



                                       
MEMORANDUM 

SUBJECT: Updates to MOVES for the Tier 3 NPRM Analysis

FROM:  Megan Beardsley, Environmental Scientist
Assessment and Standards Division 

TO: Docket Number EPA-HQ-OAR- 2011-0135 

DATE: March 11, 2013 

   The U.S. Environmental Protection Agency (EPA) is proposing new vehicle emission standards and fuel quality standards to reduce emissions of criteria pollutants.  Collectively, these proposed standards are referred to as "Tier 3 vehicle standards" and "Tier 3 fuel standards."   
   As part of this analysis we estimated the impact of the Tier 3 proposal on the nationwide highway vehicle emissions of criteria pollutants, criteria pollutant precursors and selected air toxics.  We made these estimates using an updated version of the EPA's Motor Vehicle Emission Simulator (MOVES2010). 
   The attached document details how the MOVES default code and databases were updated for the Tier 3 NPRM analysis.  
                                       


                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                                       
                 Updates to MOVES for the Tier 3 NPRM Analysis
                                       
                                  March 2013
                                       
                                       
                                       
                                       
                       Air Quality & Modeling Center
                      Assessment & Standards Division
                  Office of Transportation & Air Quality
                                   U.S. EPA

                                       

I.	Introduction	4
II.	Fuel Effect Updates	5
A.	Fuel Sulfur Impacts	5
1.	New Research of Fuel Sulfur Effects on Emissions	5
2.	Implementation in MOVES	6
B.	Fuel Effects from EPAct Study	7
1.	EPAct Fuel Effects on Criteria Pollutants	7
2.	EPAct Fuel Effects on Selected Air Toxics	9
3.	Additional Adjustments for Benzene and 1,3-Butadiene	12
C.	Additional Fuel-Related Changes to MOVES	12
III.	Evaporative Emission Updates	13
A.	Permeation	13
B.	Vapor Venting	14
1.	Update of Vapor Venting Algorithms with new DELTA Model	14
2.	Multiple Day Diurnals	14
3.	Vapor Leaks	15
4.	Summary of Updated MOVES Inputs for Vapor Venting	18
IV.	Update of Air Toxics and HC Speciation	21
A.	Algorithms for 2003-and-Earlier Gasoline Vehicles	21
B.	Estimating Emissions for Additional Air Toxics	23
1.	Additional Air Toxics:  Inputs for Gasoline Vehicles	24
a)	Exhaust	24
b)	Evaporative and Permeation	28
2.	Additional Air Toxics:  Inputs for Emissions from Diesel Vehicles	30
a)	Exhaust	30
b)	Evaporative	34
C.	Hydrocarbon Speciation Factors	35
1.	HC Speciation of Exhaust Emissions	36
2.	HC Speciation of Permeation Emissions	38
D.	Additional Updates to Air Toxics and HC Speciation	38
V.	Updated Heavy Duty and Diesel Emission Rates	38
A.	Changes to Heavy Duty Gasoline Vehicle Rates	38
1.	Regulatory Class Definition and Vehicle Composition	39
2.	Estimating Emission Rates	40
B.	Changes to Heavy- and Light-Duty Diesel Vehicle Rates	43
VI.	Other Changes	43
A.	Changes to Model Year Groups	44
B.	Changes to County Allocations	44
C.	Repair of Errors	44
D.	Changes to Facilitate Large Numbers of Runs	44
VII.	Impact of MOVES Changes	45
References	46

Introduction

   The U.S. Environmental Protection Agency (EPA) is proposing new vehicle emission standards and fuel quality standards to reduce emissions of criteria pollutants.  Collectively, these proposed standards are referred to as "Tier 3 vehicle standards" and "Tier 3 fuel standards."   
   As part of this analysis we estimated the impact of the Tier 3 proposal on the nationwide highway vehicle emissions of criteria pollutants, criteria pollutant precursors and selected air toxics.  We made these estimates using an updated version of the EPA's Motor Vehicle Emission Simulator (MOVES2010). This document describes how the MOVES default code and databases were updated for the Tier 3 NPRM analysis.  
   MOVES2010 is EPA's official model for estimating emissions from cars and trucks.  At the time of its release, it incorporated EPA's most up-to-date information on gasoline vehicle emissions.  Exhaust emission rates for HC, CO and NOx were developed based on an analysis of state inspection/maintenance and roadside remote sensing data from millions of vehicles.  Emissions of particulate matter are based on EPA's Kansas City gasoline PM study.,  Evaporative emission rates were updated based on extensive evaporative testing conducted by EPA and the Coordinating Research Council (CRC) .   
   However, since the release of MOVES2010, we have continued to collect and analyze emission data.  In particular, we have completed two major studies of fuel effects on emissions in Tier 2 light duty gasoline vehicles, and an important collection of studies on evaporative emissions.  Because fuels and evaporative emissions are affected by Tier 3 standards, it was important to include this data in our analysis.   Sections below describe this data in detail and explain how it was incorporated into an April 14, 2011 version of the MOVES model software (MOVESGHGSource20110414a) and a May 12 version of the MOVES default database (MOVESTier3db20110512).  Collectively, this memo refers to this software and database as MOVEST3NPRM.
   When we updated the MOVES fuel effects and evaporative emissions, we also had an opportunity to improve the model's emission estimates in other areas.  Notably, we updated emissions for the air toxics and hydrocarbon species addressed in MOVES2010, and we expanded MOVES capabilities to include estimating all the air toxic pollutants in the National Emission Inventory (NEI) and National Air Toxics Assessment (NATA) that are emitted by mobile sources. We also made some updates to heavy duty emission rates for diesel and gasoline vehicles.  The sulfate, sulfur dioxide, organic carbon and elemental carbon emission rates for 2007-and-later heavy duty diesel vehicles were updated to include information from a recent study that examined the composition of particulate emissions from advanced diesel engines. HC, CO and NOx start and running emission rates for light heavy duty gasoline vehicles were updated to fix an error in these rates for 2007-and-later emissions, And, we made a number of miscellaneous changes including changes to fleet and activity inputs, repairs of minor errors, and changes to facilitate the large number of parallel runs required for the Tier 3 NPRM analysis. 
   Details on all these changes are provided in the sections below.
   
Fuel Effect Updates
Fuel Sulfur Impacts
   MOVEST3NPRM includes updates to the effect of fuel sulfur levels below 30ppm on exhaust emissions.  In MOVES2010 these effects were based on extrapolation from data on sulfur levels above 30 ppm.   The updates made for this analysis were based on significant new data generated through EPA research on low sulfur fuels conducted in 2010 and 2011. 
New Research of Fuel Sulfur Effects on Emissions
   The EPA study assessed the emission reductions expected from the in-use Tier 2 fleet with a reduction in fuel sulfur level from current levels.  The study was designed to take into consideration what was known from prior studies on sulfur build-up in catalysts over time and the effect of periodic regeneration events that may result from high speed and load operation over the course of day-to-day driving.  The study sample consisted of 81 cars and light trucks covering model years 2007-9 , each with approximately 20,000-40,000 odometer miles.  The makes and models targeted for recruitment were chosen to be representative of high sales vehicles covering a range of types and sizes.  Test fuels were two non-ethanol gasolines with properties typical of certification fuel, one at a sulfur level of 5 ppm and the other at 28 ppm. All emissions data was collected using the Federal Test Procedure (FTP) cycle at a nominal temperature of 75°F. A draft report on this research is available.
      
   Major findings from this study include:
      
         * Reversible sulfur loading is occurring in in-use fleet of Tier 2 vehicles and has a measureable effect on emissions of NOx, hydrocarbons, and other pollutants of interest.
         * The effectiveness of high speed/load procedures in restoring catalyst efficiency is limited when operating on higher sulfur fuel.
         * Reducing fuel sulfur levels from 28 to 5 ppm is expected to achieve significant reductions in emissions of NOx, hydrocarbons, and other pollutants of interest in the in-use fleet, as summarized in Table 1, below.
            
      
Table 1 In-use emission reductions from 28 to 5 ppm sulfur


NOX (p-value)
THC
(p-value)
CO
(p-value)
NMHC
(p-value)
CH4
(p-value)
NOx+NMOG
(p-value)
PM[‡]
Bag 1
10.7%
(0.0033)
8.5%[†]
(0.0382)
7.5%[†]
(0.0552)
7.5%
(< 0.0001)
13.9%[†]
(< 0.0001)
N/A
 - 
Bag 2
59.2%
(< 0.0001)
48.8%
(< 0.0001)
 -  [‡]
44.8%[†]
(0.0260)
49.9%
(< 0.0001)
N/A
 - 
Bag 3
62.1%
(< 0.0001)
40.2%
(< 0.0001)
20.1%
(< 0.0001)
49.9%
(< 0.0001)
29.2%
(< 0.0001)
N/A
 - 
FTP Composite
23.0%[†]
(0.0180)
13.0%[†]
(0.0027)
11.9%[†]
(0.0378)
10.6%[†]
(0.0032)
25.8%[†]
(< 0.0001)
17.3%
(0.0140)
 - 
Bag 1  -  Bag 3
 -  [‡]
5.2%
(0.0063)
4.3%
(0.0689)
5.1%
(0.0107)
4.6%
(0.0514)
N/A
 - 
† Model with significant sulfur and mileage interaction term.  [‡] Sulfur level not significant at α = 0.10.  For THC Bag 1 and CH4 Bag 1, because the effect of clean-out was not statistically significant, the reduction estimates are based on the estimates of least squares means.
      
Implementation in MOVES
    The results shown in Table 1 were applied in MOVEST3NPRM for model year 2004-and-later gasoline vehicles (the nominal start of the Tier 2 phase-in) to estimate sulfur effects below 30 ppm.  The sulfur fuel effect applies multiplicatively to other gasoline fuel effects in MOVES, and applies only for sulfur levels below 30ppm. For sulfur levels above 30 ppm, and for all pre-2004 model year vehicles, the sulfur effect originally used in MOVES2010 remains in place.  Equation 1 shows the generic form of the new sulfur correction. Table 2 shows the specific values for the sulfur coefficients by pollutant, process, and vehicle type.
      
                 sulfur effect=1.0-Coeffsulfur*30-sulfurLevel
Equation 1 Low Sulfur Effect

Table 2 Low Sulfur Coefficients by Vehicle Type, Process and Pollutant
                                 Vehicle Type
                                      THC
                                      CO
                                      NOX
                                      PM
                                       
                                    Starts
                                    Running
                                    Starts
                                    Running
                                    Starts
                                    Running
                                    Starts
                                    Running
Motorcycle
                                                                              0
                                                                              0
                                                                              0
                                                                              0
                                                                              0
                                                                              0
                                                                              0
                                                                              0
Passenger Car, 
Passenger Truck & 
Light Commercial Truck
                                                                       0.002237
                                                                       0.020336
                                                                       0.001866
                                                                              0
                                                                              0
                                                                       0.024459
                                                                              0
                                                                              0
All other Vehicle Types
                                                                              0
                                                                       0.015488
                                                                              0
                                                                       0.009436
                                                                              0
                                                                       0.027266
                                                                              0
                                                                              0

   These equations were then used to fill in the database table that houses fuel effect equations in the MOVES database ("GeneralFuelRatioExpression"). This table allows the MOVES model to compute fuel effects based on the fuel properties of any fuel contained in the fuel supply and fuel formulation database tables.  
   In addition, in order to properly model fuel sulfur levels below 30 ppm using Equation 1 in MOVEST3NPRM, we had to modify an existing "floor" to the sulfur correction. In MOVES2010, the MOBILE6 fuel sulfur algorithms, which utilized log-log relationship, were extrapolated for fuels with sulfur level below 30 ppm.  To prevent highly non-linear relationships inherent in log-log equation from dominating the MOVES emission results, a fuel adjustment `floor' of 0.85 was added.  This `floor' prevented MOVES2010 from reducing the emission rate by more than 15 percent from the `base' emission rate as a result of changing the base fuel formulation (30 ppm sulfur) to some other low level (i.e., 0.01 ppm sulfur).  The floor is still needed for estimating sulfur reductions for model years 2001-2003 (which continue to be modeled with the original log-log equation), however, considering the reduction in emissions seen for fuel sulfur levels below 30 ppm in the latest EPA analysis, it was necessary in MOVEST3NPRM to change the sulfur `floor' to 0.40.   

