Impacts of the Proposed Rule on Emissions and Air Quality 

This chapter presents the overall emissions and air quality impacts of
the proposed Tier 3 standards.  Section 7.1 describes national average
impacts on criteria and toxic emissions resulting from the Tier 3
proposal.  Section 7.2 describes the air quality effects of the proposed
emission reductions.  Because the air quality analysis requires emission
inventories with greater geographical resolution than the national
average inventories, the emission inventories described in the two
sections were developed separately, as described in each portion of this
chapter.  Section 7.3 discusses the impact of the proposed program on
greenhouse gas emissions.  

Criteria and Toxic Pollutant Emission Impacts  

Overview 

This section presents the projected national emission impacts of the
Tier 3 proposal on criteria and toxic air pollutants for selected
calendar years, and the methodology used to estimate these reductions. 
The proposed fuel and vehicle standards will directly reduce emissions
of nitrogen oxides (NOX), including nitrogen dioxide (NO2), volatile
organic compounds (VOC), carbon monoxide (CO), particulate matter
(PM2.5), air toxics, and sulfur dioxide (SO2).  The implementation of
lower sulfur gasoline will reduce criteria and air toxic emissions from
the existing gasoline fleet, and cause some reductions in SO2 from the
nonroad gasoline sector.  The largest reductions come immediately
following the implementation of the fuel standard, as a significant
share of overall emissions are produced by Tier 2 and older vehicles. 
To reflect these early reductions, we are presenting emission reductions
in calendar year 2017, near the beginning of the proposed fuel program.

The proposed vehicle standards will incur reductions as the cleaner cars
and trucks begin to enter the fleet in model year 2017.  The magnitude
of reduction will grow as the contribution of these vehicles to fleet
emissions becomes more prominent – to reflect this, we are also
presenting emission reductions through calendar year 2030, when 2017 and
later model year cars and trucks comprise 80 percent of the light-duty
fleet, and travel 90 percent of vehicle miles travelled (VMT). 2030 is a
standard out-year for evaluation; it is used for air quality modeling in
this proposal as well as recent EPA rules.  However, the full impact of
the vehicle program would be realized after 2030.  For this reason, we
are also presenting emissions reductions in calendar year 2050, when the
fleet will have fully turned over to the proposed vehicle standards.  

Emission impacts presented in this section are estimated on an annual
basis, for all 50 U.S. states plus the District of Columbia, Puerto Rico
and the U.S. Virgin Islands.  The reductions from onroad sources were
estimated using an updated version of EPA’s MOVES model, as described
in detail in Section   REF _Ref305133367 \r  7.1.3 ; and the NONROAD
model for offroad sources.  Reductions were estimated compared to a
reference case that assumed continuation of the Tier 2 vehicle program
indefinitely and an average gasoline sulfur level of 30 ppm (10 ppm in
California).  

The emission inventory methodology applied to generate these national
estimates does differ from the methodology used to generate the finely
resolved emission inventories needed for the air quality modeling,
leading to some differences in absolute estimates of tons reduced
between the two analyses.  These differences are discussed in Section
7.2.  

Scenarios Modeled 

We analyzed emission impacts of the Tier 3 vehicle emissions and fuel
standards by comparing projected emissions for future years without the
Tier 3 rule (reference scenario) to projected emissions for future years
with the Tier 3 standards in place (control scenario).    REF
_Ref305137860  Table 7-1  below presents an overview of the reference
and control scenarios for calendar years 2017 and 2030.  Both scenarios
reflect the renewable fuel volumes mandated in the Energy Policy Act of
2005 (EPAct) and Energy Independence and Security Act of 2007 (EISA). 
We thus refer to this renewable fuel level as “post-EPAct/EISA”.  A
key update in the analysis of renewable fuel volumes from that performed
for the RFS2 final rule is the inclusion of gasoline containing 15
percent ethanol by volume (E15) in the fuel pool, per the approval in
2011 of a waiver which allows E15 to be used in model year 2001 and
later cars and trucks.  We estimated the market shares for E15 and
gasoline containing 10 percent ethanol by volume (E10) in the market in
2017 and 2030, as discussed in Section   REF _Ref309632651 \r \h 
7.1.3.2.1 .  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  1  Overview of
Reference and Control Scenarios

	Reference Scenario 	Control Scenario

2017	Renewable Fuels: RFS2 programa

  21.6 B gallons renewable fuels 

  (24 B ethanol-equivalent gallons):

17.8 B gallons ethanol: E10 b, E15c

Fuel Sulfur Level: 

  30 ppm (10 ppm California)

Fleet:

  100 percent Tier 2 and older vehicles	Renewable Fuels: RFS2 programa

  21.6 B gallons renewable fuels 

  (24 B ethanol-equivalent gallons):

17.8 B gallons ethanol: E10 b, E15c

Fuel Sulfur Level: 

  10 ppm

Fleet:d

  93 percent Tier 2 and older vehicles

  7 percent Tier 3 vehicles 

2030	Renewable Fuels: RFS2 programa

  30.5 B gallons renewable fuels 

  (36 B ethanol-equivalent gallons):

22.2 B gallons ethanol: E15c

Fuel Sulfur Level: 

  30 ppm (10 ppm California)

Fleet:

  100 percent Tier 2 and older vehicles

	Renewable Fuels: RFS2 programa

  30.5 B gallons renewable fuels 

  (36 B ethanol-equivalent gallons):

22.2 B gallons ethanol: E15c

Fuel Sulfur Level: 

  10 ppm

Fleet:d

  20 percent Tier 2 and older vehicles

  80 percent Tier 3 vehicles

a RFS2 primary mid-ethanol scenario, 75 FR 14670 (March 26, 2010)

b Gasoline containing 10 percent ethanol by volume

c Gasoline containing 15 percent ethanol by volume

d Fraction of the vehicle population 

Our reference scenarios assumed an average fuel sulfur level of 30 ppm
in accordance with the Tier 2 gasoline sulfur standards.  Under the
proposed Tier 3 program, federal gasoline would contain no more than 10
ppm sulfur on an annual average basis by January 1, 2017 (Section V of
the preamble), and we therefore assumed a nationwide fuel sulfur level
of 10 ppm for both future year control cases.  A more detailed
description of our fuel inputs and assumptions for this analysis can be
found in Section   REF _Ref302458260 \r  7.1.3.2 .

We assumed a continuation of the existing Tier 2 standards for model
years 2017 and later in modeling emissions for our reference scenario. 
Our Tier 3 control scenario modeled the suite of exhaust and evaporative
emission standards proposed for light-duty vehicles (LDVs), light duty
trucks (LDTs: 1-4), medium passenger vehicles (MDPVs) and large pick-ups
and vans (Class 2b and 3 trucks) described in Section IV of the
preamble, including:  

Fleet average Federal Test Procedure (FTP) NMOG+NOX standards of 30
mg/mi for LDVs, LDTs and MDPVs, phasing in from MYs 2017 to 2025

Fleet average Supplemental Federal Test Procedure (SFTP) NMOG+NOX
standards of 50 mg/mi for LDVs, LDTs and MDPVs, phasing in from MYs 2017
to 2025

Per-vehicle FTP PM standard of 3 mg/mi for LDVs, LDTs and MDPVs, phasing
in from MYs 2017 to 2022

Per-vehicle US06-only PM standard of 10 mg/mi for LDVs through LDT2s and
25 mg/mi for LDT3s and LDT4s, phasing in for MYs 2017 to 2022 

New standards for Class 2b and 3 trucks phasing in by MY 2022 including
NMOG+NOX declining fleet average, more stringent PM standards, and a
regulatory useful life of 150,000 miles

More stringent evaporative emission standards for diurnal plus hot soak
emissions, a new canister bleed test and emissions standard, a new
requirement to measure emissions using 9 RVP E15 certification test
fuel, and new requirements addressing evaporative leaks on in-use
vehicles.

The Tier 3 standards are expected to reduce onroad criteria and toxic
emissions, and to a much smaller extent, nonroad SO2 emissions, but will
not affect upstream, refueling or portable fuel container criteria or
toxic emissions.  The methodology for estimating emission impacts and
the results for onroad and nonroad emissions are described in Section
7.1.3 and Section 7.1.4, respectively.  

Implementation of the proposed Tier 3 standards is aligned with the
model year 2017-2025 Light-Duty GHG standards to achieve significant
criteria pollutant and GHG emissions reductions while providing
regulatory certainty and compliance efficiency to the auto and oil
industries.  The LD GHG standards were still in a preliminary state of
development (pre-proposal) at the time we developed our assumptions for
the Tier 3 emissions, air quality, and cost analyses presented here, so
we were not able to reflect them in these analyses.  However, our
analyses for the final Tier 3 rule will include the final LD GHG
standards in both the reference and control scenarios, and will thus
account for their impacts on the future vehicle fleet and future fuel
consumption.  We do not anticipate that this change will substantially
modify our results or conclusions for two major reasons.  First,
improved fuel efficiency does not have a direct impact on criteria
pollutant emissions.  Second, non-GHG emissions standards for light-duty
vehicles are expressed as grams per mile driven; thus, non-GHG emissions
are a function of the emission control systems rather than the amount of
fuel burned.  As a result, changes to vehicle efficiency as a result of
the LD GHG standards will not affect the emissions benefits of the Tier
3 rule.  The analysis described here accounts for the following national
onroad rules:

Tier 2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control
Requirements (65 FR 6698, February 10, 2000)

Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur
Control Requirements (66 FR 5002, January 18, 2001)

Mobile Source Air Toxics Rule (72 FR 8428, February 26, 2007)

Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel
Standard Program (75 FR 14670, March 26, 2010) 

Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate
Average Fuel Economy Standards for 2012-2016 (75 FR 25324, May 7, 2010) 

The analysis also accounts for many other national rules and standards. 
In addition, the modeling accounts for state and local rules including
local fuel standards, Inspection/Maintenance programs, Stage II
refueling controls, the National Low Emission Vehicle Program (NLEV),
and the Ozone Transport Commission (OTC) LEV Program.  See the Tier 3
emissions modeling TSD for more detail.

Onroad Emissions 

Methodology Overview

EPA’s official model for use in estimating mobile source emissions is
known as the Motor Vehicle Emission Simulator (MOVES), with the most
recent version approved for use in State Implementation Plan (SIP) and
transportation conformity analyses being MOVES2010b.  A version of
MOVES2010a, updated specifically for this analysis, was used to estimate
emissions of criteria and air toxic emissions from on-road gasoline and
diesel vehicles for the entire U.S. for the reference and control
scenarios described in Section   REF _Ref309633442 \r \h  7.1.2  above,
for calendar years 2017 through 2030, and 2050.  

The MOVES model updates affecting reference and control case emissions
were extensive, and are documented in a separate memorandum to the Tier
3 docket.  Updates made to MOVES2010a for this analysis primarily
incorporated major new research in three areas. The first involves fuel
effects on exhaust emissions from Tier-2 vehicles. The second involves
improvements in estimation of evaporative emissions from all vehicles,
including Tier-2 vehicles. The third involves accounting for the effects
of fuel sulfur level on exhaust emissions.

 The effects of changes in fuel properties on exhaust emissions of Tier
2 vehicles, which comprise the majority of the fleet by 2017, were
assessed through the results of the EPAct Phase-3 Program.  Specific
fuel properties addressed include ethanol level, aromatics, distillation
properties, and volatility (Reid Vapor Pressure, or RVP).  Methods used
to account for the effects of these properties in inventory modeling are
described in a separate memorandum to the docket. Improvements in
estimating and projecting evaporative emissions are described in this
document (see   REF _Ref340152605 \r \h  7.1.3.3.7 ). Finally, because
the updates to fuel sulfur effects are critical for estimating
reductions from the Tier 3 program, they are also presented in detail in
Section 7.1.3.4.  

Other than sulfur, these changes had more bearing on updating reference
case emissions than on the projected reductions from the Tier 3
standards.  One key exception to this, however, were updates made to
PM2.5 fuel effects that result in an increase in emissions with
increases in aromatics and T90 (reduction in E300).  As discussed in
Section 7.1.3.2.2, this update had an impact on control case emissions,
as changes in aromatics and E300 were projected by refinery modeling as
sulfur levels were reduced from 30 ppm to 10 ppm.  However, other recent
studies suggest that the PM emission increases observed are likely
linked to the impacts of heavier hydrocarbons, and in particular heavy
aromatics.,,  More study is needed to understand these relationships,
and whether the changes in refinery operation to make up for lost
octane, should it occur, would be likely to impact heavy aromatics or
not.   

In addition to fuel effects, we also improved emission estimates in
other areas. 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 repaired errors in the MOVES2010a
emission rates for NH3, NO and NO2., 

The MOVES version used for this analysis also includes an added
capability to model many hazardous air pollutants.  And other changes
were made to the MOVES2010a model to facilitate the large number of
parallel runs that needed to be done to complete the Tier 3 air quality
modeling inventories.  These changes are also detailed in the docket
memo addressing MOVES updates.  

In addition to the model updates needed to incorporate new research, a
set of custom inputs were developed to allow MOVES to model the
reference and control scenarios.  Some of these inputs were required to
reflect regional variations in fuels for both the reference and control
scenarios, as discussed in detail in Section 7.1.3.2.  Other inputs were
required to model the vehicle program for exhaust and evaporative
emissions, discussed in Section 7.1.3.3.  

The national emission inventories presented in this section were
developed with a simpler and quicker method than we used for air quality
modeling, because the abbreviated approach enabled analysis for calendar
years in addition to 2017 and 2030, and makes the analysis easier for
stakeholders and other commenters to replicate.  National emission
inventories were developed using the pre-aggregation feature of MOVES. 
For criteria pollutants, the default pre-aggregation level of ‘hour’
was selected, which condenses the county-level temperatures into a
single national average temperature by hour of the day.  While the model
and many of the inputs are identical for the emission inventory modeling
performed for the air quality analysis, the pre-aggregation approach is
coarser than the approach described in Section 7.2 used to develop the
gridded/hourly emission inventories needed for air quality modeling.  In
addition to this difference in temperature resolution (national average
vs. gridded hourly temperature), the national emission inventory
analysis also used information contained in the MOVES2010a default
database for vehicle miles travelled, fleet age distributions, activity
inputs (e.g., speeds), temperatures, emission standards, and
inspection/maintenance programs.  As discussed in Section 7.2, this
contrasts with the air quality modeling inventory methodology, which
used data supplied by state and regional modelers for many of these
inputs, employed hourly meteorological data, and accounts for Section
177 states adoption of California LEV standards in the modeling
baseline.  

 To assure that adequate temperature resolution was incorporated into
the national emission inventory processes, MOVES was run separately for
January and July, and annual emissions were extrapolated (for all
pollutants except PM) by scaling up each month by a factor of 5.88.  For
PM, to offset the disproportionate effect of the colder temperature
January results, a scaling factor of 7.5 was applied to July and 4.3 to
January; these factors were determined based on analysis of annual PM
emissions during modeling for the RFS2 rule. The updated MOVES version,
and all inputs and outputs that produced the results presented in
Section   REF _Ref305058784 \w \h  7.1.5  of this Chapter are contained
in the Tier 3 rulemaking docket.  

Fuel Inputs 

Estimating national emission inventories required translation of the
reference and control fuel scenarios presented in Table 7-1 into a
discrete set of fuels (defined by RVP, sulfur content, ethanol level,
aromatics content, olefin content, T50 and T90), and the market share of
these fuels, by month and county.  These data were converted into
“fuel supply” database tables used by MOVES to estimate emission
inventories.  Even for the national emission inventories calculated at a
pre-aggregated level, these county-level fuel supply tables are retained
to develop composite emissions that reflect the market share of the
entire set of fuels that define the U.S. fuel pool.  The crux of
estimating emission impacts for the Tier 3 fuel program was the
development of fuel supply database tables that reflected the difference
between the reference and control scenarios, discussed in the following
sections.    

Reference Scenario

The reference scenario was developed to reflect updated assumptions
regarding the ethanol blends used to meet the EISA-mandated volumes. 
For simplicity, we assumed the same biofuel volumes analyzed in the RFS2
final rule, referred to in that rule as the primary “mid-ethanol”
case.  The RFS2 analysis also considered a “low-ethanol” and a
“high-ethanol” scenario; the difference being an increase in
cellulosic ethanol and a decrease in cellulosic diesel going from the
low-ethanol to the high-ethanol scenario. We believe that the RFS2
renewable fuel volumes still bracket the realm of realistic potential
fuel scenarios assuming the EISA mandate is met.  However, given the
practical limitations of conducting air quality modeling for all three
scenarios, we focused our Tier 3 modeling on the primary mid-ethanol
case.  

We relied on biofuel volumes taken directly from the RFS2 mid-ethanol
case, which made projections for every year between 2010 and 2022.  We
assumed that renewable fuel production/consumption would remain constant
after EISA reaches full implementation in 2022.  Accordingly, for 2030,
we relied on 2022 biofuel volumes from the RFS2 primary mid-ethanol
case. 

In the RFS2 final rule, we assumed that all ethanol would be consumed as
either E10 or E85.  Since then, EPA issued a waiver permitting 15 volume
percent ethanol blends (E15) to be used in 2001 and newer model year
light-duty vehicles.  While E15 has only limited commercial availability
currently, EPA believes it may compete favorably with E85 in the
marketplace and could become a major gasoline blend in the future to
meet the mandated RFS2 volumes.  Accordingly, the reference scenario
assumes an increasing utilization of E15 over time as infrastructure
ramps up and owners become more aware of their vehicle’s ethanol usage
potential.  

  To estimate future E15 fuel consumption, we started by estimating the
fraction of vehicles capable of legally refueling on E15 (or higher
ethanol blends) in the future.  To do this, we relied on many vehicle
assumptions made under the RFS2 final rule.  For flexible fuel vehicles
(FFVs), we started with EPA certification data and assumed that the
“Detroit 3” (GM, Ford and Chrysler) would follow through with their
voluntary commitment to produce 50 percent FFVs by 2012.   We also
assumed that the Detroit 3 would continue to comprise approximately 45
percent of total light-duty vehicle sales.  In addition, we relied on
total light-duty vehicle sales projections from EIA’s Annual Energy
Outlook (AEO).  Factoring in EPA’s projected vehicle scrappage rates,
vehicle miles traveled (VMT) by model year (MY), and fuel economy
assumptions based on EPA’s 2012-2016 Light-Duty Greenhouse Gas Rule,
we estimate that by 2017, 13.6 percent of gasoline demand would come
from FFVs, 68.7 percent would come from 2001 and later model year
non-FFVs, and 17.7 percent would come from legacy vehicles, nonroad,
motorcycles and other small engines not approved for E15.  By 2030, the
contributions would be 22.0, 66.8, and 11.2 percent, respectively.  A
graph of the projected gasoline fuel fractions over time is provided in 
 REF _Ref306266426 \h  \* MERGEFORMAT  Figure 7-1 . 

 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  1  Gasoline Fuel
Fraction by Vehicle Type

The next step was to estimate how often E15-capable vehicles would fill
up on E15.  E15 utilization is a function of how accessible E15 is to
consumers (i.e., number of stations offering it) and how often
E15-capable vehicles choose to fill up on it (based on retail pricing
and other factors).  We assumed that E15 utilization for 2001 and later
model year vehicles would ramp up to 50 percent by 2017, increasing to
almost 100 percent utilization by 2030.  We assumed that new nonroad
equipment (and other small engines not covered by the waiver) would be
designed for and/or certified on E15 in the future.  As such, we assumed
that E15 utilization in nonroad equipment would ramp up from zero
percent in 2017 to almost 100 percent by 2030.  This resulted in a very
small amount of E85 needing to be used in 2017 and 2030 to reach the
RFS2 primary mid-ethanol volumes. FFV owners were assumed to utilize E15
at the same rate as 2001 and later model year conventional vehicles when
not refueling on E85.  When not refueling on E15 (or E85), all
E15-capable vehicles were assumed to refuel on E10.  As a simplifying
assumption, we assumed E0 usage was negligible.    REF _Ref306185659  \*
MERGEFORMAT  Figure 7-2  shows the E15 and E85 utilization assumptions
and   REF _Ref310923926 \h  Figure 7-3 shows the resulting fuel volumes
by year. Total fuel volumes were computed based on motor gasoline energy
projections provided in the Annual Energy Outlook (AEO) 2011 early
release.

 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  2 
Post-EPAct/EISA Fuel Utilization Assumptions/Results

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  3  Assumed
Post-EPAct/EISA Gasoline Fuel Volumes by Calendar Year

As shown above, our post-EPAct/EISA ethanol blend assumptions (i.e.,
ramp up of E15) are capable of meeting the RFS2 primary mid-ethanol
scenario with little to no E85 needed – especially in our 2017 and
2030 modeling years.  Therefore, for emissions and air quality modeling,
we made the simplifying assumption that E85 use would be negligible. 
While E85 used in FFVs could be part of the future biofuel picture, we
believe that given EPA’s waiver decision and the ability for E15 to
compete in today’s gasoline marketplace, E15 is a viable ethanol blend
for meeting the RFS2 volumes.  

These projected fuel scenarios do not reflect the impact of the LD GHG
standards, as the rule was still in a preliminary state at the time we
needed to finalize our assumptions for these Tier 3 analyses.  The LD
GHG standards will reduce gasoline demand in 2017 and beyond which has
two impacts on our analysis.  First, the reduced fuel consumption
associated with the LD GHG standards will likely result in somewhat
lower costs for sulfur control, because less gasoline will need to be
desulfurized.  Second, if ethanol volumes are constant but gasoline
demand is reduced due to the LD GHG standards, perhaps as much as five
percent of the fuel pool could be E85 or other higher-level ethanol
blends by 2030.  However, we do not expect this to significantly impact
criteria pollutant reference or control case emissions because recent
data on NMOG and NOX emissions do not show significant differences when
run on E85,,  and the percentage of the fleet impacted is small. There
could be a small impact on sulfur related costs and emission reductions
due to an effect of E85 on overall fuel pool sulfur concentrations,
depending on what could be assumed for future E85 (or other higher level
ethanol blend) sulfur levels.  An increase in E85 volumes could also
lead to changes in toxic emissions, with increases in acetaldehyde
emissions from ethanol combustion, but decreases in most other toxics
due to dilution.  However, EPA air quality modeling suggests that
changes in ambient levels of acetaldehyde are likely to be much smaller
than changes in direct emissions.,  This is because emissions of
acetaldehyde precursors, particularly alkenes, are lower in E85
emissions.  For the Tier 3 final rule's analysis, we will be including
the impact of the LD GHG standards, and we will revisit our assumed
future fuel scenarios accordingly, including higher-level ethanol blends
as appropriate.

For this analysis we assumed that E15 would first appear in significant
volumes in reformulated gasoline (RFG) areas.  All RFG is subject to a
de facto 7-psi RVP standard.  And since neither E10 nor E15 can take
advantage of the 1-psi waiver in RFG areas, the two fuels are
essentially the same with respect to vapor pressure limits.  In
conventional gasoline (CG) areas, E15 is held to a 9-psi standard
whereas E10 is allowed to be 10-psi in-use (through the use of a 1-psi
waiver).  As a result, in order for E15 to enter into CG areas, refiners
would need to remove butanes/pentanes to make an 8-psi blendstock to
account for the 1-psi increase associated with ethanol blending. 
Conversely, refiners could keep their CG blendstocks around 9-psi to
blend up 10-psi E10 which is eligible for the 1-psi waiver and thus is
considered 9-psi for compliance purposes.  Due to the additional actions
needed to market E15 during the summer in CG areas, it makes sense that
more E15 would be used in RFG areas, at least in the near term. 
Accordingly, we assumed that in 2017, 75 percent of RFG would be E15 and
the balance of E15 would be used in CG (resulting in 25 percent E15
utilization and 75 percent E10 utilization in those areas).  By 2030,
since we modeled a much larger volume of ethanol, E15 was assumed to be
used virtually everywhere.

Converting these assumptions into post-EPAct/EISA fuel properties on a
county-by-county basis required an assessment of fuel properties prior
to the EPAct/EISA requirements for 2017 and 2030, termed
“pre-EPAct/EISA”.   We developed these starting with 2005 fuel
volumes as the baseline for our analysis of renewable fuel impacts
because these volumes were in use immediately before EPAct was enacted,
creating volume requirements for 2006, and because they provide
consistency with the base case used for air quality modeling described
in Section 7.2.  In 2005, gasoline contained over 4 billion gallons of
ethanol.  In 2005, gasoline also contained approximately 2 billion
gallons of MTBE.

For translation into the fuel supply inputs needed by MOVES, the
MOVES2010a default fuel supply for 2005 (based on the National Emission
Inventory) was updated to serve as a better basis for the other modeling
cases. These updates are detailed below.

In 2005, California fuel was required to have an ethanol level of 5.7
percent for all fuel. The MOVES2010a default fuel supply listed
California counties with a 57 percent market share of E10 and a 43
percent market share of E0. The fuel supplies for these counties were
refined to indicate E5.7 at 100 percent market share. Also, Alameda
County, California, had anomalous fuel property data. The fuel
properties for this county were replaced with an average of the fuel
properties of the neighboring counties.

Counties in the southeast Michigan fuel program area erroneously had
been assigned higher sulfur and benzene levels than the rest of the
state of Michigan. Sulfur and benzene for these counties were corrected
to an average of all other Michigan counties. Some Michigan counties
also had very low levels of MTBE (<0.3 percent), despite an MTBE ban in
Michigan in 2005. These fuels have been corrected to a 0 percent MTBE
conventional fuel to properly reflect the absence of MTBE in Michigan.

Many counties in the 2005 default fuel supply had fuel RVP levels that
were drastically inconsistent with the mandated level for RVP in that
county for a given summer month. Summer RVP levels for all counties were
corrected to their mandated level (including a 0.25psi compliance
margin) based on knowledge of fuel programs in place for 2005. Winter
RVP levels remain unchanged from the 2005 default fuel supply.

The corrections above were applied to the months of January (winter fuel
properties) and July (summer fuel properties) for the 2005 base case.
The fuel properties for these two months were then duplicated to other
months to complete the 2005 fuel supply for all months. The corrected
January fuel supply was duplicated to February, March, April, October,
November and December; the corrected July fuel supply was duplicated to
May, June, August and September. Although this duplication eliminates
fuel property changes during the shoulder seasons in April and
September, these intermediate month fuel properties were used only when
modeling refueling for the air quality inventories; we expect that this
simplification will not significantly impact the modeling run results.
Diesel fuel properties were not adjusted in the 2005 fuel supply.

The 2017 pre-EPACT/EISA base was developed using the corrected 2005 fuel
supply discussed above as a foundation. From this fuel supply, benzene
and sulfur levels for all counties were further corrected to properly
reflect the introduction of the Control of Hazardous Air Pollutants from
Mobile Sources (MSAT2) (2007) rule and the Tier 2 Motor Vehicle
Emissions Standards and Gasoline Sulfur Control Requirements (1999).
Benzene corrections were made by PADD following results from the MSAT2
analysis of downstream benzene levels. No other fuel property changes
were found to change significantly with a change in benzene levels.
Benzene levels by PADD follow in   REF _Ref309635866 \h  Table 7-2 
below:

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  2 - MSAT2
Downstream Fuel Benzene Levels

PADD	CG	RFG

1	0.61	0.54

2	0.63	0.60

3	0.63	0.54

4	0.86	N/A

5	0.65	0.61

CA	N/A	0.62

Sulfur corrections were made to all counties based on the default sulfur
level found in the 2005 fuel supply. Counties with a sulfur level higher
than 30 ppm were reduced to 30 ppm to reflect the gasoline sulfur
standards of the Tier 2 rule (counties subjected to lower fuel sulfur
standards, such as in California, were not changed). Refinery modeling
showed that there is a significant effect on aromatics level when sulfur
is reduced. Corrections to the aromatics level based on refinery
modeling for counties with reduced sulfur level were made as follows:

1) A “high sulfur” aromatics level was determined using the
following equation:

Equation   STYLEREF 1 \s  7 -  SEQ Equation \* ARABIC \s 1  1  Aromatics
Level from Initial Sulfur Concentration

2) A “low sulfur” aromatics level was determined using the same
equation, substituting 30ppm for the initial sulfur level of the county

3) An aromatics delta was calculated by subtracting the “low sulfur”
aromatics level from the “high sulfur” aromatics level

4) This aromatics delta was applied as a correction for sulfur reduction
to the original aromatics level for the county as appearing in the 2005
fuel supply

Diesel fuel sulfur levels were also adjusted to 30ppm to reflect low
sulfur diesel levels. There were no other changes from the 2005 base
case to the 2017 base case. The fuel properties for the 2030 base case
were identical to the 2017 base case for every county; the only
modification made to create the 2030 pre-EPAct/EISA base case is a
change in the year of the fuel supply.

The post-EPAct/EISA cases used as a reference for Tier 3 were developed
using the pre-EPAct/EISA cases for 2017 and 2030 as a foundation, and
were then created by inserting the ethanol market share assumptions for
each county discussed at the beginning of this section.  Fuel properties
for all counties were then adjusted based on relationships found in
refinery modeling for this additional ethanol. The process used for
adjusting ethanol market shares and fuel properties for every county is
described below:

1) Counties with multiple fuels were aggregated to one set of temporary
fuel properties using the market shares of these fuels. Counties with
only one fuel remain unchanged in this step.

2) Two new fuels, a 10 percent ethanol blend and a 15 percent ethanol
blend, were created from the temporary fuel properties by raising the
ethanol level from the temporary aggregate to the appropriate new
ethanol level.

3) For each new fuel, other fuel properties were adjusted based on
refinery modeling to reflect changes due to increased ethanol use. These
fuel property changes (on a percent ethanol basis) are shown in   REF
_Ref309636151 \h  Table 7-3  below:

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  3 - Reference Case
Fuel Property Changes

	Summer	Winter

Fuel Propertya	E10	E15	E10	E15

E200	0.54834	0.76481	0.73693	1.04067

E300	0.13946	0.09549	0.00634	0.39788

Aromatics	-0.40887	-0.46684	-0.50555	-0.49956

Olefins	-0.13788	-0.15915	-0.25198	-0.15296

Note:

a These fuel property changes are listed as a per-ethanol percent
change.

4) Market shares for the E10 and E15 fuels were adjusted to reflect RFS2
level ethanol use. As explained above, in RFG counties, E10 was added to
the fuel supply at a 25 percent market share, E15 was added to the fuel
supply at a 75 percent market share. For counties not using RFG, E10 was
added to the fuel supply at a 75 percent market share, and E15 was added
at a 25 percent market share. In 2030, the E10 market share is set at 1
percent and E15 market share set at 99 percent, regardless of fuel
program.  

5) RVP levels for the new E10 and E15 fuels were corrected to reflect
E15 fuel not receiving a 1psi waiver for maximum RVP level. As discussed
for the base case fuel supplies in the previous section, RVP levels were
adjusted to within a 0.25 psi compliance margin depending on county fuel
programs. 

The result of this effort was two alternate fuel supply databases tables
for use in MOVES, reflecting the reference case fuels in 2017 and 2030
– these tables were used for the development of national emissions
inventory as well as air quality modeling.  New fuel formulations were
also required for MOVES, and were created for each of the new E10 and
E15 fuels created for the 2017 and 2030 reference case. 

Control Scenario

The Tier 3 control fuel scenarios for the years 2017 and 2030 used the
fuel supplies constructed for the 2017 and 2030 reference cases
described in the previous section as a foundation.  To develop the
control scenario fuel supplies, we modified the reference fuel supplies
to reflect the sulfur program proposed in the Tier 3 control case by
reducing sulfur from 30 ppm to 10 ppm for all gasoline.  Associated fuel
properties determined by refinery modeling were also adjusted to reflect
the implications of sulfur reductions on other fuel properties, such as
an increase in aromatics and decrease in olefins and distillation
properties, as shown in   REF _Ref302483106 \h  Table 7-4  below. These
changes were made to every county with fuel exceeding a sulfur level of
10ppm. No changes were made to the diesel fuel supply for the control
scenario.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  4  Tier 3 Control
Case Sulfur Fuel Property Changes

Fuel Property	Summer	Winter

E200 (%)	-0.78	-1.27

E300 (%)	-0.75	-0.68

Aromatics (%)	0.63	0.48

Olefins (%)	-0.82	-1.12

The result of this effort was two additional alternate fuel supply
databases tables for use in MOVES, reflecting the control case fuels in
2017 and 2030; these tables were used for the development of national
emissions inventory as well as air quality modeling.  For the national
emission inventories the 2017 fuel supply was applied to 2018 through
2021 as well, to approximate the fuel supply prior to full RFS2
implementation; the 2030 fuel supplies were applied to 2022 through
2029, reflecting full RFS2 implementation.  New fuel formulations were
also required for MOVES, and were created for the low sulfur versions of
the E10 and E15 fuels created for the 2017 and 2030 reference case. 

Upon further analysis, we believe that the increase in aromatics and
reduction in E300 shown in Table 7-4 is unlikely.  The process of
hydrotreating in the fluid catalytic cracking (FCC) of gasoline to
reduce sulfur tends to saturate olefins formed by the FCC unit and which
are present in FCC gasoline, thus reducing its octane.  Because refiners
have historically been pressed to make as much octane as possible to
supply market needs, various technologies and catalyst formulations have
been developed for hydrotreating FCC gasoline to avoid and/or minimize
this loss of octane.  Nevertheless, our analysis estimated that
desulfurizing gasoline to an average of 10 ppm would result in about a
half number loss in the octane of FCC gasoline, and we also
conservatively assumed that this octane loss would need to be
compensated for with other changes to gasoline, such as increased
reformate (an aromatic rich stream), isomerate and  alkylate.  These
assumptions are conservative and not consistent with more recent trends.
 For example, there is less demand for octane due to the dramatic rise
in the use of ethanol (which has very high octane), and the declining
demand for gasoline.  It is not clear that the loss in octane resulting
from gasoline desulfurization would need to be compensated for.  In
fact, there is considerable “give away” of octane in the gasoline
pool today as ethanol is still often splash blended on top of finished
gasoline instead of blending it with a sub-octane gasoline blendstock. 
Nevertheless, for our Tier 3 analysis, we conservatively assumed that
the full half number loss in the octane of FCC gasoline would have to be
compensated for by the refinery and not made up by blending in more
ethanol.  In particular, in running the refinery model to make up for
the lost octane, we conservatively constrained the refinery model such
that the only options for additional octane were changes within the
refinery.  Additional ethanol was not a modeled option, despite the fact
that ethanol will be providing additional octane to the gasoline pool as
ethanol use continues to rise through 2017 and beyond.

