EPA420-S-99-001
MOBILE6
A Revised Model for Estimation of Highway Vehicle Emissions
Presented at the
Air & Waste Management Association
Eighth Annual Conference on Emission Inventories
Emission Inventory: Living in a Global Environment
December 9, 1998
New Orleans, Louisiana
Prepared by
David J. Brzezinski and Terry P. Newell
US EPA Office of Mobile Sources
Assessment and Modeling Division
2000 Traverwood Drive
Ann Arbor MI 48105
Abstract
The Environmental Protection Agency (EPA) highway vehicle emission factor model
provides average in-use fleet emission factors for three criteria pollutants [volatile organic
compounds (VOC), a precursor of ground-level ozone; carbon monoxide (CO); and oxides of
nitrogen (NOx)], for gas and diesel cars and light- and heavy-duty trucks, and motorcycles, for
calendar years between 1970 and 2050, under various conditions affecting in-use emission levels
(e.g., ambient temperatures, average traffic speeds) as specified by the modeler. It is used by
EPA in evaluating highway mobile source control strategies, by States (except California) and
local and regional planning agencies in developing emission inventories and control strategies for
State Implementation Plans (SIPs) under the Clean Air Act, and in the development of
environmental impact statements.
EPA is now in the process of revising the MOBILE model. MOBILE6, currently
scheduled for release in December 1999, will differ significantly in both structure and data
requirements from current versions of the model (MOBILESa and 5b). MOBILE6 will
incorporate updated basic emission rates, off-cycle ("real world") driving patterns and emissions,
separation of start and running emissions, improved correction factors, and updated fleet
information. It will also include impacts of recently promulgated regulations not included in
MOBILESa, and will provide improved input and output features.
AWMA December 9, 1998 New Orleans, Louisiana
-------
Introduction
Under requirements of the Clean Air Act (CAA) amendments of 1990, Section 1301, EPA
(as represented by the Administrator) is required to "...review and, if necessary, revise, the
methods ("emission factors") use for purposes of this Act to estimate the quantity of emissions of
carbon monoxide, volatile organic compounds, and oxides of nitrogen from sources of such air
pollutants (including area sources and mobile sources)." For the case of highway vehicles, these
emission factors are calculated using a computer model; the most recent major update to this
model, MOBILESa, was released for use in March 1993. Since that time, two additional updates
to this model have been released, MOBILE5a_H and 5b; the latter were released in August 1997.
EPA's Office of Mobile Sources (OMS) is now working on a major update to this model. The
new version, MOBILE6, is currently planned for final release in December 1999.
In addition to the need for updating the model's underlying emission test data to reflect
testing and analyses that have been performed since MOBILESb, OMS is aware that the output
of the existing model is not always well suited to the needs of the large and diverse audience for
the model. In particular, there is a clear need for better integration of emission factor and air
quality modeling with transportation planning and analyses. One example of this is the need for
emission factors representing time periods less than a full day (such as the need for hourly
emission estimates for use in Urban Airshed Modeling), and for emission factors that better
represent variations in traffic flow patterns (and hence emission rates) across various roadway (or
facility) types. As discussed below, OMS is making significant changes to the model structure
and output to address such concerns.
Finally, OMS is developing this revision to the MOBILE model using a much more
intensive process of obtaining outside review during development. Among the steps being taken
is the extensive provision of opportunities for stakeholder review of various components of the
model and underlying analyses at an earlier stage of the process, and obtaining formal peer
review of such components and analyses where appropriate. The reader is referred to the OMS
Home Page on the World Wide Web (http://www.epa.gov/oms) for additional information on the
development of MOBILE6, including information on the review processes being utilized to
ensure that all components of the pending model revision are widely disseminated.
The following sections describe the updates and revisions being made to the MOBILE
model, the types and sources of data being used in these updates (where applicable), changes to
the input data requirements and output report format and content, and the stakeholder and peer
review processes being used in the development of MOBILE6. There is still much work to be
done in most areas of the model; thus, it is not possible at this time to provide "bottom line"
answers as to exactly what impacts these changes will have on highway vehicle emission factor
estimates and emission inventories developed from those factors. Where possible, indications
are provided as to the direction and approximate magnitude of each revision. Only after more
work has been completed and reviewed will it be possible to provide an estimate of the impact on
emission estimates of all of the proposed changes in the aggregate.
AWMA December 9, 1998 2 New Orleans, Louisiana
-------
As part of the development of MOBILE6, OMS has established a section within its web
site (http://www.epa.oms/oms) that includes drafts of analyses and papers that have been
prepared. These reports are posted for outside (stakeholder) review and comment. That section
of the web site is accessed by selecting "MOBILE6" from the choices available at the OMS home
page (web address shown above). Papers cited having report numbers of the form M6.xxx.yyy,
where "xxx" represents a topical area (e.g., "EXH" represents papers dealing with analysis of
exhaust emissions, "EVP" represents papers dealing with analysis of evaporative emissions, and
so forth) and "yyy" indicates a number within that series (001, 002, etc.), can be accessed at the
OMS web site. OMS also provides a list server (an electronic e-mail mailing list) to announce
the posting of new reports and other mobile source modeling developments. The OMS web site
contains information on how to subscribe to the EPAMOBILENEWS list server.
Discussion
Emission data updates
Reevaluation of in-use deterioration (for light-duty vehicles and light-duty trucks) and
update of basic emission rate equations (for light-duty vehicles, light-duty trucks, and
heavy-duty vehicles)
EPA is nearing completion of a major reassessment of the magnitude of in-use
deterioration for light-duty vehicles (LDVs) and light-duty trucks (LDTs). "Deterioration" here
is defined as the rate of increase in emission rates as a combination of emission from both
properly maintained, non-tampered vehicles and high emitting vehicles as a function of
accumulated vehicle mileage. Basic emission rate equations in the MOBILE model express
emissions for such vehicles, specific to unique combinations of vehicle type/model
year/pollutant, as a "zero-mile" level (expressed in grams per mile, or g/mi) and one or more
rates of deterioration [expressed in g/mi per 10,000 miles accumulated mileage, or (g/mi)/10K
mi]. The deterioration rates used in MOBILES were substantially higher than those used in the
preceding MOBILE4.1 model, as evidenced by the test data collected and analyzed in the late
1980s and early 1990s. However, since the release of MOBILE4.1, evidence that these
deterioration rates were not applicable to newer (post-1990 model year) LDVs and LDTs began
to appear from a variety of sources.