Fuel Effects from EPAct Study

   As required by the Energy Policy Act of 2005 (EPAct), the EPA conducted an extensive evaluation of the effects of fuel properties other than sulfur on emissions from Tier 2 gasoline vehicles.  The EPAct exhaust research program and analysis are extensively documented in the "EPAct Test Program Report"  and "EPAct Analysis Report."  Information about the study is not repeated here.   Because the fuel effects in MOVES were based on analysis of the Tier 0 and Tier 1 vehicles typical of the 1990s, we felt it was important to incorporate the EPAct study results into our air quality analysis.  However, given the many months of work needed for this analysis, we had to develop the fuel inputs for MOVEST3NPRM before the EPAct analysis described the second report was finalized.  This section describes the preliminary analysis results that were used to generate the fuel effects inputs used in MOVEST3NPRM.   Note, information on how the EPAct data was used to update hydrocarbon speciation rates is provided in Section IV.C, below.
EPAct Fuel Effects on Criteria Pollutants
   
      MOVES estimates emissions immediately following engine starts ("start emissions") as distinct from those from warmed-up operation ("running emissions"), and allows different fuel effects for these two modes of operation.   Two sets of exhaust fuel effect coefficients were therefore employed in the model; one set of coefficients was applied to represent start emissions and a second set was applied to represent hot-running emissions. In some cases fuel effects estimated for these two processes differed substantially, as the effects of fuel properties on start emissions are dominated by changes in combustion and catalyst warm-up, while the impact of running emissions is dictated by catalyst efficiency when fully operational.  
      The MOVES model accounts for the effect of fuel properties on vehicle emissions by estimating emissions from vehicles operating on a "base fuel," and multiplying these "base emissions" by a fuel adjustment to represent emissions on various in-use fuels.  Different fuel adjustments have been developed to represent each pollutant and emission process.  The fuel adjustment varies depending on the properties of the in-use fuel.  It also varies depending on the type of vehicle, the model year and the age of the vehicle.  The fuel effects used in MOVES2010 are documented in an EPA technical report.  The paragraphs below explain how these fuel effects were updated for purposes of the current analysis.
      MOVEST3NPRMallows a very flexible input data format for incorporating and applying coefficients within a wide variety of mathematical forms. These "fuel-effect ratio expressions" can include up to 32,000 characters and are stored in a database table dedicated to this purpose (GeneralFuelRatioExpression).
 	The statistical models generated using EPAct results allow estimation of emissions effects related to five fuel properties: ethanol content (%), aromatics content (%), RVP (psi), T50 (°F) and T90 (°F), as well as selected interaction terms among these five parameters.  The statistical models generated from the EPAct exhaust data follow the general structure of Equation 2below. In this equation, β denotes a model coefficient, ZetOH and ZT90 denote "standardized" fuel terms for these properties, and ZetOHxT90 denotes a "standardized" term for an etOHxT90 interaction term.  For simplicity, only the terms for ethanol, T90 and their interaction have been shown; linear and interaction terms for the remaining three properties are not shown here. Finally, the term sε[2] represents the residual error or "mean square error" for the model. Note that the subsets of the potential terms vary by emission and process, depending on the results of the statistical model fitting. Relative fuel effects are calculated by applying the models to specific "in-use" fuels and pre-defined "base fuels" and taking the ratio of the emissions on the in-use fuel to those on the base fuel, as shown in Equation 2 . 

                                       
Equation 2.  fuel effect= eCoeffetOH*etOHLevelnormalized+Coeffarom*aromLevelnormalized+...+(0.5*variance)Generic Form of Statistical Model for Emissions in relation to Selected Fuel Properties.

                                       
Equation 3 Generic Form of Relative Fuel Effect
      
       Table 3 shows the coefficients for models used in Equation 3 Generic Form of Relative Fuel Effect
       for four individual pollutants, including total hydrocarbons (THC) and the criteria pollutants CO, NOx, and PM.  The application of these models has been integrated into MOVEST3NPRM.  For implementation in MOVES, Equation 3  is input directly into the GeneralFuelRatioExpression table using sets of coefficients from Table 3.  The table presents two sets of coefficients for each pollutant, representing the effects of the fuel properties on start and running exhaust emissions, respectively. Because these coefficients apply to "standardized" fuel properties, as mentioned above, the magnitude and sign of each term gives a sense of the influence of each term in the estimation of that pollutant relative to the others. For example, for PM start emissions, changes in aromatics and T90 are very influential, and are positively related with PM emissions (e.g., when aromatics increase, PM increases).  For THC start emissions, on the other hand, the negative sign of the coefficient indicates that the relationship is inverse, i.e., increases in RVP are associated with decreases in THC.   It is also important to note that the values of the coefficients presented in Table 3 represent the status of analysis and interpretation of the EPAct results at the time they were applied to this analysis. Thus, they sometimes differ from values presented the final analysis report[9], which represent the outcome of additional analysis and interpretation.

Table 3.  Coefficients for Models representing Regulated Gaseous Emissions and Particulated, from the EPAct  Program.
Parameter
                                      THC
                                      CO
                                      NOX
                                      PM

                                     Start
                                    Running
                                     Start
                                    Running
                                     Start
                                    Running
                                     Start
                                    Running
Intercept
                                                                       -0.86577
                                                                       -4.90925
                                                                       1.347166
                                                                       -1.38957
                                                                       -2.85935
                                                                       -4.56921
                                                                              
                                                                              
etOH
                                                                       0.067934
                                                                       0.068143
                                                                       -0.06967
                                                                              0
                                                                       0.067502
                                                                       0.062989
                                                                         0.1807
                                                                         0.1158
aromatics
                                                                       0.083434
                                                                       -0.02281
                                                                       -0.01156
                                                                       0.097982
                                                                       0.133931
                                                                       0.044062
                                                                         0.3792
                                                                         0.1988
RVP
                                                                       -0.04669
                                                                       -0.03953
                                                                       0.020631
                                                                       0.028383
                                                                              
                                                                              
                                                                              
                                                                              
T50
                                                                       0.148984
                                                                       0.071538
                                                                       0.021592
                                                                       0.024856
                                                                       0.047821
                                                                              
                                                                         0.1004
                                                                        -0.0037
T90
                                                                       0.014347
                                                                       0.066813
                                                                       -0.14685
                                                                       0.041802
                                                                              
                                                                              
                                                                         0.3034
                                                                         0.0953
etOHxetOH
                                                                       0.040652
                                                                              
                                                                       0.085346
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
T50xT50
                                                                       0.075548
                                                                       0.030543
                                                                        0.07222
                                                                              
                                                                              
                                                                              
                                                                         0.0806
                                                                              
etOHxArom
                                                                       0.022971
                                                                       0.040277
                                                                       0.062441
                                                                              
                                                                       -0.02369
                                                                              
                                                                              
                                                                              
T50xT90
                                                                       0.050029
                                                                       0.028862
                                                                              
                                                                       0.021773
                                                                              
                                                                              
                                                                              
                                                                         0.1059
etOHxT50
                                                                       0.034265
                                                                              
                                                                       0.111736
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
etOHxT90
                                                                       0.047815
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
                                                                         0.0647
                                                                              
aromxRVP
                                                                       0.025524
                                                                              
                                                                         0.0537
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
aromxT50
                                                                              
                                                                       0.033858
                                                                       0.058589
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
aromxT90
                                                                       0.022605
                                                                              
                                                                       0.038612
                                                                              
                                                                              
                                                                              
                                                                         0.0665
                                                                              
variance
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
                                                                         1.0443
                                                                              
  
EPAct Fuel Effects on Selected Air Toxics
      For use in MOVEST3NPRM, inputs were also developed and applied to estimate emissions of selected toxic compounds and ethanol, but the approach differed from that described above for hydrocarbons and the criteria pollutants. Rather than estimating the change in emissions for one compound relative to two different fuels, the approach for toxics estimates the toxic emissions as a fraction of theVOC emissions on the same fuel.  The compounds selected for this analysis include acetaldehyde, formaldehyde, acrolein, benzene and 1,3-butadiene, all identified as national or regional risk drivers in the 2005 National-scale Air Toxics Assessments (NATA).  
      Table 4 shows coefficients for statistical models for ethanol and the five toxics. These models were developed using techniques similar to the models shown in Table 3, using speciated results obtained from the EPAct exhaust program. The coefficients for the toxics models are interpreted and applied in the same way.  However, two differences must be noted:  as resources limited the number of tests for which hydrocarbons could be speciated, some statistical models were developed using smaller subsets of data and are thus simpler in structure, including only subsets of the four linear terms etOH, arom, T50 and T90 and no second-order terms. In the table, models fit using reduced data subsets contain no cells for the terms not included, and the `---" symbol indicates that RVP was not included.  
      In addition, as VOC are calculated as non-methane organic gases (NMOG) minus ethane (H3CCH3), it was necessary to develop analogous models for are both NMOG and ethane. Coefficients for these two models are shown in Table 5. Models for these two compounds were calculated in the same manner as those shown in Table 3 and Table 4.  Note that models for both start and running ethane were developed using reduced data subsets.
      Thus, if we let Xβtoxic, XαNMOG and Xθethane represent models for a selected toxic compound, NMOG, and ethane, respectively, calculated by applying Equation 2 to each compound for a specified fuel,  the toxic emissions as a fraction of  VOC emissions is given by
                                       
Equation 4.  General expression for toxic fractions.
Also note that models representing running emissions are not available for three compounds: acrolein, benzene and 1,3-butadiene. For these compounds, the relevant subsets of data were inadequate to allow model fitting. Therefore, for these compounds, running emissions were represented as constant fractions of VOC, with values derived from the available data.  For acrolein, benzene and 1,3-butadiene, the values of the fractions were 0.00077, 0.047 and 0.0, respectively. 

Table 4. Fuel Effect Coefficients for Selected Toxic Pollutants from EPAct Exhaust Program
Parameter
                                 Acetaldehyde
                                 Formaldehyde
                                   Acrolein
                                    Ethanol
                                    Benzene
                                 1,3-Butadiene

                                     Start
                                    Running
                                     Start
                                    Running
                                     Start
                                     Start
                                    Running
                                     Start
                                     Start
Intercept
                                                                       -5.23229
                                                                       -9.41919
                                                                       -5.97707
                                                                       -8.65741
                                                                       -7.93381
                                                                       -4.90805
                                                                       -9.56335
                                                                        -4.1074
                                                                       -5.91748
etOH
                                                                        0.81449
                                                                       0.190954
                                                                       0.229862
                                                                       0.078035
                                                                       0.247572
                                                                       1.462653
                                                                       0.816632
                                                                              
                                                                       -0.12725
arom
                                                                       0.034835
                                                                              
                                                                       0.028224
                                                                              
                                                                       0.112172
                                                                              
                                                                              
                                                                       0.403208
                                                                              
RVP
                                                                        -0.0417
                                                                          --- 
                                                                       -0.04718
                                                                          --- 
                                                                        -0.0645
                                                                       -0.06054
                                                                          --- 
                                                                          --- 
                                                                          --- 
T50
                                                                       0.086703
                                                                              
                                                                       0.167177
                                                                              
                                                                       0.188036
                                                                       0.070291
                                                                              
                                                                              
                                                                       0.039682
T90
                                                                       0.038013
                                                                              
                                                                        0.13018
                                                                       -0.08322
                                                                       0.248876
                                                                       -0.09923
                                                                              
                                                                              
                                                                              
etOH^2
                                                                       -0.16694
                                                                              
                                                                              
                                                                              
                                                                        -0.0831
                                                                       -0.49699
                                                                              
                                                                              
                                                                              
T50^2
                                                                       0.066653
                                                                              
                                                                       0.052617
                                                                              
                                                                              
                                                                       0.110804
                                                                              
                                                                              
                                                                              
etOH*arom
                                                                       0.018402
                                                                              