Furthermore, the increase in aromatics and decrease in E300 may simply
be a function of additional assumptions in the refinery modeling.  We
did not model any revamps in FCC feed hydrotreaters.  For some
refineries, revamping FCC pretreaters is expected to realize the
decrease in FCC gasoline sulfur levels without any decrease in FCC
gasoline octane levels.  Furthermore, when refiners are faced with
tighter sulfur standards, one strategy for compliance would be to
undercut the heaviest portion of the FCC naphtha, which is also the
portion highest in sulfur, into either jet fuel or diesel fuel.  By
doing so, refiners would not only reduce the hydrotreating severity of
their FCC posttreaters (reducing olefin saturation and some of the
octane loss), but they would be increasing E300 (lightening up the
gasoline pool).  One check of our refinery modeling output, which shows
increasing aromatics and decreasing E300, would be to compare the output
of our refinery modeling with that of other studies.  Our refinery
modeling estimated a 0.48 volume percent increase in the aromatic
content and 0.68 percent reduction in the E300 of gasoline in the
winter, and a 0.63 volume percent increase in the aromatic content and
0.75 percent reduction in the E300 of gasoline in the summer.  Recent
modeling performed by Mathpro for the International Council on Clean
Transportation (ICCT) showed a 0.1 volume percent increase in aromatic
content and a 0.2 volume percent decrease in E300 in the summer, and a
0.3 volume percent decrease in aromatics and 0.8 percent volume percent
increase in E300 in the winter. Overall, the Mathpro refinery modeling
showed an annual average decrease in aromatics and increase in E300. 
Recent modeling by Baker & O’Brien for API (Sensitivity Case 3) showed
about half the increase in aromatics that our analysis showed, however
the API study also included the octane impacts of lower RVP, so we could
not determine whether modeling by API would show an increase in
aromatics or not if it solely modeled sulfur control. API’s addendum
to its original study which added an additional control case that solely
modeled a 10 ppm gasoline sulfur standard (no change in RVP), showed a
0.2 volume percent decrease in aromatics during the summertime (no
wintertime fuel quality data was presented, nor was any distillation
data presented either summer or winter).  Furthermore, after we
completed this analysis for the NPRM, we discovered that the LP refinery
cost model that we licensed to use required some improvements to
correctly characterize the qualities of the light and heavy naphtha
streams from the reformer to improve its estimation of E300 and
aromatics content.  Thus, unlike our modeling results shown in Table
7-4, which show a meaningful impact on aromatics and E300, we believe,
consistent with the Mathpro and Baker and O’Brien refinery modeling
studies, that there will be little to no change.  Note that these
improvements are not expected to have any impact on the cost estimates
made by the refinery model. 

The air quality analysis included the changes to aromatics and E300
shown in Table 7-4.  However, because of the concerns above, we did not
reflect these changes in the national emission inventories presented in
Section 7.1.5.

Vehicle Program Inputs

Modeling the controls introduced by the Tier 3 vehicle program required
the development of another set of alternate MOVES database tables to
reflect each aspect of the proposed Tier 3 program.  These database
tables included: 

Gaseous exhaust emissions rates (HC/CO/NOX) for light duty cars, trucks,
and light-heavy trucks (gas and diesel) to reflect the proposed Tier 3
FTP and US06 standards and their phase-in. 

Elemental carbon (EC) and organic carbon (OC) exhaust emission rates for
light duty cars, trucks, and light-heavy trucks (gas and diesel) to
reflect the proposed Tier 3 FTP and US06 PM standards and phase-in. 

Evaporative HC permeation emission rates to reflect the proposed diurnal
test standard, certification fuel, and phase-in.

 Leak prevalence rates for tank vapor venting and liquid leaks to
reflect proposed requirements for evaporative leak detection.  

The development of these alternative emission rates is discussed below
by pollutant, fuel and vehicle regulatory class.  

Gasoline LD HC/CO/NOX Exhaust 

Gaseous emission rates in MOVES are contained in a database table
(EmissionRateByAge) that expresses emission rates as mass per time,
distinguished by emission process (start and running), fuel type (gas
and diesel), vehicle regulatory class (LDV, LDT, Light HD, etc.), model
year, age, and operating mode (power/speed for running, vehicle soak
time for start).  Developing these rates on Tier 3 vehicles required
accounting for expected changes in each of these dimensions.  

The development of Tier 3 emission rates followed the same procedures
used to develop National LEV (NLEV, covering model years 2001-2003) and
Tier 2 rates (covering model years 2004 and later) in the default MOVES
database, as described in the documentation for development of
light-duty exhaust emission rates for MOVES2010 (known as the “MOVES
Light-Duty report”).  However, specific modifications were made to
represent the introduction of Tier 3 standards, summarized below.  Where
no modifications to methods were made, we will refer the reader to the
appropriate section of the MOVES2010 report. In particular, see Section
1.3.4.  

MOVES emission rates are estimated by standard level, model year, age,
and vehicle regulatory class.  There are separate rates for areas with
Inspection/Maintenance programs (I/M) and those without.  Developing the
rates involves six steps, listed below.

1.  Project average Federal Test Procedure (FTP) results by standard
level and vehicle regulatory class.  As in the development of the
default MOVES2010 database outlined in the MOVES Light Duty Report, we
made use of data measured on the FTP cycle in the course of EPA’s
In-use Verification Program (IUVP) to project emissions under the
proposed standards.  For Tier 3, we developed estimates of FTP results
for Tier 3 vehicles based on IUVP data from Tier 2 Bin 2 and 3 vehicles,
including composite results, “cold-start” emissions” (Bag1 minus
Bag3) and “hot-running” emissions (Bag 2 FTP and US06).

2.  Develop phase-in assumptions for model years (MY) 2017 – 2031, by
standard level, vehicle class and model year, including phase-in
assumptions representing the introduction of Tier 3 standards.  

3.  Merge FTP results and Phase-in assumptions.  For running emissions,
calculate weighted ratios of emissions in each model year relative to
those for cars (LDV) in MY2000, which represent Tier 1 LDV (as discussed
in the MOVES Light Duty report, default MOVES rates were projected
forward based on model year 2000 data from state I/M data, in
conjunction with IUVP data for later model years).  

4.  Estimate Emissions by Operating Mode.  Calculate emissions by
operating mode in each model year by multiplying the MY2000 emission
rates by the weighted ratio for each model year.  We assume that the
emissions control at high power (outside ranges of speed and
acceleration covered by Bag 2 of the FTP) is not as effective as at
lower power (within the range of speed and acceleration covered by Bag
2). 

5.  Apply Deterioration to estimate emissions for three additional age
groups (4-5, 6-7 and 8-9). We assume that Tier 3 vehicles will
deteriorate similarly to other vehicles, when viewed in logarithmic
terms, but we modified deterioration to represent a useful life of
150,000 miles, as opposed to a useful life of 120K miles, assumed for
Tier 2 and NLEV vehicles.  This is the outcome of applying ln-linear
deterioration to the rates developed in steps 1-4.  For the remaining
three groups (10-14, 15-19 and 20+), emissions are assumed to stabilize
as described in the MOVES2010 report.

6. Estimate non-I/M reference rates.  The rates in steps 1-6 represent
rates under a reference inspection/maintenance (I/M) program. 
Corresponding non-I/M rates are calculated by applying the ratios
applied to the Tier 1 and pre-Tier 1 rates.

Each of these six steps is described in more detail below.  Addition
information is available in a separate memo available in the docket.

Average FTP Results (Step 1) (Standard)

Our projected emissions for Tier 3 vehicles are driven by the proposed
NMOG+NOX standard, set at 30 mg/mi. However, because MOVES projects NOX
and THC emissions separately, we apportioned the aggregate standard into
NMOG and NOX components, which we will refer to as the “effective
standards” for each pollutant.  For purposes of apportionment, we
assumed that NMOG control would pose a greater technical challenge than
NOX control.  Accordingly, we assumed “effective standards” for NMOG
and NOX would be 20 mg/mi and 10 mg/mi, respectively. To implement this
assumption, we further assumed that for NOX, vehicles would be
effectively brought into Bin 2, and that for NMOG, vehicles would be
brought to a level between Bin 2 and Bin 3, but closer to Bin 2.

In addition, MOVES models start and running processes separately.  It is
therefore necessary to translate the composite standard into start and
running components.  One component represents a “cold start” on the
FTP cycle, represented as “Bag1 minus Bag3” emissions.  A second
component represents “hot-running” emissions, represented by the
hot-running phase of the FTP (Bag 2). A third component represents
emissions on the US06 cycle, representing emissions at high speed and
power.

Estimated FTP and US06 emissions levels for hydrocarbons (NMOG and NMHC)
are shown in   REF _Ref305667131 \h  Table 7-5 , for several Tier 2 Bins
and for Tier 3.  Values for all standards except Tier 3 are identical to
those used to develop rates in the default database.  The values for
Tier 3 are calculated as a weighted average of those for Bins 2 and 3,
using   REF _Ref305666070 \h  Equation 7-2 .

 

Equation   STYLEREF 1 \s  7 -  SEQ Equation \* ARABIC \s 1  2 

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  5   Hydrocarbons
(HC): Useful-Life FTP Standards and Associated Cold-Start and
Hot-Running Results on the FTP and US06 Cycles. Values for the FTP and
US06 represent NMOG and NMHC, respectively.

Bin	Useful-life Standard

(mg/mi)	FTP Compositea

(mg/mi)	FTP Cold Starta 

(mg)	FTP hot Runninga 

(Bag 2) 

(mg/mi)	US06b 

(mg/mi)

8	125	41.3	591	3.56	35.8

5	90	35.5	534	2.63	35.8

4	70	24.8	383	2.28	35.8

3	55	21.5	329	1.74	35.8

2	10	5.6	87	0.42	2.6

Tier 3c	20	9.2	142	0.7	10.0

a Values represent “non-methane organic gases” (NMOG).

b Values represent “non-methane hydrocarbons” (NMHC).

c Values for Tier 3 calculated using   REF _Ref311790073 \h  Equation
7-2 .

Under a general assumption that CO standards are not forcing, but that
CO emissions tend to track NMOG emissions, corresponding values for CO
were calculated in the same manner, and are presented in   REF
_Ref305667212 \h  Table 7-6 .

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  6  CO: Useful-Life
FTP Standards and Associated Cold-Start and Hot-Running Results on the
FTP and US06 Cycles.

Bin	Useful-life Standard

(mg/mi)	FTP Composite

(mg/mi)	Cold Start 

(mg)	FTP hot Running 

(Bag 2) 

(mg/mi)	US06 

(mg/mi)

8	4,200	861	6,680	451	2,895

5	4,200	606	5,510	238	2,895

4	4,200	537	5,500	201	2,895

3	2,100	463	3,470	119	2,895

2	2,100	235	1,620	70	948

Tier 3a	2,100	286	2,040	81	1,390

a Values for Tier 3 calculated using   REF _Ref311790073  Equation 7-2 .

Corresponding results for NOX are presented in   REF _Ref305667276 \h 
Table 7-7 .  In contrast to HC and CO, the values for Tier 2 Bin 2 were
adopted for Tier 3, as the FTP composite of 5.5 mg/mi suggests that Bin
2 vehicles can meet the “effective standard” of 10 mg/mi with a
reasonable compliance margin.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  7   NOX:
Useful-Life FTP Standards and Associated Cold-Start and Hot-Running
Results on the FTP and US06 Cycles.

Bin	Useful-life Standard

(mg/mi)	FTP Composite

(mg/mi)	Cold Start 

(mg)	FTP hot Running 

(Bag 2) 

(mg/mi)	US06 

(mg/mi)

8	200	64.2	418	35.1	61.3

5	70	21.2	165	8.2	45.9

4	40	8.7	90	4.7	30.6

3	30	5.7	71	3.8	30.6

2	20	5.5	67	0.4	18.4







	Tier 3	10	5.5	67	0.4	18.4

Develop Phase-In Assumptions (Step 2)

We designed phase-in assumptions so as to project compliance with the
Tier 3 fleet average NMOG+NOX requirements.  The requirements are
illustrated in   REF _Ref305668037 \h  Figure 7-4 . The phase-in begins
in model year 2017 and ends in model year 2025.  Note the sharp drop in
emissions at the outset of the Tier 3 phase-in, also that the truck
standards (LDT2,3,4) are slightly higher than the lighter vehicles’
(LDV-T1).  After 2017, the reduction in the fleet average is linear. 
The fleet averages for cars and trucks no longer differ at the
completion of the phase-in.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  4  NMOG+NOX FTP
Fleet Average Requirements during Phase-In of the Tier 3 Exhaust
Emissions Standards for Light-Duty Vehicles.

Merge Cycle Results and Phase-In Assumptions (Step 3)

The goal of this step is to calculate weighted averages of the FTP
(cold-start and hot-running) results for all standards in each model
year, with the emissions results weighted by applicable phase-in
fractions. We do this step for each vehicle class separately, then
weight the four truck classes together using a set of fractions also
derived from the weighted sales estimates.  

Start and running emissions in each model year are simply calculated as
weighted averages of the emissions estimates and the phase-in fractions.
 The resulting weighted start estimates are used directly to represent
cold-start emissions for young vehicles in each model year (ages 0-3).
For running emissions, however, the averages are not used directly;
rather, each is expressed as a ratio to the corresponding Tier 1 value.

Estimate Emissions by Operating Mode (Step 4)

To project emissions for the 2016-and-later vehicles, we divided the
operating modes for running exhaust into two groups. These groups
represent the ranges of speed and power covered by the hot-running phase
(Bag 2) of the FTP standards (< ~20 kW/Mg), and the ranges covered by
the SFTP standards (primarily the US06 cycle). For convenience, we refer
to these two regions as “the hot-running FTP region” and “US06
region,” respectively (See   REF _Ref305668986 \h  Figure 7-5 ). Data
measured on the SC03 cycle did not play a role in emission rate
development.

 To estimate emissions by operating mode, the approach was to multiply
the emission rates for MY 2000, representing Tier 1, by a specific ratio
for each model year from 2016 to 2025, to represent emissions for that
model year.  

To estimate rates for the US06 modes, we followed a procedure similar to
that for the “FTP” modes, but using the “US06” columns in   REF
_Ref305667131 \h  Table 7-5  through   REF _Ref305667276 \h  Table 7-7 .
 For HC and CO, we used   REF _Ref305666070 \h  Equation 7-2 , as
before.  For NOX, we applied the Bin-2 values.    REF _Ref305669681 \h 
Figure 7-6  and   REF _Ref305669684 \h  Figure 7-7  show application of
the ratios to the FTP and US06 operating modes in model years 2010,
2017, and 2025, representing fully phased-in Tier 2 standards, an
interim year during the Tier 3 phase–in, and the fully phased-in Tier
3 standards, respectively.  Figure 7-3 displays the information on
linear scale, highlighting the differences at the higher operating
modes, while   REF _Ref305669684 \h  Figure 7-7  shows the same
information on a logarithmic scale, illustrating the patterns for the
lower operating modes.

  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  5  Operating
modes for running Exhaust Emissions, divided broadly into “hot-running
FTP” and “US06” regions.

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  6  Projected
Emission Rates for Cars in Operating modes 21-30, vs. VSP, in ageGroup
0-3 years, for three model years,  for (a) CO, (b) THC and (c) NOX
(LINEAR SCALE).

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  7  Projected
Emission Rates for Cars in Operating modes 21-30, vs. VSP, in ageGroup
0-3 years, for Four Model Years,  for (a) CO, (b) THC and (c) NOX
(LOGARITHMIC SCALE).

Apply Deterioration (Step 5)

Based on our extensive emissions analysis during MOVES2010 development,
we assume that deterioration for different technologies is best
represented by a multiplicative model, in which different technologies,
represented by successive model-year groups, show similar deterioration
in relative terms but markedly different deterioration in absolute
terms.  We implemented this approach by translating emissions for the
0-3 age group, as calculated above, into natural logarithms and applying
uniform logarithmic age trends to all model-year groups.  We derived
logarithmic deterioration slopes for Tier 1 vehicles (MY 1996-98) and
applied them to Tier 2 vehicles.  In this process we applied the same
logarithmic slope to each operating mode, which is an extension of the
multiplicative deterioration assumption. For 2017 and later model year
vehicles, the deterioration assumptions were modified to represent the
extension of the full useful life (FUL), which is increased from 120,000
mi to 150,000 mi. However, we did not extrapolate the deterioration
trend beyond the 8-9 year age group, as we know that emissions tend to
stabilize beyond this age, while the ln-linear emissions model would
project an increasingly steep and unrealistic exponential emissions
trend.  For the 10-14, 15-19 and 20+ age groups, the “stabilization of
emissions with age” was estimated as for MOVES2010 (MOVES Light Duty
report, section 1.3.3.7).

Estimate Non-I/M References (Step 6)

Completion of the preceding steps provided a set of rates representing
I/M reference rates for MY 2016-2025.  As a final step, we estimated
non-I/M reference rates by applying the same ratios used in MOVES2010
(section 1.3.3.6).

Start Emissions

The values for “Cold Start” shown in Tables 8- 4 through 6 above
were used to represent cold-start emissions for the various standard
levels.  These are designated as opModeID=108 in the emissionRateByAge
table; emission rates for starts following shorter soak periods were
developed by applying standard soak curves (found in the MOVES Light
Duty report) to the updated cold start rates.  Deterioration was applied
to start emissions, using the same approach as used for developing
MOVES2010 base rates discussed in the MOVES Light Duty report.  Start
deterioration is expressed relative to deterioration for running
emissions.

Final Estimates of Composite FTP and US06

In producing emission inventory estimates, MOVES combines emission rates
with activity patterns derived from surveys of in-use vehicles.  These
emissions do not necessarily correlate directly with the test procedures
used for compliance; for example, in-use activity shows that more miles
are driven per start event than assumed on the FTP.  Likewise, the US06
cycle is focused on compliance, and represents a relatively small
portion of in-use driving.  However, to give a relative sense of the
changes projected by the proposed Tier 3 standards, emissions can be
constructed for FTP composite and US06 from MOVES emission rates for the
Tier 2 (labeled as model year 2010) and Tier 3 (labeled as model year
2025) cases.  These are shown in Figures 7-5 through 7-8 below.  Note
that the Tier 3 rates shown below are for the MOVES base fuel of 30 ppm.
 In modeling the control scenarios on 10 ppm, these emission rates were
further lowered by the sulfur reductions outlined in Section 7.1.3.4.1. 


Figure    STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  8  FTP
Composite NOX emissions for reference (2010) and Tier 3 (2025)
constructed from MOVES rates

Figure 7-6 FTP Composite THC emissions constructed from MOVES rates

Figure 7-7 US06 NOX emissions constructed from MOVES rates

Figure 7-8 US06 THC emissions constructed from MOVES rates

Diesel LD HC/CO/NOX Exhaust

Emission rates representing light-duty diesel vehicles under Tier 3
standards were calculated identically to those representing gasoline
vehicles, with the exception that the “effective standards” were set
differently.  Again, diesel vehicles are projected to meet the same
NMOG+NOX standard as gasoline vehicles (30 mg/mi). However, for diesel
vehicles, we assumed that light-duty vehicles would meet Bin-2 standards
following completion of the phase in. Accordingly, the “effective
standards” for NMOG and NOX were set at 10 and 20 mg/mi, respectively.
 As mentioned, all remaining steps were conducted as described in
7.1.3.3.1 above. As a result of the different effective standards,
however, the ratios and other numeric results specific to diesel
vehicles vary slightly from their counterparts for gasoline vehicles.

Gasoline MD HC/CO/NOX

The proposed Tier 3 program will affect not just light duty vehicles
(below 8,500 pounds GVW), but also chassis certified vehicles between
8,500 and 14,000 pounds.  These vehicles are referred to here as medium
duty vehicles, but are also commonly known as Class 2b and 3 heavy
trucks.  In MOVES, these vehicles are designated regulatory class 41. 
Regulatory class 41 also captures certain other vehicles, namely engine
certified trucks and medium duty passenger vehicles, which are not
regulated underneath the proposed medium duty standards.  As described
in the reference case medium duty updates to MOVES, we assumed that
during this timeframe engine certified vehicles and medium-duty
passenger vehicles (MDPV) are five percent and fifteen percent of the
regulatory class respectively.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  8  Population
Percentage

 Category	Percent of Reg Class 41

MDPV	15%

Class 2B	60%

Class 3	20%

Engine Certified	5%

The Class 2b and Class 3 vehicle program was modeled to begin in model
year 2017 and fully phase in during the 2022 model year (  REF
_Ref303164571 \h  Figure 7-9 ).  This yields an aggregate standard of
0.178 g/mile NMOG+NOX for Class 2b vehicles and 0.247 gram/mile for
Class 3 vehicles in 2022.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  9  Class 2B and
3 Standard Phase-in

We combined the proposed Tier 3 phase-in for Class 2b and 3 vehicles
with the existing emission standards MDPV and engine certified vehicles
that comprise MOVES regulatory class 41.  For this analysis, we assumed
that MDPVs met the Tier 3 SULEV 30 standard, and that engine certified
vehicles would perform at 1.2x their standard on the chassis FTP. 
Calculated using Table 7-8, the weighted average of the Class 2b, Class
3, MDPV, and engine standards is 0.181.  To account for the real world
performance of these vehicles, we related this average back to Tier 2
light duty vehicles, for which we have a significant amount of data from
EPA’s in-use verification program. Using the same tool that was used
for developing the Tier 2 and Tier 3 light duty emission rates, we
modeled MOVES Regulatory Class 41 in 2022 as 90 percent Bin 5 and 10
percent Bin 8 vehicles.  For the phase-in years of 2017-2021, we
calculated new MOVES rates as a weighted average of the Tier 3 rates and
the existing MOVES rates for regulatory class 41 such that the
appropriate weighted composite was calculated each year   REF
_Ref302485461 \h  Table 7-9 .  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  9  Phase-in of MD
Tier 3 Rates

Model Year	Tier 3 Rate	MOVES DB Rates	Composite Standard (g/mile)

2017	34%	66%	0.33

2018	49%	51%	0.29

2019	62%	38%	0.26

2020	75%	25%	0.24

2021	87%	13%	0.21

2022	100%	0%	0.18

The CO standards for MD vehicles are less stringent than those for Tier
2 Bin 5 and Bin 8 vehicles.  For Bin 5 and Bin 8 vehicles, the CO
standard is 4.2 g/mile.  For engine certified vehicles, the standard is
approximately 17.3 grams per mile (14.4 grams per bhp multiplied by
1.2),  and for the Tier 3 MD vehicles, the standard ranges from 4.2 to
7.3 g/mile.  Using the same weighted averages as before, we calculated
an aggregate CO standard of 4.4 grams/mile, which is 5.5 percent higher
than the Tier 2 Bin 5/8 standards.  To compensate for the lower CO
emissions in the Tier 2 vehicles that were used to develop the Tier 3 MD
emission rates, we multiplied the running CO rates by 1.1 and the start
CO rates by 1.05.  

Like in the light duty vehicles, deterioration modeled to represent a
150,000 mile useful life.  The same methodology was used for light duty
and medium duty vehicles.

Diesel MD HC/CO/NOX

For medium duty diesel vehicles, the emission rates currently in MOVES
are significantly below the proposed Tier 3 HC and CO standards.  When
MOVES is used to generate a simulated FTP estimate for NMHC, the model
calculates a rate of approximately 0.05 grams per mile, while the
simulated FTP estimate for CO is less than 1 gram/mile.  Consequently,
we assumed no HC and CO emission benefits from Tier 3 standards on MD
diesel vehicles.

By contrast, we estimate that the Tier 3 NOX standard will produce a
reduction in diesel Class 2b and Class 3 NOX emissions.  Because data on
current NOX emissions are limited, as there is little in-use data on MY
2010 and 2011 vehicles which use selective catalytic reduction as a NOX
control strategy, we used a proportional approach to estimate the Tier 3
effect, reducing NOX in proportion to the change in the emission
standard.  Because emission standards tend to impact start and running
emissions differently, we applied a greater portion of the reduction to
running emissions and a smaller reduction to start emissions.  These
reductions were phased-in over the same schedule as for gasoline
vehicles, as detailed in   REF _Ref306259193  Table 7-10 .  Also, to
account for the change in “useful life”, we duplicated the Tier 3
age 0-3 NOX rates to the 4-5 year age-group.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  10  Phase-in of MD
Diesel Tier 3 NOX Rates

Model Year	Tier 3 Phase In	Reduction in  NOX Running Emission Rate
Reduction in NOX Start Emission Rate

2017	20%	12%	5%

2018	38%	23%	9%

2019	54%	33%	12%

2020	69%	42%	16%

2021	85%	52%	19%

2022	100%	61%	23%

Gasoline PM 2.5

Tier 3 will reduce direct particulate matter (PM2.5) emissions from
light-duty vehicles through a full-useful life (FUL) PM standard on the
FTP of 3 mg/mi. Additionally, the Tier 3 standards will include more
stringent SFTP PM standards for light-duty vehicles, with a 10 mg/mi
standard on the US06 for light-duty passenger vehicles, and 20 mg/mi on
the US06 for light-duty trucks. These standards are targeting several
processes that contribute to particulate matter in light-duty gasoline
vehicles: cold starts, high-power operation, and deterioration of engine
and emission control technology over the life of the vehicle. 

To achieve the FUL PM standards without sacrificing fuel economy, it is
anticipated that manufacturers may try a variety of strategies including
reducing lubrication oil consumption over the life of the vehicle. For
our analysis, we projected that reductions in lube oil consumption would
reduce the organic carbon (OC) fraction of PM by 30 percent for both
cold start and running emissions. These reductions are based on an
analysis of the Kansas City Light-duty Vehicle Emissions Study, which
collected PM emissions from a randomly recruited sample of vehicles in
the Kansas City area in 2005. Using chemical tracers found in both the
lube oil and the particulate samples, the lube oil contribution to PM
was estimated to be 25 percent of the PM emissions, primarily from the
OC fraction of PM. The 25 percent reduction is estimated by weighting
the sample to represent the distribution of vehicle ages and vehicle
types in the Kansas City Metropolitan Area in 2004.  

From this analysis, a 30 percent OC reduction is projected in Tier 3
gasoline vehicles due to decreased lube oil consumption, which
represents approximately a 25 percent PM reduction for gasoline vehicles
during cold start and running emission processes. The modeling
assumptions and overview of the Kansas City analysis are located in the
Supporting Technical Document: Estimated Reductions in Particulate
Matter Emissions from Light-duty and Medium-duty Gasoline Vehicles
through Implementation of Tier 3 Regulations.

For the modeling runs described here, the OC reductions are anticipated
to be phased-in over a four year period, with the full-implementation
occurring in 2020. The OC reductions are 7.5 (2017), 15 (2018), 22.5
(2019), and 30 percent (2020). The percent reductions are applied to
cold-start and running emission processes, to passenger car and truck
sources, and to PM emissions as the vehicles age.

Diesel PM

The Tier 3 controls were modeled as having no impact on light duty
diesel PM emissions.  

Gasoline Evaporative Emissions

The proposed Tier 3 evaporative program, requiring lower emissions on
the diurnal test procedure on E15 certification fuel and strengthened
in-use detection of vapor leaks, is projected to cause a significant
reduction in evaporative hydrocarbon emissions.  For this analysis,
tighter diurnal standards in conjunction with E15 certification fuel
were attributed to reductions in evaporative permeation emissions, since
the current certification standards are aimed at not allowing any vented
vapor emissions during the test.  The new requirements for in-use leak
detection were modeled as resulting in a reduced prevalence (frequency
rate) of fuel system vapor and liquid leaks.

Permeation Improvements

Permeation emissions include fuel vapors that escape from a vehicle
through micro pores in the various fuel system components and materials.
Tier 3 will reduce the allowable emissions from this process. Light duty
vehicles will see a reduction from 0.50 g/test to 0.300 g/test,
approximately a 40 percent reduction. 

The Tier 2 permeation rate in MOVES is 0.0102 g/hour on E0 fuel. 
Analysis of the impact of ethanol on permeation emissions conducted as
part of the RFS2 final rule, and included in MOVES2010, suggests that
the use of E15 as Tier 3 certification test fuel will effectively double
permeation emissions over the test procedure. Therefore, the combination
of lowering the vehicle standard and certifying on a fuel with higher
propensity to permeate must be accounted for in Tier 3 permeation rates.


The Tier 3 rate is developed by estimating permeation emissions over one
day of diurnal activity (65F-105F) on an ethanol-containing fuel using
various reductions in the base rate. The total permeation emissions for
the day should equal about 75 percent of the standard (~0.225g) as the
other 25 percent of the standard can be attributed to the Hot Soak
portion of the certification test. The result is a Tier 3 permeation
rate of 0.0026g/hour (a 75 percent reduction from the Tier 2 rate).

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  11  Tier 3
Permeation Rates 

Model Year	Tier 3 Phase-in	Permeation (g/hr)

Tier 2	0%	0.0102

2016_2017	40%	0.0072

2018_2019	60%	0.0056

2020_2021	80%	0.0041

2022	100%	0.0026

Reduced frequency of vapor leaks

EPA, in conjunction with the state of Colorado and the Coordinating
Research Council, undertook multiple research programs to help quantify
the prevalence of evaporative system leaks in the real world, and the
emissions they cause.,  The proposed evaporative leak provisions grew
from this work, and informed the emission inventory contribution of
evaporative leaks, and the level of reductions possible from an in-use
program focused on reducing the incidence of these leaks.  To establish
the reference case, the frequency of evaporative system leaks were
estimated from the prevalence of high evaporative emission vehicles in
the Colorado field study.  Because the Colorado study was not able to
distinguish leaks from other high evaporative emission sources in the
broader population, we analyzed data from CRC’s E-77 program and found
that insufficient canister purge is also contributing to high
evaporative emissions, as this causes the canisters to become
oversaturated, venting vapor into the atmosphere.  

Because the Colorado data suggested high evaporative emissions on newer
vehicles as well as old, we assumed that the leak prevalence rates from
the Colorado study included a degree of emissions from insufficient
purge. Because insufficient purge would not be a function of age, we
assume most (all but 1 percent) of the Colorado prevalence rates at age
0-3 are due to insufficient purge, which is assumed constant with age. 
The increase in high evaporative emissions incidence with age was then
attributed to an increasing prevalence of leaks. This is illustrated in
the chart to the left in   REF _Ref316558277  Figure 7-10 .

To model the control scenario, because the proposed leak requirement
would only address leaks and not insufficient purge, only the leak
portion of high evaporative emissions were reduced.  Colorado data on a
subset of vehicles that were selected for more detailed observation and
testing suggested that 70 percent of the evaporative leaks detected were
due to durability of the evaporative and/or fuel system – e.g.
problems like corroded fuel lines, filler neck, cracked hoses etc. that
could be addressed with improved durability (the other 30 percent were
due to issues beyond the manufacturers control, such as improper
maintenance or missing gas caps).  We estimated that the leak program
would lead to a 70 percent reduction in the occurrence of vapor leaks. 
This is reflected in the chart on the right in   REF _Ref316558277 
Figure 7-10 ; note that the emissions attributed to insufficient purge
were not changed from the reference case.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  10  Vapor Leak
Occurrence Assumed for Reference (Left) and Tier 3 (Right)

Reduced frequency of liquid leaks

Similar to vapor leaks, we expect a reduction in the occurrence of
liquid leaks due to improved system design and integrity. We believe
that liquid leaks occurring in advanced evaporative systems will be
primarily caused by tampering and mal-maintenance. Therefore we have
reduced the frequency rate for leaks for vehicles less than 15 years of
age, and expect vehicles older than 15 to have the same rate of leak
occurrence as current technologies.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  12  Reductions of
Liquid Leaks in Tier 3

Age	Operating	Hot Soak	Cold Soak

0-9	45%	45%	45%

10-14	30%	30%	30%

15-19	0%	0%	0%

20+	0%	0%	0%

Updates to MOVES Sulfur Effects

In order to evaluate the emission impacts of the proposed sulfur
standards, the version of MOVES used for this analysis made significant
updates to the effect of fuel sulfur levels below 30ppm on exhaust
emissions.  In MOVES2010a these effects were based on extrapolation of
data on sulfur levels above 30 ppm.  The updates made for this analysis
were based on significant new data generated from EPA research conducted
in 2010-11, summarized below.  A draft report on this research is
available in the docket.

EPA Sulfur Research Program

Fuel sulfur content has long been understood to affect the performance
of emission aftertreatment catalysts in light duty vehicles, where the
sulfur and/or its oxides adsorb to the active precious metal sites,
reducing the catalyst’s efficiency in destroying harmful pollutants. 
This can severely impair the effectiveness of the catalyst to convert
the products of combustion, leading to increases in these emissions
relative to a “clean” catalyst.  The quantity of sulfur present on
the catalyst at any given time is a function of its temperature and the
fuel sulfur level, with elevated catalyst temperatures and lower fuel
sulfur concentration both reducing sulfur loading.  Numerous studies
have shown the direct impact of fuel sulfur levels above 30 ppm on
emissions; these formed the basis of the Tier 2 rulemaking, which
considered the impact of sulfur in terms of immediate impact, and
irreversible impact due to permanent catalyst damage. 

With the advent of the Tier 2 sulfur standards, new research has focused
on the emission reduction potential of lowering sulfur levels below 30
ppm, particularly on Tier 2 technology vehicles, under the hypothesis
that increased reliance on the catalytic convertor would result in a
higher sensitivity to sulfur accumulation.  A study conducted by EPA and
the auto industry on nine Tier 2 vehicles in support of the Mobile
Source Air Toxics (MSAT) rule, found significant reductions in NOX, CO
and total HC when the vehicles were tested on low sulfur fuel, relative
to 32 ppm fuel.  In particular, the study found a nearly 50 percent
increase in NOX when sulfur was increased from 6 ppm to 32 ppm.  Another
recent study by Umicore showed reductions of 41 percent for NOX and 17
percent for HC on a PZEV operating on fuel with 33 ppm and 3 ppm sulfur.
 Both of these studies conducted testing on high and low sulfur after
running the test vehicles through test cycles meant to clean the
catalyst from the effects of prior sulfur exposure.  