This reevaluation of in-use deterioration rates has been performed under the advice and
guidance of the Mobile Source Technical Advisory Subcommittee of the Clean Air Act Advisory
Committee, established under the Federal Advisory Committee Act (FACA). The results of this
work have not yet been approved for final release and use in the model; however, the general
trend is clear that in-use deterioration from newer vehicles (approximately model year 1990 and
newer) will be less than had been modeled in MOBILES. This change is likely to result in lower
emission factors for all pollutants from LDVs and LDTs, for all calendar years from 1991
forward. This change is also likely to have the effect of reducing the benefits available from the
imposition of in-use inspection and maintenance (I/M) programs.
AWMA December 9, 1998 3 New Orleans, Louisiana
-------
However, this change is not occurring in isolation; a large number of other revisions and
updates, as briefly described in the following sections, are also being implemented for
MOBILE6. At this time, it is impossible to quantify the overall impact of all of the changes
being made on the emission factors calculated by the model.
Off-cycle effects
The Federal Test Procedure (FTP) has been used as the basic test cycle for emissions
measurements (including certification of new vehicles to applicable exhaust standards and
characterization of emissions from in-use vehicles) since the early 1970s. Among the most
important shortcomings of the FTP are the limits on speed and acceleration rates. At the time
that the FTP was developed, limitations on dynamometers prevented the inclusion of vehicle
speeds higher than about 60 mph, or of acceleration rates greater than 3.4 mph/sec. Particularly
since the repeal of the national 55 mph speed limit, a significant fraction of total vehicle miles
traveled (VMT) are accumulated at speeds greater than 55 mph, and acceleration rates well over
3.4 mph/s (up to at least 8 mph/s) are not uncommon. Both of these have a significant impact on
emissions. Driving behavior not fully captured by the FTP is referred to as off-cycle (or non-
FTP) driving.
EPA has developed a new test procedure (the Supplemental FTP, or SFTP) in response to
requirements of the 1990 Clean Air Act Amendments, which includes vehicle operation at higher
speeds and higher acceleration rates. Future model years will be required to meet emission
standards using the SFTP and will be required to mitigate the emission impacts of off-cycle
driving behavior.
Related to off-cycle driving is the impact of vehicle air conditioning (A/C) on emission
rates. Historically, EPA has simulated the impacts of A/C during emission testing by increasing
the load on the vehicle. More recently, testing has shown that these simulations do not
adequately capture the true impacts of A/C use on emissions. When vehicles are tested on a
chassis dynamometer with the A/C actually "on" in an environmentally controlled testing cell,
the emissions impacts observed have been considerably greater than those seen by simulating the
impact of A/C on emissions by the added load.
MOBILE6 is being developed to account for the impacts of driving behaviors not
represented by the FTP (off-cycle), the impacts of newer vehicles being certified using the SFTP,
and the impacts of A/C on emissions. While much work remains to be done, the broad outlines
of the impacts of these changes can be characterized. Special emission testing was done using a
variety of new driving cycles that include off-cycle behavior. Inclusion of non-FTP ("off-cycle"
or "real-world") driving patterns will result in a significant increase, relative to MOBILES
estimates, for all pollutants. This effect will be most pronounced for evaluation years in the
1990s. As new vehicles are certified using the SFTP, beginning in the 2001 model year, this
impact will be greatly reduced; however, only after 25 model years (the window of model years
in operation that is used in the MOBILE model's emission factor calculations) have passed after
AWMA December 9, 1998 4 New Orleans, Louisiana
-------
implementation of certification using the SFTP will the emission factor increases associated with
modeling such off-cycle driving be completely eliminated.
With respect to the modeling of A/C impacts, the reader is referred to the reports "Air
Conditioning Activity Effects in MOBILE6" (M6.ACE.001)2 and "Air Conditioning Correction
Factors in MOBILE6 (M6.ACE.002)3. EPA intends that a number of parameters affecting the
magnitude of the A/C effect on emission factors will be under the control of the modeler through
specifications in the input data file. The testing used to develop these emission effects included
extreme ambient conditions (i.e., very high temperatures, humidity levels, full solar load)
intended to capture the impact of all vehicles equipped with A/C operating full time. User
control of these parameters will allow the impacts on emissions of A/C use to be "scaled down"
to lesser levels, representative of less extreme ambient conditions. Since practical constraints
prohibited the testing of vehicles over a range of temperatures, humidity levels, and solar loads,
both with and without the A/C in operation, an approach to modeling the effects of A/C at
intermediate conditions is developed and presented in the above paper.
Update fuel effects on emissions
MOBILES includes the impacts on in-use emissions of several properties of gasoline:
volatility (as measured by Reid vapor pressure), oxygenate content (for both alcohol/gasoline and
ether/gasoline blends), and sulfur level. Such modeling is required in order to account for the
fact that new vehicle certification, and most in-use emission factor testing (excluding that aimed
at characterization of fuel property impacts on emissions), is performed using specially
formulated fuel that meets certain requirements. Over the past few model updates, EPA has
incorporated volatility (RVP) effects on both exhaust and evaporative emissions into the model
and provided the model user with the ability to specify in-use RVP levels. MOBILES also
provided for modeler control of fuel oxygen content and market share. In addition, EPA adjusted
the basic emission rates for all gasoline-fueled vehicle types to account for other fuel properties
(such as sulfur content) in the industry average fuel that differed from the EPA test fuel
specifications. However, no user control of sulfur levels or other fuel parameters was included
for MOBILES or earlier versions of the model.