                                                                       0.016505
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
etOH*RVP
                                                                        0.02194
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
                                                                              
T50*T90
                                                                       0.039593
                                                                              
                                                                       0.034887
                                                                              
                                                                       0.059859
                                                                              
                                                                              
                                                                              
                                                                              
etOH*T50
                                                                              
                                                                              
                                                                       -0.01627
                                                                              
                                                                       -0.11858
                                                                              
                                                                              
                                                                              
                                                                              
etOH*T90
                                                                              
                                                                              
                                                                       0.020043
                                                                              
                                                                       0.046167
                                                                              
                                                                              
                                                                              
                                                                              
Variance (sε2)
                                                                         0.2034
                                                                        0.46902
                                                                         0.4765
                                                                        0.45999
                                                                        0.46611
                                                                       0.701256
                                                                       1.927463
                                                                         0.4635
                                                                        0.13059

Table 5.  Fuel Effect Coefficients for NMOG and Ethane, from EPAct Exhaust Program
Parameter
                                     NMOG
                                    Ethane

                                     Start
                                    Running
                                     Start 
                                    Running
Intercept
                                                                        -0.9574
                                                                        -5.5218
                                                                        -4.3412
                                                                        -7.7241
etOH
                                                                        0.08783
                                                                        0.09213
                                                                        0.05222
                                                                               
arom
                                                                         0.1037
                                                                        0.02477
                                                                        -0.1925
                                                                        -0.1092
RVP
                                                                       -0.04745
                                                                       -0.03752
                                                                            ---
                                                                            ---
T50
                                                                         0.1535
                                                                       0.092137
                                                                         0.1830
                                                                         0.1452
T90
                                                                       0.009353
                                                                         0.1114
                                                                               
                                                                         0.1270
etOH^2
                                                                         0.0414
                                                                               
                                                                               
                                                                               
T50^2
                                                                        0.07934
                                                                        0.06726
                                                                               
                                                                               
etOH*arom
                                                                        0.02172
                                                                        0.07561
                                                                               
                                                                               
etOH*RVP
                                                                               
                                                                        -0.0303
                                                                               
                                                                               
T50*T90
                                                                        0.05249
                                                                               
                                                                               
                                                                               
etOH*T50
                                                                        0.03658
                                                                        0.04477
                                                                               
                                                                               
etOH*T90
                                                                        0.04798
                                                                               
                                                                               
                                                                               
Arom*RVP
                                                                        0.02616
                                                                               
                                                                               
                                                                               
Arom*T50
                                                                               
                                                                        0.07454
                                                                               
                                                                               
Arom*T90
                                                                        0.02022
                                                                        0.03430
                                                                               
                                                                               
RVP*T90
                                                                               
                                                                               
                                                                               
                                                                               
Variance (sε2)
                                                                         0.1949
                                                                          1.853
                                                                         0.2750
                                                                          2.822
Additional Adjustments for Benzene and 1,3-Butadiene
      For two compounds, benzene and 1,3-butadiene, additional refinements were applied to address limitations of the EPAct study design. For benzene, the issue was that the fuel matrix included aromatics generally, but not benzene specifically. We considered it inadequate to model benzene in exhaust without explicitly accounting for benzene levels in fuel, so we chose to develop a post-hoc refinement using data external to the EPAct program.  In this case the source was a program conducted in support of the 2007 MSAT2 rule.  This program performed measurements on nine Tier 2 certified vehicles on fuels with benzene levels ranging from 0.6 to 1.1 percent by weight.[,]  With benzene represented as a fraction of VOC (as in Equation 4) denoted as fbenzene, a value modified to account for benzene levels in different fuels (f*benzene) is calculated as shown in  Equation 5 where xbenzene is the benzene level for the fuel modeled (weight percent),  A is the mean benzene level in the EPAct exhaust program fuel set (0.66 weight percent), and B is a coefficient, taking a value of 0.24. 

                                       
Equation 5.  Benzene Correction
Similarly, given the importance of olefins to estimation of emissions for 1,3-butadiene, and the fact that the EPAct exhaust program study design did not incorporate olefins as a factor, we developed a post-hoc adjustment explicitly accounting for olefin level. This adjustment was derived by varying olefin levels in the Complex Model and fitting a polynomial trend to the results. Starting with an unadjusted toxic fraction for 1,3-butadiene (fbuta), the modified fraction f[*]buta is calculated using Equation 6, in which xolefin is the olefin level, and A, B, C and D are coefficients, taking values of 0.000008, 0.0002, 0.0069 and 0.008823, respectively.
                                       
                                       
Equation 6.  1,3-Butadiene Correction
      To estimate emissions for other gaseous air toxics, additional updates were incorporated into MOVES for this analysis. Additional compounds include 2,2,4-trimethylpentane, ethylbenzene, n-hexane, methyl-tert-butyl ether, propionaldehyde, styrene, toluene, and xylenes.  For these pollutants, toxic-to-VOC ratios were derived from speciation profiles and applied to VOC as estimated by MOVES.  These ratios varied between pre-2004 and 2004 and later vehicles, and also among E0, E10 and E15 blends.  For more information, see Section IV.B. below.

Additional Fuel-Related Changes to MOVES

   Quantifying emissions from the proposed Tier 3 rule required a number of other minor changes to the MOVES2010 fuel databases and fuel effects code.
   Because MOVES2010 models distinct fuel supplies only through 2012, we updated the fuel supply support tables to supporting tables to allow distinct fuel supplies for calendar years through 2050.  Also, as described in Section III.A below, we updated the permeation calculation to allow modeling of E15 fuels.
   Additional fuel related changes to the MOVES model include the addition of T50 and T90 as fuel parameters in the FuelFormulation table (these parameters will be calculated from E200 and E300 if not present, otherwise the MOVES model will default to using T50 and T90 values); a correction to an error in the fuel effects generator; and a correction to ensure that ethanol emissions are modeled as zero in counties where ethanol fuel blends are not present. Other fuel-related changes included replacing the "NULL" values in the default fuel descriptions with actual values, and deleting most fuels from the MOVES default database, since they were provided in county-specific databases.   We also changed the operation of the fuel importer to require users to import all fuel formulations needed for a given county, year, and month.


Evaporative Emission Updates
   Evaporative emissions account for a significant portion of the gaseous hydrocarbon inventory.  For MOVEST3NPRM, the evaporative emission calculations were updated significantly.   We changed the calculation of permeation emissions to account for the E15 fuels modeled in the reference and control scenarios.  More importantly, we updated the calculation of vapor venting emissions to account for results from a multi-year research program that improved our understanding of the vapor venting process, the emission consequences of long-duration parking ("multi-day diurnals") and the prevalence of vapor leaks. These updates are detailed in the sections below.  We made no changes to the modeling of liquid leaks.
Permeation
   MOVES2010 already included effects of E10 on permeation emissions for newer technology vehicles, based on data collected by Coordinating Research Council (CRC) in the E-65.3 program and as part of the more recent CRC E-77 series of evaporative emissions programs. These data showed a significant change in the relative impact of E10, from a 65 percent increase for pre-enhanced vehicles, to a 213 percent increase for newer technology vehicles.   The CRC E-77 analysis also confirmed the E-65.3 finding that there is no significant difference between emission effects on E5.7 and E10.  
   The use of E15 in the reference and control fuel supplies modeled for the Tier 3 NPRM required that MOVES model E15 permeation effects as well.  The E-77 program evaluated permeation emissions on E20 as well as E10, and found that there was not a significant difference in permeation emissions between these two blends.  These data, in conjunction with the E-65.3 data discussed above, suggest that any level of ethanol in the fuel results in large increases in permeation emissions, but that the magnitude of increases are largely insensitive to how much ethanol is in the fuel.  As a result, in MOVEST3NPRM, we applied the E10 permeation effects to E15 as well.  
   These data also indicated significant impacts on toxics that differed from other types of evaporative emissions.  The toxics impacts were not analyzed in time for inclusion in MOVES2010, but were included in MOVEST3NPRM, as explained in Section IV.B.1.b) ,below. 
Vapor Venting
   When a vehicle's fuel system warms (due to ambient temperature rise while a vehicle is parked, or from vehicle operation), fuel vapors are generated. These may escape into the atmosphere if the capacity of the evaporative control system is exceeded, or through system leaks.  
   In particular, as a vehicle sits, without driving, through multiple diurnal cycles, the carbon canister accumulates vapor every day. It can only adsorb vapor until it reaches its finite capacity; then it begins to vent to the atmosphere. A canister with degraded/damaged carbon may have reduced capacity, but eventually every canister will vent to the atmosphere when the carbon reaches saturation. Upon reaching saturation, a canister acts in a relatively predictable manner. During cooling hours, the canister backpurges to the fuel tank and regains some of its capacity. Then, during the subsequent warming period, that capacity is re-filled with vapor and any excess vapor generated will escape to the atmosphere.
   MOVES2010 has a specific Tank Vapor Venting emission module to account for vapor venting emissions, but does not account for soaking periods of longer than 24 hours.  Based on new research findings, updates to MOVES were required to account for longer duration diurnal events, to improve the distinction between evaporative canister breakthrough and vapor leaks, and to reflect new vapor venting emissions data from the CRC E-77 program.  
      
Update of Vapor Venting Algorithms with new DELTA Model
   MOVES2010 estimates vapor venting emissions by calculating vapor generated as a function of temperature rise and fuel RVP, and a test-data-based relationship between the amount of vapor vented compared to the amount generated based on available diurnal emission test data. The MOVES2010 evaporative emissions estimates were based on in-use data collected between 1994 and 1996. This dataset provides SHED evaporative emissions (in grams) for a large set of vehicles undergoing a single diurnal cycle.  Although the data used in MOVES 2010 provided a large and robust dataset for one day of emissions, and the resulting emissions rates provided a good estimate of average overall venting emissions, we wanted to improve our ability to model emissions in situations where the cannister capacity might be exceeded, such as when vehicle uses high RVP fuels (e.g. E15) or parks for multiple days.
   To improve MOVES' ability to model this breakthrough effect, the "Diurnal Emissions Lost to Atmosphere" (DELTA) model was developed.This model applied theoretical vapor generation models to data from the CRC E-77 program on 9 and 10 RVP fuel, to produce vapor venting emission rates for MOVES.  
Multiple Day Diurnals
   The DELTA model described above provided a way for us to estimate emissions for long-duration parking (multi-day diurnals) episodes, but did not address the question of the frequency of such events.  While there have been many activity studies that track vehicle operation, very few have focused on the question of vehicle inactivity.  The only source of data we found to address this question was a study performed by Georgia Tech, which instrumented vehicles for two year time periods, and also surveyed participating households on the driving patterns of all vehicles in a household.  It was found that 16 percent of all household vehicles surveyed drove less than 3,000 miles per year. We lack information to know how many of these vehicles are driven on frequent short trips or very infrequent longer trips, but assuming that at least half of these vehicles have long soaks or drive without purging the canister, we used this value to estimate that at a given time 8 percent of the fleet has been soaking for 5 days or longer (the maximum soak length we've included in MOVES).  For implementation in MOVES, the number of vehicles experiencing day 2 through 5 of a diurnal are modeled as a percent of the vehicles that have been soaking since the beginning of the day. The resulting estimate of the percent of vehicles soaking for multiple days, by time of day, is shown in below.
   
                                       
                                       
Figure 1  Percentage of Vehicles Soaking By Time of Day
   Although a relatively small number of vehicles soak beyond one day, they are important because their emissions are disproportionately high.  The MOVEST3NPRM code was modified to incorporate the DELTA model and to model the complex processes of backpurge and canister capacity over multiple days.  
Vapor Leaks

   MOVES was also updated to reflect the prevalence of vehicles with evaporative leaks based on field research conducted in Denver between 2008 and 2010.  In MOVES2010, emissions for properly functioning and leaking vehicles were averaged into a single rate. In MOVEST3NPRM, we have distinguished the emission rate for a properly functioning evaporative control system from the rate for vehicles with leaks. A leaking vehicle essentially 'bypasses' the canister system, and generated vapor will escape through the leak location.
   