Both of these studies showed the emission reduction potential of lower
sulfur fuel on Tier 2 and later technology vehicles over the FTP cycle. 
However, assessing the potential for reduction on the in-use fleet
requires understanding how sulfur exposure over time impacts emissions,
and the state of sulfur loading for the typical vehicle in field.  In
response, EPA conducted a new study to assess the emission reductions
expected from the in-use Tier 2 fleet with a reduction in fuel sulfur
level from current levels.  It 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
higher speed and load operation over the course of day-to-day driving.  

The study sample described in this analysis consisted of 81 cars and
light trucks recruited from owners in southeast Michigan, covering model
years 2007-9 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.  A nominal
concentration of approximately 25 ppm was targeted for the high level to
be representative of retail fuel available to the public in the vehicle
recruiting area.  All emissions data was collected using the FTP cycle
at a nominal temperature of 75°F.

Using the 28 ppm test fuel, emissions data were collected from vehicles
in their as-received state, and then following a high-speed/load
“clean-out” procedure consisting of two back-to-back US06 cycles
intended to reduce sulfur loading in the catalyst.  A statistical
analysis of this data showed highly significant reductions in several
pollutants including NOX and hydrocarbons (Table 7-11), suggesting that
reversible sulfur loading exists in the in-use fleet and has a
measurable effect on aftertreatment performance; for example, Bag 2 NOX
emissions dropped 32 percent between the pre- and post-cleanout tests on
28 ppm fuel.

Table 7-11 Average Clean-out Effect on In-use Emissions using 28 ppm
Test Fuel

	NOx

(p-value)	THC

(p-value)	CO

(p-value)	NMHC

(p-value)	CH4

(p-value)	PM

(p-value)

Bag 1	–	–	4.7% (0.0737)	–	–	15.4%

(< 0.0001)

Bag 2	31.9% (0.0009)	16.5% (0.0024)	–	17.8% (0.0181)	15.3% (0.0015)
–

Bag 3	38.3% (<0.0001)	21.4% (<0.0001)	19.5% (0.0011)	27.8% (<0.0001)
12.0% (<0.0001)	24.5% (<0.0001)

FTP Composite	11.4% (<0.0001)	4.1% (0.0187)	7.6% (0.0008)	3.0% (0.0751)
6.9% (0.0003)	13.7% (<0.0001)

Bag 1 – Bag 3	–	–	4.2%

(0.0714)	–	–	–

Where no reduction estimate is provided, the clean-out effect is not
significant at α = 0.10.

To assess the impact of lower sulfur fuel on in-use emissions, a
representative subset of vehicles was kept to conduct testing on 28 ppm
and 5 ppm fuel with accumulated mileage.  A first step in this portion
of the study was to assess differences in the effectiveness of the
clean-out procedure when done using different fuel sulfur levels.  Table
7-12 presents a comparison of emissions immediately following (<50
miles) the clean-out procedures at the low vs. high sulfur level.  These
results show significant emission reductions for the 5 ppm fuel relative
to the 28 ppm fuel immediately after this clean-out; for example, Bag 2
NOX emissions were 47 percent lower on the 5 ppm fuel vs. the 28 ppm
fuel.  This indicates that the catalyst is not fully desulfurized, even
after a clean out procedure, as long as there is sulfur in the fuel. 

Table 7-12  Reduction in emissions from 28 ppm to 5 ppm immediately
following clean-out

	NOx

(p-value)	THC

(p-value)	CO

(p-value)	NMHC

(p-value)	CH4

(p-value)	PM‡

Bag 1	5.9%

(0.0896)	5.4%

(0.0118)	7.3%

(0.0023)	4.6%

(0.0465)	11.1%

(<0.0001)	–

Bag 2	47.3%

(0.0010)	40.2%

(<0.0001)	– ‡	34.4%

(0.0041)	53.6%

(<0.0001)	–

Bag 3	51.2%

(<0.0001)	35.0%

(<0.0001)	10.1%

(0.0988)	45.0%

(<0.0001)	25.4%

(<0.0001)	–

FTP Composite	17.7%

(0.0001)	11.2%

(<0.0001)	8.3%

(0.0003)	8.8%

(0.0003)	21.4%

(<0.0001)	–

Bag 1 – Bag 3	– ‡	– ‡	5.8%

α = 0.10.  

To assess the overall in-use reduction between high and low sulfur fuel,
a mixed model analysis of all data as a function of fuel sulfur level
and miles driven after cleanout was performed.  This analysis found
highly significant reductions for several pollutants, as shown in   REF
_Ref309110540 \h  Table 7-13 ; reductions for Bag 2 NOX were
particularly high, estimated at 59 percent between 28ppm and 5ppm
overall For some pollutants, such as Bag 2 NOX, the model fitting did
not find a significant miles-by-sulfur interaction, suggesting the
relative differences were not dependent on miles driven after clean-out.
 Other results, such as Bag 1 hydrocarbons, did show a significant
miles-by-sulfur interaction.  In this case, determination of a sulfur
level effect for the in-use fleet required estimation of a
miles-equivalent level of sulfur loading, which can be gleaned from the
cleanout results obtained from the baseline testing on the vehicles
as-received.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  13 :  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%

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.

	Major findings from this study include:

Reversible sulfur loading is occurring in the 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.

The overall reductions found in this study are in agreement with other
low sulfur studies conducted on Tier 2 vehicles.  The magnitude of NOX
and HC reductions found in this study when switching from 28 ppm to 5
ppm fuel  are consistent with those found in MSAT and Umicore studies
mentioned above.  

Implementation in MOVES

 The results shown in Table 7-13 were applied in MOVES 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 MOVES remains in place.   REF _Ref302460557 \h 
Equation 7-3  shows the generic form of the new sulfur correction;   REF
_Ref302460892 \h  Table 7-14  shows the specific values for the sulfur
coefficients by pollutant, process, and vehicle type.

 

Equation   STYLEREF 1 \s  7 -  SEQ Equation \* ARABIC \s 1  3  Low
Sulfur Effect

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  14  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.  

Nonroad Emissions 

The nonroad sector includes a wide-range of mobile emission sources
ranging from locomotives and construction equipment to hand-held lawn
tools. In the nonroad sector, the only emissions that are directly
affected by the proposed Tier 3 regulation are the emissions from
gasoline-powered equipment such as lawn-mowers, recreational boats and
all-terrain vehicles.  Their SO2 emissions are reduced with the proposed
decrease in gasoline sulfur levels.  As with onroad, reference and
control case emissions were generated using the fuel supply inputs
reflecting the post-EISA/EPAct renewable fuel volumes.  

Gasoline and land-based diesel nonroad emissions were estimated using
EPA’s NONROAD2008 model, as run by the EPA’s consolidated modeling
system known as the National Mobile Inventory Model (NMIM).  The fuels
in the NMIM database, NCD2010201Tier 3, were developed from the
reference and control fuels used for onroad vehicles, as described in
Section 7.1.3. In 2005, onroad and nonroad gasoline formulations are
assumed to be identical. In 2017, E10 and E15 ethanol blends are
available in every county, but nonroad equipment is assumed to use E10
only. In 2030, we assume almost all nonroad equipment can use either E10
or E15, so onroad and nonroad gasoline formulations are again assumed to
be identical.  For all years, the reference case included the higher
sulfur reference gasoline and the control case met the proposed sulfur
limits.  

Since aircraft, locomotive and commercial marine emission sources do not
burn gasoline, their emission factors are unaffected by the changes in
gasoline fuels that were developed for this rule.  Hence, their
emissions are the same for both the reference and control cases.  The
emissions from these sources used for this rule are the same as they
were for the Heavy-Duty Greenhouse Gas Rule (2011).  The procedures for
calculating emissions from locomotives and C1/C2 commercial marine were
developed for the Locomotive Marine Rule (2008) and are detailed in the
RIA for that rule.  The procedures used for calculating C3 commercial
marine emissions are those developed in the recent C3 Rule (2010).

Criteria and Toxic Emission Impact Results 

The proposed Tier 3 rule will reduce NOX (including NO2), VOC, CO, and
SO2 from all gasoline-powered on road vehicles immediately upon
implementation of lower sulfur fuel, and will further reduce these
emissions as well as PM2.5 from cars, light trucks and light heavy-duty
trucks (gas and diesel) as tighter emission standards from these
vehicles phase in.  There also will be reductions in SO2 emissions from
the nonroad gasoline fleet as a result of sulfur standards.  The
reductions are summarized in this section for each pollutant.  

NOX reductions are shown in   REF _Ref306283279  Table 7-15  for
Calendar Years 2017 through 2030, and 2050 when the fleet will have
turned over completed to Tier 3 standards.  We project significant
reductions immediately upon implementation of the program, growing to a
nearly 30 percent reduction in onroad emissions by 2030, and nearly 40
percent reduction in onroad emissions with full fleet turnover by 2050. 


Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  15  Tier 3 NOX
Reductions by Calendar Year (Annual U.S. Short Tons)

Year	Onroad mobile reference	Onroad mobile with control	Reduction 
Percent reduction in onroad

2017	3,452,314	3,167,934	284,381	8.2%

2018	3,179,942	2,890,474	289,469	9.1%

2019	2,942,238	2,646,150	296,088	10.1%

2020	2,732,380	2,429,517	302,863	11.1%

2021	2,557,573	2,241,209	316,364	12.4%

2022	2,427,305	2,091,971	335,335	13.8%

2023	2,304,619	1,950,687	353,932	15.4%

2024	2,196,463	1,823,335	373,128	17.0%

2025	2,115,326	1,717,083	398,244	18.8%

2026	2,040,468	1,617,809	422,659	20.7%

2027	1,989,071	1,540,270	448,801	22.6%

2028	1,942,891	1,469,181	473,710	24.4%

2029	1,913,717	1,413,482	500,234	26.1%

2030	1,890,403	1,365,613	524,790	27.8%

2050	2,410,373	1,507,204	903,170	37.5%

  REF _Ref306283294  Table 7-16  shows the reduction in NOX emissions,
in annual short tons, projected in calendar years 2017 and 2030.  The
reductions are split into those attributable to the introduction of low
sulfur fuel in the pre-Tier 3 fleet (defined for this analysis as model
years prior to 2017); and reductions attributable to vehicle standards
enabled by low sulfur fuel (model year 2017 and later).  As shown, in
2017 over 90 percent of the program reductions are coming from lower
sulfur gasoline on the fleet already on the road.  By 2030, nearly 90
percent of the reduction is coming from 2017 and later model year
vehicles, with remaining reduction coming from lower sulfur fuel on
pre-Tier 3 vehicles.  We project that about one tenth of the reductions
from the Tier 3 fleet in 2030 are attributable to the proposed Tier 3
heavy-duty emission standards.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  16  Projected NOX
Reductions from Tier 3 Program (Annual U.S. Short Tons)

	2017	2030

Total reduction	284,381	524,790

Reduction from pre-Tier 3 fleet due to sulfur standard	264,653	66,286

Reduction from Tier 3 fleet due to vehicle and sulfur standards 	19,728
458,504

VOC reductions are shown in   REF _Ref306283308  Table 7-17  for
Calendar Years 2017 through 2030, and 2050 when the fleet will have
turned over completed to Tier 3 standards.  We project reductions of
over 40,000 tons (3 percent of the onroad fleet emissions) immediately
upon implementation of the program, growing to a 23 percent reduction in
onroad emissions by 2030, and 36 percent reduction in onroad emissions
with full fleet turnover by 2050.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  17  Tier 3 VOC
Reductions by Calendar Year (Annual U.S. Short Tons)

CY	Onroad mobile reference	Onroad mobile with control	Reduction 	Percent
reduction in onroad

2017	1,614,384	1,569,603	44,782	2.8%

2018	1,506,001	1,456,342	49,659	3.3%

2019	1,411,787	1,356,874	54,913	3.9%

2020	1,333,867	1,272,497	61,370	4.6%

2021	1,266,250	1,195,112	71,138	5.6%

2022	1,211,710	1,128,740	82,970	6.8%

2023	1,163,986	1,067,208	96,778	8.3%

2024	1,123,476	1,012,446	111,029	9.9%

2025	1,084,012	955,680	128,332	11.8%

2026	1,045,808	899,298	146,509	14.0%

2027	1,025,657	858,927	166,730	16.3%

2028	1,004,922	818,008	186,913	18.6%

2029	988,529	781,665	206,863	20.9%

2030	977,067	751,040	226,028	23.1%

2050	1,151,908	735,548	416,361	36.1%

  REF _Ref306283336  Table 7-18  shows the VOC reductions in 2017 and
2030 split into those attributable to the pre-Tier 3 fleet, and the Tier
3 fleet.  The Tier 3 fleet reductions are further subdivided into the
contribution of the proposed exhaust and evaporative standards.  In 2017
over 80 percent of the program reductions are coming from lower sulfur
gasoline on the fleet already on the road.  By 2030, over 90 percent of
the reduction is coming from 2017 and later model year vehicles, with
remaining reduction coming from lower sulfur fuel on pre-Tier 3
vehicles.  The evaporative standards account for about one quarter of
the Tier 3 fleet reductions in 2030. 

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  18  Projected VOC
Reductions from Tier 3 Program (Annual U.S. Short Tons)

	2017	2030

Total reduction	44,782	226,028

Reduction from pre-Tier 3 fleet due to sulfur standard	39,561	13,739

Reduction from Tier 3 fleet due to vehicle and sulfur standards	5,222
212,289



Exhaust	41,433	168,264

Evaporative 	3,349	57,764

CO reductions are shown in   REF _Ref306283352  Table 7-19  for Calendar
Years 2017 through 2030, and 2050 when the fleet will have turned over
completed to Tier 3 standards.  We project significant reductions
immediately upon implementation of the program, growing to a 30 percent
reduction in onroad emissions by 2030, and 46 percent reduction in
onroad emissions with full fleet turnover by 2050.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  19  Tier 3 CO
Reductions by Calendar Year (Annual U.S. Short Tons)

CY	Onroad mobile reference	Onroad mobile with control	Reduction 	Percent
reduction in onroad

2017	20,915,593	20,168,910	746,683	3.6%

2018	20,297,089	19,454,190	842,899	4.2%

2019	19,782,476	18,830,656	951,820	4.8%

2020	19,374,712	18,295,010	1,079,702	5.6%

2021	19,122,796	17,747,325	1,375,471	7.2%

2022	18,830,157	17,119,383	1,710,775	9.1%

2023	18,736,417	16,610,543	2,125,874	11.3%

2024	18,647,649	16,095,479	2,552,170	13.7%

2025	18,647,177	15,584,593	3,062,585	16.4%

2026	18,621,982	15,054,696	3,567,286	19.2%

2027	18,699,029	14,558,766	4,140,263	22.1%

2028	18,745,355	14,047,003	4,698,352	25.1%

2029	18,850,303	13,594,884	5,255,420	27.9%

2030	18,951,626	13,186,263	5,765,362	30.4%

2050	24,839,365	13,339,791	11,499,574	46.3%

  REF _Ref306283370  Table 7-20  shows the reductions for CO, broken
down by pre- and post-Tier 3 in the manner described for NOX and VOC
above.  The immediate reductions in the onroad fleet from sulfur control
comprise about 80 percent of total reductions in 2017.  By 2030, the
proposed vehicle standard, enabled by low sulfur fuel, are accounting
for 98 percent of program reductions.  Of the Tier 3 vehicle standard
reductions in 2030, we estimate that about 5 percent are contributed by
the proposed heavy-duty tailpipe standards.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  20  CO Reductions
from Tier 3 Program (Annual U.S. Short Tons)

	2017	2030

Total reduction	746,683	5,765,362

Reduction from pre-Tier 3 fleet due to sulfur standard	608,502	139,074

Reduction from Tier 3 fleet due to vehicle and sulfur standards	138,181
5,626,288

Direct PM2.5 impacts are shown in   REF _Ref306283378  Table 7-21  for
calendar years 2017 through 2030, and 2050 when the fleet will have
turned over completed to Tier 3 standards.  For direct PM, the impact
shown is solely from the proposed tailpipe standards.  Thus, unlike
other pollutants, reductions do not become significant until the fleet
has turned over to cleaner vehicles.  By 2030, we project a reduction of
about 7,500 tons annually, which represents approximately 10 percent of
the onroad direct PM2.5 inventory.  However, since the PM standards are
mainly focused on improving engine durability through the end of a
vehicle’s useful life, the relative reduction in onroad emissions is
projected to grow to 17 percent with full fleet turnover in 2050.  

As discussed in Section 7.2.1.1, the control scenario emissions
inventory prepared for air quality modeling included the impact of an
increase in aromatics as sulfur is reduced  from 30 ppm to 10 ppm fuel. 
While this assumption had a minor effect on control case emissions for
several pollutants, the effect was more visible for direct PM2.5
emissions, as it resulted in a projected increase in emissions (roughly
700 tons nationwide) in 2017.  As discussed in Section 7.1.3.2.2, this
emissions increase results from a series of conservative assumptions and
uncertainties related to fuel parameters in 2017, and is not expected to
occur in reality.    As a result, we have not included the PM emissions
increase from sulfur reduction in the emission inventory impacts shown
in Table 7-21.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  21  Tier 3 PM2.5
Reductions by Calendar Year (Annual U.S. Short Tons)

CY	Onroad mobile reference	Onroad mobile with control	Reduction 	Percent
reduction in onroad

2017	115,098	114,977	121	0.1%

2018	107,295	106,932	362	0.3%

2019	100,885	100,162	724	0.7%

2020	95,192	93,992	1,200	1.3%

2021	90,480	88,767	1,713	1.9%

2022	85,144	82,943	2,201	2.6%

2023	81,859	79,081	2,778	3.4%

2024	78,955	75,561	3,393	4.3%

2025	76,935	72,890	4,045	5.3%

2026	75,204	70,502	4,702	6.3%

2027	73,880	68,496	5,384	7.3%

2028	72,504	66,431	6,073	8.4%

2029	71,990	65,210	6,780	9.4%

2030	71,554	64,096	7,458	10.4%

 	 



	2050	92,895	77,279	15,616	16.8%

Emissions of air toxics also would be reduced by the proposed sulfur,
exhaust and evaporative standards.  Air toxics are generally a subset of
compounds making up VOC, so the reduction trends tend to track the VOC
reductions presented above.  Table 7-22 presents reductions for certain
gaseous air toxics and polycyclic aromatic hydrocarbons (PAHs),
reflecting reductions of a few percent in 2017, and 20 to 40 percent of
onroad emissions, depending on the individual pollutant, in 2030.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  22  Reductions for
Certain Individual Compounds (Annual U.S. Short Tons)

	2017 Reduction 	Percent reduction in onroad	2030 Reduction 	Percent
reduction in onroad

Acetaldehyde	762 	3%	4,414 	26%

Formaldehyde	727	3%	2,707 	12%

Acrolein	23 	1%	184 	15%

1,3-Butadiene	322 	5%	1,087 	37%

Benzene	1,625 	4%	8,581 	36%

Naphthalene	96 	2%	420 	17%

Ethanol	2,684	2%	27,821 	24%

2,2,4-Trimethylpentane	840	2%	5,616	18%

Ethyl Benzene	724	3%	3,204	22%

Hexane	857	3%	3,525	30%

Propionaldehyde	29	2%	102	20%

Styrene	98	4%	755	38%

Toluene	3,504	3%	16,965	22%

Xylene	2,856	3%	14,238	22%

PAHs	7	1%	67	17%

The totals shown in Table 7-23 represent the sum of all toxic species in
Table 7-22.  As shown, in 2030 the overall on-road inventory of total
toxics would be reduced by over 20 percent, with nearly one third of the
vehicle program reductions coming from the proposed evaporative
standards.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  23  Reductions in
Total Mobile Source Air Toxics (Annual U.S. Short Tons)

	2017	2030

Total reduction	15,156	89,685

Reduction from pre-Tier 3 fleet due to sulfur standard	13,184	5,022

Reduction from Tier 3 fleet due to vehicle and sulfur standards	1,972
84,663

Exhaust	13,748	64,144

Evaporative 	1,408	25,541

SO2 emissions from mobile sources are a direct function of sulfur in the
fuel, and reducing sulfur in gasoline would result in immediate
reductions in SO2 from the on and off-road fleet.  The reductions, shown
in   REF _Ref306283840  Table 7-24 , represent a roughly 50 percent
reduction in onroad SO2 emissions.  The breakdown of the relative
contribution of on-road vehicles and off-road equipment is shown; the
contribution of off-road sources is a function of off-road gasoline
consumption accounting for approximately 5 percent of overall gasoline
use.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  24  Projected SO2
Reductions from Tier 3 Program (Annual U.S. Short Tons)

	2017	2030

Total reduction	16,261	17,267

Reduction from onroad vehicles due to sulfur standard	15,494	16,370

Reduction from off-road equipment due to sulfur standard	767	897

Percent reduction in onroad SO2  emissions	51%	51%

Criteria and Toxic Pollutant Air Quality Impacts 

Emission Inventories for Air Quality Modeling 

As noted in Section 7.1, emission inventories for air quality modeling
were required for the entire U.S. by 12 km grid cell and hour of the day
for each day of the year, requiring a methodology with much greater
detail than the national emission inventories presented above.  While
most of the modeling tools and inputs used for estimating national
emission inventories were also used in developing inputs for air quality
modeling, the application of these tools (particularly MOVES) to produce
the gridded / hourly emissions was quite different, and in essence a
separate analysis.  As explained in Section   REF _Ref310244172 \r \h 
7.2.1.1 , the different analyses generated different onroad inventory
totals, but the relative reduction from reference to control scenarios
was consistent.  The summary of the methodology for each sector is
contained in the following sections; for brevity, details of the process
for developing air-quality ready emission inventories are available in a
separate technical support document.

Onroad Emissions 

For the onroad vehicle emissions inputs to our air quality modeling, we
used an emission inventory approach that provided more temporal and
geographical resolution than the approach used for the national
inventories described above.  While modeling at this level is
time-consuming and resource intensive, this detail is needed when
generating inputs to air quality models because it allows us much more
precision in accounting for local ambient temperatures and local fuel
properties in our air quality modeling.  For this purpose, we used
county-specific inputs and tools that integrated the MOVES model of
onroad emissions with the Sparse Matrix Operator Kernel Emissions tool
(SMOKE) emission inventory model to take advantage of the gridded hourly
temperature information used in air quality modeling. 

In particular, we used an automated process to run MOVES to produce
emission factors by temperature and speed for the fleet mix, fuels, and
I/M program for more than 100 “representing counties”, to which
every other county could be mapped.  The emission factors then were
multiplied by activity at the grid-cell-hour level to produce gridded
hourly emissions for the entire continental U.S.  These emissions were
input into the Community Multiscale Air Quality Modeling System (CMAQ). 
We summarize this approach in the sections below.  

We used the same approach to model both the reference and the control
cases, except, for the Tier 3 control case, we used the Tier 3 emission
rates and fuels developed for the national inventories and described in
Section   REF _Ref305133367 \r \h  \* MERGEFORMAT  7.1.3 .

Because of differences in methodology, particularly the treatment of
vehicle speed distributions and the non-linear temperature effects in
MOVES, the more detailed approach used for the air quality inventory
produced different emission estimates than those described in the
national inventory section above.  The two sets of results are compared
in Table 7-25 below.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  25  Comparison of
Calendar Year 2030 Onroad Emission National Inventories and Inventories
Used for Air Quality Modeling [U.S. Short tons]

Pollutant	Reference	Control

	National Inventory	Air Quality Inventory	Difference AQ vs. NI	National
Inventory	Air Quality Inventory	Difference AQ vs. NI

NOX	1,890,403	1,846,571	-2%	1,365,613	1,371,925	0%

VOC	977,067	911,513	-7%	751,040	699,592	-7%

CO	18,951,626	17,021,674	-10%	13,186,263	11,984,061	-9%

PM2.5	71,554	88,516	24%	64,917	83,842	29%

Benzene	23,654	22,221	-6%	15,073	14,352	-5%

Ethanol	107,912	116,762	8%	80,091	89,574	12%

Acrolein	1,223	863	-29%	1,039	699	-33%

1,3-Butadiene	2,915	2,932	1%	1,828	1,955	7%

Formaldehyde	21,967	14,810	-33%	19,260	12,270	-36%

Acetaldehyde	16,757	13,926	-17%	12,343	9,946	-19%

SO2	31,983	30,526	-5%	15,613	15,068	-3%

The differences between the national inventories and air quality
inventories reflect the non-linear response to the more detailed
handling of temperature and other local variables such as speed in the
air quality inventory; this is pronounced in pollutants with strong
temperature sensitivities in MOVES, such as PM2.5, where the finer
temperature resolution in the air quality approach produced
significantly higher emissions than the aggregate national inventory
approach.  

Because the reference and control case emissions rates were the same for
the national inventory and air quality inventory runs, the percent
reductions due to the proposed Tier 3 rule are very similar, as shown in
Table 7-26.  The exception is PM2.5, where the air quality inventory
shows a slight increase in emissions in 2017.  As discussed in Section
7.1.3.2.2, this increases resulted from a series of conservative
assumptions and uncertainties related to fuel parameters in 2017 which
we do not expect to occur in reality.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  26   Comparison of
Emission Reductions from Reference to Control Case in “National” and
“Air Quality” Onroad Inventories

	2017	2030

Pollutant	National Inventory Reduction	Air Quality Inventory Reduction
National Inventory Reduction	Air Quality Inventory Reduction

NOX	-8%	-7%	-28%	-26%

VOC	-3%	-3%	-25%	-23%

CO	-4%	-3%	-30%	-30%

PM2.5	-0.1%	1%	-10%	-5%

Benzene	-4%	-4%	-36%	-35%

Ethanol	-3%	-2%	-26%	-23%

Acrolein	-1%	-1%	-15%	-19%

1,3-Butadiene	-5%	-5%	-37%	-33%

Formaldehyde	-2%	-3%	-12%	-17%

Acetaldehyde	-3%	-3%	-26%	-29%

SO2	-51%	-50%	-51%	-51%

The following sections summarize the analysis done to generate the air
quality inventories.

Representing Counties

Air quality modeling requires emission inventories for nearly all of the
more than 3,000 counties in the United States.  Although EPA compiles
county-specific databases for all counties in the nation, actual
county-specific data is rare.  Instead, much of our “county” data is
based on state-wide estimates or national defaults.  For this proposal,
rather than explicitly model every county in the nation, we have done
detailed modeling for some counties and less detailed estimates for the
other counties.  

In this approach, we group counties that have similar properties that
would result in similar emission rates.  We explicitly model only one
county in the group (the "representing" county) to determine emission
rates.  These rates are then used in combination with county specific
activity and meteorology data, to generate inventories for all of the
counties in the group.  This approach dramatically reduces the number of
modeling runs required to generate inventories and still takes into
account differences between counties.

The grouping of counties was based on several characteristics as
summarized in Table 7-27 below.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  27 
Characteristics for Representing County Groupings

County Grouping Characteristic	Description

PADD	Petroleum Administration for Defense Districts (PADDs).  PADD 1 is
divided into three sub-PADD groupings and each sub-group is treated as a
separate PADD (1a, 1b and 1c).  Each state belongs to a PADD and all
counties in any state are within the same PADD.

Fuel Parameters	Average gasoline fuel properties for January and July
2005, including RVP, sulfur level, ethanol fraction and percent benzene

Emission Standards	Some states have adopted California highway vehicle
emission standards or plan to adopt them. Since implementation of the
standards varies, each state with California standards is treated
separately.

Inspection/Maintenance Programs	Counties were grouped within a state
according to whether or not they had an I/M program.  All I/M programs
within a state were considered as a single program, even though each
county may be administered separately and have a different program
design.



Altitude	Counties were categorized as high or low altitude based on the
criteria set forth by EPA certification procedures (4,000 feet above sea
level).

The result is a set of 106 county groups with similar fuel, emission
standards, altitude and I/M programs.  For each group, the county with
the highest VMT was chosen as the representing county.  Of these, only
103 were needed to model the 48 states included in the air quality
analysis inventory. 

For each county group, SMOKE-MOVES generated a set of rates that varied
by vehicle type, speed and temperature, thus we did not need to consider
the fleet mix, speed or temperature range in our grouping
characteristics.  This greatly increases the number of counties that can
be in each grouping, and reduces the number of MOVES runs required.  

More detail on the process for selecting representative counties and a
list of all of the 3,322 counties in the nation and the counties
selected to represent is provided in the emission inventory technical
support document.

SMOKE-MOVES

The official EPA highway vehicle emissions model (MOVES) was updated as
described in Section   REF _Ref305133367 \r \h  7.1.3  for national
emission inventory development, but in order to take advantage of the
gridded hourly temperature information used in air quality modeling,
MOVES and SMOKE have been integrated into an inventory generation system
called SMOKE-MOVES.  MOVES can be run in “inventory mode” to
calculate the mass of pollutant emissions, as was done for the national
inventories, or in “emission rate” mode, in which it calculates
emissions in grams per mile (for running emissions) or grams per vehicle
(for start and evaporative emissions).  For our air quality runs, we
used the rates approach.  This creates a set of “lookup tables” with
emission rates by temperature, speed, pollutant, and vehicle class
(Source Classification Code (SCC)). SMOKE then transforms these rates
into emission inventories for the air quality modeling by multiplying
these emission factors by activity specific to each grid cell hour,  

The SMOKE-MOVES process generates MOVES run specification files to
produce the emission rate lookup tables (in MOVES there are three per
run to cover all emission processes: Rate Per Distance, Rate per
Vehicle, and Rate per Profile) covering the range of temperatures
needed, across each combination of fuel and I/M program in the nation. 
For a given scenario, this resulted in over 16,000 run specification
files.  A series of post-processing scripts were developed to take the
raw MOVES emission rate table results and translate it into the emission
rates tables needed by SMOKE to produce mass emissions by 12 km grid and
hour of the day, for an entire year.  Note, an update to these
post-processing scripts was made between the reference case and control
case runs, which inadvertently introduced a small inconsistency in
emission rates between reference and control for about one-third of the
counties.  Specifically, there were 1,218 counties, out of 3,109 total
counties, which were impacted.  The result was that for some counties,
control case emission rates were a few percent higher than they should
have been; fixing this error would increase the magnitude of reduction
in the air quality analysis.  

For expediency, MOVES lookup tables were generated for July and January
to get the full range of temperatures needed for an entire year’s
worth of meteorology data.  This efficiency step introduces uncertainty
because it does not account for fuel “shoulder” seasons in the fall
and spring, where the actual fuel pool is a blend of winter and summer
fuel.  This is mainly an issue for fuel RVP, which is not changing
between the reference and control scenarios.  

Inputs to MOVES

The county-level fuel-property inputs for the air quality runs were the
same as for the national inventories described in Section   REF
_Ref305133367 \r \h  7.1.3 .  However, for the air quality runs, we were
able to use grid-level temperatures.  We also needed county-specific
information on vehicle populations, age distributions, and
inspection-maintenance programs for each of the representing counties. 
The source data for each of these inputs is described below.

Temperature and Humidity

Ambient temperature can have a large impact on emissions.  Cold
temperatures are associated with high start emissions for many
pollutants.  High temperatures are associated with greater running
emissions due to the higher engine load of air conditioning.  High
temperatures also are associated with higher evaporative emissions. 
And, of course, the interaction between emissions and ambient
temperatures is an important consideration in air quality modeling. 
Thus accurately accounting for ambient temperatures was important for
our air quality modeling work.

 The gridded meteorological input data for the entire year of 2005 were
derived from simulations of the Pennsylvania State University / National
Center for Atmospheric Research Mesoscale Model.  This model, commonly
referred to as MM5, is a limited-area, nonhydrostatic, terrain-following
system that solves for the full set of physical and thermodynamic
equations which govern atmospheric motions.  A description of how this
tool was used to determine temperatures is found in the documentation
for the recent Heavy-Duty Greenhouse Gas rule.

SMOKE-MOVES uses the MM5 temperatures and the county groups described in
Section   REF _Ref310237613 \r \h  7.2.1.1.1  to generate a list of all
the possible temperatures and temperature profiles that are needed in
the lookup tables.  (Temperature profiles are vectors of 24 temperatures
that describe how temperatures change over a day. They are needed to
estimate vapor venting emissions.)  To do this, SMOKE-MOVES determines
the minimum and maximum temperatures in the county group for January and
for July, and the minimum and maximum temperatures for each hour of the
day.  It then generates a list of all possible temperatures between
these limits, using a five degree Fahrenheit interval.  The model also
uses these temperatures (using a 10 degree interval) to develops a
collection of possible temperature profiles.

SMOKE-MOVES then runs MOVES for each of the listed temperatures and
temperature profiles, generating emission rate look-up tables that cover
the desired temperature ranges.  Finally, the original grid cell
temperatures are used to find the appropriate emission rate for each
cell.  The 2005 temperatures were used for all scenarios.

The treatment of humidity is simpler.  SMOKE-MOVES calculates an average
day-time (6am to 6pm) relative humidity for the county group for July
and for January. The appropriate (July or January) humidity is used for
all runs of the county group.