The primary updates to the effect of fuels planned for MOBILE6 are in the areas of
oxygenate and sulfur effects. (The only change in the volatility corrections is the separation of
RVP effects for start and for running emissions, as noted in the section on that topic below.) For
oxygenates, the user of the model specifies the market share (fraction of all gasoline consumed in
the area being modeled) of alcohol blend fuels and of ether blend fuels, and the average oxygen
content (in weight percent) of each of these oxygenated fuel types. This user control will be
maintained in MOBILE6; however, on the basis of new test data, the impact of fuel oxygen
content on carbon monoxide (CO) emissions is being reduced, relative to that seen in
MOBILESa. Additional information on the data used for this analysis and the resulting changes
in the modeled impact of oxygenated fuels can be found at the MOBILE6 section of the OMS
web site, in the paper "Fuel Oxygen Effects on Exhaust CO Emissions"(M6.FUL.002)4.
AWMA December 9, 1998 5 New Orleans, Louisiana
-------
Considerable interest in the impact of fuel sulfur levels on in-use emissions has
developed in the last few years, as emission standards continue to be tightened and the wide
variation in sulfur levels in different parts of the country are better known. MOBILES contains
an average correction factor for sulfur content, increasing emission factors slightly to reflect the
nationwide average sulfur content of 339 ppm relative to the test fuel maximum sulfur content of
50 ppm. In response to demand from States and other parties who need to have this impact better
characterized, and to be in a position to evaluate the potential benefits of sulfur control
regulations, OMS is planning to add the capability for the user to input average fuel sulfur
content data for MOBILE6. This will enable the modeler to specify an in-use sulfur level, within
the limits of the model, and have the resulting impact reflected in the emission factors calculated
by the model. These analyses have not yet completed internal review and are not yet available at
the OMS web site. The range of sulfur levels that the user of MOBILE6 will be able to specify,
while subject to change, is likely to be from 30 to 600 ppm.
Roadway (facility) type and average speed correction factors
MOBILES and previous versions of the model have based emission factor estimates on a
single value of "average" speed. The average speed of the FTP driving cycle (the LA4 at 19.6
mph), provided the basic emission rates for MOBILES and was considered to be "uncorrected"
for speed. When the modeler indicated an average speed other than 19.6 mph, the base emission
factors were corrected to that speed. The range of permissible average speeds in MOBILES is
2.5 to 65 mph. However, different roadways (facility types in transportation terminology) can
have very different patterns, in terms of frequency and magnitude of accelerations (for example),
while still maintaining the same average speed. Consider the case of "average speed" of 25 mph
on a roadway with a posted limit of 30 mph, and on a freeway with a posted limit of 65 mph; in
the first case, if traffic is averaging 25 mph, it is essentially free flowing (except for signals,
stops, and the like), while in the second case, traffic is very congested, likely includes stop-and-
go driving, and may include brief excursions to speeds nearing 65 mph.
While the available data do not permit OMS to resolve this issue for all speeds on all
facility types, we are taking the first major steps toward incorporating facility types into the
emission factors produced by the model. This will provide more realistic emissions estimates for
two major facility types (arterials and freeways) on which much VMT is accumulated, and will
provide a significant step toward better integration of emission factor and traffic/transportation
models.
MOBILE6 as now planned will include an emission adjustment for "local roadways" and
for freeway on/off ramps, and will enable the modeler to specify (within limits) the average
speed to be assumed for arterials and freeways. Area-wide emission factors for each pollutant
and vehicle type will be calculated as a weighted sum of the four facility-specific emission
factors (local/collectors, arterials, freeways, and ramps), with the user able to control the
weighting factors used or using a national average set of weighting factors included within the
model. The emission factors for the various facility types, and (in the cases of arterials and
AWMA December 9, 1998 6 New Orleans, Louisiana
-------
freeways) at various average speeds, are based on vehicle testing over a series of facility cycles
intended to better represent driving patterns actually observed on each respective facility type.
The development of the facility cycles is detailed in an EPA report "Development of Speed
Correction Cycles"5 prepared by Sierra Research under contract to OMS. This report
(M6.SPD.001) can be found at the MOBILE6 section of the OMS web site.
Diurnal evaporative emissions based on real-time diurnal testing
Diurnal evaporative emissions are those emissions of gasoline vapors that occur as a
result of increasing ambient temperatures during times that a vehicle is not being operated. The
increased temperatures lead to increased fuel tank temperatures, leading to vapor generation and
increased fuel system pressure, and hence to vapor release. When the vehicle is running, any
diurnal vapors are routed to the engine intake manifold and burned. The base diurnal emission
factors calculated by earlier versions of the MOBILE model were determined using a test
procedure that simulated the diurnal temperature rise from 60-84°F (16-29°C) over a period of
one hour. Results from testing over other temperature ranges and using other fuel RVP levels
were used to develop estimated diurnal emission factors for other temperature rises and fuel
volatilities.
Along with the implementation of enhanced evaporative emission control regulations,
EPA and others have tested vehicles over the "real-time diurnal" test, in which emissions are
measured over a period of hours with the temperature of the vehicle and SHED (sealed housing
for evaporative determination) being increased over real time (as contrasted to forcing the heat
build to occur over one hour using a heat blanket on the vehicle fuel tank). The results of this
testing, which also included hourly measurements of diurnal emissions, are being used to develop
new and more realistic diurnal emission factors for use in MOBILE6. The hourly measurements
during these real-time diurnal tests are also being used in the development of hour-by-hour
emission factor outputs, which are described under "Structural Changes" later in this paper.