      In the Denver field studies, a remote sensing device (RSD) was used to recruit vehicles which were prepped and measured in a PSHED (Portable Sealed Housing for Evaporative Determination). The vehicles' hydrocarbon emissions were recorded for 15 minutes and leaking vehicles were inspected to identify the cause/source of the vapor leak. Using the individual vehicle data along with the data on the population of vehicles passing the RSD, the prevalence of vapor leaks in the fleet was estimated (see Figure 2).

      We have classified a "vapor leaker" as any vehicle that would fail the enhanced evaporative standard of 2 grams. The standard sums the emissions from the worst day of a 3-day diurnal test and the hot soak. To develop a surrogate standard for a 15 minute hot soak test, we used knowledge of certification testing to attribute 0.4g of the 2g standard to the hot soak portion, and, 0.3 grams of 0.4 grams to the first 15 minutes of the hour-long hot soak test. This approach results in 0.3g as a surrogate standard for a 15 minute hot soak. 




Figure 2 Estimated Distribution of PSHED values in the Ken Caryl Fleet by Model Year Group
      
      When observing the difference between two points in the above figure, one can determine how many vehicles fall into a particular emission range. For example, in model year group 1981-1995, 2.6% of vehicles are leaking at more than 20g/15min and 6.5% of vehicles are leaking at  more than 10g/15min. Subtracting these two values yields that 3.9% of vehicles in the model year group have a leak between 10g/15min and 20g/15min. This process reveals that older vehicles not only have more leaks, but larger leaks.
      
      Because the detailed, by model year, data only contain prevalence rates for PSHED measurements as low as 1.0g/15min, failure rates are extrapolated to 0.3g/15min. Using aggregate data from the Ken Caryl station, it is found that 0.3g/15min PSHED measurements are 50% more prevalent than 1.0g/15min PSHED measurements.

      Linear extrapolation and interpolation is used to model vapor leak prevalence for ages and model years not represented in this cross-sectional study. We assume newer cars will have lower leak prevalence than older cars. We also divide model year groups at the year new technology was introduced.  However, because we know vapor leaks also occur due to tampering, mal-maintenance, and other non-durability related issues, technology is not the only factor in a vehicle having a vapor leak. The older model year regressions always  have a more severe deterioration than newer model years. Failure rates at different severities are included below for two of the different age groups (see Figures 3 and 4). 


Figure 3 Extrapolated Prevalence Rates for Binning Leaks in MOVES, Age 0-3


Figure 4 Extrapolated Prevalence Rates for Binning Leaks in MOVES, Age 20+

   
Summary of Updated MOVES Inputs for Vapor Venting
   The MOVES cumtvvcoeffs table in the default database was updated to include the new inputs required to more accurately model cold soak vapor venting emissions and separate non-leaking from leaking vehicles. The new fields are described below in 
   Table 6.

Table 6 New Columns in `cumtvvcoeffs' Table
Column Name
Brief Explanation
How it's used
backPurgeFactor
The percent of canister capacity regained during the diurnal cooling cycle. 
For example: A canister with a gasoline vapor capacity of 200 grams fills to 100 grams in the first day. With a backPurgeFactor of 35%, it will begin day 2 filled to 65 grams.
averageCanister
This is the sales-weighted average Gasoline Working Capacity. Canisters have typically increased in size over time and with new regulations.
When vapor generated > averageCanister capacity, emissions are released to the environment. This is also the maximum value to which backPurgeFactor can be applied.
tvvEquation
An equation representing how much vapor is vented per unit vapor generated for non-leaking vehicles
These emissions are weighted with the leaking vehicle emissions estimated with leakEquation.
leakEquation
A linear equation representing how much vapor is vented per unit vapor generated for leaking vehicles
These emissions are weighted with the vapor venting emissions estimated with tvvEquation.
leakFraction
The fraction of vehicles leaking. Used to weight together tvvEquation and leakEquation.  The leak fraction is a function of model year and age.
weightedLine = leakFraction*leakEquation+(1-leakFraction)*tvvEquation
tankSize
The Sales-weighted average size of gas tank.
The tankSize is a key factor in calculating how much vapor is generated. A larger tank with the same tankFillFraction as a smaller tank will generate more vapor due to a larger headspace.
tankFillFraction
The average fill level of gas tanks.
This is applied to the tankSize to calculate how much headspace is available for vapor generation.

      A key improvement facilitated by these new coefficients is the development of tvvEquation and leakEquation. The emission curves that were previously in the model were drawn in grams of vapor generated per gallon of headspace, and only intended for one day of emissions. The new curves use total grams as the unit and are intended to model the complex processes of backpurge, canister capacity, and multiple days. The tvvEquations are complex exponential curves created using data from the E-77 programs using the DELTA model. The leak equations are linear equations developed using emission rates of known leakers. The data on leaking vehicles comes from the E-77 program and Colorado High Evap program. These two curves are generated in the database for each source type, model year group, and age group. Each set of curves has a leak fraction (representing the percent of leaking vehicles in that population) that is used to weight the two curves together to produce a composite curve.
   
      We model mission rates for vapor leaks as a linear relationship between total tank vapor generated (TVG) and tank vapor vented (TVV). TVG is calculated using the Wade-Reddy equation which determines TVG as a function of fuel RVP, fuel ethanol content, temperature rise, and tank fill level. There are two datasets used to develop vapor leak emission rates. The E-77 program supplied information on high emitting vehicles, with known fuel properties and implanted leaks, that were tested on the California diurnal cycle. The Colorado program provided information on 16 leaking vehicles. They performed a 6-hour SHED test consisting of a controlled heat rise from 72F-96F. This test provides us with enough information to calculate TVG. For both programs, the Sealed Housing for Evaporative Determination (SHED) emission measurements along with calculated TVG obtain enough information to apply a linear regression model of the relationship between fuel vapor generated and fuel vapor vented.
   
      Vapor leak emissions span many orders of magnitude given their diverse locations and sizes. Since our leak prevalence estimates were measured at high altitude in Denver, we must use a correction factor for the way we bin the leaking vehicles. Using the Wade-Reddy equation, we can calculate that Denver E10 fuel generates 1.41 times as much vapor as Sea Level gasoline. For example, a vapor leak at the rate of .3g/15min in Denver will equate to a .21g/15min leak at Sea Level. The bins used to categorize vapor leak severity as well as the average vapor leak emission rate for that bin are listed in the table below.

Table 7 Vapor leak emission rates by bin
                           Denver bins (grams/15min)
                                Sea Level bins
                        Grams vented / grams generated
                                     0.3-2
                                   0.21-1.42
                                    0.1191
                                      2-5
                                   1.42-3.55
                                    0.2686
                                     5-10
                                   3.55-7.09
                                    0.6492
                                    >10
                                   >7.09
                                    1.3273
   
      The hot soak measurements from the E-77 programs and the PSHED (Portable SHED) measurements from the Denver program are used to bin each vehicle. In addition, the corresponding diurnal test provides information necessary to determining the leak emission rate.

For each model year and age group combination, the vapor leak prevalence for each severity bin is applied to the corresponding vapor leak emission rate. The result is a vapor leak emission rate representative of all vapor leaks greater than 0.3g/15min. This representative leak emission rate is then weighted by the leak prevalence with the non-leaking emission rate. See Figure 5.


Figure 5 Example of how leak emission rates are calculated using prevalence rate bins

Update of Air Toxics and HC Speciation
   
   The calculation of air toxics and the speciation of hydrocarbons was updated in the version of MOVES used for the Tier 3NPRM analysis.  As described in Section II.B above, the EPAct analysis led to changes in the fuel effects for some major air toxics.   Other changes to MOVES included the capability to estimate emissions for many air toxics not available in MOVES2010a (but now available in MOVES2010b).  The section below details these changes.
   Also, the documentation of toxic calculations for MOVES2010 did not include details on the calculation of some toxics for 2003-and-earlier gasoline vehicles.  These calculations are unchanged from MOVES2010, but are documented here for completeness.
Algorithms for 2003-and-Earlier Gasoline Vehicles
   
   For 2003-and-earlier model year gasoline vehicles, the algorithms for calculating emissions of major air toxics are unchanged from MOVES2010a.  The algorithms used for benzene, 1,3-butadiene, formaldehyde, and acetaldehyde are documented in this subsection for completeness.
     For 2003-and-earlier light-duty cars and trucks, all heavy-duty gasoline vehicles and all motorcycles, algorithms used to calculate toxic ratios are the same as in MOBILE6.2.  In MOBILE6.2 toxic ratios between light-duty gasoline vehicles and heavy-duty gasoline vehicles differed; however, since the only available data for heavy-duty gasoline vehicles were extremely limited, and mostly from vehicles with either no emission controls or early oxidation catalysts, for MOVES we decided to apply toxic to VOC ratios from light-duty gasoline vehicles to VOC emissions for this vehicle class.
    	Toxic to VOC ratios for benzene, 1,3-butadiene, formaldehyde, and acetaldehyde exhaust emissions from light-duty gasoline vehicles with three-way or three-way plus oxidation catalysts were estimated using algorithms developed for the Complex Model for Reformulated Gasoline.  For MTBE, a draft fuel effects model based on the Complex Model database was used.  This model is based on algorithms from underlying data of nearly 900 observations.  However, instead of using model algorithms directly, MOBILE6.2 was run at different fuel MTBE levels and a curve of fuel MTBE to MTBE/VOC ratios was developed.  This curve was then used in MOVES.  It should be noted that the sulfur effects terms in the algorithms were not used; instead, sulfur impacts on toxic emissions were assumed to be proportional to the sulfur impacts on total VOC estimated by MOVES.  The Complex Model algorithms are based on data from vehicles representing a 1990 model year fleet.      
   For benzene, 1,3-butadiene, formaldehyde, and acetaldehyde, the algorithms are based on about 1800 observations.  The algorithms are applied by stratifying the light-duty gasoline fleet into ten Technology Groups and applying the algorithms individually to each group (this is known as the unconsolidated Complex Model).  The ten groups are formed as a combination of fuel system, catalyst type, air injection (y/n toggle), EGR, and Normal / High emitter status.  These groups are listed in Table 8. The first nine groups represent only normal emitting vehicles.  The tenth group represents all of the high emitters, regardless of technology.  Separate toxic ratios are calculated for each of the groups and weighted together by the fraction of the fleet attributable to each technology group in MOBILE6.  
   

Table 8 Technology groups in the Complex Model for Reformulated Gasoline
                        
                         Technology Group Definitions
                               Technology Group
                                  Fuel System
                                   Catalyst
                                 Air Injection
                                      EGR
                                       1
                                      PFI
                                     3WAY
                                    NO AIR
                                      EGR
                                       2
                                      PFI
                                     3WAY
                                    NO AIR
                                    NO EGR
                                       3
                                      TBI
                                     3WAY
                                    NO AIR
                                      EGR
                                       4
                                      PFI
                                   3WAY + OX
                                      AIR
                                      EGR
                                       5
                                      PFI
                                     3WAY
                                      AIR
                                      EGR
                                       6
                                      TBI
                                     3WAY
                                      AIR
                                      EGR
                                       7
                                      TBI
                                   3WAY + OX
                                      AIR
                                      EGR
                                       8
                                      TBI
                                     3WAY
                                    NO AIR
                                    NO EGR
                                       9
                                     CARB
                                   3WAY + OX
                                      AIR
                                      EGR
                                 High Emitters
                                      ALL
                                      ALL
                                      ALL
                                      ALL

PFI = port fuel Injection, TBI = throttle body injection, CARB = carburetor, 3WAY = three way catalyst, 3WAY + OX = three way plus oxidation catalyst, ERG = exhaust gas recirculation


   For light-duty gasoline vehicles with oxidation catalysts only or no catalysts, toxic to VOC ratios for these pollutants are determined using algorithms derived from a more limited data set of about 50 vehicles tested on a baseline fuel and a small number tested on reformulated fuels.  Data were not available to develop algorithms for ETBE and TAME blends; thus, the algorithms for ethanol oxygenated gasoline were used for ETBE blends, and the algorithms for MTBE oxygenated gasoline were used for MTBE blends.  
   For vehicles running on 10 percent ethanol, 2.39 percent of exhaust VOC was estimated to be ethanol.  This estimate is based on tests on nine vehicles from four test programs., , ,   MOVES also has ethanol to VOC ratios for E5 and E8, based on linear interpolation from the E10 value (0.01195 for E5 and 0.01912 for E8).