Vehicle Population Inputs

Vehicle population data is a required input for MOVES when modeling on a
county basis.  Using the technical guidance provided to states by EPA, a
contractor generated appropriate estimates for vehicle populations for
use in the MOVES databases using the county specific VMT and national
average ratios of vehicle populations versus vehicle VMT from the MOVES
application.  This method is described in Section 3.3 of the document,
"Technical Guidance on the Use of MOVES2010 for Emission Inventory
Preparation in State Implementation Plans and Transportation Conformity"
(EPA-420-B-10-023, April 2010), which is available on the EPA web site
at: http://www.epa.gov/otaq/models/moves/index.htm

Other Local Inputs 

In addition to temperature, vehicle population and fuels, we also needed
inputs such as age distribution and Inspection Maintenance program
descriptions for each of the representing counties.  These inputs are
required for the model to run at the county level and provided an
opportunity to assure that the model was properly accounting for the
most recent available local data.  These county inputs were derived from
the inputs used for the National Emissions Inventory (NEI).  This
inventory covers the 50 United States (U.S.), Washington DC, Puerto Rico
and U.S. Virgin Islands. The NEI was created by the U.S. Environmental
Protection Agency's (EPA's) Emission Inventory Group (EIG) in Research
Triangle Park, North Carolina, in cooperation with the Office of
Transportation and Air Quality in Ann Arbor, Michigan. The inputs for
the NEI are stored in the National Mobile Inventory Model (NMIM) county
database (NCD).  Details of how the NCD was developed are documented for
the NEI.  These inputs were then converted to a format consistent with
MOVES.

Parallel Processing on the “Cloud”  

Providing the level of detail desired for the air quality modeling
required an enormous amount of data. Even with the “representing
county” approach, we ran MOVES over 83,000 times to support Tier3 rule
making.  Processing just one of the Tier 3 scenarios required 16,604
runs, which if run serially would take over 200 days.  Early on, we
recognized that this would be infeasible, even with the fastest
computers available to us.  Therefore, we developed a Linux-based
environment at Amazon Web Services that enabled us to process the Tier 3
base case in less than 48 hours.  We split the 16,604 runs into 206
batches (103 representative counties, for January and July). We then ran
206 Linux processor instances in parallel, with each instance processing
80 - 120 individual MOVES runs.  

VMT, Population, and Speed

In addition to the lookup tables, SMOKE requires county VMT, population,
and average speed by road type to calculate the necessary emissions for
air quality modeling.  

VMT by county and Source Classification Code (SCC) was developed using
MOVES2010a and the National County Database.  MOVES2010a has the EPA’s
most recent projections of VMT growth at the national level, based on
estimates from the Annual Energy Outlook.  The National County Database
(NCD20101201) has our most recent estimates of 2005 VMT and our best
estimates of allocation of VMT from national to the county level. 
Accordingly, for the 2005 base year, our estimates of VMT by county and
SCC were taken directly from the NCD.  For the 2017 and 2030
inventories, we ran MOVES2010a with default inputs to generate total
national VMT by SCC.  But, because MOVES uses a static (1999) default
allocation of VMT to county, we did not use MOVES for these allocations.
 Instead, the 2017 county VMT was created by interpolating between the
NCD VMT values for 2015 and those for 2020 and computing the NCD
fraction for each county, then multiplying these fractions by the MOVES
VMT.  The 2030 county allocation was computed similarly, using the NCD
VMT for 2030.  The VMT was also adjusted to account for increased onroad
transportation of ethanol fuels and the resulting increase in travel by
large tanker trucks.  For both the reference and control scenarios,
impacts of this activity on emissions from tank trucks (Class 8) are
accounted by adjusting VMT used in SMOKE-MOVES.  The VMT adjustments
were derived from the Oak Ridge National Laboratory analysis of ethanol
transport, scaled to account for the ethanol volumes. 

Vehicle populations by county and SCC were developed similarly to the
VMT, using MOVES to generate national totals for each year and using the
NCD to allocate to county.  However, the NCD does not include population
estimates, so we used MOVES to generate the 2005 national population and
we assumed that, for each calendar year (2005, 2017 and 2030) and for
each SCC, the allocation of  national vehicle population to county was
proportional to the allocation of VMT (summed across roadtypes).  

The average speeds provided to SMOKE for each county were derived from
the default national average speed distributions found in the default
MOVES2010a database AvgSpeedDistribution table.  These average speeds
are the average speeds developed for the previous EPA highway vehicle
emission factor model, MOBILE6.  In MOVES, there is a distribution of
average speeds for each hour of the day for each road type.  The average
speeds in these distributions were used to calculate an overall average
speed for each hour of the day.  These hourly average speeds were
weighted together using the default national average hourly vehicle
miles traveled (VMT) distribution found in the MOVES default database
HourlyVMTFraction table, to calculate an average speed for each road
type.  This average speed by road type was provided to SMOKE for each
county.

Nonroad Emissions

The “primary” nonroad emissions used in air quality modeling are
identical to those used for national inventories as presented in Section
  REF _Ref306284340 \r  7.1.4  above.  The NMIM model was run to
generate county-month inventories by SCC, which were processed to
gridded-hourly emissions by SMOKE.  For more details on SMOKE processing
of nonroad emissions, see the emissions modeling technical support
document.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  28  Comparison of
Calendar Year 2030 Nonroad Emission National Inventories and Inventories
Used for Air Quality Modeling [U.S. Short tons]

Pollutant	Reference	Control

	National Inventory	Air Quality Inventory	Difference	National Inventory
Air Quality Inventory	Difference

NOX	765,026	765,026	0.0%	765,026	765,026	0.0%

VOC	1,209,534	1,209,534	0.0%	1,209,452	1,209,452	0.0%

CO	12,921,772	12,921,772	0.0%	12,921,772	12,921,772	0.0%

PM2.5	68,308	68,308	0.0%	68,308	68,308	0.0%

Benzene	23,246	23,390	0.6%	23,245	23,389	0.6%

Acrolein	571	570	-0.2%	571	570	-0.2%

1,3-Butadiene	1,772	1,774	0.1%	1,772	1,774	0.1%

Formaldehyde	13,580	13,522	-0.4%	13,580	13,522	-0.4%

Acetaldehyde	7,899	7,888	-0.1%	7,899	7,888	-0.1%

SO2	3,154	3,154	0.0%	2,257	2,257	0.0%

Refueling

Methodology

This section describes how the emission inventories for refueling from
on-road vehicles in calendar years 2017 and 2030 for Tier 3 reference
and control cases were generated for air quality modeling.  The
refueling inventory includes emissions from spillage loss and
displacement vapor loss.  The displacement emissions vary from scenario
to scenario depending on the RVP of the modeled fuels.  For this
analysis, the refueling emissions were estimated using the revised
version of EPA’s Motor Vehicle Emissions Simulator (MOVES2010a) at the
county level for all twelve months.  We used the same fuel supply inputs
as we used for the onroad inventories described in Section   REF
_Ref302458260 \r \h  7.1.3.2 .

 As for onroad emissions, described above, we used a “representing
county” approach to reduce MOVES runtime.  Additional information on
the use of representing counties to model refueling emissions is
available in a technical support document.

Emission Inventory Results

The annual refueling emission inventories for air quality modeling for
the lower 48 states are shown in   REF _Ref305740452 \h  Table 7-29 ,
along with the percent changes between the two scenarios. The Tier 3
proposal to eliminate the RVP waiver for E10 fuels reduces refueling
emissions in 2017, but has no impact in 2030, where we model E15 as the
only gasoline available for onroad vehicles in both the reference and
control cases.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  29  Refueling
Emissions for Tier 3 reference and control (U.S. Short Tons)

	2017	2030

	Reference	Control	Percent Reduction	Reference	Control	Percent Reduction

VOC	75,397	74,235	-1.5%	56,405	56,402	0.0%

Benzene	158	160	1.6%	90	90	0.0%

Ethanol	8,582	8,426	-1.8%	7,143	7,142	0.0%

Portable Fuel Container and Upstream Emissions

While the Tier 3 proposed rule has no impact on portable fuel container
(PFC) or upstream emissions associated with fuel production and
transport/distribution, a significant number of modifications were made
to the 2005v.4.2 platform inventory to account for impacts of renewable
fuel requirements under EISA in the reference case air quality
inventory.  These modifications are described in detail in a memorandum
to the docket.  Modifications to point and nonpoint inventories include
adjustments to agricultural emissions, increases in emissions associated
with production of corn ethanol, cellulosic ethanol, cellulosic diesel,
and biodiesel, decreases in petroleum refinery emissions to account for
gasoline displacement, and increased vapor loss emissions from transport
of ethanol and gasoline/ethanol fuel blends.  Modifications to mobile
source inventories include increases in combustion emissions from water,
rail and truck transport of biofuels.  PFC emissions were adjusted to
account for impacts of RVP changes associated with use of
gasoline/ethanol blends.

Hydrocarbon Speciation Profiles and SMOKE

We used the Community Multi-scale Air Quality (CMAQ) model, described in
detail in the following section, to conduct air quality modeling for
this analysis.  The SMOKE tool is used to process emission inventories
for air quality modeling.  Specifically, SMOKE converts our air quality
emissions inventories into CMAQ-ready inputs by transforming the
emission inventories based on the temporal allocation, chemical
speciation, and spatial allocation requirements of CMAQ. In processing
our Tier 3 emissions inventories for CMAQ, SMOKE uses hydrocarbon
speciation profiles to break total hydrocarbons down into individual
constituent compounds and create the needed chemical speciation inputs
required for CMAQ.  Given the complexity of the atmospheric chemistry,
the hydrocarbon speciation can have an important influence on the air
quality modeling results.  We recently created a number of new
hydrocarbon speciation profiles for vehicle exhaust and evaporative
emissions and headspace vapor. These profiles were created using data
from the EPAct test program, CRC’s E-77 series of evaporative
emissions programs described in Section 7.2 above,, and recent
measurements of speciated gasoline headspace vapors collected by EPA’s
Office of Research and Development (ORD).  Mobile source hydrocarbon
speciation profiles used in this analysis are listed in a memo to the
docket.  

Air Quality Modeling Methodology

Air quality models use mathematical and numerical techniques to simulate
the physical and chemical processes that affect air pollutants as they
disperse and react in the atmosphere. Based on inputs of meteorological
data and source information, these models are designed to characterize
primary pollutants that are emitted directly into the atmosphere and
secondary pollutants that are formed as a result of complex chemical
reactions within the atmosphere.  Photochemical air quality models have
become widely recognized and routinely utilized tools for regulatory
analysis by assessing the effectiveness of control strategies.  These
models are applied at multiple spatial scales - local, regional,
national, and global.  This section provides detailed information on the
photochemical model used for our air quality analysis (the Community
Multi-scale Air Quality (CMAQ) model), atmospheric reactions and the
role of chemical mechanisms in modeling, and model uncertainties and
limitations.  Further discussion of the modeling methodology is included
in the Air Quality Modeling Technical Support Document (AQM TSD) found
in the docket for this rule.  Results of the air quality modeling are
presented in Section   REF _Ref309219793 \n \h  7.2.4 .

Modeling Methodology

A national-scale air quality modeling analysis was performed to estimate
future year 8-hour ozone concentrations, annual PM2.5 concentrations,
24-hour PM2.5 concentrations, annual NO2 concentrations, air toxics
concentrations, visibility levels and nitrogen and sulfur deposition
levels for 2017 and 2030.  The 2005-based CMAQ modeling platform was
used as the basis for the air quality modeling for this proposed rule. 
This platform represents a structured system of connected
modeling-related tools and data that provide a consistent and
transparent basis for assessing the air quality response to projected
changes in emissions.  The base year of data used to construct this
platform includes emissions and meteorology for 2005.  The platform was
developed by the U.S. EPA’s Office of Air Quality Planning and
Standards in collaboration with the Office of Research and Development
and is intended to support a variety of regulatory and research model
applications and analyses.

The CMAQ modeling system is a non-proprietary, publicly available,
peer-reviewed, state-of-the-science, three-dimensional, grid-based
Eulerian air quality model designed to estimate the formation and fate
of oxidant precursors, primary and secondary PM concentrations, acid
deposition, and air toxics, over regional and urban spatial scales for
given input sets of meteorological conditions and emissions.,,  The CMAQ
model version 4.7 was most recently peer-reviewed in February of 2009
for the U.S. EPA.  The CMAQ model is a well-known and well-respected
tool and has been used in numerous national and international
applications.,,  This 2005 multi-pollutant modeling platform used the
most recent CMAQ code available at the time of air quality modeling
(CMAQ version 4.7.1) with a minor internal change made by the U.S. EPA
CMAQ model developers intended to speed model runtimes when only a small
subset of toxics species are of interest.  

CMAQ includes many science modules that simulate the emission,
production, decay, deposition and transport of organic and inorganic
gas-phase and particle-phase pollutants in the atmosphere.  We used CMAQ
v4.7.1 which reflects updates to version 4.7 to improve the underlying
science.  These include aqueous chemistry mass conservation
improvements, improved vertical convective mixing and lowered CB05
mechanism unit yields for acrolein from 1,3-butadiene tracer reactions
which were updated to be consistent with laboratory measurements. 
Section   REF _Ref305053973 \n  7.2.3  of this draft RIA discusses the
chemical mechanism and SOA formation. 

Model Domain and Configuration

The CMAQ modeling domain encompasses all of the lower 48 States and
portions of Canada and Mexico.  The modeling domain is made up of a
large continental U.S. 36 kilometer (km) grid and two 12 km grids (an
Eastern U.S. and a Western U.S. domain), as shown in   REF _Ref306258174
 Figure 7-11 .  The modeling domain contains 14 vertical layers with the
top of the modeling domain at about 16,200 meters, or 100 millibars (mb)
of atmospheric pressure.

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  11  Map of the
CMAQ Modeling Domain

Model Inputs

The key inputs to the CMAQ model include emissions from anthropogenic
and biogenic sources, meteorological data, and initial and boundary
conditions.  The CMAQ meteorological input files were derived from
simulations of the Pennsylvania State University/National Center for
Atmospheric Research Mesoscale Model for the entire year of 2005 over
model domains that are slightly larger than those shown in   REF
_Ref306258174  Figure 7-11 .  This model, commonly referred to as MM5,
is a limited-area, nonhydrostatic, terrain-following system that solves
for the full set of physical and thermodynamic equations which govern
atmospheric motions.  The meteorology for the national 36 km grid and
the two 12 km grids were developed by EPA and are described in more
detail within the AQM TSD.  The meteorological outputs from MM5 were
processed to create model-ready inputs for CMAQ using the
Meteorology-Chemistry Interface Processor (MCIP) version 3.4. Outputs
include: horizontal wind components (i.e., speed and direction),
temperature, moisture, vertical diffusion rates, and rainfall rates for
each grid cell in each vertical layer.

The lateral boundary and initial species concentrations are provided by
a three-dimensional global atmospheric chemistry model, the GEOS-CHEM
model.  The global GEOS-CHEM model simulates atmospheric chemical and
physical processes driven by assimilated meteorological observations
from the NASA’s Goddard Earth Observing System (GEOS).  This model was
run for 2005 with a grid resolution of 2 degree x 2.5 degree
(latitude-longitude) and 20 vertical layers.  The predictions were used
to provide one-way dynamic boundary conditions at three-hour intervals
and an initial concentration field for the 36 km CMAQ simulations.  The
future base conditions from the 36 km coarse grid modeling were used as
the initial/boundary state for all subsequent 12 km finer grid modeling.

The emissions inputs used for the 2005 base year and each of the future
year base cases and control scenarios analyzed for this rule are
summarized in Section   REF _Ref305059196 \n \h  7.2.1  of this draft
RIA.

CMAQ Evaluation

An operational model performance evaluation for ozone, PM2.5 and its
related speciated components (e.g., sulfate, nitrate, elemental carbon,
organic carbon, etc.), nitrate and sulfate deposition, and specific air
toxics (formaldehyde, acetaldehyde, benzene, 1,3-butadiene, and
acrolein) was conducted using 2005 state/local monitoring data in order
to estimate the ability of the CMAQ modeling system to replicate base
year concentrations.  Model performance statistics were calculated for
observed/predicted pairs of daily/monthly/seasonal/annual
concentrations.  Statistics were generated for the following geographic
groupings: domain wide, Eastern vs. Western (divided along the 100th
meridian), and each Regional Planning Organization (RPO) region.  The
“acceptability” of model performance was judged by comparing our
results to those found in recent regional PM2.5 model applications for
other, non-EPA studies.  Overall, the performance for the 2005 modeling
platform is within the range or close to that of these other
applications.  The performance of the CMAQ modeling was evaluated over a
2005 base case.  The model was able to reproduce historical
concentrations of ozone and PM2.5 over land with low bias and error
results.  Model predictions of annual formaldehyde, acetaldehyde and
benzene showed relatively small bias and error results when compared to
observations.  The model yielded larger bias and error results for 1,3
butadiene and acrolein based on limited monitoring sites.  A more
detailed summary of the 2005 CMAQ model performance evaluation is
available within the AQM TSD found in the docket of this rule.

Model Simulation Scenarios

As part of our analysis for this rulemaking, the CMAQ modeling system
was used to calculate 8-hour ozone concentrations, daily and annual
PM2.5 concentrations, annual NO2 concentrations, annual and seasonal
(summer and winter) air toxics concentrations, visibility levels and
annual nitrogen and sulfur deposition total levels for each of the
following emissions scenarios:

- 2005 base year

- 2017 Tier 3 reference case 

- 2017 Tier 3 control case 

- 2030 Tier 3 reference case 

- 2030 Tier 3 control case 

The emission inventories used in the air quality and benefits modeling
are different from the proposed rule inventories due to the considerable
length of time required to conduct the modeling.  As noted above,
emission inventories for air quality modeling were required for the
entire U.S. by 12 km grid cell and hour of the day for each day of the
year, requiring a methodology of much greater detail than the national
emission inventories presented in Section 7.1.  While most of the
modeling tools and inputs used for estimating national emission
inventories were also used in developing inputs for air quality modeling
as well, the application of these tools (particularly MOVES) to produce
the gridded / hourly emissions was quite different, and in essence a
separate analysis.  As explained in Section   REF _Ref310244172 \r \h 
7.2.1.1 , the different analyses generated different onroad inventory
totals, but the reduction from reference to control scenarios was
consistent.  The emission inventories used for air quality modeling are
discussed in Section   REF _Ref305059196 \n \h  7.2.1  of this draft
RIA.  The emissions modeling TSD, found in the docket for this rule
(EPA-HQ-OAR-2011-0135), contains a detailed discussion of the emissions
inputs used in our air quality modeling.  

We use the predictions from the model in a relative sense by combining
the 2005 base-year predictions with predictions from each future-year
scenario and applying these modeled ratios to ambient air quality
observations to estimate 8-hour ozone concentrations, daily and annual
PM2.5 concentrations, annual NO2 concentrations and visibility
impairment for each of the 2017 and 2030 scenarios.  The ambient air
quality observations are average conditions, on a site-by-site basis,
for a period centered around the model base year (i.e., 2003-2007).  

The projected daily and annual PM2.5 design values were calculated using
the Speciated Modeled Attainment Test (SMAT) approach.  The SMAT uses a
Federal Reference Method (FRM) mass construction methodology that
results in reduced nitrates (relative to the amount measured by routine
speciation networks), higher mass associated with sulfates (reflecting
water included in FRM measurements), and a measure of organic
carbonaceous mass that is derived from the difference between measured
PM2.5 and its non-carbon components.  This characterization of PM2.5
mass also reflects crustal material and other minor constituents.  The
resulting characterization provides a complete mass balance.  It does
not have any unknown mass that is sometimes presented as the difference
between measured PM2.5 mass and the characterized chemical components
derived from routine speciation measurements.  However, the assumption
that all mass difference is organic carbon has not been validated in
many areas of the U.S.  The SMAT methodology uses the following PM2.5
species components: sulfates, nitrates, ammonium, organic carbon mass,
elemental carbon, crustal, water, and blank mass (a fixed value of 0.5
µg/m3).  More complete details of the SMAT procedures can be found in
the report "Procedures for Estimating Future PM2.5 Values for the CAIR
Final Rule by Application of the (Revised) Speciated Modeled Attainment
Test (SMAT)".  For this latest analysis, several datasets and techniques
were updated.  These changes are fully described within the technical
support document for the Final Transport Rule AQM TSD.  The projected
8-hour ozone design values were calculated using the approach identified
in EPA's guidance on air quality modeling attainment demonstrations.  

Additionally, we conducted an analysis to compare the absolute and
percent differences between the future year reference and control cases
for annual and seasonal ethanol, formaldehyde, acetaldehyde, benzene,
1,3-butadiene, and acrolein, as well as annual nitrate and sulfate
deposition.  These data were not compared in a relative sense due to the
limited observational data available.  

Chemical Mechanisms in Modeling

This rule presents inventories for NOX, VOC, CO, PM2.5,  SO2, NH3, and
seven air toxics: benzene, 1,3-butadiene, formaldehyde, acetaldehyde,
ethanol, naphthalene and acrolein.  The air toxics are explicit model
species in the CMAQv4.7 model with carbon bond 5 (CB05) mechanisms. 
Emissions of all the pollutants included in the rule inventories, except
ethanol, were generated using the Motor Vehicle Emissions Simulator
(MOVES) VOC emissions and toxic-to-VOC ratios calculated using EPAct
data.  Ethanol emissions for air quality modeling were based on
speciation of VOC using different ethanol profiles (E0, E10 and E85)
(see Section   REF _Ref306278901 \r  7.2.1.5  for more information).  In
addition to direct emissions, photochemical processes mechanisms are
responsible for formation of some of these compounds in the atmosphere
from precursor emissions.  For some pollutants such as PM, formaldehyde,
and acetaldehyde, many photochemical processes are involved.  CMAQ
therefore also requires inventories for a large number of other air
toxics and precursor pollutants.  Methods used to develop the air
quality inventories can be found in Section   REF _Ref305059196 \r 
7.2.1 .

  In the CB05 mechanism, the chemistry of thousands of different VOCs in
the atmosphere are represented by a much smaller number of model species
which characterize the general behavior of a subset of chemical bond
types; this condensation is necessary to allow the use of complex
photochemistry in a fully 3-D air quality model.

Complete combustion of ethanol in fuel produces carbon dioxide (CO2) and
water (H2O). Incomplete combustion results in the production of other
air pollutants, such as acetaldehyde and other aldehydes, and the
release of unburned ethanol.  Ethanol is also present in evaporative
emissions.  In the atmosphere, ethanol from unburned fuel and
evaporative emissions can undergo photodegradation to form aldehydes
(acetaldehyde and formaldehyde) and peroxyacetyl nitrate (PAN), and also
plays a role in ground-level ozone formation.  Mechanisms for these
reactions are included in CMAQ.  Additionally, alkenes and other
hydrocarbons are considered because any increase in acetyl peroxy
radicals due to ethanol increases might be counterbalanced by a decrease
in radicals resulting from decreases in other hydrocarbons.

CMAQ includes 63 inorganic reactions to account for the cycling of all
relevant oxidized nitrogen species and cycling of radicals, including
the termination of NO2 and formation of nitric acid (HNO3) without PAN
formation.

 + ∙OH + M → HNO3 + M			k = 1.19 x 10-11 cm3molecule-1s-1  

The CB05 mechanism also includes more than 90 organic reactions that
include alternate pathways for the formation of acetyl peroxy radical,
such as by reaction of ethene and other alkenes, alkanes, and aromatics.
 Alternate reactions of acetyl peroxy radical, such as oxidation of NO
to form NO2, which again leads to ozone formation, are also included.

Atmospheric reactions and chemical mechanisms involving several key
formation pathways are discussed in more detail in the following
sections.  

Acetaldehyde

Acetaldehyde is the main photodegradation product of ethanol, as well as
other precursor hydrocarbons.  Acetaldehyde is also a product of fuel
combustion.  In the atmosphere, acetaldehyde can react with the OH
radical and O2 to form the acetyl peroxy radical [CH3C(O)OO∙].  When
NOX is present in the atmosphere this radical species can then further
react with nitric oxide (NO), to produce formaldehyde (HCHO), or with
nitrogen dioxide (NO2), to produce PAN [CH3C(O)OONO2].  An overview of
these reactions and the corresponding reaction rates are provided below.


CH3CHO + ∙OH → CH3C∙O + H2O		k = 1.5 x 10-11 cm3molecule-1s-1  

CH3C∙O + O2 + M → CH3C(O)OO∙ + M

CH3C(O)OO∙ + NO → CH3C(O)O∙ + NO2		k = 2.0 x 10-11
cm3molecule-1s-1  

CH3C(O)O∙ → ∙CH3 + CO2	

∙CH3 + O2 + M → CH3OO∙ + M 

CH3OO∙ + NO → CH3O∙ + NO2

CH3O∙ + O2 → HCHO + HO2

CH3C(O)OO∙ + NO2 + M → CH3C(O)OONO2 + M  k = 1.0 x 10-11
cm3molecule-1s-1  

Acetaldehyde can react with the NO3 radical, ground state oxygen atom
(O3P) and chlorine, although these reactions are much slower. 
Acetaldehyde can also photolyze (hν), which predominantly produces
∙CH3 (which reacts as shown above to form CH3OO∙) and HCO (which
rapidly forms HO2 and CO):

CH3CHO + hν +2 O2 → CH3OO∙ +HO2 + CO		( = 240-380 nm 

As mentioned above, CH3OO∙ can react in the atmosphere to produce
formaldehyde (HCHO).  Formaldehyde is also a product of hydrocarbon
combustion.  In the atmosphere, the most important reactions of
formaldehyde are photolysis and reaction with the OH, with atmospheric
lifetimes of approximately 3 hours and 13 hours, respectively. 
Formaldehyde can also react with NO3 radical, ground state oxygen atom
(O3P) and chlorine, although these reactions are much slower. 
Formaldehyde is removed mainly by photolysis whereas the higher
aldehydes, those with two or more carbons such as acetaldehyde, react
predominantly with OH radicals.  The photolysis of formaldehyde is an
important source of new hydroperoxy radicals (HO2), which can lead to
ozone formation and regenerate OH radicals.  

HCHO + hν + 2 O2 → 2 HO2 + CO		( = 240-360 nm 

HO2 + NO → NO2+ OH

Photolysis of HCHO can also proceed by a competing pathway which makes
only stable products: H2 and CO. 

CB05 mechanisms for acetaldehyde formation warrant a detailed discussion
given the increase in vehicle and engine exhaust emissions for this
pollutant and ethanol, which can form acetaldehyde in the air. 
Acetaldehyde is represented explicitly in the CB05 chemical mechanism,
by the ALD2 model species, which can be both formed from other VOCs and
can decay via reactions with oxidants and radicals.  The reaction rates
for acetaldehyde, as well as for the inorganic reactions that produce
and cycle radicals, and the representative reactions of other VOCs have
all been updated to be consistent with recommendations in the
literature.

The decay reactions of acetaldehyde are fewer in number and can be
characterized well because they are explicit representations.  In CB05,
acetaldehyde can photolyze in the presence of sunlight or react with
molecular oxygen (O3(P)), hydroxyl radical (OH), or nitrate radicals. 
The reaction rates are based on expert recommendations, and the
photolysis rate is from IUPAC recommendations. 

In CMAQ v4.7, the acetaldehyde that is formed from photochemical
reactions is tracked separately from that which is due to direct
emission and transport of direct emissions.  In CB05, there are 25
different reactions that form acetaldehyde in molar yields ranging from
0.02 (ozone reacting with lumped products from isoprene oxidation) to
2.0 (cross reaction of acylperoxy radicals, CXO3).  The specific parent
VOCs that contribute the most to acetaldehyde concentrations vary
spatially and temporally depending on characteristics of the ambient
air, but alkenes in particular are found to play a large role.  The IOLE
model species, which represents internal carbon-carbon double bonds, has
high emissions and relatively high yields of acetaldehyde.  The OLE
model species, representing terminal carbon double bonds, also plays a
role because it has high emissions although lower acetaldehyde yields. 
Production from peroxyproprional nitrate and other peroxyacylnitrates
(PANX) and aldehydes with 3 or more carbon atoms can in some instances
increase acetaldehyde but because they also are a sink of radicals,
their effect is smaller.  Thus, the amount of acetaldehyde (and
formaldehyde as well) formed in the ambient air as well as emitted in
the exhaust (the latter being accounted for in emission inventories) is
affected by changes in these precursor compounds due to the addition of
ethanol to fuels (e.g., decreases in alkenes would cause some decrease
of acetaldehyde, and to a larger extent, formaldehyde).  

 reaction with the hydroxyl radical (∙OH).  This reaction produces
acetaldehyde (CH3CHO) with a 90 percent yield.  The lifetime of ethanol
in the atmosphere can be calculated from the rate coefficient, k, and
due to reaction with the OH radical, occurs on the order of a day in
polluted urban areas or several days in unpolluted areas.   

In CB05, reaction of one molecule of ethanol yields 0.90 molecules of
acetaldehyde.  It assumes the majority of the reaction occurs through
H-atom abstraction of the more weakly-bonded methylene group, which
reacts with oxygen to form acetaldehyde and hydroperoxy radical (HO2),
and the remainder of the reaction occurs at the –CH3 and –OH groups,
creating formaldehyde (HCHO), oxidizing NO to NO2 (represented by model
species XO2) and creating glycoaldehyde, which is represented as ALDX:

CH3CHOH + OH → HO2 + 0.90 CH3CHO + 0.05 ALDX + 0.10 HCHO + 0.10 XO2

Secondary Organic Aerosols (SOA)

Secondary organic aerosol (SOA) chemistry research described below has
led to implementation of new pathways for secondary organic aerosol
(SOA) in CMAQ 4.7, based on recommendations of Edney et al. and the
recent work of Carlton et al.,   In previous versions of CMAQ, all SOA
was semivolatile and resulted from the oxidation of compounds emitted
entirely in the gas-phase.  In CMAQ v4.7, parameters in existing
pathways were revised and new formation mechanisms were added.  Some of
the new pathways, such as low-NOX oxidation of aromatics and
particle-phase oligomerization, result in nonvolatile SOA.

Organic aerosol can be classified as either primary or secondary
depending on whether it is emitted into the atmosphere as a particle
(primary organic aerosol, POA) or formed in the atmosphere (SOA).  SOA
precursors include volatile organic compounds (VOCs) as well as
low-volatility compounds that can react to form even lower volatility
compounds. Current research suggests SOA contributes significantly to
ambient organic aerosol (OA) concentrations, and in Southeast and
Midwest States may make up more than 50 percent (although the
contribution varies from area to area) of the organic fraction of PM2.5
during the summer (but less in the winter).,  A wide range of laboratory
studies conducted over the past twenty years show that anthropogenic
aromatic hydrocarbons and long-chained alkanes, along with biogenic
isoprene, monoterpenes, and sesquiterpenes, contribute to SOA
formation.,,,,  Modeling studies, as well as carbon isotope
measurements, indicate that a significant fraction of SOA results from
the oxidation of biogenic hydrocarbons.,  Based on parameters derived
from laboratory chamber experiments, SOA chemical mechanisms have been
developed and integrated into air quality models such as the CMAQ model
and have been used to predict OA concentrations.  

Over the past 10 years, ambient OA concentrations have been routinely
measured in the U.S. and some of these data have been used to determine,
by employing source/receptor methods, the contributions of the major OA
sources, including biomass burning and vehicular gasoline and diesel
exhaust.  Since mobile sources are a significant source of VOC
emissions, currently accounting for almost 40 percent of anthropogenic
VOC, mobile sources are also an important source of SOA, particularly in
populated areas.

Toluene is an important contributor to anthropogenic SOA.  Mobile
sources are the most significant contributor to ambient toluene
concentrations as shown by analyses done for the 2005 National Air
Toxics Assessment (NATA) and the Mobile Source Air Toxics (MSAT) Rule. 
The 2005 NATA indicates that onroad and nonroad mobile sources accounted
for almost 60 percent (1.46 µg/m3) of the total average nationwide
ambient concentration of toluene (2.48 µg/m3), when the contribution
of the estimated “background” is apportioned among source sectors.

The amount of toluene in gasoline influences the amount of toluene
emitted in vehicle exhaust and evaporative emissions, although, like
benzene, some toluene is formed in the combustion process.  In turn,
levels of toluene and other aromatics in gasoline are potentially
influenced by the amount of ethanol blended into the fuel.  Due to the
high octane quality of ethanol, it greatly reduces the need for and
levels of other high-octane components such as aromatics including
toluene (which is the major aromatic compound in gasoline).  Since
toluene contributes to SOA and the toluene level of gasoline is
decreasing, it is important to assess the effect of these reductions on
ambient PM.

In addition to toluene, other mobile-source hydrocarbons such as
benzene, xylene, and alkanes form SOA.  Similar to toluene, the SOA
produced by benzene and xylene from low-NOX pathways is expected to be
less volatile and be produced in higher yields than SOA from high- NOX
conditions.   Alkanes form SOA with higher yields resulting from the
oxidation of longer chain as well as cyclic alkanes.

It is unlikely that ethanol would form directly from SOA or affect SOA
formation indirectly through changes in the radical populations from
increasing ethanol exhaust.  Nevertheless, scientists at the U.S.
EPA’s Office of Research and Development recently directed experiments
to investigate ethanol’s SOA forming potential.  The experiments were
conducted under conditions where peroxy radical reactions would dominate
over reaction with NO (i.e., irradiations performed in the absence of
NOX and OH produced from the photolysis of hydrogen peroxide). This was
the most likely scenario under which SOA formation could occur, since a
highly oxygenated C4 organic would be potentially made.  As expected, no
SOA was produced. From these experiments, the upper limit for the
aerosol yield would have been less than 0.01 percent based on scanning
mobility particle sizer (SMPS) data.  Given the expected negative result
based on these initial smog chamber experiments, these data were not
published.

In general, measurements of organic aerosol represent the sum of POA and
SOA and the fraction of aerosol that is secondary in nature can only be
estimated. One of the most widely applied method of estimating total
ambient SOA concentrations is the EC tracer method using ambient data
which estimates the OC/EC ratio in primary source emissions.,  SOA
concentrations have also been estimated using OM (organic mass) to OC
(organic carbon) ratios, which can indicate that SOA formation has
occurred, or by subtracting the source/receptor-based total primary
organic aerosol (POA) from the measured OC concentration.  Aerosol mass
spectrometer (AMS) measurements along with positive matrix factorization
(PMF) can also be used to identify surrogates for POA and SOA in ambient
as well as chamber experiments. Such methods, however, may not be
quantitatively accurate and provide no information on the contribution
of individual biogenic and anthropogenic SOA sources, which is critical
information needed to assess the impact of specific sources and the
associated health risk.  These methods assume that OM containing
additional mass from oxidation of OC comes about largely (or solely)
from SOA formation.  In particular, the contributions of anthropogenic
SOA sources, including those of aromatic precursors, are required to
determine exposures and risks associated with replacing fossil fuels
with biofuels.