The use of the real-time diurnal test results is also providing a means for improving the
estimates of multi-day diurnal emissions (that is, those diurnal emissions generated on the second
or greater consecutive day that a vehicle is not driven, and therefore is not providing an
opportunity for the carbon canister used to control such emissions to be purged). Some vehicles
have been tested over "real time" diurnal cycles spanning three or more days, uninterrupted by
driving (and without canister purging). MOBILES, in general terms, assumed that a second
consecutive no-drive day resulted in diurnal emissions approximately twice the level of a single
day's diurnal, and that third (or greater) consecutive no-driving days resulted in diurnal emissions
at uncontrolled levels (that is, these emissions were similar to those produced by vehicles
completing lacking evaporative emission control systems). The actual measurements indicate
that these effects are overstated. MOBILE6 will reflect the new evaporative emission estimates.
Considerable information on real-time diurnal testing and how the results are being used
in the development of MOBILE6 is presented in the paper "Modeling Hourly Diurnal Emissions
AWMA December 9, 1998 7 New Orleans, Louisiana
-------
and Interrupted Diurnal Emissions Based on Real-Time Diurnal Data" (M6.EVP.002)6.
Additional analyses and papers on the topics of real-time diurnal emissions, multi-day diurnal
emissions, and related topics are still under internal OMS review, and are scheduled to be
completed and posted on in the MOBILE6 section of the OMS Web site early in 1999.
Update hot soak evaporative emission factors for RVP < 9.0 psi
"Hot soak" emissions are those evaporative emissions generated at the end of a vehicle
trip ("trip-end" emissions) due to increased temperatures in the fuel tank, fuel delivery and
evaporative emission control lines, and engine compartment. These emissions are primarily a
function of trip duration, but are also dependent on ambient temperature and fuel volatility (as
measured by Reid vapor pressure (RVP)). MOBILES and previous versions have based the hot
soak emission factors on testing over a range of temperatures and fuel RVP levels, but have not
included test data from "low volatility" (i.e., < 9.0 psi RVP) gasolines. MOBILE6 will include
revised estimates of hot soak emissions for cases where the in-use (user specified) fuel volatility
is less than 9.0 psi RVP.
Considering the relatively large data sets used for estimation of hot soak emission rates in
the 9.0-11.5 psi RVP range, particularly at 9.0 psi RVP (the volatility of vehicle certification test
fuel), compared to the still relatively small quantity of data available for RVPs < 9.0 psi, the new
emission estimates for RVP < 9.0 psi were matched to the current (MOBILES) estimates at
9.0 psi for any given ambient temperature. This revision is not expected to make a large
difference in hot soak emission estimates, but should result in more accurate emission factors for
RVPs < 9.0 psi, which represents virtually the entire country during the ozone formation season.
Details of the data used and analyses performed are presented in the report "Update of Hot Soak
Emissions Analysis" (M6.EVP.004)7, prepared for EPA under contract by ARCADIS Geraghty
& Miller. This report will be made available for review in the MOBILE6 section of the OMS
Web site.
Update heavy-duty engine emission conversion factors
Heavy-duty engines (HDEs) are defined as engines used in completed vehicles, gas or
diesel, that have a gross vehicle weight rating (GVWR) over 8,500 Ibs. Emissions from such
engines are regulated in terms of grams per brake horsepower-hour (g/bhp-hr). For use in
emission factor calculations, such as are performed by the MOBILE model, and in the
development of emission inventories based on such emission factors, the need is for emissions
expressed in terms of grams per mile (g/mi). "Conversion factors," in units of bhp-hr/mi, are
used to convert emissions from g/bhp-hr to g/mi for both gas and diesel HDEs in the MOBILE
model.
Such conversion factors are a function of several variables, notably in-use fuel
consumption rates (generally expressed as brake specific fuel consumption, or BSFC), fuel
economy (miles per gallon), fuel density, and non-engine related fuel economy improvements
AWMA December 9, 1998 8 New Orleans, Louisiana
-------
(e.g., low rolling resistance tires, reductions in aerodynamic drag). Over the years, as each of the
variables affecting the conversion factors change, the conversion factors themselves change,
generally in the direction of reducing the in-use g/mi emission factor associated with a given
g/bhp-hr emission rate. MOBILE6 will include updated conversion factors for all gas and diesel
HDE vehicles. Much of the new data used in this analysis came from the 1992 Truck Inventory
and Use Survey (TIUS)8, and is detailed in a pair of reports prepared by ARCADIS Geraghty &
Miller under contract to EPA: "Update Heavy-Duty Engine Emission Conversion Factors for
MOBILE6: Analysis of Fuel Economy, Non-Engine Fuel Economy Improvements, and Fuel
Densities" (M6.HDE.002)9 and "Update Heavy-Duty Engine Emission Conversion Factors for
MOBILE6: Analysis of BSFCs and Calculation of Heavy-Duty Engine Emission Conversion
Factors (M6.HDE.004)10. Both of these reports have been posted in the MOBILE6 section of the
QMS Web site.
Update fleet characterization data
The term "fleet characterization data" refers to data describing the overall population of
registered vehicles in use and the rates at which they are used (registration distributions and
annual mileage accumulation rates by age and vehicle type). Emission data is specific to various
types and ages of vehicles, and the relative contribution of vehicles of differing ages is a function
of how many miles such vehicles are operated. In the MOBILE model, the two main types of
fleet characterization data are registration distributions by age and annual mileage accumulation
rates by age, each of which is defined distinctly for each vehicle type included in the model.
Registration distributions by age, for each vehicle type, are a set of fractions that sum to 1.0 and
describe what fraction of all vehicles of type X are of age Y at a given point in time. The
MOBILE model explicitly accounts for 25 model years for each vehicle type, with the small
number of vehicles greater than 25 years of age included within "age 25." Annual mileage
accumulation rates describe how many miles per year vehicles of type X and age Y are operated.
All of the registration distributions and annual mileage accumulation rates by age are
being updated for MOBILE6. Details of this analysis and the sources of data used are provided
in the report "Update of Fleet Characterization Data for Use in MOBILE6" (M6.FLT.002)11,
which was prepared by Arcadis Geraghty & Miller under contract to EPA and is posted in the
MOBILE6 section of the QMS Web site.