 Estimating Emissions for Additional Air Toxics

   The version of MOVES used for the Tier 3 NPRM analysis was expanded to include all air toxic pollutants in the National Emission Inventory (NEI) and National Air Toxics Assessment (NATA) that are emitted by mobile sources.  This list of pollutants is provided in Table 9.  These pollutants are organized into four categories:

   1) Gaseous hydrocarbons
   2) Polycyclic aromatic hydrocarbons (PAHs)  -  These hydrocarbons which contain fused aromatic rings can be found in the gas phase, particle phase, or both, depending on properties of the compound, particle characteristics, and atmospheric conditions
   3) Dioxins and furans  -  polychlorinated organic compounds which are persistent and bioaccumulative
   4) Metals

   Emissions for gaseous HAPs and PAHs are reported in analyses for the Tier 3 regulatory analysis.  The pollutant "xylenes" represents the sum of emissions from three isomers of xylene:  o-xylene, m-xylene, and p-xylene.  Emissions of metals and dioxins and furans are not reported, since MOVES estimates these using straight emission rates which do not change with fuel properties or criteria pollutant emission levels.  Thus, scenarios being evaluated will have no impact on inventories for these pollutants.  
   
Table 9 Air toxics included in MOVESTier3NPRM
Pollutant
                              MOVES Pollutant ID
CAS Number
Gaseous Hydrocarbons
                                       

1,3-Butadiene
                                      24
106990
2,2,4-Trimethylpentane
                                      40
540841
Acetaldehyde
                                      26
75070
Acrolein
                                      27
107028
Benzene
                                      20
71432
Ethanol
                                      21
64175
Ethyl Benzene
                                      52
100414
Formaldehyde
                                      25
50000
Hexane
                                      55
110543
Methyl Tert Butyl Ether (MTBE)
                                      22
1634044
Propionaldehyde
                                      58
123386
Styrene
                                      60
100425
Toluene
                                      61
108883
Xylenes
                                      62
1330207
PAHs
                                       

Acenaphthene
                                      41
83329
Acenaphthylene
                                      42
208968
Anthracene
                                      44
120127
Benz(a)anthracene
                                      45
56553
Benzo(a)pyrene
                                      46
50328
Benzo(b)fluoranthene
                                      47
205992
Benzo(g,h,i)perylene
                                      48
191242
Benzo(k)fluoranthene
                                      49
207089
Chrysene
                                      50
218019
Dibenzo(a,h)anthracene
                                      51
53703
Fluoranthene
                                      53
206440
Fluorene
                                      54
86737
Indeno(1,2,3,c,d)pyrene
                                      56
193395
Naphthalene
                                      23
91203
Phenanthrene
                                      57
85018
Pyrene
                                      59
129000

Additional Air Toxics:  Inputs for Gasoline Vehicles

      Data methodologies and sources are provided for gasoline vehicles running on E0, E10, E15 and E20 gasoline.
Exhaust
    Exhaust inputs developed for the additional air toxics do not vary by temperature or operating mode.  While empirical data suggest toxic to VOC and toxic to PM ratios differ for start versus running emissions, we decided not to vary these inputs in the model by operating mode for two key reasons.  First, for advanced technology vehicles, HAP emissions during running operation are often below levels of quantitation in test programs.   Thus, using data from running operation only would necessitate heavy reliance on imputed data.  Second, with the large number of toxics modeled, applying separate inputs would be place additional computational demands on the model.  Ambient temperature also affects these ratios, but we assumed toxic emission of gaseous toxics and PAHs change proportionally with VOC and PM as temperature varies.  In general, data are inadequate to account for temperature effects on toxic to VOC and toxic to PM ratios.  Emission rates for metals are assumed to be independent of operating mode and temperature; very little data exist to account for impacts of these parameters.

   For 2003 and earlier vehicles, toxic to VOC ratios for E0 an E10 were developed by Sierra Research using estimates from EPA's SPECIATE 4.2 database (Table 10).  Toxic to VOC ratios for MTBE blends, however, were obtained from the National County Database in the National Mobile Inventory Model (NMIM).  NMIM 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. Ratios for LDG vehicles were used for all gasoline vehicle classes.  These ratios were based on older speciation profiles than the E0 and E10 data.  Separate ratios are used for each of the following four categories of gasoline blends:
      
1.  Baseline Gasoline. All cases that do not fall into categories 2-4 below. 
2.  10 perecent Ethanol Gasoline - Used where the fuel contains ethanol which is greater than or equal to 5 percent by volume or ETBE greater than or equal to 5 percent by volume. 
3.  Winter Oxygenate MTBE/TAME Gasoline (MTBE1) - Used where the fuel contains MTBE which is greater than or equal to 12 percent by volume or TAME greater than or equal to 13percent by volume. 
4.  RFG/MTBE/TAME (MTBE2) - Used where the fuel is RFG and where the fuel contains oxygenate greater than 5percent by volume and where the fuel contains MTBE which is less than 12percent by volume or TAME less than 13 percent by volume. 

Table 10 Toxic to VOC ratios for selected air toxics from Tier 1 and earlier gasoline vehicles and trucks
                                Pollutant
Name
                                      E0
                                      E10
                                       
                                     MTBE1
                                       
                                     MTBE2
                            2,2,4-Trimethylpentane
                                    0.01823
                                    0.01849
                                    0.04327
                                    0.04327
                                   Acrolein
                                    0.0006
                                    0.0006
                                    0.0006
                                    0.0006
                                 Ethyl Benzene
                                    0.02147
                                    0.01932
                                    0.01398
                                    0.01484
                                    Hexane
                                    0.01570
                                    0.01593
                                    0.00861
                                    0.00888
                                Propionaldehyde
                                    0.00086
                                    0.00086
                                    0.00073
                                    0.00073
                                    Styrene
                                    0.00108
                                    0.00097
                                    0.00328
                                    0.00340
                                    Toluene
                                    0.09619
                                    0.08657
                                    0.09873
                                    0.10494
                                    Xylene 
                                    0.07814
                                    0.07032
                                    0.05557
                                    0.05910

   For 2004-and-later vehicles running on E0 and E10, and for all vehicles running on E15, ratios for additional gaseous HAPs for E0, E10 and E15 were obtained from Phase 1 of the EPAct test program.  These profiles were based on tests from 3 vehicles.  Toxic to VOC ratios for E10 are used in MOVES for all gasoline above 5 percent ethanol by volume.  Ratios are provided in Table 11.

Table 11 Toxic to VOC ratios for selected air toxics from Tier 2 gasoline vehicles and trucks
                                       
                                Pollutant
Name
                                      E0
                                      E10
                                       
                                      E15
                            2,2,4-Trimethylpentane
                                    0.03188
                                    0.01227
                                    0.0220
                                 Ethyl Benzene
                                    0.01683
                                    0.01660
                                    0.0157
                                    Hexane
                                    0.00279
                                    0.02911
                                    0.0110
                                Propionaldehyde
                                    0.00122
                                    0.00054
                                    0.0006
                                    Styrene
                                    0.00085
                                    0.00083
                                    0.0046
                                    Toluene
                                    0.07542
                                    0.07440
                                    0.0727
                                    Xylene 
                                    0.06127
                                    0.06047
                                    0.0690
                                       
   In the NMIM model, PAH emissions were estimated as ratios to PM.  This approach was used, even though PAHs are found in the gas, semi-volatile and particle phase, because there is generally reasonable correlation between PAH and PM emissions.[,]   However, for MOVESTier 3NPRM, PAH mass emissions are apportioned into the gas and particle phase, and gas phase PAHs are estimated using toxic/VOC ratios and particle phase PAHs were estimated using toxic to OC2.5 (organic carbon less than 2.5 microns) ratios.  Although partitioning of PAHs into the gas and particle phases depends on concentration, temperature and other factors, MOVES applies one set of allocation factors under all conditions in order to streamline data processing.  These allocation factors were developed by Sierra Research using estimates from EPA's SPECIATE 4.2 database and information on compounds' physical and chemical properties.  Table 12 provides molecular weights and allocation factors.  The allocations from SPECIATE were based on medium duty diesel engine data.  Using PAH and VOC emissions data from Norbeck et al. (1998) and the allocation factors in Table 12, the toxic/VOC ratios in  Table 14 were developed.  Norbeck et al. data were also used to develop toxic/PM10 ratios.  These ratios were then converted to toxic/OC2.5 ratios for start and running operation.  These conversions were made using assumptions that 90 percent of PM10 is PM2.5, 87.3 percent of running PM2.5 is OC, and 66.6 percent of start PM2.5 is OC.  OC percentages of PM2.5 were obtained from MOVES runs.  PAH to VOC and PAH to OC2.5 ratios do not vary between E0 and E10 fuels.

Table 12 PAH gas/particle phase allocation factors
                                  PAH Species
                               Molecular
Weight
                              Allocation Fraction
                                       
                                       
                                   PM Phase 
                                 Gaseous Phase
Acenapthene
                                      154
                                       0
                                       1
Acenaphthylene
                                      152
                                       0
                                       1
Anthracene
                                      178
                                     0.466
                                     0.534
Benz(a)anthracene
                                      228
                                     0.723
                                     0.277
Benzo(a)pyrene
                                      252
                                       1
                                       0
Benzo(b)fluoranthene
                                      252
                                       1
                                       0
Benzo(g,h,i)perylene
                                      276
                                     0.773
                                     0.227
Benzo(k)fluoranthene
                                      252
                                       1
                                       0
Chrysene
                                      228
                                     0.823
                                     0.177
Dibenzo(a,h)anthracene
                                      278
                                       1
                                       0
Fluoranthene
                                      202
                                     0.516
                                     0.484
Fluorene
                                      166
                                     0.215
                                     0.785
Indeno(1,2,3-cd)pyrene
                                      276
                                       1
                                       0
Phenanthrene
                                      178
                                     0.335
                                     0.665
Pyrene
                                      202
                                     0.552
                                     0.448

















	Table 13  PAH/VOC and PAH/PM ratios for gasoline vehicles and trucks.
PAH
Start Fraction of OC2.5
Running Fraction of OC2.5
Fraction of VOC
Benzo(a)anthracene
                                                                      0.0001237
                                                                      0.0000944
                                                                      0.0000007
Benzo(a)pyrene
                                                                      0.0001711
                                                                      0.0001306
                                                                      0.0000000
Benzo(b)fluoranthene
                                                                      0.0002032
                                                                      0.0001551
                                                                      0.0000000
Benzo(k)fluoranthene
                                                                      0.0002032
                                                                      0.0001551
                                                                      0.0000000
Chrysene
                                                                      0.0001408
                                                                      0.0001075
                                                                      0.0000005
Dibenz(a,h)anthracene
                                                                      0.0000000
                                                                      0.0000000
                                                                      0.0000000
Indeno(1,2,3-cd)pyrene
                                                                      0.0001283
                                                                      0.0000979
                                                                      0.0000000
Acenaphthene
                                                                      0.0000000
                                                                      0.0000000
                                                                      0.0000185
Acenaphthalene
                                                                      0.0000000
                                                                      0.0000000
                                                                      0.0001040
Anthracene
                                                                      0.0006577
                                                                      0.0005020
                                                                      0.0000114
Benzo(ghi)perylene
                                                                      0.0003306
                                                                      0.0002524
                                                                      0.0000015
Fluoranthene
                                                                      0.0007834
                                                                      0.0005980
                                                                      0.0000111
Fluorene
                                                                      0.0005425
                                                                      0.0004141
                                                                      0.0000300
Naphthalene
                                                                      0.0000000
                                                                      0.0000000
                                                                      0.0022928
Phenanthrene
                                                                      0.0014184
                                                                      0.0010827
                                                                      0.0000427
Pyrene
                                                                      0.0011450
                                                                      0.0008740
                                                                      0.0000141
                                       
                                       
Evaporative and Permeation
   Toxic to VOC ratios for evaporative non-permeation emissions of benzene, ethanol, MTBE and naphthalene remain unchanged from those in MOVES2010.   In summary, benzene and MTBE ratios are estimated using algorithms originally developed for MOBILE 6.2.  However, evaporative emissions processes for MOVES differ from those in MOBILE6.2.  Thus, algorithms for hot soak in MOBILE6.2 are used for vapor venting and refueling vapor loss in MOVES, and algorithms for running loss are used for fuel leaks and refueling spillage loss (Table 14).  Ratios for naphthalene and ethanol are given in Table 15.  
   Ratios for the additional air toxics found in evaporative non-permeation emissions (2,2,4-trimethylpentane, ethyl benzene, hexane, propionaldehyde, toluene and xylenes) were obtained from profiles developed for EPA by Environ Corporation, using data from the Auto/Oil test program conducted in the early 1990's.  These ratios are also given in Table 15.  Conventional gasoline ratios are also used for MTBE oxygenated gasoline.  Since no emissions data exist for evaporative and permeation emissions from E15 vehicles, E15 and E10 fuel speciation data from the EPAct test program were used to adjust the E10 evaporative emissions speciation.  For all pollutants except benzene and MTBE, ratios are the same for all types of non-permeation evaporative emissions.  The ratios for 10 percent ethanol are used for all fuels with greater than or equal to 5 percent ethanol and less than 12 percent.