Upon release into the atmosphere, numerous VOC compounds can react with
free radicals in the atmosphere to form SOA.  While this has been
investigated in the laboratory, there is relatively little information
available on the specific chemical composition of SOA compounds
themselves from specific VOC precursors.  This absence of compositional
data from the precursors has largely prevented the identification of
aromatically-derived SOA in ambient samples which, in turn, has
prevented observation-based measurements of the aromatic and other SOA
contributions to ambient PM levels.

β-caryophyllene, the latter three of which are emitted by vegetation
and are more significant sources of SOA than toluene.  Smog chamber work
can also be used to investigate SOA chemical formation mechanisms.,,,

Although these concentrations are only estimates, due to the assumption
that the mass fractions of the smog chamber SOA samples using these
tracers are equal to those in the ambient atmosphere, there are
presently no other means available for estimating the SOA concentrations
originating from individual SOA precursors.  Among the tracer compounds
observed in ambient PM2.5 samples are two tracer compounds that have
been identified in smog chamber aromatic SOA samples.  To date, these
aromatic tracer compounds have been identified, in the laboratory, for
toluene and m-xylene SOA.  Additional work is underway by the EPA to
determine whether these tracers are also formed by benzene and other
alkylbenzenes (including o-xylene, p-xylene, 1,2,4-trimethylbenzene, and
ethylbenzene).

One caveat regarding this work is that a large number of VOCs emitted
into the atmosphere, which have the potential to form SOA, have not yet
been studied in this way.  It is possible that these unstudied compounds
produce SOA species which are being used as tracers for other VOCs. 
This means that the present work could overestimate the amount of SOA
formed in the atmosphere by the VOCs studied to date.  This approach may
also estimate entire hydrocarbon classes (e.g., all
methylsubstituted-monoaromatics or all monoterpenes) and not individual
precursor hydrocarbons.  Thus the tracers could be broadly
representative and not indicative of individual precursors.  This is
still unknown.  Also, anthropogenic precursors play a role in formation
of atmospheric radicals and aerosol acidity, and these factors influence
SOA formation from biogenic hydrocarbons.  This anthropogenic and
biogenic interaction, important to EPA and others, needs further study. 
The issue of SOA formation from aromatic precursors is an important one
to which EPA and others are paying significant attention.  

The aromatic tracer compounds and their mass fractions have also been
used to estimate monthly ambient aromatic SOA concentrations from March
2004 to February 2005 in five U.S. Midwestern cities.  The annual
tracer-based SOA concentration estimates were 0.15, 0.18, 0.13, 0.15,
and 0.19 μg carbon/m3 for Bondville, IL, East St. Louis, IL,
Northbrook, IL, Cincinnati, OH and Detroit, MI, respectively, with the
highest concentrations occurring in the summer.  On average, the
aromatic SOA concentrations made up 17  percent of the total SOA
concentration.  Thus, this work suggests that we are finding ambient PM
levels on an annual basis of about 0.15 μg/m3 associated with present
toluene levels in the ambient air in these Midwest cities.  Based on
preliminary analysis of recent laboratory experiments, it appears the
toluene tracer could also be formed during photooxidation of some of the
xylenes.

Over the past decade a variety of modeling studies have been conducted
to predict ambient SOA levels. While early studies focused on the
contribution of biogenic monoterpenes, additional precursors, such as
sesquiterpenes, isoprene, benzene, toluene, and xylene, have been
implemented in atmospheric models such as GEOS-Chem, PMCAMx, and CMAQ.,
, , , ,,  Studies have indicated that ambient OC levels may be
underestimated by current model parameterizations. While the treatment
of new precursors has likely reduced the model/measurement bias,
underestimates can persist. In general, modeling studies focus on
comparing the sum of the POA and SOA concentrations with ambient OC or
estimated OA concentrations. Without a method to attribute measured OC
to different sources or precursors, identifying causes of the
underestimates in modeled OC via model/measurement comparisons can be
challenging. Oxidation of low-volatility organic compounds as well as
particle-phase reactions resulting from acidity have been explored as
potential missing sources of OC in models.,

Ozone

+ M → CH3C(O)OO∙ + NO2 + M		k = 3.3 x 10-4 s-1 

The reaction above shows how NO2 is released in the thermal
decomposition of PAN, along with a peroxy radical which can oxidize NO
to NO2 as previously shown in Section 7.2.3.1.  NO2 can also be formed
in photodegradation reactions where NO is converted to NO2 (see OH
radical reaction of acetaldehyde in Section   REF _Ref306260314 \r 
7.2.3.1 ).  In both cases, NO2 further photolyzes to produce ozone (O3).

NO2 + hν → NO + O(3P)			( = 300-800 nm 

O(3P) + O2 + M → O3 + M

The temperature sensitivity of PAN allows it to be stable enough at low
temperatures to be transported long distances before decomposing to
release NO2.  NO2 can then participate in ozone formation in regions
remote from the original NOX source.  A discussion of CB05 mechanisms
for ozone formation can be found in Yarwood et al. (2005).

Another important way that ethanol fuels contribute to ozone formation
is by increasing the formation of new radicals through increases in
formaldehyde and acetaldehyde.  As shown in Section   REF _Ref306260314
\r  7.2.3.1 , the photolysis of both aldehydes results in two molecules
of either hydroperoxy radical or methylperoxy radical, both of which
oxidize NO to NO2 leading to ozone formation.  

Uncertainties Associated with Chemical Mechanisms

A key source of uncertainty with respect to the air quality modeling
results is the photochemical mechanisms in CMAQ 4.7.1.  Pollutants such
as ozone, PM, acetaldehyde, formaldehyde, acrolein, and 1,3-butadiene
can be formed secondarily through atmospheric chemical processes.  Since
secondarily formed pollutants can result from many different reaction
pathways, there are uncertainties associated with each pathway. 
Simplifications of chemistry must be made in order to handle reactions
of thousands of chemicals in the atmosphere.  Mechanisms for formation
of ozone, PM, acetaldehyde and peroxyacetyl nitrate (PAN) are discussed
in Section   REF _Ref305053973 \r  7.2.3 .  

For PM, there are a number of uncertainties associated with SOA
formation that should be addressed explicitly.  As mentioned in Section 
 REF _Ref305053973 \r  7.2.3 , a large number of VOCs emitted into the
atmosphere, which have the potential to form SOA, have not yet been
studied in detail.  In addition, the amount of ambient SOA that comes
from benzene is uncertain.  Simplifications to the SOA treatment in CMAQ
have also been made in order to preserve computational efficiency. 
These simplifications are described in release notes for CMAQ 4.7 on the
Community Modeling and Analysis System (CMAS) website.  

Impacts of the Proposed Rule on Air Quality

Air quality modeling performed for this proposed rule estimates the
changes in ambient concentrations of PM2.5, ozone and NO2, as well as
changes in ambient concentrations of ethanol and the following air
toxics: acetaldehyde, acrolein, benzene, 1,3-butadiene, and
formaldehyde.  The air quality modeling results also include changes in
deposition of nitrogen and sulfur and changes in visibility levels due
to this proposed rule.  

This section describes current ambient levels of the modeled pollutants
and presents the projected future ambient levels resulting from the
proposed rule.  

Ozone

As described in Section 6.2.1 of this draft RIA, ozone causes adverse
health effects, and the EPA has set national ambient air quality
standards (NAAQS) to protect against those health effects.  In this
section, we present information on current and model-projected future
ozone levels.

Current Levels of Ozone

  REF _Ref304990261 \h  Figure 7-12  shows a snapshot of measured ozone
concentrations in 2010. The highest ozone concentrations were located in
California.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  12  Ozone
Concentrations (fourth highest daily maximum 8-hour concentration) in
ppm for 2010

The primary and secondary NAAQS for ozone are 8-hour standards set at
0.075 ppm.  The most recent revision to the ozone standards was in 2008;
the previous 8-hour ozone standards, set in 1997, had been set at 0.08
ppm.  In 2004, the U.S. EPA designated nonattainment areas for the 1997
8-hour ozone NAAQS (69 FR 23858, April 30, 2004).  As of December 14,
2012, there were 41 8-hour ozone nonattainment areas for the 1997 ozone
NAAQS, composed of 221 full or partial counties, with a total population
of over 129 million.  Nonattainment areas for the 1997 8-hour ozone
NAAQS are pictured in   REF _Ref304990291 \h  Figure 7-13 . 
Nonattainment designations for the 2008 ozone standards were finalized
on April 30, 2012 and May 31, 2012.  These designations include 46
areas, composed of 227 full or partial counties, with a population of
over 123 million.  Nonattainment areas for the 2008 ozone NAAQS are
pictured in Figure 7-14.  As of December 14, 2012, over 138 million
people are living in ozone nonattainment areas.

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  13  1997 8-hour
Ozone Nonattainment Areas

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  14  2008 8-hour
Ozone Nonattainment Areas

States with ozone nonattainment areas are required to take action to
bring those areas into attainment in the future.  The attainment date
assigned to an ozone nonattainment area is based on the area’s
classification.  Most ozone nonattainment areas are required to attain
the 1997 8-hour ozone NAAQS in the 2007 to 2013 time frame and then to
maintain it thereafter.  The attainment dates for areas designated
nonattainment for the 2008 8-hour ozone NAAQS are in the 2015 to 2032
timeframe, depending on the severity of the problem in each area.  In
addition, EPA is working to complete the current review of the ozone
NAAQS by mid-2014.  If EPA revises the ozone standards in 2014 pursuant
to that review, the attainment dates associated with areas designated
nonattainment for that NAAQS would likely be in the 2019 to 2036
timeframe, depending on the severity of the problem in each area.

Projected Levels of Ozone Without this Proposed Rule

EPA has already adopted many mobile source emission control programs
that are expected to reduce ambient ozone levels.  These control
programs include the New Marine Compression-Ignition Engines at or Above
30 Liters per Cylinder Rule (75 FR 22895, April 30, 2010), the Marine
Spark-Ignition and Small Spark-Ignition Engine Rule (73 FR 59034,
October 8, 2008), the Locomotive and Marine Rule (73 FR 25098, May 6,
2008), the Clean Air Nonroad Diesel Rule (69 FR 38957, June 29, 2004),
the Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel
Sulfur Control Requirements (66 FR 5002, January 18, 2001) and the Tier
2 Motor Vehicle Emissions Standards and Gasoline Sulfur Control
Requirements (65 FR 6698, February 10, 2000).  As a result of these and
other federal, state and local programs, 8-hour ozone levels are
expected to improve in the future.  However, even with the
implementation of all current state and federal regulations, there are
projected to be counties violating the ozone NAAQS well into the future.
 Thus additional federal control programs, such as Tier 3, can assist
areas with attainment dates in 2017 and beyond in attaining the NAAQS as
expeditiously as practicable and may relieve areas with already
stringent local regulations from some of the burden associated with
adopting additional local controls.

The air quality modeling projects that in 2017, with all current
controls in effect but excluding the emissions changes expected to occur
as a result of this proposed action or any other additional controls, at
least 40 counties, with a projected population of almost 50 million
people, would have projected design values above the level of the 2008
8-hour ozone standard of 75 ppb.  Even in 2030 the modeling projects
there will be 12 counties with a population of almost 32 million people
that would have projected design values above the level of the 2008
8-hour ozone standard of 75 ppb without additional controls.  Since the
emission changes from this proposal go into effect during the period
when some areas are still working to attain the ozone NAAQS, the
projected emission changes will help state and local agencies in their
effort to attain and maintain the ozone standard.  In the following
section we discuss the projected ozone reductions associated with the
proposed standards.  

Projected Levels of Ozone With this Proposed Rule

This section summarizes the results of our modeling of ozone air quality
impacts in the future with the proposed standards.  Specifically, for
the years 2017 and 2030 we compare a reference scenario (a scenario
without the proposed standards) to a control scenario that includes the
proposed standards.  Our modeling indicates that ozone design value
concentrations will decrease dramatically in many areas of the country
as a result of this proposal and in some places those decreases will be
enough to move the projected design values from being above the NAAQS to
being below the NAAQS.  Additional information on the emissions
reductions that are projected with this final action is available in
Section 7.2.1 of this draft RIA.  

  REF _Ref308781205 \h  Figure 7-15  and   REF _Ref304990551 \h  Figure
7-16  present the changes in 8-hour ozone design value concentrations in
2017 and 2030 respectively.  Note that the projected results for 2017 do
not include California, while the projected results for 2030 do.  This
issue does not have a significant impact on the AQ modeling results for
the rest of the country. 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  15  Projected
Change in 2017 8-hour Ozone Design Values Between the Reference Case and
Control Case

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  16  Projected
Change in 2030 8-hour Ozone Design Values Between the Reference Case and
Control Case

As can be seen in   REF _Ref308781205 \h  Figure 7-15 , the majority of
the design value decreases in 2017 are between 0.5 and 1.0 ppb.  The
projected population-weighted average design value concentration without
the proposed rule is 71.3 ppb in 2017.  There are also 7 counties with
projected 8-hour ozone design value decreases of more than 1 ppb; these
counties are in Arizona, Texas and Tennessee.  The maximum projected
decrease in an 8-hour ozone design value in 2017 is 1.09 ppb in Tarrant
County, Texas near Dallas, which is projected to be above the ozone
standard.    REF _Ref304990551 \h  Figure 7-16  presents the ozone
design value changes for 2030.  The projected population-weighted
average design value concentration without the proposed rule is 66.7 ppb
in 2030.  In 2030 the ozone design value decreases are larger than in
2017; most decreases are projected to be between 1.0 and 1.5 ppb, and
over 200 counties have design values with projected decreases greater
than 1.5 ppb.  The maximum projected decrease in an 8-hour ozone design
value in 2030 is 3.2 ppb in Maricopa County, Arizona, where Phoenix is
located.  

  REF _Ref304990674 \h  Table 7-30  and   REF _Ref308342119 \h  Table
7-31  show the average change, due to this proposed rule, in 2017 and
2030 8-hour ozone design values for: (1) all counties with 2005 baseline
design values, (2) counties with 2005 baseline design values that
exceeded the 2008 ozone standard, (3) counties with 2005 baseline design
values that did not exceed the 2008 standard, but were within 10 percent
of it, (4) counties with 2017/2030 design values that exceeded the 2008
ozone standard, and (5) counties with 2017/2030 design values that did
not exceed the standard, but were within 10 percent of it.  Counties
within 10 percent of the standard are intended to reflect counties that
although not violating the standards, will also be impacted by changes
in ozone as they work to ensure long-term maintenance of the ozone
NAAQS.  All of these metrics show a decrease in 2017 and 2030,
indicating in five different ways the overall improvement in air
quality.  

On a population-weighted basis, the average modeled future-year 8-hour
ozone design values are projected to decrease by 0.47 ppb in 2017 and
1.55 ppb in 2030.  On a population-weighted basis, design values in
those counties that are projected to be above the 2008 ozone standard in
2017 and 2030 are projected to decrease by 0.30 and 1.62 ppb
respectively due to the proposed standards.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  30  Average Change
in Projected 8-hour Ozone Design Value in 2017 

Averagea	Number of U.S. Counties	2020 Population	Change in 2017 design
value (ppb)

All	676

	238,026,106

	-0.50

All, population-weighted

	-0.47

Counties whose 2005 base year is violating the 2008 8-hour ozone
standard	393

	176,910,535

	-0.56

Counties whose 2005 base year is violating the 2008 8-hour ozone
standard, population-weighted

	-0.51

Counties whose 2005 base year is within 10 percent of the 2008 8-hour
ozone standard	201

	40,516,171

	-0.47

Counties whose 2005 base year is within 10 percent of the 2008 8-hour
ozone standard, population-weighted

	-0.42

Counties whose 2017 control case is violating the 2008 8-hour ozone
standard	37

	47,659,433

	-0.35

Counties whose 2017 control case is violating the 2008 8-hour ozone
standard, population-weighted

	-0.30

Counties whose 2017 control case is within 10 percent of the 2008 8-hour
ozone standard	124

	68,625,934

	-0.51

Counties whose 2017 control case is within 10 percent of the 2008 8-hour
ozone standard, population-weighted

	-0.49

Notes:

a Averages are over counties with 2005 modeled design values. 

b Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  31  Average Change
in Projected 8-hour Ozone Design Value in 2030 

Averagea	Number of U.S. Counties	2030 Population	Change in 2030 design
value (ppb)

All	676

	261,497,900

	-1.35

All, population-weighted

	-1.55

Counties whose 2005 base year is violating the 2008 8-hour ozone
standard	393

	194,118,748

	-1.54

Counties whose 2005 base year is violating the 2008 8-hour ozone
standard, population-weighted

	-1.69

Counties whose 2005 base year is within 10 percent of the 2008 8-hour
ozone standard	201

	44,436,103

	-1.18

Counties whose 2005 base year is within 10 percent of the 2008 8-hour
ozone standard, population-weighted

	-1.25

Counties whose 2030 control case is violating the 2008 8-hour ozone
standard	10

	30,619,714

	-1.49

Counties whose 2030 control case is violating the 2008 8-hour ozone
standard, population-weighted

	-1.62

Counties whose 2030 control case is within 10 percent of the 2008 8-hour
ozone standard	40

	21,541,863

	-1.37

Counties whose 2030 control case is within 10 percent of the 2008 8-hour
ozone standard, population-weighted

	-1.50

Notes:

a Averages are over counties with 2005 modeled design values 

b Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD.

There are still 12 counties, most of them in California, that are
projected to have 8-hour ozone design values above the 2008 NAAQS in
2030 without the proposed standards or any other additional controls in
place.    REF _Ref308781709 \h  Table 7-32  below presents the changes
in design values for these counties.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  32  Change in
Ozone Design Values (ppb) for Counties Projected to be Above the 2008
Ozone NAAQS in 2030

County Name	Change in 8-hour Ozone Design Value (ppb)	Population in
2030a

San Bernardino, California	-1.91	2,784,490

Riverside, California	-1.68	2,614,198

Los Angeles, California	-1.68	10,742,722

Kern, California	-1.00	981,806

Orange, California	-1.13	4,431,071

Harris, Texas	-2.05	5,268,889

Tulare, California	-1.00	528,663

Suffolk, New York	-1.33	1,705,822

Fresno, California	-1.15	1,196,950

Harford, Maryland	-1.91	365,103

Brazoria, Texas	-1.96	364,257

Hudson, New Jersey	-0.88	747,407

Note:

a Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD.

The proposed rule would also reduce ozone design values in some counties
from above the level of the standard to below it.  In 2017, ozone design
values in three counties (Bucks County in Pennsylvania, Arlington County
in Virginia and St Louis County in Missouri) move from being above the
standard to below.  The projected population in these three counties in
2017 is almost 2 million people. There are two more counties whose
design values would be reduced from above the standard to below by the
proposed rule in 2030.  These counties are Hudson County in New Jersey
and Brazoria County in Texas.  The projected population in these two
counties in 2030 is over 1 million people.     

In terms of modeling accuracy, the count of modeled nonattainment
counties is much less certain than the average changes in air quality.
For example, actions by states to meet their SIP obligations would not
be expected to significantly change the overall concentration changes
induced by this final rule, but they could substantially change the
number of counties in or out of attainment. If state actions resulted in
an increase in the number of areas that are very close to, but still
above, the NAAQS, then this rule might bring many of those counties down
sufficiently to change their attainment status. On the other hand, if
future state actions brought several counties we project to be very
close to the standard into attainment status, then the air quality
improvements from this rule might change the actual attainment status of
very few counties. Bearing this limitation in mind, our modeling
indicates that the emission reductions from this proposed rule will
decrease the number of ozone nonattainment counties by 3 in 2017 and by
2 in 2030, without consideration of new state or local programs.

As described in Section 6.1.2.1 of this draft RIA the science of ozone
formation, transport, and accumulation is complex.  The air quality
modeling projects ozone decreases as a result of emissions changes from
the proposed fuel and vehicle standards.  This change in ozone results
from interactions between photochemistry, background concentrations of
VOC and NOX, local emissions and meteorology.  

There is one county in 2017 that is projected to have an increase in
modeled ozone design value concentration (Multnomah County, Oregon,
where Portland is located).  When NOX levels are relatively high and VOC
levels relatively low, NOX forms inorganic nitrates (i.e., particles)
but relatively little ozone.  Such conditions are called
“NOX-saturated.”  Under these conditions, VOC reductions are
effective in reducing ozone, but NOX reductions can actually increase
local ozone under certain circumstances.  We believe that this is the
case in Multnomah County, Oregon in 2017.  In 2030, when the fleet would
be composed of vehicles meeting the new standards and the NOX and VOC
emissions reductions are larger, this ozone disbenefit is eliminated,
and the design values for all the modeled counties are decreasing.

Particulate Matter

As described in Section 6.2.2 of this draft RIA, PM causes adverse
health effects, and the EPA has set national ambient air quality
standards (NAAQS) to protect against those health effects.  In this
section we present information on current and model-projected future PM
levels.

Current Levels of PM

  REF _Ref304987760 \h  Figure 7-17  and   REF _Ref308782074 \h  Figure
7-18  respectively show a snapshot of annual and 24-hour PM2.5
concentrations in 2010.  In 2010, the highest annual average PM2.5
concentrations were in California, Indiana, Pennsylvania and Hawaii and
the highest 24-hour PM2.5 concentrations were in California and Alaska.

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  17  Annual
Average PM2.5 Concentrations in µg/m3 for 2010 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  18   24-hour
(98th percentile 24- hour concentrations) PM2.5 Concentrations in µg/m3
for 2010

.0 μg/m3) and a 24-hour standard (35 μg/m3).  The most recent
revisions to these standards were in 1997, 2006 and in December 2012. 
The December 2012 rule revised the level of the annual PM2.5 standard
from 15 μg/m3 to 12 μg/m3.  

In 2005 the U.S. EPA designated nonattainment areas for the 1997 PM2.5
NAAQS (70 FR 19844, April 14, 2005).  As of December 14, 2012, over 91
million people lived in the 35 areas that are designated as
nonattainment for the 1997 PM2.5 NAAQS.  These PM2.5 nonattainment areas
are comprised of 191 full or partial counties.  Nonattainment areas for
the 1997 PM2.5 NAAQS are pictured in Figure 7 19.  EPA anticipates
making attainment/nonattainment designations for the 2012 PM2.5 NAAQS by
December 2014, with those designations likely becoming effective in
early 2015.   On October 8, 2009, the EPA issued final nonattainment
area designations for the 2006 24-hour PM2.5 NAAQS (74 FR 58688,
November 13, 2009).  These designations include 32 areas composed of 121
full or partial counties, with a population of over 70 million. 
Nonattainment areas for the 2006 PM2.5 NAAQS are pictured in Figure 7
20.  In total, there are 50 PM2.5 nonattainment areas with a population
of over 105 million people.   

States with PM2.5 nonattainment areas will be required to take action to
bring those areas into attainment in the future.  Most 1997 PM2.5
nonattainment areas are required to attain the 1997 PM2.5 NAAQS in the
2009 to 2014 time frame and then required to maintain the 1997 PM2.5
NAAQS thereafter.  The 2006 24-hour PM2.5 nonattainment areas will be
required to attain the 2006 24-hour PM2.5 NAAQS in the 2015 to 2019 time
frame and then be required to maintain the 2006 24-hour PM2.5 NAAQS
thereafter.  2012 PM2.5 nonattainment areas will likely be required to
attain the 2012 PM2.5 NAAQS in the 2021 to 2025 time frame, depending on
the severity of an area’s fine particle pollution problems and the
availability of pollution controls.

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  19  1997 PM2.5
Nonattainment Areas

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  20  2006 PM2.5
Nonattainment Areas

As of December 14, 2012, over 29 million people live in the 46 areas
that are designated as nonattainment for the PM10 NAAQS.  There are 39
full or partial counties that make up the PM10 nonattainment areas. 
Nonattainment areas for the PM10 NAAQS are pictured in   REF
_Ref304988152 \h  Figure 7-21 .

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  21  PM10
Nonattainment Areas

Projected Levels of PM2.5 Without this Proposed Rule

EPA has already adopted many mobile source emission control programs
that are expected to reduce ambient PM levels.  These control programs
include the Heavy-Duty Greenhouse Gas Rule (76 FR 57106, September 15,
2011), the New Marine Compression-Ignition Engines at or Above 30 Liters
per Cylinder Rule (75 FR 22895, April 30, 2010), the Marine
Spark-Ignition and Small Spark-Ignition Engine Rule (73 FR 59034,
October 8, 2008), the Locomotive and Marine Compression-Ignition Engine
Rule (73 FR 25098, May 6, 2008), the Clean Air Nonroad Diesel (69 FR
38957, June 29, 2004), the Heavy-Duty Engine and Vehicle Standards and
Highway Diesel Fuel Sulfur Control Requirements (66 FR 5002, January 18,
2001) and the Tier 2 Motor Vehicle Emissions Standards and Gasoline
Sulfur Control Requirements (65 FR 6698, February 10, 2000).  As a
result of these and other federal, state and local programs, the number
of areas that fail to meet the PM2.5 NAAQS in the future is expected to
decrease.  However, even with the implementation of all current state
and federal regulations, there are projected to be counties violating
the PM2.5 NAAQS well into the future.  Thus additional federal control
programs, such as Tier 3, can assist areas with attainment dates in 2017
and beyond in attaining the NAAQS as expeditiously as practicable and
may relieve areas with already stringent local regulations from some of
the burden associated with adopting additional local controls.  

The air quality modeling conducted for this proposal projects that in
2030, with all current controls in effect but excluding the emissions
changes expected to occur as a result of this proposal or any other
additional controls, at least 14 counties, with a projected population
of over 28 million people, would have projected design values above the
level of the annual standard of 12.0 µg/m3 and at least 21 counties,
with a projected population of over 31 million people, would have
projected design values above the level of the 2006 24-hour standard of
35 µg/m3.  Since the emission changes from this proposed action would
go into effect during the period when some areas are still working to
attain the PM2.5 NAAQS, the projected emission changes will help state
and local agencies in their effort to attain and maintain the PM2.5
standard.  In the following section we discuss projected PM2.5
reductions from these proposed standards.  

Projected Annual Average Levels of PM2.5 With this Proposed Rule

This section summarizes the results of our modeling of annual average
PM2.5 air quality impacts in the future due to the standards proposed in
this action.  Specifically, for the years 2017 and 2030 we compare a
reference scenario (a scenario without the proposed standards) to a
control scenario that includes the proposed standards.  Our modeling
indicates that by 2030 annual PM2.5 design values in the majority of the
modeled counties would decrease due to the proposed standards.  The
decreases in annual PM2.5 design values are likely due to the projected
reductions in primary PM2.5, NOX, SOX and VOC emissions.  Additional
information on the emissions reductions that are projected with this
proposed action is available in Section   REF _Ref305059196 \n \h  7.2.1
 of this draft RIA.

It is important to note that, as discussed in Section 7.1.5 and 7.2.1.1,
the control scenario emissions inventory prepared for air quality
modeling included direct PM2.5 vehicle emissions increases that we do
not expect to occur in reality.  These increases resulted from a series
of conservative assumptions and uncertainties related to fuel parameters
in 2017, and also an emissions processing issue which erroneously
increased direct PM emissions in about one third of modeled counties
(see Section   REF _Ref310433834 \w \h  7.2.1.1.2  for more detail). 
Because our air quality modeling assumes this increase, our air quality
results overestimate ambient PM and underestimate the reductions that
would result from the proposed Tier 3 standards.

    REF _Ref304988783 \h  Figure 7-22  and   REF _Ref304988791 \h 
Figure 7-23  present the changes in annual PM2.5 design values in 2017
and 2030 respectively.  Note that the projected results for 2017 do not
include California, while the projected results for 2030 do.   This
issue does not have a significant impact on the AQ modeling results for
the rest of the country. 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  22  Projected
Change in 2017 Annual PM2.5 Design Values Between the Reference Case and
Control Case

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  23  Projected
Change in 2030 Annual PM2.5 Design Values Between the Reference Case and
Control Case

The projected population-weighted average design value concentration
without the proposed rule is 9.3 µg/m3 in 2017.  As shown in   REF
_Ref304988783 \h  \* MERGEFORMAT  Figure 7-22 , we project that in 2017
seven counties will have design value decreases of between 0.01 µg/m3
and 0.05 µg/m3.  These counties are in Utah, Pennsylvania and
Wisconsin.  The maximum projected decrease in a 2017 annual PM2.5 design
value is 0.03 µg/m3 in Weber County, Utah.  As mentioned above the
decreases in ambient annual PM2.5 concentrations are due to reductions
in NOX, SOX and VOCs and the subsequent reductions in secondarily formed
PM due to this proposed rule in 2017, which offset the small increases
in direct PM emissions that were modeled but we do not expect to occur
(see Section   REF _Ref309220854 \r \h  7.1.5  and Section   REF
_Ref310433834 \w \h  7.2.1.1.2  of this draft RIA for more detail).  As
a result, the projected decreases in design values are underestimates of
the actual effects of the proposed rule.  There are a few counties with
projected small increases in annual PM2.5 in 2017, but as explained, we
do not expect that these localized small increases will actually happen.


The projected population-weighted average design value concentration
without the proposed rule is 9.5 µg/m3 in 2030.    REF _Ref304988791 \h
 Figure 7-23  presents the annual PM2.5 design value changes in 2030. 
In 2030 all the modeled counties have decreases in annual PM2.5 design
values.  The annual PM2.5 design value decreases in 2030 are larger than
the decreases in 2017; most design values are projected to decrease
between 0.01 and 0.05 µg/m3 and over 100 additional counties have
projected design value decreases greater than 0.05 µg/m3.  The maximum
projected decrease in an annual PM2.5 design value in 2030 is 0.20
µg/m3 in Tulare County, California.  

   REF _Ref349804941  Table 7-33  and   REF _Ref309292824 \h  Table 7-34
 show the average change in 2017 and 2030 annual PM2.5 design values
for: (1) all counties with 2005 baseline design values, (2) counties
with 2005 baseline design values that exceeded the 2012 annual PM2.5
standard, (3) counties with 2005 baseline design values that did not
exceed the 2012 standard, but were within 10 percent of it, (4) counties
with 2017/2030 design values that exceeded the 2012 annual PM2.5
standard, and (5) counties with 2017/2030 design values that did not
exceed the standard, but were within 10 percent of it.  Counties within
10 percent of the standard are intended to reflect counties that
although not violating the standards, will also be impacted by changes
in PM2.5 as they work to ensure long-term maintenance of the annual
PM2.5 NAAQS.  All of these metrics show either no change or a small
increase or decrease in 2017 and show a decrease in 2030; these results
are underestimates for the reasons explained above.  The rows in   REF
_Ref349804941  Table 7-33  that compare the 2017 control case to the
standard do not include California as we were not able to model design
value changes in CA due to an inventory error.  On a population-weighted
basis, there is no change (0.00 µg/m3) in the average modeled
future-year annual PM2.5 design values in 2017 and a 0.06 µg/m3
decrease in 2030.  On a population-weighted basis design values in those
counties that are projected to be above the annual PM2.5 standard in
2030 would decrease by 0.11 µg/m3 due to the proposed standards.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  33  Average Change
in 2017 Annual PM2.5 Design Value as a Result of the Proposed Rule

AVERAGE		Number of U.S. Counties	2020 Population	Change in 2017 design
value (µg/m3) 

All	576	226,563,544	0.00

All, population-weighted

	0.00

Counties whose 2005 base year is violating the annual PM2.5 standard	314
140,706,274	0.00

Counties whose 2005 base year is violating the annual PM2.5  standard,
population-weighted

	0.00

Counties whose 2005 base year is within 10 percent of the annual PM2.5 
standard	83	29,253,128	-0.01

Counties whose 2005 base year is within 10 percent of the annual PM2.5 
standard, population-weighted

	0.00

Counties whose 2017 control case is violating the annual PM2.5  standard
	4	1,992,506	0.01

Counties whose 2017 control case is violating the annual PM2.5 
standard, population-weighted

	0.01

Counties whose 2017 control case is within 10 percent of the annual
PM2.5  standard	13	12,158,988	-0.01

Counties whose 2017 control case is within 10 percent of the annual
PM2.5  standard, population-weighted

	-0.01

Notes:

a Averages are over counties with 2005 modeled design values 

b Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  34  Average Change
in 2030 Annual PM2.5 Design Value as a Result of the Proposed Rule

AVERAGE		Number of U.S. Counties	2030 Population	Change in 2030 design
value (µg/m3) 

All	576	    247,415,381 	-0.05

All, population-weighted

	-0.06

Counties whose 2005 base year is violating the annual PM2.5 standard	314
      152,109,569 	-0.05

Counties whose 2005 base year is violating the annual PM2.5  standard,
population-weighted

	-0.07

Counties whose 2005 base year is within 10 percent of the annual PM2.5 
standard	83	      31,863,376 	-0.05

Counties whose 2005 base year is within 10 percent of the annual PM2.5 
standard, population-weighted

	-0.05

Counties whose 2030 control case is violating the annual PM2.5  standard
	14	      28,624,758 	-0.11

Counties whose 2030 control case is violating the annual PM2.5 
standard, population-weighted

	-0.10

Counties whose 2030 control case is within 10 percent of the annual
PM2.5  standard	28	      23,840,272 	-0.07

Counties whose 2030 control case is within 10 percent of the annual
PM2.5  standard, population-weighted

	-0.09

Notes:

a Averages are over counties with 2005 modeled design values 

b Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD.