Other activity rates
While the primary activity rate used in the development of highway mobile source
emission inventories is total vehicle miles traveled (VMT), which is multiplied by appropriate
emission factors in g/mi to yield total emissions, there are a number of other activity rates used
within the MOBILE model. Most of these are used to weight together emission rates
corresponding to specific situations experienced by some fraction of all vehicles in order to
obtain "overall average" emission rates. Examples of these types of activity factors include the
distribution of soak times (engine off time) preceding starts, used to estimate average start
AWMA December 9, 1998 9 New Orleans, Louisiana
-------
emissions (see following section) and average hot soak (trip end) emissions, and average trip
durations (lengths, in terms of time rather than distance), which determine the rate of running
loss emissions.
Many of these "internal" activity rates are being updated for MOBILE6 using recent
instrumented vehicle data. For more information, the reader is referred to the following reports,
all of which are available at the MOBILE6 section of the OMS Web site: "Soak Length Activity
Factors for Start Emissions"12, "Soak Length Activity Factors for Hot Soak Emissions"13, and
"Trip Length Activity Factors for Running Loss and Exhaust Running Emissions."14
Structural changes
A number of changes being made for MOBILE6 are not to the emission factors per se,
but rather to the structure of the model in terms of how emission factors are calculated and
expressed, what types of input are required of (or can optionally be provided by) the modeler, and
how the output is presented. The most important of these are briefly described in this section.
Separation of start and running emissions
The MOBILE model historically has provided a single emission factor in g/mi for any
unique combination of vehicle type, pollutant, and scenario (defined by user-specified conditions,
such as ambient temperature and fuel RVP, and requested year of evaluation). The effects of
engine operating mode (cold start, hot start, hot stabilized operation) have been handled through
the use of "operating mode fractions," which define the fraction of VMT assumed to be
accumulated in each operating mode. Through manipulation of the operating mode fraction
inputs, the modeler was able to obtain emission factors corresponding to any combination of cold
start, hot start, and hot stabilized operation. The most widely used sets of fractions included
some of each operating mode, in order to represent overall area-wide daily average emissions.
As the transportation and air quality sectors have been better integrated in recent years,
there is a growing need for a more detailed breakout of daily emissions. To respond to this need,
OMS is revising the model to produce separated "start" and "running" emission factors.
"Running" emission factors will continue to be represented in terms of g/mi emission rates, and
will be based only on hot stabilized engine operation (which in turn represents most VMT in
most areas at most times). "Start" emissions will be presented in terms of emission increments;
that is, the added emissions resulting from vehicle start-ups. This will facilitate the modeling
needs of many users, especially those requiring spatial and temporal allocation of emissions. The
user will have the ability to specify a distribution of engine off (soak) times, while the model will
contain a default distribution representing national average conditions. Thus, the previous (and
somewhat arbitrary) distinction between "cold starts" and "hot starts" will be replaced by a more
realistic representation of the range of conditions (i.e., distribution of soak times) preceding
engine starts, and these start emissions can then be allocated to specific locations within an area
and specific times of day.
AWMA December 9, 1998 10 New Orleans, Louisiana
-------
For more information on how OMS has analyzed start and running emissions and
determined the distribution of soak times preceding vehicle starts, see the reports "Soak Length
Activity Factors for Start Emissions"12, "Determination of Start Emissions as a Function of
Mileage and Soak Time for 1981-1993 Model Year Light-Duty Vehicles"15, and "Determination
of Hot Running Emissions from FTP Bag Emissions"16, all of which are available at the
MOBILE6 section of the OMS Web site.
Expansion of vehicle classes
The MOBILE model has provided emission factors for eight vehicle types in past
versions: light-duty gasoline vehicles (LDGV, passenger cars with gasoline engines up to
6000 Ib GVW), light-duty gas trucks (LDGT) I and LDGT II (gasoline-fueled light trucks up to
6000 Ib GVW, and 6001-8500 Ib GVW, respectively), heavy-duty gasoline vehicles (HDGV,
trucks 8501 Ib GVW and up equipped with heavy-duty gasoline engines), light-duty diesel
vehicles (LDDV, passenger cars with diesel engines up to 6000 Ib GVW), light-duty diesel trucks
(LDDT, up to 8500 Ib GVW), heavy-duty diesel vehicles (HDDV, trucks 8501 Ib GVW and up
equipped with heavy-duty diesel engines), and motorcycles (MC, includes only motorcycles
certified for on-highway use and assumed to be all gasoline-fueled). The LDGT I and LDGT II
breakdown was included because there are different emission standards applicable to these two
subgroups; there is no similar break for diesel LDTs as the same emission factors apply to all
diesel LDTs up to 8500 Ib GVW.
MOBILE6 will provide distinct emission factor calculations for a wider range of vehicle
categories, in part due to evolving regulations (i.e., the breakdown of LDGTs into four rather
than two subgroups), and in part due to the needs of the modeling community (i.e., breakdown of
HDGVs and HDDVs by GVW categories). The modeler will still be provided with the familiar
eight vehicle types as listed above, but will be able to obtain a breakdown of emission factors by
finer distinctions for vehicle categories. The current EPA proposal is: LDGVs, LDDVs,
LDDTs, and MCs remain defined as currently, with no further breakdown are available; LDGTs
can be divided into four categories, up from the current two, in accordance with emission
standards issued under the 1990 Clean Air Act amendments; and HDGVs and HDDVs can be
divided into a number of smaller groups. For HDGVs, separate emission factors will be
available for Class lib (8501-10,000 Ib GVW), Class IE (10,001-14,000 Ib GVW), Class IV
(14,001-16,000 Ib GVW), Class V (16,001-19,500 Ib GVW), Class VI (19,501-26,000 Ib GVW),
Class VII (26,001-33,000 Ib GVW), and Class VIE (33,001 Ib GVW and up), as well as for
gasoline-fueled buses (including all types of buses). For HDDVs, a similar breakdown will be
available (Classes lib through VH), with Class VHI further divided into Class VHIa (33,001-
60,000 Ib GVW) and VHIb (60,001 Ib GVW and up), and buses further divided into diesel school
buses and diesel transit/commercial buses).