Table 14 Toxic/VOC algorithms for benzene and MTBE gasoline vehicle evaporative emissions


                                   Pollutant

                                    Process

                      Toxic Fraction Equation (Toxic/VOC)

                                    Benzene

                         Vapor Venting/Refueling Vapor
                              Fuel Leaks/Spillage

                (-0.03420*OXY - 0.080274*RVP + 1.4448)*BNZ/100
                (-0.03420*OXY - 0.080274*RVP + 1.4448)*BNZ/100

                                     MTBE

                         Vapor Venting/Refueling Vapor
                              Fuel Leaks/Spillage

                        (24.205 - 1.746*RVP)*MTBE/1000
                        (17.8538 - 1.6622*RVP)*MTBE/1000
 
Note:	OXY = wt percent oxygen
RVP = Reid vapor pressure in psi
BNZ = vol percent benzene
MTBE = vol percent MTBE


Table 15 Evaporative Toxic/VOC ratios for other toxics from gasoline vehicles (Auto/Oil)

Pollutant
                                      E0
E10
E15
Ethanol
                                    0.00000
                                    0.11896
                                   0.193458
Naphthalene
                                    0.00040
                                    0.00040
                                   0.000000
2,2,4-Trimethylpentane
                                    0.01984
                                    0.03354
                                   0.053129
Ethyl Benzene
                                    0.02521
                                    0.01721
                                   0.016616
Hexane
                                    0.02217
                                    0.02536
                                   0.007478
Toluene
                                    0.09643
                                    0.14336
                                   0.140632
Xylene
                                    0.07999
                                    0.06423
                                   0.057348
                                       
   
   The composition of permeation emissions differs significantly from other types of evaporative emissions. Work to better characterize these permeation emissions was recently conducted by Southwest Research Institute for EPA and the Coordinating Research Council.  Data from 3-day diurnal tests on vehicles meeting Tier 1 and near zero evaporative emission standards were used.  Toxic to VOC ratios for E0 and E10 were estimated by averaging data from gasolines of different RVPs.  Resulting ratios are provided in Table 16, for all compounds except benzene, MTBE and naphthalene.  For benzene, the diurnal emissions algorithm from MOBILE6.2 was used instead, since this algorithm account for impacts of changing oxygenate, RVP and fuel benzene levels.  However, a study of permeation emissions suggests that the ratio of benzene from permeation to total VOC is about 1.77 times higher than the ratio associated with evaporation.  Thus the diurnal emissions algorithm was multiplied by 1.77.  This algorithm is:

             1.77[(-0.02895*OXY - 0.080274*RVP + 1.3758)*BNZ/100]
Equation 7 Diurnal Emissions of Benzene

In the absence of data on permeation emissions for MTBE, the resting loss algorithm from MOBILE6.2 was used:

                          (22.198-1.746RVP)*MTBE/1000
Equation 8 Diurnal Emissions of MTBE
For naphthalene, the toxic to VOC ratio for non-permeation evaporative emissions was also applied to permeation.

   For E15, model inputs were developed by interpolating between E10 and E20 permeation toxic/VOC ratios.  The E10 and E20 ratios were obtained from the CRC E-77-2b and CRC E-77-2c test programs.,  Inputs are provided in Table 16.

Table 16 Gasoline vehicle permeation toxic/VOC ratios for air toxics except benzene, naphthalene and MTBE (CRC E-77-2b)
                                       
Pollutant
E0
E10
2,2,4-Trimethylpentane
                                                                          0.036
                                                                          0.024
Ethyl Benzene
                                                                          0.003
                                                                          0.001
Hexane
                                                                          0.050
                                                                          0.065
Toluene
                                                                          0.110
                                                                          0.101
Xylene
                                                                          0.016
                                                                          0.011
Ethanol
 
                                                                          0.202
                                       
Additional Air Toxics:  Inputs for Emissions from Diesel Vehicles
   Toxic/VOC ratios, PAH/VOC ratios, PAH/PM ratios and metal emission factors were developed for exhaust emissions from heavy-duty diesel vehicles and applied to all diesel vehicle categories.  There are no separate emission ratios or factors for diesel engines running on biodiesel fuels or synthetic diesel fuels, due to very limited data. 
   
Exhaust
   
   The composition of VOCs for heavy-duty diesel engines without model year 2007-and-later emission controls versus those engines with such controls varies significantly.  Thus, we developed one set of toxic to VOC ratios for pre-2007 diesel engines and another set for 2007 and later engines.  Since extended idle emissions associated with auxiliary power units (APUs) are not subject to 2007 standards, toxic to VOC ratios for pre-2007 diesel engines were used for them.  Since light-duty diesels comprise a very small portion of the fleet, the same ratios were applied to all diesel vehicle classes to streamline modeling
   
   EPA relied on a database compiled for the Coordinating Research Council (CRC E-75) and National Renewable Energy Laboratory (NREL) to develop toxic to VOC ratios for pre-2007 engines.  This database was developed from a literature survey and included data from 13 different studies.  The studies included in this database were conducted in a number of different countries, included heavy-duty and light-duty engines, a variety of diesel and biodiesel fuels, and a number of different operating modes and cycles.  For 2,2,4-trimethylpentane, hexane, propionaldehyde, and toluene, toxic to VOC ratios developed by Sierra Research from CRC E-75 data were used.  The methodology they used to develop ratios is described in detail in their technical report.  Data from tests using non-conventional diesel fuel (Fischer-Tropsch, bioDiesel, ethanol-Diesel blends, emulsified fuel, European blends, and other obvious research fuels) were excluded, as were data from non-heavy duty engines.  The ratios are provided in Table 10.
   
   Toxic to VOC ratios for benzene, 1,3-butadiene, formaldehyde, acetaldehyde, acrolein, naphthalene, ethylbenzene, styrene and xylenes  were developed by EPA from the E-75 database.  We relied on United States data from heavy-duty diesel engines running on conventional diesel fuels, collected on test-cycles representative of real world operation.  Some studies measured emissions over distance, while other studies measure emissions relative to brake horsepower.  For studies which measured emissions relative to distance, we calculated mean emissions per mile for toxics and VOC, then calculated a ratio of toxics to VOC.  For studies which measured emissions relative to brake horsepower-hour, we calculated mean emissions per brake horsepower-hour for toxics and VOC, then calculated a second ratio of toxics to VOC.  We then calculated a composite ratio using sample size to weight the two ratios.  The resulting ratios are provided in Table 17.
   
Table 17 Toxic to VOC Ratios for pre-2007 diesel engines (CRC E-75)
                                       
Pollutant
                                   Toxic/VOC
1,3-Butadiene
                                   0.002918
2,2,4-Trimethylpentane
                                   0.001808
Acetaldehyde
                                   0.035559
Acrolein
                                   0.006622
Benzene
                                   0.007835
Ethyl Benzene
                                   0.002655
Formaldehyde
                                   0.078225
Hexane
                                    0.00197
Propionaldehyde
                                    0.00468
Styrene
                                   0.001312
Toluene
                                    0.00433
Xylenes
                                   0.003784
   
   For 2007 and later heavy-duty diesels, and 2004 and later light-duty diesels which meet Tier 2 vehicle standards, advanced emission controls change the composition of VOCs.  For these engines, we relied on speciated emissions data from the Advanced Collaborative Emissions Study (ACES), directed by the Health Effects Institute and Coordinating Research Council, with participation from a range of government and private sector sponsors.  Detailed emissions data from the study were provided to EPA at the request of the Coordinating Research Council.
   The data were collected on four engines on several test cycles with low sulfur diesel fuel.  EPA used data from a 16-hour transient cycle.  Toxic to VOC ratios obtained from the ACES data are provided in Table 18.
   

Table 18 Toxic to VOC Ratios for 2007 and later diesel engines

Pollutant
                                   Toxic/VOC
1,3-Butadiene
                                    0.00080
2,2,4-Trimethylpentane
                                    0.00782
Acetaldehyde
                                    0.06934
Acrolein
                                    0.00999
Benzene
                                    0.01291
Ethyl Benzene
                                    0.00627
Formaldehyde
                                    0.21744
Hexane
                                    0.00541
Propionaldehyde
                                    0.00314
Styrene
                                    0.00000
Toluene
                                    0.02999
Xylenes
                                    0.03800

   As with gasoline vehicles, PAH mass emissions from diesel engines were apportioned into the gas and particle phase, using a single set of allocation factors for all conditions.  Gas phase PAHs were estimated using toxic/VOC ratios and particle phase PAHs were estimated using toxic to OC2.5 ratios.  Toxic to VOC and toxic to PM2.5 ratios for pre-2007 diesel engines were developed by EPA from the E-75 database.  Toxic to PM2.5 ratios were converted to toxic to OC2.5 ratios using OC2.5 to PM2.5 percentages from MOVES.  We relied on United States data from heavy-duty diesel engines running on conventional diesel fuels, collected on test-cycles representative of real world operation.  It should be noted that for some PAHs, there were substantially more data than for others; thus the level of confidence in emission factors varies among individual compounds.  For instance, while data from 66 tests were available for acenaphthene, data from only two tests were available for dibenz(a,h)anthracene.  Table 19 provides VOC and PM ratios for PAHs from these older technology engines. 

Table 19 PAH/VOC and PAH/ OC2.5 ratios for pre-2007 diesel engines

PAH
Start Fraction of OC2.5
Running Fraction of OC2.5
Fraction of HC
Benzo(a)anthracene
0.0002100
0.0005806
                                                                      0.0000445
Benzo(a)pyrene
0.0000789
0.0002180
                                                                      0.0000000
Benzo(b)fluoranthene
0.0000234
0.0000648
                                                                      0.0000000
Benzo(k)fluoranthene
0.0000033
0.0000091
                                                                      0.0000000
Chrysene
0.0001327
0.0003667
                                                                      0.0000235
Dibenz(a,h)anthracene
0.0000032
0.0000087
                                                                      0.0000000
Indeno(1,2,3-cd)pyrene
0.0000060
0.0000166
                                                                      0.0000000
Acenaphthene
0.0000000
0.0000000
                                                                      0.0003210
Acenaphthalene
0.0000000
0.0000000
                                                                      0.0005009
Anthracene
0.0001058
0.0002924
                                                                      0.0002353
Benzo(ghi)perylene
0.0000038
0.0000104
                                                                      0.0000276
Fluoranthene
0.0004058
0.0011217
                                                                      0.0006108
Fluorene
0.0001818
0.0005025
                                                                      0.0005914
Naphthalene
0.0000000
0.0000000
                                                                      0.0090464
Phenanthrene
0.0004187
0.0011574
                                                                      0.0019446
Pyrene
0.0005865
0.0016213
                                                                      0.0007577


   
   For 2007 and later diesels, advanced emission controls change the composition and reduce the total mass of PAHs.  For these engines, we relied on speciated emissions data from the ACES study.  Table 20 provides provides VOC and OC2.5 ratios for PAHs from these newer technology engines.
   