There are fourteen counties in California that are projected to have
annual PM2.5 design values above the NAAQS in 2030 without the proposed
standards or any other additional standards in place.    REF
_Ref309305355 \h  Table 7-35  below presents the changes in design
values for these counties.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  35  Change in
Annual PM2.5 Design Values (µg/m3) for Counties Projected to be Above
the Annual PM2.5 NAAQS in 2030

County Name	Change in Annual PM2.5 Design Value (µg/m3)	Population in
2030a

Riverside County, California	-0.14	2,614,198 

San Bernardino County, California	-0.11	2,784,489 

Kern County, California	-0.19	981,806 

Tulare County, California	-0.2	528,662 

Fresno County, California	-0.16	1,196,949 

Kings County, California	-0.17	195,067 

Los Angeles County, California	-0.09	10,742,722 

Allegheny County, Pennsylvania	-0.05	1,234,930 

Lincoln County, Montana	-0.02	20,454 

Jefferson County, Alabama	-0.05	697,239 

Wayne County, Michigan	-0.09	1,838,269 

Santa Cruz County, Arizona	-0.03	55,393 

Orange County, California	-0.06	4,431,070 

 Note:

a Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD.

Projected 24-hour Average Levels of PM2.5 With this Proposed Rule

This section summarizes the results of our modeling of 24-hour PM2.5 air
quality impacts in the future due to the proposed rule.  Specifically,
for the years 2017 and 2030 we compare a reference scenario (a scenario
without the proposed standards) to a control scenario that includes the
proposed standards.  Our modeling indicates that by 2030 24-hour PM2.5
design values in the majority of the modeled counties would decrease due
to the proposed standards.  The decreases in 24-hour PM2.5 design values
are likely due to the projected reductions in primary PM2.5, NOX, SOX
and VOCs.  Additional information on the emissions reductions that are
projected with this proposed action is available in Section   REF
_Ref305059196 \n \h  \* MERGEFORMAT  7.2.1  of this draft RIA. 

It is important to note that, as discussed in Section 7.1.5 and 7.2.1.1,
the control scenario emissions inventory prepared for air quality
modeling included direct PM2.5 vehicle emissions increases that we do
not expect to occur in reality.  These increases resulted from a series
of conservative assumptions and uncertainties related to fuel parameters
in 2017, and also an emissions processing issue which erroneously
increased direct PM emissions in about one third of modeled counties
(see Section   REF _Ref310433834 \w \h  7.2.1.1.2  for more detail). 
Because our air quality modeling assumes this increase, our air quality
results overestimate ambient PM and underestimate the reductions that
would result from the proposed Tier 3 standards.

   REF _Ref308691780 \h  Figure 7-24  and   REF _Ref308691958 \h  Figure
7-25  present the changes in 24-hour PM2.5 design values in 2017 and
2030 respectively.  Note that the projected results for 2017 do not
include California, while the projected results for 2030 do.   This
issue does not have a significant impact on the AQ modeling results for
the rest of the country.

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  24  Projected
Change in 2017 24-hour PM2.5 Design Values Between the Reference Case
and the Control Case

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  25  Projected
Change in 2030 24-hour PM2.5 Design Values Between the Reference Case
and the Control Case

The projected population-weighted average design value concentration
without the proposed rule is 23.4 µg/m3 in 2017.  As shown in   REF
_Ref304989753 \h  Figure 7-24 , in 2017 there are 72 counties with
projected 24-hour PM2.5 design value decreases greater than 0.05 µg/m3.
 These counties are in Utah, Pennsylvania and scattered throughout the
Midwest.  The maximum projected decrease in a 2017 24-hour PM2.5 design
value is 0.20 µg/m3 in Tooele County, Utah.  As mentioned above, the
decreases in ambient annual PM2.5 concentrations are due to reductions
in NOX, SOX and VOCs and the subsequent reductions in secondarily formed
PM due to this proposed rule in 2017, which offset the small increases
in direct PM emissions that were modeled but we do not expect to occur
(see Section   REF _Ref309220854 \r \h  7.1.5  and Section   REF
_Ref310433834 \w \h  7.2.1.1.2  of this draft RIA for more detail).  As
a result, the projected decreases in design values are underestimates of
the actual effects of the proposed rule.  There are some counties with
projected small increases in 24-hour PM2.5 in 2017, but as explained, we
do not expect that these localized small increases will actually happen.


The projected population-weighted average design value concentration
without the proposed rule is 24.3 µg/m3 in 2030.    REF _Ref308691958
\h  Figure 7-25  presents the 24-hour PM2.5 design value changes in
2030.  In 2030 the 24-hour PM2.5 design value decreases are larger; most
design values are projected to decrease between 0.05 and 0.15 µg/m3 and
almost 200 counties have projected design value decreases greater than
0.15 µg/m3.  The maximum projected decrease in a 24-hour PM2.5 design
value in 2030 is 0.81 µg/m3 in Cache County, Utah.  As shown in   REF
_Ref308691958 \h  \* MERGEFORMAT  Figure 7-25 , design values in 68
counties would decrease by more than 0.25 µg/m3.  These counties are in
Idaho, Nevada, Montana, Louisiana, northern Utah, and the upper Midwest.
 The decreases in 24-hour PM2.5 design values that are projected in some
counties are likely due to emission reductions related to reductions in
PM2.5 precursor emissions (NOX, SOX, and VOCs).  There is one county,
Richmond County, Georgia, with a projected 24-hour PM2.5 design value
increase of less than 0.15 µg/m3.  Additional information on the
emissions reductions that are projected with this proposed action is
available in Section   REF _Ref305059196 \n \h  7.2.1 .   

  REF _Ref304990050 \h  Table 7-36  shows the average change in 2030
24-hour PM2.5 design values for: (1) all counties with 2005 baseline
design values, (2) counties with 2005 baseline design values that
exceeded the 24-hour PM2.5 standard, (3) counties with 2005 baseline
design values that did not exceed the standard, but were within 10
percent of it, (4) counties with 2030 design values that exceeded the
24-hour PM2.5 standard, and (5) counties with 2030 design values that
did not exceed the standard, but were within 10 percent of it.  Counties
within 10 percent of the standard are intended to reflect counties that
although not violating the standards, will also be impacted by changes
in PM2.5 as they work to ensure long-term maintenance of the 24-hour
PM2.5 NAAQS.  On a population-weighted basis, the average modeled
future-year 24-hour PM2.5 design values are projected to decrease by
0.14 µg/m3 due to the proposed standards.  On a population-weighted
basis, 24-hour PM2.5 design values in those counties that are projected
to be above the 24-hour PM2.5 standard in 2030 would decrease by 0.13
µg/m3.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  36  Average Change
in Projected 24-hour PM2.5 Design Values in 2030

Averagea	Number of U.S. Counties	2030 Populationb	Change in 2030 design
value (µg/m3)

All	569	245,111,480	-0.13

All, population-weighted

	-0.14

Counties whose 2005 base year is violating the 2006 24-hour PM2.5
standard	108	91,474,036	-0.19

Counties whose 2005 base year is violating the 2006 24-hour PM2.5
standard, population-weighted

	-0.15

Counties whose 2005 base year is within 10 percent of the 2006 24-hour
PM2.5 standard	140	53,990,060	-0.16

Counties whose 2005 base year is within 10 percent of the 2006 24-hour
PM2.5 standard, population-weighted

	-0.17

Counties whose 2030 control case is violating the 2006 24-hour PM2.5
standard 	21	31,002,272	-0.13

Counties whose 2030 control case is violating the 2006 24-hour PM2.5
standard, population-weighted

	-0.05

Counties whose 2030 control case is within 10 percent of the 2006
24-hour PM2.5 standard	7	4,212,913	-0.12

Counties whose 2030 control case is within 10 percent of the 2006
24-hour PM2.5 standard, population-weighted

	-0.10

Note:

a Averages are over counties with 2005 modeled design values 

b Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD.

There are 21 counties that are projected to have 24-hour PM2.5 design
values above the NAAQS in 2030 without the proposed standards or any
other additional controls in place.    REF _Ref309300916 \h  Table 7-37 
below presents the changes in design values for these counties.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  37  Change in
24-hour PM2.5 Design Values (µg/m3) for Counties Projected to be Above
the 24-hour PM2.5 NAAQS in 2030

County Name	Change in 24-hour PM2.5 Design Value (µg/m3)	Population in
2030a

Riverside County, California	0.00	2,614,198

Kern County, California	0.00	981,806

San Bernardino County, California	0.00	2,784,490

Fresno County, California	0.00	1,196,950

Sacramento County, California	0.00	1,856,971

Cache County, Utah	-0.81	141,446

Kings County, California	0.00	195,067

Lane County, Oregon	-0.12	460,993

Los Angeles County, California	0.00	10,742,722

Allegheny County, Pennsylvania	-0.11	1,234,931

Tulare County, California	0.00	528,663

Lake County, Montana	-0.07	40,126

Stanislaus County, California	0.00	688,246

Ravalli County, Montana	-0.19	63,914

Missoula County, Montana	-0.27	141,264

Klamath County, Oregon	-0.13	77,200

Salt Lake County, Utah	-0.59	1,431,946

Lincoln County, Montana	-0.12	20,454

Butte County, California	0.00	287,236

Pierce County, Washington	-0.24	1,082,579

Orange County, California	0.00	4,431,071

Note:

a Population numbers based on Woods & Poole data.  Woods & Poole
Economics, Inc. 2001.  Population by Single Year of Age CD. 

Nitrogen Dioxide

As described in Section 6.2.3 of this draft RIA, NO2 causes adverse
health effects, and the EPA has set national ambient air quality
standards (NAAQS) to protect against those health effects.  In this
section we present information on current and model-projected future NO2
levels.

Current Levels of NO2

Between 2003 and 2005, national mean concentrations of NO2 were about 15
ppb for averaging periods ranging from a day to a year.  There are two
NAAQS for NO2: an annual standard (53 ppb) and a 1-hour standard (100
ppb).  The primary NAAQS for NO2 was revised in January 2010. EPA
completed area designations in January 2012 and there are currently no
nonattainment areas.  The designations were based on the existing
community-wide monitoring network.  Once the expanded network of NO2
monitors is fully deployed and three years of air quality data have been
collected, EPA intends to redesignate areas in 2016 or 2017, as
appropriate, based on the air quality data from the new monitoring
network.,

Projected Levels of NO2 without this Proposed Rule

EPA has already adopted many mobile source emission control programs
that are expected to reduce ambient NO2 levels.  These control programs
include the Heavy-Duty Greenhouse Gas Rule (76 FR 57106, September 15,
2011), New Marine Compression-Ignition Engines at or Above 30 Liters per
Cylinder Rule (75 FR 22895, April 30, 2010), the Locomotive and Marine
Compression-Ignition Engine Rule (73 FR 25098, May 6, 2008), the Clean
Air Nonroad Diesel (69 FR 38957, June 29, 2004), the Heavy-Duty Engine
and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements (66 FR 5002, January 18, 2001) and the Tier 2 Motor Vehicle
Emissions Standards and Gasoline Sulfur Control Requirements (65 FR
6698, February 10, 2000).  As a result of these and other federal, state
and local programs, ambient concentrations of NO2 in the future are
expected to decrease.  However, even with the implementation of all
current state and federal regulations, there are projected to be parts
of the country with elevated NO2 concentrations in the future.  

As mentioned in Section 7.2.4.3.1, nonattainment designations for the
NO2 NAAQS will potentially be occurring in 2016 or 2017.  The emission
changes from this proposal would go into effect during the period when
any NO2 nonattainment areas would be working to attain the NO2 NAAQS.  

Projected Levels of NO2 with this Proposed Rule

This section summarizes the results of our modeling of annual average
NO2 air quality impacts in the future due to the proposed standards. 
Specifically, for the years 2017 and 2030 we compare a reference
scenario (a scenario without the proposed standards) to a control
scenario that includes the proposed standards.    REF _Ref312064796 \h 
Figure 7-26  and   REF _Ref310422960 \h  Figure 7-27  present the
changes in annual NO2 concentrations in 2017 and 2030 respectively.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  26  Projected
Change in 2017 Annual NO2 Concentrations Between the Reference Case and
Control Case

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  27  Projected
Change in 2030 Annual NO2 Concentrations Between the Reference Case and
Control Case

As shown in   REF _Ref310422960 \h  \* MERGEFORMAT  Figure 7-27 , our
modeling indicates that by 2030 annual NO2 concentrations in the
majority of the country would decrease less than 0.1 ppb due to this
proposal.  However decreases in annual NO2 concentrations are greater
than 0.3 ppb in most urban areas.  These emissions reductions would also
likely decrease 1-hour NO2 concentrations and help any potential
nonattainment areas to attain and maintain the standard.  Note that the
projected results for 2017 do not include California, while the
projected results for 2030 do.   This issue does not have a significant
impact on the AQ modeling results for the rest of the country.  

Air Toxics

As described in Section 6.2.5 of this draft RIA, air toxics cause
adverse health effects.  In this section we present information on
current and model-projected future levels of air toxics.

Current Levels of Air Toxics

The majority of Americans continue to be exposed to ambient
concentrations of air toxics at levels which have the potential to cause
adverse health effects.  The levels of air toxics to which people are
exposed vary depending on where people live and work and the kinds of
activities in which they engage, as discussed in detail in U.S. EPA’s
most recent Mobile Source Air Toxics (MSAT) Rule.  In order to identify
and prioritize air toxics, emission source types and locations which are
of greatest potential concern, U. S. EPA conducts the National-Scale Air
Toxics Assessment (NATA).  The most recent NATA was conducted for
calendar year 2005, and was released in March 2011.  NATA for 2005
includes four steps:

1) 	Compiling a national emissions inventory of air toxics emissions
from outdoor sources 

2)  Estimating ambient concentrations of air toxics across the United
States 

3)  Estimating population exposures across the United States 

4) 	Characterizing potential public health risk due to inhalation of air
toxics including both cancer and noncancer effects

  REF _Ref309626693 \h  Figure 7-28  and   REF _Ref309626702 \h  \*
MERGEFORMAT  Figure 7-29  depict estimated tract-level carcinogenic risk
and noncancer respiratory hazard from the assessment.  The respiratory
hazard is dominated by a single pollutant, acrolein. 

According to the NATA for 2005, mobile sources were responsible for 43
percent of outdoor toxic emissions and over 50 percent of the cancer
risk and noncancer hazard attributable to direct emissions from mobile
and stationary sources.,,  Mobile sources are also large contributors to
precursor emissions which react to form secondary concentrations of air
toxics. Formaldehyde is the largest contributor to cancer risk of all 80
pollutants quantitatively assessed in the 2005 NATA, and mobile sources
were responsible for over 40 percent of primary emissions of this
pollutant in 2005, and are major contributors to formaldehyde precursor
emissions.  Benzene is also a large contributor to cancer risk, and
mobile sources account for over 70 percent of ambient exposure.  Over
the years, EPA has implemented a number of mobile source and fuel
controls which have resulted in VOC reductions, which also reduced
formaldehyde, benzene and other air toxic emissions.  

 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  28  Tract Level
Average Carcinogenic Risk, 2005 NATA

 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  29  County Level
Average Noncancer Hazard Index, 2005 NATA

Projected Levels of Air Toxics

In the following sections, we describe results of our modeling of air
toxics levels in the future with the proposed standards.  Although there
are a large number of compounds which are considered air toxics, we
focused on those which were identified as national and regional-scale
cancer and noncancer risk drivers in past NATA assessments and were also
likely to be significantly impacted by the standards.  These compounds
include benzene, 1,3-butadiene, formaldehyde, acetaldehyde, and
acrolein.  Impacts on ethanol concentrations were also included in our
analyses.  Information on the air quality modeling methodology is
contained in Section 7.2.2.  Additional detail can be found in the air
quality modeling technical support document (AQM TSD) in the docket for
this rule.  Additional maps, including the seasonal concentration maps
for 2030 and 2017, are included in Appendix 7.A.

It should be noted that EPA has adopted many mobile source emission
control programs that are expected to reduce ambient air toxics levels. 
These control programs include the Heavy-duty Onboard Diagnostic Rule
(74 FR 8310, February 24, 2009), Small SI and Marine SI Engine Rule (73
FR 59034, October 8, 2008), Locomotive and Commercial Marine Rule (73 FR
25098, May 6, 2008), Mobile Source Air Toxics Rule (72 FR 8428, February
26, 2007), Clean Air Nonroad Diesel Rule (69 FR 38957, June 29, 2004),
Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur
Control Requirements (66 FR 5002, Jan. 18, 2001) and the Tier 2 Motor
Vehicle Emissions Standards and Gasoline Sulfur Control Requirements (65
FR 6698, Feb. 10, 2000).  As a result of these programs, the ambient
concentration of air toxics in the future is expected to decrease.  The
reference case and control case scenarios include these controls.  

Our modeling indicates that the impacts of the proposed standards
include generally small decreases in ambient concentrations of air
toxics, with the greatest reductions in urban areas.  Air toxics
pollutants dominated by primary emissions (or a decay product of a
directly emitted pollutant), such as benzene and 1,3-butadiene, have the
largest impacts.  Air toxics that primarily result from photochemical
transformation, such as formaldehyde and acetaldehyde, are not impacted
as much as those dominated by direct emissions.  Our modeling shows
decreases in ambient air toxics concentrations for both 2017 and 2030. 
Reductions are greater in 2030, when Tier 3 cars and trucks would
contribute nearly 90 percent of fleet-wide vehicle miles travelled, than
in 2017, which is the first year of the proposed program.  However, our
modeling projects there would be small immediate reductions in ambient
concentrations of air toxics due to the proposed sulfur controls in
2017.  Furthermore, the full reduction of the vehicle program would be
realized after 2030, when the fleet has fully turned over to Tier 3
vehicles.  Because overall impacts are relatively small in both future
years, we concluded that assessing exposure to ambient concentrations
and conducting a quantitative risk assessment of air toxic impacts was
not warranted.  However, we did develop population metrics, including
the population living in areas with increases or decreases in
concentrations of various magnitudes.  

Benzene

Our modeling projects that the proposed standards would have a notable
impact on ambient benzene concentrations.  In 2017, the first year the
proposed Tier 3 standards take effect, ambient benzene reductions are
generally between 0.001 and 0.01 µg/m³, or between 1 and 2.5 percent
in some areas (  REF _Ref309743776  Figure 7-30 ).  In 2030, our
modeling projects that the proposal would decrease ambient benzene
concentrations across much of the country on the order of 1 to 5
percent, with reductions ranging from 10 to 25 percent in some urban
areas (  REF _Ref254785068 \h  Figure 7-31 ).  Absolute decreases in
ambient concentrations of benzene are generally between 0.001 and 0.01
µg/m³ in rural areas and as much as 0.1 µg/m³ in urban areas (  REF
_Ref254785068 \h  Figure 7-31 ).   

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  30  Changes in
Benzene Ambient Concentrations Between the Reference Case and the
Control Case in 2017: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  31  Changes in
Benzene Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

1,3-Butadiene

μg/m³).  In 2030, reductions of 1,3-butadiene concentrations range
between 1 and 25 percent, with decreases of at least 0.005 μg/m³ in
urban areas (  REF _Ref254787393 \h  Figure 7-33 ).  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  32  Changes in
1,3-Butadiene Ambient Concentrations Between the Reference Case and the
Control Case in 2017: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  33  Changes in
1,3-Butadiene Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Acrolein 

Our modeling indicates the proposed standards would reduce ambient
concentrations of acrolein in 2017 and 2030.    REF _Ref309745558 
Figure 7-34  shows decreases in ambient concentrations of acrolein
generally between 1 and 2.5 percent across the parts of the country in
2017, corresponding to small decreases in absolute concentrations (less
than 0.001 μg/m³). Reductions of acrolein concentrations in 2030 range
between 1 and 25 percent, with decreases as high as 0.003 μg/m³ in a
few urban areas.

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  34  Changes in
Acrolein Ambient Concentrations Between the Reference Case and the
Control Case in 2017: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  35  Changes in
Acrolein Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Ethanol

Our modeling projects that the proposed standards would slightly
decrease ambient ethanol concentrations in 2030, with negligible impact
in 2017.  As shown in   REF _Ref309744692  Figure 7-36 , in 2017, annual
percent changes in ambient concentrations of ethanol are less than 1
percent across the country, with absolute concentrations of ± 0.01 ppb.
 In 2030, some parts of the country, especially urban areas, are
projected to have reductions in ethanol concentrations on the order of 1
to 5 percent as a result of the proposal (  REF _Ref254784021 \h  Figure
7-37 ).    REF _Ref254784021 \h  Figure 7-37  also shows that absolute
decreases in ambient concentrations of ethanol are generally between
0.001 and 0.1 ppb in 2030 with decreases in a few urban areas as high as
0.2 ppb.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  36  Changes in
Ethanol Ambient Concentrations Between the Reference Case and the
Control Case in 2017: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  37  Changes in
Ethanol Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Formaldehyde 

Our modeling projects that formaldehyde concentrations would slightly
decrease in parts of the country (mainly urban areas) as a result of the
proposal.  As shown in   REF _Ref309744295  Figure 7-38  and   REF
_Ref254781656 \h  Figure 7-39 , annual percent changes in ambient
concentrations of formaldehyde are less than 1 percent across much of
the country for 2017 but are on the order of 1 to 5 percent in 2030 in
some urban areas as a result of the proposal.    REF _Ref309744295 
Figure 7-38  and   REF _Ref254781656 \h  Figure 7-39  also show that
absolute changes in ambient concentrations of formaldehyde are generally
between 0.001 and 0.01 µg/m³ in both years, with some areas as high as
0.1 µg/m³ in 2030.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  38  Changes in
Formaldehyde Ambient Concentrations Between the Reference Case and the
Control Case in 2017: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  39  Changes in
Formaldehyde Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Acetaldehyde 

Our air quality modeling shows  annual percent changes in ambient
concentrations of acetaldehyde of generally less than 1 percent across
the U.S., although the proposal may decrease acetaldehyde concentrations
in some urban areas by 1 to 2.5 percent in 2030 (  REF _Ref254781030 \h 
\* MERGEFORMAT  Figure 7-40 ).  Changes in ambient concentrations of
acetaldehyde are generally in the range of 0.01 µg/m³ to -0.01 µg/m³
with decreases happening in the more populated areas and increases
happening in more rural areas (  REF _Ref254781030 \h  Figure 7-40 ). 

The complex photochemistry associated with NOX emissions and
acetaldehyde formation appears to be the explanation for the split
between increased rural concentrations and decreased urban
concentrations. In the atmosphere, acetaldehyde precursors react with
NOX to form peroxyacylnitrate (PAN).  Reducing NOX allows acetaldehyde
precursors to be available to form acetaldehyde instead.  This
phenomenon is more prevalent in rural areas where NOX is low.  The
chemistry involved is further described by a recent study done by
EPA’s Office of Research and Development and Region 3 evaluating the
complex effects of reducing multiple emissions on reactive air toxics
and criteria pollutants.  

  SHAPE  \* MERGEFORMAT   

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  40  Changes in
Acetaldehyde Ambient Concentrations Between the Reference Case and the
Control Case in 2017: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

  SHAPE  \* MERGEFORMAT   

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  41  Changes in
Acetaldehyde Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in
µg/m³ (right) 

Population Metrics

Although the reductions in ambient air toxics concentrations expected
from the proposed Tier 3 standards are generally small, they are
projected to benefit the majority of the U.S. population.  As shown in  
REF _Ref309629013  Table 7-38 , over 80 percent of the total U.S.
population is projected to experience a decrease in ambient benzene and
acrolein concentrations of at least 2.5 percent.  More than 85 percent
of the population is projected to experience decrease in 1,3-butadiene
concentrations of at least 5 percent.    REF _Ref309629013  Table 7-38 
also shows that over 80 percent of the U.S population is projected to
experience at least a 1 percent decrease in ambient ethanol
concentrations, and over a 60 percent would experience a similar
decrease in ambient formaldehyde concentrations with the proposed
standards.  

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  38  Percent of
Total Population Experiencing Changes in Annual Ambient Concentrations
of Toxic Pollutants in 2030 as a Result of the Proposed Standards

Percent Change	Benzene	Acrolein	1,3-Butadiene	Formaldehyde	Ethanol
Acetaldehyde

≤ -50







> -50 to ≤ -25

	0.1%



	> -25 to ≤ -10	2.8%	0.7%	56.8%



	> -10 to ≤ -5	23.7%	36.8%	30.8%



	> -5 to ≤ -2.5	54.5%	43.7%	7.1%	1.2%	33.0%	0.3%

> -2.5 to ≤ -1	17.7%	15.3%	3.4%	63.2%	55.3%	25.1%

> -1 to < 1	1.4%	3.5%	1.7%	35.6%	11.6%	74.6%

 ≥ 1 to < 2.5

	0.0%



	 ≥ 2.5  to < 5







≥ 5 to < 10







≥ 10 to < 25







≥ 25 to < 50







≥ 50 







Of note, the proposed rule is expected to decrease population exposure
to acrolein which is currently a national risk driver for noncancer
respiratory health effects as described in Section 7.2.5.4.1.  Our
modeling projects that acrolein concentrations would decrease to levels
below the inhalation reference concentration for acrolein (0.02 µg/m³)
for over 5 million people in 2030, meaning that as a result of the
proposed Tier 3 standards, 5 million fewer Americans will be exposed to
ambient levels of acrolein high enough to present a potential for
adverse health effects.  The inhalation reference concentration for
acrolein and other risk drivers is described in Section 6.1.5.6.  In
addition, the decrease in population exposure to the toxic compounds in
Table 7-36 will decrease cancer risks that are described in Section
6.1.5. 

Visibility

As described in Section 6.3.1 of this draft RIA, PM also causes adverse
visibility effects, and the EPA has set national ambient air quality
standards (NAAQS) and regional haze rules to protect against visibility
impairment.  In this section we present information on current and
model-projected future visibility levels at Mandatory Class I Federal
Areas.

Current Visibility Levels

Designated PM2.5 nonattainment areas indicate that, as of July 20, 2012,
over 105 million people live in nonattainment areas for the PM2.5 NAAQS.
 Thus, at least these populations would likely be experiencing
visibility impairment, as well as many thousands of individuals who
travel to these areas.  In addition, while visibility trends have
improved in Mandatory Class I Federal areas, these areas continue to
suffer from visibility impairment.  Calculated from light extinction
efficiencies from Trijonis et al. (1987, 1988), annual average visual
range under natural conditions in the East is estimated to be 150 km ±
45 km (i.e., 65 to 120 miles) and 230 km ± 35 km (i.e., 120 to 165
miles) in the West.,,  In summary, visibility impairment is experienced
throughout the U.S., in multi-state regions, urban areas, and remote
Mandatory Class I Federal areas. 

Projected Visibility Levels

Air quality modeling conducted for the final action was used to project
visibility conditions in 139 Mandatory Class I Federal areas across the
U.S.  The results show that in 2030 all the modeled areas would continue
to have annual average deciview levels above background and the proposed
rule would improve visibility in all these areas.  The average
visibility on the 20 percent worst days at all modeled Mandatory Class I
Federal areas is projected to improve by 0.04 deciviews, or 0.28
percent, in 2030.  The greatest improvement in visibilities will be seen
in Joshua Tree National Monument, where visibility is projected to
improve by 0.99 percent (0.16 DV) in 2030 due to the proposed standards.
   REF _Ref304971769 \h  Table 7-39  contains the full visibility
results from 2030 for the 139 analyzed areas.

Table   STYLEREF 1 \s  7 -  SEQ Table \* ARABIC \s 1  39  Visibility
Levels (in Deciviews) for Mandatory Class I Federal Areas on the 20
Percent Worst Days with and without this Proposed Rule

Class 1 Area

(20% worst days)	State	2005 Baseline Visibility (dv)a	2017 Reference 
2017 Tier 3 Control	2030 Reference 	2030 Tier 3 Control	Natural
Background