EPA is still working on these new vehicle categories and subcategories and how emission
factors for each will be calculated. As additional information becomes available it will be posted
in the MOBILE6 section of the OMS Web page.
AWMA December 9, 1998 11 New Orleans, Louisiana
-------
Inclusion of CNG vehicles
There is an increasing demand for emission factor calculations for alternatively fueled
vehicles. Of the various alternative fuels being developed, only compressed natural gas (CNG)
vehicles are currently in use in sufficient numbers with sufficient data to allow EPA to estimate
emission factors for such vehicles. At the option of the modeler, a fraction of each of the vehicle
types outlined above can be specified as being CNG vehicles in MOBILE6, and the resulting
emission factors by vehicle category will account for the different emissions from CNG vehicles
(relative to their gasoline-fueled or diesel counterparts), weighted by their fractions of each
category. At this time there are no data to determine if CNG vehicles are operated differently
than the gas or diesel vehicles that they replace, thus the activity levels, annual mileage
accumulation rates by age, and other non-emission parameters of such vehicles will be assumed
the same as the gas (or diesel) counterparts in MOBILE6.
Hourly calculation/output of emission factors
MOBILES and earlier versions of the model have provided emission factor estimates that
are based on an entire day (daily average emission factors). Among the assumptions made that
result in the emission factors being "daily" averages are that the diurnal emissions are based on
the user input minimum and maximum temperatures, and that the temperature used for correction
of exhaust and other emission factors is based on a trip-weighted average (that is, the model uses
a typical daily temperature profile and a national average distribution of VMT over the 24 hours
of the day to estimate a single value of temperature that, when used to correct emissions,
approximates the same result that would be obtained by estimating emission factors for each of
24 hourly temperatures and weighting those factors by the fraction of daily VMT occurring in
each of the 24 hours). This is further discussed in an earlier EPA technical report.17 When users
seek emission factors for shorter time periods, such as hourly emissions needed for airshed
modeling, the model can be instructed to correct all non-diurnal emission factors for the user
input value of ambient temperature; however, there are no provisions for estimating diurnal
emissions on an hourly basis.
Due to the increasing need on the part of some emission modelers to estimate emissions
on an hourly basis, MOBILE6 is being developed to include the ability to provide both daily
average emission factors (as is now done) and as hourly emission factors for each of the 24 hours
of the day. The data from the real time diurnal testing, already discussed above, will be used to
develop hourly diurnal emission rates. The user will be provided the option of supplying a daily
minimum and maximum temperature, in which case the hourly temperatures will be calculated
using the typical daily temperature profile already included in MOBILES; or, or supplying 24
hourly temperatures, in which case the distribution of VMT over the 24 hours will be combined
with these user-specified temperatures in order to calculate the daily average emission factors.
Exhaust emissions and other categories of non-exhaust emissions will also be calculated on an
hourly basis. Daily emissions will be calculated as a weighted sum of these hourly estimates.
AWMA December 9, 1998 12 New Orleans, Louisiana
-------
Work on this aspect of MOBILE6 is still underway, and has not been completed to a
point that would allow more detail to be provided at this time. As in other parts of the model that
are still in relatively early stages of development, EPA will post any technical reports and other
analyses in the MOBILE6 section of the OMS Web site as they are completed.
Input/Output changes
Use of labeled input/elimination of control flags
Perhaps the biggest change in the MOBILE model from the perspective of the modeler
preparing input data files is the planned elimination of the "control flags" and the use of labeled
input. Rather than a set series of integer-valued flags specifying types of input data to be used,
output items to be included, and the like, there will be a set of default assumptions made; if there
is no contrary or supplementary information in the input file, these defaults will be used. This
will have a number of benefits in terms of simplicity of data files, easing the preparation of the
files, and assisting the modeler (and others) in reading and interpreting these files.
"Labeled input" refers to the ability of the MOBILE6 program to recognize certain
characters or character strings and to operate accordingly. For example, the optional ability to
have more detailed vehicle class factors provided in the output along with the more familiar eight
vehicle class average emission factors was discussed above. In the absence of instructions, the
model will produce eight average emission factors only. However, if a line is included in the
input file that says "Heavy Duty Breakout" then the output will include both the standard eight
average emission factors and emission factors for each of the heavy duty vehicle classes. Since
the program will recognize this instruction no matter where in the input data file it appears, the
use of labeled input will also eliminate the requirement that every component of the input file
appear in a precisely specified ordering. Most input files for MOBILE6 will be considerably
shorter than those used for earlier versions, as the file will only have to contain information that
the modeler wishes to specify (that is, if the default is to be used for a specific aspect of the
modeling, then no information relating to that aspect need appear in the input data file).
The use of labeled input will make reading and interpreting MOBILE6 input data files
much easier than in the past. In addition, the modeler will have much more extensive abilities to
document the input assumptions in the resulting output: Any line in the input data file that
begins with an asterisk (*) in column 1 will be interpreted as a "comment," which will be read in
and reprinted as part of the corresponding output file without having any impact on the
calculations. Finally, the more complicated possible user inputs will be handled through the use
of external data files, as described below.
Much remains to be determined with respect to the exact labels to be used, how much
variation in labels will be permitted (that is, will still be recognized and correctly interpreted by
the model), and other details. A more complete outline of the plans for MOBILE6 in this area
was presented at the MOBILE6 public workshops held in Ann Arbor, MI in 1997. The
AWMA December 9, 1998 13 New Orleans, Louisiana
-------
presentations made at those workshops are all available for review in the MOBILE6 section of
the QMS Web site.