Table 20 PAH/VOC and PAH/PM ratios for 2007 and later diesel engines

PAH
Start Fraction of OC2.5
Running Fraction of OC2.5
Fraction of HC
Benzo(a)anthracene
0.0000015
0.0000010
                                                                      0.0000003
Benzo(a)pyrene
0.0000061
0.0000042
                                                                      0.0000000
Benzo(b)fluoranthene
0.0000026
0.0000018
                                                                      0.0000000
Benzo(k)fluoranthene
0.0000026
0.0000018
                                                                      0.0000000
Chrysene
0.0000046
0.0000031
                                                                      0.0000005
Dibenz(a,h)anthracene
0.0000018
0.0000013
                                                                      0.0000000
Indeno(1,2,3-cd)pyrene
0.0000009
0.0000006
                                                                      0.0000000
Acenaphthene
0.0000000
0.0000000
                                                                      0.0000526
Acenaphthalene
0.0000000
0.0000000
                                                                      0.0000853
Anthracene
0.0000488
0.0000334
                                                                      0.0000304
Benzo(ghi)perylene
0.0000004
0.0000003
                                                                      0.0000002
Fluoranthene
0.0000896
0.0000613
                                                                      0.0000457
Fluorene
0.0000990
0.0000677
                                                                      0.0001963
Naphthalene
0.0000000
0.0000000
                                                                      0.0163278
Phenanthrene
0.0007886
0.0005395
                                                                      0.0008507
Pyrene
0.0000859
0.0000588
                                                                      0.0000379
                                       

Evaporative

   MOVES estimates evaporative emissions associated with spillage for diesel vehicles.  Since there are no speciated emissions of diesel spillage emissions, we developed toxic to VOC ratios based on a diesel headspace profile, profile number 4547 from the SPECIATE database.  These ratios are provided in Table 21.

Table 21 Toxic to VOC ratios for diesel spillage emissions
                                       
Pollutant
Toxic/VOC
2,2,4-Trimethylpentane
                                                                        0.00974
Ethyl Benzene
                                                                        0.00324
Hexane
                                                                        0.01076
Toluene
                                                                        0.01419
Xylene
                                                                        0.01222
Benzene
                                                                        0.00410
                                       
Hydrocarbon Speciation Factors

   MOVES provides estimates of hydrocarbon emissions in a number of different groupings or "hydrocarbon speciations."  The choice depends on application.   The following metrics are commonly used in EPA and produced by the MOVES model as documented in a MOVES technical report:
      
   Total hydrocarbons (THC) -- Organic compounds in exhaust, as measured by a flame ionization detector (FID). A FID does not accurately measure all of the mass of oxygenated organic gases, such as ethanol, acetaldehyde, and formaldehyde.
   Non-methane hydrocarbons (NMHC)  -  THC minus the mass of methane.
   Total organic gases (TOG) --  All organic gas compounds emitted to the atmosphere, including all of the mass of oxygenated organic gases.
   Non-methane organic gases (NMOG)  -  TOG minus the mass of ethane. 
   Volatile organic compounds (VOC)  -  Organic gas compounds that participate in atmospheric photochemical reactions.  Does not include the mass of methane and ethane.  Does include organic compounds not measured by the FID.

   Basic exhaust emission rates in MOVES are expressed as THC.  NMHC is then estimated by applying a methane adjustment provided in the MOVES database.  The following equation is used to calculate NMHC in MOVES: 

NMHC=THC-CH4
Equation 9 Calculating NMHC from THC

   The methane estimates used in the NMHC calculation were updated from MOVES2010 to MOVES2010a as described in an EPA technical report. The methane estimates for gasoline vehicles of model year 2004-and-later were further updated for MOVEST3NPRM as described in Section IV.C.1 below.   
   NMOG, TOG, and VOC are other MOVES outputs.  These metrics are obtained by applying NMHC-to-NMOG and NMHC-to-VOC speciation factors. Although TOG emissions were not calculated for the Tier 3 NPRM analysis, this version of MOVES changed the way TOG is calculated, setting it to equal NMOG  plus methane.  
   NMHC to VOC factors are calculated using the following equation:





	
Equation 10 NMHC to VOC

NMHC to NMOG factors are calculated using the following equation:






Where:

CF = carbon fraction
MPC = mass per carbon
FIDX = FID response factor

Equation 11 NMHC to NMOG

   The FID does not measure any of the mass of formaldehyde; thus for that pollutant a FID response factor of zero was used.  The FID measures roughly half the mass of acetaldehyde; thus a FID response of 0.5 was used.  For ethanol, a FID response of 0.74 was used, based on analyses done at Southwest Research Institute for the EPAct test program.
   
HC Speciation of Exhaust Emissions

   Exhaust speciation factors for pre-Tier 2 gasoline vehicles remain unchanged from MOVES2010.  However, since exhaust hydrocarbon emissions from Tier 2-and-later vehicles are substantially different in composition from pre-Tier 2 vehicles, new speciation factors were developed, based on EPAct Phase 3 data.  The new factors are summarized in Table 23 below.  
   The EPAct data also was used to update methane/total hydrocarbon ratios for 2004-and-later gasoline light-duty cars and trucks.  Separate ratios were calculated for running and starts, keeping the age effect in MOVES2010.  Updated methane to THC ratios for light-duty cars and trucks for model years 2004-and-later are provided in Table 22. In the process of updating these rates, we also identified and fixed some errors and missing values found in these ratios for pre-2004 gasoline vehicles.  

Table 22 Methane/THC ratios for gasoline light-duty cars and trucks (MY2004 and later)
                                       
                             Light-Duty Cars (LDV)
                            Light-Duty Trucks (LDT)
                                   Age Group
                                    Running
                                    Starts
                                    Running
                                    Starts
                                    0  -  3
                                   0.454054
                                   0.082814
                                   0.520114
                                   0.107257
                                    4  -  5
                                   0.235549
                                   0.050193
                                   0.269818
                                   0.066915
                                    6  -  7
                                   0.207525
                                    0.04565
                                   0.237717
                                   0.059709
                                    8  -  9
                                    0.17483
                                   0.040697
                                   0.200266
                                   0.053998
                                   10  -  14
                                    0.11926
                                   0.030697
                                   0.136611
                                    0.04364
                                   15  -  19
                                   0.101667
                                   0.028455
                                   0.116458
                                   0.040392
                                     20 +
                                   0.100044
                                   0.028016
                                    0.1146
                                   0.038834

   
   More recent and extensive data were available for pre-2007 diesel engines than those used in earlier versions of MOVES, and for 2007-and-later diesel engines, data were available from the Advanced Collaborative Emissions Study (ACES).,  The new diesel speciation factors and the speciation factors for gasoline vehicles are summarized in Table 23.

Table 23 Exhaust hydrocarbon speciation factors

Engine Type
Fuel Subtype
Operating Mode
Speciation Factor
NMHC to VOC



 Gasoline Tier 2
E0 to E10
Start
0.9787 + (0.0068 * wt. % oxygen)
Gasoline Tier 2
E0 to E10
Running
0.9148 - (0.0013 *  wt. % oxygen)
Gasoline
E15
Start
1.0162
Gasoline
E15
Running
0.9049
Diesel  -  pre-2007
All
Start and Running
1.1243
Diesel  -  2007+
All
Start and Running
1.3058
NMHC to NMOG



 Gasoline Tier 2
E0 to E10
Start
1.0078 + (0.0082 * wt. % oxygen)
Gasoline Tier 2
E0 to E10
Running
1.0149 + (0.0028 *  wt. % oxygen)
Gasoline
E15
Start
1.0495
Gasoline
E15
Running
1.0318
Diesel  -  pre-2007
All
Start and Running
1.1455
Diesel  -  2007+
All
Start and Running
1.3431

HC Speciation of Permeation Emissions
   
   Since no methane or ethane are found in evaporative or permeation emissions, THC is equivalent to NMHC, and VOC is equivalent to NMOG and TOG.  Thus, speciation factors are only needed to convert THC to VOC, to account for the mass of ethanol not measured by the FID:
      
      
      
       
      
      
      
Equation 12 VOC speciation for permeation
   For evaporative emissions, THC to VOC factors are unchanged from earlier versions of MOVES.  However, factors were developed for permeation emissions, based on data from the CRC E-77 test program.[,]  These factors are provided in Table 24.

Table 24 Gasoline Vehicle Permeation hydrocarbon THC to VOC speciation factors
                                       
                                 Fuel Subtype
                               Speciation Factor
                                   E0 to E10
                         1.0000+(0.0360* wt. % oxygen)
                                      E15
                                    1.1755
                                       
Additional Updates to Air Toxics and HC Speciation

   We made a number of other small changes to air toxics and HC speciation for MOVEST3NPRM.   We updated the General Fuel Expression Ratio table to fix a bug in acrolein running and start emissions for model years 1960 through 2003.  We fixed a small bug in the calculation of ethanol emissions, and we fixed an error in the HC speciation factors for diesel and biodiesel running,start and extended idle emissions. 
Updated Heavy Duty and Diesel Emission Rates
Changes to Heavy Duty Gasoline Vehicle Rates
   The "2007 Heavy Duty rule" reduced both start and running emissions from heavy duty on gasoline vehicles.  The reduction to running rates was inadvertently omitted from the MOVES2010 model.  This subsection documents the detailed analysis that was conducted to include this rule in MOVEST3NPRM, and also reflect the requirement that Medium Duty Passenger Vehicles (MDPVs) be certified under Tier 2 regulations.  While the 2007 Heavy Duty rule affects emissions in all Heavy Duty (HD) gasoline categories, there are very few gasoline vehicles in the largest classes.  As a result, revisions were made only to MOVES regulatory class 41, which are commonly called light HD or Medium Duty vehicles (8,500 to 14,000 pound GVWR). 
      
Regulatory Class Definition and Vehicle Composition
   Gasoline vehicles in MOVES regulatory class 41 are a mixture of engine certified HD vehicles, chassis certified HD vehicles, and MDPVs.  Each has separate set regulations governing their emissions, which are summarized below (Table 25). 

Table 25 Useful Life FTP Standards
                                                                               
                                     MDPV
                                (Tier 2 Bin 5)
                            8.5k  -  10k (Class 2B)
                                    10k-14k
                                   (Class 3)
                                    Engine
                                   Certified
 Units
                                    g/mile
                                    g/mile
                                    g/mile
                                   g/bhp-hr
Fully Phased in MY
                                     2009
                                     2008
                                     2008
                                     2010
HC
                                   0.09 NMOG
                                   0.195NMHC
                                  0.230 NMHC
                                   0.14 NMHC
CO
                                      4.2
                                      7.3
                                      8.1
                                     14.4
NOx
                                     0.07
                                      0.2
                                      0.4
                                      0.2

   The relative proportions of the vehicles within this weight class vary each year depending on economic demand.  Consequently, we estimated proportions based on existing data and engineering judgment.  MOBILE6 documentation indicates that MDPVs are approximately 16 percent of the gasoline 8,500 to 10,000 truck class. Based on the estimate that MDPV sales have increased since the 2003 release of the MOBILE6 report, we project that MDPVs are 15 percent of total MOVES regulatory class 41.  The MOBILE6 document further asserts that more than 95 percent of class 2B trucks are chassis certified.  Consequently, we estimate that 5 percent of all vehicles in the 41 regulatory class are engine certified.  Based on analysis from the recent medium and heavy duty greenhouse gas rulemaking, we assume that sales of 2B class trucks vehicles were triple that of 3 class trucks (Table 26). This is roughly consistent with recent model year sales totals, and yields the sales fractions below.
   