Sipsey Wilderness	AL	29.03	21.67	21.80	21.84	21.76	11.39

Caney Creek Wilderness	AR	26.36	21.00	21.03	21.10	21.02	11.33

Upper Buffalo Wilderness	AR	26.27	21.24	21.30	21.35	21.28	11.28

Chiricahua NM	AZ	12.89	12.29	12.28	12.23	12.21	6.92

Chiricahua Wilderness	AZ	12.89	12.27	12.27	12.22	12.20	6.91

Galiuro Wilderness	AZ	12.89	12.37	12.35	12.21	12.15	6.88

Grand Canyon NP	AZ	11.86	11.03	11.02	10.89	10.86	6.95

Mazatzal Wilderness	AZ	13.95	12.87	12.84	12.61	12.55	6.91

Mount Baldy Wilderness	AZ	11.32	10.91	10.91	10.85	10.84	6.95

Petrified Forest NP	AZ	13.56	12.90	12.89	12.75	12.72	6.97

Pine Mountain Wilderness	AZ	13.95	12.81	12.78	12.54	12.48	6.92

Saguaro NM	AZ	14.39	13.72	13.71	13.57	13.55	6.84

Sierra Ancha Wilderness	AZ	14.45	13.55	13.53	13.33	13.28	6.92

Superstition Wilderness	AZ	14.15	13.15	13.13	12.99	12.93	6.88

Sycamore Canyon Wilderness	AZ	15.45	14.83	14.81	14.70	14.67	6.96

Agua Tibia Wilderness	CA	22.36	18.87	18.87	18.19	18.09	7.17

Ansel Adams Wilderness (Minarets)	CA	15.24	14.48	14.48	14.29	14.25	7.12

Caribou Wilderness	CA	13.65	12.75	12.75	12.61	12.57	7.29

Cucamonga Wilderness	CA	18.44	15.83	15.83	15.42	15.32	7.17

Desolation Wilderness	CA	12.87	11.89	11.88	11.76	11.73	7.13

Emigrant Wilderness	CA	16.87	15.86	15.85	15.69	15.65	7.14

Hoover Wilderness	CA	11.61	11.05	11.05	10.95	10.94	7.12

John Muir Wilderness	CA	15.24	14.41	14.42	14.25	14.22	7.14

Joshua Tree NM	CA	18.90	16.74	16.72	16.14	15.98	7.08

Kaiser Wilderness	CA	15.24	14.21	14.21	13.99	13.95	7.13

Kings Canyon NP	CA	23.73	22.29	22.29	21.99	21.91	7.13

Lassen Volcanic NP	CA	13.65	12.78	12.77	12.61	12.56	7.31

Lava Beds NM	CA	14.13	13.14	13.13	13.20	13.17	7.49

Mokelumne Wilderness	CA	12.87	12.04	12.03	11.91	11.88	7.14

Pinnacles NM	CA	17.90	15.64	15.61	15.43	15.31	7.34

Point Reyes NS	CA	22.40	20.89	20.87	21.08	20.99	7.39

Redwood NP	CA	18.55	17.99	17.97	17.77	17.73	7.81

San Gabriel Wilderness	CA	18.44	15.86	15.86	15.37	15.26	7.17

San Gorgonio Wilderness	CA	21.43	19.50	19.50	19.01	18.89	7.10

San Jacinto Wilderness	CA	21.43	18.70	18.71	17.67	17.52	7.12

San Rafael Wilderness	CA	19.43	17.63	17.61	17.30	17.21	7.28

Sequoia NP	CA	23.73	21.90	21.88	21.52	21.42	7.13

South Warner Wilderness	CA	14.13	13.34	13.33	13.30	13.27	7.32

Thousand Lakes Wilderness	CA	13.65	12.76	12.75	12.60	12.55	7.32

Ventana Wilderness	CA	17.90	16.48	16.45	16.21	16.07	7.32

Yosemite NP	CA	16.87	15.87	15.86	15.71	15.67	7.14

Black Canyon of the Gunnison NM	CO	10.00	9.32	9.32	9.29	9.28	7.06

Eagles Nest Wilderness	CO	8.82	8.27	8.26	8.22	8.20	7.08

Flat Tops Wilderness	CO	8.82	8.43	8.43	8.39	8.38	7.07

Great Sand Dunes NM	CO	11.82	11.34	11.34	11.31	11.30	7.10

La Garita Wilderness	CO	10.00	9.59	9.59	9.54	9.54	7.06

Maroon Bells-Snowmass Wilderness	CO	8.82	8.38	8.38	8.36	8.35	7.07

Mesa Verde NP	CO	12.14	11.46	11.45	11.48	11.46	7.09

Mount Zirkel Wilderness	CO	9.72	9.29	9.29	9.28	9.27	7.08

Rawah Wilderness	CO	9.72	9.29	9.29	9.26	9.25	7.08

Rocky Mountain NP	CO	12.85	12.37	12.36	12.34	12.32	7.05

Weminuche Wilderness	CO	10.00	9.58	9.58	9.51	9.51	7.06

West Elk Wilderness	CO	8.82	8.35	8.35	8.33	8.32	7.07

Everglades NP	FL	22.31	19.30	19.06	19.10	19.04	11.15

Okefenokee	GA	27.13	21.29	21.44	21.47	21.40	11.45

Wolf Island	GA	27.13	21.10	21.12	21.18	21.12	11.42

Craters of the Moon NM	ID	14.06	13.30	13.30	13.10	13.05	7.13

Sawtooth Wilderness	ID	14.97	14.75	14.74	14.75	14.75	7.15

Selway-Bitterroot Wilderness	ID	17.11	16.83	16.83	16.86	16.85	7.32

Mammoth Cave NP	KY	31.37	22.87	23.09	23.14	23.07	11.53

Acadia NP	ME	22.89	18.51	18.80	18.83	18.80	11.45

Moosehorn	ME	21.72	17.81	18.01	18.03	18.01	11.36

Roosevelt Campobello International Park	ME	21.72	17.68	17.95	17.96	17.94
11.36

Isle Royale NP	MI	20.74	18.69	18.64	18.74	18.68	11.22

Seney	MI	24.16	21.32	21.33	21.44	21.35	11.37

Boundary Waters Canoe Area	MN	20.20	17.13	17.05	17.16	17.10	11.21

Voyageurs NP	MN	19.27	17.03	16.95	17.05	16.99	11.09

Hercules-Glades Wilderness	MO	26.75	21.92	21.99	22.04	21.97	11.27

Anaconda-Pintler Wilderness	MT	17.11	16.73	16.72	16.77	16.77	7.28

Bob Marshall Wilderness	MT	16.13	15.71	15.71	15.74	15.73	7.36

Cabinet Mountains Wilderness	MT	14.31	13.74	13.74	13.79	13.78	7.43

Gates of the Mountains Wilderness	MT	11.94	11.56	11.56	11.57	11.57	7.22

Glacier NP	MT	19.62	18.81	18.82	18.81	18.81	7.56

Medicine Lake	MT	18.21	17.65	17.65	17.58	17.57	7.30

Mission Mountains Wilderness	MT	16.13	15.57	15.57	15.62	15.61	7.39

Red Rock Lakes	MT	11.19	10.78	10.78	10.74	10.72	7.14

Scapegoat Wilderness	MT	16.13	15.68	15.68	15.70	15.70	7.29

UL Bend	MT	15.49	15.17	15.17	15.13	15.12	7.18

Linville Gorge Wilderness	NC	28.77	20.85	21.23	21.24	21.19	11.43

Shining Rock Wilderness	NC	28.54	20.47	20.78	20.81	20.73	11.45

Lostwood	ND	19.57	18.48	18.28	18.39	18.36	7.33

Theodore Roosevelt NP	ND	17.74	16.71	16.51	16.61	16.58	7.31

Great Gulf Wilderness	NH	22.82	16.73	17.01	17.05	17.01	11.31

Presidential Range-Dry River Wilderness	NH	22.82	16.68	16.97	17.00	16.96
11.33

Brigantine	NJ	29.01	21.56	21.88	21.93	21.84	11.28

Bandelier NM	NM	11.97	10.89	10.88	10.77	10.75	7.02

Bosque del Apache	NM	13.81	12.78	12.76	12.63	12.61	6.97

Carlsbad Caverns NP	NM	17.19	14.93	15.00	15.03	15.00	7.02

Gila Wilderness	NM	13.12	12.57	12.56	12.53	12.52	6.95

Pecos Wilderness	NM	9.60	9.08	9.07	9.01	8.99	7.04

Salt Creek	NM	18.27	16.70	16.70	16.67	16.64	6.99

San Pedro Parks Wilderness	NM	10.42	9.87	9.87	9.78	9.77	7.03

Wheeler Peak Wilderness	NM	9.60	8.92	8.92	8.82	8.80	7.07

White Mountain Wilderness	NM	13.01	12.16	12.16	12.20	12.19	6.98

Jarbidge Wilderness	NV	12.26	11.98	11.98	11.98	11.97	7.10

Wichita Mountains	OK	23.81	19.38	19.18	19.29	19.18	11.07

Crater Lake NP	OR	13.21	12.52	12.51	12.62	12.60	7.71

Diamond Peak Wilderness	OR	13.21	12.45	12.44	12.56	12.54	7.77

Eagle Cap Wilderness	OR	17.34	16.36	16.36	16.51	16.48	7.34

Gearhart Mountain Wilderness	OR	13.21	12.69	12.68	12.68	12.66	7.46

Hells Canyon Wilderness	OR	19.00	18.00	18.00	17.82	17.76	7.32

Kalmiopsis Wilderness	OR	16.38	15.48	15.46	15.60	15.56	7.71

Mount Hood Wilderness	OR	14.68	13.33	13.30	13.53	13.46	7.77

Mount Jefferson Wilderness	OR	15.80	14.96	14.95	15.12	15.09	7.81

Mount Washington Wilderness	OR	15.80	14.95	14.93	15.12	15.09	7.89

Mountain Lakes Wilderness	OR	13.21	12.44	12.43	12.57	12.55	7.57

Strawberry Mountain Wilderness	OR	17.34	16.66	16.65	16.55	16.51	7.49

Three Sisters Wilderness	OR	15.80	15.02	15.01	15.18	15.15	7.87

Cape Romain	SC	26.48	20.61	20.72	20.76	20.69	11.36

Badlands NP	SD	17.14	15.56	15.50	15.56	15.54	7.30

Wind Cave NP	SD	15.84	14.81	14.72	14.77	14.75	7.24

Great Smoky Mountains NP	TN	30.28	22.32	22.57	22.62	22.52	11.44

Joyce-Kilmer-Slickrock Wilderness	TN	30.28	22.03	22.29	22.34	22.25	11.45

Big Bend NP	TX	17.30	15.76	15.71	15.75	15.72	6.93

Guadalupe Mountains NP	TX	17.19	14.95	15.03	15.06	15.03	7.03

Arches NP	UT	10.77	10.13	10.13	10.20	10.18	6.99

Bryce Canyon NP	UT	11.62	10.95	10.95	10.93	10.93	6.99

Canyonlands NP	UT	10.77	10.15	10.13	10.24	10.21	7.01

Capitol Reef NP	UT	10.86	10.46	10.46	10.53	10.52	7.03

James River Face Wilderness	VA	29.12	20.45	20.61	20.65	20.56	11.24

Shenandoah NP	VA	29.31	20.24	20.67	20.69	20.61	11.25

Lye Brook Wilderness	VT	24.45	17.72	17.75	17.80	17.67	11.25

Alpine Lake Wilderness	WA	16.99	15.59	15.55	15.35	15.22	7.86

Glacier Peak Wilderness	WA	13.29	12.26	12.25	12.26	12.24	7.80

Goat Rocks Wilderness	WA	12.67	11.54	11.52	11.60	11.56	7.82

Mount Adams Wilderness	WA	12.67	11.57	11.56	11.65	11.61	7.78

Mount Rainier NP	WA	17.07	15.77	15.75	15.80	15.75	7.90

North Cascades NP	WA	13.29	12.24	12.23	12.18	12.17	7.78

Olympic NP	WA	15.83	14.63	14.61	14.71	14.65	7.88

Pasayten Wilderness	WA	15.35	14.34	14.32	14.48	14.46	7.77

Dolly Sods Wilderness	WV	29.05	20.23	20.82	20.84	20.79	11.32

Otter Creek Wilderness	WV	29.05	20.34	20.90	20.93	20.87	11.33

Bridger Wilderness	WY	10.73	10.38	10.38	10.39	10.39	7.08

Fitzpatrick Wilderness	WY	10.73	10.38	10.38	10.38	10.38	7.09

Grand Teton NP	WY	11.19	10.73	10.72	10.68	10.66	7.09

North Absaroka Wilderness	WY	11.30	10.99	10.99	10.97	10.97	7.09

Teton Wilderness	WY	11.19	10.81	10.80	10.77	10.75	7.09

Washakie Wilderness	WY	11.30	10.99	10.99	10.97	10.96	7.09

Yellowstone NP	WY	11.19	10.76	10.76	10.70	10.69	7.12

a The level of visibility impairment in an area is based on the
light-extinction coefficient and a unitless visibility index, called a
“deciview”, which is used in the valuation of visibility.  The
deciview metric provides a scale for perceived visual changes over the
entire range of conditions, from clear to hazy.  Under many scenic
conditions, the average person can generally perceive a change of one
deciview.  The higher the deciview value, the worse the visibility. 
Thus, an improvement in visibility is a decrease in deciview value.

Deposition of Nitrogen and Sulfur

As described in Section 6.3.2.1 of this draft RIA, deposition of
nitrogen and sulfur can cause adverse environmental effects.  In this
section we present information on current and model-projected future
nitrogen and sulfur deposition levels.

Current Levels of Nitrogen and Sulfur Deposition

Over the past two decades, the EPA has undertaken numerous efforts to
reduce nitrogen and sulfur deposition across the U.S.  Analyses of
long-term monitoring data for the U.S. show that deposition of both
nitrogen and sulfur compounds has decreased over the last 17 years.  The
data show that reductions were more substantial for sulfur compounds
than for nitrogen compounds.  In the eastern U.S., where data are most
abundant, total sulfur deposition decreased by about 44 percent between
1990 and 2007, while total nitrogen deposition decreased by 25 percent
over the same time frame.  These numbers are generated by the U.S.
national monitoring network and they likely underestimate nitrogen
deposition because neither ammonia nor organic nitrogen is measured. 
Although total nitrogen and sulfur deposition has decreased over time,
many areas continue to be negatively impacted by deposition.  Deposition
of inorganic nitrogen and sulfur species routinely measured in the U.S.
between 2005 and 2007 were as high as 9.6 kilograms of nitrogen per
hectare (kg N/ha) averaged over three years and 20.8 kilograms of sulfur
per hectare (kg S/ha) averaged over three years.   

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  42  Total Sulfur
Deposition in the Contiguous U.S., 1989-1991 and 2005 -2007

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  43  Total
Nitrogen Deposition in the Contiguous U.S., 1989-1991 and 2005-2007

Projected Levels of Nitrogen and Sulfur Deposition

Our air quality modeling projects decreases in both nitrogen and sulfur
deposition due to this proposed rule.    REF _Ref304975144 \h  Figure
7-44  shows that for nitrogen deposition by 2030 the proposed standards
would result in annual percent decreases of more than 5 percent in most
urban areas with decreases of more than 7 percent in urban areas in
Nevada, Arizona and Florida.  In addition, smaller decreases, in the 1
to 1.5 percent range, would occur over most of the rest of the country. 


Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  44  Percent
Change in Annual Total Nitrogen over the U.S. Modeling Domain as a
Result of the Proposed Standards

  REF _Ref304975404 \h  Figure 7-45  shows that for sulfur deposition
the proposed standards will result in annual percent decreases of more
than 2 percent in some areas in 2030.  The decreases in sulfur
deposition are likely due to projected reductions in the sulfur level in
fuel.  Minimal changes in sulfur deposition, ranging from decreases of
less than 0.5 percent to no change, are projected for the rest of the
country.  

Figure   STYLEREF 1 \s  7 -  SEQ Figure \* ARABIC \s 1  45  Percent
Change in Annual Total Sulfur over the U.S. Modeling Domain as a Result
of the Proposed Standards

Greenhouse Gas Emission Impacts 

Reductions in nitrous oxide (N2O) emissions and methane (CH4) emissions,
both potent greenhouse gas emissions (with global warming potentials 298
and 25 times greater than CO2, respectively), are projected for gasoline
cars and trucks due to the proposed sulfur and tailpipe standards. 
These projections are based on studies that provide a basis for
reductions in N2O and CH4 emissions due to the Tier 3 sulfur and vehicle
standards.  With respect to sulfur, a study published in 2004 by the
University of California at Riverside found a 29 percent reduction in
N2O emissions over the FTP and a 50 percent reduction over the US06 when
sulfur was reduced from 30 to 5 ppm.  EPA’s sulfur study, detailed in
Section 7.2, found a 25 percent reduction in CH4 emissions when sulfur
was reduced from 28 to 5 ppm (the EPA study did not measure N2O
emissions).  

Several studies have also established that reductions in tailpipe
standards for NOX and HC results in reductions in and N2O and CH4,
respectively.  N2O is unique in that it is not formed primarily during
combustion, but in the catalyst during catalyst warm-up, before the
catalyst reaches the temperatures required for full effectiveness (known
as “light-off”).  Improvements to catalyst technology required to
meet lower emission standards reduce the time required for the catalyst
to reach light-off, which reduces the window of N2O formation.  Studies
conducted by EPA and Environment Canada found that  N2O emission are
lower on vehicles certified to more stringent NOX emission standards., 
A study by Meffert, et al. established a strong correlation between
improvements in NOX catalytic conversion efficiency and reductions in
N2O emissions.  A study published by Behrentz, et al. in 2004 examined
the relationship between N2O and NOX from data collected by the
California Air Resources Board on cars and light trucks ranging from the
mid-1980s through early 2000’s Low Emission Vehicle (LEV) technology. 
The study reported an N2O: NOX ratio of 0.095±0.035 (with the lower end
of the range comprised of older oxidation catalyst technologies, and the
higher end of the range comprised of modern three-way catalyst
technologies), and supported the application of N2O: NOX ratios to NOX
emissions as a reasonable method for estimating N2O emission
inventories.  Subsequent analysis of this dataset by Meszler also found
that for vehicles equipped with more modern controls, N2O emissions
increased with vehicle mileage, suggesting deterioration in N2O
emissions as vehicles age.    

The  Meszler and Environment Canada studies cited above also established
that vehicles certified to more stringent HC standards emit less CH4, 
even though HC standards from Tier 1 and later do not include methane in
the regulated standards.  This trend is also reflected in the MOVES
model, based on analysis of correlation between CH4 and HC emissions. 
MOVES estimates methane as a function of total HC emissions, so the CH4
emission inventories account for effects such as deterioration,
temperature, aggressive driving, and reductions in tailpipe emission
standards.  Because of this, CH4 reductions from the Tier 3 program can
be estimated directly by MOVES, as a function of reductions in HC from
the  sulfur and vehicle standards (although this will provide a
conservative estimate of reductions, as the percent reduction in CH4
from using low sulfur fuel is about double that for total HC (Table
7-13)).  The MOVES results are shown in Table 7-40.

Table 7-40: Estimated Reduction in CH4 from Tier 3 Program (MMTCO2eq)

	2017	2030

Reference case onroad mobile emissions	2.1	2.5

Control case onroad mobile emissions	2.0	2.0

Reduction	0.1	0.5

MOVES N2O emissions are based directly on a limited sample of N2O
emission data, rather than  linking N2O emissions to NOX emissions as
suggested by Behrentz; as a result, the model does not estimate
potential N2O reduction concurrent with the  Tier 3 program.  Because of
this, the MOVES-based inventories are significantly lower than
inventories that take into account the N2O:NOX link,  because they do
not account for factors affecting light-duty NOX emissions, like
deterioration or aggressive driving, and lower NOX standards.  To
estimate N2O reductions, we have calculated N2O reductions due to
vehicle standards on the Tier 3 fleet (2017 and later model year
vehicles), using the 0.095 N2O:NOX factor from Behrentz.  For the
pre-Tier 3 fleet, we then bounded the N2O reductions due to sulfur
control using two methods:  1) applying the 0.095 N2O: NOX ratio
directly to the sulfur-related NOX reductions reported in Section 7.2;
and 2) applying the percent reduction in N2O  from the UC Riverside
study  to current MOVES inventory estimates for onroad gasoline
vehicles.  These two methods are outlined in Table 7-41, along with the
range of reductions that result.  

Table 7-41: Estimated Reduction in N2O from the Tier 3 Program

	2017	2030

Reduction from Tier 3 fleet

NOX reduction from Tier 3 fleet due to vehicle and sulfur standards
(U.S. Short Tons)	19,728	458,504

N2O reduction based on N2O:NOX ratio of 0.095 (MMTCO2eqa )	0.5	11.8

Reduction from pre-Tier 3 fleet

Method 1



Reference case onroad gasoline NOX emissions (U.S. Short Tons)	1,844,772
980,679

Reference case onroad gasoline N2O emissions based on N2O:NOX ratio of
0.095 (MMTCO2eqa )	47.4	25.2

Reduction from pre-Tier 3 fleet due to sulfur standard (U.S. Short Tons)
264,653	66,286

N2O reduction based on N2O:NOX ratio of 0.095 (MMTCO2eq a )	6.8	1.7

Method 2



U.S. onroad gasoline N2O emissions from pre-Tier 3 fleet using MOVES
(MMTCO2e)	10.3	2.0

Percent reduction in N2O going from 30 to 10 ppm	23% b	23%

N2O reduction (MMTCO2eq)	2.4	0.5

Total Range of N2O Reduction (MMTCO2eq)	2.9 –7.3	12.3-13.5

a Using GWP of 298 

b 29 percent from 25ppm sulfur reduction in UC Riverside study scaled to
20ppm reduction

Summing the results from Tables 7-40 and 7-41, the range of reductions
calculated for CH4 and N2O is from 3.0 to 7.4 million metric tons of
carbon dioxide equivalent (MMTCO2eq) in 2017, growing to 12.8 to 14.0
MMTCO2eq in 2030.  These reductions would be offset to some degree by
CO2 emissions associated with higher energy use required in the process
of removing sulfur.  To assess refinery permitting implications, EPA has
conducted an analysis of these refinery impacts, and estimates an
increase of up to 4.6 MMTCO2eq million metric tons of carbon dioxide
equivalent (MMTCO2e) in 2017 with the implementation of the lower sulfur
standards.  The actual increase will be lower, since refineries will not
be operating at their permit capacity.  The actual increase will also be
a function of several factors, including technology options selected by
the refineries and the projected use of averaging, banking and trading
in avoiding the need for investments at some refineries. As a result,
4.6 MMTCO2e represents an upper-bound estimate of the possible increase
in refinery CO2 emissions due to the need for additional process heat
and hydrogen production to enable additional hydrotreating.  

In 2017, the range of potential decrease in CH4  and N2O emissions
overlaps with the range of projected increase in CO2 from refinery
processes,  suggesting that a net increase or decrease in GHG emissions
cannot be quantified with certainty.  However, we estimate the program
would result in net GHG reductions as the program continues into the
future, as shown by our 2030 estimates.  

References

 The MOVES updates are reflected in a version of the MOVES model code
(April 14, 2011 Version a) and a concurrently updated version of the
MOVES default database (May 12, 2011), available in the Tier 3 docket 
As these updates are still draft, this code and/or database are not
approved for official use in SIP and conformity analyses.  

 The changes to the NO and NO2 rates did not impact the total NOX
emissions, but facilitated the output of separate results for nitric
oxide and nitrogen dioxide. 

 For air toxics, pre-aggregation level of ‘month’ was used instead
to improve the model run time.  Monthly pre-aggregation averages the
temperatures of all selected days and hours into a national average
hourly temperature for the month.  

 On Nov. 4 2010, EPA issued a partial waiver for MY2007 and newer
light-duty vehicles (75 FR 68094).  On January 26, 2011, EPA extended
the waiver to MY 2001-2006 light-duty vehicles (76 FR 4662).

 Refer to Section 1.7 of the RFS2 RIA for more on our FFV/vehicle
assumptions.

 We assumed that non-domestic automakers would produce around 2 percent
FFVs in current and future years.

 75 FR 25324 (May 7, 2010).

  The processing of control case sulfur levels nationwide introduced an
error in California counties.  Control case fuels were set to 10ppm in
all California counties, whereas the reference case sulfur levels ranged
from 8-19ppm. This led to small changes in emissions due to fuel changes
between reference and control, whereas in reality we expect no change in
California fuel.  The error had a negligible effect on national emission
totals, though the 2017 air quality modeling results captured regional
California impacts associated with the error that we judged not valid. 
While this error also affected 2030 air quality changes related to the
Tier 3 rule, the impact was far outweighed by emission reductions from
the vehicle program projected for this year. Our analysis therefore
includes the 2030 air quality impacts in California.      

 Due to a calculation error, an NMOG+ NOX standard of 0.177 was used for
Regulatory Class 41.  As a result, a slightly different phase-in of the
NMOG+NOX and CO rates was used in the proposal modeling.  This error
overstates the NMOG+ NOX reductions for this vehicle class by
approximately 2.5% in 2022.   

 By basing the data on light duty vehicles, it is possible that we are
misstating the emission profile of these larger vehicles, but as
emissions decrease, it is also possible that the emission profile for
the larger vehicles will more closely resemble that of light duty
vehicles.  

 One of the updates to MOVES for this analysis was to enable direct
input of the leak prevalence rates.

  This is an approximate breakdown, as there will be some NOX emission
reduction from heavy-duty gasoline vehicles greater than 14,000 pounds
beyond the 2017 model year that are counted in the “Tier 3 fleet”
here

 PAHs represents the sum of the following 15 PAH compounds:
acenaphthene, acenaphthalene, anthracene, benz(a)anthracene,
benzo(a)pyrene, benzo(b)fluoranthene, benze(g,h,i)perylene,
benzo(k)fluoranthene, chrysene, dibenzo(a,h)anthracene, fluoranthene,
fluorine, indeno(1,2,3,cd)pyrene, phenanthrene, and pyrene.  These PAHs
are included in EPA’s national emissions inventory (NEI).

 For more information, please see the website for SMOKE: 
http://www.smoke-model.org/index.cfm.  

 Allen, D., Burns, D., Chock, D., Kumar, N., Lamb, B., Moran, M.
(February 2009). Report on the Peer Review of the Atmospheric Modeling
and Analysis Division, NERL/ORD/EPA.  U.S. EPA, Research Triangle Park,
NC.,  http://www.epa.gov/amad/peer/2009_AMAD_PeerReviewReport.pdf.

 CMAQ version 4.7 was released on December, 2008.  It is available from
the Community Modeling and Analysis System (CMAS) as well as previous
peer-review reports at: http://www.cmascenter.org.

 Regional Planning Organization regions include:  Mid-Atlantic/Northeast
Visibility Union (MANE-VU), Midwest Regional Planning Organization –
Lake Michigan Air Directors Consortium (MWRPO-LADCO), Visibility
Improvement State and Tribal Association of the Southeast (VISTAS),
Central States Regional Air Partnership (CENRAP), and Western Regional
Air Partnership (WRAP).

 These other modeling studies represent a wide range of modeling
analyses which cover various models, model configurations, domains,
years and/or episodes, chemical mechanisms, and aerosol modules.

 All rate coefficients are listed at 298 K and, if applicable, 1 bar of
air.

 Acetaldehyde is not the only source of acetyl peroxy radicals in the
atmosphere. For example, dicarbonyl compounds (methylglyoxal, biacetyl,
and others) also form acetyl radicals, which can further react to form
peroxyacetyl nitrate (PAN).

 All rate coefficients are listed at 298 K and, if applicable, 1 bar of
air.

 All rate coefficients are listed at 298 K and, if applicable, 1 bar of
air.

 Many aromatic hydrocarbons, particularly those present in high
percentages in gasoline (toluene, m-, o-, p-xylene, and 1,3,5-,
1,2,4-trimethylbenzene), form methylglyoxal and biacetyl, which are also
strong generators of acetyl radicals (Smith, D.F., T.E. Kleindienst,
C.D. McIver (1999) Primary product distribution from the reaction of OH
with m-, p-xylene and 1,2,4- and 1,3,5-Trimethylbenzene. J. Atmos.
Chem., 34: 339- 364.).

 All rate coefficients are listed at 298 K and, if applicable, 1 bar of
air.

 From U.S. EPA, 2011.  Our Nation’s Air: Status and Trends through
2010. EPA-454/R-12-001. February 2012.  Available at:
http://www.epa.gov/airtrends/2011/.  

 A nonattainment area is defined in the Clean Air Act (CAA) as an area
that is violating an ambient standard or is contributing to a nearby
area that is violating the standard.

 The 138 million total is calculated by summing, without double
counting, the 1997 and 2008 ozone nonattainment populations contained in
the Summary Nonattainment Area Population Exposure report
(http://www.epa.gov/oar/oaqps/greenbk/popexp.html).  If there is a
population associated with both the 1997 and 2008 nonattainment areas,
and they are not the same, then the larger of the two populations is
included in the sum.

 The Los Angeles South Coast Air Basin 8-hour ozone nonattainment area
and the San Joaquin Valley Air Basin 8-hour ozone nonattainment area are
designated as extreme and will have to attain before June 15, 2024.  The
Sacramento, Coachella Valley, Western Mojave and Houston 8-hour ozone
nonattainment areas are designated as severe and will have to attain by
June 15, 2019.  

 The processing of control case sulfur levels nationwide introduced an
error in California counties.  Control case fuels were set to 10 ppm in
all California counties, whereas the reference case sulfur levels ranged
from 8-19 ppm. This led to small changes in emissions due to fuel
changes between reference and control, whereas in reality we expect no
change in California fuel.  The error had a negligible effect on
national emission totals, though the 2017 air quality modeling results
captured regional California impacts associated with the error that we
judged not valid.

 From U.S. EPA, 2011.  Our Nation’s Air: Status and Trends through
2010. EPA-454/R-12-001. February 2012.  Available at:
http://www.epa.gov/airtrends/2011/.  

 From U.S. EPA, 2011.  Our Nation’s Air: Status and Trends through
2010. EPA-454/R-12-001. February 2012.  Available at:
http://www.epa.gov/airtrends/2011/.  

 Data come from Summary Nonattainment Area Population Exposure Report,
current as of December 14, 2012 at:  HYPERLINK
"http://www.epa.gov/oar/oaqps/greenbk/popexp.html"
http://www.epa.gov/oar/oaqps/greenbk/popexp.html  and contained in
Docket EPA-HQ-OAR-2011-0135.

 An annual PM2.5 design value is the concentration that determines
whether a monitoring site meets the annual NAAQS for PM2.5.  The full
details involved in calculating an annual PM2.5 design value are given
in appendix N of 40 CFR part 50.

 The processing of control case sulfur levels nationwide introduced an
error in California counties.  Control case fuels were set to 10 ppm in
all California counties, whereas the reference case sulfur levels ranged
from 8-19 ppm. This led to small changes in emissions due to fuel
changes between reference and control, whereas in reality we expect no
change in California fuel.  The error had a negligible effect on
national emission totals, though the 2017 air quality modeling results
captured regional California impacts associated with the error that we
judged not valid.

 The processing of control case sulfur levels nationwide introduced an
error in California counties.  Control case fuels were set to 10 ppm in
all California counties, whereas the reference case sulfur levels ranged
from 8-19 ppm. This led to small changes in emissions due to fuel
changes between reference and control, whereas in reality we expect no
change in California fuel.  The error had a negligible effect on
national emission totals, though the 2017 air quality modeling results
captured regional California impacts associated with the error that we
judged not valid.

 A 24-hour PM2.5 design value is the concentration that determines
whether a monitoring site meets the 24-hour NAAQS for PM2.5.  The full
details involved in calculating a 24-hour PM2.5 design value are given
in appendix N of 40 CFR part 50.

 The processing of control case sulfur levels nationwide introduced an
error in California counties.  Control case fuels were set to 10 ppm in
all California counties, whereas the reference case sulfur levels ranged
from 8-19 ppm. This led to small changes in emissions due to fuel
changes between reference and control, whereas in reality we expect no
change in California fuel.  The error had a negligible effect on
national emission totals, though the 2017 air quality modeling results
captured regional California impacts associated with the error that we
judged not valid.

 EPA, 2010.  Final Regulatory Impact Analysis (RIA) for the NO2 National
Ambient Air Quality Standards (NAAQS). 
http://www.epa.gov/ttn/ecas/regdata/RIAs/FinalNO2RIAfulldocument.pdf

 The processing of control case sulfur levels nationwide introduced an
error in California counties.  Control case fuels were set to 10ppm in
all California counties, whereas the reference case sulfur levels ranged
from 8-19ppm. This led to small changes in emissions due to fuel changes
between reference and control, whereas in reality we expect no change in
California fuel.  The error had a negligible effect on national emission
totals, though the 2017 air quality modeling  results captured regional
California impacts associated with the error that we judged not valid.

 NATA also includes estimates of risk attributable to background
concentrations, which includes contributions from long-range transport,
persistent air toxics, and natural sources; as well as secondary
concentrations, where toxics are formed via secondary formation.  Mobile
sources substantially contribute to long-range transport and secondarily
formed air toxics.

 NATA relies on a Guassian plume model, Assessment System for Population
Exposure Nationwide (ASPEN), to estimate toxic air pollutant
concentrations. Projected air toxics concentrations presented in this
final action were modeled with CMAQ 4.7.1.

 The level of visibility impairment in an area is based on the
light-extinction coefficient and a unitless visibility index, called a
“deciview”, which is used in the valuation of visibility.  The
deciview metric provides a scale for perceived visual changes over the
entire range of conditions, from clear to hazy.  Under many scenic
conditions, the average person can generally perceive a change of one
deciview.  The higher the deciview value, the worse the visibility. 
Thus, an improvement in visibility is a decrease in deciview value.

 U.S. EPA. (2012). U.S. EPA’s Report on the Environment. Data accessed
online February 15, 2012 at:
http://cfpub.epa.gov/eroe/index.cfm?fuseaction=detail.viewPDF&ch=46&lSho
wInd=0&subtop=341&lv=list.listByChapter&r=216610 and contained in Docket
EPA-HQ-OAR-2011-0135.  

 U.S. EPA. (2012). U.S. EPA's Report on the Environment. Data accessed
online February 15, 2012 at:
http://cfpub.epa.gov/eroe/index.cfm?fuseaction=detail.viewPDF&ch=46&lSho
wInd=0&subtop=341&lv=list.listByChapter&r=216610 and contained in Docket
EPA-HQ-OAR-2011-0135.  

 The global warming potentials (GWP) used in this rule are consistent
with the 100-year time frame values in the 2007 Intergovernmental Panel
on Climate Change (IPCC) Fourth Assessment Report (AR4).  At this time,
the 1996 IPCC Second Assessment Report (SAR) 100-year GWP values are
used in the official U.S. greenhouse gas inventory submission to the
United Nations Framework Convention on Climate Change (per the reporting
requirements under that international convention, which were last
updated in 2006) .  N2O has a 100-year GWP of 298 and CH4 has a 100-year
GWP of 25 according to the 2007 IPCC AR4.

Chapter 1 

*** E.O. 12866 Review – Revised Version – Do Not Cite, Quote, or
Release During Review ***

Page   PAGE  2 

  PAGE   \* MERGEFORMAT  7-1 

 2017 and Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions
and Corporate Average Fuel Economy Standards; Final Rule (77 FR
62623–63200), October 2012.

 MOVES (Motor Vehicle Emission Simulator) website:
http://www.epa.gov/otaq/models/moves/index.htm

 U.S. EPA. 2013. “Memorandum to Docket:  Updates to MOVES for the Tier
3 NPRM”

 U.S. EPA. 2012. Memorandum to Docket: “Development of fuel
adjustments and toxic fractions for use in MOVES2010a using draft
statistical models generated using results from the Phase-3 EPAct
Project.”  

 Particle Emissions from a 2009 Gasoline Direct Injection Engine Using
Different Commercially Available Fuels, SAE Paper 2010-01-2117, Imad A.
Khalek and Thomas Bougher, October 25, 2010.

 Development of a Predictive Model for Gasoline Vehicle Particulate
Matter Emissiosn, SAE Paper 2010-01-2115, Koichiro Aikawa and Takayuki
Sakurai, October 25, 2010.

 Masashi Iisuka, Japan Petroleum Energy Center, “Effect of Fuel
Properties on Emissions from Direct Injection Gasoline Vehicle,” 5th
Asian Petroleum Technology Symposium, Jakarta, Indonesia, January 25,
2007.

 Khalek, I., Bougher, T., and Merritt, P. M.  2009.  Phase 1 of the
Advanced Collaborative Emissions Study.  Prepared by Southwest Research
Institute for the Coordinating Research Council and the Health Effects
Institute, June 2009.  Available at  HYPERLINK "http://www.crcao.org"
www.crcao.org .

  U.S. EPA. 2013. “Memorandum to Docket:  Updates to MOVES for the
Tier 3 NPRM”

 U.S. EPA, 2008, Memorandum to Docket: Using MOVES to Generate
Inventories for the RFS2 NPRM.

 Regulation of Fuels and Fuel Additives:  Changes to Renewable Fuel
Standard Program; Final Rule (75 FR 14670), March 2010.

 RFS2 Final Rule Regulatory Impact Analysis, Chapter 3

 Coordinating Research Council (2011), Exhaust and Evaporative Emissions
Testing of Flexible Fuel Vehicles, CRC E-80

 R. Cook,  et al.  (2011).  Air quality impacts of increased use of
ethanol under the United States’ Energy Independence and Security Act.
 Atmospheric Environment, 45: 7714-7724.

 Luecken, D. J. et al. 2011.  Regional sources of atmospheric
formaldehyde and acetaldehyde, and implications for atmospheric
modeling.  Atmospheric Environment, in press.
doi:10.1016/j.atmosenv.2011.10.005

 "DOCUMENTATION FOR THE 2005 MOBILE NATIONAL EMISSIONS INVENTORY,
VERSION 2"
ftp://ftp.epa.gov/EmisInventory/2005_nei/mobile/2005_mobile_nei_version_
2_report.pdf

 Mathpro (October 2011). Refining Economics of a National Low Sulfur,
Low RVP Gasoline Standard, Performed for The International Council for
Clean Transportation,  Available at:
http://www.theicct.org/sites/default/files/publications/ICCT04_Tier3_Rep
ort_Final_v4_All.pdf.  Accessed December 12, 2011. 

 Baker and O’Brien (July 2011).Potential Supply and Cost Impacts of
Lower Sulfur, Lower RVP Gasoline; prepared for The American Petroleum
Institute. Available at:
http://www.api.org/Newsroom/upload/110715_LowerSulfur_LowerRVP_Final.pdf
.  Accessed December 12, 2011

 U.S.. EPA.2011. Development of Emission Rates for Light-Duty Vehicles
in the Motor Vehicle Emissions Simulator (MOVES2010): Final Report. 
(EPA-420-R-11-011)
http://www.epa.gov/otaq/models/moves/documents/420r11011.pdf.

 U.S. EPA. 2011. Memorandum to Docket: Tier 3 Emission Rates for
Gasoline Light Duty HC/CO/NOx Exhaust

 U.S. EPA. 2013. “Memorandum to Docket:  Updates to MOVES for the Tier
3 NPRM”

 U.S. EPA. 2013. “Memorandum to Docket:  Updates to MOVES for the Tier
3 NPRM”

 U.S. EPA. 2013. “Memorandum to Docket:  Updates to MOVES for the Tier
3 NPRM”

 Sonntag, Darrell. 2013.  Supporting Technical Document:  Estimated
Reductions in Particulate Matter Emissions from Light-duty and
Medium-duty Gasoline vehicles through Implementation of Tier 3
Regulations.

 Defries, et. al. 2011. “Estimates of the Fraction of the Fleet with
High Evaporative Emissions based on the Ken Caryl Station “(Denver,
Colorado) Field Study, Eastern Research Group report to EPA
0218.04.004.001

 Haskew, Harold, et al. 2011.  “Vehicle Evaporative Emission
Mechanisms:  A Pilot Study,”  available at   HYPERLINK
"http://crcao.org/reports/recentstudies2008/E-77PilotStudy/E-77PilotStud
yFinalReport6.24.08.pdf" 
http://crcao.org/reports/recentstudies2008/E-77PilotStudy/E-77PilotStudy
FinalReport6.24.08.pdf 

U.S. EPA. 2011.  MOVES2010 Fuel Adjustment and Air Toxic Emission
Calculation Algorithm –Development and Results (EPA-420-R-11-009)
http://www.epa.gov/otaq/models/moves/documents/420r11009.pdf

 U.S. EPA.  2013. "The Effects of Fuel Sulfur Level on Emissions from
Tier 2 Vehicles In-Use" (draft).

 See Chapter III and Appendix B of the Regulatory Impact Analysis for
the Tier2/Sulfur Final Rule EPA 420-R-99-023. 