External data files (e.g., fleet characteristics data)
Certain input data that is commonly used, but not always required, in estimating emission
factors is particularly long and complex, and so is usually considered onerous to provide and
interpret. The major examples of these types of input data are the specification of alternate
registration distributions and annual mileage accumulation rates by age and vehicle type. In
MOBILE6, these (and possibly other) types of input data will not appear directly in the input data
file prepared for MOBILE6. Rather, they will be provided in external (to the model) data files,
and only the location of the external data file will need to be provided in the MOBILE input file.
This will provide two significant benefits to the modeler. First, the information (for example, the
local registration distributions) will only have to be developed and formatted once; after that, any
MOBILE runs that require inclusion of the alternate registration distributions will only have to
note the name and location (file path) of that external file.
As in the case of the labeled input, there are still a number of details to be worked out in
terms of exactly how this will be implemented in MOBILE6. An overview of the concept was
presented at the 1997 MOBILE6 public workshops, and can be viewed at the OMS Web site. As
EPA completes more work in this area, additional documentation will be posted there.
Modified outputs (one descriptive, one electronic/database-type output)
MOBILES continued to provide the modeler with four options for the emission factor
output reports, as had been done with earlier versions: a "short" and a "long" form of descriptive
output, and a "short" and "long" form of numeric output. The descriptive output formats are
intended for reading and interpretation by persons; they contained extensive labeling, echoing of
input information, and other text enabling the reader to follow what was presented. The numeric
outputs are intended to facilitate use of the output file as input to other programs; the labels and
other text included are minimal, with the focus being on consistent columns of data so as to be
read in by other programs.
Given the extensive changes being made, the improvements in computer hardware and
software since MOBILES was developed, and EPA's understanding of the needs of the modeling
community, we have decided against attempting to maintain all four of these output format
choices for MOBILE6. Instead, attention is being focused on providing one high-quality, easily
read and understood "descriptive" format, and one electronic (numeric) output. The descriptive
output will be based on what is currently known as the "long" descriptive output format; it will
be best viewed landscape style, and will include extensive, clear labeling of all information.
Combined with the changes to the input files (labeled input, external data files), it should be
easily read and interpreted, even without having the corresponding input file at hand. As noted,
it will also contain echoing of all comments (lines starting with "*" in column 1) provided by the
AWMA December 9, 1998 14 New Orleans, Louisiana
-------
preparer of the input data file.
The new electronic output is being developed in such a way as to be easily read into
common database programs. The output will use ASCII characters making it also readable by
spreadsheet programs, such as Lotus 1-2-3 and Excel, as well as to facilitate its input to other
calculation programs that require highway vehicle emission factors, such as CALINE or
CAL3QHC.
The precise details of how these two output reports will appear cannot be definitively
specified until we are much closer to completing revisions to the model. Overviews of EPA's
intent in this area were presented at the 1997 public workshops on MOBILE6, and are available
for review at the OMS Web site. As we complete more work and more decisions on details in
this area are made, additional information and documentation will be posted there.
Stakeholder and Peer Review
The development of revisions to the MOBILE model is a lengthy and involved process.
In the past, the primary means by which OMS has involved its stakeholders in the process has
been public workshops. At these workshops, OMS presents preliminary plans, analyses, and
results, and seeks the input and reaction of attendees. While these workshops are open to the
general public and announced in the Federal Register, the bulk of attendees have been
representatives of the regulated industries (auto and engine manufacturers, petroleum companies,
and trade associations for these industries), contractors, and model users (State and local air
quality officials and staff, EPA Regional Office representatives). EPA has always encouraged
attendees to submit comments on the materials presented at these workshops, and has considered
those comments in further refinement and finalizing of model revisions.
However, considerable criticisms have been voiced that these workshops alone are
inadequate to provide sufficient opportunities for affected parties to comment on revisions to the
model and to influence its final form and content. In developing MOBILE6, OMS is taking
unprecedented steps to provide more opportunities for interested parties to provide input to the
model revision process. In addition to continuing to conduct public workshops, as we have done
in the past, OMS has taken several other approaches to maximizing opportunities for comment
on the model.
First, OMS has established an electronic list server devoted to news and information
concerning mobile source emission factor development, revision, and use. This list server is
used to disseminate information and guidance concerning the model and its use to all interested
parties using computer based e-mail. Updates and corrections to the existing model, MOBILESa,
are issued over the list server, and guidance on using the model for specific applications is
announced there. In addition, information concerning the development of MOBILE6 is sent over
the list server, notably announcements that new material has been made available for review and
comment by stakeholders. This stakeholder review process is discussed below.
AWMA December 9, 1998 15 New Orleans, Louisiana
-------
Second, we have established a section within the OMS Web site, as has been noted earlier
in this paper, devoted exclusively to MOBILE6 development and stakeholder review. As
proposals are developed, analyses completed, and progress made on other aspects of the revised
model, EPA posts reports and other materials on the Web site for stakeholder review. Formal
reports are numbered to facilitate tracking of the material, and announcements of such postings
are sent over the list server. Each report dealing with MOBILE6, when posted, provides a sixty
day comment period during which comments are accepted from any interested individuals and
organizations. This addresses the concerns that have been raised that in the workshop format,
EPA only presents material that has largely been finalized, and the stakeholder review process
provides earlier and more frequent chances for comment on proposals, preliminary results, and
intended modeling approaches. All comments submitted as part of the stakeholder review
process will be fully considered, and responses prepared and posted, before final changes are
implemented in MOBILE6. More information about the stakeholder review process, means of
submitting comments, and related information can be found at the Web site
(http://www.epa.gov/oms/m6.htm).
Finally, EPA is obtaining formal peer review of most significant analyses and
documentation intended for use in MOBILE6. While difficulties have been encountered in
locating a sufficient pool of independent experts for reviewing components of the model, this
process appears to be working, and is providing EPA with more and more independent reaction
and comment than has been true in the past. The primary peer review coordinator for MOBILE6
analyses is Mr. Venkatesh Rao of the Assessment and Modeling Division. As with other related
topics, additional information on peer review and its role in the development of MOBILE6 can
be found in the modeling section of the OMS Web site.