Table 26 Population Percentage
 
 percent of Reg Class 41
MDPV
                                                                            15%
Class 2B
                                                                            60%
Class 3
                                                                            20%
Engine Certified
                                                                             5%

Estimating Emission Rates
      
   We weighted the individual FTP standards together using the proportions shown in Table 26, which yields an aggregate MOVES gasoline FTP useful life standard shown below.  While we do not expect the model to produce these rates, it is a useful benchmark for the modeled rates.  The distinction between NMHC and NMOG was ignored in calculating the aggregate FTP, and would have yielded only minor variation in the final aggregate FTP (Table 27).
   
Table 27 Aggregate Useful Life FTP
 
Reg Class 41 g/mile
NMOG
                                                                           0.18
CO
                                                                           7.49
NOx
                                                                           0.22

   We then compared this aggregate FTP standard to the emissions calculated by MOVES 2010a.  The Physical Emission Rate Estimator (PERE), modified to produce Scaled Tractive Power (STP) distributions, was used to generate the bin mix of a regulatory class 41 vehicle on the federal test procedure (FTP) drive cycle.  The STP modification required shifting the vehicle weight to match Source Type 32 in MOVES (2.06 Tons).   We pulled rates by hour from the MOVES DB for the age 0 -3 group, and added in a cold start (operating mode 108) and a hot start (operating mode 102) from the MOVES database. The modified version of PERE produced the operating mode distribution shown in Table 28.
      .
Table 28 Operating Mode Bin Mix for a MD FTP
STPpbin
N
 percent
vspbin
N
 percent
                                                                              0
                                                                            160
                                                                            12%
                                                                             25
                                                                             41
                                                                             3%
                                                                              1
                                                                            258
                                                                            19%
                                                                             27
                                                                             49
                                                                             4%
                                                                             11
                                                                             94
                                                                             7%
                                                                             28
                                                                             17
                                                                             1%
                                                                             12
                                                                             68
                                                                             5%
                                                                             29
                                                                             13
                                                                             1%
                                                                             13
                                                                             70
                                                                             5%
                                                                             30
                                                                             15
                                                                             1%
                                                                             14
                                                                             36
                                                                             3%
                                                                             33
                                                                             13
                                                                             1%
                                                                             15
                                                                             48
                                                                             3%
                                                                             35
                                                                             12
                                                                             1%
                                                                             16
                                                                            141
                                                                            10%
                                                                             37
                                                                             13
                                                                             1%
                                                                             21
                                                                             68
                                                                             5%
                                                                             38
                                                                             17
                                                                             1%
                                                                             22
                                                                             44
                                                                             3%
                                                                             39
                                                                             15
                                                                             1%
                                                                             23
                                                                             97
                                                                             7%
                                                                             40
                                                                              6
                                                                             0%
                                                                             24
                                                                             77
                                                                             6%

                                                                               
                                                                               
                                                                               
Total
                                                                           1372
                                                                           100%

   Using this operating mode distribution, we constructed a simulated FTP out of four components (bag 1/3 running, cold start, hot start, and bag 2 running).  We constructed bag 1 (cold start + bag 1/3 running) and bag 3 (hot start + bag 1/3 running)  and weighted the resulting components together according to the FTP formula, and compared the 2007 and later rates in MOVES to the aggregate standard shown in Table 29.  The MOVES 2010a estimates at age 3 are 2.5x to 11x the standard, which indicates that the average vehicle is significantly out of compliance with the 2010 standard.

Table 29 Comparison between MOVES DB FTP and Aggregate FTP 
 
MOVES FTP 
Aggregate FTP
                              MOVES /Aggregate FTP
NMOG
                                                                           0.45
                                                                           0.18
                                                                           2.51
CO
                                                                          16.21
                                                                           7.37
                                                                           2.20
NOx
                                                                           2.53
                                                                           0.23
                                                                          11.06

Based on this analysis, we reviewed In Use Verification Program (IUVP) data from MYs 2004-2008 vehicles with estimated test weights of 7,500 pounds-or-above to determine the average in-use compliance ratio.  Non-sales weighted, and after accounting for the individual standards applicable to each vehicle, this yielded average ratios of FTP to the standard of 0.42 (NMOG) and 0.23 (NOx) (Table 30and 

Figure 6).  MOVES, which includes all vehicles and is a broader sample than the IUVP program, would be expected to yield somewhat higher emissions.  Further, while the Heavy Duty 2007 rule does not fully phase-in until 2010, we would anticipate that manufacturers would achieve similar compliance ratios as in past MYs.    

Table 30 Average Compliance Margin and Headroom
 
                                    Average
                                    Average
 
                              Ratio FTP/Standard
                                   Headroom
NMOG
                                                                           0.42
                                                                           0.58
Nox
                                                                           0.23
                                                                           0.77

                                       
                                       
Figure 6 Distribution of IUVP FTP Tests
      
   To produce similar compliance ratios in MOVES, we applied the following ratios to the running and start rates, incorporating the assumption that it is much more difficult to reduce start emission than running emissions.  

Table 31 Ratios applied to MOVES Rates
 
HC
CO
Nox
Running
                                                                            0.2
                                                                            0.2
                                                                           0.05
Start
                                                                            0.4
                                                                            0.4
                                                                            0.3

   Applying these ratios against the MOVES2010 default rates yields the new regulatory class 41 rates.  These are shown compared against the aggregate FTP from Table 27.

Table 32 Comparison of FTPs using Preliminary New MOVES Rates
 
MOVES FTP 
Aggregate FTP
                              MOVES /Aggregate FTP
NMOG
                                                                           0.10
                                                                           0.18
                                                                           0.56
CO
                                                                           4.72
                                                                           7.37
                                                                           0.64
NOx
                                                                           0.13
                                                                           0.23
                                                                           0.58

   This process yielded a number of medium duty rates, which by operating mode, were lower than the Light Duty Tier 2 Bin 8 rates.  Specifically, most of the higher power modes such as 27-30 and 37-40 were reduced to below the Bin 8 rates.  These operating modes are likely the least affected by the MY 2010 HD Gas standards, which include no SFTP component.
   
   To account for the lack of SFTP control on these vehicles we bounded the emission rates so that rates for individual operating mode could be lower than Bin 8 rates * 1.2, and no higher than the MD rates currently in MOVES.  This yielded rates that are slightly higher than the unbounded rates, with NOx is slightly over the Aggregate FTP standard.  These rates were phased in at a rate of 25 percent per year between MY 2007 and 2010 and placed in the MOVES emission rate database.
      
Table 33 Comparison of Outputs after accounting for SFTP emissions
      
 
MOVES FTP 
Aggregate FTP
 MOVES /Aggregate FTP
NMOG
                                                                           0.13
                                                                           0.18
                                                                           0.74
CO
                                                                           5.31
                                                                           7.37
                                                                           0.72
NOx
                                                                           0.28
                                                                           0.23
                                                                           1.20


Changes to Heavy- and Light-Duty Diesel Vehicle Rates
   We updated sulfate, sulfur dioxide, organic carbon and elemental carbon emissions for 2007-and-later diesel vehicles based on data from the ACES study. In 2007-and-later diesel vehicles, more of the fuel sulfur is converted to SO4 and less to SO2 compared with earlier diesel vehicles.  However, in this update, total PM is unchanged, so organic carbon is reduced.  The details of the data, its analysis, and its application in MOVES are explained in a separate memo., 
   

 Other Changes
   Other changes made for MOVEST3NPRM did not impact MOVES results, but improved our ability to generate the necessary emission rates for air quality modeling.  For example, we made finer divisions in MOVES model year groups to allow the phase-in of emission standards. More important in this category of changes were changes to allow MOVES to run a cloud computing environment.   This was necessary because evaporative emissions were expected to be important in this analysis, so we wanted to be sure to capture fine distinctions in fuel characteristics and ambient temperatures.  Instead of running MOVES to generate state-level inventories as we did for the recent Heavy Duty Greenhouse Gasoline rule analysis, we ran MOVES in emission rate mode and calculated rates for "representative counties."   As explained in the RIA for this proposal, these rates were then applied to other similar counties. This approach allowed us great precision in modeling temperatures, but it required more than 14,000 runs for each scenario-- an unprecedented requirement for computing power.  
   To meet this requirement, we adapted MOVES to run in a cloud computing environment with a Linux operating system.  We improved MOVES error handling and the ability to distribute MOVES functions among computers.  We truncated the IMCoverage table to facilitate the input of local inspection and maintenance data for each county.  We adjusted the zoneroadtype and sccroadtypedistribution table to assure that all roadtypes could be modeled in all counties.  We also made changes to reduce the runtime for each run.
Changes to Model Year Groups

   MOVES2010a grouped emissions for 2021-and-later model years together with a single emission rate.  Because the Tier 3 analysis had to estimate the impact of phasing in a new emission standard in this time period, MOVES was revised to use year-by-year model year groups for all years 2021 through 2030.  We also introduced a new model year group of 2031-2050.  This change required modification of many MOVES tables, but had no impact on emission estimates for the baseline and reference case.
Changes to County Allocations

   In the tables zoneroadtype and sccroadtypedistribution, zeros were replaced with the value 10-15.  This facilitates our approach of using representative counties; it allows the model to calculate emission rates on roadtypes that do not exist in a given county so they can be applied to a similar county that does have that roadtype.
Repair of Errors
   An error was found in the algorithm for aggregating Inspection /Maintenance effects to the state-level.  This was fixed, facilitating the use of the MOVES model at the national level for inventory calculations. We also updated the modelyear groups used for ammonia extended idle emissions to allow the model to correctly generate these emissions. And we updated model year groups for sulfate, NO and NO2 to allow the model to correctly estimate emissions for model year 2021. 
Changes to Facilitate Large Numbers of Runs

   For air quality modeling, we ran MOVES for a very large number of counties.  To facilitate this, we adapted MOVES to improve the model's operation in diverse computing environments, including large computer clusters and services available over the Internet from commercial providers.   
   The MOVES air quality runs were performed with IMCoverage tables for each county.  This change avoided conflicts between the default and the input data. 
   MOVES now provides a message when users create a user input database with the same name as an existing database, and we added the ability to omit roadtypes when modeling evaporative processes, thus facilitating faster production of RatePerProfile and RatePerVehicle output.  We improved the use of indices to speed model runs.
   We made many changes to the MOVES code to facilitate running MOVES on Linux operating systems, to allow the operation of MOVES with local MySQL, to allow separate processing of "TODO" and "Done" files, to better manage independent "worker" computers, to improve the MOVES log file, and to allow longer names for MOVES worker and master computers
   None of the changes described in this subsection changed MOVES emission results.
Impact of MOVES Changes
   To assess the overall impact of the MOVES changes described above in emissions with the updated version of MOVES were evaluated by conducting a national run with MOVES2010a and the updated MOVES version, using the reference case fuels described the Tier 3 RIA for 2017.  The results are shown in Table 34 Impact of MOVES updates to U.S. emissions in 2017 (annual tons).

Table 34 Impact of MOVES updates to U.S. emissions in 2017 (annual tons)

MOVES2010a
MOVEST3NPRM
Percent Change
NOx
                                                                      3,584,299
                                                                      3,452,314
                                                                            -4%
VOC
                                                                      1,922,560
                                                                      1,614,384
                                                                           -16%
CO
                                                                     23,663,807
                                                                     20,915,593
                                                                           -12%
PM
                                                                        111,928
                                                                        115,098
                                                                             3%
Benzene
                                                                         39,567
                                                                         39,695
                                                                             0%
1,3-Butadiene
                                                                          6,950
                                                                          6,022
                                                                           -13%
Acetaldehyde
                                                                         23,433
                                                                         23,918
                                                                             2%
Formaldehyde
                                                                         25,496
                                                                         29,446
                                                                            15%
Acrolein
                                                                          1,193
                                                                          1,814
                                                                            52%
Ethanol
                                                                         66,563
                                                                        111,629
                                                                            68%
Naphthalene
                                                                          4,786
                                                                          3,953
                                                                           -17%

   As illustrated in the table, the revision to air toxics inputs had a major impact on the emissions of these compounds.  Moreover, the changes between MOVES2010 and MOVEST3NPRM led to a substantial reduction in the emissions of CO and VOC.  These reductions will be seen in both the reference and control cases of the Tier 3 NPRM analysis.



 
 References