 See Chapter 6 of the Regulatory Impact Analysis for the Control of
Hazardous Air Pollutants from Mobile Sources Final Rule, EPA
420-R-07-002.

 Ball D., Clark D., Moser D. 2011. “Effects of Fuel Sulfur on FTP NOx
Emissions from a PZEV 4 Cylinder Application”. SAE 2011 World Congress
Paper 2011-01-0300.  SAE International: Warrendale, PA.

 U.S. EPA.  2005 EPA’s National Inventory Model (NMIM), A Consolidated
Emissions Modeling System for MOBILE6 and NONROAD; EPA420-R-05-024;
Office of Transportation and Air Quality, Ann Arbor, MI.   HYPERLINK
"http://www.epa.gov/otaq/models/nmim/420r05024.pdf" 
http://www.epa.gov/otaq/models/nmim/420r05024.pdf  

 Final Rulemaking: Greenhouse Gas Emissions Standards and Fuel
Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles
(signed August 9, 2011). 
http://www.epa.gov/oms/climate/regulations.htm#1-2

 Final Rule: Control of Emissions of Air Pollution from Locomotives and
Marine Compression-Ignition Engines Less Than 30 Liters per Cylinder
(published May 6, 2008 and republished June 30, 2008). 
http://www.epa.gov/oms/locomotives.htm#2008final

 Final Rule: Control of Emissions from New Marine Compression-Ignition
Engines at or Above 30 Liters per Cylinder (published April 30, 2010). 
http://www.epa.gov/oms/oceanvessels.htm#car-ems.

 U.S. Energy Information Administration (2011),  Annual Energy Outlook
2011, DOE/EIA-0383(2011)

 U.S. EPA 2011, Tier 3 Emissions Inventory for Air Quality Modeling
Technical Support Document

 U.S. EPA 2011.  Tier 3 Emissions Inventory for Air Quality Modeling
Technical Support Document

 User’s Guide for the SMOKE–MOVES Integration Tool, July 14, 2010. 
Available at http://www.smoke-model.org/smoke_moves_tool/.

 For more information, please see the website for SMOKE:   HYPERLINK
"http://www.smoke-model.org/index.cfm"
http://www.smoke-model.org/index.cfm .  

  Michaels, H. (2011, December).  Documentation of Smoke-MOVES
Post-Processing Script Inconsistency. Memorandum to the docket.

 Grell, G., J. Dudhia, and D. Stauffer, 1994: A Description of the
Fifth-Generation Penn State/NCAR Mesoscale Model (MM5),
NCAR/TN-398+STR., 138 pp, National Center for Atmospheric Research,
Boulder CO.

 "Air Quality Modeling Technical Support Document: Heavy-Duty Vehicle
Greenhouse Gas Emission Standards Final Rule", (EPA-454-R-11-004, August
2011) (http://www.epa.gov/otaq/climate/documents/454r11004.pdf)

 "Documentation for the 2005 Mobile National Emissions Inventory,
Version 2",  (December 2008)
(ftp://ftp.epa.gov/EmisInventory/2005_nei/mobile/2005_mobile_nei_version
_2_report.pdf)

 U.S. EPA. 2010 MOVES2010 Highway Vehicle Population and Activity Data, 
EPA-420-R-10-026, http://www.epa.gov/otaq/models/moves/420r10026.pdf

 Documentation for the 2005 Mobile National Emissions Inventory, Version
2",  (December 2008)
(ftp://ftp.epa.gov/EmisInventory/2005_nei/mobile/2005_mobile_nei_version
_2_report.pdf)

 U. S. EPA. 2011.  Spreadsheet 
“Comparison_VMT_future_15jul2011.xlsx.”  Docket EPA-HQ-OAR-2011-0135

 "Development of Methodology for Estimating VMT Weighting by Facility
Type", (EPA420-R-01-009, M6.SPD.003, April 2001) . Prepared for EPA by
Systems Applications International, Inc., ICF Consulting Group.
(http://www.epa.gov/otaq/models/mobile6/r01009.pdf)

 U.S. EPA 2011.  Tier 3 Emissions Inventory for Air Quality Modeling
Technical Support Document.

 U.S. EPA 2011.  Tier 3 Emissions Inventory for Air Quality Modeling
Technical Support Document .

 Cook, R. 2012.  Development of Air Quality Reference Case Upstream and
Portable Fuel Container Inventories for Tier 3 Proposal. Memorandum to
Docket EPA-HQ-OAR-2010-0162.

U.S. EPA. 2009.  Exhaust Emission Profiles for EPA SPECIATE Database:
Energy Policy Act (EPAct) Low-Level Ethanol Fuel Blends and Tier 2
Light-Duty Vehicles.  EPA-20-R-09-002, June 2009. 

 U.S. EPA. 2010. Evaporative Emissions from In-Use Vehicles: Test Fleet
Expansion (CRC E-77-2b). EPA-420-R-10-025. Available at:
http://www.epa.gov/otaq/emission-factors-research/420r10025.pdf.
Accessed December 15, 2011. 

 Coordinating Research Council. 2010. Study To Determine Evaporative
Emission Breakdown, Including Permeation Effects and Diurnal Emissions,
Using E20 Fuels on Aging Enhanced Evaporative Emissions Certified
Vehicles (CRC E-77-2c). Available at:
http://www.crcao.com/reports/recentstudies2011/E-77-2c/E-77-2c%20Final%2
0Report%20for%20sure%201-28-11.pdf.  Accessed December 15, 2011. 

 U.S. EPA. 2011.  Hydrocarbon Composition of Gasoline Vapor Emissions
from Enclosed Fuel Tanks. EPA-420-R-11-018.  Available in the docket to
this rule. 

 U.S. EPA. 2013, January. Mobile Source Hydrocarbon Speciation Profiles
for the Tier 3 Rule NPRM and Anti-backsliding Study AQ Modeling. 
Memorandum to the docket.

 Byun, D.W., Ching, J. K.S. 1999. Science algorithms of the EPA models-3
community multiscale air quality (CMAQ) modeling system. Washington, DC:
U.S. Environmental Protection Agency, Office of Research and
Development. 

 Byun, D.W., Schere, K.L. (2006). Review of the Governing Equations,
Computational Algorithms, and Other Components of the Models-3 Community
Multiscale Air Quality (CMAQ) Modeling System. Journal of Applied
Mechanics Reviews, 59(2), 51-77. 

 Dennis, R.L., Byun, D.W., Novak, J.H., Galluppi, K.J., Coats, C.J., and
Vouk, M.A. (1996). The next generation of integrated air quality
modeling: EPA’s Models-3, Atmospheric Environment, 30, 1925-1938. 

 Hogrefe, C., Biswas, J., Lynn, B., Civerolo, K., Ku, J.Y., Rosenthal,
J., et al. (2004). Simulating regional-scale ozone climatology over the
eastern United States: model evaluation results. Atmospheric
Environment, 38(17), 2627-2638. 

 Lin, M., Oki, T., Holloway, T., Streets, D.G., Bengtsson, M., Kanae, S.
(2008). Long-range transport of acidifying substances in East Asia-Part
I: Model evaluation and sensitivity studies. Atmospheric Environment,
42(24), 5939-5955.  

 U.S. Environmental Protection Agency. (2008). Technical support
document for the final locomotive/marine rule: Air quality modeling
analyses. Research Triangle Park, N.C.: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Air Quality
Assessment Division. 

 Grell, G., Dudhia, J., Stauffer, D. (1994). A Description of the
Fifth-Generation Penn State/NCAR Mesoscale Model (MM5),
NCAR/TN-398+STR., 138 pp, National Center for Atmospheric Research,
Boulder CO. 

 Grell, G., Dudhia, J., Stauffer, D. (1994). A Description of the
Fifth-Generation Penn State/NCAR Mesoscale Model (MM5),
NCAR/TN-398+STR., 138 pp, National Center for Atmospheric Research,
Boulder CO. 

 Byun, D.W., Ching, J. K.S. (1999). Science algorithms of EPA Models-3
Community Multiscale Air Quality (CMAQ) modeling system,
EPA/600/R-99/030, Office of Research and Development).  Please also see:
  HYPERLINK "http://www.cmascenter.org/" http://www.cmascenter.org/ . 

 Le Sager, P. Yantosca, B., Carouge, C. (2008). GEOS-CHEM v8-01-02
User’s Guide, Atmospheric Chemistry Modeling Group, Harvard
University, Cambridge, MA, December 18, 2008. 

 U.S. EPA, (2004), Procedures for Estimating Future PM2.5 Values for the
CAIR Final Rule by Application of the (Revised) Speciated Modeled
Attainment Test (SMAT)- Updated 11/8/04.  

 U.S. EPA, (2011), Final Cross State Air Pollution Rule Air Quality
Modeling TSD. 

 U.S. EPA (2007) Guidance on the Use of Models and Other Analyses For
Demonstrating Attainment of Air Quality Goals for Ozone, PM2.5, and
Regional Haze; EPA-454/B-07-002; Research Triangle Park, NC; April 2007.


 Yarwood G, Rao S, Yocke M, Whitten GZ (2005) Updates to the Carbon Bond
Chemical Mechanism: CB05. Final Report to the U.S. EPA, RT-0400675,
December 8, 2005.
http://www.camx.com/publ/pdfs/CB05_Final_Report_120805.pdf.  

 U.S. EPA. 2013. “Memorandum to Docket:  Updates to MOVES for the Tier
3 NPRM”

 Dodge, M.C., 2000. Chemical oxidant mechanisms for air quality
modeling:  critical review. Atmospheric Environment 34, 2103-2130.  

 Atkinson, R., Baulch, D.L., Cox, R.A., Crowley, J.N., Hampson, R.F.
Jr., Hynes, R.G., Jenkin, M.E., Kerr, J.A., Rossi, M.J., Troe, J. (2005)
Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry -
IUPAC Subcommittee on Gas Kinetic Data Evaluation for Atmospheric
Chemistry. July 2005 web version.
http://www.iupac-kinetic.ch.cam.ac.uk/index.html. 

 Atkinson, R., Baulch, D.L., Cox, R.A., Crowley, J.N., Hampson, R.F.
Jr., Hynes, R.G., Jenkin, M.E., Kerr, J.A., Rossi, M.J., Troe, J. (2005)
Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry -
IUPAC Subcommittee on Gas Kinetic Data Evaluation for Atmospheric
Chemistry. July 2005 web version.
http://www.iupac-kinetic.ch.cam.ac.uk/index.html. 

 Atkinson, R., Baulch, D.L., Cox, R.A., Crowley, J.N., Hampson, R.F.
Jr., Hynes, R.G., Jenkin, M.E., Kerr, J.A., Rossi, M.J., Troe, J. (2005)
Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry -
IUPAC Subcommittee on Gas Kinetic Data Evaluation for Atmospheric
Chemistry. July 2005 web version.
http://www.iupac-kinetic.ch.cam.ac.uk/index.html. 

 Atkinson, R., Baulch, D.L., Cox, R.A., Crowley, J.N., Hampson, R.F.
Jr., Hynes, R.G., Jenkin, M.E., Kerr, J.A., Rossi, M.J., Troe, J. (2005)
Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry -
IUPAC Subcommittee on Gas Kinetic Data Evaluation for Atmospheric
Chemistry. July 2005 web version.
http://www.iupac-kinetic.ch.cam.ac.uk/index.html. 

 Sander, S.P., Friedl, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E.,
Orkin, V.L., Moortgat, G.K., Ravishankara, A.R., Kolb, C.E., Molina,
M.J., Finlayson-Pitts, B.J. (2003) Chemical Kinetics and Photochemical
Data for use in Atmospheric Studies, Evaluation Number 14. NASA Jet
Propulsion Laboratory http://jpldataeval.jpl.nasa.gov/index.html.  

 J.G. Calvert, A. Mellouki, J.J. Orlando, M.J. Pilling, and T.J.
Wallington, 2011. The mechanisms of atmospheric oxidation of the
oxygenates. Oxford University Press, New York/Oxford.

 Sander, S.P., Friedl, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E.,
Orkin, V.L., Moortgat, G.K., Ravishankara, A.R., Kolb, C.E., Molina,
M.J., Finlayson-Pitts, B.J. (2003) Chemical Kinetics and Photochemical
Data for use in Atmospheric Studies, Evaluation Number 14. NASA Jet
Propulsion Laboratory http://jpldataeval.jpl.nasa.gov/index.html. 

 Yarwood, G., Rao, S., Yocke, M., Whitten, G.Z., 2005. Updates to the
Carbon Bond Mechanism: CB05.  Final Report to the U.S. EPA, RT-0400675.
Yocke and Company, Novato, CA. 

 Luecken, D.J., Phillips, S., Sarwar, G., Jang, C., 2008b. Effects of
using the CB05 vs. SAPRC99 vs. CB4 chemical mechanism on model
predictions: Ozone and gas-phase photochemical precursor concentrations.
Atmospheric Environment 42, 5805-5820.  

 Sander, S.P., Friedl, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E.,
Orkin, V.L., Moortgat, G.K., Ravishankara, A.R., Kolb, C.E., Molina,
M.J., Finlayson-Pitts, B.J., 2003. Chemical Kinetics and Photochemical
Data for use in Atmospheric Studies, Evaluation Number 14. NASA Jet
Propulsion Laboratory.  

 Sander, S.P., Friedl, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E.,
Orkin, V.L., Moortgat, G.K., Ravishankara, A.R., Kolb, C.E., Molina,
M.J., Finlayson-Pitts, B.J., 2003. Chemical Kinetics and Photochemical
Data for use in Atmospheric Studies, Evaluation Number 14. NASA Jet
Propulsion Laboratory.  

 Atkinson R, Arey J (2003) Atmospheric Degradation of Volatile Organic
Compounds. Chem Rev 103: 4605-4638.  

 Atkinson, R., Baulch, D.L., Cox, R.A., Crowley, J.N., Hampson, R.F.
Jr., Hynes, R.G., Jenkin, M.E., Kerr, J.A., Rossi, M.J., Troe, J. (2005)
Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry -
IUPAC Subcommittee on Gas Kinetic Data Evaluation for Atmospheric
Chemistry. July 2005 web version.
http://www.iupac-kinetic.ch.cam.ac.uk/index.html. 

 Edney, E. O., T. E. Kleindienst, M. Lewandowski, and J. H. Offenberg,
2007. Updated SOA chemical mechanism for the Community Multi-Scale Air
Quality model, EPA 600/X-07/025, U.S. EPA, Research Triangle Park, NC. 

 Carlton, A.G., B. J. Turpin, K. Altieri, S. Seitzinger, R. Mathur, S.
Roselle, R. J. Weber, (2008),  CMAQ model performance enhanced when
in-cloud SOA is included: comparisons of OC predictions with
measurements, Environ. Sci. Technol. 42 (23), 8798–8802.  

 Lewandowski M, M Jaoui , JH Offenberg , TE Kleindienst, EO Edney, RJ
Sheesley, JJ Schauer (2008)  Primary and secondary contributions to
ambient PM in the midwestern United States, Environ Sci Technol
42(9):3303-3309.
http://pubs.acs.org/cgi-bin/article.cgi/esthag/2008/42/i09/html/es072041
2.html. 

 Kleindienst TE, M Jaoui, M Lewandowski, JH Offenberg, EO Edney (2007)
Estimates of the contributions of biogenic and anthropogenic
hydrocarbons to secondary organic aerosol at a southeastern U.S.
location, Atmos Environ 41(37):8288-8300.  

 Offenberg JH, CW Lewis, M Lewandowski, M Jaoui, TE Kleindienst, EO
Edney (2007) Contributions of Toluene and -pinene to SOA Formed in an
Irradiated Toluene/-pinene,NOx/Air Mixture: Comparison of Results
Using 14C Content and SOA Organic Tracer Methods, Environ Sci Technol
41: 3972-3976.  

 Pandis, S.N., Harley, R.A., Cass, G.R., Seinfeld, J.H. (1992) Secondary
organic aerosol formation and transport. Atmos Environ 26, 2269–2282. 

 Takekawa, H. Minoura, H. Yamazaki, S. (2003) Temperature dependence of
secondary organic aerosol formation by photo-oxidation of hydrocarbons.
Atmos Environ 37: 3413–3424. 

 Kleeman, M.J., Ying, Q., Lu, J., Mysliwiec, M.J., Griffin, R.J., Chen,
J., Clegg, S. (2007) Source apportionment of secondary organic aerosol
during a severe photochemical smog episode. Atmos Environ 41: 576–591.
 

 Robinson, A. L.; Donahue, N. M.; Shrivastava, M.; Weitkamp, E. A.;
Sage, A. M.; Grieshop, A. P.; Lane, T. E.; Pierce, J. R.; Pandis, S. N.
(2007) Rethinking organic aerosol: Semivolatile emissions and
photochemical aging. Science 315: 1259-1262. 

 Griffin, R. J.; Cocker, D. R.; Seinfeld, J. H.; Dabdub, D. (1999)
Estimate of global atmospheric organic aerosol from oxidation of
biogenic hydrocarbons. Geophys Res Lett 26 ( 17) 2721– 2724

 Lewis, C. W.; Klouda, G. A.; Ellenson, W. D. (2004) Radiocarbon
measurement of the biogenic contribution to summertime PM-2.5 ambient
aerosol in Nashville, TN. Atmos Environ 38 ( 35) 6053– 6061.

 Byun DW, Schere, KL (2006) Review of the Governing Equations,
Computational Algorithms, and Other Components of the Models-3 Community
Multiscale Air Quality (CMAQ) Modeling System, J Applied Mechanics
Reviews 59: 51-76. 

 U. S. EPA (2010) Our Nations Air, Status and Trends through 2008. EPA
454/R-09-002, February 2010. http://www.epa.gov/airtrends/2010.  

 U. S. EPA. (2011) 2005 National-Scale Air Toxics Assessment. 
http://www.epa.gov/ttn/atw/nata2005/risksum.html.  

 U. S. EPA (2007) Regulatory Impact Analysis for the Control of
Hazardous Air Pollutants from Mobile Sources Rule, Chapter 3, Air
Quality and Resulting Health and Welfare Effects of Air Pollution from
Mobile Sources. 72 FR 8428, February 26, 2007.
http://www.epa.gov/otaq/regs/toxics/420r07002.pdf.  

 Lim, Y.B., Ziemann, P.J. (2009) Effects of Molecular Structure on
Aerosol Yields from OH Radical-Initiated Reactions of Linear, Branched,
and Cyclic Alkanes in the Presence of NOX. Environ Sci Technol 43 (7):
2328-2334. 

 Kleindienst, T.E. (2008) Hypothetical SOA Production from Ethanol
Photooxidation.  Memo to the Docket EPA-HQ-OAR-2005-0161. 

 Turpin, B.J., Huntzicker, J.J., Larson, S.M., Cass, G.R. (1991) Los
Angeles Summer Midday Particulate Carbon: Primary and Secondary Aerosol.
Environ Sci Technol 25: 1788-1793. 

 Turpin, B.J., Huntzicker, J.J. (1995) Identification of Secondary
Organic Aerosol Episodes and Quantitation of Primary and Secondary
Organic Aerosol Concentrations During SCAQS. Atmos Environ 29(23):
3527-3544. 

 Bae M-S, Schauer JJ, Turner JR (2006) Estimation of the Monthly Average
Ratios of Organic Mass to Organic Carbon for Fine Particulate Matter at
an Urban Site, Aerosol Sci Technol 40(12): 1123-1139.
http://dx.doi.org/10.1080/02786820601004085.  

 Kleindienst TE, M Jaoui, M Lewandowski, JH Offenberg, EO Edney (2007)
Estimates of the contributions of biogenic and anthropogenic
hydrocarbons to secondary organic aerosol at a southeastern U.S.
location. Atmos Environ 41(37):8288-8300.  

 CW Lewis, M Lewandowski, M Jaoui, TE Kleindienst, EO Edney (2007)
Contributions of Toluene and -pinene to SOA Formed in an Irradiated
Toluene/-pinene,NOx/Air Mixture: Comparison of Results Using 14C
Content and SOA Organic Tracer Methods, Environ Sci Technol 41:
3972-3976.  

idation of -pinene. Environ Sci Technol 41(5): 1628-1634. 



"

,

]

_

†

Õ

×

ä

õ

฀

 

-

?

G

Y

o

p

’

”

h

h

 h

  h

 h

 h

hj

hj

hj

hj

h°r

)

)

)

O

h'

O

O

O

O

h'

 hãX

j

 hãX

”ÿ÷p

”ÿ÷p

”ÿ÷p

”ÿ÷p

”ÿ÷p

”ÿ÷p

摧ᚙ´ЀG摧厡

h

h

 h

#

+

W

X

h8

h8

¶

¶

¶

¶

¶

¶

ÿ

ÿ

ÿ

ÿ

ÿ

ÿ

h¿

h¿S

h8

h£

h¿S

hÀ

hÀ

hÀ

hÀ

hÀ

hÀ

þ

h/"

hÀ

hÀ

hÀ

	ƒ

	ƒ

	ƒ

	ƒ

3

3

3

3

3

3

3

3

3

3

3

똇

3

3

3

3

3

3

3

3

3

3

܀Ⲕࠁ䛖̀ﾔҟᏃ⋧€f

 h

3

&

 h

 h

 h

 h

F

G

}

‡

‹

Œ

1

2

Ç

È

ȕ脈樃㡃

 h

H* h

&

ȕ脈樃㤽

ȕ脈樃㣀

 h

 h

 h

 h

 h

H* h

 h

 h

 h

 h

  h

 h

 h

 h

 h

 h

 h

 h

 h

H* h

 h

 h

  h

 h

 h

 h

 h

 h

H* h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

lð h

 h

 h

 h

 h

 h

 h

 h

 h

 h

  h

H* h

 h

 h

  h

 h

 h

 h

 h

 h

  h

 h

 h

 h

 h

 h

 h

 h

  h

 h

 h

 h

H*  h

 h

 h

 h

 h

 h

 h

 h

 h

&

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

lð h

 h

 h

 h

 h

 h

ȕ脈樃㦺

 h

 h

 h

 hD

 h

摧䢦æԈ਀Ц䘋

 h

 h

 h

  h

 h

  h

H

	H

	H

	H

	H

	H

	H

	H

	H

	H

	H

	H

	H

 h

 h

摧䝲ú

 h

 

¡

 h

 h

 h

 h

h

 h

h

 h

H* h

 

 

+

,

ᔌꅨꡭᘀ써坶✀,

-

.

/

0

J

K

M

N

O

a

e

f

m

o

r

“

¡

¨

©

´

¸

ï

ð

ó

ô

÷

ø

'¡

ª

Â

Ò

÷

ü

*

+

4

7

B

C

E

G

H

K

L

O

V

W

š

Ä

Å

B

E

P

V

②朂⡤漶洀ĤȀÅ

 h

  h

 h

  h

 h

  h

 h

 h

7þ

7þ

7þ

7þ

7þ

7þ

7þ

7þ

7þ

7þ

7þ

7þ

7þ

7þ

 h

 h

H* h

h

 h

h

 h

 h

 h

 h

 h

 h

 h

 h

 h

ᘀ﹨靻唀Ĉč脈䠄栖凸

 h

 h

H* h

 h

 h

 h

 h

H* h

 h~	

  h

 h

 h

  h

 h

 h

  h

H* h

 h

 h

 h

 h

 h

h

 h

h

 h

 h

  h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

 h

  h

 h

 h

 h

 h

昀Ĵ瑹ᢰǹ

昀Ĵ瑹ᢰǹ

 h

híg

hÿ

hÀ

hÀ

 hó

 hó

 hó

!

#

$

%

’

•

Ú

Û

Ü

Q

q

 h´

 hH'

倆㄀$摧篾ЀP摧篾

 hª

iated isoprene/NOX/SO2/air mixtures and their detection in ambient PM2.5
samples collected in the Eastern United States. Atmos Environ 39:
5281-5289.  

 Jaoui M, TE Kleindienst, M Lewandowski, JH Offenberg, EO Edney (2005)
Identification and quantification of aerosol polar oxygenated compounds
bearing carboxylic or hydroxyl groups. 2. Organic tracer compounds from
monoterpenes. Environ Sci Technol 39: 5661-5673. 

 Kleindienst TE, TS Conver, CD McIver, EO Edney (2004) Determination of
secondary organic aerosol products from the photooxidation of toluene
and their implications in ambient PM2.5. J Atmos  Chem 47: 70-100. 

 Kleindienst TE, TS Conver, CD McIver, EO Edney (2004) Determination of
secondary organic aerosol products from the photooxidation of toluene
and their implication in ambient PM2.5,  J Atmos Chem 47: 70-100.  

 Lewandowski M, M Jaoui , JH Offenberg , TE Kleindienst, EO Edney, RJ
Sheesley, JJ Schauer (2008)  Primary and secondary contributions to
ambient PM in the midwestern United States, Environ Sci Technol
42(9):3303-3309.
http://pubs.acs.org/cgi-bin/article.cgi/esthag/2008/42/i09/html/es072041
2.html.  

 Kleindienst TE, M Jaoui, M Lewandowski, JH Offenberg, EO Edney (2007)
Estimates of the contributions of biogenic and anthropogenic
hydrocarbons to secondary organic aerosol at a southeastern U.S.
location. Atmos Environ 41(37):8288-8300.  

 Henze DK, JH Seinfeld (2006) Global secondary organic aerosol from
isoprene oxidation. Geophys Res Lett 33: L09812.
doi:10.1029/2006GL025976.  

 Hildebrandt, L., Donahue1, N. M, Pandis1, S. N. (2009) High formation
of secondary organic aerosol from the photo-oxidation of toluene.  Atmos
Chem Phys 9: 2973-2986. Docket EPA-HQ-OAR-2011-0135 9. 

 Ng, N. L., Kroll, J. H., Chan, A. W. H., Chabra, P. S., Flagan, R. C.,
Seinfield, J. H., Secondary organic aerosol formation from m-xylene,
toluene, and benzene,  Atmospheric Chemistry and Physics Discussion, 7,
3909-3922, 2007.  Docket EPA-HQ-OAR-2011-0135

 Henze, D. K., Seinfeld, J. H., Ng, N. L., Kroll, J. H.,
Fu, T.-M., Jacob, D. J., and Heald, C. L. (2008) Global modeling of
secondary organic aerosol formation from aromatic hydrocarbons: high-
vs. low-yield pathways, Atmos. Chem. Phys., 8, 2405-2420,
doi:10.5194/acp-8-2405-2008

 Lane, T. E., Donahue, N.M. and Pandis, S.N. (2008) Simulating secondary
organic aerosol formation using the volatility basis-set approach in a
chemical transport model, Atmos. Environ., 42, 7439-7451, doi:
10.1016/j.atmosenv.2008.06.026

 Carlton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.O., Sarwar, g.,
Pinder, R.W., Pouliot, G.A., Houyoux, M., (2010). Model Representation
of Secondary Organic Aerosol in CMAQv4.7. Environ Sci Technol 44(22),
8553-8560.

 Parikh, H.M., Carlton, A.G., Vizuete, W., and Kamen, R.M. (2011)
Modeling secondary organic aerosol using a dynamic partitioning approach
incorporating particle aqueous-phase chemistry, Atmospheric Environment,
45, 1126-1137.

 Volkamer, R., J.L. Jimenez, F. SanMartini,K.Dzepina,Q. Zhang,D.
Salcedo,L. T. Molina, D. R.Worsnop, and M. J. Molina (2006), Secondary
organic aerosol formation from anthropogenic air pollution: Rapid and
higher than expected, Geophys. Res. Lett., 33, L17811,
doi:10.1029/2006GL026899.

 Carlton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.O., Sarwar, g.,
Pinder, R.W., Pouliot, G.A., Houyoux, M., (2010). Model Representation
of Secondary Organic Aerosol in CMAQv4.7. Environ Sci Technol 44(22),
8553-8560.

 Robinson, A. L.; Donahue, N. M.; Shrivastava, M.; Weitkamp, E. A.;
Sage, A. M.; Grieshop, A. P.; Lane, T. E.; Pierce, J. R.; Pandis, S. N.
(2007) Rethinking organic aerosol: Semivolatile emissions and
photochemical aging. Science 315: 1259-1262. Docket EPA-HQ-OAR-2011-0135

 Carlton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.O., Sarwar, g.,
Pinder, R.W., Pouliot, G.A., Houyoux, M., (2010). Model Representation
of Secondary Organic Aerosol in CMAQv4.7. Environ Sci Technol 44(22),
8553-8560.

 Atkinson, R., Baulch, D.L., Cox, R.A., Crowley, J.N., Hampson, R.F.
Jr., Hynes, R.G., Jenkin, M.E., Kerr, J.A., Rossi, M.J., Troe, J. (2005)
Evaluated Kinetic and Photochemical Data for Atmospheric Chemistry -
IUPAC Subcommittee on Gas Kinetic Data Evaluation for Atmospheric
Chemistry. July 2005 web version.
http://www.iupac-kinetic.ch.cam.ac.uk/index.html. Docket
EPA-HQ-OAR-2011-0135

 Sander, S.P., Friedl, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E.,
Orkin, V.L., Moortgat, G.K., Ravishankara, A.R., Kolb, C.E., Molina,
M.J., Finlayson-Pitts, B.J. (2003) Chemical Kinetics and Photochemical
Data for use in Atmospheric Studies, Evaluation Number 14. NASA Jet
Propulsion Laboratory. http://jpldataeval.jpl.nasa.gov/index.html.
Docket EPA-HQ-OAR-2011-0135

 Finlayson-Pitts BJ, Pitts JN Jr. (1986) Atmospheric Chemistry:
Fundamentals and Experimental Techniques, Wiley, New York.

 Yarwood G, Rao S, Yocke M, Whitten GZ (2005) Updates to the Carbon Bond
Chemical Mechanism: CB05. Final Report to the U.S. EPA, RT-0400675,
December 8, 2005.
http://www.camx.com/publ/pdfs/CB05_Final_Report_120805.pdf.  Docket
EPA-HQ-OAR-2011-0135

 http://www.cmascenter.org/help/model_docs/cmaq/4.7/RELEASE_NOTES.txt

 77FR 30088 (May 21, 2012)

 US EPA (2012).  National Ambient Air Quality Standards for Particulate
Matter.  http://www.epa.gov/PM/2012/finalrule.pdf

 US EPA (2012).  Fact Sheet: Implementing the Standards.
http://www.epa.gov/airquality/particlepollution/2012/decfsimp.pdf

 U.S. EPA. (2007). PM2.5 National Ambient Air Quality Standard
Implementation Rule (Final).  Washington, DC: U.S. EPA. 72 FR 20586,
April 25, 2007.

 PM Standards Revision – 2006: Timeline. 
http://www.epa.gov/PM/naaqsrev2006.html#timeline 

 U.S. EPA. (2012).  Fact Sheet – Air Quality Designations for the 2010
Primary Nitrogen Dioxide (NO2) National Ambient Air Quality Standards. 
http://www.epa.gov/airquality/nitrogenoxides/designations/pdfs/20120120F
S.pdf

  U.S. Environmental Protection Agency (2012).  Revision to Ambient
Nitrogen Dioxide Monitoring Requirements.  77 FR 64244, October 19,
2012.

 U. S. Environmental Protection Agency (2007). Control of Hazardous Air
Pollutants from Mobile Sources; Final Rule.  72 FR 8434, February 26,
2007. 

 U. S. Environmental Protection Agency (2007). Control of Hazardous Air
Pollutants from Mobile Sources; Final Rule.  72 FR 8434, February 26,
2007. 

 U.S. EPA. (2011) 2005 National-Scale Air Toxics Assessment.   HYPERLINK
"http://www.epa.gov/ttn/atw/n" http://www.epa.gov/ttn/atw/n ata2005/. 
Docket EPA-HQ-OAR-2011-0135

 U.S. EPA. (2011) Summary of Results for the 2005 National-Scale
Assessment.  http://www.epa.gov/ttn/atw/nata2005/05pdf/sum_results.pdf. 

 Luecken, D,J, Clmorel, A.J. 2008. Codependencies of Reactive Air Toxic
and Criteria Pollutants on Emission Reductions. J. Air & Waste Manage.
Assoc. 58:693–701. DOI:10.3155/1047-3289.58.5.693

 U.S. Environmental Protection Agency (EPA). 2009. Integrated Science
Assessment for Particulate Matter. U.S. Environmental Protection Agency.
Research Triangle Park. EPA/600/R-08/139F

 Trijonis, J.C. et al. 1987. Preliminary extinction budget results from
the RESOLVE program, pp. 872-883. In: P.J. Bhardwaja, et. al. Visibility
Protection Research and Policy Aspects.  Air Pollution Control Assoc.,
Pittsburgh, PA.

 Trijonis, J.C. et al. 1988. RESOLVE Project Final Report: Visibility
conditions and Causes of Visibility Degradation in the Mojave Desert of
California. NWC TP #6869. Naval Weapons Center, China Lake, CA.

 Irving, Patricia M., e.d., 1991. Acid Deposition: State of Science and
Technology, Volume III, Terrestrial, Materials, Health, and Visibility
Effects, The U.S. National Acid Precipitation Assessment Program,
Chapter 24, page 24–76.

 Huai, et al. (2004) Estimates of the emission rates of nitrous oxide
from light-duty vehicles using different chassis dynamometer test cycles
 Atmospheric Environment 6621-6629 

 Michaels, H. (1998) Emissions of Nitrous Oxide from Highway Mobile
Sources, U.S. EPA EPA420-R-98-009

 Graham, L. Greenhouse Gas Emissions from 1997-2005 Model Year Light
Duty Vehicles Environment Canada ERMD Report #04-44 

 Meffert, et. al (2000) Analysis of Nitrous Oxide Emissions from Light
Duty Passenger Cars, SAE 2000-01-1952

 Behrentz, et al. (2004), Measurements of nitrous oxide emissions from
light-duty motor vehicles: a pilot study  Atmospheric Environment
4291-4303 

 Meszler, D. (2004), Light Duty Vehicle Methane and Nitrous Oxide
Emissions: Greenhouse Gas Impacts Study for Northeast States Center for
a Clean Air Future 