The overall benefits of the list server and the expanded stakeholder and peer review
processes are expected to be a more sound model, with greater confidence in the results provided
by the model. Users of the model, and those who make decisions based in part on results of such
modeling, will be able to have more confidence in the emission factors produced by MOBILE6,
and hence in their decisions based on such modeling.
Summary
In many ways, MOBILE6 will be the most extensive revision of the highway vehicle
emission factor model undertaken since MOBILE1. Virtually all emission rates and correction
factors are being revised and updated, the structure of the input and output are being revised to
account for changes both in the underlying analyses and assumptions used in the model and the
expanding audience for use of the model, and the process for obtaining outside review and input
is more extensive than has been attempted in the past. While MOBILE6 will not provide "the"
answer to the question "What are the levels of emissions from highway vehicles in use?," and
will not address every concern that has been raised by those using the model, we are confident
that it will provide better, more accurate emission factor estimates than did earlier versions, and
that those using these results will be able to have more confidence in them than in the past.
AWMA December 9, 1998 16 New Orleans, Louisiana
-------
In the time remaining before MOBILE6 is finalized and released, the continued use of the
current MOBILESa and MOBILESb models (as applicable) will be accepted by EPA for all
analyses (including State Implementation Plan (SIP) inventories and other submittals). As in all
cases where the MOBILE model is revised, there will be "transition issues" that will require EPA
to develop and issue guidance covering the details of switching from use of the previous model
to the new model. Such information will be disseminated over the list server discussed earlier,
posted on the OMS Web site, and where appropriate announced in the Federal Register.
References
1. "Environmental Law: Volume 1" (Committee on Commerce United States House of
Representatives), April 1997.
2. "Air Conditioning Activity Effects in MOBILE6," Report M6.ACE.001, John Koupal,
U.S. EPA Office of Mobile Sources, Assessment and Modeling Division.
3. "Air Conditioning Correction Factors in MOBILE6," Report M6.ACE.002, John Koupal,
U.S. EPA Office of Mobile Sources, Assessment and Modeling Division.
4. "Fuel Oxygen Effects on Exhaust CO Emissions," Report M6.FUL.002, Venkatesh Rao,
U.S. EPA Office of Mobile Sources, Assessment and Modeling Division.
5. "Development of Speed Correction Cycles," Report M6.SPD.001, prepared for U.S. EPA
Office of Mobile Sources, Assessment and Modeling Division by Thomas C Carlson and
Thomas C. Austin, Sierra Research, Inc., Sacramento CA, April 30, 1997.
6. "Modeling Hourly Diurnal Emissions and Interrupted Diurnal Emissions Based on
Real-Time Diurnal Data," Report M6.EVP.002, Larry Landman, U.S. EPA Office of
Mobile Sources, Assessment and Modeling Division.
7. "Update of Hot Soak Emissions Analysis," Report M6.EVP.004, prepared for U. S. EPA
Office of Mobile Sources, Assessment and Modeling Division, under contract by Dr.
Louis Browning, ARCADIS Geraghty & Miller, Mountain View CA.
8. "1992 Census of Transportation Truck Inventory and Use Survey, United States," Report
TC92-T-52, U.S. Department of Commerce, Bureau of the Census, May 1995.
9. "Update Heavy-Duty Engine Emission Conversion Factors for MOBILE6: Analysis of
Fuel Economy, Non-Engine Fuel Economy Improvements, and Fuel Densities," Report
M6.HDE.002, prepared for U.S. EPA Office of Mobile Sources, Assessment and
Modeling Division, by Dr. Louis Browning, ARCADIS Geraghty & Miller, Mountain
View CA.
AWMA December 9, 1998 17 New Orleans, Louisiana
-------
10. "Update Heavy-Duty Engine Emission Conversion Factors for MOBILE6: Analysis of
BSFCs and Calculation of Heavy-Duty Engine Emission Conversion Factors," Report
M6.FtDE.004, prepared for U.S. EPA Office of Mobile Sources, Assessment and
Modeling Division, by Dr. Louis Browning, ARCADIS Geraghty & Miller, Mountain
View CA.
11. "Update of Fleet Characterization Data for Use in MOBILE6," Report M6.FLT.002,
prepared for U.S. EPA Office of Mobile Sources, Assessment and Modeling Division, by
Dr. Louis Browning, ARCADIS Geraghty & Miller, Mountain View CA.
12. "Soak Length Activity Factors for Start Emissions," Report M6.FLT.003, Edward L.
Glover and David J. Brzezinski, U.S. EPA Office of Mobile Sources, Assessment and
Modeling Division.
13. "Soak Length Activity Factors for Hot Soak Emissions," Report M6.FLT.004, Edward L.
Glover and David J. Brzezinski, U.S. EPA Office of Mobile Sources, Assessment and
Modeling Division.
14. "Trip Length Activity Factors for Running Loss and Exhaust Running Emissions," Report
M6.FLT.005, Edward L. Glover and David J. Brzezinski, U.S. EPA Office of Mobile
Sources, Assessment and Modeling Division.
15. "Determination of Start Emissions as a Function of Mileage and Soak Time for 1981-
1993 Model Year Light-Duty Vehicles," Report M6.STE.003, Ed Glover, Penny Carey,
Phil Enns, and David J. Brzezinski, U.S. EPA Office of Mobile Sources, Assessment and
Modeling Division
16. "Determination of Hot Running Emissions from FTP Bag Emissions"Report
M6.STE.002, David J Brzezinski, Ed Glover, and Phil Enns, U.S. EPA Office of Mobile
Sources, Assessment and Modeling Division.
17. "Estimation of Trip- and Emission-Weighted Temperatures for MOBILE4," Report EPA-
AA-TEB-EF-90-01, Celia Shih, U.S. EPA Office of Mobile Sources, Assessment and
Modeling Division, January 1990.
AWMA December 9, 1998 18 New Orleans, Louisiana
------- |