transportation and Air Quality, Ann
Arbor, MI, 2003.
9. "Reformulated Gasoline Survey Data for 2000," U.S. Environmental Protection Agency,
Office of Transportation and Air Quality, Ann Arbor, Michigan, 2000 Internet address:
http://www.epa.gov/otaqyconsumer/fuels/mtbe/oxy-95-00.pdf
10. "U.S. EPA Oxygenated Fuel Program Summary, 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, Michigan. October 2001. Internet
address: htti^/www^epa,^^
1 1 . Thompson, Ramo, and Woolridge (TRW) or National Institute for Petroleum and Energy
Research (NIPER) Fuel Survey, 1999.
12. "North American Gasoline and Diesel Fuel Survey," Alliance of Automobile Manufacturers,
1999.
23
-------
13. "U.S. EPA Oxygenated Fuel Program Summary, 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, Michigan. October 2001. Internet
address: htt|3^/wwwj^^^
14. Federal Highway Administration (FHWA) website for oxygenated fuel sale
percentage. Table MF-33E - Estimated Use of Gasohol and Table MF-21 - Motor-Fuel Use,
1999. Internet address: htt|L^Zww\^^
15. "National Mobile Inventory Model (NMIM) Base and Future Year County Database
Documentation and Quality Assurance Procedures," prepared by Eastern Research Group for
U.S. Environmental Protection Agency, Office of Transportation and Air Quality, May 2003,
(EPA contract No. 68-COO-l 12, Work Assignment 3-05).
16. "Documentation for the Onroad National Emissions Inventory (NEI) For Base Years 1970-
2002," prepared for: Office of Air Quality Planning and Standards, Emission Factor and
Inventory Group, U.S. Environmental Protection Agency, Research Triangle Park, NC 2771 1,
prepared by: E.H. Pechan & Associates, Inc., 5528-B Hempstead Way, Springfield, VA 22151,
January 2004.
17. "Instructions to State and Local Agencies for Updating the County -Level Database from
EPA's National Mobile Inventory Model: Technical Memorandum," prepared by Maureen
Mullen, E.H. Pechan and Associates, Springfield, VA, for Emission Factors and Inventory
Group, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC 2771 1, November, 2004. It is available on the web at
ftp: //ftp. epa.gov/EmisInventory /draftnei2002/mobile/nmim_related/nmim_technical_memorandu
m_1104.pdf.
18. Documentation for the Onroad National Emissions Inventory (NEI) For Base Years 1970-
2002 Prepared for: Office of Air Quality Planning and Standards Emission Factor and Inventory
Group U.S. Environmental Protection Agency Research Triangle Park, NC 2771 1. Prepared by:
E.H. Pechan & Associates, Inc., 5528-B Hempstead Way, Springfield, VA 22151, January 2004.
http://www.epa.gov/ttii/chief/iiet/1999inveiitory.htmltffiiial3haps
19. "Documentation for Aircraft, Commercial Marine Vessel, Locomotive, and Other Nonroad
Components in the National Emissions Inventory, Volumes I and II," prepared by Eastern
Research Group for U.S. Environmental Protection Agency, October 7, 2003. Available from
http://www.epa.gov/ttn/chief/net/1999iiiventory.htmltffinal3haps
20. Exhaust and Crankcase Emission Factors for Nonroad Engine Modeling — Compression-
Ignition, U.S. Environmental Protection Agency, Office of Transportation and Air Quality,
Report No. EPA420-P-02-016, November 2002. May be found on the web at
http://www.epa.gov/otaq/nonrdmdl.htmtfdocs.
24
------- |